Extended abdominopelvic MRI versus CT at the time of adnexal mass characterization for assessing radiologic peritoneal cancer index (PCI) prior to cytoreductive surgery

• Published in: Abdominal radiology (New York) Vol. 44; no. 6; pp. 2254 - 2261
• Main Authors: Gadelhak, Basma, Tawfik, Ahmed M., Saleh, Gehad A., Batouty, Nihal M., Sobh, Donia M., Hamdy, Omar, Refky, Basel
• Format: Journal Article
• Language: English
• Published: New York Springer US 01-06-2019
• Subjects: ABDOMEN; COMPUTERIZED TOMOGRAPHY; Gastroenterology; Hepatology; Imaging; Medicine; Medicine & Public Health; NEOPLASMS; NMR IMAGING; OVARIES; PATIENTS; PELVIS; Peritoneum; Radiology; RADIOLOGY AND NUCLEAR MEDICINE; SURGERY; Peritoneal cancer index; Ovarian and fallopian tube malignancy; MRI; Diffusion weighted imaging
• ISSN: 2366-004X; 2366-0058; 2366-0058
• DOI: 10.1007/s00261-019-01939-y

Purpose
To evaluate whether extending the MRI scan to include the abdomen at the time of adnexal mass characterization could replace additional CT for peritoneal cancer index (PCI) assessment.
Methods
After institutional review board approval for this prospective study, 36 consecutive females with ovarian and FT malignancies were included. All patients signed an informed consent. Patients underwent preoperative CT (32 patients) and MRI (36 patients). Images were interpreted by 2 independent observers. Surgical data were available in 27 patients. Region-by-region analysis was performed for detection rates of peritoneal carcinomatosis (PC). Inter-observer agreement for each region was evaluated by kappa statistics. Radiologic PCI was calculated by CT and MRI independently and inter-observer agreement for CT and MRI as well as agreement between radiologic and surgical PCI were evaluated by weighted-kappa statistics.
Results
On region-by-region analysis, the highest detection rates of PC were noted at the central abdomen and pelvis. Detection rates were higher by MRI than CT, mainly in bowel serosal surface, pelvis, and right upper abdomen regions. Inter-observer agreement of MRI was higher than CT in most regions. The median PCI by CT was 5 and 4 for the first and second observers (range 0–21 for both observers), respectively. The median PCI by MRI was 6 (range 0–23 for both observers). The inter-observer agreement of PCI was excellent by both CT and MRI (
k
 = 0.876 and
k
 = 0.912, respectively). The agreement between CT and surgical PCI was 0.660 and 0.590 for the first and second observers, respectively. The agreement between MRI and surgical PCI was 0.797 and 0.798 for the first and second observers, respectively.
Conclusions
Extending MRI scan to include the abdomen at the time of adnexal mass characterization allows accurate estimation of PC, with better results than CT, obviating the need for dedicated CT scan of abdomen and pelvis for imaging of PC.