AI and Cancer Research ── Colorectal ── 2026-11-09
Colorectal cancer: screening, treatment, trials and AI
- Abstract
Tumor-associated macrophages (TAMs) are key regulators of the tumor microenvironment in colorectal cancer (CRC), yet their functional heterogeneity remains poorly understood, especially regarding the interplay between cellular senescence and lipid metabolism. We integrated multi-omics and single-cell transcriptomic data to characterize a novel macrophage subpopulation termed senescence-associated lipid-metabolic (SALM) macrophages. Using eight machine learning algorithms, a prognostic model based on SALM signatures was developed and internally evaluated via nested cross-validation in a combined TCGA–GSE39582 cohort. Spatial multiplex immunofluorescence and qPCR were employed for experimental validation. SALM macrophages exhibit elevated IL-10/STAT3 signaling, terminal differentiation, and a core gene module enriched in senescence and lipid metabolism pathways. They interact strongly with T cells via immunosuppressive ligands IL-10 and MIF. Spatial analysis confirmed p16 + CD68 + IL-10 + cell enrichment at the tumor front. The SALM-based prognostic model achieved high accuracy (AUC = 0.972), and a corresponding nomogram showed potential for personalized prognosis prediction. SALM macrophages drive immunosuppression in CRC through a senescence-lipid metabolism axis, serving as a novel prognostic biomarker and potential therapeutic target.
Journal IF-equivalent: 5.9 (OpenAlex 2-year mean citedness, value as of 2026-11-04, retrieved 2026-11-05; not the official Clarivate IF)Reference: Zhao B, Liao Y, Ye S, Jiang N, Lu W, Chen Y, et al. Multi-omics and machine learning reveal a senescence-associated lipid-metabolic macrophage phenotype driving immunosuppression and poor prognosis in colorectal cancer. Sci Rep. 2026 Nov 7 [Epub ahead of print]. doi:10.1038/s41598-026-80104-0.Checked: Abstract only - Abstract
Colorectal cancer (CRC) remains a leading cause of cancer-related mortality in the UK. Early detection and removal of adenomatous polyps are essential to improving outcomes. Artificial intelligence-assisted colonoscopy tools, such as GI Genius (GIG), may enhance adenoma detection rates (ADR) and reduce the likelihood of missed lesions. A patient-level simulation cost-effectiveness model was developed to estimate the economic and clinical impact of GIG-assisted colonoscopy compared with standard colonoscopy alone in a symptomatic for suspected CRC population in England. The model adopted a National Health Service and Personal Social Services perspective, with a lifetime time horizon. Key model inputs were informed by the NAIAD trial, which evaluated the impact of GIG-assisted colonoscopy on ADR, compared with standard colonoscopy. Base case and probabilistic sensitivity analysis (PSA) were conducted to estimate the incremental cost-effectiveness ratio and the probability of cost effectiveness. Base case results suggest that GIG is a cost-effective intervention, associated with per-person cost savings of £69 and quality-adjusted life year (QALY) gains of 0.046. This was primarily due to increased ADR and subsequent reduced CRC incidence. Scenario analyses testing uncertainty around the sensitivity of colonoscopy, clinical outcomes for those initially undiagnosed, differences in outcomes across experienced and non-experienced endoscopists, and the total costs of GIG produced findings that were consistent with the base case and did not materially change the estimated costs or QALYs. PSA results indicate that GIG is cost effective in 53% of simulations at the £20,000 per QALY gained threshold. This uncertainty is attributable to the marginal change in QALY outcomes between arms of the model, meaning that minor fluctuations in QALY outcomes between PSA iterations lead to large variations in the ICER. The results highlight that GIG is a cost-saving intervention that can improve health outcomes by increasing ADRs. Scenario analyses broadly support the base case findings, suggesting that the conclusions were generally consistent across the alternative assumptions explored. PSA results highlighted considerable decision uncertainty, driven by the small incremental differences in QALYs. PSA results highlight considerable decision uncertainty, driven by the small incremental differences in QALYs.
Journal IF-equivalent: 4.9 (OpenAlex 2-year mean citedness, value as of 2026-11-07, retrieved 2026-11-08; not the official Clarivate IF)Reference: Curtis-Gretton L, Varghese AA, Watts K, Green M, Holmes H. The economic and clinical impact of an AI-assisted tool for colonoscopy to support diagnosis of colorectal cancer in symptomatic patients in England. BMC Digit Health. 2026 Nov 6;4(1):79. doi:10.1186/s44247-026-00314-4.Checked: Abstract only - Abstract
BACKGROUND: Prognostic models for colon adenocarcinoma (COAD) frequently rely on predefined biologic gene sets or a single analytic framework, and few studies simultaneously address algorithm selection, biologic interpretation, comparison with conventional staging, and benchmarking against published COAD signatures. This study aimed to develop and biologically characterize a random survival forest (RSF) prognostic model for COAD. METHODS: RNA-sequencing (RNA-seq) and clinical data from 455 The Cancer Genome Atlas (TCGA)-COAD patients were analyzed in a training cohort (n=324) and an internal test cohort (n=131). A 12-gene signature was derived through differential-expression screening, log-rank testing, and univariate Cox analysis. Eight survival algorithms were compared, and the best-performing RSF model generated the final risk score. Additional analyses included Gene Ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, variable-importance and SHapley Additive exPlanations (SHAP) interpretation, Kaplan-Meier and subgroup interaction analyses, Cox regression, comparison of RSF vs. tumor-node-metastasis (TNM) vs. RSF + TNM using time-dependent receiver operating characteristic (ROC), concordance index (C-index), net reclassification improvement (NRI), and integrated discrimination improvement (IDI), Immuno-Oncology Biological Research (IOBR)-based immune-signature profiling, and benchmarking against published COAD signatures. RESULTS: All 12 model genes mapped successfully for enrichment analysis, yielding 115 significant GO biological-process terms, 33 cellular-component terms, 25 molecular-function terms, and 4 KEGG pathways centered on telomere regulation, chromosome organization, and transcription-related programs. RSF showed near-perfect internal discrimination, with 1-, 3-, and 5-year areas under the ROC curves (AUCs) of 0.999, 0.999, and 0.998 in the training set, but external performance was more modest in the test set (0.631, 0.608, and 0.552). PPFIA4 , GTF2H4 , H4C5 , ARL6IP4 , LRP2 , and TERT were consistently prioritized by variable importance (VIMP) and SHAP. Only pathologic M in training and lymphatic invasion in testing showed significant interaction with RSF score. In the test cohort, TNM and RSF + TNM outperformed RSF alone, and adding TNM to RSF improved NRI at 1 and 3 years. IOBR analyses identified 72 significant correlations in training and 57 in testing, dominated by exosome-, glycan-, metabolism-, and T-cell-related signatures. CONCLUSIONS: The 12-gene RSF score captures biologically meaningful transcriptomic and immune-related signals in COAD, but its most defensible translational role is as a complement to TNM rather than a replacement. Rigorous external RNA-seq validation remains necessary.
Journal IF-equivalent: 1.4 (OpenAlex 2-year mean citedness, value as of 2026-10-13, retrieved 2026-10-15; not the official Clarivate IF)Reference: Zhang J, Tan J, Lu M, Luo D, Yang M, Gao F, et al. Development and biological characterization of a random survival forest prognostic model for colon adenocarcinoma using eight machine learning algorithms. Transl Cancer Res. 2026 Oct;15(10):765-5. doi:10.21037/tcr-2026-0877.Checked: Abstract only