CT attenuation of liver metastases before targeted therapy is a prognostic factor of overall survival in colorectal cancer patients. Results from the randomised, open-label FIRE-3/AIO KRK0306 trial.

ABSTRACT

OBJECTIVES: To assess the prognostic value of pre-therapeutic computed tomography (CT) attenuation of liver metastases for overall survival (OS) in metastatic colorectal cancer (mCRC). METHODS: In the open-label, randomised, prospective phase-III FIRE-3 trial, patients with histologically confirmed mCRC received fluorouracil (5-FU), leucovorin and irinotecan (FOLFIRI) with either cetuximab or bevacizumab. Participating patients gave written informed consent prior to study entry. In CT at baseline (portal venous phase, slice thickness ≤5 mm), mean attenuation [Hounsfield units (HU)] of liver metastases was retrospectively assessed by semi-automated volumetry. Its prognostic influence on OS was analysed in Kaplan-Meier-analysis and Cox proportional hazard regression and an optimal threshold was determined. RESULTS: In FIRE-3, 592 patients were enrolled between 2007 and 2012. Among the 347 patients eligible for liver volumetry, median baseline CT attenuation of liver metastases was 59.67 HU [interquartile range (IQR), 49.13, 68.85]. Increased attenuation was associated with longer OS {per 10 HU hazard ratio (HR), 0.85 [95% confidence interval (CI), 0.78, 0.93], p < 0.001}. The optimised threshold (≥61.62 HU) was a strong predictor for increased OS [median, 21.3 vs 30.6 months; HR, 0.61 (95% CI, 0.47, 0.80), p < 0.001]. Multivariate regression controlling for correlated and further prognostic factors confirmed this [HR, 0.60 (95% CI, 0.45, 0.81), p = 0.001]. Furthermore, mean attenuation ≥61.62 HU was significantly associated with increased early tumour shrinkage (p = 0.002) and increased depth of response (p = 0.012). CONCLUSIONS: Increased mean baseline CT attenuation of liver metastases may identify mCRC patients with prolonged OS and better tumour response. KEY POINTS • In colorectal cancer, increased attenuation of liver metastases in baseline computed tomography is a prognostic factor for prolonged OS (p < 0.001). • A threshold of ≥61.62 HU was determined as optimal cut-off to identify patients with prolonged OS (p < 0.001), early tumour shrinkage (p = 0.002) and increased depth of response (p = 0.012).