A head-to-head comparison of 68Ga-FAPI PET/MRI(CT) and contrast-enhanced CT in Gastric Cancer Initial Staging

Introduction: ⁶⁸Ga-FAPI PET has been proven to perform well in the diagnosis and staging of diverse solid malignant tumors, including the diagnosis of gastric cancer. However, the value of ⁶⁸Ga-FAPI PET for gastric cancer staging remains uncertain. Contrast-enhanced CT(CECT) is recommended by NCCN guidelines as conventional imaging for gastric cancer staging. This study aimed to evaluate the performance of ⁶⁸Ga-DOTA-FAPI-04 (⁶⁸Ga-FAPI) PET/MRI(CT) for the diagnosis of primary tumor and initial staging in patients with gastric carcinomas and to compare with CECT.

Methods: Patients who underwent ⁶⁸Ga-FAPI PET/MRI(CT) from June 2020 to June 2023 were retrospectively collected. Inclusion criteria: (1) histologically proven gastric carcinomas; (2) underwent CECT within 1 month; (3) no anti-tumor treatment before ⁶⁸Ga-FAPI PET/MRI(CT) or CECT and between the two modalities. Exclusion criteria: (1) histologically proven non epithelial gastric malignancies including gastric lymphoma, etc; (2) suffering from other malignancies; (3) lack of clinical or follow-up data. Histopathology or follow-up data served as the standard. The detection rate of ⁶⁸Ga-FAPI PET/MRI(CT) for primary tumors was evaluated, and the diagnostic performance for regional lymph nodes and distant metastasis was assessed and compared with CECT. The patients received D2 lymphadenectomy were analyzed to compare the detection efficiency of ⁶⁸Ga-FAPI PET/MRI and CECT for regional lymph nodes metastasis. Peritoneal Cancer Index (PCI) was used for the evaluation of peritoneal metastases. The Mcnemar test was used to compare the detection efficiency of two modalities.

Results: A total of 62 patients (range: 25–85 y) were enrolled, including 43 men and 19 women. The detection rate of ⁶⁸Ga-FAPI PET for primary tumor was 98.4% (61/62). The missed case was T1 stage. Fifty-five patients were finally diagnosed with regional lymph node metastasis. The sensitivity and accuracy of ⁶⁸Ga-FAPI PET/MRI(CT) were 76.4% (42/55) and 79.0% (49/62), respectively, which are lower than those of CECT (85.5% [47/55] and 85.5% [53/62]), although there was no significant difference(p>0.05). Among the 18 patients received D2 lymphadenectomy, ⁶⁸Ga-FAPI PET/MRI and CECT both correctly diagnosed 11 cases and misdiagnosed 7 cases. ⁶⁸Ga-FAPI PET/MRI correctly staged 7 cases and underestimated 11 cases. CECT correctly staged 6 cases, underestimated 10 cases, and overestimated 2 cases. The sensitivity of ⁶⁸Ga-FAPI PET/MRI for regional lymph node metastases was similar to CECT (21.2% [21/99] vs. 20.2% [20/99], p>0.05). Excluding 3 patients who only underwent abdominal CECT, 32 patients were finally diagnosed with distant metastasis. The sensitivity, specificity and accuracy of ⁶⁸Ga-FAPI PET/MRI(CT) and CECT for distant metastasis showed no significant difference (100% [32/32] vs. 87.5% [28/32], 92.6% [25/27] vs. 96.4% [26/27], 96.6% [57/59] vs. 91.5% [54/59], p>0.05). Based on lesions,⁶⁸Ga-FAPI PET/MRI(CT) detected more distant lymph nodes (35 vs.32), bone (25 vs.13), liver (11 vs. 10) and ovarian (6 vs.4) metastases. Compared with CECT, ⁶⁸Ga-FAPI PET/MRI(CT) observed more areas of peritoneal metastases in each patient(5 vs. 3, p<0.01) and led to higher median PCI score(10 vs. 6, p<0.01). Ultimately, ⁶⁸Ga-FAPI PET/MRI(CT) revised 2 cases of regional lymph node metastasis and 3 cases of distant metastasis diagnosis compared with CECT, which increased the clinical staging of 3 cases from stage III to IVB and changed their managements.

Conclusions: In this group of gastric carcinoma patients, ⁶⁸Ga-FAPI PET/MRI(CT) has high sensitivity for primary tumors. The diagnostic performance of ⁶⁸Ga-FAPI PET/MRI(CT) for distant metastasis was comparable to CECT based on patient, but outperformed CECT in visualizing the lesions. However, due to the false negatives, ⁶⁸Ga-FAPI PET/MRI(CT) has no advantage over CECT in the evaluation of regional lymph nodes. More large sample multi-center prospective head-to-head comparisons are needed for further validation.

• Download figure
• Open in new tab
• Download powerpoint