Report from SMI: Dynamic MR imaging predicts responsiveness in rectal cancer patients

New evidence presented this week at the 2005 Society for Molecular Imaging confirms that perfusion index measures obtained with dynamic contrast-enhanced MRI predict whether rectal cancer patients will benefit from chemoradiation.

Microcirculatory measurements possible with DCE-MRI are key to reliable prediction of therapy responses, according to Dr. Alexander De Vries, an imaging researcher in the department of radiotherapy and radio-oncology at the Medical University of Innsbruck in Austria.

De Vries examined the relative diagnostic power for fMRI apparent diffusion coefficient (ADC) mapping and perfusion indexing (PI) for two groups of patients with grade II rectal cancer. Both groups were prescribed chemoradiation preceding surgical resection.

Morphology MRI had shown that the primary tumor had not invaded the perirectal tissue in any of the cases. No evidence of metastatic disease or secondary malignancies was found. Previous therapy had not altered microcirculation associated with the tumors.

De Vries generated perfusion indexes for 79 patients based on the DCE-MRI performed several days before chemoradiation therapy and resection. Using evaluation criteria developed in the past five years in his Innsbruck lab, he classified 38 patients as responders and 41 as nonresponders.

Responders had a mean PI value of 7.4 mL/min/100 g. The 41 nonresponders had a mean PI of 9.8.

The sensitivity and specificity of PI were 73% and 100%, respectively, and its positive predictive value was 100%. ADC mapping predicted a response among 18 patients and a lack of response among 16 subjects with the disease.

High PI values most likely correlate with increased tissue permeability, suggesting increased angiogenic activity and the development of microcirculation, according to De Vries. High ADC measures suggest the presence of necrotic tissue and rapid cell proliferation.

The results led De Vries to conclude that PI measurements alone can be trusted to predict which patients would benefit from chemoradiation. None of the 30 patients classified as nonresponders using PI measurements ultimately responded positively to chemoradiation therapy. Eleven patients thought to be responders ultimately did not respond to the therapy.

Without other risk factors such as positive lymph nodes, 75% of the patients could have been spared the side effects and cost of undergoing chemoradiation, De Vries said.

"We think that the success of chemotherapy is primarily determined by the extent of microcirculation that can be measured prior to therapy," he said. "We think that dynamic contrast MRI is superior to morphologic MR in understanding tumor behavior. It also can be helpful on an individual basis regarding when chemoradiation will be helpful."