Comparison of FDG-PET/CT and MR with diffusion-weighted imaging for assessing peritoneal carcinomatosis from gastrointestinal malignancy.

Abstract

To assess the accuracy of FDG-PET/CT and MR with diffusion-weighted imaging (MR-DWI) for diagnosing peritoneal carcinomatosis (PC) from gastrointestinal malignancies. Thirty consecutive patients referred for staging of gastrointestinal malignancy underwent FDG-PET/CT and MR-DWI in this retrospective study. Extent of PC was characterised by dividing the peritoneal cavity into three sites in each patient: right and left supramesocolic areas and inframesocolic level (total 90 sites). Presence of PC was confirmed either by surgery (18/30) or by follow-up (12/30). PC was confirmed in 19 patients (19/30). At a total of 90 sites, 27 showed proven PC. On a patient-based analysis, sensitivity, specificity, PPV, NPV and accuracy were respectively 84%, 73%, 84%, 73% and 80% for PET/CT and 84%, 82%, 89%, 75% and 83% for MR-DWI. On a site-based analysis, overall sensitivity and specificity of PET/CT (63%, 90%) and MR-DWI (74%, 97%) were not statistically different (P = 0.27). In the supramesocolic area, MR-DWI detected more sites involved than PET/CT (7/9 vs. 4/9). The sensitivities of PET and MR were lower for subcentimetre tumour implants (42%, 50%). Interobserver agreement was very good for PET/CT and good for MR-DWI. FDG-PET/CT and MR-DWI showed similar high accuracy in diagnosing PC. Both techniques underestimated the real extent of PC because of decreased sensitivity for subcentimetre lesions. FDG-PET/CT and MR-DWI showed similar high accuracy for diagnosing peritoneal carcinomatosis. • In the supramesocolic area, MR-DWI could be more sensitive than PET/CT. • Both techniques showed lower sensitivity for subcentimetre lesions. • Interobserver agreement was very good for PET/CT and good for MR-DWI.