Imaging Pearls ❯ 3D and Workflow ❯ Radiomics
|
-- OR -- |
|
- ”Radiomics is changing the world of medicine and more specifically the world of oncology. Early diagnosis and treatment improve the prognosis of patients with cancer. After treatment, the evaluation of the response will determine future treatments. In oncology, every change in treatment means a loss of ther- apeutic options and this is key in pancreatic cancer. Radiomics has been developed in oncology in the early diagnosis and differential diagnosis of benign and malignant lesions, in the evaluation of response, in the prediction of possible side effects, marking the risk of recurrence, survival and prognosis of the disease. Some studies have validated its use to differentiate normal tissues from tumor tissues with high sensitivity and specificity, and to differentiate cystic lesions and pancreatic neuroendocrine tumor grades with texture parameters. In addition, these parameters have been related to survival in patients with pancreatic cancer and to response to radiotherapy and chemotherapy. This review aimed to establish the current status of the use of radiomics in pancreatic cancer and future perspectives.”
Radiomics in pancreatic cancer for oncologist: Present and future
Carolina de la Pinta
Hepatobiliary & Pancreatic Diseases International 21 (2022) 356–361 - “Chu et al. used radiomic features of CT images to differentiate pancreatic adenocarcinoma and normal pancreatic tissues in a series of patients with a radiological and pathological diagnosis, and the study included a training cohort and a validation cohort. Accuracy, sensitivity and specificity were calculated. Patients were classified with a sensitivity of 100% and a specificity of 98.5%. This would allow a more precise definition of tumor areas, which is very important to local treatment strategies.”
Radiomics in pancreatic cancer for oncologist: Present and future
Carolina de la Pinta
Hepatobiliary & Pancreatic Diseases International 21 (2022) 356–361 - “Dmitriev et al. differentiated four types of cysts by com- bining demographic variables with radiomic characteristics of in- tensity and shape, achieving differentiation of 84% of the lesions. Wei et al. analyzed cyst images in preoperative tests to differentiate SCNs from other pancreatic cystic lesions (PCLs) includ- ing 17 intensity and texture features (T-range, wavelet intensity, T-median, and wavelet neighbourhood gray-tone difference matrix busyness) and clinical features. Adequate classification was achieved in 76% of patients and 84% in a validation cohort of 60 patients. Yang et al. evaluated variable slice images, 2 and 5 mm, without affecting feature extraction. In the validation group the accuracy was 74% in patients with 2-mm slice and 83% in 5- mm slice. ”
Radiomics in pancreatic cancer for oncologist: Present and future
Carolina de la Pinta
Hepatobiliary & Pancreatic Diseases International 21 (2022) 356–361 - “Yamashita et al. demonstrated that differences in contrast- enhanced CT acquisition affected the results of the radiomic study leading to changes in segmentation and its reproducibility and comparability between series . The study did not demonstrate statistically significant differences in CT model, pixel spacing, and contrast administration ratio. The study suggests that radiologists are more or less sensitive to CT acquisition parameters, demonstrating the importance of adjusting for these variables to established protocols. Furthermore, this study support the hypothesis of the usefulness of a semi-automated segmentation tool previously trained by several radiologists that can homogenize these varia- tions. Standardization of protocols is therefore important, in addition to external validation. Also many of the comparisons between diagnostic entities using radiomics are subjective and not clinically applicable. For example, the distinction between pancreatic adenocarcinoma and pancreatic neuroendocrine tumors alone.”
Radiomics in pancreatic cancer for oncologist: Present and future
Carolina de la Pinta
Hepatobiliary & Pancreatic Diseases International 21 (2022) 356–361 - ”Radiomics is a promising non-invasive tool for the diagnosis and clinical management of pancreatic tumors. The usefulness of radiomics has been studied in the differential diagnosis of benign, premalignant and malignant lesions in the pancreas. In addition, in patients with neoadjuvant pancreatic cancer, it can help in the more precise definition of lesions for radiotherapy and assessment of response. Radiomics provides a more adequate and reproducible measurement of the tumor than other methods. In addition, the combination of radiomics and genomics has a promising future. However, image acquisition protocols and radiomic analysis sys- tems need to be standardized and validation cohorts are needed. Further studies are needed to consolidate the available data.”
Radiomics in pancreatic cancer for oncologist: Present and future
Carolina de la Pinta
Hepatobiliary & Pancreatic Diseases International 21 (2022) 356–361
- “This scoping review has provided evidence that 12 artificial intelligence-based machine learning models have sufficient capacity to evaluate the risk of malignancy in IPMN. However, the methodological quality of the included studies is inadequate, and the clinical value of the proposed models has not been proven. As a result, caution should be advised when interpreting these results, and the findings must be corroborated by additional high-quality studies. Future research should focus on developing rigorous models and investigating their usefulness in clinical practice to ensure that they are dependable tools for assessing the risk of malignancy in IPMN.”
Artificial intelligence-based models to assess the risk of malignancy on radiological imaging in patients with intraductal papillary mucinous neoplasm of the pancreas: scoping review
Alberto Balduzzi et al.
Br J Surg. 2023 Jul 4:znad201. doi: 10.1093/bjs/znad201. (in press) - “Most patients diagnosed with IPMN will be kept under surveillance, aimed at monitoring progression of the cyst, which may require surgical resection in highly selected patients. Still, the risk of clinicians missing IPMN progression to malignancy is concerning5, with burdensome consequences for the patient. This concern must be balanced against the risk of complications after major pancreatic surgery. Therefore, patient selection is crucial both to avoid unnecessary surgery for benign lesions, and to continue surveillance safely. Typically, diagnostic imaging plays a central role in guiding patient selection for, and the timing of, surgery. However, current imaging approaches fall short for optimal decision-making.”
Artificial intelligence-based models to assess the risk of malignancy on radiological imaging in patients with intraductal papillary mucinous neoplasm of the pancreas: scoping review
Alberto Balduzzi et al.
Br J Surg. 2023 Jul 4:znad201. doi: 10.1093/bjs/znad201. (in press) - “Future research should concentrate on developing methodologically sound, generalizable, and clinically validated models. Multiple methodological elements are frequently missed or ignored, as is evident from the mRQS scores of the research included. Once robust and generalizable models have been constructed, their performance and value should be validated in clinical settings. Currently available studies have focused on assessing the discriminative performance of machine learning models for malignant IPMNs. However, ideally, models would exclude the presence of malignancy with a high negative predictive value and ‘safely’ advise surveillance in patients who would have been selected for surgical treatment according to current criteria. This would represent a true added value to current clinical practice.”
Artificial intelligence-based models to assess the risk of malignancy on radiological imaging in patients with intraductal papillary mucinous neoplasm of the pancreas: scoping review
Alberto Balduzzi et al.
Br J Surg. 2023 Jul 4:znad201. doi: 10.1093/bjs/znad201. (in press)
- “Rapid advances in automated methods for extracting large numbers of quantitative features from medical images have led to tremendous growth of publications reporting on radiomic analyses. Translation of these research studies into clinical practice can be hindered by biases introduced during the design, analysis, or reporting of the studies. Herein, the authors review biases, sources of variability, and pitfalls that frequently arise in radiomic research, with an emphasis on study design and statistical analysis considerations. Drawing on existing work in the statistical, radiologic, and machine learning literature, approaches for avoiding these pitfalls are described.”
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press) - Summary
This review highlights biases and inappropriate methods used in Radiomic research that can lead to erroneous conclusions; address in these issues will accelerate translation of research to clinical practice and have the potential to positively impact patient care.
Essentials
• Many radiomic research studies are hindered by systematic biases.
• In addition to ongoing initiatives for standardization, improvements in study design, data collection, rigorous statistical analysis,and thorough reporting are needed in radiomic research.
• Insight into potential problems and suggestions for how to circumvent common pitfalls in radiomic studies are provided.
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press) - “It is not always possible to safeguard against all potential sources of study bias in radiomics research. Therefore, it is imperativethat researchers thoroughly report on their imaging data (ie,Digital Imaging and Communications in Medicine [DICOM]header information), methodology, limitations, and any other potential sources of variability. Rigorous reporting enablesresearchers to build on others’ results and protects against failed attempts to replicate spurious and overstated results. For instance, Eslami et al included a detailed description of their methodology in their supplementary material.”
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press) - “Radiomic analyses are highly susceptible to bias arising from multiple sources. A unifying theme behind the biases and pitfalls we have outlined is that they can all lead to incorrect inference and a model that erroneously includes or excludes imaging features and, ultimately, performs poorly. While not meant to be an allencompassing list, the issues we have highlighted arise frequently. Some, such as overfitting and lack of adjusting for multiple testing, are particularly relevant in radiomic studies. Others are issues that may arise equally as frequently in other types of studies but have been highlighted here because we have noticed a lack of awareness of these issues among investigators conducting radiomic studies. All are issues that are broadly applicable to many studies, including those where features are derived by the computer using convolution neural network (deep) approaches. In any analysis, the challenge is to identify the most relevant sources of bias and measurement error.”
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press) - "Although software packages to implement analyses are readily available and increasingly user friendly, if they are not implemented with the necessary expertise or correct guidance, there is a high risk that incorrect conclusions will be drawn from the work. The field of radiomics lies at the intersection of medicine, computer science, and statistics. We contend that to produce clinically meaningful results that positively impact patient care and minimize biases and pitfalls, radiomic analysis requires a multidisciplinary approach with a research team that includes individuals with multiple areas of expertise.”
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press) 
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press)
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press)
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press)
Radiomic Analysis: Study Design, Statistical Analysis, and Other Bias Mitigation Strategies
Chaya S. Moskowitz et al.
Radiology 2022; (in press)
- Purpose: To develop and validate an effective model to differentiate NF-pNET from PDAC.
Conclusion: The integrated model outperformed the model based on clinicoradiological features alone and was comparable to the model based on the radiomic signature alone with respect to the differential diagnosis of atypical NF-pNET and PDAC. The nomogram achieved an optimal preoperative, noninvasive differential diagnosis between atypical pNET and PDAC, which can better inform therapeutic choice in clinical practice.
Differentiation of atypical non-functional pancreatic neuroendocrine tumor T and pancreatic ductal adenocarcinoma using CT based radiomics
Ming He et al.
European Journal of Radiology 117 (2019) 102–111 - “The therapeutic strategies and prognoses differ significantly between these two major pancreatic solid lesion subtypes, which make the correct differentiation of PDAC from pNET a major issue in clinical practice, especially for atypical cases. For pNET, enucleation is possible, and patients with liver metastasis and with preoperative vascular abutment or invasion can still benefit from surgical resection. For PDAC, more radical surgery is needed, which entails higher post-operative complications and risks; surgery is contraindicated for pa- tients with liver metastasis or vascular invasion.”
Differentiation of atypical non-functional pancreatic neuroendocrine tumor T and pancreatic ductal adenocarcinoma using CT based radiomics
Ming He et al.
European Journal of Radiology 117 (2019) 102–111 - In the present study, our hypothesis was that a radiomics-based model represented with a nomogram that integrated clinicoradiological features and the radiomic signature would improve the differential diagnostic performance between atypical NF-pNET and PDAC, which is difficult to achieve in clinical practice. Therefore, we aimed to develop and validate an effective model and represent it with a nomogram to differentiate NF-pNET from PDAC.
Differentiation of atypical non-functional pancreatic neuroendocrine tumor T and pancreatic ductal adenocarcinoma using CT based radiomics
Ming He et al.
European Journal of Radiology 117 (2019) 102–111 
Differentiation of atypical non-functional pancreatic neuroendocrine tumor T and pancreatic ductal adenocarcinoma using CT based radiomics
Ming He et al.
European Journal of Radiology 117 (2019) 102–111- In conclusion, the integrated model outperformed the model based on the clinicoradiological features alone and performed comparably to the model based on the radiomic signature alone in the differential diagnosis of atypical NF-pNET versus PDAC. The nomogram achieved an optimal preoperative, noninvasive differential diagnosis between atypical pNET and PDAC, which facilitates informed therapeutic choices in clinical practice.
Differentiation of atypical non-functional pancreatic neuroendocrine tumor T and pancreatic ductal adenocarcinoma using CT based radiomics
Ming He et al.
European Journal of Radiology 117 (2019) 102–111 - Objectives: To assess whether radiomics signature can identify aggressive behavior and predict recurrence of hepatocellular carcinoma (HCC) after liver transplantation.
Results: The stable radiomics features associated with the recurrence of HCC were simply found in arterial phase and portal phase. The prediction model based on the radiomics features extracted from the arterial phase showed better prediction performance than the portal vein phase or the fusion signature combining both of arterial and portal vein phase. A radiomics nomogram based on combined model consisting of the radiomics signature and clinical risk factors showed good predictive performance for RFS with a C-index of 0.785 (95% confidence interval [CI]: 0.674-0.895) in the training dataset and 0.789 (95% CI: 0.620-0.957) in the validation dataset. The calibration curves showed agreement in both training (p = 0.121) and validation cohorts (p = 0.164).
Conclusions: Radiomics signature extracted from CT images may be a potential imaging biomarker for liver cancer invasion and enable accurate prediction of HCC recurrence after liver transplantation.
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - “Microvascular invasion (MVI) and consequently, early tumor recurrence, is noted in 20% of patients undergoing liver transplantation, reducing the 5-year survival from 80% to 40% after liver transplantation . Recurrence has been the main factor that affects the curative effect of HCC after liver transplantation. Prospective identification of recurrence in patients with HCC thus has direct implications for organ allocation, surgical techniques, prognosis, and public policy.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - “Previous reports have indicated that some radiomic features extracted from tumor regions of CT images, especially characteristics of arterial and portal phases, can surpass traditional indicators such as Barcelona Clinic Liver Cancer (BCLC) for assessing the efficacy of HCC hepatectomy. These high-throughput characteristics related to MVI (Microvascular invasion) or prognosis of HCC may be potential independent predictors of HCC recurrence after liver transplantation.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - Radiomics signature was built with the rad-score of arterial phase. A combined predictive model was built by incorporating the clinical risk factors and radiomics signature with multivariable Cox regression model. Patients were finally stratified into high-risk and low-risk groups based on the combined model with cut-off values at the median of the training dataset in arterial phase. The Log- rank test was computed to compare the two separated KM survival curves.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - “A multi-feature-based radiomics signature was identified to be an effective biomarker for the prediction of HCC recurrence after liver transplantation in this study, with potential prognosis value for individualized RFS. The radiomics signature successfully stratified patients with HCC into high-risk and low-risk groups, with significant differences in RFS. Furthermore, the radiomics nomogram based on the combined model which integrates effective clinical characteristics and radiomics signature showed good discrimination and prominent predictive performance for RFS in HCC patients after liver transplantation.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - “For patients with high recurrence, liver transplantation is not recommended, so that the scarce donor liver resources can be allocated to patients who urgently need liver transplantation and have a good prognosis.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - “ It is conducive to fully exploit potential images using artificial intelligence and big data technology to overcome limitations of traditional evaluation methods, improve the predictive performance of HCC recurrence after liver transplantation, facilitate the rational distribution of donors in clinical practice, and conduct neces- sary intraoperative or postoperative prophylactic treatment for patients with high recurrence.”
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 - Although the usefulness of the proposed nomogram lacks external validation, calibration curve analysis demonstrated that the radiomics nomogram was consistent with actual observations for the probability of tumor recurrence at 1, 2, or 3 years. This indicates that radiomics signature has the potential to be used as a biomarker for recurrence in HCC patients after liver transplantation. Further, radiomics can be used to predict MVI and histopathological differentiation closely related to prognosis of HCC. Radiomics technique is expected to become an important cornerstone of accurate diagnosis and treatment of HCC.
Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation
Donghui Guo et al.
European Journal of Radiology 117 (2019) 33–40 
- Purpose The aim of this study was to investigate the relationship between CT imaging phenotypes and genetic and biological characteristics in pancreatic ductal adenocarcinoma (PDAC).
Conclusions In this study, we demonstrate that in PDAC SMAD4 status and tumor stromal content can be predicted using radiomic analysis of preoperative CT imaging. These data show an association between resectable PDAC imaging features and underlying tumor biology and their potential for future precision medicine.
CT radiomics associations with genotype and stromal content in pancreatic ductal adenocarcinoma
Marc A. Attiyeh et al.
Abdominal Radiology (in press) https://doi.org/10.1007/s00261-019-02112-1 - “In this study, we identified radiomic features associated with PDAC genetic alterations and stromal content. These associations show the potential of using noninvasive imaging on pre-surgical pancreas cancer patients for precision medicine. Linking radiomic features to underlying tumor biology is an area of great interest, given the ubiquity of diagnostic imaging and the challenges and costs in performing molecular analyses. Recent DNA/RNA sequencing studies have provided in-depth insights into individual tumor’s genetic makeup and have demonstrated some prognostic power for survival that may guide personalized treatment.”
CT radiomics associations with genotype and stromal content in pancreatic ductal adenocarcinoma
Marc A. Attiyeh et al.
Abdominal Radiology (in press) https://doi.org/10.1007/s00261-019-02112-1 - “In conclusion, we demonstrate that radiomic features extracted from clinical CT images are associated with genotype, the number of altered genes, and stromal content in PDAC. These associations may underlie the observation that PDAC imaging features are associated with survival. Further studies will be needed to increase sample size and perform external validation."
CT radiomics associations with genotype and stromal content in pancreatic ductal adenocarcinoma
Marc A. Attiyeh et al.
Abdominal Radiology (in press) https://doi.org/10.1007/s00261-019-02112-1 - OBJECTIVE. The purpose of this study is to develop and evaluate an unenhanced CT–based radiomics model to predict brain metastasis (BM) in patients with category T1 lung adenocarcinoma.
CONCLUSION. A CT-based radiomics model presented good predictive performance and great potential for predicting BM in patients with category T1 lung adenocarcinoma. As a promising adjuvant tool, it can be helpful for guiding BM screening and thus benefiting personalized surveillance for patients with lung cancer.
CT-Based Radiomics Model for Predicting Brain Metastasis in Category T1 Lung Adenocarcinoma
Chen A et al.
AJR 2019; 213:134–139 - “Non–small cell lung cancer (NSCLC) accounts for approximately 80–85% of all lung cancers. Adenocarcinoma (ADC) is the most common histologic subtype of NSCLC, accounting for almost 50% of all lung cancer. The prognosis of lung cancer mostly depends on the TNM system, which may determine therapy. The brain is one of the most frequent sites of extrathoracic metastases, and brain metastasis (BM) is a very important prognostic factor for the M category.”
CT-Based Radiomics Model for Predicting Brain Metastasis in Category T1 Lung Adenocarcinoma
Chen A et al.
AJR 2019; 213:134–139 - “With the rapid development of radiomics in oncology, many predictive radiomics models have been developed to predict the preoperative stage and prognosis of various malignant tumors, and all these radiomics signatures implied superior prediction po- tential. To our knowledge, the association between CT-based radiomics features and the risk factors of BM in category T1 ADC has not been established.”
CT-Based Radiomics Model for Predicting Brain Metastasis in Category T1 Lung Adenocarcinoma
Chen A et al.
AJR 2019; 213:134–139 - “In conclusion, the CT-based radiomics model had a powerful predictive performance and great potential for predicting BM in patients with category T1 lung ADC. As a promising adjuvant tool, it could guide therapeutic strategies for and personalized surveillance of patients.”
CT-Based Radiomics Model for Predicting Brain Metastasis in Category T1 Lung Adenocarcinoma
Chen A et al.
AJR 2019; 213:134–139 - Background: Intratumor heterogeneity in lung cancer may influence outcomes. CT radiomics seeks to assess tumor features to provide detailed imaging features. However, CT radiomic features vary according to the reconstruction kernel used for image generation.
Purpose: To investigate the effect of different reconstruction kernels on radiomic features and assess whether image conversion using a convolutional neural network (CNN) could improve reproducibility of radiomic features between different kernels.
Conclusion: Chest CT image conversion using a convolutional neural network effectively reduced the effect of two different reconstruction kernels and may improve the reproducibility of radiomic features in pulmonary nodules or masses.
Deep Learning–based Image Conversion of CT Reconstruction Kernels Improves Radiomics Reproducibility for Pulmonary Nodules or Masses
Choe J et al.
Radiology 2019; 00:1–9 • https://doi.org/10.1148/radiol.2019181960 - * For evaluation of pulmonary nodules and masses, reproducibility of radiomic features using different CT reconstruction kernels is low (concordance correlation coefficient [CCC], 0.38). Only 107 of 702 features (15.2%; by excluding wavelet features, 16 of 78 [20.5%]) showed a CCC of 0.85 or higher.
* Texture and wavelet features are more affected than tumor intensity features by different kernels (CCC, 0.82 6 0.17 for tumor intensity features, 0.61 6 0.17 for texture features, and 0.35 6 0.33 for wavelet features) because these features are sensitive to changes in image spatial and density resolutions.
Deep Learning–based Image Conversion of CT Reconstruction Kernels Improves Radiomics Reproducibility for Pulmonary Nodules or Masses
Choe J et al.
Radiology 2019; 00:1–9 • https://doi.org/10.1148/radiol.2019181960 
Deep Learning–based Image Conversion of CT Reconstruction Kernels Improves Radiomics Reproducibility for Pulmonary Nodules or Masses
Choe J et al.
Radiology 2019; 00:1–9 • https://doi.org/10.1148/radiol.2019181960- “Radiomics converts medical images such as CT scans into high-throughput quantitative data that can be used to improve diagnostic, prognostic, and predictive accuracy. Indeed, radiomics can provide objective and comprehensive information regarding whole or subregions of cancers in a noninvasive and repeatable way. Radiomic signatures have been demonstrated to reflect intratumoral heterogeneity and to be associated with gene-expression profiles, both of which can serve as important prognostic factors.”
Can Artificial Intelligence Fix the Reproducibility Problem of Radiomics?
Park CM
Radiology 2019; 00:1–2 • https://doi.org/10.1148/radiol.2019191154 - Radiomics: Feature Variability can be due to;
- image reconstruction algorithms
- Slice thickness (.75mm vs 3mm vs 5mm)
- Use of iodinated contrast agent
- Delivery of iodinated contrast agent ( injection rate, contrast volume and timing of acquisition) - Although we previously believed that different reconstruction kernels could not be used interchangeably in radiomics research, Choe and colleagues showed that it may be possible to compare radiomics features from CT images with different reconstruction kernels using this novel CNN-based kernel- conversion algorithm.”
Can Artificial Intelligence Fix the Reproducibility Problem of Radiomics?
Park CM
Radiology 2019; 00:1–2 • https://doi.org/10.1148/radiol.2019191154 - “In conclusion, a convolutional neural network–based kernel-conversion algorithm dramatically improved the similarity of CT radiomic features obtained using different reconstruction kernels. Deep learning–based approaches are expected to substantially contribute to the applicability of radiomics. The study by Choe et al will serve as one of the first steps toward bigger strides in radiomics.
Can Artificial Intelligence Fix the Reproducibility Problem of Radiomics?
Park CM
Radiology 2019; 00:1–2 • https://doi.org/10.1148/radiol.2019191154 - Objective To develop and validate a radiomics-based nomogram for preoperatively predicting grade 1 and grade 2/3 tumors in patients with pancreatic neuroendocrine tumors (PNETs). Results The fusion radiomic signature has significant association with histologic grade (p < 0.001). The nomogram integrating independent clinical risk factor tumor margin and fusion radiomic signature showed strong discrimination with an area under the curve (AUC) of 0.974 (95% CI 0.950–0.998) in the training cohort and 0.902 (95% CI 0.798–1.000) in the validation cohort with good calibration. Decision curve analysis verified the clinical usefulness of the predictive nomogram.
Conclusion We proposed a comprehensive nomogram consisting of tumor margin and fusion radiomic signature as a powerful tool to predict grade 1 and grade 2/3 PNET preoperatively and assist the clinical decision-making for PNET patients.
CT radiomics may predict the grade of pancreatic neuroendocrine tumors: a multicenter study
Gu D et al.
European Radiology June 2019: https://doi.org/10.1007/s00330-019-06176-x - Key Points
• Radiomic signature has strong discriminatory ability for the histologic grade of PNETs.
• Arterial and portal venous phase CT imaging are complementary for the prediction of PNET grading.
• The comprehensive nomogram outperformed clinical factors in assisting therapy strategy in PNET patients.
CT radiomics may predict the grade of pancreatic neuroendocrine tumors: a multicenter study
Gu D et al.
European Radiology June 2019: https://doi.org/10.1007/s00330-019-06176-x - ”Thus, in this multicenter study, we build a radiomic-based predictive model to noninvasively and operatively achieve PNET grading using CT images. Meanwhile, we would also explore the predictive value of clinical and radiological variables, as comparisons with the radiomic signature. A final combined model integrating both radiomic and clinical factors is expected to accurately classify PNET grading.”
CT radiomics may predict the grade of pancreatic neuroendocrine tumors: a multicenter study
Gu D et al.
European Radiology June 2019: https://doi.org/10.1007/s00330-019-06176-x 
CT radiomics may predict the grade of pancreatic neuroendocrine tumors: a multicenter study
Gu D et al.
European Radiology June 2019: https://doi.org/10.1007/s00330-019-06176-x- ”For the fusion radiomic signature, we build a multivariable logistic regression model using the two single-phase radiomic signatures. The fusion radiomic signature outperformed either of the single-phase radiomic signatures. Potential reasons for this finding may be that the combination of the two phases could show the vascularity of PNETs more accurately than only one phase. The fusion signature could also provide more textural information in the tumor microenvironments since the most effective features from the two phases in this study were texture features.”
CT radiomics may predict the grade of pancreatic neuroendocrine tumors: a multicenter study
Gu D et al.
European Radiology June 2019: https://doi.org/10.1007/s00330-019-06176-x 
- Objectives: The primary aim of this study was to determine if computed tomographic (CT) texture analysis measurements of the tumor are independently associated with progression-free survival (PFS) and overall survival (OS) in patients with unresectable pancreatic ductal adenocarcinoma (PDAC), including both unresectable locally advanced and metastatic PDAC, who were treated with chemotherapy.
Conclusions: Pretreatment CT quantitative imaging biomarkers from texture analysis are associated with PFS and OS in patients with unresectable PDAC who were treated with chemotherapy, and the combination of pre- treatment texture parameters and tumor size have the potential to perform better in survival models than imaging biomarker alone.
Unresectable pancreatic ductal adenocarcinoma: Role of CT quantitative T imaging biomarkers for predicting outcomes of patients treated with chemotherapy
Cheng, Si-Hang et al.
European Journal of Radiology, Volume 113, 188 - 197 - “This may reveal subtle differences imperceptible to the naked eye, thereby compensating for the limitations of conventional CT . CT texture analysis relies on objective computer-aided evaluation of gray-level patterns within lesions to assess tumor heterogeneity quantitatively in terms of numerous relevant parameters, which has been used in the prediction of various cancer prognosis.”
Unresectable pancreatic ductal adenocarcinoma: Role of CT quantitative T imaging biomarkers for predicting outcomes of patients treated with chemotherapy
Cheng, Si-Hang et al.
European Journal of Radiology, Volume 113, 188 - 197 - ”The CT texture parameter measured in this study included (1) mean gray-level intensity (Mean, brightness); (2) standard deviation (SD, spread of distribution); (3) entropy (irregularity or complexity of pixel intensity in space); (4) mean of positive pixels (MPP); (5) skewness (symmetry of the pixel intensity distribution); (6) kurtosis (sharpness or pointedness of the pixel intensity distribution).”
Unresectable pancreatic ductal adenocarcinoma: Role of CT quantitative T imaging biomarkers for predicting outcomes of patients treated with chemotherapy
Cheng, Si-Hang et al.
European Journal of Radiology, Volume 113, 188 - 197 - In conclusion, instead of post-chemotherapy texture parameters or Δ value, pre-chemotherapy could provide more information about tumor biology. Therefore, using pre-chemotherapy texture of unresectable PDAC to predict survival is more accurate and reliable. Furthermore, texture analysis as a noninvasive image-processing tool has the potential to select patients with good prognosis before therapy, indicating a promising prospect of clinical application in the future.
Unresectable pancreatic ductal adenocarcinoma: Role of CT quantitative T imaging biomarkers for predicting outcomes of patients treated with chemotherapy
Cheng, Si-Hang et al.
European Journal of Radiology, Volume 113, 188 - 197 - “In our study, SD has been demonstrated to be closely associated with both PFS and OS, and higher SD, which indicated higher intratumoral heterogeneity, predicted better survival outcome in patients with unresectable PDAC. However, in many cancers, increased tumor heterogeneity is associated with worse outcomes. Hypoxia and necrosis, correlated with impaired response to chemotherapy and radiotherapy, are likely to occur in tumors with low levels of angiogenesis, which were closely associated with SD value. In addition, tumor necrosis, which can reflect the presence of hypoxia, was in- vestigated by previous study to verify its significant value in predicting outcome in patients with PDAC, and multivariate survival analysis showed that necrosis was an independent predictor of poor outcome in terms of both disease-free survival (DFS) and disease-specific survival.”
Unresectable pancreatic ductal adenocarcinoma: Role of CT quantitative T imaging biomarkers for predicting outcomes of patients treated with chemotherapy
Cheng, Si-Hang et al.
European Journal of Radiology, Volume 113, 188 - 197 - Purpose: Grades of pancreatic neuroendocrine neoplasms (PNENs) are associated with the choice of treatment strategies. Texture analysis has been used in tumor diagnosis and staging evaluation. In this study, we aim to evaluate the potential ability of texture parameters in differentiation of PNENs grades.
Results: There were significant differences in tumor margin, pancreatic duct dilatation, lymph nodes invasion, size, portal enhancement ratio (PER), arterial enhancement ratio (AER), mean grey-level intensity, kurtosis, entropy, and uniformity among G1, G2, and pancreatic neuroendocrine carcinoma (PNEC) G3 (p < 0.01).
Conclusions: Our data indicated that texture parameters have potential in grading PNENs, in particular in differentiating PNEC G3 from PNETs G1/G2.
Textural analysis on contrast-enhanced CT in pancreatic neuroendocrine neoplasms: association with WHO grade
Chuangen Guo et al.
Abdom Radiol (2019) 44:576–585 - “The texture parameters included mean grey-level inten- sity, kurtosis, skewness, entropy, and uniformity which had been used in previous reports. The mean of each parameter was calculated automatically. The mathematical expression and means of texture parame- ters were reported in the previous study. The inter- observer agreements in ROIs were assessed with Conger’s kappa test. For those cases with interobserver disagreement, texture analysis was not performed until the two readers reach a consensus.”
Textural analysis on contrast-enhanced CT in pancreatic neuroendocrine neoplasms: association with WHO grade
Chuangen Guo et al.
Abdom Radiol (2019) 44:576–585 
Textural analysis on contrast-enhanced CT in pancreatic neuroendocrine neoplasms: association with WHO grade
Chuangen Guo et al.
Abdom Radiol (2019) 44:576–585- “The higher standard deviations in texture parameters were present in PNET G2 and PNEC. PNET G2 and PNEC showed more heterogeneous than PNET G1 on contrast-enhanced imaging [12]. PNET G2 usually showed marked enhancement. Mild enhancement was also found in PNETs G2. PNEC usually showed mild enhancement. However, marked enhancement also can be found in PNEC. Therefore, the standard deviation is higher in PNET G2/PNEC than PNET G1.”
Textural analysis on contrast-enhanced CT in pancreatic neuroendocrine neoplasms: association with WHO grade
Chuangen Guo et al.
Abdom Radiol (2019) 44:576–585 - In conclusion, our data indicate that tumor size, pancreatic duct dilation, local invasion/metastases, AER, and PER have potential for differentiating PNEC G3 from PNET G1/G2. Moreover, our data indicate that texture analysis on contrast-enhanced CT images may represent as promising, non-invasive biomarkers to evaluate the pathologic grade of PNENs.”
Textural analysis on contrast-enhanced CT in pancreatic neuroendocrine neoplasms: association with WHO grade
Chuangen Guo et al.
Abdom Radiol (2019) 44:576–585 - Objectives: This study was designed to estimate the performance of textural features derived from contrast- enhanced CT in the differential diagnosis of pancreatic serous cystadenomas and pancreatic mucinous cystadenomas.
Conclusions: In conclusion, our study provided preliminary evidence that textural features derived from CT images were useful in differential diagnosis of pancreatic mucinous cystadenomas and serous cystadenomas, which may provide a non-invasive approach to determine whether surgery is needed in clinical practice. However, multicentre studies with larger sample size are needed to confirm these results.
Discrimination of Pancreatic Serous Cystadenomas From Mucinous Cystadenomas With CT Textural Features: Based on Machine Learning
Jing Yang et al.
Front Oncol. 2019 Jun 12;9:494. doi: 10.3389/fonc.2019.00494 - Methods: Fifty-three patients with pancreatic serous cystadenoma and 25 patients with pancreatic mucinous cystadenoma were included. Textural parameters of the pancreatic neoplasms were extracted using the LIFEx software, and were analyzed using random forest and Least Absolute Shrinkage and Selection Operator (LASSO) methods. Patients were randomly divided into training and validation sets with a ratio of 4:1; random forest method was adopted to constructed a diagnostic prediction model. Scoring metrics included sensitivity, specificity, accuracy, and AUC.
Results: Radiomics features extracted from contrast-enhanced CT were able to discriminate pancreatic mucinous cystadenomas from serous cystadenomas in both the training group (slice thickness of 2 mm, AUC 0.77, sensitivity 0.95, specificity 0.83, accuracy 0.85; slice thickness of 5 mm, AUC 0.72, sensitivity 0.90, specificity 0.84, accuracy 0.86) and the validation group (slice thickness of 2 mm, AUC 0.66, sensitivity 0.86, specificity 0.71, accuracy 0.74; slice thickness of 5 mm, AUC 0.75, sensitivity 0.85, specificity 0.83, accuracy 0.83).
Discrimination of Pancreatic Serous Cystadenomas From Mucinous Cystadenomas With CT Textural Features: Based on Machine Learning
Jing Yang et al.
Front Oncol. 2019 Jun 12;9:494. doi: 10.3389/fonc.2019.00494 
Discrimination of Pancreatic Serous Cystadenomas From Mucinous Cystadenomas With CT Textural Features: Based on Machine Learning
Jing Yang et al.
Front Oncol. 2019 Jun 12;9:494. doi: 10.3389/fonc.2019.00494- “The textural parameters were obtained using Local Image features Extraction (LIFEx) software in the portal vein phase CT images. A two-dimensional region of interest (ROI) was delineated around the boundary of tumor lesion in each layer of transaxial CT images to form a three-dimensional ROI. ROIs were drawn independently by two radiologists, who were unaware of the diagnosis of patients. Minimum, maximum, mean, and standard deviation of the density values inside the ROI were calculated. From these primary calculations, geometry based and histogram based features, the Gray-level co- occurrence matrix (GLCM), the Neighborhood gray-level different matrix (NGLDM), the Gray level run length matrix (GLRLM) and the Gray level zone length matrix (GLZLM) were obtained.”
Discrimination of Pancreatic Serous Cystadenomas From Mucinous Cystadenomas With CT Textural Features: Based on Machine Learning
Jing Yang et al.
Front Oncol. 2019 Jun 12;9:494. doi: 10.3389/fonc.2019.00494
- “Advances in the management of genitourinary neoplasms have resulted in a trend towards providing patients with personalized care. Texture analysis of medical images, is one of the tools that is being explored to provide information such as detection and characterization of tumors, determining their aggressiveness including grade and metastatic potential and for prediction of survival rates and risk of recurrence. In this article we review the basic principles of texture analysis and then detail its current role in imaging of individual neoplasms of the genitourinary system.”
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - "Radiogenomics, which has been defined as a science of identifying associations between imaging features and genomic characteristics of a disease, assumes great importance in this era of precision medicine. There is an increasing interest in providing quantitative information along with subjective interpretation of radiologic images.”
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - Texture segmentation is a technique to divide the image into various segments that differ in their broad textural properties. This can help delineate anatomical structures or pathological lesions and is often a part of computer aided diagnosis. Examples include lung nodule and colonic polyp detection. Some of the face detection softwares in our everyday social media platforms also use this technique.”
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - “Texture analysis seeks to provide more detailed information of the textural properties of the image, usually within a region of interest drawn by the observer. While several methods of texture analysis exist, the most commonly employed ones are based on statistical analysis of gray scale values of the pixels in the image. In this regard, it is useful to understand some of the basic principles of these statistical methods.”
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - Texture Analysis
- First order statistics
- Second order statistics
- Higher order statistics - First Order Statistics
These methods take into consideration individual pixel intensity values and do not analyze the relationships between different pixels. This usually starts with a histogram that plots gray scale values of individual pixels against their number/percentage. A variety of information such as mean, standard deviation (SD), variation, and percentiles can be obtained from the graph obtained. Mean of positive pixels (MPP) is a self-explanatory term that refers to the mean of pixels with value >0. In addition, features such as skew and kurtosis of the graph can also be evaluated. Skew is a measure of asymmetry of the distribution of variables about the mean. A positively skewed graph tapers (tail) to the right side and a negatively skewed graph tapers (tail) to the left side. Kurtosis is a measure of the sharpness of the peak of a distribution curve. Higher Kurtosis denotes a sharper peak and vice versa. - Second Order Statistics
These methods analyze the inter- relationship or co-occurrence of pixel intensities. Several methods exist to analyze second order statistics, a few of which are summarized below:
- Gradient A gradient image plots the difference in gray scale values between adjacent pixels. A higher change in pixel intensity results in a higher gradient value at that point, and a lower change in a lower value. From this image, one can again estimate first order statistical infor- mation like mean, SD, etc.
- Run-length matrix (RLM) This involves estimating how many times one can find groups of pixels that have the same gray scale value, in a specified direction, for a specified group size. For example, in Fig. 3, in the vertical direction, pixel value 1 has occurred thrice by itself and once in group sizes of two and three.
- Gray-level co-occurrence matrix (GLCM) This involves calculating how many times different pixel values occur next to each other in a specified direction. For example, in Fig. 4, in the vertical direction, pixel value of 1 is followed by 2 three times whereas pixel value 1 is followed by 3 two times. 
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510- How do you do textural analysis?
One of the inherent advantages of texture analysis is the ability to apply it retrospectively to any image. Both commercially available software and ones that are devel- oped within the institution are used. There is no need to modify the imaging protocols or techniques. A filter may be applied to the image with lower filter settings corresponding to fine textural features and higher filter setting corresponding to medium or coarse textural features. Usually the software allows the user to select the filter settings or may provide textural features at every filter setting. A region of interest is manually drawn over the pathology and the software computes and displays the results. These results are then associated with multiple parameters. - Textural Analysis for Renal Pathology: Applications
- Renal neoplasms (RCC vs AML)
- Renal neoplasms (Clear Cell vs. Papillary)
- Response to chemotherapy
- Predicting survival - Limitations of texture analysis
Texture analysis, like the field of radiogenomics in general, must deal with a problem of plenty. The numerous and varied parameters that can be assessed, combined with the small sample size in several of these studies, makes analysis both challenging and difficult to interpret. For the same reasons, these studies are prone to type I errors. Type I error is to falsely imply an association that does not exist. Chalkidou et el applied a statistical analysis method to fifteen studies that reported an association between patient outcomes and texture features on PET and CT, and identified probability of type I error to be as high as 76%. - Limitations of texture analysis
- Scanner variability
- Scan parameters including;
-- slice thickness
-- reconstruction filters
-- Voxel size and gray levels
-- radiation dose
-- type of image reconstruction method used (filtered back projection or iterative reconstruction) - An increasingly exciting aspect of medical imaging is artificial intelligence. This has been varyingly described both as a threat to the field of radiology and as a valuable tool for radiologists. Nevertheless computer aided detection and machine learning are here to stay and are likely to play an important role in imaging in the years to come. Texture analysis is one of the many tools used by the deep learning processes of artificial intelligence and in a matter of time it may be incorporated into clinical care, along with other quantitative imaging methods.
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - The discipline of radiology continues to evolve, paralleling advances in technology and clinical medicine. In the field of genitourinary oncology, texture analysis has been shown to aid in the detection of neoplasms, in the characterization of their pathological subtypes and grades, in predicting survival rates and in estimating risk of recurrence. Texture analysis is one essential tool that we must arm ourselves with as we endeavor to detect the unseen and continue to serve as the eye of medicine.
A review of the principles of texture analysis and its role in imaging of genitourinary neoplasms
Richard Thomas et al.
Abdominal Radiology (2019) 44:2501–2510 - Purpose To evaluate the feasibility of using computed tomography texture analysis (CTTA) parameters for predicting malignant risk grade and mitosis index of gastrointestinal stromal tumors (GISTs), compared with visual inspection.
Conclusion Using TexRAD, MPP and kurtosis are feasible in predicting risk grade and mitosis index of GISTs. CTTA demonstrated meaningful accuracy in preoperative risk stratification of GISTs.
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
Choi IY et al.
Abdominal Radiology (2019) 44:2346–2356 - “In the NCCN guideline, if the size of the tumor is more than 10 cm or the mitosis count is more than 5 in high- power fields (HPFs), the gastric GIST is classified as a high risk. For tumors in the small and large intestines, if the tumor size exceeds 5 cm, or the mitosis count is more than 5 in 50 HPFs, the tumor is classified as a high risk.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
Choi IY et al.
Abdominal Radiology (2019) 44:2346–2356 - “Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. About 10–30% of GISTs are clinically malignant, but all GISTs are considered to have some degree of malignant potential. The malignancy risk of GISTs is deter- mined based on the mitotic index, tumor size, and site of the lesion according to National Comprehensive Cancer Network (NCCN) guidelines.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
Choi IY et al.
Abdominal Radiology (2019) 44:2346–2356 - “In conclusions, CTTA parameters, such as MPP and kurtosis can be useful in predict the risk grade and mitosis index of GISTs. In our study, texture analysis parameters demonstrated meaningful accuracy in preoperative diagnosis of tumor risk stratification and can be used as imaging biomarkers for determination of tumor grade.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
Choi IY et al.
Abdominal Radiology (2019) 44:2346–2356
- OBJECTIVE. The objective of our study was to determine the utility of radiomics features in differentiating CT cases of pancreatic ductal adenocarcinoma (PDAC) from normal pancreas.
RESULTS. Mean tumor size was 4.1 ± 1.7 (SD) cm. The overall accuracy of the random forest binary classification was 99.2% (124/125), and AUC was 99.9%. All PDAC cases (60/60) were correctly classified. One case from a renal donor was misclassified as PDAC (1/65). The sensitivity was 100%, and specificity was 98.5%.
CONCLUSION. Radiomics features extracted from whole pancreas can be used to differentiate between CT cases from patients with PDAC and healthy control subjects with normal pancreas.
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Fishman EK et al
AJR 2019; 213:1–9 - RESULTS. Mean tumor size was 4.1 ± 1.7 (SD) cm. The overall accuracy of the random forest binary classification was 99.2% (124/125), and AUC was 99.9%. All PDAC cases (60/60) were correctly classified. One case from a renal donor was misclassified as PDAC (1/65). The sensitivity was 100%, and specificity was 98.5%.
CONCLUSION. Radiomics features extracted from whole pancreas can be used to differentiate between CT cases from patients with PDAC and healthy control subjects with normal pancreas.
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S et al.
AJR 2019; 213:1–9 - “CT features of early PDAC can be subtle and missed by even experienced radiologists. Early signs of PDAC such as pancreatic parenchyma inhomogeneity and loss of normal fatty marbling of the pancreas have been described on retrospective CT review up to 34 months before the diagnosis of PDAC. Quantitative analysis of these imaging features offers the potential for computer-aided diagnosis of PDAC.”
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Kawamoto S, Fouladi DF, Shayesteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK.
AJR Am J Roentgenol. 2019 Apr 23:1-9. - “This study aimed to tackle the second goal—differentiation of abnormal from normal pancreatic tissue using segmentation of the entire pancreas (i.e., without relying on separate segmentation of the tumor region). Our results showed that, after manual segmentation of pancreas boundaries, radiomics features and the random forest classifier were highly accurate in differentiating PDAC cases from normal control cases (sensitivity, 100%; specificity, 98.5%; accuracy, 99.2%). The radiomics features most relevant to differentiate PDAC from normal pancreas were based on shape and textural heterogeneity.”
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Kawamoto S, Fouladi DF, Shayesteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK.
AJR Am J Roentgenol. 2019 Apr 23:1-9. - “Given the high accuracy of automatic pan- creas segmentation by existing algorithms, these algorithms could be used to generate the boundaries for pancreas segmentation, and then the radiomics feature analysis algorithm could be performed to differentiate PDAC from normal pancreas. Some technical hurdles need to be overcome before these complex algorithms can be combined, but we anticipate that will be possible in the near future.”
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Kawamoto S, Fouladi DF, Shayesteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK.
AJR Am J Roentgenol. 2019 Apr 23:1-9. - ”All of the scans in the current study were obtained at a single institution on units manufactured by a single vendor using matched protocols and the same reconstruction algorithm. Differences in image acquisition, reconstruction, segmentation, and feature extraction can affect radiomics features and results. There is currently no standardization in the optimal protocol for imaging acquisition and postprocessing for radiomics analysis.”
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Kawamoto S, Fouladi DF, Shayesteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK.
AJR Am J Roentgenol. 2019 Apr 23:1-9. - “This preliminary study showed that the radiomics features extracted from the whole pancreas can be used to differentiate between CT images of patients with PDAC and CT images of healthy control subjects. There is the potential to combine this algorithm with automatic organ segmentation algorithms for automatic detection of PDAC.”
Utility of CT Radiomics Features in Differentiation of Pancreatic Ductal Adenocarcinoma From Normal Pancreatic Tissue
Chu LC, Park S, Kawamoto S, Fouladi DF, Shayesteh S, Zinreich ES, Graves JS, Horton KM, Hruban RH, Yuille AL, Kinzler KW, Vogelstein B, Fishman EK.
AJR Am J Roentgenol. 2019 Apr 23:1-9. - OBJECTIVE. The objective of our study was to investigate the potential influence of intra- and interobserver manual segmentation variability on the reliability of single-slice–based 2D CT texture analysis of renal masses.
CONCLUSION. Single-slice–based 2D CT texture analysis of renal masses is sensitive to intra- and interobserver manual segmentation variability. Therefore, it may lead to nonreproducible results in radiomic analysis unless a reliability analysis is considered in the workflow.
Reliability of Single-Slice–Based 2D CT Texture Analysis of Renal Masses: Influence of Intra- and Interobserver Manual Segmentation Variability on Radiomic Feature Reproducibility
Kocak B et al.
AJR 2019; 213:1–7 - “Lately, texture analysis has been an active area of research in the field of radiomics, suggesting that it can be used in predicting tumor subtypes, tumor stage, tumor grade, response to treatment, genomic profile, and overall survival . Nonetheless, recent evi- dence also suggests that conclusions must be treated with caution because several texture parameters may have reproducibility problems, which is an important challenge for building reliable predictive models to be used in clinical practice.”
Reliability of Single-Slice–Based 2D CT Texture Analysis of Renal Masses: Influence of Intra- and Interobserver Manual Segmentation Variability on Radiomic Feature Reproducibility
Kocak B et al.
AJR 2019; 213:1–7 - Although whole-tumor segmentation is known to be the most representative for tumor texture , it is considered an impractical and time-consuming process to be used in clinical routine, particularly in large tumors such as renal masses. For renal tumors, there has been a trend toward using a single image slice along with manual segmentation in an attempt to bring texture analysis into a daily routine.
Reliability of Single-Slice–Based 2D CT Texture Analysis of Renal Masses: Influence of Intra- and Interobserver Manual Segmentation Variability on Radiomic Feature Reproducibility
Kocak B et al.
AJR 2019; 213:1–7 - “In conclusion, single-slice–based 2D CT texture analysis of RCCs is sensitive to intra-and interobserver manual segmentation variability, which may lead to nonreproducible results in radiomic analysis. Therefore, a reliability analysis with as much and heterogeneous data as possible must be incorporated into every scientific research study using this technique. Otherwise, the radiomic studies of renal masses without a reliability analysis might lead to a chain of nonreproducible outcomes in terms of selected texture features and statistical models created, which might further influence the generalizability and replicability of the findings of the radiomic studies. bring texture analysis into a daily routine.”
Reliability of Single-Slice–Based 2D CT Texture Analysis of Renal Masses: Influence of Intra- and Interobserver Manual Segmentation Variability on Radiomic Feature Reproducibility
Kocak B et al.
AJR 2019; 213:1–7 - In addition, CECT provides more texture features with good to excellent interobserver reliability than unenhanced CT does. Filtered and transformed images might be useful for reducing the influence of manual segmentation variations on single-slice–based 2D CT texture analysis, yielding more features with good to excellent reliability than original images do.
Reliability of Single-Slice–Based 2D CT Texture Analysis of Renal Masses: Influence of Intra- and Interobserver Manual Segmentation Variability on Radiomic Feature Reproducibility
Kocak B et al.
AJR 2019; 213:1–7
- Purpose To evaluate the feasibility of using computed tomography texture analysis (CTTA) parameters for predicting malignant risk grade and mitosis index of gastrointestinal stromal tumors (GISTs), compared with visual inspection. Conclusion Using TexRAD, MPP and kurtosis are feasible in predicting risk grade and mitosis index of GISTs. CTTA demonstrated meaningful accuracy in preoperative risk stratification of GISTs.
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “Three to four parameters at different scales were significantly different according to the risk grade, mitosis rate, and the presence or absence of necrosis (p < 0.041). MPP at fine or medium scale (r = − 0.547 to − 393) and kurtosis at coarse scale (r = 0.424–0.454) correlated significantly with risk grade (p < 0.001). HG-GIST was best differentiated from LG-GIST by MPP at SSF 2 (AUC, 0.782), and kurtosis at SSF 4 (AUC, 0.779) (all p<0.001). CT features predictive of HG-GIST were density lower than or equal to that of the erector spinae muscles on enhanced images (OR 2.1; p = 0.037; AUC, 0.59), necrosis (OR, 6.1; p < 0.001; AUC, 0.70), heterogeneity (OR, 4.3; p < 0.001; AUC, 0.67), and mucosal ulceration (OR, 3.3; p = 0.002; AUC, 0.62).”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. About 10–30% of GISTs are clinically malignant, but all GISTs are considered to have some degree of malignant potential. The malignancy risk of GISTs is determined based on the mitotic index, tumor size, and site of the lesion according to National Comprehensive Can- cer Network (NCCN) guidelines.
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “In the NCCN guideline, if the size of the tumor is more than 10 cm or the mitosis count is more than 5 in high- power fields (HPFs), the gastric GIST is classified as a high risk. For tumors in the small and large intestines, if the tumor size exceeds 5 cm, or the mitosis count is more than 5 in 50 HPFs, the tumor is classified as a high risk.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “The mitotic index is the key feature for assessing malignancy risk and is considered one of the most important prognostic factors for GISTs. However, the mitotic index cannot be assessed without histological examination. Preoperative endoscopic biopsy of a mass is commonly performed but it has a risk of hemorrhage. Furthermore, a core-needle biopsy may be inconclusive if a necrotic or hemorrhagic portion of the tumor is sampled and has the limitation of sampling bias.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “CT texture analysis (CTTA) is a novel imaging postprocessing tool used to measure tissue heterogeneity that may not be perceptible to the naked eye. Texture analysis makes available an objective and quantitative evaluation of tumor heterogeneity by analyzing the distribution and relationship of pixel gray levels in the image.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “We hypothesized that CTTA could provide quantitative information about the malignancy risk and mitotic count in GISTs. Therefore, the purpose of our study was to the evaluate feasibility of using CTTA parameters for predicting the malignancy risk and mitotic index of GISTs, in comparison with the visual inspection.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4- This study showed that CTTA parameters may be helpful in preoperative stratification of malignancy risk of GIST. We found that HG-GISTs were characterized by significantly lower mean, SD, and MPP, and higher skewness and kurto- sis, compared with LG-GISTs. ROC analysis demonstrated that MPP at a fine SSF and kurtosis at a coarse SSF showed the highest diagnostic performance for differentiating HG- GISTs from LG-GISTs. These texture analysis parameters were superior to conventional parameters, such as mean density or SD, in the differential diagnosis of HG-GISTs and LG-GISTs.
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - “In conclusions, CTTA parameters, such as MPP and kurtosis can be useful in predict the risk grade and mitosis index of GISTs. In our study, texture analysis parameters demonstrated meaningful accuracy in preoperative diagno- sis of tumor risk stratification and can be used as imaging biomarkers for determination of tumor grade.”
Feasibility of using computed tomography texture analysis parameters as imaging biomarkers for predicting risk grade of gastrointestinal stromal tumors: comparison with visual inspection
In Young Choi et al.
Abdominal Radiology https://doi.org/10.1007/s00261-019-01995-4 - Purpose: To explore the value of CT texture analysis (CTTA) for differentiation of focal nodular hyperplasia (FNH) from hepatocellular adenoma (HCA) on contrast-enhanced CT(CECT).
Results: On HAP images, mean, mpp, and skewness were significantly higher in FNH than in HCA on unfiltered images (p £ 0.007); SD, entropy, and mpp on filtered analysis (p £ 0.006). On PVP, mean, mpp, and skewness in FNH were significantly different from HCA (p £ 0.001) on unfiltered images, while entropy and kurtosis were significantly higher in FNH on filtered images (p £ 0.018). The multivariate logistic regression analysis indicated that the mean, mpp, and entropy of medium-level and coarse-level filtered images on HAP were independent predictors for the diagnosis of HCA and a model based on all these parameters showed the largest AUROC (0.824).
Conclusions: Multiple explored CTTA parameters are significantly different between FNH and HCA on CECT.
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330 - Purpose: To explore the value of CT texture analysis (CTTA) for differentiation of focal nodular hyperplasia (FNH) from hepatocellular adenoma (HCA) on contrast-enhanced CT(CECT).
Conclusions: Multiple explored CTTA parameters are significantly different between FNH and HCA on CECT.
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330 - “In conclusion, multiple explored CTTA parameters are significantly different in FNH compared to HCA. These results justify further larger-cohort studies, preferably multicenter, to validate the use of texture analysis in differentiating benign liver lesions and to increase the diagnostic value of CECT.”
Evaluation of texture analysis for the differential diagnosis of focal nodular hyperplasia from hepatocellular adenoma on contrast-enhanced CT images
Roberto Cannella et al.
Abdom Radiol (2019) 44:1323–1330
- “Radiogenomics as a distinct, new field within medical imaging. Radiogenomics aims to correlate imaging characteristics (i.e., the imaging phenotype) with underlying genes, mutations, and expression patterns. As such, radiogenomics represents the evolution of radiology-pathology correlation from
the anatomic-histologic level to the genetic level and also characterizes the interface of biologic systems approaches and imaging.”
Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and Treatment Katja Pinker et al. Radiology 2018; 287:732–747 - “Importantly, radiogenomics is not equivalent to radiomics. Radiomics refers to the conversion of digital medical images to higher dimensional data and the subsequent mining for hypothesis generation and/or testing and ultimately an improved decision support.”
Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and Treatment Katja Pinker et al. Radiology 2018; 287:732–747 - “Texture features are computer- extracted imaging features of special interest. Texture analysis quantifies the internal morphology and three-dimensional structure of the lesion of interest. Texture analysis involves four steps: feature extraction, texture discrimination, texture classification, and, if necessary, shape reconstruction . In feature extraction, calculations are made on the basis of statistical, structural, model-based signal processing, and a numerical value is generated for each specific texture variable of interest.”
Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and Treatment Katja Pinker et al. Radiology 2018; 287:732–747 - “Radiogenomic generation of big data also poses challenges. It is a nontrivial issue to store, manage, extract, analyze, integrate, visualize, and communicate information from the myriad of data representations of cancer. Such multifactorial and heterogeneous data must be integrated in a standardized, cost-effective, and secure manner.”
Precision Medicine and Radiogenomics in Breast Cancer: New Approaches toward Diagnosis and Treatment Katja Pinker et al. Radiology 2018; 287:732–747