CTPA Performance Challenges Approach to Pulmonary Embolism Treatment
A recent study comparing CT pulmonary angiography (CTPA) and V/Q scanning to diagnose pulmonary embolism (PE) raises important questions about the clinical implications of the vastly increased information provided by CTPA.
The study, published in the Dec. 19, 2007, issue of The Journal of the American Medical Association (JAMA), indicated that CTPA appears to be a safe alternative to V/Q scanning for excluding PE but CTPA also detects significantly more clots than V/Q. Therein lies a potential problem, said the study authors, pointing out that many of these additional diagnoses may be clinically insignificant and lead to potentially dangerous and costly treatment, with an attendant increase in radiation exposure.
John R. Mayo, M.D., a professor of radiology and cardiology at the University of British Columbia and not involved in the study, agreed. "What we have is a brand new frontier," said Dr. Mayo, a former chair of the chest subcommittee of the RSNA Scientific Program Committee. "Now that we have better imaging, we're finding things with poorly understood clinical implications."
For 30 years, V/Q lung scanning was the noninvasive imaging procedure of choice in patients with suspected PE, with a normal V/Q scan excluding a PE diagnosis and a high probability scan having an 85 to 90 percent positive predictive value. However, as the JAMA researchers note, in most patients for whom pulmonary embolism is suspected, V/Q scans are either of low or intermediate probability—such diagnostic uncertainty is a major limitation of V/Q scanning.
CTPA emerged over the last decade as an alternative noninvasive test, despite concerns about reported sensitivity ranging from 53 to 100 percent. Sensitivity has increased as modern multislice CT technology has replaced early single-slice scanners.
CTPA Finds Significantly More Pulmonary Emboli
"Our study was initiated by concerns about the widespread adoption of CTPA for diagnosis of PE, despite the relative lack of data to support its use to exclude the diagnosis of PE," said lead author David R. Anderson, M.D., a professor of medicine, pathology, and community health and epidemiology, head of the Division of Hematology and deputy head of the Department of Medicine at Dalhousie University in Halifax, Nova Scotia.
Investigators assigned 716 patients showing symptoms of possible PE to undergo V/Q scanning and another 701 to undergo CTPA. In the CTPA group, 133 patients (19.2 percent) were diagnosed as having PE in the initial evaluation, versus 101 (14.2 percent) in the V/Q scan group. Of those patients in whom PE was considered excluded, two patients randomized to CTPA and six randomized to V/Q scanning developed venous thromboembolism in follow-up, including one patient with fatal PE in the V/Q group.
"The study demonstrated that CTPA was at least as safe as the previous standard of V/Q scanning to exclude diagnosis of pulmonary embolism, which was reassuring," said Dr. Anderson, noting that CTPA has become a standard tool to evaluate patients with suspected PE at his institution. "We expected more patients would be diagnosed with PE using CTPA than V/Q, but we were surprised about the magnitude of the difference that we observed."
The researchers concluded that a diagnostic-management strategy, using CTPA in combination with consideration of clinical probability, D-dimer testing and leg venous ultrasonography, was not inferior to one using V/Q scanning to exclude the diagnosis of PE.
They added, however, that "further research is required to confirm whether some pulmonary emboli detected by CTPA may be clinically unimportant—the equivalent of deep vein thrombosis isolated to the calf veins—and not require anticoagulant therapy."
Definition of Disease Burden Challenged
Determining the clinical and therapeutic significance of small subsegmental defects detected by CTPA is indeed important, as most radiology centers in the U.S. are using CTPA and not V/Q as their frontline test, said Ioannis Vlahos, M.D., an assistant professor of radiology at New York University and a consultant at St. George's Hospital in London. Rather, V/Q scanning is commonly in use as part of a screening algorithm in those areas—including the U.K.—that have limited access to CT technology, he said.
"In the U.S., CTPA performed with multidetector CT is the number one investigational modality, without a doubt," said Dr. Vlahos, a co-investigator on a study of CTPA and PE detection presented at RSNA 2007.
"The problem has been that, through the years, CTPA has been difficult to prove as our gold standard because, in general, the standards we use for detection of pulmonary thrombotic disease are flawed to some degree," said Dr. Vlahos.
Dr. Mayo agreed. "There is a lack of knowledge as to what constitutes enough of a disease burden to warrant treatment," he said.
Dr. Vlahos acknowledged that a small fraction of patients—specifically those with contrast allergies, those who have recently received contrast for other studies and pregnant women who should avoid radiation exposure—may be better served by V/Q scanning or MR imaging versus CTPA.
In most cases, however, CTPA is considered an attractive modality, offering more than just sensitivity and specificity in detecting PE, said Dr. Vlahos. "When the case is normal or definitively abnormal, V/Q is useful," he said. "The problem is that the majority of cases do not fall into absolutely normal or high probability categories. They fall somewhere in between, which necessitates obtaining additional tests, such as CTPA or even an old-fashioned catheter pulmonary angiogram."
CTPA not only provides a definitive answer as to the absence or presence of PE, said Dr. Vlahos, but also can identify equally important alternative causes of symptoms, such as cardiac problems.
Reevaluation of Therapeutic Approach Needed
"CTPA is a superb test—the data indicate that it is better than V/Q scanning and can safely replace it," said Dr. Mayo, adding that advances in technology and protocol are ever increasing CTPA performance. "Because it is such a good test, we need to reevaluate how to use this new information in determining who should be treated—that is, how much clot burden we need to see before we treat."
Dr. Mayo said the University of British Columbia uses a 64-slice CT scanner for PE studies. Previous studies at the institution using single- and 4-slice scanners revealed sensitivity of approximately 90 percent and specificity of 95 percent in detecting PE. Ongoing clinical follow up indicated that the sensitivity and specificity for PE to the subsegmental level remained at greater than 90 percent.
Drs. Mayo and Vlahos don't dispute the idea that some pulmonary emboli detected by CTPA are of unknown clinical significance. That fact, however, should not diminish CTPA as the preferred test for PE, they said.
"As with all radiologic studies, a conscious balance of risk to benefit should be considered by the referring clinician and all studies should be obtained with the lowest radiation dose that provides acceptable image quality," said Dr. Mayo.
 |  |
Contrast material–enhanced 16–detector row CT yielded coronal volume renderings in (a) anterocranial and (b) anterior perspectives, which allow intuitive visualization of the location and extent of embolus (arrows).
Radiology 2004; 230:329-337. © RSNA, 2004. All rights reserved. Reprinted with permission.
Special Focus Session Focuses on Pulmonary Embolism
"Imaging Algorithms in Pulmonary Embolism," an RSNA 2008 special focus session, will consider the results of the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED) 2 study as it details algorithms proposed to evaluate patients with suspected venous thromboembolism. Also addressed in the session will be methods for optimizing CT pulmonary angiography (CTPA) and reducing the radiation exposure patients receive, as well as the role of alternate tests. Moderators are Ella A. Kazerooni, M.D., and Joel E. Fishman, M.D., Ph.D. RSNA 2008 registration is under way for RSNA and AAPM members. Registration begins May 19 for non-members. More information is available at RSNA.org/register.