RADIOMICS BASED BREAST MALIGNANCY INDEX TO DIFFERENTIATE PATHOLOGICAL CHANGES DUE TO NEOADJUVANT CHEMOTHERAPY

Abstract

The leading cause of deaths among women in the world is Breast Cancer. Neoadjuvant chemotherapy (NAC) offers effective treatment results, thus reducing tumor aggression and allowing treatment monitoring. The Dynamic Contrast Enhanced (DCE) MRI plays a vital role in assessing the treatment response due to NAC. However, quantifying the treatment response in low-grade tumours is visually challenging. Radiomics is an evolving field of medical imaging that reflects the histopathological variations in breast tissues. Integrating radiomics with breast DCE-MRI provides clinically useful measures in evaluating the NAC response. In this work, we have formulated an index called Radiomics based Breast Malignancy Index (RBMI) using texture and Haar wavelets to differentiate the radiological differences of breast tissue due to NAC. The statistically significant radiomic features extracted from 20 DCE-MR images obtained using TCIA database were used in the calculation of RBMI. Results show that, RBMI could statistically differentiate (p=0.007) the treatment response between visit-1 & 2 due to NAC with mean and standard deviation values of 334706.5949 ± 93952.5123 and 296354.9720 ± 77120.6718 respectively. Hence, RBMI seems to be a clinically adjunct measure in evaluating the treatment response of breast cancer due to NAC.