Prognostic value of clinical and radiomic parameters in patients with liver metastases from uveal melanoma

Mael Lever ¹, Simon Bogner ², Melina Giousmas ²

⁰Department of Ophthalmology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Essen, Germany.

Pigment Cell & Melanoma Research +

|

July 12, 2024

View abstract on PubMed

Summary

Identifying prognostic biomarkers for uveal melanoma with liver metastases is crucial. Machine learning models show clinical, hepatic tumor burden, and radiomic parameters independently predict patient outcomes.

Area Of Science

Oncology
Medical Imaging
Machine Learning

Background

Uveal melanoma frequently metastasizes to the liver, with limited survival after diagnosis.
Identifying prognostic biomarkers is essential for risk stratification and personalized treatment.

Purpose Of The Study

To develop prognostic models for uveal melanoma patients with hepatic metastases.
To evaluate the independent and combined prognostic value of clinical, quantitative hepatic tumor burden, and radiomic parameters.

Main Methods

Retrospective analysis of 101 uveal melanoma patients with newly diagnosed hepatic metastases.
Application of Cox-Lasso regression machine learning for risk stratification.
Evaluation of clinical, laboratory, quantitative hepatic tumor burden, and radiomic parameters.

Main Results

Substantial binary risk stratification was achieved using clinical, hepatic tumor burden, or radiomic parameters individually.
Combining parameter domains did not improve prognostic separation.
Key clinical predictors for time-to-treatment failure and overall survival were identified.

Conclusions

Clinical, radiological, and radiomic parameters possess comparable and independent prognostic value in uveal melanoma with hepatic metastases.
Machine learning models can effectively stratify risk in this patient population.
Further research into these parameters may enhance patient counseling and therapeutic strategies.

Keywords:

biomarker hepatic metastases machine learning prognosis radiomics uveal melanoma