Artificial intelligence for breast cancer detection in mammography: experience of use of the ScreenPoint Medical Transpara system in 310 Japanese women.

BACKGROUND: To compare the breast cancer detection performance in digital mammograms of a panel of three unaided human readers (HR) versus a stand-alone artificial intelligence (AI)-based Transpara system in a population of Japanese women. METHODS: The subjects were 310 Japanese female outpatients who underwent digital mammographic examinations between January 2018 and October 2018. A panel of three HR provided a Breast Imaging Reporting and Data System (BI-RADS) score, and Transpara system provided an interactive decision support score and an examination-based cancer likelihood score. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were compared under each of reading conditions. RESULTS: The AUC was higher for human readers than with stand-alone Transpara system (human readers 0.816; Transpara system 0.706; difference 0.11; P < 0.001). The sensitivity of the unaided HR for diagnosis was 89% and specificity was 86%. The sensitivity of stand-alone Transpara system for cutoff scores of 4 and 7 were 93% and 85%, and specificities were 45% and 67%, respectively. CONCLUSIONS: Although the diagnostic performance of Transpara system was statistically lower than that of HR, the recent advances in AI algorithms are expected to reduce the difference between computers and human experts in detecting breast cancer.

Does integrated shimming improve lesion detection in whole-body diffusion-weighted examinations of patients with breast cancer?

PURPOSE: To evaluate the feasibility of proposed integrated slice-by-slice shimming (iShim) for whole-body diffusion weighted imaging (WB-DWI) in comparison to conventional 3D shim in patients with breast cancer. MATERIALS AND METHODS: Retrospective analysis of 116 consecutive patients (116 lesions) who underwent whole-body PET/MR using iShim (iShim group) were performed and compared with historical control of 103 patients (105 lesions) using 3D Shim (3D Shim group). RESULTS: As compared with dynamic contrast-enhanced (DCE) breast MRI, the apparent diffusion coefficient (ADC) value could not be determined for 15 (14%) of the 105 lesions of the 3D shim group and for 10 (9%) of the 116 lesions on iShim group. The intergroup difference failed to reach statistical significance (P = 0.1843). On the other hand, there was a significant difference in the frequencies of PET-positive and DWI-negative lesions between the 3D shim and iShim group (8.6% vs. 1.7%, respectively, P = 0.01942). CONCLUSION: In regard to detectability of breast cancers by DWI, iShim may allow improved detectability as compared to conventional 3D shim.

Simultaneous whole-body and breast 18F-FDG PET/MRI examinations in patients with breast cancer: a comparison of apparent diffusion coefficients and maximum standardized uptake values.

PURPOSE: To compare standardized uptake value (SUV) and apparent diffusion coefficient (ADC) values acquired using a PET/MRI scanner in breast cancer patients. MATERIALS AND METHODS: Whole-body PET/MRI and breast PET/MRI were performed in 108 consecutive patients. Ninety-four patients who had a total of 100 breast cancers were analyzed. SUVmax and ADCmean acquired using breast PET/MRI were compared with pathologic prognostic factors. RESULTS: All the lesions were visually detectable using PET and diffusion-weighted imaging (DWI) on breast PET/MRI; however, lesions were visually undetectable on whole-body DWI in 13 patients (13%) or on whole-body PET in 7 patients (7%). An analysis of ADCmean and SUVmax demonstrated a statistically significant correlation between whole-body imaging and breast imaging (rho = 0.613, p < 0.001 and rho = 0.928, p < 0.001, respectively). In a univariate analysis, SUVmax was significantly correlated with HER2 status (p < 0.001), Ki-67 (p = 0.014), tumor size (p = 0.0177), and nuclear grade (p = 0.0448). In multiple regression analysis, only tumor size (p = 0.00701) was shown to independently influence SUVmax. CONCLUSION: Prone breast imaging was more sensitive than whole-body PET/MRI for detection of breast cancers. Both SUVmax and ADCmean showed limited correlation with pathologic prognostic factors.