Accuracy of liver-specific magnetic resonance imaging as a predictor of chemotherapy-associated hepatic cellular injury prior to liver resection

• Published in: European journal of surgical oncology 2009-10, Vol.35 (10), p.1085-1091
• Main Authors: O'Rourke, T.R, Welsh, F.K.S, Tekkis, P.P, Lyle, N, Mustajab, A, John, T.G, Peppercorn, D, Rees, M
• Format: Article
• Language: English
• Subjects: Adult; Aged; Antineoplastic Agents - adverse effects; Chemical and Drug Induced Liver Injury - etiology; Chemical and Drug Induced Liver Injury - pathology; Chemotherapy; Colorectal Neoplasms - pathology; Contrast Media; Dextrans; Fatty Liver - chemically induced; Fatty Liver - pathology; Female; Ferrosoferric Oxide; Hematology, Oncology and Palliative Medicine; Hepatectomy; Hepatotoxicity; Humans; Linear Models; Liver Diseases - pathology; Liver Neoplasms - drug therapy; Liver Neoplasms - pathology; Liver Neoplasms - secondary; Liver Neoplasms - surgery; Magnetic Resonance Imaging - methods; Magnetite Nanoparticles; Male; Middle Aged; MRI; Multivariate Analysis; Neoadjuvant Therapy - adverse effects; Preoperative Care; Sensitivity; Sensitivity and Specificity; Single-Blind Method; Specificity; Surgery
• ISSN: 0748-7983; 1532-2157
• DOI: 10.1016/j.ejso.2009.01.015

Abstract Background/aims To evaluate the diagnostic precision of chemical-shift imaging MRI and ferucarbotran-enhanced MRI for hepatic parenchymal injury prior to hepatic resection for colorectal metastases. Methods Preoperative MRI criteria were used to score 37 patients with colorectal liver metastases by two independent radiologists, blinded to outcomes, for signal drop-out on chemical-shift imaging MRI and ferucarbotran uptake and compared to blinded standardized histopathological endpoints of steatosis, steatohepatitis and sinusoidal dilatation. Sensitivity, specificity, predictive values and the area under the receiver operating characteristic curve (AUC) were calculated for the MRI sequences. Results On histology, severe steatosis, steatohepatitis and sinusoidal dilation were evident in 6 (16.2%), 4 (10.8%) and 9 (24.3%) patients respectively. Chemical-shift imaging MRI had a positive predictive value (PPV) of 100% for severe steatosis, 80% for steatohepatitis and zero for sinusoidal dilatation, with an AUC of 1.0, 0.99 and 0.36 respectively. Ferucarbotran-enhanced MRI had a 100% PPV for the detection of severe sinusoidal dilatation, with an AUC of 0.61. Conclusions This study demonstrates that liver-specific MRI can accurately predict the severity of pre-existing hepatic injury. Moreover, it may play a key role in planning the timing and extent of chemotherapy and hepatic resection for colorectal metastases.