Abstract PS1-06-24: Single MIP Image Evaluation for Predicting Pathologic Response in Patients Receiving Combination Neoadjuvant Immunotherapy/Chemotherapy: A Reader Study

Abstract Purpose: The purpose of this study was to compare diagnostic accuracy of single-slice maximum intensity projection (MIP) image evaluation with standard MRI in determining pathologic response for patients undergoing combination neoadjuvant immunotherapy/chemotherapy (NAI/NAC). Materials and Methods: Four single-institution NAI/NAC trials were evaluated retrospectively for MRI and pathologic responses. Diagnostic reports were evaluated to determine MRI complete vs partial response (mCR vs mPR) in comparison to pathologic complete vs partial response (pCR vs pPR). Next, early vs late phase MRI analysis was performed with three breast-fellowship trained radiologists blinded to MRI and pathologic reports. The readers categorized cases as mCR vs mPR based on single, random-ordered, early or late-phase maximum intensity projection (MIP) subtracted T1 post-contrast images. Diagnostic accuracy was calculated, and inter-reader agreement was evaluated using Kappa analysis. Results: A total of 77 patients were included, with 154 cases (early and late) utilized for reader analysis. This included 20 patients with triple negative breast cancer and 57 patients with HER2+ breast cancer. Pathologic complete response was achieved in 44/77 (57%). MRI reports categorizing mCR as a true predictor of pCR demonstrated 75% accuracy (95% Confidence Interval, CI, 68-82%) with sensitivity 89% and specificity 58%. Averaged results from the MIP reader study demonstrated accuracy of 70% (CI 62-77%, reader accuracies 68-71%), not significantly different than the MRI report (kappa 0.620). Average reading time was 6.6 seconds per case. Comparison of diagnostic accuracy for early vs late MIP evaluations yielded similar results (accuracy 73% vs 68%) which were not significantly different (CI 61-82% vs 56-78%, respectively). Conclusion: Post-NAI/NAC pre-surgical MRIs demonstrated moderate accuracy in predicting pCR, despite the potential for confounding inflammatory changes with immunotherapy in the neoadjuvant regimen. In addition, accuracy was similar when evaluating a single early-phase MIP image, which could support investigation into abbreviated studies in the future. Clinical Relevance: Single-slice early-phase MIP image evaluation of post-NAI/NAC pre-surgical breast cancer patients demonstrates moderate accuracy in predicting pathologic response. Citation Format: R. J. Weinfurtner, D. Ataya, S. Falcon, K. Funaro, B. Asha, B. Czerniecki, S. Hatem, H. Hyo. Single MIP Image Evaluation for Predicting Pathologic Response in Patients Receiving Combination Neoadjuvant Immunotherapy/Chemotherapy: A Reader Study [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS1-06-24.