By Erik L. Ridley, AuntMinnie staff writer
November 11, 2013

Tuesday, December 3 | 3:50 p.m.-4:00 p.m. | SSJ06-06 | Room S404CD
Researchers from University Hospitals in Cleveland have found that a next-generation computer-aided detection (CAD) platform could potentially speed up the diagnosis of lung cancer.

While the early versions of lung CAD algorithms developed by Riverain Technologies were limited by high false-positive rates, a newer version has yielded encouraging results in terms of sensitivity and false positives, said presenter Dr. Robert Gilkeson, vice president of clinical research at University Hospitals.

The researchers retrospectively tested version 5.2 of Riverain's OnGuard software (now known as ClearRead +Detect) on a collection of very subtle lung cancers that had initially been missed by radiologists on chest radiographs. There was an average time interval of 577 days (range, 1 day to 8 years) between the index chest radiograph and eventual diagnosis.

The CAD algorithm detected 28 of the 41 lesions for a sensitivity of 67%.

"In the future, I would like to assess the CAD algorithm in our growing CT lung cancer screening program," Gilkeson said. "A prospective trial to assess its ability to detect actionable lung nodules seen on screening low-dose CT would be particularly helpful. This would be best served in a larger multicenter trial to better assess true effect on patient survival."

Last Updated hh 11/7/2013 4:40:23 PM