Skip to content
Review Open access

Radiomics and deep learning for immune checkpoint inhibitor outcomes: imaging-derived immune phenotyping and clinical translation across solid tumors

Sep 2026 · Frontiers in Immunology · Vol 17 · 0 citations · 67 references
Medicine

TL;DR

This narrative review critically synthesizes the evidence across solid tumors and compares handcrafted radiomics, classical machine learning, convolutional neural networks, transformer-based approaches, self-supervised learning, and emerging foundation models, and discusses explainable AI, longitudinal imaging, pseudoprogression, and hyperprogressive disease.

Abstract

Immune checkpoint inhibitors (ICIs) have transformed the treatment of multiple solid tumors, but durable benefit remains restricted to a subset of patients and conventional size-based imaging does not fully capture immune-related response patterns. Radiomics and deep learning offer a route to convert routine CT, MRI, and PET/CT into quantitative biomarkers of tumor phenotype, whole-body heterogeneity, and temporal treatment response. This narrative review critically synthesizes the evidence across solid tumors with an emphasis on four questions that are often separated in the literature: what added information is provided by different AI approaches, whether imaging signatures can be interpreted as non-invasive immune phenotypes, how methodological choices determine reproducibility, and what evidence is required for clinical translation. CT currently has the broadest outcome-prediction literature, particularly in non-small cell lung cancer; MRI provides complementary diffusion, perfusion, and soft-tissue information in hepatocellular carcinoma and brain metastases; and PET/CT adds metabolic and whole-body measures that may reflect both tumor burden and immune-metabolic state. Associations with CD8-positive T-cell infiltration, PD-L1 expression, tumor immune microenvironment states, response, progression-free survival, and overall survival support biological plausibility, but they do not establish treatment-specific predictive utility. We compare handcrafted radiomics, classical machine learning, convolutional neural networks, transformer-based approaches, self-supervised learning, and emerging foundation models, and discuss explainable AI, longitudinal imaging, pseudoprogression, and hyperprogressive disease. Major barriers remain small retrospective cohorts, high-dimensional feature selection, data leakage, scanner and protocol effects, treatment heterogeneity, selective reporting, weak calibration, and limited external or prospective validation. At present, AI-derived imaging biomarkers are best considered investigational decision-support tools. Their most credible near-term role is to complement, rather than replace, established clinical and molecular biomarkers; multimodal strategies that combine imaging with tissue, circulating, and clinical data remain promising but require demonstration of incremental and prospective clinical utility.

Read PDF

Similar papers

Review Open access Sep 2026

Deep learning-based radiomics and pathomics for decoding tumor microenvironment and predicting immunotherapy outcomes in gastric cancer

Although immune checkpoint inhibitors have reshaped the therapeutic landscape of gastric cancer, durable responses are achieved in only a minority of patients. The tumor microenvironment, particularly immune-cell composition and spatial organization, critically modulates cancer progression and the efficacy of immune ch...

Xin-Yi Shen, Ze-Yu Liu, Wen-Juan Yu et al. · 0 citations
Open access Sep 2026

Tumor-Intrinsic Blood and Imaging Correlatives in Advanced Prostate Cancer Treated with Combination Radiopharmaceutical Therapy and Immunotherapy.

The PRINCE trial showed the clinical activity for 177Lu-PSMA-617 in combination with pembrolizumab for metastatic castration-resistant prostate cancer. To refine patient selection and improve response monitoring strategies to this combination, we investigated candidate tumor-intrinsic biomarkers of treatment response a...

S. Tolmeijer, Edmond M. Kwan, Andrei Gafita et al. · 0 citations
Review Open access Oct 2026

Spatial Characteristics of the Immune Microenvironment in Brain Metastases Provide Critical Insights for Developing Predictive Markers for Anti-PD-1 Therapy

Brain metastases (BrM) exhibit variable intracranial responses to immune checkpoint blockade (ICB), and reliable biomarkers of benefit remain uncertain, possibly reflecting the distinctive brain microenvironment. This review discusses multiplex protein imaging, array-based spatial transcriptomics, and in situ RNA imagi...

Xiu-Qi Wang, Zhen-Yu Yang · 0 citations
Open access Sep 2026

Integration of MRI radiomics with immune profiling to predict response to CDK6 inhibitors in high-grade glioma

Background Predictive biomarkers are needed to identify patients with high-grade glioma who benefit from cyclin-dependent kinase 6 (CDK6) inhibition. We hypothesized that an integrated Radiomics–Immune Score (R/I-Score), combining MRI-derived tumor heterogeneity with peripheral immune profiling, would predict progressi...

Jian-Fang Wang, Peng Li, Jin-Hu Li et al. · 0 citations
Review Sep 2026

Subtype-Agnostic Biomarkers of Immunotherapy Response in Lung Cancer.

This review addresses the limitations of currently used biomarkers across lung cancer histologies and describes emerging immune-related biomarkers encompassing innate immune infiltration, antigen presentation capacity, cytotoxic T-cell activation, and inflammatory signaling that have shown promise as histology-independ...

E. Gobbini, Subhamoy Chakraborty, I. Vathiotis et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.