google ads
Search

Everything you need to know about Computed Tomography (CT) & CT Scanning

May 2022 Imaging Pearls - Educational Tools | CT Scanning | CT Imaging | CT Scan Protocols - CTisus
Imaging Pearls ❯ May 2022

-- OR --

Adrenal

  • “The 2017 American College of Radiology White Paper was the most used guideline, yet the management of indeterminate adrenal incidentalomas was highly variable with no single management option reaching a majority. Hormonal evaluation and endocrinology consultation was most often rarely or never recommended. The results of the survey indicate wide variability in the interpretation of imaging findings and management recommendations for incidental adrenal nodules among surveyed radiologists. Further standardization of adrenal incidentaloma guidelines and education of radiologists is needed.”
    Management of incidental adrenal nodules: a survey of abdominal radiologists conducted by the Society of Abdominal Radiology Disease‐Focused Panel on Adrenal Neoplasms  
    Michael T. Corwin et al.
    Abdominal Radiology (2022) 47:1360–1368 

  • Management of incidental adrenal nodules: a survey of abdominal radiologists conducted by the Society of Abdominal Radiology Disease‐Focused Panel on Adrenal Neoplasms  
    Michael T. Corwin et al.
    Abdominal Radiology (2022) 47:1360–1368 
  • “The majority of respondents either rarely or never rec- ommend hormonal evaluation or endocrinology consultation when describing an adrenal incidentaloma. Both the American Association of Clinical Endocrinologists and the European Society of Endocrinology recommend hormonal evaluation to determine the functional activity in all patients with adrenal incidentalomas. The 2017 ACR white paper advises consideration for biochemical evaluation for most incidentalomas as adrenal hyperfunction may not be clinically evident."
    Management of incidental adrenal nodules: a survey of abdominal radiologists conducted by the Society of Abdominal Radiology Disease‐Focused Panel on Adrenal Neoplasms  
    Michael T. Corwin et al.
    Abdominal Radiology (2022) 47:1360–1368 
  • Causes of Spontaneous Bleeds: Adrenal
    - Adenoma
    - Neuroblastoma
    - Primary adrenal carcinoma
    - Pheochromocytoma
    - Metastases (RCC)
Chest

  • Objectives: To evaluate and compare the diagnostic performances of a commercialized artificial intelligence (AI) algorithm for diagnosing pulmonary embolism (PE) on CT pulmonary angiogram (CTPA) with those of emergency radiologists in routine clinical practice.
    Methods: This was an IRB-approved retrospective multicentric study including patients with suspected PE from September to December 2019 (i.e., during a preliminary evaluation period of an approved AI algorithm). CTPA quality and conclusions by emergency radiologists were retrieved from radiological reports. The gold standard was a retrospective review of CTPA, radiological and clinical reports, AI outputs, and patient outcomes. Diagnostic performance metrics for AI and radiologists were assessed in the entire cohort and depending on CTPA quality.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2 
  • Results: Overall, 1202 patients were included (median age: 66.2 years). PE prevalence was 15.8% (190/1202). The AI algorithm detected 219 suspicious PEs, of which 176 were true PEs, including 19 true PEs missed by radiologists. In the cohort, the highest sensitivity and negative predictive values (NPVs) were obtained with AI (92.6% versus 90% and 98.6% versus 98.1%, respectively), while the highest specificity and positive predictive value (PPV) were found with radiologists (99.1% versus 95.8% and 95% versus 80.4%, respectively). Accuracy, specificity, and PPV were significantly higher for radiologists except in subcohorts with poor-to-average injection quality. Radiologists positively evaluated the AI algorithm to improve their diagnostic comfort (55/79 [69.6%]).  
    Conclusion: Instead of replacing radiologists, AI for PE detection appears to be a safety net in emergency radiology practice due to high sensitivity and NPV, thereby increasing the self-confidence of radiologists.
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022 https://doi.org/10.1007/s00330-022-08645-2 
  • Key Points
    • Both the AI algorithm and emergency radiologists showed excellent performance in diagnosing PE on CTPA (sensitivity and specificity ≥ 90%; accuracy ≥ 95%).
    • The AI algorithm for PE detection can help increase the sensitivity and NPV of emergency radiologists in clinical practice, especially in cases of poor-to-moderate injection quality.
    • Emergency radiologists recommended the use of AI for PE detection in satisfaction surveys to increase their confidence and comfort in their final diagnosis.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  • “Indeed, in the entire cohort-2019, AIDOC captured 19 PEs that were not diagnosed by radiologists in 19 distinct patients. In other words, the AI algorithm could correct a misdiagnosed PE approximately every 63 CTPAs (≈1202/19). This estimation must be considered in parallel with the high number of CTPAs required by emergency physicians (≈18,000 CTPAs in 2020 in our group—so approximately 285 [≈18000/1202 × 19] true PEs detected by AI but initially misdiagnosed by radiologists in 2020) and with human and financial consequences of missed PEs [32]. Indeed, mortality and recurrence rates for untreated or missed PE range between 5 and 30%.”  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  •  “In conclusion, this study confirms the high diagnostic performances of AI algorithms relying on DCNN to diagnose PE on CTPA in a large multicentric retrospective emergency series. It also underscores where and how AI algorithms could better support (or “augment”) radiologists, i.e., for poor- quality examinations and by increasing their diagnostic con- fidence through the high sensitivity and high NPV of AI. Thus, our work provides more scientific ground for the concept of “AI-augmented” radiologists instead of supporting the theory of radiologists’ replacement by AI.”
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
Colon

  • “In addition to surgery, chemotherapy is a cornerstone in cancer treatment. Chemotherapeutic agents are rapidly evolving as both cytotoxic agents and newer molecular-targeted therapies with the aim of inhibiting cancer growth and proliferation. However, these drugs can cause toxicities or complications in the bowel as its high mitotic rate makes it particularly sensitive to the toxicity of chemotherapy. The spectrum of bowel complications can range from mild or no symptoms to urgent or emergent.”
    Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    Sitthipong Srisajjakul, Patcharin Prapaisilp, Sirikan Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  
  • "Enteritis may occur after exposure to chemotherapies. It most commonly associated with fluorouracil and leucovorin, mTOR inhibitors, immune modulators, and tyrosine kinase inhibitors. The segment of the small bowel that is most affected is the ileum, but the entire small bowel may be involved. Diarrhea is the clinical clue.”
    Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    Sitthipong Srisajjakul, Patcharin Prapaisilp, Sirikan Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310   
  • “Neutropenic colitis (typhlitis) is an acute inflammation of the cecum associated with neutropenia and is highly fatal. The ascending colon and terminal ileum may be involved. These conditions are typically identified in patients with leukemia or lymphoma who are receiving chemotherapy, or in patients who are undergoing stem cell transplantation. Neutropenia, intestinal mucosal injury, and ischemia with a superimposed infection have been postulated to be the mechanisms.”
    Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    Sitthipong Srisajjakul, Patcharin Prapaisilp, Sirikan Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  
  • "Fulminant colitis (toxic megacolon) is a life-threatening condition that can be seen in patients with severe colitis. It has been reported in patients with solid tumors treated with cyclophosphamide, fluorouracil, and epirubicin. However, it is a well-recognized consequence of inflammatory bowel disease, ischemic colitis, and infectious colitis. Inflammation that extends deep into the mucosa and damages the muscularis propria will result in colonic dilatation and absent colonic haustration.”
    Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    Sitthipong Srisajjakul, Patcharin Prapaisilp, Sirikan Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  
  • "As a result of chemotherapy, cancer patients are sus- ceptible to immunosuppression. This can cause lymphoid depletion and mucosal disruption in the bowel wall. In turn, normal intraluminal gas is able to pass into the bowel wall without frank evidence of bowel ischemia (benign, or non-life-threatening, PI). Lymphoid tissue and Peyer’s patches are rich in the distal ileum and colon, and their depletion has been postulated to be the cause of benign PI. Common chemotherapeutic agents associated with PI are tyrosine kinase inhibitors such as imatinib and suni- tinib. They can damage the intestinal mucosa, impair regenerative ability, and disrupt normal motility.”
    Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    Sitthipong Srisajjakul, Patcharin Prapaisilp, Sirikan Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  

  • Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    S Srisajjakul, P Prapaisilp, S Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  

  • Drug‐induced bowel complications and toxicities: imaging findings and pearls  
    S Srisajjakul, P Prapaisilp, S Bangchokdee  
    Abdominal Radiology (2022) 47:1298–1310  
Deep Learning

  • “Even in geographies that have a reputation for high quality care (such as metropolitan Boston and metropolitan New York) there is a five times greater chance of death from acute myocardial infarction (heart attack), depending on the hospital one chooses. Across the United States, on average, patients are twice as likely to die in the lowest-performing hospitals. This includes a 2.3-fold difference in heart attack mortalities. There are even greater differences in safety. The top 10% of hospitals are 10 times safer than bottom 10%. Patients are 18 times more likely to suffer a bloodstream infection from a central venous catheter when treated at poor-performing hospitals.”  
    4 Actions to Reduce Medical Errors in U.S> Hospitals
    Troussaint JS, Segel KT
    Harvard Business Review April 20, 2022
  • 1. Make patient and staff safety a top priority
    2.Establish a national  safety organization
    3. create a national reporting mechanism
    4. turn on EHR’s machine leaning symptoms
    4 Actions to Reduce Medical Errors in U.S> Hospitals
    Troussaint JS, Segel KT
    Harvard Business Review April 20, 2022
  • “The use of artificial intelligence (AI) has grown dramatically in the past few years in the United States and worldwide, with more than 300 AI-enabled devices approved by the U.S. Food and Drug Administration (FDA). Most of these AI-enabled applications focus on helping radiologists with detection, triage, and prioritization of tasks by using data from a single point, but clinical practice often encompasses a dynamic scenario wherein physicians make decisions on the basis of longitudinal information. Unfortunately, benchmark data sets incorporating clinical and radiologic data from several points are scarce, and, therefore, the machine learning community has not focused on developing methods and architectures suitable for these tasks. Current AI algorithms are not suited to tackle key image interpretation tasks that require comparisons to previous examinations. Focusing on the curation of data sets and algorithm development that allow for comparisons at different points will be required to advance the range of relevant tasks covered by future AI-enabled FDA-cleared devices.”
    The Need for Medical Artificial Intelligence That Incorporates Prior Images  
    Julián N. Acosta et al.  
    Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 
  • “Most of these AI-enabled applications focus on helping radiologists with detection, triage, and prioritization of tasks by using data from a single point, but clinical practice often encompasses a dynamic scenario wherein physicians make decisions on the basis of longitudinal information. Unfortunately, benchmark data sets incorporating clinical and radiologic data from several points are scarce, and, therefore, the machine learning community has not focused on developing methods and architectures suitable for these tasks. Current AI algorithms are not suited to tackle key image interpretation tasks that require comparisons to previous examinations.”
    The Need for Medical Artificial Intelligence That Incorporates Prior Images  
    Julián N. Acosta et al.  
    Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 
  • Essentials  
    • Most artificial intelligence (AI)–enabled devices that are approved by the U.S. Food and Drug Administration (FDA) and are avail- able to date address tasks by considering only a single point.  
    • Clinical tasks involve dynamic scenarios, and diagnostic and prognostic decisions often rely on the combination of prior and current information.  
    • The development of benchmark data sets and algorithms that leverage prior examinations have the potential to improve the range of tasks covered by FDA-approved medical AI devices.  
    The Need for Medical Artificial Intelligence That Incorporates Prior Images  
    Julián N. Acosta et al.  
    Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830 
  • “In summary, even though physicians routinely perform comparisons with prior examinations when interpreting images in clinical practice, only a few artificial intelligence (AI) algorithms currently available are able to incorporate information from more than one point to help in these critical tasks. The curation of high-quality data sets with longitudinal clinical and imaging data, and the development of AI algorithms capable of solving a wider range of problems, will be essential to provide meaningful improvements in clinical workflows.”
    The Need for Medical Artificial Intelligence That Incorporates Prior Images  
    Julián N. Acosta et al.  
    Radiology 2022; 000:1–6 • https://doi.org/10.1148/radiol.212830  
  • “For early cancer detection, the percentage of confidence that comes with a diagnostic decision made by the algorithm might appear straightforward (for example, 80% confidence that lung cancer is present), but the process behind this number is very complex and understandably may not be apparent to the user. It is therefore not difficult to understand that there might be resistance to adoption of such strategies and the fear of overdiagnosis. It is important to understand that AI will not remove the need for physicians and experts to interpret the findings, provide a global picture of patient health, spot related diseases, and come up with a final diagnosis.”
    The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 
  • "What the introduction of AI algorithms might do, providing that data management and safety regulations are in place, is reduce the cost and time needed to diagnose the disease. This will enable health practitioners to spend more time developing efficient and holistic treatment protocols, and will make state-of-art diagnostics more affordable. Furthermore, AI can be a training tool that provides immediate specialist feedback to generalists so that, in time, they may perform at an expert level.”
    The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 
  • “Medical diagnostic imaging studies frequently detect findings that require further evaluation. An initiative at Northwestern Medicine was designed to prevent delays and improve outcomes by engineering reliable follow-up of radiographic findings. An artificial intelligence natural language processing (NLP) system was developed to identify radiology reports containing lung- and adrenal-related findings requiring follow-up. Over 13 months, more than 570,000 imaging studies were screened, of which more than 29,000 were flagged as containing lung-related follow-up recommendations, representing a 5.1% rate of lung-related findings occurrence on relevant imaging studies and an average of 70 findings flagged per day. Northwestern’s prospective clinical validation of the system, the first of its kind, demonstrated a sensitivity of 77.1%, specificity of 99.5%, and positive predictive value of 90.3% for lung findings requiring follow-up. To date, the workflow has generated nearly 5,000 interactions with ordering physicians and has tracked more than 2,400 follow-ups to completion. The authors conclude that NLP demonstrates significant potential to improve reliable follow-up to imaging findings and, thus, to reduce preventable morbidity in lung pathology and other high-risk and problem-prone areas of medicine.”  
    Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 
  •  “An artificial intelligence natural language processing (NLP) system was developed to identify radiology reports containing lung- and adrenal-related findings requiring follow-up. Over 13 months, more than 570,000 imaging studies were screened, of which more than 29,000 were flagged as containing lung-related follow-up recommendations, representing a 5.1% rate of lung-related findings occurrence on relevant imaging studies and an average of 70 findings flagged per day. Northwestern’s prospective clinical validation of the system, the first of its kind, demonstrated a sensitivity of 77.1%, specificity of 99.5%, and positive predictive value of 90.3% for lung findings requiring follow-up. To date, the workflow has generated nearly 5,000 interactions with ordering physicians and has tracked more than 2,400 follow-ups to completion. The authors conclude that NLP demonstrates significant potential to improve reliable follow-up to imaging findings and, thus, to reduce preventable morbidity in lung pathology and other high-risk and problem-prone areas of medicine.”  
    Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 
  • "Considering these clinical needs, artificial intelligence (AI) is well suited to the detection and reporting of follow-up recommendations because of the large volume of imaging studies requiring screening and the relatively standardized language employed by radiologists in preparing reports. Natural language processing (NLP) methods, including text pattern-matching and traditional machine-learning techniques, have been developed for this task. In this article, we use the term traditional machine learning to refer to all machine-learning methods that are not deep learning, and these terms will be defined in detail in the sections that follow. More recently, novel deep-learning methods for NLP have shown great promise for the detection of follow-up recommendations However, methods reported to date are limited by the size of the data sets used for model training, as well as by a lack of prospective evaluation and implementation in clinical settings.”
    Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 

  • Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 

  • Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 
  • “We will continue to build and test new and improved NLP models and clinical workflow adjustments. Review of model misclassifications has allowed for identification of radiology reports that may be particularly difficult to classify and retraining on updated data sets collected as part of this effort will continually improve performance. The dedicated annotation system and EHR infrastructure in place facilitate streamlined model prototyping, evaluation, and deployment. Moreover, current efforts are underway to expand this system to hepatic, thyroid, and ovarian findings requiring follow-up. Finally, as the Result Management system continues to mature and the system tracks more follow-ups to completion, we aim to further characterize its impact on patient outcomes.”  
    Preventing Delayed and Missed Care by Applying Artificial Intelligence to Trigger Radiology Imaging Follow-up  
    Jane Domingo et al.
    NEJM Catalyst Vol. 3 No. 4 | April 2022 DOI: 10.1056/CAT.21.0469 
  • AI and Liability
    - Who is responsible for the accuracy of an AI system when it makes an error?
    - What is the liability of the Radiologist when using AI?
    - What is the liability of the health system that purchases an AI product?
  • “Developers of health care AI products face the risk of product liability lawsuits when their products injure patients, whether injuries arise from defective manufacturing, defective design, or failure to warn users about mitigable dangers.16 Physicians may also face risks from patient injuries stemming from the use of AI, including faulty recommendations or inadequate monitoring. Similarly, hospitals or health systems may face liability as coordinating providers of health care or on the basis of inadequate care in supplying AI tools — an analogy to familiar forms of medical liability for providing inadequate facilities or negligently credentialing a physician practicing at the hospital. Such risks may reduce incentives to adopt AI tools.”
    AI Insurance: How Liability Insurance Can Drive the Responsible Adoption of Artificial Intelligence in Health Care  
    Ariel Dora Stern et al.  
    NEJM Catalyst Vol. 3 No. 4 | April 2022  DOI: 10.1056/CAT.21.0242 
  • "AI liability insurance would reduce the liability risk to developers, physicians, and hospitals. Insurance is a tool for managing risk, allowing the insurance policy holders to benefit from pooling risk with others. Insurance providers are intermediaries that play an organizing role in creating these pools and performing actuarial assessment of associated risks. While many types of insurance exist in the health care context, our focus in this article is entirely on AI liability insurance rather than coverage for health care services.”
    AI Insurance: How Liability Insurance Can Drive the Responsible Adoption of Artificial Intelligence in Health Care  
    Ariel Dora Stern et al.  
    NEJM Catalyst Vol. 3 No. 4 | April 2022  DOI: 10.1056/CAT.21.0242 
  • "The credentialing function of insurance will thus reinforce the patient-centered incentives of AI developers Consequently, this insurance may alleviate health care provider concerns, at least to the point at which they are willing to adopt the AI technology. Indeed, this should be the case regardless of whether the AI manufacturer or the health care provider is the holder of the insurance policy, as long as such a policy can be purchased. However, the price and implicit value of insurance are likely to be passed through. For example, a manufacturer selling an AI tool that comes with liability insurance will be able to command a higher price than for the same tool without such insurance. Insurers may also require ongoing performance data from AI developers, whether they are in house or commercial; such data could be well beyond those needed to meet the requirements of regulatory premarket review.28 While insurers do not provide the same level of centralized review that regulators do, they may well serve a more context-sensitive, hands-on evaluative role focused on both quantifying and reducing risk — a role that may be especially important given the questionable generalizability of many current-generation AI systems.”
    AI Insurance: How Liability Insurance Can Drive the Responsible Adoption of Artificial Intelligence in Health Care  
    Ariel Dora Stern et al.  
    NEJM Catalyst Vol. 3 No. 4 | April 2022  DOI: 10.1056/CAT.21.0242 
  • Background: Radiomics and radiogenomics are two words that recur often in language of radiologists, nuclear doctors and medical physicists especially in oncology field. Radiomics is the technique of medical images analysis to extract quantitative data that are not detected by human eye.
    Results: Several studies showed that radiomics is very promising. However, there were some critical issues: poor standardization and generalization of radiomics results, data-quality control, repeatability, reproducibility, database balancing and issues related to model overfitting.
    Conclusions: Radiomics procedure should made considered all pitfalls and challenges to obtain robust and reproducible results that could be generalized in other patients cohort.  
    Radiomics in medical imaging: pitfalls and challenges in clinical management  
    Roberta Fusco et al.
    Japanese Journal of Radiology https://doi.org/10.1007/s11604-022-01271-4 

  • Radiomics in medical imaging: pitfalls and challenges in clinical management  
    Roberta Fusco et al.
    Japanese Journal of Radiology https://doi.org/10.1007/s11604-022-01271-4 
  • "The features can be morphological, of first, second and higher statistical orders. The morphological metrics describe the shape of segmented volume of interest and its geometric characteristics such as volume, largest diameter in different orthogonal directions, surface, compactness and sphericity. First-order statistics features describe the distribution of individual voxel values such as mean, median, maximum, minimum values, skewness (asymmetry), kurtosis (flatness), uniformity, and entropy. Second-order statistics features are obtained calculating the statistical inter-relationships between neighboring voxels providing a measure of spatial arrangement and of intra-lesion heterogeneity.”  
    Radiomics in medical imaging: pitfalls and challenges in clinical management  
    Roberta Fusco et al.
    Japanese Journal of Radiology https://doi.org/10.1007/s11604-022-01271-4 
  • "Radiomics is the technique of medical images analysis to extract quantitative data that are not detected by human eye. In practice, visual analysis manages to extract only about 10% of the information contained in a digital medical image. If, on the other hand, these images are analyzed in detail by powerful computers through complex mathematical algorithms, it is possible to obtain objective quantitative data, capable of providing information on the underlying patho-physiological phenomena, inaccessible to simple visual analysis. Therefore, in the field of medicine, radiomics is a method that extracts a large number of features from medical images using data-characterization algorithms.”
    Radiomics in medical imaging: pitfalls and challenges in clinical management  
    Roberta Fusco et al.
    Japanese Journal of Radiology https://doi.org/10.1007/s11604-022-01271-4 
  • "Although several studies showed as Radiomic is very promising, there has been poor standardization and generalization of radiomics results, which limit the translation of this method into clinical setting. Clear limitations of this field are emerging, especially with regard to data-quality control, repeatability, reproducibility, generalizability of results, and issues related to model overfitting.”
    Radiomics in medical imaging: pitfalls and challenges in clinical management  
    Roberta Fusco et al.
    Japanese Journal of Radiology https://doi.org/10.1007/s11604-022-01271-4   
  • “Moreover, patients must be properly informed about the relevant concepts. Many patients are unfamiliar with the concept of over-diagnosis and therefore may be unable to weigh the relative risk of unnecessary diagnosis and treatment against the risk failing to discover a cancer. Moreover, patients may not always have preferences about such outcomes. There must still be a sensible default decision threshold that can be used in cases in which patients choose to withhold their attitudes or simply have no preferences.”
    Clinical decisions using AI must consider patient values  
    Jonathan Birch, Kathleen A. Creel, Abhinav K.  
    Nature Medicine | VOL 28 | Feb 2022 | 226–235 
  • “A risk-profiling questionnaire suitable for cancer screening would probe the patient’s attitudes about the risk of over-diagnosis, false-positive and false-negative results, and over-treatment versus under-treatment, and the expected value to the patient of additional years of life of varying quality levels. The questionnaire might also ask patients to respond to statements such as ‘I would rather risk surgical complications to treat a benign tumor than risk missing a cancerous tumor’.”
    Clinical decisions using AI must consider patient values  
    Jonathan Birch, Kathleen A. Creel, Abhinav K.  
    Nature Medicine | VOL 28 | Feb 2022 | 226–235
  • “Relatively weak evidence supporting the use of AI in routine clinical practice health care settings, AI models continue to be marketed and deployed. A recent example is the Epic Sepsis Model. While this model was widely implemented in hundreds of US hospitals, a recent study showed that it performed significantly worse in correctly identifying patients with early sepsis and improving patient outcomes in a clinical setting compared with performance observed during development of the model.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence  
    Cornelius A. James,  et al.
    JAMA Published online March 21, 2022
  • “AI will soon become ubiquitous in health care. Building on lessons learned as implementation strategies continue to be devised, it will be essential to consider the key role of clinicians as end users of AI-developed algorithms, processes, and risk predictors. It is imperative that clinicians have the knowledge and skills to assess and determine the appropriate application of AI outputs, for their own clinical practice and for their patients. Rather than being replaced by AI, these new technologies will create new roles and responsibilities for clinicians.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence  
    Cornelius A. James,  et al.
    JAMA Published online March 21, 2022
  • “One must ask, then, what is the burden of value a radiology AI product must provide to justify purchase? The answer depends on many factors including the health care setting and its purchasing structure, the health care payer system, and patient distribution, but a nearly universal thread is that the software must provide financial return on investment. The challenge is matching the “return” to the “investor.” When these two parties are mismatched, the cost justification for one group’s investment for another group’s benefit rarely occurs.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • “If an AI model can increase the throughput for ex- amination interpretation, the practice can now absorb this additional volume without an additional radiolo- gist hire. Similarly, models that triage low complexity or negative studies can be used to route these exami- nations to physician extenders and reduce costs for a group by over 75% while maintaining imaging revenue. Conversely, AI models that close the loop on patient follow-up or detect incidental findings have no financial benefit for a private practice and therefore are less likely to be paid for by the radiology group.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • "Similarly, driving additional outpatient referrals into the radiology department (ie, additional examinations) can also be significant revenue generators. For example, closing the loop on an incidental adrenal nodule can result in an additional triple-phase CT or MRI examinations and thousands of dollars in additional revenue while providing standard of care for the patient. Models that provide opportunistic screening such as scoring of coronary artery calcium on routine nongated chest CT can identify high-risk patients for cardiology referral, some of whom may ultimately receive advanced interventions. Capture or retention of a patient into the health system provides significant revenue streams, and in each of these cases the financial incentive and champions for adoption of these models are outside of the radiology department. Ultimately, these models have little impact on radiology workflow or the finances of a radiology practice.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • "In the emergency setting in which community hospitals may not have 24-7 radiologist coverage, AI models can help increase patient throughput and lead to significant cost savings. Many emergency pro- viders must currently choose the lesser of two evils—have patients wait overnight for examination re- ports or independently interpret examinations to guide patient disposition. Deployment of AI models in these settings can increase confidence in discharging patients for negative examinations or help quickly flag emergent findings that require immediate intervention.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • "Nevertheless, at least two companies have recently secured additional CPT codes for use of AI software—one for detection of vertebral compression fracture on CT (www.zebramedical.com) and another for scoring trabecular bone health on bone densitometry examinations (www.nanox.vision) to improve risk stratification for osteopenic and osteoporotic patients. However, it is important to note that reimbursements for AI software may have unintended consequences on the reimbursement for radiology examination interpretation, particularly for cases in which AI software reduces the average interpretation time of the examination.”  
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • "Lastly, although the majority of care in the United States is based on fee-for-service, there are a few domestic (eg, Veterans Affairs Hospitals) and many more international examples of vertical payment models in which there are in- centives to improve quality as a means to reduce cost. In these set- tings, a win-win-win is possible for patients, payers, and physicians. For example, in a fee-for-service system, a model that reduces unnecessary biopsies in screening mammography is good for patients and payers but may face barriers to adoption because it decreases hospital and practice revenue. However, the same model is a single-payer system is a win-win-win: patients receive better care, physicians have decreased workload, and payers significantly reduce costs. For this reason, many AI companies have seen wider adoption in Europe and Asia as compared with the United States.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • “In summary, the path to adoption of radiology AI is complex but must be viewed through a realistic lens that considers the economic truths of the health care system. Software that improves quality alone without a secondary benefit in efficiency, referrals, or another revenue stream is difficult to justify in a fee-for-service model, but these same models are being actively adopted within single-payer systems domestically and internationally. Commercial radiology AI vendors should consider these dynamics when developing models for various markets and tailor their value propositions to the needs of the potential customer. Ultimately, this will increase AI adoption and transform radiology AI into a financially sustainable tool for radiologists, hospitals systems, and most importantly patients.”
    The Business of Artificial Intelligence in Radiology Has Little to Do With Radiologists  
    Trivedi H
    J Am Coll Radiol. 2022 Feb 15:S1546-1440(22)00113-2.
  • “Artificial intelligence (AI) and machine learning (ML) are poised to transform the way health care is delivered. AI is the use of computers to simulate intelligent tasks typically performed by humans. ML is a domain of AI that involves computers automatically learning from data without a priori programming. While AI has been critiqued as being in its “hype cycle” (throughout this article, AI will be used as shorthand for AI and ML), over time, it is likely that every medical specialty will be influenced by AI, and some will be transformed. As AI takes on a larger role in clinical practice, it is clear that multiple levels of oversight are needed. However, even with appropriate outside oversight, the importance of clinician review and trust of these technologies cannot be overstated.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • "Importantly, equipping clinicians with the skills, resources, and support necessary to use AI-based technologies is now recognized as essential to successful deployment of AI in health care. To do so, clinicians need to have a realistic understanding of the potential uses and limitations of medical AI applications. Overlooking this fact risks clinician cynicism and suboptimal patient outcomes.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • "Despite relatively weak evidence supporting the use of AI in routine clinical practice health care settings, AI models continue to be marketed and deployed. A recent example is the Epic Sepsis Model. While this model was widely implemented in hundreds of US hospitals, a recent study showed that it performed significantly worse in correctly identifying patients with early sepsis and improving patient outcomes in a clinical set- ting compared with performance observed during development of the model.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • "At about the time that the EBM movement was being launched, a parallel movement began promoting shared decision-making between patients and clinicians. As AI-based predictions and algorithms continue to inform medical decisions, patients and clinicians must rethink shared decision-making as decisions may well now involve a new member of the team—an AI-derived algorithm. Ultimately, clinicians will bear much of the responsibility to successfully broker the triadic relationship between patients, the computer, and themselves.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • "Ultimately, clinicians will bear much of the responsibility to successfully broker the triadic relationship between patients, the computer, and themselves. Clinicians will need to explain the role that AI has in their reasoning and recommendations. Over time, this relationship is likely to change, with the possibility of some decisions being made directly by patients and families based on AI recommenda- tions, bypassing the clinician. Navigating this transition—and finding the appropriate role for credentialed experts in it—will be a significant challenge in a health care system transformed by AI.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • "AI will soon become ubiquitous in health care. Building on lessons learned as implementation strategies continue to be devised, it will be essential to consider the key role of clinicians as end users of AI-developed algorithms, processes, and risk predictors. It is imperative that clinicians have the knowledge and skills to assess and determine the appropriate application of AI outputs, for their own clinical practice and for their patients. Rather than being replaced by AI, these new technologies will create new roles and responsibilities for clinicians.”
    Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence
    James CA, Wachter  RM, Woolliscroft JO
    JAMA March 2022 doi:10.1001/jama.2022.3580 
  • “Our recent research, which is going to be published in Management Science, identifies a much better approach that would be relatively easy to implement. Using the same relative benchmarking idea that is currently used to incentivize cost reduction, hospitals’ waiting times should be measured (as they are) and benchmarked against the national (risk-adjusted) average waiting time of patients with similar conditions. Hospitals that exhibit shorter waiting times than the average should be financially rewarded, while underperforming hospitals should be penalized.”
    To Reduce Emergency Room Wait Times, Tie Them to Payments
    Nicos Savva  and Tolga Tezcan
    HBR February 06, 2019
  • “Our research shows that such financial and outcomes-based incentives create indirect competition on waiting times and have the same effect on outcomes as direct competition has on other service points, without patients needing to exercise choice. This solution would work without requiring the regulator to figure out the thorny question: What is an acceptable waiting time and how much would it cost?”
    To Reduce Emergency Room Wait Times, Tie Them to Payments
    Nicos Savva  and Tolga Tezcan
    HBR February 06, 2019
  • Key Points  
    Question: Can machine learning–based medical directives (MLMDs) be used to autonomously order testing at triage for common pediatric presentations in the emergency department?  
    Findings: This decision analytical model analyzing 77 219 presentations of children to an emergency department noted that the best-performing MLMD models obtained high area-under- receiver-operator curve and positive predictive values across 6 pediatric emergency department use cases. The implementation of MLMD using these thresholds may help streamline care for 22.3% of all patient visits.  
    Meaning: The findings of this study suggest MLMDs can autonomously order diagnostic testing for pediatric patients at triage with high positive predictive values and minimal over testing; model explainability can be provided to clinicians and patients regarding why a test is ordered, allowing for transparency and trust to be built with artificial intelligence systems.  
    Assessment of Machine Learning–Based Medical Directives to Expedite Care in Pediatric Emergency Medicine
    Devin Singh et al.  
    JAMA Network Open. 2022;5(3):e222599. doi:10.1001/jamanetworkopen.2022.2599 
  • OBJECTIVE  To explore the use of machine learning–based medical directives(MLMDs)to automate diagnostic testing at triage for patients with common pediatric ED diagnoses.  
    EXPOSURE Machine learning models were trained to predict the need for urinary dip stick testing, electrocardiogram, abdominal ultrasonography, testicular ultrasonography, bilirubin level testing, and forearm radiographs.  
    CONCLUSIONS AND RELEVANCE The findings of this  study suggest the potential for clinical automation using MLMDs. When integrated into clinical workflows, MLMDs may have the potential to autonomously order common ED tests early in a patient’s visit with explainability provided to patients and clinicians.  
    Assessment of Machine Learning–Based Medical Directives to Expedite Care in Pediatric Emergency Medicine
    Devin Singh et al.  
    JAMA Network Open. 2022;5(3):e222599. 
  • “Overcrowding in emergency departments (EDs) and prolonged wait times are common challenges associated with poor health outcomes globally. A retrospective cross-sectional study reported that few EDs achieve recommended wait times, and innovative strategies to improve patient flow are required to address these challenges. The typical pathway for patients with stable vital signs in an ED involves a triage assessment followed by transfer to a waiting area. As ED assessment rooms become available, patients are moved into the department and then wait until they can be assessed by a health care practitioner (HCP). From here, an HCP will order tests to rule in or out suspected differential diagnoses as needed. This process triggers another sequence of waiting for the test to be conducted and for results to be processed before further reassessment and treatment are provided.”
    Assessment of Machine Learning–Based Medical Directives to Expedite Care in Pediatric Emergency Medicine
    Devin Singh et al.  
    JAMA Network Open. 2022;5(3):e222599. 
  • “The findings of this study suggest that segments of health care in EDs can be automated, adding both efficiency and consistency to the way care is delivered to patients at scale. This service can be achieved through the development of MLMDs programmed to have high PPVs and low FPRs. When integrated into clinical workflow using an augmented dual-pathway system, automation can be achieved without overtesting.”
    Assessment of Machine Learning–Based Medical Directives to Expedite Care in Pediatric Emergency Medicine
    Devin Singh et al.  
    JAMA Network Open. 2022;5(3):e222599. 
  • "To quantify the initial association of MLMDs with wait times for patients, we measured the time between triage completion (ie, the time when an MLMD would activate) and when a test order is made in the EHR by an HCP. As noted in the Table, this time difference represents the potential efficiency gained within this segment of the patient journey if the directive is ordered at triage. Using these data, the weighted mean reduction was approximately 165 minutes per patient when the MLMD was activated.”
    Assessment of Machine Learning–Based Medical Directives to Expedite Care in Pediatric Emergency Medicine
    Devin Singh et al.  
    JAMA Network Open. 2022;5(3):e222599. 
  • Purpose: To develop DL algorithms for the automated classification of benign versus malignant ovarian tumors assessed with US and to compare algorithm performance to Ovarian-Adnexal Reporting and Data System (O-RADS) and subjective expert assessment for malignancy.  
    Results: A total of 422 women (mean age, 46.4 years 6 14.8 [SD]) with 304 benign and 118 malignant tumors were included; there were 337 women in the training and validation data set and 85 women in the test data set. DLfeature had an AUC of 0.93 (95% CI: 0.85, 0.97) for classifying malignant from benign ovarian tumors, comparable with O-RADS (AUC, 0.92; 95% CI: 0.85, 0.97; P  .88) and expert assessment (AUC, 0.97; 95% CI: 0.91, 0.99; P  .07), and similar to DLdecision (AUC, 0.90; 95% CI: 0.82, 0.96; P  .29). DLdecision, DLfeature, O-RADS, and expert assessment achieved sensitivities of 92%, 92%, 92%, and 96%, respectively, and specificities of 80%, 85%, 89%, and 87%, respectively, for malignancy.  
    Conclusion: Deep learning algorithms developed by using multimodal US images may distinguish malignant from benign ovarian tumors with diagnostic performance comparable to expert subjective and Ovarian-Adnexal Reporting and Data System assessment.  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 
  • Key Results  
    • In this retrospective study of 422 women, US-based deep learning with feature fusion (DLfeature) was comparable to Ovarian-Adnexal Reporting and Data System (O-RADS) risk categorization (area under the receiver operating characteristic curve [AUC], 0.93 vs 0.92, respectively; P  .88) and subjective expert assessment (AUC, 0.93 vs 0.97, respectively; P  .07) in distinguishing malignant from benign ovarian tumors.  
    • DLfeature, O-RADS, and expert assessment achieved sensitivities of 92%, 92%, and 96%, respectively, and specificities of 85%, 89%, and 87%, respectively, for malignancy.  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 
  • “Ovarian cancer is the second most common cause of cancer-related death worldwide among women, with a 5-year survival rate less than 45%. Early detection and accurate characterization of ovarian tumors is important for optimal patient treatment (3,4). Benign tumors can be treated conservatively, avoiding unnecessary costs and overtreatment, and preserving fertility. However, malignant tumors require referral to gynecologic oncology, appropriate staging, and consideration for radical surgery. To provide individualized and effective treatment op- tions, it is critical to be able to distinguish benign and malignant ovarian tumors with high accuracy.”
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 

  • Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367  
  • "In conclusion, we demonstrated that deep learning algorithms based on multimodal US images may predict ovarian malignancy with high diagnostic performance comparable to that of expert subjective and Ovarian-Adnexal Reporting and Data System assessment.”  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367  
  • Objectives: To evaluate and compare the diagnostic performances of a commercialized artificial intelligence (AI) algorithm for diagnosing pulmonary embolism (PE) on CT pulmonary angiogram (CTPA) with those of emergency radiologists in routine clinical practice.
    Methods: This was an IRB-approved retrospective multicentric study including patients with suspected PE from September to December 2019 (i.e., during a preliminary evaluation period of an approved AI algorithm). CTPA quality and conclusions by emergency radiologists were retrieved from radiological reports. The gold standard was a retrospective review of CTPA, radiological and clinical reports, AI outputs, and patient outcomes. Diagnostic performance metrics for AI and radiologists were assessed in the entire cohort and depending on CTPA quality.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2 
  • Results: Overall, 1202 patients were included (median age: 66.2 years). PE prevalence was 15.8% (190/1202). The AI algorithm detected 219 suspicious PEs, of which 176 were true PEs, including 19 true PEs missed by radiologists. In the cohort, the highest sensitivity and negative predictive values (NPVs) were obtained with AI (92.6% versus 90% and 98.6% versus 98.1%, respectively), while the highest specificity and positive predictive value (PPV) were found with radiologists (99.1% versus 95.8% and 95% versus 80.4%, respectively). Accuracy, specificity, and PPV were significantly higher for radiologists except in subcohorts with poor-to-average injection quality. Radiologists positively evaluated the AI algorithm to improve their diagnostic comfort (55/79 [69.6%]).  
    Conclusion: Instead of replacing radiologists, AI for PE detection appears to be a safety net in emergency radiology practice due to high sensitivity and NPV, thereby increasing the self-confidence of radiologists.
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022 https://doi.org/10.1007/s00330-022-08645-2 
  • Key Points
    • Both the AI algorithm and emergency radiologists showed excellent performance in diagnosing PE on CTPA (sensitivity and specificity ≥ 90%; accuracy ≥ 95%).
    • The AI algorithm for PE detection can help increase the sensitivity and NPV of emergency radiologists in clinical practice, especially in cases of poor-to-moderate injection quality.
    • Emergency radiologists recommended the use of AI for PE detection in satisfaction surveys to increase their confidence and comfort in their final diagnosis.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  • “Indeed, in the entire cohort-2019, AIDOC captured 19 PEs that were not diagnosed by radiologists in 19 distinct patients. In other words, the AI algorithm could correct a misdiagnosed PE approximately every 63 CTPAs (≈1202/19). This estimation must be considered in parallel with the high number of CTPAs required by emergency physicians (≈18,000 CTPAs in 2020 in our group—so approximately 285 [≈18000/1202 × 19] true PEs detected by AI but initially misdiagnosed by radiologists in 2020) and with human and financial consequences of missed PEs [32]. Indeed, mortality and recurrence rates for untreated or missed PE range between 5 and 30%.”  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  •  “In conclusion, this study confirms the high diagnostic performances of AI algorithms relying on DCNN to diagnose PE on CTPA in a large multicentric retrospective emergency series. It also underscores where and how AI algorithms could better support (or “augment”) radiologists, i.e., for poor- quality examinations and by increasing their diagnostic con- fidence through the high sensitivity and high NPV of AI. Thus, our work provides more scientific ground for the concept of “AI-augmented” radiologists instead of supporting the theory of radiologists’ replacement by AI.”
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  • “Hip fractures are a major cause of morbidity and mortality in the elderly, and incur high health and social care costs. Given projected population ageing, the number of incident hip fractures is predicted to increase globally. As fracture classification strongly determines the chosen surgical treatment, differences in fracture classification influence patient outcomes and treatment costs. We aimed to create a machine learning method for identifying and classifying hip fractures, and to compare its performance to experienced human observers. We used 3659 hip radiographs, classified by at least two expert clinicians. The machine learning method was able to classify hip fractures with 19% greater accuracy than humans, achieving overall accuracy of 92%.”
    Machine learning outperforms  clinical experts in classification  of hip fractures  
    E. A. Murphy et al.
    Scientific Reports (Nature) (2022) 12:2058 
  • "In this work, we have demonstrated that a trained neural network can classify hip fractures with 19% increased accuracy compared to human observers with experience of hip fracture classification in a clinical setting. In the work presented here, we used as ground truth the classification of 3,659 hip radiographs by at least two (and up to five) experts to achieve consensus. Thus, this analysis is a prototype only and a more extensive study is needed before this approach can be fully transformed to a clinical application. We envisage that this approach could be used clinically and aid in the diagnosis and in the treatment of patients who sustain hip fractures.”
    Machine learning outperforms  clinical experts in classification  of hip fractures  
    E. A. Murphy et al.
    Scientific Reports (Nature) (2022) 12:2058 
  • Background: Patients with fractures are a common emergency presentation and may be misdiagnosed at radiologic imaging. An increasing number of studies apply artificial intelligence (AI) techniques to fracture detection as an adjunct to clinician diagnosis.
    Purpose: To perform a systematic review and meta-analysis comparing the diagnostic performance in fracture detection between AI and clinicians in peer-reviewed publications and the gray literature (ie, articles published on preprint repositories).  
    Materials and Methods: A search of multiple electronic databases between January 2018 and July 2020 (updated June 2021) was per- formed that included any primary research studies that developed and/or validated AI for the purposes of fracture detection at any imaging modality and excluded studies that evaluated image segmentation algorithms. Meta-analysis with a hierarchical model to calculate pooled sensitivity and specificity was used. Risk of bias was assessed by using a modified Prediction Model Study Risk of Bias Assessment Tool, or PROBAST, checklist.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • Results: Included for analysis were 42 studies, with 115 contingency tables extracted from 32 studies (55 061 images). Thirty-seven studies identified fractures on radiographs and five studies identified fractures on CT images. For internal validation test sets, the pooled sensitivity was 92% (95% CI: 88, 93) for AI and 91% (95% CI: 85, 95) for clinicians, and the pooled specificity was 91% (95% CI: 88, 93) for AI and 92% (95% CI: 89, 92) for clinicians. For external validation test sets, the pooled sensitivity was 91% (95% CI: 84, 95) for AI and 94% (95% CI: 90, 96) for clinicians, and the pooled specificity was 91% (95% CI: 81, 95) for AI and 94% (95% CI: 91, 95) for clinicians. There were no statistically significant differences between clinician and AI performance. There were 22 of 42 (52%) studies that were judged to have high risk of bias. Meta-regression identified multiple sources of heterogeneity in the data, including risk of bias and fracture type.  
    Conclusion: Artificial intelligence (AI) and clinicians had comparable reported diagnostic performance in fracture detection, suggesting that AI technology holds promise as a diagnostic adjunct in future clinical practice.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • Summary  
    Artificial intelligence is noninferior to clinicians in terms of diagnostic performance in fracture detection, showing promise as a useful diagnostic tool.  
    Key Results  
    • In a systematic review and meta-analysis of 42 studies (37 studies with radiography and five studies with CT), the pooled diagnostic performance from the use of artificial intelligence (AI) to detect fractures had a sensitivity of 92% and 91% and specificity of 91% and 91%, on internal and external validation, respectively.
    • Clinician performance had comparable performance to AI in fracture detection (sensitivity 91%, 92%; specificity 94%, 94%).  
    • Only 13 studies externally validated results, and only one study evaluated AI performance in a prospective clinical trial.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • “Current artificial intelligence (AI) is designed as a diagnostic adjunct and may improve workflow through screening or prioritizing images on worklists and highlighting regions of interest for a reporting radiologist. AI may also improve diagnostic certainty through acting as a “second reader” for clinicians or as an interim report prior to radiologist interpretation. However, it is not a replacement for the clinical workflow, and clinicians must understand AI performance and exercise judgement in interpreting algorithm output. We advocate for transparent reporting of study methods and results as crucial to AI integration. By addressing these areas for development, deep learning has potential to streamline fracture diagnosis in a way that is safe and sustainable for patients and health care systems.”
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785
  • Background: Proximal femoral fractures are an important clinical and public health issue associated with substantial morbidity and early mortality. Artificial intelligence might offer improved diagnostic accuracy for these fractures, but typical approaches to testing of artificial intelligence models can underestimate the risks of artificial intelligence- based diagnostic systems.  
    Methods: We present a preclinical evaluation of a deep learning model intended to detect proximal femoral fractures in frontal x-ray films in emergency department patients, trained on films from the Royal Adelaide Hospital (Adelaide, SA, Australia). This evaluation included a reader study comparing the performance of the model against five radiologists (three musculoskeletal specialists and two general radiologists) on a dataset of 200 fracture cases and 200 non-fractures (also from the Royal Adelaide Hospital), an external validation study using a dataset obtained from Stanford University Medical Center, CA, USA, and an algorithmic audit to detect any unusual or unexpected model behaviour.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Findings: In the reader study, the area under the receiver operating characteristic curve (AUC) for the performance of the deep learning model was 0·994 (95% CI 0·988–0·999) compared with an AUC of 0·969 (0·960–0·978) for the five radiologists. This strong model performance was maintained on external validation, with an AUC of 0·980 (0·931–1·000). However, the preclinical evaluation identified barriers to safe deployment, including a substantial shift in the model operating point on external validation and an increased error rate on cases with abnormal bones (eg, Paget’s disease).  
    Interpretation: The model outperformed the radiologists tested and maintained performance on external validation, but showed several unexpected limitations during further testing. Thorough preclinical evaluation of artificial intelligence models, including algorithmic auditing, can reveal unexpected and potentially harmful behaviour even in high-performance artificial intelligence systems, which can inform future clinical testing and deployment decisions.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Interpretation: The model outperformed the radiologists tested and maintained performance on external validation, but showed several unexpected limitations during further testing. Thorough preclinical evaluation of artificial intelligence models, including algorithmic auditing, can reveal unexpected and potentially harmful behaviour even in high-performance artificial intelligence systems, which can inform future clinical testing and deployment decisions.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Added value of this study:  This study presents a thorough preclinical evaluation of a medical artificial intelligence system (trained to detect proximal femoral fractures on plain film imaging). Despite high performance of the model, which outperformed human experts in the task of proximal femoral fracture detection, an evaluation including algorithmic auditing showed unexpected and potentially harmful algorithmic behaviour.  
    Implications of all the available evidence:  Thorough evaluation of artificial intelligence systems, including algorithmic auditing, can identify barriers to safe artificial intelligence deployment that might not be appreciated during standard preclinical testing and which could cause significant harm. Regulators, medical governance bodies, and professional groups should consider the need for more comprehensive preclinical testing of artificial intelligence before clinical deployment.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • “We note that although our model shows high performance, and does not appear to deviate from human performance in prespecified subgroups,it does still make the occasional inhuman error (eg, misdiagnosing a highly displaced fracture). We also note on saliency mapping that although the model reproduces some recognizable aspects of human practice (eg, it appears to pay attention to Shenton’s line), the visualizations nonetheless raise concerns about the regions that are not highlighted in the heatmaps. In particular, the saliency maps almost never show strong activity along the outer region of the femoral neck, even in cases where the cortex in this area is clearly disrupted.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • "Our study evaluated a high-performance proximal femoral fracture detection deep learning model, which outperforms highly trained clinical specialists in diagnostic conditions, as well as other clinical readers in normal clinical conditions. The performance of the artificial intelligence system was maintained when applied to an external validation sample, and a thorough analysis of the behaviour of the artificial intelligence system shows that it is mostly consistent with that of human experts. We also characterized the occasional aberrant or unexpected behaviour of the artificial intelligence model which could inform future clinical testing protocols. We next intend to test our model in a clinical environment, in the form of an interventional randomised controlled trial.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • "Our study had a number of limitations. First, the deep learning model itself is limited by being unable to act on cases with implanted metalwork (although our system is able to automatically identify these cases and exclude them from analysis). Second, the sample size of the MRMC study was limited by the availability of readers; we determined a total dataset of 400 cases (200 positive and 200 negative cases) was as many as we could reasonably expect the readers to review, and only five radiologists reviewed the cases under diagnostic conditions as defined in the local standards of practice.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Objectives: To develop and validate machine learning models to distinguish between benign and malignant bone lesions and compare the performance to radiologists.
    Results: The best machine learning model was based on an artificial neural network (ANN) combining both radiomic and demographic information achieving 80% and 75% accuracy at 75% and 90% sensitivity with 0.79 and 0.90 AUC on the internal and external test set, respectively. In comparison, the radiology residents achieved 71% and 65% accuracy at 61% and 35% sensitivity while the radiologists specialized in musculoskeletal tumor imaging achieved an 84% and 83% accuracy at 90% and 81% sensitivity, respectively.  
    Conclusions: An ANN combining radiomic features and demographic information showed the best performance in distinguishing between benign and malignant bone lesions. The model showed lower accuracy compared to specialized radiologists, while accuracy was higher or similar compared to residents.  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • “In this study, machine learning models based on radiomics and demographic information were developed and validated to distinguish between benign and malignant bone lesions on radiographs and compared to radiologists on an external test set. Overall, machine learning models using the combination of radiomics and demographic information showed a higher diagnostic accuracy than machine learning models using radiomics or demographic information only. The best model was based on an ANN that used both radiomics and demographic information. On an external test set, this model demonstrated lower accuracy compared to radiologists specialized in musculoskeletal tumor imaging, while accuracy was higher or similar compared to radiology residents.”
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • Results: The best machine learning model was based on an artificial neural network (ANN) combining both radiomic and demographic information achieving 80% and 75% accuracy at 75% and 90% sensitivity with 0.79 and 0.90 AUC on the internal and external test set, respectively. In comparison, the radiology residents achieved 71% and 65% accuracy at 61% and 35% sensitivity while the radiologists specialized in musculoskeletal tumor imaging achieved an 84% and 83% accuracy at 90% and 81% sensitivity, respectively.  
    Conclusions: An ANN combining radiomic features and demographic information showed the best performance in distinguishing between benign and malignant bone lesions. The model showed lower accuracy compared to specialized radiologists, while accuracy was higher or similar compared to residents.  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • “In conclusion, a machine learning model using both radiomic features and demographic information was developed that showed high accuracy and discriminatory power for the distinction between benign and malignant bone tumors on radiographs of patients that underwent biopsy. The best model was based on an ANN that used both radiomics and demographic information resulting in an accuracy higher or similar compared to radiology residents. A model such as this may enhance diagnostic decision-making especially for radiologists or physicians with limited experience and may therefore improve the diagnostic work up of bone tumors.”  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
Kidney

  • Causes of Spontaneous Bleeds: Kidney
    - Clear cell renal cell carcinoma
    - AML
    - Oncocytoma
    - TCC
  • “RCC is the most common malignant renal tumor, with clear cell carcinoma being the most common subtype. Also, RCC is the most common malignant renal neoplasm associated with intratumoral hemorrhage or perirenal hematoma. In addition, hematuria may be observed in cases where the tumor extends into the renal pelvis. The tumor subtype is not a good predictor of hemorrhage, as both clear cell RCC and papillary RCC, the most common subtypes, usually contain hemorrhagic areas. Also, unlike angiomyolipoma (AML), tumor size is not a reliable predictor for bleeding in RCCs. Rarely, an arterio-venous fistula may develop within larger tumors, which may increase the risk of spontaneous bleeding. Tumor thrombus in renal veins may also predispose the tumor to bleed due to increased intratumoral pressure.”
    Imaging findings of spontaneous intraabdominal hemorrhage: neoplastic and non‐neoplastic causes  
    Sevtap Arslan et al.
    Abdominal Radiology (2022) 47:1473–1502 
Liver

  • “Approximately 7% of benign HAs will progress to HCC. Whenever a distinctive nodule is noted within a HA lesion, the suspicion for malignancy should be high. The probability of developing cancer is highest if the patient is male, has a β-catenin subtype, or has a large tumor > 5 cm. The most significant risk factor is male sex, which increases the chance of malignant transformation into HCC by ten- fold, with a 10-year cumulative risk of up to 60%. The risk of HCC increases exponentially with tumor size, yet is unrelated to the number of HAs. The β-catenin subtype is most frequently implicated in malignant transformation, with an incidence of up to 50%.”
    A Scoping Review of the Classification, Diagnosis, and Management of Hepatic Adenomas  
    Hassan Aziz et al.  
    Journal of Gastrointestinal Surgery (2022) 26:965–978 
  • "Patients with HA require surveillance imaging as HCC can develop in a pre-existing lesion over many years. HA patients taking exogenous androgens (for example, patients with Fanconi anemia) should especially be monitored for signs of malignancy. OCPs and liver glycogen diseases have also been implicated in the pathogenesis of HA but have not been as strongly associated with malignant transformation. HA-related HCC has a better prognosis than other HCC variants because it is usually detected at an early stage and resected with negative margins. Furthermore, patients with HA tend to have a relatively normal background liver, while patients with HCC tend to have significant liver disease.”
    A Scoping Review of the Classification, Diagnosis, and Management of Hepatic Adenomas  
    Hassan Aziz et al.  
    Journal of Gastrointestinal Surgery (2022) 26:965–978 
  • "Given that spontaneous regression of HAs after discontinuing OCPs may occur, often the first step in the management of HAs in asymptomatic women without evidence of β-catenin activation is the cessation of estrogen-containing medications. A subsequent 6-month observation period, even in patients with HAs ≥ 5 cm, may be a reasonable approach. To this point, Klompenhouwer et al. reported that 58.5% of HAs regressed to less than 5 cm at a median of 104 weeks after cessation of OCPs. Larger tumors took longer to regress, and there was no correlation between complications and observation time, suggesting that post-hormonal surveillance can be extended up to 12 months. For individuals with a nodule ≥ 5 cm or a lesion that demonstrates continued growth on repeat imaging, surgical resection is warranted.”
    A Scoping Review of the Classification, Diagnosis, and Management of Hepatic Adenomas  
    Hassan Aziz et al.  
    Journal of Gastrointestinal Surgery (2022) 26:965–978 
  • Causes of Spontaneous Bleeds: Liver
    - Hepatoma
    - Hepatic adenoma
    - Focal nodular hyperplasia
    - Metastases (NET, RCC, melanoma)
    - Hepatic angiosarcoma
    - Hepatoblastoma
  • "Hepatic angiosarcoma (HA), albeit rare, is the most common malignant mesenchymal tumor of the liver and accounts for 2% of all primary liver tumors. Exposure to thorotrast, vinyl chloride, arsenic, anabolic steroids, and radiation have all been mentioned in etiopathogenesis. It is more common in males and mostly seen in the sixth or seventh decades of life. HAs are biologically very aggressive and often metastatic at the initial diagnosis. Lungs and spleen are the most common sites for metastases. On gross pathological evaluation, the tumor is characterized by remarkable necrosis and intratumoral bleeding.”
    Imaging findings of spontaneous intraabdominal hemorrhage: neoplastic and non‐neoplastic causes  
    Sevtap Arslan et al.
    Abdominal Radiology (2022) 47:1473–1502 
  • "Although the most common cause of spontaneous hepatic hemorrhage in patients with cirrhosis is HCC rupture, spontaneous hepatic hemorrhage without malignant changes may also be rarely seen in cirrhosis. The underlying mechanism is either the rupture of macronodular cirrhosis or rupture of the lymphatic vessels and varicose veins due to portal hypertension. Extreme caution should be exercised in these patients not to overlook an underlying HCC.”
    Imaging findings of spontaneous intraabdominal hemorrhage: neoplastic and non‐neoplastic causes  
    Sevtap Arslan et al.
    Abdominal Radiology (2022) 47:1473–1502 
Musculoskeletal

  • “Hip fractures are a major cause of morbidity and mortality in the elderly, and incur high health and social care costs. Given projected population ageing, the number of incident hip fractures is predicted to increase globally. As fracture classification strongly determines the chosen surgical treatment, differences in fracture classification influence patient outcomes and treatment costs. We aimed to create a machine learning method for identifying and classifying hip fractures, and to compare its performance to experienced human observers. We used 3659 hip radiographs, classified by at least two expert clinicians. The machine learning method was able to classify hip fractures with 19% greater accuracy than humans, achieving overall accuracy of 92%.”
    Machine learning outperforms  clinical experts in classification  of hip fractures  
    E. A. Murphy et al.
    Scientific Reports (Nature) (2022) 12:2058 
  • "In this work, we have demonstrated that a trained neural network can classify hip fractures with 19% increased accuracy compared to human observers with experience of hip fracture classification in a clinical setting. In the work presented here, we used as ground truth the classification of 3,659 hip radiographs by at least two (and up to five) experts to achieve consensus. Thus, this analysis is a prototype only and a more extensive study is needed before this approach can be fully transformed to a clinical application. We envisage that this approach could be used clinically and aid in the diagnosis and in the treatment of patients who sustain hip fractures.”
    Machine learning outperforms  clinical experts in classification  of hip fractures  
    E. A. Murphy et al.
    Scientific Reports (Nature) (2022) 12:2058 
  • Background: Patients with fractures are a common emergency presentation and may be misdiagnosed at radiologic imaging. An increasing number of studies apply artificial intelligence (AI) techniques to fracture detection as an adjunct to clinician diagnosis.
    Purpose: To perform a systematic review and meta-analysis comparing the diagnostic performance in fracture detection between AI and clinicians in peer-reviewed publications and the gray literature (ie, articles published on preprint repositories).  
    Materials and Methods: A search of multiple electronic databases between January 2018 and July 2020 (updated June 2021) was per- formed that included any primary research studies that developed and/or validated AI for the purposes of fracture detection at any imaging modality and excluded studies that evaluated image segmentation algorithms. Meta-analysis with a hierarchical model to calculate pooled sensitivity and specificity was used. Risk of bias was assessed by using a modified Prediction Model Study Risk of Bias Assessment Tool, or PROBAST, checklist.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • Results: Included for analysis were 42 studies, with 115 contingency tables extracted from 32 studies (55 061 images). Thirty-seven studies identified fractures on radiographs and five studies identified fractures on CT images. For internal validation test sets, the pooled sensitivity was 92% (95% CI: 88, 93) for AI and 91% (95% CI: 85, 95) for clinicians, and the pooled specificity was 91% (95% CI: 88, 93) for AI and 92% (95% CI: 89, 92) for clinicians. For external validation test sets, the pooled sensitivity was 91% (95% CI: 84, 95) for AI and 94% (95% CI: 90, 96) for clinicians, and the pooled specificity was 91% (95% CI: 81, 95) for AI and 94% (95% CI: 91, 95) for clinicians. There were no statistically significant differences between clinician and AI performance. There were 22 of 42 (52%) studies that were judged to have high risk of bias. Meta-regression identified multiple sources of heterogeneity in the data, including risk of bias and fracture type.  
    Conclusion: Artificial intelligence (AI) and clinicians had comparable reported diagnostic performance in fracture detection, suggesting that AI technology holds promise as a diagnostic adjunct in future clinical practice.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • Summary  
    Artificial intelligence is noninferior to clinicians in terms of diagnostic performance in fracture detection, showing promise as a useful diagnostic tool.  
    Key Results  
    • In a systematic review and meta-analysis of 42 studies (37 studies with radiography and five studies with CT), the pooled diagnostic performance from the use of artificial intelligence (AI) to detect fractures had a sensitivity of 92% and 91% and specificity of 91% and 91%, on internal and external validation, respectively.
    • Clinician performance had comparable performance to AI in fracture detection (sensitivity 91%, 92%; specificity 94%, 94%).  
    • Only 13 studies externally validated results, and only one study evaluated AI performance in a prospective clinical trial.  
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785 
  • “Current artificial intelligence (AI) is designed as a diagnostic adjunct and may improve workflow through screening or prioritizing images on worklists and highlighting regions of interest for a reporting radiologist. AI may also improve diagnostic certainty through acting as a “second reader” for clinicians or as an interim report prior to radiologist interpretation. However, it is not a replacement for the clinical workflow, and clinicians must understand AI performance and exercise judgement in interpreting algorithm output. We advocate for transparent reporting of study methods and results as crucial to AI integration. By addressing these areas for development, deep learning has potential to streamline fracture diagnosis in a way that is safe and sustainable for patients and health care systems.”
    Artificial Intelligence in Fracture Detection:  A Systematic Review and Meta-Analysis  
    Kuo RYL et al.
    Radiology 2022; 000:1–13 • https://doi.org/10.1148/radiol.211785
  • Background: Proximal femoral fractures are an important clinical and public health issue associated with substantial morbidity and early mortality. Artificial intelligence might offer improved diagnostic accuracy for these fractures, but typical approaches to testing of artificial intelligence models can underestimate the risks of artificial intelligence- based diagnostic systems.  
    Methods: We present a preclinical evaluation of a deep learning model intended to detect proximal femoral fractures in frontal x-ray films in emergency department patients, trained on films from the Royal Adelaide Hospital (Adelaide, SA, Australia). This evaluation included a reader study comparing the performance of the model against five radiologists (three musculoskeletal specialists and two general radiologists) on a dataset of 200 fracture cases and 200 non-fractures (also from the Royal Adelaide Hospital), an external validation study using a dataset obtained from Stanford University Medical Center, CA, USA, and an algorithmic audit to detect any unusual or unexpected model behaviour.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Findings: In the reader study, the area under the receiver operating characteristic curve (AUC) for the performance of the deep learning model was 0·994 (95% CI 0·988–0·999) compared with an AUC of 0·969 (0·960–0·978) for the five radiologists. This strong model performance was maintained on external validation, with an AUC of 0·980 (0·931–1·000). However, the preclinical evaluation identified barriers to safe deployment, including a substantial shift in the model operating point on external validation and an increased error rate on cases with abnormal bones (eg, Paget’s disease).  
    Interpretation: The model outperformed the radiologists tested and maintained performance on external validation, but showed several unexpected limitations during further testing. Thorough preclinical evaluation of artificial intelligence models, including algorithmic auditing, can reveal unexpected and potentially harmful behaviour even in high-performance artificial intelligence systems, which can inform future clinical testing and deployment decisions.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Interpretation: The model outperformed the radiologists tested and maintained performance on external validation, but showed several unexpected limitations during further testing. Thorough preclinical evaluation of artificial intelligence models, including algorithmic auditing, can reveal unexpected and potentially harmful behaviour even in high-performance artificial intelligence systems, which can inform future clinical testing and deployment decisions.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Added value of this study:  This study presents a thorough preclinical evaluation of a medical artificial intelligence system (trained to detect proximal femoral fractures on plain film imaging). Despite high performance of the model, which outperformed human experts in the task of proximal femoral fracture detection, an evaluation including algorithmic auditing showed unexpected and potentially harmful algorithmic behaviour.  
    Implications of all the available evidence:  Thorough evaluation of artificial intelligence systems, including algorithmic auditing, can identify barriers to safe artificial intelligence deployment that might not be appreciated during standard preclinical testing and which could cause significant harm. Regulators, medical governance bodies, and professional groups should consider the need for more comprehensive preclinical testing of artificial intelligence before clinical deployment.  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • “We note that although our model shows high performance, and does not appear to deviate from human performance in prespecified subgroups,it does still make the occasional inhuman error (eg, misdiagnosing a highly displaced fracture). We also note on saliency mapping that although the model reproduces some recognizable aspects of human practice (eg, it appears to pay attention to Shenton’s line), the visualizations nonetheless raise concerns about the regions that are not highlighted in the heatmaps. In particular, the saliency maps almost never show strong activity along the outer region of the femoral neck, even in cases where the cortex in this area is clearly disrupted.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • "Our study evaluated a high-performance proximal femoral fracture detection deep learning model, which outperforms highly trained clinical specialists in diagnostic conditions, as well as other clinical readers in normal clinical conditions. The performance of the artificial intelligence system was maintained when applied to an external validation sample, and a thorough analysis of the behaviour of the artificial intelligence system shows that it is mostly consistent with that of human experts. We also characterized the occasional aberrant or unexpected behaviour of the artificial intelligence model which could inform future clinical testing protocols. We next intend to test our model in a clinical environment, in the form of an interventional randomised controlled trial.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • "Our study had a number of limitations. First, the deep learning model itself is limited by being unable to act on cases with implanted metalwork (although our system is able to automatically identify these cases and exclude them from analysis). Second, the sample size of the MRMC study was limited by the availability of readers; we determined a total dataset of 400 cases (200 positive and 200 negative cases) was as many as we could reasonably expect the readers to review, and only five radiologists reviewed the cases under diagnostic conditions as defined in the local standards of practice.”  
    Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study
    Lauren Oakden-Rayner et al.
    www.thelancet.com/digital-health Published online April 5, 2022 https://doi.org/10.1016/S2589-7500(22)00004-8  
  • Objectives: To develop and validate machine learning models to distinguish between benign and malignant bone lesions and compare the performance to radiologists.
    Results: The best machine learning model was based on an artificial neural network (ANN) combining both radiomic and demographic information achieving 80% and 75% accuracy at 75% and 90% sensitivity with 0.79 and 0.90 AUC on the internal and external test set, respectively. In comparison, the radiology residents achieved 71% and 65% accuracy at 61% and 35% sensitivity while the radiologists specialized in musculoskeletal tumor imaging achieved an 84% and 83% accuracy at 90% and 81% sensitivity, respectively.  
    Conclusions: An ANN combining radiomic features and demographic information showed the best performance in distinguishing between benign and malignant bone lesions. The model showed lower accuracy compared to specialized radiologists, while accuracy was higher or similar compared to residents.  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • “In this study, machine learning models based on radiomics and demographic information were developed and validated to distinguish between benign and malignant bone lesions on radiographs and compared to radiologists on an external test set. Overall, machine learning models using the combination of radiomics and demographic information showed a higher diagnostic accuracy than machine learning models using radiomics or demographic information only. The best model was based on an ANN that used both radiomics and demographic information. On an external test set, this model demonstrated lower accuracy compared to radiologists specialized in musculoskeletal tumor imaging, while accuracy was higher or similar compared to radiology residents.”
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • Results: The best machine learning model was based on an artificial neural network (ANN) combining both radiomic and demographic information achieving 80% and 75% accuracy at 75% and 90% sensitivity with 0.79 and 0.90 AUC on the internal and external test set, respectively. In comparison, the radiology residents achieved 71% and 65% accuracy at 61% and 35% sensitivity while the radiologists specialized in musculoskeletal tumor imaging achieved an 84% and 83% accuracy at 90% and 81% sensitivity, respectively.  
    Conclusions: An ANN combining radiomic features and demographic information showed the best performance in distinguishing between benign and malignant bone lesions. The model showed lower accuracy compared to specialized radiologists, while accuracy was higher or similar compared to residents.  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
  • “In conclusion, a machine learning model using both radiomic features and demographic information was developed that showed high accuracy and discriminatory power for the distinction between benign and malignant bone tumors on radiographs of patients that underwent biopsy. The best model was based on an ANN that used both radiomics and demographic information resulting in an accuracy higher or similar compared to radiology residents. A model such as this may enhance diagnostic decision-making especially for radiologists or physicians with limited experience and may therefore improve the diagnostic work up of bone tumors.”  
    Development and evaluation of machine learning models based on X-ray radiomics for the classification and differentiation of malignant and benign bone tumors  
    Claudio E. von Schacky et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08764-w 
OB GYN

  • Purpose: To develop DL algorithms for the automated classification of benign versus malignant ovarian tumors assessed with US and to compare algorithm performance to Ovarian-Adnexal Reporting and Data System (O-RADS) and subjective expert assessment for malignancy.  
    Results: A total of 422 women (mean age, 46.4 years 6 14.8 [SD]) with 304 benign and 118 malignant tumors were included; there were 337 women in the training and validation data set and 85 women in the test data set. DLfeature had an AUC of 0.93 (95% CI: 0.85, 0.97) for classifying malignant from benign ovarian tumors, comparable with O-RADS (AUC, 0.92; 95% CI: 0.85, 0.97; P  .88) and expert assessment (AUC, 0.97; 95% CI: 0.91, 0.99; P  .07), and similar to DLdecision (AUC, 0.90; 95% CI: 0.82, 0.96; P  .29). DLdecision, DLfeature, O-RADS, and expert assessment achieved sensitivities of 92%, 92%, 92%, and 96%, respectively, and specificities of 80%, 85%, 89%, and 87%, respectively, for malignancy.  
    Conclusion: Deep learning algorithms developed by using multimodal US images may distinguish malignant from benign ovarian tumors with diagnostic performance comparable to expert subjective and Ovarian-Adnexal Reporting and Data System assessment.  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 
  • Key Results  
    • In this retrospective study of 422 women, US-based deep learning with feature fusion (DLfeature) was comparable to Ovarian-Adnexal Reporting and Data System (O-RADS) risk categorization (area under the receiver operating characteristic curve [AUC], 0.93 vs 0.92, respectively; P  .88) and subjective expert assessment (AUC, 0.93 vs 0.97, respectively; P  .07) in distinguishing malignant from benign ovarian tumors.  
    • DLfeature, O-RADS, and expert assessment achieved sensitivities of 92%, 92%, and 96%, respectively, and specificities of 85%, 89%, and 87%, respectively, for malignancy.  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 
  • “Ovarian cancer is the second most common cause of cancer-related death worldwide among women, with a 5-year survival rate less than 45%. Early detection and accurate characterization of ovarian tumors is important for optimal patient treatment (3,4). Benign tumors can be treated conservatively, avoiding unnecessary costs and overtreatment, and preserving fertility. However, malignant tumors require referral to gynecologic oncology, appropriate staging, and consideration for radical surgery. To provide individualized and effective treatment op- tions, it is critical to be able to distinguish benign and malignant ovarian tumors with high accuracy.”
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367 

  • Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367  
  • "In conclusion, we demonstrated that deep learning algorithms based on multimodal US images may predict ovarian malignancy with high diagnostic performance comparable to that of expert subjective and Ovarian-Adnexal Reporting and Data System assessment.”  
    Deep Learning Prediction of Ovarian Malignancy at US Compared with O-RADS and Expert Assessment  
    Hui Chen et al.
    Radiology 2022; 000:1–8 • https://doi.org/10.1148/radiol.211367  
Pancreas

  • Causes of Spontaneous Bleeds: Pancreas
    - Serous cystadenoma
    - IPMN
    - Mucinous cystadenoma
    - Neuroendocrine tumor
Practice Management

  • Objective: To examine our website and social media audience data to define the number of African users and review the effects of COVID-19 on our viewership from Africa, and look ahead to potential opportunities.
    Conclusion: We have been successful in achieving a 27% increase in website traffic from Africa to our radiology social media offerings. This increase since the onset of the COVID-19 pandemic is encouraging, especially when noting that we do not have any paid advertising. We look forward to reaching a larger African audience to deliver radiology education in the future.  
    Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press

  • Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press 

  • Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press

  • Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press

  • Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press
  • "Reflecting on these data, and analyzing CTisus Facebook, YouTube, and Twitter statistics, we consider the limited internet accessibility as a likely factor in our total viewership from Africa. Specifically, in Central, Eastern, and Western African countries, only 8%, 10%, and 16% of their respective populations are active social media users.21 More- over, Eritrea, Niger, Central African Republic, Chad, Malawi, South Sudan, Democratic Republic of Congo, and Ethiopia ranked among the 10 countries with the lowest levels worldwide of social media use.21 Despite limited internet access, though, countries, such as Malawi, South Sudan, Democratic Republic of Congo, and Ethiopia were key contributors to our social media success. We are satisfied by our growing audience from Africa and encouraged that such interest has risen since the onset of the COVID-19 pandemic, especially when noting that we do not have any paid advertising; therefore, interest in our website and social media is generated by users them- selves through word searches or word of mouth.”
    Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press
  • “While reaching a larger African audience to deliver radiology education remains a goal for our team, broadening internet access in Africa for larger purposes, such as increasing awareness of urgent public healthcare issues (including the promotion of COVID-19 vaccination as, it is hoped, becomes more available), remains of greater importance. As it is, though, we are encouraged by the increasing interest from Africa in our radiology social media offerings and look forward to continued growth in the future. Further, we hope that our work helps to provide valuable information for education and improved patient care. “ .”
    Radiology Without Borders: Identifying Global Reach of Radiology Social Media in Africa  
    Elias Lugo-Fagundo, BS, Edmund M. Weisberg, MS, MBE*, Lilly Kauffman, BA, Elliot K. Fishman, MD  
    Current Problems in Diagnostic Radiology (2022) in press
  • “Radiological errors can be classified according to the reporting process as pre-reporting, reporting or post- reporting errors. Pre-reporting errors consist of tech- nical issues and procedure-related problems, whereas post-reporting errors are mainly caused by poor communication between radiologists and clinicians. Reporting errors are directly related to radiologists and can be categorized into two parts. "Perceptual errors" are more common and related to the fact that the present finding is not noticed, while "interpretative errors" are influenced by cognitive biases that can contribute to false reasoning.”
    Errors, discrepancies and underlying bias in radiology with case examples: a pictorial review
    Omer Onder et al.  
    Insights Imaging (2021) 12:51 https://doi.org/10.1186/s13244-021-00986-8 
  • “Even in geographies that have a reputation for high quality care (such as metropolitan Boston and metropolitan New York) there is a five times greater chance of death from acute myocardial infarction (heart attack), depending on the hospital one chooses. Across the United States, on average, patients are twice as likely to die in the lowest-performing hospitals. This includes a 2.3-fold difference in heart attack mortalities. There are even greater differences in safety. The top 10% of hospitals are 10 times safer than bottom 10%. Patients are 18 times more likely to suffer a bloodstream infection from a central venous catheter when treated at poor-performing hospitals.”  
    4 Actions to Reduce Medical Errors in U.S> Hospitals
    Troussaint JS, Segel KT
    Harvard Business Review April 20, 2022
  • 1. Make patient and staff safety a top priority
    2.Establish a national  safety organization
    3. create a national reporting mechanism
    4. turn on EHR’s machine leaning symptoms
    4 Actions to Reduce Medical Errors in U.S> Hospitals
    Troussaint JS, Segel KT
    Harvard Business Review April 20, 2022
  • “For early cancer detection, the percentage of confidence that comes with a diagnostic decision made by the algorithm might appear straightforward (for example, 80% confidence that lung cancer is present), but the process behind this number is very complex and understandably may not be apparent to the user. It is therefore not difficult to understand that there might be resistance to adoption of such strategies and the fear of overdiagnosis. It is important to understand that AI will not remove the need for physicians and experts to interpret the findings, provide a global picture of patient health, spot related diseases, and come up with a final diagnosis.”
    The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 
  • "What the introduction of AI algorithms might do, providing that data management and safety regulations are in place, is reduce the cost and time needed to diagnose the disease. This will enable health practitioners to spend more time developing efficient and holistic treatment protocols, and will make state-of-art diagnostics more affordable. Furthermore, AI can be a training tool that provides immediate specialist feedback to generalists so that, in time, they may perform at an expert level.”
    The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 

  • The future of early cancer detection  
    Rebecca C. Fitzgerald et al.  
    Nature Medicine| VOL 28 | APRIL 2022 | 666–677 
Small Bowel

  • “Small bowel neuroendocrine neoplasms (SB-NENs) are now the most common small bowel tumor, surpassing adenocarcinoma and comprising almost 40% of small bowel malignancies. Historically known as “carcinoid” tumors, the classification of NENs has evolved over the years, with increased recognition of their diverse clinical and pathologic profile. SB-NENs were previously classified as foregut duodenal or midgut jejunoileal neoplasms. Using this prior framework, duodenal NENs are uncommon, accounting for less than 3% of gastrointestinal neuroendocrine tumors, and are most often identified incidentally. Most are non- functional, and a small minority are associated with multiple endocrine neoplasia type 1 (MEN 1), MEN type 4, or neu- rofibromatosis type 1 syndromes. In comparison, jejunoileal tumors are much more common, with more than 70% of tumors originating within 100 cm of the ileocecal valve.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  • "Patients with SB-NENs are most often diagnosed in the sixth decade, without a gender predilection, and often incidentally on imaging. If patients present with symptoms, the clinical presentation in large part relates to the site of origin as well as underlying tumor burden. Patients with jejunoileal tumors may be asymptomatic, have long standing vague abdominal symptoms, or present with complications of local tumor progression or distant metastases, including carcinoid syndrome. The clinical and imaging findings of early stage disease can be subtle. Most jeju- noileal NETs are typically grade 1 or 2 tumors, with a more indolent course; this can lead to significant diagnostic delays. Duodenal tumors are more heterogeneous and even well differentiated tumors may present with early metastatic disease.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  •  “Small bowel NENs almost always produce biologically active peptides such as serotonin, histamine, and neurokinin A. However, these peptides are normally metabolized by the liver. The classic flushing, diarrhea, and bronchospasm of carcinoid syndrome occurs in up to 20% of patients, almost exclusively when liver and other distant metastases are present.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  • “Functional imaging has evolved and now primarily consists of positron emission tomography (PET) CT. Newer radiotracers take advantage of the high degree of somatostatin receptor expression exhibited by well differentiated NETs, and offer improved sensitivity for detection of disease. Gallium (68Ga) labelled octreotide analogs are widely in use, with the FDA having approved 68Ga-DOTA- octreotate (DOTATATE) for imaging of NETs. Fluorode- oxyglucose (FDG) 18 PET CT has a limited role in the assessment of grade 1 and 2 well differentiated neuroen- docrine tumors. However grade 3 NETs and PDNECs have variable somatostatin receptor expression, higher mitotic rates, and higher glucose utilization. Therefore, these tumors are often better imaged with conventional FDG 18 PET CT as opposed to DOTATATE PET CT.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  • "Similarly, the use of CT enterography allows for better small bowel distention with neutral oral contrast medium. This improves tumor-to- background contrast resolution and improves the sensitivity for detecting small bowel tumors to over 90%. Combining multiphase imaging with enterography further improves the detection of NET. This approach has been comparable to, if not better than video capsule endoscopy, which may be more limited in identifying submucosal lesions such as NEN.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  • "While tumor size, grade, and serosal invasion all play a role in the likelihood of developing metastases, almost half of patients with primary tumors under one centimeter in size still have mesenteric disease. The most relevant imaging finding in the setting of nodal disease, is whether involved lymph nodes are within the potential surgical field. Cytoreductive operations for jejunoileal NENs include resection of these locoregional mesenteric nodal metastases both for accurate staging and symptomatic improvement.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
  • "Small bowel neuroendocrine neoplasms have been increasingly identified and are now the most common small bowel tumor. Although frequently metastatic at presentation, initial surgical cytoreduction has demonstrated a survival benefit in well differentiated NET and may aid in symptom control. The radiologic findings of the primary tumor, nodal spread of disease, and distant metastases have significant impact on the management of these patients. A multimodality and interdisciplinary approach are necessary to determine the ideal treatment strategy for patients with small bowel NEN.”
    Small bowel neuroendocrine neoplasm: what surgeons want to know  
    Akshya Gupta et al.
    Abdominal Radiology  2022(in press)
Spleen

  • Causes of Spontaneous Bleeds: Spleen
    - Angiosarcoma
    - Prior pseudocyst from pancreatitis
  • "The most common primary tumor of the spleen asso- ciated with hemorrhage is, although rare, angiosarcoma. Tumor rupture is not rare in these patients, with a reported prevalence of 25%. Metastatic foci within liver parenchyma are commonly detected at the initial diagnosis. On cross-sectional imaging, these tumors are typically heterogeneously hyperenhancing masses with internal necrosis and hemorrhage. These imaging features characteristically reflect the biologically aggressive nature of these tumors.”
    Imaging findings of spontaneous intraabdominal hemorrhage: neoplastic and non‐neoplastic causes  
    Sevtap Arslan et al.
    Abdominal Radiology (2022) 47:1473–1502 
Vascular

  • Objectives: To evaluate and compare the diagnostic performances of a commercialized artificial intelligence (AI) algorithm for diagnosing pulmonary embolism (PE) on CT pulmonary angiogram (CTPA) with those of emergency radiologists in routine clinical practice.
    Methods: This was an IRB-approved retrospective multicentric study including patients with suspected PE from September to December 2019 (i.e., during a preliminary evaluation period of an approved AI algorithm). CTPA quality and conclusions by emergency radiologists were retrieved from radiological reports. The gold standard was a retrospective review of CTPA, radiological and clinical reports, AI outputs, and patient outcomes. Diagnostic performance metrics for AI and radiologists were assessed in the entire cohort and depending on CTPA quality.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2 
  • Results: Overall, 1202 patients were included (median age: 66.2 years). PE prevalence was 15.8% (190/1202). The AI algorithm detected 219 suspicious PEs, of which 176 were true PEs, including 19 true PEs missed by radiologists. In the cohort, the highest sensitivity and negative predictive values (NPVs) were obtained with AI (92.6% versus 90% and 98.6% versus 98.1%, respectively), while the highest specificity and positive predictive value (PPV) were found with radiologists (99.1% versus 95.8% and 95% versus 80.4%, respectively). Accuracy, specificity, and PPV were significantly higher for radiologists except in subcohorts with poor-to-average injection quality. Radiologists positively evaluated the AI algorithm to improve their diagnostic comfort (55/79 [69.6%]).  
    Conclusion: Instead of replacing radiologists, AI for PE detection appears to be a safety net in emergency radiology practice due to high sensitivity and NPV, thereby increasing the self-confidence of radiologists.
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022 https://doi.org/10.1007/s00330-022-08645-2 
  • Key Points
    • Both the AI algorithm and emergency radiologists showed excellent performance in diagnosing PE on CTPA (sensitivity and specificity ≥ 90%; accuracy ≥ 95%).
    • The AI algorithm for PE detection can help increase the sensitivity and NPV of emergency radiologists in clinical practice, especially in cases of poor-to-moderate injection quality.
    • Emergency radiologists recommended the use of AI for PE detection in satisfaction surveys to increase their confidence and comfort in their final diagnosis.  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  • “Indeed, in the entire cohort-2019, AIDOC captured 19 PEs that were not diagnosed by radiologists in 19 distinct patients. In other words, the AI algorithm could correct a misdiagnosed PE approximately every 63 CTPAs (≈1202/19). This estimation must be considered in parallel with the high number of CTPAs required by emergency physicians (≈18,000 CTPAs in 2020 in our group—so approximately 285 [≈18000/1202 × 19] true PEs detected by AI but initially misdiagnosed by radiologists in 2020) and with human and financial consequences of missed PEs [32]. Indeed, mortality and recurrence rates for untreated or missed PE range between 5 and 30%.”  
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
  •  “In conclusion, this study confirms the high diagnostic performances of AI algorithms relying on DCNN to diagnose PE on CTPA in a large multicentric retrospective emergency series. It also underscores where and how AI algorithms could better support (or “augment”) radiologists, i.e., for poor- quality examinations and by increasing their diagnostic con- fidence through the high sensitivity and high NPV of AI. Thus, our work provides more scientific ground for the concept of “AI-augmented” radiologists instead of supporting the theory of radiologists’ replacement by AI.”
    How artificial intelligence improves radiological interpretation in suspected pulmonary embolism  
    Alexandre Ben Cheikh et al.
    European Radiology 2022https://doi.org/10.1007/s00330-022-08645-2
© 1999-2022 Elliot K. Fishman, MD, FACR. All rights reserved.