Use of preoperative liver biopsy to predict livers that have avoided chemotherapy-induced liver injury.

Abstract

393 Background: Liver parenchymal damage caused by neoadjuvant chemotherapy may increase the surgical risk. The aim of the present study was to assess the performance of preoperative liver biopsy (PLB) in predicting the status of the liver parenchyma at the time of resection. Methods: Among 40 patients who underwent liver resection after neoadjuvant chemotherapy for colorectal liver metastases (CLM) in 2009, 28 had percutaneous PLB from 2 to 6 weeks after onset of the treatment. Patients underwent a mean number of 9 cycles of chemotherapy (range 4-22). Chemotherapy included oxaliplatin (n=22), irinotecan (n=10), and bevacizumab (n=12). PLB and surgical specimens were compared according to the presence and rate of steatosis, sinusoidal obstruction syndrome (SOS) and chemotherapy-associated steatohepatitis (CASH). Results: Surgical specimen analysis showed steatosis in 12 (43%) patients (ranging from 5 to 80%), SOS in 12 patients (43%) and CASH in 1 patient (3%). The sensitivity, specificity, positive predictive value and negative predictive value of PLB for steatosis were 67%, 81%, 73% and 77%, respectively. It was 50%, 94%, 86% and 71%, respectively for SOS and 0%, 100%, 0% and 96%, respectively for CASH. Performance of PLB was not related to the type of chemotherapy. Among the 28 patients, 8 patients had no abnormalities on PLB. In 7 of them, surgical specimen analysis was considered normal. Conclusions: The accuracy of PLB in predicting chemotherapy- induced liver injury appears to be low. It is useful, however, given its high specificity when the PBL is normal, for encouraging hepatic resection without delay. No significant financial relationships to disclose.