20.500.12556/RUL-178236 Impact of pectoral muscle removal on deep-learning-based breast cancer risk prediction Objective. State-of-the-art breast cancer risk (BCR) prediction models have been originally trained on mammograms with pectoral muscle (PM) included. This study investigated whether excluding PM during training/fine-tuning improves the model’s BCR discrimination performance, calibration, and robustness. Approach. First, the Original deep learning model (MIRAI), trained on the US (Massachusetts General Hospital) data, was validated, and the relative contribution of PM to BCR predictions was evaluated using saliency maps. Additionally, 23 792 mammograms from the Slovenian screening program were collected and two datasets were created, with and without screening positive exams. The original MIRAI was then fine-tuned on the training/fine-tuning set of Slovenian mammograms with and without PM, creating Fine-tuned MIRAI models. In total, four models (Original MIRAI with PM, Original MIRAI without PM, Fine-tuned MIRAI with PM, Fine-tuned MIRAI without PM) were compared on a test set in terms of discrimination performance for 1–5 Year BCR (evaluating area under the curve), calibration performance (measured with expected calibration error—ECE) and robustness to incremental PM removals/additions, and to incremental breast tissue removals. Results. The relative contribution of PM to the BCR prediction on the Original MIRAI model was low (∼5%); however, there were significant outliers where the relative contribution was more than 50%. The removal of PM did not impact the 1–5 Year BCR discrimination performance of the Original MIRAI (with screening positive exams: 0.77–0.91, without screening positive exams: 0.64–0.67). Fine-tuned MIRAI on mammograms with PM removed achieved significantly higher 1-5 Year BCR discrimination performance (with screening positive exams: 0.82–0.93, without screening positive exams: 0.71–0.79). After recalibration, all models had similar ECE (with screening positive exams: 0.04–0.05, without screening positive exams: 0.02–0.03). Significance. Improved BCR discrimination performance can be achieved when the model is trained/fine-tuned on mammograms with PM removed. breast cancer risk calibration mammography pectoral muscle deep learning convolutional neural networks robustness dejavniki tveganja rak dojke kalibracija mamografija true false true Angleški jezik Slovenski jezik Članek v reviji 2026-01-21 15:22:24 2026-01-21 15:22:27 2026-01-22 04:36:27 0000-00-00 00:00:00 2025 0 0 19 str. no. 5, art. 055006 Vol. 70 2025 0000-00-00 Zaloznikova Objavljeno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 614 0031-9155 10.1088/1361-6560/adb367 226342403 Klanecek_2025_Phys._Med._Biol._70_055006.pdf Klanecek_2025_Phys._Med._Biol._70_055006.pdf 1 68418E36D574579170953A534DC2A64F f86c54f8e35af2013acaf6b39bc37607a96b0f5ba1e3b88b91bb07fc5d530d68 4880a56f-f6d5-11f0-bf92-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=225881 https://iopscience.iop.org/article/10.1088/1361-6560/adb367 1 8f5a626b-f6d4-11f0-bf92-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=225876 Fakulteta za matematiko in fiziko 0 0 0