METHODS: We included clinical and ultrasonographic attributes from 1526 breast lesions classified as BI-RADS 3, 4a, 4b, 4c, 5, and 6 that underwent US-guided breast biopsy in four institutions. We selected the most informative attributes to train nine machine learning models, ensemble models and models with tuned threshold to make inferences about the diagnosis of BI-RADS 4a and 4b lesions (validation dataset). We tested the performance of the final model with 403 new suspicious lesions.
RESULTS: The most informative attributes were shape, margin, orientation and size of the lesions, the resistance index of the internal vessel, the age of the patient and the presence of a palpable lump. The highest mean negative predictive value (NPV) was achieved with the K-Nearest Neighbors algorithm (97.9%). Making ensembles did not improve the performance. Tuning the threshold did improve the performance of the models and we chose the algorithm XGBoost with the tuned threshold as the final one. The tested performance of the final model was: NPV 98.1%, false negative 1.9%, positive predictive value 77.1%, false positive 22.9%. Applying this final model, we would have missed 2 of the 231 malignant lesions of the test dataset (0.8%).
CONCLUSIONS: Machine learning can help physicians predict malignancy in suspicious breast lesions identified by the US. Our final model would be able to avoid 60.4% of the biopsies in benign lesions missing less than 1% of the cancer cases.
方法:我们纳入了1526个乳腺病变的临床和超声特征,这些病变被分类为BI-RADS3、4a,4b,4c,在四个机构中接受US引导的乳房活检的5和6。我们选择了信息最丰富的属性来训练九种机器学习模型,集成模型和具有调谐阈值的模型,以推断BI-RADS4a和4b病变的诊断(验证数据集)。我们用403个新的可疑病变测试了最终模型的性能。
结果:信息最多的属性是形状,margin,病变的方向和大小,内部血管的阻力指数,患者的年龄和明显的肿块的存在。K-最近邻算法实现了最高的平均阴性预测值(NPV)(97.9%)。合奏并没有提高性能。调整阈值确实提高了模型的性能,我们选择了具有调整阈值的算法XGBoost作为最终阈值。最终模型的测试性能为:净现值98.1%,假阴性1.9%,阳性预测值77.1%,假阳性22.9%。应用这个最终模型,我们会错过测试数据集的231个恶性病变中的2个(0.8%).
结论:机器学习可以帮助医生预测美国确定的可疑乳腺病变的恶性程度。我们的最终模型将能够避免60.4%的良性病变中的活检缺失少于1%的癌症病例。