Imaging Disease Biology: A Systematic Review on Radiological Predictors of Response to Chemotherapy in Colorectal Liver Metastasis

Introduction: Pathological response to neoadjuvant chemotherapy in colorectal liver metastases (CRLM) strongly correlates with disease-free and overall survival. While prior clinicopathological predictive models have been described, these provide only an approximation of disease biology. Pre-operative imaging offers a non-invasive morphological and functional assessment of response, with certain parameters outperforming traditional metrics. We present a systematic review of pre-operative PET-CT, CT and MRI in predicting metabolic response to chemotherapy and survival in colorectal liver metastasis. Methods: A systematic review was undertaken based on PRISMA guidelines. PubMed, Embase and Cochrane were searched up to 31/7/2021 for studies on the accuracy of PET-CT, CT and MRI in predicting metabolic response to chemotherapy and/or survival outcomes in CRLM. Results: From 515 citations, 22 studies were included. Amongst all modalities, PET-CT parameters were most predictive of outcomes. On PET-CT, reductions in maximum-standardized-uptake-value (SUVmax), total-lesion-glycolysis (TLG), tumor-glycolytic-volume (TGV) and metabolic-tumor-volume (MTV) were strongly correlated with chemotherapy response and survival. Predictive parameters on CT included lesion uniformity and entropy. On MRI, apparent-diffusion-coefficient (ADC) was predictive. Conclusion: Radiological quantification of metabolic activity as a predictor of chemotherapy response and survival is an emerging field of interest, with several parameters surpassing RECIST in prognostication. In this study, we report an increased heterogeneity of outcomes on CT and MRI. However, FDG-PET-CT parameters such as SUVmax, TLG, TGV and MTV were most predictive of response and survival. Overall, PET-CT provides an efficacious, non-invasive measure of disease biology and should be incorporated in multimodal management of colorectal liver metastases.