Abstract Objectives and rationale: Early response prediction during neoadjuvant therapy (NAT) for breast cancer (BC) enables response-guided precision treatment. Imaging-based post-NAT evaluation may reduce dependence on pathological complete response (pCR) confirmation and facilitate surgical de-escalation. Fibroblast activation protein inhibitor (FAPI) PET/MRI represents a promising molecular imaging modality targeting the tumor microenvironment. This study aimed to evaluate the efficacy of 18F-FAPI-42 PET/MRI for early and late prediction of pCR to NAT in BC. Methods: 45 consecutive patients with newly diagnosed BC who met the criteria for NAT were prospectively enrolled and underwent 18F-FAPI-42 PET/MRI before treatment initiation (PET1), after two cycles of NAT (PET2), and before post-NAT surgery (PET3). SUVmax was measured in the primary tumor and metastatic lymph nodes (MLNs). The tumor-to-background ratio (TBR) was calculated as the lesion SUVmax divided by the SUVmean of contralateral normal breast tissue. The percentage changes in SUVmax and TBR from baseline were calculated after 2 NAT cycles (ΔSUVmax1 and ΔTBR1) and before surgery (ΔSUVmax2 and ΔTBR2). Predictive performance of 18F-FAPI uptake was analyzed by receiver operating characteristic curve analysis and the area under the curve (AUC). The Youden index determined optimal cutoffs. Results: 18 patients (40%) achieved pCR, while 27 patients (60%) exhibited residual disease. Among 40 patients with MLNs, 16 (40%) attained axillary pCR. By molecular subtypes, pCR rates were 21.1% (4/19) in HR-positive/HER2-negative, 52.4% (11/21) in HER2-enriched, and 60.0% (3/5) in TNBC cases. The ΔSUVmax of primary tumors (r = 0.636; P 0.001) and MLNs (r = 0.560; P 0.001) were significantly correlated between PET2 and PET3. Patients achieving pCR demonstrated significantly lower absolute SUVmax and TBR values at both PET2 and PET3 compared to non-pCR cases (P 0.05). Furthermore, greater reductions in both SUVmax and TBR at any time point were strongly associated with pCR (P 0.05). In predicting pCR, the ΔSUVmax1 (AUC = 0.792; sensitivity = 55.6%; specificity = 92.6%) and ΔTBR1 (AUC = 0.774; sensitivity = 61.1%; specificity = 100%) of the primary tumor had the highest specificity; the TBR3 (AUC = 0.855; sensitivity = 94.1%; specificity = 76.0%) and ΔTBR2 (AUC = 0.815; sensitivity = 94.1%; specificity = 72.0%) of the primary tumor showed the highest sensitivity. In the HR-positive/HER2-negative subtype, parameters derived from PET2 (SUVmax2 and TBR2; AUC = 0.950; sensitivity = 100%; specificity = 86.7%) and all parameters derived from PET3 (SUVmax3, ΔSUVmax2, TBR3, and ΔTBR2; AUC = 1.000; sensitivity = 100%; specificity = 100%) of the primary tumor showed optimal AUC in predicting pCR; both the ΔSUVmax1 and ΔSUVmax2 of MLNs achieved 100% sensitivity for axillary pCR prediction. Conclusions: Early-phase 18F-FAPI-42 PET/MRI demonstrates high specificity in predicting pCR, which may facilitate the selection of alternative NAT regimens for non-pCR cases. Preoperative 18F-FAPI-42 PET/MRI shows high sensitivity for pCR prediction, which can identify potential candidates for surgery de-escalation. 18F-FAPI-42 PET/MRI exhibits excellent potential for both early and late-phase prediction of pathological response to NAT in HR-positive/HER2-negative BC. Citation Format: S. Liu, Y. Zhang, Y. Yang, D. Zhai, X. Kuang, Y. Shi, L. Yu, Y. Zhang, N. Shao, X. Zhang, Y. Lin. The Early and Late Predictive Value of 18F-FAPI-42 PET/MRI for Neoadjuvant Therapy Response in Breast Cancer: A Prospective 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 PS1-06-05.
Liu et al. (Tue,) studied this question.