Towards data-driven decision-making for breast cancer patients undergoing mastectomy and reconstruction: accurate prediction of individual patient-reported outcomes at 2-year follow-up using machine learning

Purpose Post-surgical satisfaction with breasts is a key outcome for women undergoing cancer-related mastectomy and reconstruction. Current decision-making relies on group-level evidence, which may not offer optimal choice of treatment for individuals. We developed and validated machine learning algorithms to predict individual post-surgical breast-satisfaction which may facilitate optimal decision-making in breast cancer.

Methods We collected data from 3058 patients across 11 sites in North America. We trained and validated 4 algorithms to predict 2-year patient-reported BREAST-Q satisfaction with breasts: Regularized regression, Support Vector Machine, Neural Network, and Regression Tree. Algorithms were evaluated in terms of their accuracy and area under the receiver operating characteristics curve (AUC).

NCT01723423

Results Machine learning algorithms were able to accurately predict changes in women’s satisfaction with breasts (see table).

Low satisfaction with reconstruction at follow-up was most strongly associated with high preoperative satisfaction with breasts, radiation during or after reconstruction, nipple-sparing mastectomy, implant-based reconstruction, chemotherapy, unilateral mastectomy, and obesity.

Conclusion We reveal the crucial role of patient-reported outcomes in determining post-operative outcomes and that machine learning algorithms are suitable to identify individuals who might benefit from alternative treatment decisions than suggested by group-level evidence. We provide an easy-to-use tool based on our algorithms

importdemo.com

Tab.1 Evaluation of algorithms trained to predict significant changes in satisfaction with breasts at two-year follow up

	lower than baseline - Accuracy	AUC	greater than baseline - Accuracy	AUC
Regularized regression	.90 (.86 - .92)	.87 (.83 - .91)	.83 (.79 - .86)	.78 (.73 - .83)
Regression Tree	.90 (.87 - .92)	.77 (.70 -.83)	.83 (.79 - .86)	.67 (.61 - .73)
Support Vector Machine	.89 (.86 - .92)	.84 (.78 - .89)	.82 (.78 - .86)	.73 (.67 - .79)
Neural Network	.90 (.87 - .92)	.87 (.82 - .92)	.84 (.80 - .87)	.78 (.72 - .83)

Publikationsverlauf

Artikel online veröffentlicht:
07. Oktober 2020