Performance of imaging modalities in diagnosis of liver metastases from colorectal cancer: a systematic review and meta-analysis.

Surgery of liver metastases can be effective, and the appropriate selection of surgical candidates relies first on imaging. Different techniques are available, but information on their relative performance is unclear. The aim of this overview is to assess the imaging modality performance in the diagnosis of colorectal cancer (CRC) liver metastases. MEDLINE and EMBASE were searched for articles published from January 2000 to August 2008. Eligible trials had to be conducted on patients with diagnosis/suspicion of CRC liver metastases, comparing more than two modalities among MRI, computed tomography (CT), positron emission tomography using fluoro-18-deoxyglucose (FDG-PET), ultrasonography (US). Pooled estimates of sensitivity, specificity were calculated and pair-wise comparisons were performed. Of 6030 screened articles, 25 were eligible. Sensitivity and specificity on a per-patient basis for US, CT, MRI, and FDG-PET were 63.0% and 97.6%, 74.8% and 95.6%, 81.1% and 97.2, and 93.8% and 98.7%, respectively. On a per-lesion basis, sensitivity was 86.3%, 82.6%, 86.3%, and 86.0%, respectively. Specificity was reported in few studies. MRI showed a better sensitivity than CT in per-patient (odds ratio [OR]: 0.69; 95% confidence interval [CI]: 0.47-0.99; P = 0.05) and in per-lesion analysis (OR: 0.66; 95% CI: 0.55-0.80; P < 0.0001). In per-lesion analysis, the difference was higher when liver-specific contrast agents were administered. Available evidence supports the MRI use for the detection of CRC liver metastases.