CADx catches missed pulmonary embolisms

A prototype computer aided detection (CADx) program for identifying pulmonary embolisms (PEs) on CT angiography (CTA) was able to detect a PE in more than 75 percent of cases in which an acute PE had been missed in clinical practice, according to a study published in the January issue of the American Journal of Roentgenology.

“Given its performance in this study, we would recommend using the software as a second reader in pulmonary CTA cases initially interpreted as negative by a radiologist,” wrote Seth J. Kligerman, MD, of the University of Maryland School of Medicine in Baltimore, and colleagues.

PE-CADx has failed to gain traction due to concerns about false-positives and a perception that overlooked PEs are uncommon, explained the authors.

In order to evaluate the effectiveness of such a system, Kligerman and colleagues retrospectively assessed thousands of CTA studies conducted between January 2009 and July 2012, ultimately finding 53 studies with an overlooked acute PE. These studies were analyzed using a noncommercially available prototype PE-CADx program (Philips Healthcare).

In 41 of the cases (77.4 percent), the PE-CADx program was able to identify at least one PE, including correctly marking at least one PE in all 23 cases with multiple PEs missed on the initial interpretation. The program detected a PE in 60 percent of cases with a solitary PE.

The PE-CADx program demonstrated decreasing sensitivity for small vessels, with a per-study sensitivity of 91.7 percent for segmental PEs and 65.5 percent for subsegmental PEs, reported the authors.

The number of false positives per case was directly related to image quality, and the overall average was 3.8 false-positive marks per case. “Although most [false-positive] marks would likely be recognized and dismissed by skilled interpreters, there is the possibility that negative PE studies could be incorrectly interpreted as positive, leading to unnecessary treatment and a potentially adverse outcome,” wrote Kligerman and colleagues.

With regard to outcomes for the patients whose scans were included in the study, follow-up data showed that many of the patients with overlooked PE went untreated and had increased clot burden down the line, with 14 patients developing new PEs.

Kligerman and colleagues wrote that the program’s 37.5 second average processing time would allow it to be incorporated into the daily workflow of many practices, particularly if run in the background during review of the study, should the program be approved for use by the FDA.