[Skip to Navigation]
Sign In
Invited Critique
May 2001

Spiral Computed Tomography for the Diagnosis of Pulmonary Embolism in Critically Ill Surgical Patients—Invited Critique

Arch Surg. 2001;136(5):511. doi:10.1001/archsurg.136.5.511

There seems to be a great deal of interest in, and enthusiasm for, the use of spiral computed tomography in the diagnosis of acute pulmonary embolism (PE), particularly among our radiological colleagues. With regard to critically ill surgical patients, such enthusiasm may be somewhat premature, as pointed out by Velmahos et al. In this small series of 22 surgical intensive care unit (SICU) patients, spiral computed tomographic pulmonary angiography (CTPA) achieved overall sensitivity and specificity rates of only 45% and 82%, respectively, when compared with conventional pulmonary angiography (PA). In fact, even duplex ultrasonography was more accurate in the prediction of PE than was CTPA. In addition to this sobering message, several other points deserve emphasis. First, in this patient population, the classic clinical findings associated with PE are of no value in identifying those patients who are ultimately found to have sustained PE. Therefore, diagnostic testing is mandatory and in most cases necessitates a trip to the radiology department, with all of its attendant risks. The second point concerns the low sensitivity rate (in this study, 33%) of CTPA in the detection of peripheral PE and the clinical significance thereof. Although there is considerable debate regarding the physiological significance of subsegmental PE, in patients with limited pulmonary reserve, even small peripheral emboli may be associated with morbidity and mortality. Should these subsegmental emboli go undiagnosed and untreated, the patient is clearly at risk for central, potentially fatal PE. Thus, the poor performance of CTPA in identifying these subsegmental emboli is of great significance in this patient population. Third, in addition to patient population, technical factors such as collimation, reconstruction interval, contrast dose, site of contrast injection, scan delay time, and breath-holding requirements may affect the accuracy of CTPA. The true value of this technology cannot be adequately assessed in the absence of a standardized technique for performing CTPA, and clear diagnostic criteria for its interpretation.

Add or change institution