Comparison of diagnostic performance between MRI and ⁶⁸Ga-FAPI-04 PET/CT for stratified preoperative TNM staging of rectal cancer: a single-centre retrospective cohort study

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Background MRI exhibits inherent limitations for preoperative TNM staging of rectal cancer, partially limiting its clinical utility. This head-to-head study compared the performance of ⁶⁸Ga-FAPI-04 PET/CT and MRI for stratified preoperative TNM staging of rectal cancer. We further assessed the predictive efficacy of quantitative PET/CT parameters for differentiating T stages, and explored whether these quantitative indicators could act as an adjunct to MRI staging. Methods This single-center retrospective study recruited 81 patients diagnosed with rectal cancer, all of whom received whole-abdominal MRI and whole-body ⁶⁸Ga-FAPI-04 PET/CT scanning. Multiple quantitative metabolic parameters, including SUVmax, SUVmean, MTV, TLU, and TBR, were extracted for subsequent analyses. McNemar’s test was adopted to compare the sensitivity, specificity and diagnostic accuracy of the two imaging modalities in preoperative TN staging. Regarding M staging evaluation, the detection performance of metastatic lesions was only compared within the overlapping abdominopelvic scanning range of the two imaging protocols. All lesions located outside this overlapping coverage zone were merely described qualitatively without inter-group comparison. Spearman correlation analysis and binary logistic regression were applied to explore the correlations between T stage and the aforementioned quantitative parameters. ROC and DCA curves were plotted to quantify diagnostic performance and corresponding clinical net benefit, respectively. Results In T2-stage assessment, MRI demonstrated superior specificity and accuracy relative to PET/CT (85.25% vs. 59.02%; 86.42% vs. 62.96%). Similarly, MRI yielded a greater overall accuracy for T3 staging(83.95% vs. 69.14%). All the above intergroup differences reached statistical significance (all P < 0.001). For N1 nodal staging, PET/CT showed higher specificity (87.72% vs. 73.68%, P = 0.039) and accuracy (80.25% vs. 62.96%, P = 0.004) compared with MRI. Within the overlapping abdominopelvic scan range, PET/CT identified more metastatic lesions than MRI (6 versus 3). Additionally, PET/CT detected 14 metastatic lesions that were invisible on MRI beyond the abdominopelvic coverage area. Distinct values of all PET/CT quantitative parameters were observed across different T categories (all P < 0.001). TLU presented the most prominent positive linear association with T stage ( r = 0.752, P < 0.001). Binary logistic regression analysis preliminarily indicated a statistical correlation between TLU and T3 disease status (all P < 0.05). When distinguishing T1/T2 low-stage tumors from T3 lesions, TLU generated an AUC value of 0.889. The optimal threshold of 64.354 provided a sensitivity of 81.00% and specificity of 86.20%. Combined parameter groups including SUVmax+TLU and ADC+TLU produced AUCs of 0.883 and 0.935, respectively. At their respective optimal cut-off values, both combined parameter groups yielded a sensitivity of 78.60% and specificity of 93.10%. Decision curve analysis further indicated that the ADC+TLU combined parameter group provided the maximum net clinical benefit throughout the 0–50% probability threshold interval. Conclusion MRI remains the primary imaging modality for local T-staging evaluation, especially for T2 and T3 rectal lesions. ⁶⁸Ga-FAPI-04 PET/CT achieves superior diagnostic performance in N1 nodal staging and the identification of distant metastases. Quantitative indicators derived from ⁶⁸Ga-FAPI-04 PET/CT, particularly TLU and multiple combined indicators, preliminarily indicate favorable diagnostic ability and carry potential auxiliary value to assist MRI in differentiating distinct T-stage subgroups.