Is routine magnetic resonance imaging justified for the early detection of resectable liver metastases from colorectal cancer?

PURPOSE: This study was designed to determine whether routine follow-up by magnetic resonance imaging improves the detection of resectable liver metastases from colorectal cancer and patients' survival. METHODS: Patients who underwent curative surgery for colorectal cancer were included in a program of liver surveillance by routine magnetic resonance imaging, in addition to the standard follow-up protocol consisting of clinical examination and biochemical tests. The median follow-up was 41 (interquartile range, 30-53) months, with a median magnetic resonance imaging surveillance period of 20 (interquartile range, 12-27) months. Cases were analyzed for mode of diagnosis, resectability, and overall survival. RESULTS: Liver metastases were found in 37 (13 percent) of 293 patients studied. Magnetic resonance imaging diagnosed hepatic metastases with 84 percent sensitivity and 90 percent specificity. In 28 (76 percent) patients, carcinoembryonic antigen and/or liver function tests were abnormally elevated and 5 patients (14 percent) were symptomatic. Hepatic resection was possible in only nine patients (24 percent). Magnetic resonance imaging detected all resectable cases, whereas traditional follow-up would have missed three (33 percent) cases suitable for surgery. CONCLUSIONS: Although magnetic resonance imaging surveillance increased the number of patients suitable for liver resection by 50 percent, these represented only 1 percent of the patients included in the study. Whether these results are enough to justify the allocation of expensive resources is controversial.