AI identifies pancreatic cancer frequently missed on CT

"Researchers have recently developed an artificial intelligence tool that can spot pancreatic cancer on CT in its earliest stages.

Specifically, the computer-aided detection (CAD) tool is capable of identifying lesions that are less than 2 cm, before the cancer has progressed to the point of diminishing a positive prognosis, experts recently reported in Radiology.

“Because the prognosis worsens precipitously once the tumor grows larger than 2 cm, early detection represents the most effective strategy to improve the dismal prognosis of PC,” co-senior author of the study Weichung Wang, of the Taipei, Taiwan; Institute of Applied Mathematical Sciences, and colleagues explained. “CT is the major imaging modality used to help detect PC, but its sensitivity for small tumors is modest, with approximately 40% of tumors smaller than 2 cm being missed.”

The CAD tool was developed using an internal test set of 546 patients with pancreatic cancer and 733 control patients. On this test set, the tool achieved 90% sensitivity and 96% specificity.

Next, the tool was validated on the internal test set, in addition to 1,473 CT exams from multiple facilities in Taiwan. The tool again performed well in distinguishing pancreatic cancer from controls, at 90% sensitivity and 93% specificity.

In addition to detecting pancreatic cancer on par with radiologists, the tool also achieved 75% sensitivity for identifying lesions less than 2 cm in size—a feat that researchers involved in the study suggested could render the tool a valuable assistive device for radiologists in the future...." Lire la suite