Dual-energy digital radiography improves pneumothorax detection

Pneumothoraces, especially small ones, can be better detected by bone-subtracted dual-energy digital radiographs than by conventional chest radiography, according to a French study published in Academic Radiology. This is especially true for clinicians — whether radiologists or non-radiologists — less experienced in identifying pneumothoraces.

In the study conducted by radiologists from the Radiology and Imaging Department Guilloz of the Hôpital Central in Nancy, 36 physicians interpreted three sets of dual-energy chest radiographs of 40 patients, 19 with confirmed pneumothorax and 21 without. The physician mix included 4 radiologists, 11 emergency physicians, 16 residents specializing in radiology, pneumonology, or general medicine, and 5 interns. They hypothesized that bone subtraction with dual-energy digital subtraction might improve detection rates of pneumothorax and help improve diagnostic performance. They felt that if either or both hypotheses provided to be correct, the findings could be used to help improve management of patients with suspected pneumothorax and also reduce the need for chest computed tomography (CT) exams.

The confirmed pneumothoraces were located on both the left (12 cases) and the right (7 cases) sides of the chest. A thoracic radiologist calculated the pneumothorax volume, which ranged from 5.5% to 39.1%. Fourteen pneumothoraces were small (less than 20%) and five were medium/large (greater than 20%).

Three “interpretation” sessions of the 40 exams were conducted at two-week intervals, displaying first a standard radiograph from the high-energy exposure, followed by a bone image with soft-tissue suppression, and then a soft-tissue image with suppression of the osseous structures. The readers viewed high-resolution (1920 x 1200 pixel) video images calibrated to be identical to a diagnostic workstation in a room with luminance comparable to a radiology reading room. The 40 images from each “set” were displayed in 15-second intervals, with the order of presentation of the images changed each time. Participants were asked to identify if they identified a pneumothorax or not.

Under these conditions, with only 15 seconds to make an interpretation, the mean number of accurate interpretations was 30.9%, 35%, and 33.8% respectively. The highest mean sensitivity and specificity was achieved with the bone-suppressed image set, at 82% and 92%, compared to 70% and 84% for the standard radiograph set. Identification of pneumothorax improved for each reader subgroup, but especially for inexperienced and less experienced readers.

“These results indicated that bone-subtracted dual-energy chest radiographs can reduce the impact of the reader’s dual energy subtraction can reduce the impact of the reader’s expertise in the detection of pneumothorax....[It’s} use...would seem to be of particular interest for chest radiographs performed at an emergency department, where nonradiologists are usually the first, and sometimes the only ones, to interpret these images,” wrote lead author Ayla Urbaneja, MD.

REFERENCE

1. Urbaneja A, Dodin G, Hossu G, et al. Added Value of Bone Subtraction in Dual-energy Digital Radiography in the Detection of Pneumothorax: Impact of Reader Expertise and Medical Specialty. Acad Radiol. 2018 25;1:82-87.