Preoperative classification of histopathological growth patterns in colorectal liver metastases using prospective 18 F-FDG PET/CT

Abstract Colorectal liver metastases (CLM) exhibit distinct histopathological growth patterns (HGPs) at the metastasis-liver interface, clinically relevant for prognosis and potential predictive value. This study investigates the feasibility of using Fluorine-18 Fluorodeoxyglucose Positron Emission Tomography Computed Tomography ( 18 F-FDG PET/CT) to preoperatively classify HGPs in CLM and secondly investigate PET parameters as prognostic markers for overall survival (OS). In a prospective study at Umeå University Hospital (June 2013-December 2017), 78 patients with resectable CLM underwent contrast-enhanced liver MRI and 18 F-FDG PET/CT imaging. HGP classification was performed on tissue sections according to 2017 international guidelines. PET features (SUV mean [standardised uptake value], SUV max , metabolic tumour volume [MTV] and first-order texture metrics) were analysed to distinguish encapsulating (eHGP) from replacement (rHGP) patterns. In pretreated patients, multivariate logistic regression including both metabolic and texture parameters identified significant associations for parameters from the fixed level threshold algorithm (VuePoint Sharp IR) with small odds ratios. Overall, PET parameters showed limited ability to discriminate between eHGP and rHGP. OS was superior with eHGP versus rHGP (log-rank p = 0.045). In chemonaïve patients, SUV-based metrics showed unadjusted exploratory associations with OS after liver surgery, varying by reconstruction/delineation and were not significant after correction for multiple testing. FDG-PET-derived parameters were not reliably associated with HGP and did not support non-invasive HGP prediction in this cohort. eHGP was associated with superior survival, underscoring the value of HGP assessment for prognostication while alternative non-invasive biomarkers remain to be identified.