The use of newer chemotherapeutic agents before resection of colorectal cancer liver metastases has been linked with parenchymal liver injury, in particular preoperative irinotecan and oxaliplatin with chemotherapy-associated steatohepatitis (CASH) and vascular parenchymal injury, respectively. We retrospectively assessed 334 cases from 2002 to 2007 and correlated pathologic findings with chemotherapy use and perioperative course. Features of fatty liver disease were graded according to established schemes, and several features of vascular injury, including sinusoidal dilation, nodular regenerative hyperplasia and parenchymal extinction lesions (PELs), were also scored semiquantitatively and a combined vascular injury (CVI) score was determined. Moderate and severe fatty injury was uncommon with steatohepatitis detected in 8 cases (2.4%), 7 of whom did not receive chemotherapy. Multivariate analysis showed steatosis greater than 33% and steatohepatitis were independently associated with Body Mass Index of 30 or more (P<0.001) but not chemotherapy. Vascular injuries were detected in 117 cases, were significantly associated with oxaliplatin, and the combined assessment of vascular features (a CVI score of 3 or more) was more strongly associated with oxaliplatin (P=0.0004) than any one feature in isolation. Perioperative outcome was not associated with parenchymal injury or preoperative chemotherapy. We conclude that although CASH is uncommon in this population vascular injury is frequently seen in resection specimens, but pathologic examination limited to sinusoidal dilation misses the majority of these. Semiquantitative measurement enables reproducible assessment of vascular injuries, allows comparison between studies, and may help inform future treatment decisions in patients with limited hepatic reserve.