Insights into pet-based radiogenomics in oncology: an updated systematic review

This study systematically reviews current evidence on radiogenomics applied to positron emission tomography (PET) imaging across oncological diseases. The primary objective is to evaluate how PET-based radiogenomics aids in understanding tumor biology, prognostic stratification, and clinical outcome prediction, while identifying methodological challenges in the field.

A systematic review was conducted following PRISMA guidelines, focusing on English-language studies indexed in Scopus, PubMed, and Web of Science until October 31, 2024. Inclusion criteria targeted original research articles involving human oncology studies using radiomics and genomics in a comprehensive "omics" framework. Data extraction included patient cohorts, radiopharmaceuticals and statistical methods. Studies were assessed for methodological rigor and reporting quality according to radiomics quality scores (RQS 2.0).

Eighteen studies involving 1780 patients were included, with 75.8% focused on lung cancer. Most studies were retrospective (72.2%) and single-center (77.7%). The primary radiopharmaceutical was [¹⁸F]FDG (88.8%). Key findings demonstrated correlations between PET-derived radiomic features and genomic alterations, such as KRAS, EGFR, and TGFβ mutations in lung cancer, and prognostic biomarkers in other malignancies. However, systemic shortcomings, including limited external validation, low reproducibility, and inadequate harmonization, were prevalent. None of the studies exceeded 50% of the RQS maximum score.

PET-based radiogenomics holds significant potential for advancing precision oncology by capturing tumor heterogeneity and improving prognostic stratification. However, methodological limitations, particularly regarding study design and data transparency, hinder its clinical applicability. Future research should prioritize multicentric designs, robust external validations, and enhanced standardization to fully realize the discipline's potential.