Predicting treatment response with MRI

A new scoring system for colorectal cancer liver metastases based on gadoxetic acid-enhanced MRI can help to predict treatment response to chemotherapy and possibly also estimate overall survival.

Colorectal cancer (CRC) is the second most common cancer in Europe. The liver is the most common metastatic site in patients with CRC. Nuno Costa (Medizinische Universität Wien, Austria) presented results of a retrospective study aiming to create a scoring system that can predict treatment response of CRC liver metastases to chemotherapy using gadoxetic acid-enhanced magnetic resonance imaging (gaMRI).

The scoring system

Lesions were scored based on homogeneity (homogeneous=0, mixed=1, heterogeneous=2); tumor-liver interface (sharp=0, mixed=1, ill-defined=2); and peripheral rim enhancement (≤2 mm=0, 2-4 mm=1, ≥4 mm=2). The hepatobiliary phase was used to assess homogeneity and tumor-liver interface while the peripheral rim enhancement was assessed using the arterial phase.
Scores correlate with residual vital tumor (RVT).

Patients with CRC liver metastases who underwent gaMRI after chemotherapy and before hepatic resection were included in the study. Stratification of the patient’s response was performed based on the score assigned to the lesion: optimal response (score ≤2), partial response (score >2 and <4) and no response (score ≥4). gaMRI-based response classification was compared to the percentage of residual vital tumor (RVT).

In total, 41 lesions in 30 patients were analyzed. The assignment of response groups based on the scoring system correlated well with the RVT percentage (p<0.001). In addition, response groups were analyzed for differences in overall survival (OS). There was a trend (p=0.066) towards differences in OS between patients classified as having optimal or partial response and those with no response. Patients in the no response group tended to have lower OS.


Response to therapy classified by a gaMRI-based scoring system correlated well with the RVT percentage, thereby offering a new approach to assess treatment response. Costa pointed out that “further prospective multicenter studies are needed to investigate if the pathological response translates into an accurate estimate of overall survival.”

Public Discussion

Costa was asked to comment on the chemotherapy regime received by the patients in the study and on the time point of MRI. Costa replied, that patients received different chemotherapy regimes, but chemotherapy regime was not identified as a confounder in the analysis. MRI was performed 6 to 8 weeks after start of chemotherapy. The exact time point for the MRI did not influence the results, Costa explained.