In the realm of personalized and precision medicine, quantitative metrics are needed for predicting treatment outcomes during neoadjuvant chemotherapy for breast cancer. We evaluated an automated tumor sphericity measurement as an addition to standard functional tumor volume for prediction of treatment response in a cohort of 220 patients with invasive breast cancer. Tumor sphericity captured some similar information to a visually assessed morphologic pattern score. Pre-treatment sphericity showed comparable prediction of pathologic complete response to pre-treatment volume, and added benefit in a logistic regression model. Sphericity may be useful in combination with tumor size to improve prediction of outcome.