[18F]AlF-NOTA-FAPI-04 PET/CT for Predicting Pathologic Response of Resectable Esophageal Squamous Cell Carcinoma to Neoadjuvant Camrelizumab and Chemotherapy: A Phase II Clinical Trial

This single-center, single-arm, phase II trial (ChiCTR2100050057) investigated the ability of ¹⁸F-labeled fibroblast activation protein inhibitor ([¹⁸F]AlF-NOTA-FAPI-04, denoted as ¹⁸F-FAPI) PET/CT to predict the response to neoadjuvant camrelizumab plus chemotherapy (nCC) in locally advanced esophageal squamous cell carcinoma (LA-ESCC). Methods: This study included 32 newly diagnosed LA-ESCC participants who underwent ¹⁸F-FAPI PET/CT at baseline, of whom 23 also underwent scanning after 2 cycles of nCC. The participants underwent surgery after 2 cycles of nCC. Recorded PET parameters included maximum, peak, and mean SUVs and tumor-to-background ratios (TBRs), metabolic tumor volume, and total lesion FAP expression. PET parameters were compared between patient groups with good and poor pathologic responses, and the predictive performance for treatment response was analyzed. Results: The good and poor response groups each included 16 participants (16/32, 50.0%). On ¹⁸F-FAPI PET/CT, the posttreatment SUVs were significantly lower in good responders than in poor responders, whereas the changes in SUVs with treatment were significantly higher (all P < 0.05). SUV_max (area under the curve [AUC], 0.87; P = 0.0026), SUV_peak (AUC, 0.89; P = 0.0017), SUV_mean (AUC, 0.88; P = 0.0021), TBR_max (AUC, 0.86; P = 0.0031), and TBR_mean (AUC, 0.88; P = 0.0021) after nCC were significant predictors of pathologic response to nCC, with sensitivities of 63.64%–81.82% and specificities of 83.33%–100%. Changes in SUV_max (AUC, 0.81; P = 0.0116), SUV_peak (AUC, 0.82; P = 0.0097), SUV_mean (AUC, 0.81; P = 0.0116), and TBR_mean (AUC, 0.74; P = 0.0489) also were significant predictors of the pathologic response to nCC, with sensitivities and specificities in similar ranges. Conclusion: ¹⁸F-FAPI PET/CT parameters after treatment and their changes from baseline can predict the pathologic response to nCC in LA-ESCC participants.