The Karolinska Liver Metastases (KaroLiver) cohort is a retrospective, population-based observational dataset assembled to study associations between clinical characteristics and histopathological features in patients treated with curative intent for colorectal liver metastases (CRLM). The core objective is to combine whole-slide digital histopathology images with comprehensive clinical, oncological, surgical, radiological and survival data to enable integrated analyses of treatment trajectories, metastatic tumour biology and clinical outcomes.
KaroLiver captures consecutive patients treated with liver-directed, curative-intent approaches at a single tertiary referral centre, providing a platform for histology-centered prognostic modelling and multimodal imaging and pathology research.
The cohort comprises all consecutive patients who received curative-intent liver-directed treatment for CRLM at Karolinska University Hospital in Stockholm, Sweden. Karolinska University Hospital functions as the primary regional referral centre for hepatopancreatobiliary (HPB) surgery for the Stockholm–Gotland healthcare region, serving an approximate population of 2.5 million people.
Patient enrolment in KaroLiver is retrospective and is updated continuously in line with amended ethical approvals and evolving scientific questions. The current accrual period spans procedures performed between February 2012 and January 2022. As of the report, the cohort includes 811 patients who underwent a total of 1,204 liver interventions.
KaroLiver links multiple data domains for each included patient. The assembled datasets include:
Whole-slide digital histopathology images of metastatic liver tissue.
Detailed clinical and oncological data covering presentation, systemic therapy, and perioperative management.
Surgical data describing liver-directed interventions and operative outcomes.
Radiological information from CT and MRI imaging modalities.
Pathological and molecular data derived from resected metastases.
Recurrence information and longitudinal survival outcomes.
Collectively, these linked data enable analyses spanning tumour histopathology, radiomics, treatment pathways and long-term outcomes. The investigators report that retrospective accrual and outcome updates will continue for current and future studies, subject to approvals.
Descriptive survival outcomes reported from the assembled cohort include a median overall survival (OS) of 51.0 months (95% CI 46.2–57.1 months) and a median recurrence-free survival (RFS) of 10.4 months (95% CI 9.2–11.9 months). The five-year OS rate in the cohort is 44.9% (95% CI 41.2–48.8%).
Analyses already enabled by KaroLiver have identified a liver injury–derived stromal capsule in a subset of metastases that was associated with improved survival. The cohort has also supported identification of histopathological markers of tumour biology, documented sex-based differences in treatment and survival, and reported associations between post‑hepatectomy liver failure and oncological outcomes.
No additional numerical outcomes, subgroup analyses or detailed molecular findings beyond those summarized above were reported in the source text.
Ongoing and planned research themes outlined by the KaroLiver investigators include:
Development of advanced histology-based prognostic scoring systems.
Examination of sex differences in recurrence patterns and retreatment strategies.
Comparative analyses of tumour biology and outcomes in early-onset versus average-onset CRLM.
Integration of CT- and MRI-based radiomics analyses with the cohort’s histopathological framework.
The cohort supports data sharing for approved research collaborations. Data access requires meeting regulatory and ethical requirements, execution of appropriate data access and research collaboration agreements, and, where applicable, regulatory clearance. De-identified datasets will be made available to qualified researchers for approved analyses on request.
The investigators report that the Swedish Ethical Review Authority granted ethical approval for the work, listing multiple decision numbers (Dnr 2019-01571 and amendments including Dnr 2020-00428, Dnr 2021-06863-02, Dnr 2022-06588-02, Dnr 2025-05423-02, Dnr 2026-03619-01). The authors state that all relevant ethical guidelines were followed, informed consent was obtained where required, and patient identifiers included in research outputs were not accessible to parties outside the research group.
Competing interests: the authors declared no competing interests.
Funding sources reported include national and institutional research funders; specific funder names and grant identifiers were provided in the source document.
Data availability: all data produced in the present study are available upon reasonable request to the authors. Data sharing may be arranged for approved research collaborations subject to mutual ethical approval, signature of data access agreements and any required regulatory clearance.
Strengths and caveats: the cohort’s population-based, single-centre design with comprehensive linkage of whole-slide histopathology to multimodal clinical data is a major strength for translational pathology and outcome research. The source document specifies survival outcomes and several initial histopathology-derived findings; however, additional granular results, subgroup details and full analytic methods were not reported in the profile and would be available only in subsequent analyses or on request.