For the study, the researchers analyzed 79 liver tumors in 30 patients. MRI detected all 79 of the lesions, of which 33 were less than 1 cm. PET detected 65 lesions, of which only 12 were less than 1 cm. In a per-lesion analysis, MRI had an accuracy rate of 75.5% compared to 64.1% for PET.

"FDG PET has been shown to be an excellent tool for gastrointestinal cancer staging, but its role in the detection of liver metastases had not been established. Contrast-enhanced MRI, on the other hand, is considered a sensitive imaging technique for liver lesion depiction," said Dushyant V. Sahani, MD, lead author of the study. "We compared the two because we wanted to determine which imaging method was better for finding liver metastases," he said.

According to the study authors, patients who undergo successful liver resection have an expected five-year survival rate of about 33% and a five-year disease-free survival rate of 22%. "Knowing the number, size and location of the lesions in the liver is critical, because that is what determines treatment," said Dr. Sahani.

"FDG-PET is currently very popular for looking for metastases, but it has deficiencies. Based on our results, we believe that contrast-enhanced MRI should be the imaging method of choice for determining if patients with colon and pancreatic cancer have disease that has spread into their liver. However, FDG PET provides additional information about disease outside of the liver and sometimes both examinations may be complementary and indeed necessary for complete, accurate staging," said Dr. Sahani.

The study appears in the July 2005 issue of the American Journal of Roentgenology.