Detection of liver metastases under 2 cm: comparison of different acquisition protocols in four row multidetector-CT...

This study compared different acquisition protocols performance to detect small liver metastases (<2 cm). Thirty consecutive patients with histologically proven hepatic metastases were explored by MDCT at the liver equilibrium phase by four successive acquisitions. We compared the following protocols (1-4): 5/30/1.5 (section thickness/table speed/pitch); 5/15/0.75; 5/11.25/0.75; and 2.5/15/1.5 with the same X-ray dose. The gold standard was based on patient radiological follow-up. Evolutive lesions were considered as true positive (TP). The described lesions, not found on the follow-up exams despite tumoral progression, were considered as false positive (FP). Stable lesions could not be considered as metastasis and were eliminated. One hundred and seventy-six lesions were detected: 61 TP and 91 FP. Twenty-four lesions were eliminated. The mean kappa values for protocols 1, 2, 3 and 4 were, respectively, 0.43, 0.68, 0.73 and 0.51 (0.61-0.80: substantial agreement) and the mean areas under the ROC curve were, respectively, 0.76, 0.87, 0.86 and 0.80. The results of protocols 2 and 3 were significantly superior to those of protocols 1 and 4. MDCT protocols using thin sections or an increased table speed are less efficient in detecting small metastases.