User: Guest  Login
Document type:
journal article 
Eiber, M; Fingerle, AA; Brügel, M; Gaa, J; Rummeny, EJ; Holzapfel, K 
Detection and classification of focal liver lesions in patients with colorectal cancer: Retrospective comparison of diffusion-weighted MR imaging and multi-slice CT. 
To compare the diagnostic performance of diffusion-weighted MR imaging (DWI) with multi-slice CT (MS-CT) in the detection and classification of focal liver lesions in patients with colorectal cancer.In a retrospective study 68 patients who underwent DWI at 1.5 T (b-values of 50, 300 and 600s/mm(2)) and contrast-enhanced MS-CT were analysed by two radiologists blinded to the clinical results. Imaging results were correlated with intraoperative surgical and ultrasound findings (n=24), imaging follow-up or PET (n=44). Sensitivity of DWI and MS-CT in detection of focal liver lesions was compared on a per-lesion and a per-segment basis. Receiver operator-characteristic (ROC) curves to determine the diagnostic performance and the sensitivities of correctly identifying liver metastases on a segmental base were calculated.For lesion detection, DWI was significantly superior to MS-CT both on a per-lesion (difference in sensitivities for reader 1 and 2 22.65% and 19.06%, p<0.0001) and a per-segment basis (16.86% and 11.76%, p<0.0001). Especially lesions smaller than 10mm were better detected with DWI compared to MS-CT (difference 41.10% and 29.45%, p<0.0001). ROC-analysis showed superiority for lesions classification (p<0.0001) of DWI (AUC: 0.949 and 0.951) as compared to MS-CT (AUC: 0.879 and 0.892, p<0.0001 and p=0.005). DWI was able to filter out metastatic segments with a higher sensitivity (88.2 and 86.5%) compared to MS-CT (68.0 and 67.4%, p<0.0001 and p=0.005, respectively).Compared to MS-CT DWI is both more sensitive in the detection of liver lesions and more accurate in determining the extent of metastatic disease in patients with colorectal cancer and therefore might help to optimize therapeutic management in those patients. 
Journal title abbreviation:
Eur J Radiol 
Journal volume:
Journal issue:
Pages contribution:
TUM Institution:
Institut für Radiologie