Aim: To assess the place of computed tomography (CT) of the colon in frail or elderly patients with symptoms suggestive of colon cancer.
Method: A total of 195 patients (median age 76 years) underwent CT of the abdomen and pelvis following the administration of positive oral contrast medium but no bowel preparation. All had symptoms suggestive of colon cancer. CT findings were classified as normal/diverticular disease (DD), possible colon cancer, definite colon cancer or extracolonic pathology. Accuracy of CT was assessed against patient outcome. Association between symptoms and colon cancer was assessed by chi-squared test.
Results: There were 47 deaths and median follow up for those alive was 16 months. Overall sensitivity of CT was 100% and specificity 87% for detection of colon cancer. One hundred and ten normal/DD CT examinations had no significant bowel lesion on follow up. Of 12 cases defined as 'definite cancers' on CT, there were nine colon cancers, two extracolonic cancers, and one normal. Of 23 'possible cancers' on CT, there were two colon cancers, three DD masses and 18 normal/DD. Fifty examinations had extracolonic findings including 33 (17%) cases of significant abdominal disease. CT findings led to a halt in investigations in 115 cases (59%), colonoscopy in 18 (9%) cases and surgery in 16 (8%) cases. None of the symptoms present showed a significant association with colon cancer (all P > 0.05).
Conclusion: Minimal preparation CT is a non-invasive and sensitive method for investigating colon cancer in frail or elderly patients. It has a 100% negative predictive value and also detects a large number of extracolonic lesions.
Copyright 2002 The Royal College of Radiologists.