Chest:Pulmonary Embolism Evaluation

Pulmonary embolism accounts for 10% of all hospital deaths and is a contributing factor in another 10%. Over 500,000 patients experience a pulmonary embolism each year and over 50,000 will die from pulmonary embolism. Early therapeutic intervention can improve survival sixfold compared to non-treatment. Yet, to date many patients who have pulmonary embolism proven at autopsy never had the diagnosis made pre-mortem. In many of these cases no imaging study was ever ordered and the diagnosis often not considered. Two recent studies in fact show that the prevalence of pulmonary embolism may be on the order of 1-2% of all patients who undergo CT scanning for non-pulmonary embolism causes. Gosselin et al. found that in 785 consecutive patients referred for CT that 12 or 1.5% had unsuspected pulmonary embolism. Similarly Winston et al. reviewed 1,879 consecutive contrast enhanced CT scans and found pulmonary embolism in 18 or nearly 1% of the population. The results from these studies only emphasizes why a new imaging paradigm is needed for the evaluation of suspected pulmonary embolism.

The radiologic diagnosis of pulmonary embolism was dependent on the use of chest radiographs (CXR), ventilation perfusion lung scanning (VQ scans), and pulmonary angiography. Some of the critical facts regarding these techniques can be summarized as follows:

The chest x-ray is a nonspecific and relatively insensitive screening test. The classic plain film sign of pulmonary embolism is seen in less than 7% of cases (Westermark's sign of prominent central pulmonary artery and decreased pulmonary vascularity) and in nearly half the cases when present is not specific for pulmonary embolism. The main role of a chest x-ray has been to exclude other pathology including pneumonia or an underlying neoplasm.

Nuclear scintigraphy or VQ lung scanning is a technique which provides a probability of a pulmonary embolism by defining the studies with one of three diagnosis; low probability pulmonary embolism, indeterminate study for pulmonary embolism, and high probability pulmonary embolism. Using PIOPED criteria only 40% of patients with documented pulmonary embolism will have a high probability scan and 20% will have a low probability study. If a VQ scan is read as high probability for pulmonary embolism and there is a high clinical suspicion then there is a better than 96% chance that the patient has a pulmonary embolism. Similarly if the study is read as low probability or normal there is greater than a 98% chance that the patient does not have a pulmonary embolism. However, the problem rests in that nearly 70-75% of VQ scans are read as indeterminate in real clinical practice. In addition there is up to a 30% chance if interobservor variability in scan interpretation (i.e. indeterminate vs. high probability). (4) The PIOPED study recommended a pulmonary angiogram in indeterminate cases, yet well less than 5% of these patients ever get a pulmonary angiogram.

Pulmonary Angiography has been the gold standard for the evaluation of suspected pulmonary embolism. However, its use has been limited by a number of factors including radiologist training in doing pulmonary angiography has been limited, the invasiveness of the study and its potential complications, and its increased risk in very ill patients. Although pulmonary angiography has been the accepted gold standard a closer look at the statistics finds that when subsegmental emboli are present there will be an up to one-third interobservor variation in interpretation and up to one-third error rate. (5) Even in cases with segmental emboli studies (5) have shown that per patient agreement may only occur in around 90% of cases. Stein et al. found similar results and concluded "conventional pulmonary angiography is not precise for the diagnosis of PE limited to subsegmental arteries." (6) Classic angiography then is a very accurate study but can not be considered "24 carat gold."

The introduction of spiral or helical CT with its fast acquisition times has revolutionized vascular imaging. The current state-of-the-art single detector spiral scanner can acquire data using second or subsecond (.750 sec) acquisitions, scan for a continuous 40-50 second time frame, and do this with narrow collimation studies (2-3 mm) and close interscan spacing (1-3 mm). The use of CT for the evaluation of suspected pulmonary embolism is not a new one. Institutions with the electron beam scanner found it to be ideal for the evaluation of suspected pulmonary embolism. Teigen et al. found EBCT to approach 100% accuracy for clinically significant pulmonary embolism. (7) When spiral CT was introduced several publications produced positive results and conclusions including Remy-Jardin who in 1992 concluded that "spiral volumetric CT can reliably depict thromboemboli in second to fourth division pulmonary vessels", van Rossum who in 1996 found that "spiral CT angiography is an accurate method for the depiction and exclusion of PE, with the exception of isolated subsegmental PE" and by Mayo (10) who in 1997 concluded that "in cases of pulmonary embolism, sensitivity of spiral CT is greater than that of scintigraphy. Interobservor agreement is better with spiral CT." Others were less impressed as Goodman noted that in their subsegment population that "helical CT was only 63% sensitive" and that CT has a limited role in the evaluation of acute pulmonary embolism."

Close inspection of some of the initial published results were in great part limited to the scanner used and the scan technique followed. For example today a routine CT technique for the evaluation of suspected pulmonary embolism would be a single 30-35 second breathhold, 2-3 mm collimation, images reconstructed every 2 mm and 120-150 ml of non-ionic contrast injected. Goodman used two 12 second or a single 24 second breathhold, 5-mm collimation and reconstruction every 3 mm. By 1997, Goodman was more optimistic of the role of CT after newer scanners allowed better protocols in his institution. (12) In fact numerous articles addressed the importance of technical factors in the successful use of CT for pulmonary embolism including using the proper window settings, optimization of timing of data acquisition and contrast injection, slice collimation, caudal to cranial scanning to name a few key factors that must be monitored for each study.

Several articles specifically addressed the comparison of conventional angiography and spiral CT angiography. Remy-Jardin (13) reviewed in a prospective study 75 patients who were evaluated with both CT and pulmonary angiography. The results showed that CT was 100% for positive predictive value, with a prospective sensitivity of 91%. Ten emboli were seen only with CT but not with classic pulmonary angiography. Please note that this article was from 1996 before the latest generation of CT scanners . Two recent articles show further progress making CT (14-15) the current study for the evaluation of suspected pulmonary embolism. Garg et al. concluded that "a helical CT scan can be effectively used to rule out clinically significant pulmonary emboli and may prevent further investigation or unnecessary treatment of most patients." Lefebvre et al. showed that "interobservor and intraobservor agreement in the diagnosis of pulmonary embolism with helical CT is very good despite a wide variety of experience among radiologists." Goodman et al. also concluded that "The frequency of clinical diagnoses of PE after a negative CT scan was low and similar to that after a negative or low probability V-P scan. Helical CT is a reliable imaging tool for excluding clinically important PE."

MDCT provides the next step forward by providing the ability to obtain thin sections (2-3 mm) at narrow intervals (2-3 mm) and yet perform the study with a 10-15 second breathhold (depending on the scan parameters selected). Although no article with specific statistics with MDCT data has been published its advantages are very clear-cut for the detection of pulmonary embolism. The advantages of narrow collimation to pick up clot in smaller vessels have previously been addressed by Remy-Jardin et al.

An additional advantage of MDCT is the ability to combine a CT venogram of the deep venous system with the CT pulmonary angiogram. Several recent articles have shown that CT and venography were comparable in accuracy with CT having the advantage when looking at the IVC and iliac vessels. Although the optimal time delay from the completion of the CT pulmonary angiogram to the venogram is not yet defined, between a 120-180 second delay is currently usually selected. Another question is the area to be scanned (i.e. from mid-abdomen through the calves or a more focused study such as from the iliac crests through the popliteal vessels) as well as scanning parameters (3-10 mm collimation every 5-20 mm) used. Nevertheless, it now seems apparent that a CT pulmonary angiogram combined with a CT venogram may be the combination needed to become the single examination of choice for the evaluation of suspected embolic disease. A recent article by Cham et al. nicely summarizes the subject: "Among patients suspected to have pulmonary embolism, a substantial number had DVT in the absence of pulmonary embolism. Combined pulmonary CT angiography-indirect CT venography can depict these cases with accuracy comparable to that of ultrasound and thus could have a significant effect on patient care."


"Among patients suspected to have pulmonary embolism, a substantial number had DVT in the absence of pulmonary embolism. Combined pulmonary CT angiography-indirect CT venography can depict these cases with accuracy comparable to that of ultrasound and thus could have a significant effect on patient care." Deep Venous Thrombosis: Detection by Using Indirect CT Venography
Cham MD et al.
Radiology 2000; 216:744-751

"CT of the deep venous system of the lower extremities after standard CT pulmonary angiography, performed with appropriate timing considerations, allows near maximal enhancement of the venous system in most patients without altering the optimum CT pulmonary angiography protocol." Optimization of Combined CT Pulmonary Angiography with Lower Extremity CT Venography
Yankelevitz DF et al.
AJR 2000; 174:67-69

"CT venous phase imaging at the time of pulmonary angiography is comparable to venous sonography in the evaluation of femoropopliteal DVT. The iliac veins and vena cava, vessels poorly shown on sonography but sometimes the source of significant pulmonary emboli , are also depicted by CT venography." Combined CT Venography and Pulmonary Angiography in Suspected Thromboembolic Disease: Diagnostic Accuracy for Deep Venous Thrombosis
Loud PA et al.
AJR 2000; 174:61-65

"Thin collimation helical CT provided technically acceptable examinations for pulmonary embolism in 360 patients (97%). In this population, CT revealed pulmonary embolism in 104 patients (29%), negative findings in 217 patients (59%), indeterminate findings in 39 patients (10%), and alternative diagnosis in 65% of patients with negative or inconclusive findings. The estimated false negative rate of CT was 5%." Clinical Value of Thin Collimation in the Diagnostic Work-up of Pulmonary Embolism
Remy-Jardin M et al.
AJR 2000; 175:407-411

"The frequency of clinical diagnoses of PE after a negative CT scan was low and similar to that after a negative or low probability V-P scan, Helical CT is a reliable imaging tool for excluding clinically important PE." Subsequent Pulmonary Embolism: Risk after a Negative Helical CT Pulmonary Angiogram-- Prospective Comparison with Scintigraphy
Goodman LR et al.
Radiology 2000;215:535-542

"Lung nodules can be detected with similar detection rates when viewing conventional film or videotaped helical CT images. Videotaped images incur a lower cost, an important consideration in mass screening for lung cancer." Videotaped Helical CT Images for Lung Cancer Screening
Iwano S et al.
J Comput Assist Tomogr 2000; 24(2):242-246

"Low dose CT can greatly improve the likelihood of detection of small non-calcified nodules, and thus of lung cancer at an earlier and potentially more curable stage. Although false positive CT results are common, they can be managed with little use of invasive diagnostic procedures." Early Lung Cancer Action Project: Overall Design and Findings from Baseline Screening
Henschke CI et al.
Lancet 1999; 354:99-105

"The minimum tube current required for screening helical CT differs for different locations in the lung. An ideal CT protocol for the lung should permit the tube current to be changed during helical scanning." Lung Cancer Screening: Minimum Tube Current Required for Helical CT
Itoh S et al.
Radiology 2000; 215:175-183

"The lung cancer detection rate with CT was 0.48%, significantly higher than the 0.03-0.05% for standard mass assessments (chest x-rays) done previously in the same area." (study of 5483 patients) Mass screening for lung cancer with mobile spiral computed tomography scanner
Sone S et al.
Lancet 1998; 351:1242-1245

"Unfortunately, these data caution that improved small nodule detection with screening CT may not significantly improve lung cancer mortality." Correlation of Tumor Size and Survival in Patients With Stage IA Non-small Cell Lung Cancer
Patz Jr EF et al.
Chest 2000; 117:1568-1571