Background The selection of patients for cytoreductive surgery (CRS) and hyperthermic intra-peritoneal chemotherapy infusion (HIPEC) is important, and relies heavily on imaging. However, it has been reported that Computer Tomographic (CT) scans may only achieve a low sensitivity of 33% for peritoneal disease. We propose a set of radiological criteria for pre-operative determination of resectability of peritoneal disease in peritoneal-based malignancies and validate this in our cohort of patients. Methods A retrospective review of all patients who underwent laparotomy with a view for CRS and HIPEC, at the National Cancer Centre Singapore from January 2000 to April 2010, was performed. Intra-operative Peritoneal Cancer Index (PCI) scores were recorded. The pre-operative imaging was reviewed with a senior radiologist who was blinded, and recorded the radiological PCI scores (CT-PCI) and eight additional CT prognostic factors (CT-PF). The CT-PCI and CT-PF scores were then compared with the intra-operative findings to determine the radiological accuracy. The scores and the individual prognostic factors were then evaluated for their predictive ability for unresectability. Results Comparison of the CT-PCI and PCI scores showed a concordance correlation coefficient at 0.52 (95% CI 0.34-0.7). Accuracy was increased with the addition CT-PF. The presence of omental caking and ascites were predictors of unresectability. We propose a scoring system which is able to predict for unresectable disease with a specificity of 80% and a sensitivity of 62%. Conclusion With our proposed criteria, and scoring system, the selection of patients for CRS and HIPEC can be improved, and unnecessary exploratory operations avoided.