Immunohistological factors affect breast MRI CAD accuracy

Thursday, December 5 | 11:00 a.m.-11:10 a.m. | SSQ02-04 | Room E450A
Korean researchers will highlight in this RSNA 2013 session which tumor characteristics can lead to missed cancers with breast MRI computer-aided detection (CAD) software.

After installing a CADstream breast MRI CAD system (Merge Healthcare) at their institution, the researchers found it was easier to read breast MRI exams. However, they also noticed that many cases were not being detected by the CAD software, according to Dr. Seon Hyeong Choi of Sungkyunkwan University.

Although most of their breast MRI exams are performed for preoperative evaluation in breast cancer patients, Choi and colleagues thought it was important for breast MRI screening purposes to understand the factors that affect CAD's ability to detect cancer.

Analysis of the preoperative breast MRI studies for 327 patients with histologically proven breast cancer found that 68 lesions (20%) were not detected by the CAD software. The software struggled somewhat with ductal carcinoma in situ (DCIS) cases, detecting only 31 (54.4%) of 57.

"[The study findings] suggested that you can miss many DCIS if you don't check the original images on PACS," Choi said.

Page 1 of 605
Next Page