Commercially Available Artificial Intelligence Score on Preoperative Mammography for Prediction of Future Breast...

BACKGROUND. Mammographic artificial intelligence (AI) systems have been explored for future breast cancer risk prediction. OBJECTIVE. The purpose of this study was to investigate associations between scores generated by a commercial AI system for mammographic breast cancer detection and diagnosis and development of second breast cancers and to compare the predictive performance of AI with that of existing clinical risk models. METHODS. This retrospective five-center study included 1740 patients (median age, 50.0 years) who underwent surgery for ductal carcinoma in situ (DCIS) between January 2012 and December 2017 and had at least 1 year of postoperative follow-up. Medical records were reviewed to identify second breast cancers (ipsilateral recurrences after breast-conserving surgery [BCS] or mastectomy or contralateral breast cancers). A commercial AI system for breast cancer detection and diagnosis processed preoperative mammograms. AI scores were dichotomized using the Youden index for prediction of a second breast cancer. Univariable and multivariable cause-specific hazards models with competing-risk analysis assessed associations with second breast cancers. Cumulative incidence rates (CIRs) were compared between dichotomized AI scores using log-rank tests. Time-dependent AUCs were compared between AI scores and clinical risk models incorporating pathologic information (Van Nuys prognostic index [VNPI] and Memorial Sloan Kettering Cancer Center [MSKCC] nomograms) using bootstrapping. RESULTS. Twenty-eight patients developed post-BCS ipsilateral recurrence, seven developed postmastectomy ipsilateral recurrence, and 25 developed contralateral breast cancer. AI scores were dichotomized at a threshold of 73.5% or greater. Post-BCS ipsilateral recurrence showed a significant independent association with an AI score of 73.5% or more (HR = 2.88). CIR for post-BCS ipsilateral recurrence was higher for an AI score of at least 73.5% than for an AI score of less than 73.5% at 5 years (4.13% vs 0.86%, p < .001) and 10 years (7.26% vs 3.72%, p < .001). The AUC for predicting post-BCS ipsilateral recurrence was not significantly different between AI scores and the VNPI or MSKCC nomogram at 5 years (0.70 vs 0.73 [p > .99] and 0.63 [p = .82], respectively) and 10 years (0.66 vs 0.75 [p = .66] and 0.68 [p > .99], respectively). AI scores were not associated with other second breast cancer events in hazard models and CIR analyses (p > .05). CONCLUSION. AI scores showed independent associations with ipsilateral recurrence after BCS for DCIS and had predictive performance that was not significantly different from that of existing clinical models. CLINICAL IMPACT. AI scores, readily obtained noninvasively on preoperative mammography, may help inform DCIS treatment and surveillance strategies.