Multifractal analysis of tumour microscopic images in the prediction of breast cancer chemotherapy response

Due to the individual heterogeneity, highly accurate predictors of chemotherapy response in invasive breast cancer are needed for effective chemotherapeutic management. However, predictive molecular determinants for conventional chemotherapy are only emerging and still incorporate a high degree of predictive variability. Based on such pressing need for predictive performance improvement, we explored the value of pre-therapy tumour histology image analysis to predict chemotherapy response. Fractal analysis was applied to hematoxylin/eosin stained archival tissue of diagnostic biopsies derived from 106 patients diagnosed with invasive breast cancer. The tissue was obtained prior to neoadjuvant anthracycline-based chemotherapy and patients were subsequently divided into three groups according to their actual chemotherapy response: partial pathological response (pPR), pathological complete response (pCR) and progressive/stable disease (PD/SD). It was shown that multifractal analysis of breast tumour tissue prior to chemotherapy indeed has the capacity to distinguish between histological images of the different chemotherapy responder groups with accuracies of 91.4 % for pPR, 82.9 % for pCR and 82.1 % for PD/SD. F(α)max was identified as the most important predictive parameter. It represents the maximum of multifractal spectrum f(α), where α is the Hölder’s exponent. This is the first study investigating the predictive value of multifractal analysis as a simple and cost-effective tool to predict the chemotherapy response. Improvements in chemotherapy prediction provide clinical benefit by enabling more optimal chemotherapy decisions, thus directly affecting the quality of life and survival.