Abstracts

RETURN TO ABSTRACT LISTING


E2192. Evaluation of Preoperative Computed Tomography in Staging Peritoneal Carcinomatosis Patients
Authors
  1. Baran Vardar; Dokuz Eylul University Medicine Faculty Hospital
  2. Cihan Agalar; Dokuz Eylul University Medicine Faculty Hospital
  3. Aras Canda; Dokuz Eylul University Medicine Faculty Hospital
  4. Selman Sokmen; Dokuz Eylul University Medicine Faculty Hospital
  5. Funda Obuz; Dokuz Eylul University Medicine Faculty Hospital
Objective:
It is important to preoperatively diagnose and stage peritoneal carcinomatosis (PC) patients for optimal patient selection for surgery and planning of cytoreductive surgery. The aim of this study was to investigate the correlation of preoperative computed tomography (CT) findings with intraoperative findings and evaluate the interobserver agreement between two radiologists.

Materials and Methods:
Two blinded radiologists evaluated preoperative CT-scans of 48 patients with peritoneal carcinomatosis from colorectal or ovarian origins. The peritoneal cancer index (PCI) was scored according to the Sugarbaker classification, based on lesion size and distribution. Intraoperative findings were regarded as the gold standard. The results were analyzed by the Wilcoxon rank-sum test, Spearman’s correlation test, Kendall’s tau-b test, and Cohen’s kappa test.

Results:
The mean PCI score was 12.73 (±6.92) for surgery, 11.08 (±6.34) for Radiologist 1 and 10.56 (±6.49) for Radiologist 2. A comparison of overall PCI scores showed a significant difference between surgical PCI scores and both radiologists’ preoperative CT scores (for both radiologists p=0.002). There were no significant differences between the two radiologists’ overall CT scores (p=0.123). Both radiologists highly correlated with surgical overall PCI (for Radiologist 1 r=0.860, p<0.001; for Radiologist 2 r=0.782, p<0.001) and very highly correlated between each others (r=0.921, p<0.001). Results by regions according to lesion sizes and PCI scores show a moderate and high correlation between two radiologists (correlation coefficients range from 0.660 to 0.894 at statistically significant levels p< 0.001). There were no significant differences between the two radiologists’ regional lesion sizes (p values range from 0.189 to 0.423). There were good and excellent agreements between two radiologists’ regional PCI scores (values vary between 0.624 and 0.853, p <0.001).

Conclusion:
CT is a reliable imaging technique in the preoperative staging of PC patients with its high interobserver agreement, although it underestimates intraoperative PCI.