Abstract
BACKGROUND: Accurately and promptly predicting the response of gastrointestinal stromal tumors (GISTs) to targeted therapy is essential for optimizing treatment strategies. However, some fractions of recurrent or metastatic GISTs present as non-FDG-avid lesions, limiting the value of [(18)F]fluorodeoxyglucose positron emission tomography/computed tomography ([(18)F]FDG PET/CT) in treatment evaluation. This study evaluated the efficacy of [(18)F]F-fibroblast activation protein inhibitor (FAPI)-42 [(18)F]FAPI-42) PET/CT for assessing the treatment response in recurrent or metastatic GISTs, in comparison to [(18)F]FDG PET/CT and explores a model integrating PET/CT imaging and clinical parameters to optimize the clinical use of these diagnostic tools. METHODS: Our retrospective analysis included 27 patients with recurrent or metastatic GISTs who underwent [(18)F]FAPI-42 PET/CT and [(18)F]FDG PET/CT at baseline before switching targeted therapy. Treatment response status was divided into a progression group (PG) and a non-progression group (NPG) based on the Response Criteria in Solid Tumors (RECIST) 1.1, according to the contrast-enhanced computed tomography (CT) scan at six months. [(18)F]FAPI-42 and [(18)F]FDG PET/CT parameters including the mean standardized uptake value (SUV(mean)), the standard uptake value corrected for lean body mass (SUL(peak)), the maximum standardized uptake value (SUV(max)), tumor-to-blood pool SUV ratio (TBR), tumor-to-liver SUV ratio (TLR), metabolic tumor volume (MTV)/FAPI-positive tumor volume (GTV-FAPI), total lesion glycolysis (TLG)/FAPI-positive total lesion accumulation (TLF) were correlated with the response status to identify indicative of treatment response. The predictive performance of them was quantified by generating receiver operating characteristic curves (ROC), calibration curves, and cross-validation. RESULTS: A total of 110 lesions were identified in 27 patients. Compared with PG, NPG was associated with lower levels of TBR and SUV(mean) in FDG PET/CT (TBR-FDG, SUV(mean)-FDG; P=0.033 and P=0.038, respectively), with higher SUL(peak) and TLF in FAPI PET/CT (SUL(peak)-FAPI, TLF-FAPI; P=0.10 and P=0.049, respectively). The predictive power of a composite-parameter model, including TBR-FDG, SUL(peak)-FAPI, gene mutation, and type of targeted therapy [area under the curve (AUC) =0.865], was superior to the few-parameter models incorporating TBR-FDG (AUC =0.637, P<0.001), SUL(peak)-FAPI (AUC =0.665, P<0.001) or both (AUC =0.721, P<0.001). CONCLUSIONS: Both [(18)F]FAPI-42 PET/CT and [(18)F]FDG PET/CT have value in predicting the treatment response of recurrent or metastatic GISTs. And [(18)F]FAPI-42 PET/CT offers synergistic value when used in combination with [(18)F]FDG PET/CT. Notably, the nomogram generated from the model incorporating [(18)F]FAPI-42 PET/CT, [(18)F]FDG PET/CT parameters, gene mutation, and type of targeted therapy could yield more precise predictions of the response of recurrent metastatic GISTs.