Limited diagnostic performance of imaging evaluation for staging in gastric-type endocervical adenocarcinoma: a multi-center study

Abstract

[Purpose]The purposes of the study are to assess the diagnostic performance of preoperative imaging for staging factors in gastric-type endocervical adenocarcinoma (GEA) and to compare the performance for GEA with that of usual-type endocervical adenocarcinoma (UEA) among patients preoperatively deemed locally early stage (DLES) (< T2b without distant metastasis).[Materials and methods]For this multi-center retrospective study, 58 patients were enrolled. All had undergone MRI with or without CT and FDG PET-CT preoperatively and had been pathologically diagnosed with GEA at five institutions. Based on the medical charts and radiological reports, the diagnostic performances of preoperative imaging for the International Federation of Gynecology and Obstetrics staging factors were assessed retrospectively. Next, the imaging performance was assessed in preoperatively DLES-GEA (n = 36) and DLES-UEA (n = 136, with the same inclusion criteria). The proportions of underestimation of GEA and UEA were compared using Fisher’s exact test.[Results]Imaging diagnostic performance for GEA was limited, especially for sensitivity: parametrial invasion, 0.49; vaginal invasion, 0.54; pelvic lymph node metastasis (PELNM), 0.48; para-aortic lymph node metastasis, 0.00; and peritoneal dissemination, 0.25. Among preoperatively DLES patients, the proportions of underestimation were significantly higher in GEA than in UEA; parametrial invasion, 35% vs. 5% (p < 0.01); vaginal invasion, 28% vs. 6% (p < 0.01); PELNM, 24% vs. 6% (p < 0.05); peritoneal dissemination, 6% vs. 0% (p < 0.05).[Conclusion]At present, preoperative imaging diagnostic performance for staging factors in GEA does not meet clinical expectations, especially for sensitivity. Among patients preoperatively DLES, the proportions of underestimation in GEA were significantly higher than in UEA. Future incorporation of approaches specifically emphasizing GEA is desirable to improve imaging performance.