Machine Learning Models and Multiparametric Magnetic Resonance Imaging for the Prediction of Pathologic Response to Neoadjuvant Chemotherapy in Breast Cancer

SIMPLE SUMMARY: Achieving pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer (BC) is crucial, as pCR is a surrogate marker for survival. However, only 10–30% of patients achieve it. It is therefore essential to develop tools that enable to accurately predict response. Recently, different studies have demonstrated the feasibility of applying machine learning approaches to non-invasively predict pCR before NAC administration from magnetic resonance imaging (MRI) data. Some of those models are based on radiomics, an emerging field that allows the automated extraction of clinically relevant information from radiologic images. However, the research is still at an early stage and further data are needed. Here, we determine whether the combination of imaging data (radiomic features and perfusion/diffusion imaging biomarkers) extracted from multiparametric MRIs and clinical variables can improve pCR prediction to NAC compared to models only including imaging or clinical data, potentially avoiding unnecessary treatment and delays to surgery. ABSTRACT: Background: Most breast cancer (BC) patients fail to achieve pathological complete response (pCR) after neoadjuvant chemotherapy (NAC). The aim of this study was to evaluate whether imaging features (perfusion/diffusion imaging biomarkers + radiomic features) extracted from pre-treatment multiparametric (mp)MRIs were able to predict, alone or in combination with clinical data, pCR to NAC. Methods: Patients with stage II-III BC receiving NAC and undergoing breast mpMRI were retrospectively evaluated. Imaging features were extracted from mpMRIs performed before NAC. Three different machine learning models based on imaging features, clinical data or imaging features + clinical data were trained to predict pCR. Confusion matrices and performance metrics were obtained to assess model performance. Statistical analyses were conducted to evaluate differences between responders and non-responders. Results: Fifty-eight patients (median [range] age, 52 [45–58] years) were included, of whom 12 showed pCR. The combined model improved pCR prediction compared to clinical and imaging models, yielding 91.5% of accuracy with no false positive cases and only 17% false negative results. Changes in different parameters between responders and non-responders suggested a possible increase in vascularity and reduced tumour heterogeneity in patients with pCR, with the percentile 25th of time-to-peak (TTP), a classical perfusion parameter, being able to discriminate both groups in a 75% of the cases. Conclusions: A combination of mpMRI-derived imaging features and clinical variables was able to successfully predict pCR to NAC. Specific patient profiles according to tumour vascularity and heterogeneity might explain pCR differences, where TTP could emerge as a putative surrogate marker for pCR.