Diagnostic Performance of 18F Fibroblast Activation Protein Inhibitor (FAPI)-42 PET Compared with Multiparametric MRI for Glioma Grading

Introduction: Both diffuse-weighted MR imaging (DWI) and ¹⁸F-fibroblast activation protein inhibitor (FAPI)-42 provide grading information in cerebral gliomas. This study aimed to compare the diagnostic value of ¹⁸F–FAPI-42 PET and DWI for tumor grading in patients with gliomas.

Methods: Twenty-two patients (9 newly diagnosed and 13 recurrence patients) with pathologically confirmed gliomas [7 low-grade gliomas (LGGs), and 15 high-grade gliomas (HGGs)] were investigated with ¹⁸F–FAPI PET and DWI study. The minimum apparent diffusion coefficient (ADCmin) in DWI, tumor-to-brain ratios of mean and maximum standard uptake value (TBRmean, TBRmax) in¹⁸F- FAPI-42 PET were recorded. The nomogram was constructed using multivariate logistic regression combined with ADCmin, TBRmax, and TBRmean. The sensitivity, specificity, positive predictive values (PPV), negative predictive values (NPV), and diagnostic accuracies of ADCmin, TBRmax, TBRmean, and nomogram for differentiating LGGs from HGGs were compared. The diagnostic accuracies were evaluated by receiver operating characteristic (ROC) analyses and the value of area under the curve (AUC).

Results: Twenty-two patients (median age,39.5 years; interquartile range, 32–50 years) were included, of which 73% (16 of 22) underwent open resection. There were significant differences in all parameters, including ADCmin, TBRmax, TBRmean between LGGs and HGGs (P=0.013 vs. P=0.020 vs. P=0.014, respectively). The specificities of ADCmin, TBRmean, and nomogram were similar(all 85.7%) and higher than that of TBRmax (71.4%). The sensitivities of ADCmin, TBRmax, and nomogram were similar (all 93.3%) and higher than that of TBRmean (73.3%). The diagnostic accuracies of ¹⁸F–FAPI-42 and DWI to discriminate LGGs from HGGs were similar with the highest AUC values for TBRmax and TBRmean of ¹⁸F–FAPI PET [85.7(95% CI: 68.8,100) vs. 82.4(95% CI: 64.6,100), respectively] and for ADCmin of DWI [87.6 (95% CI: 65.1,100)]. The nomogram model with an AUC of 90.5% (95% CI: 73.3,100) could further improve diagnostic accuracy.

Conclusions: Both ¹⁸F–FAPI-42 PET and DWI discriminated LGGs from HGGs with similar diagnostic performance. The nomogram model based on 18F-FAPI-42 PET and DWI might be clinically useful in non-invasive grade classification for glioma patients.