AI and Cancer Research ── Breast ── 2026-11-16
Breast cancer: treatment, trials and AI
- Abstract
Unsupervised hierarchical clustering of early-stage breast cancer glycogene expression data using the Growing Hierarchical Self Organising Representation Map (GHSORM) algorithm identified seven subclasses. These glycogene expression clusters (GECs) overlap with known breast cancer biomarkers such as hormone receptor status and immune signatures. Feature discovery via Recursive Feature Elimination with Strikeout (RFES) identified a panel of 60 glycogenes that can predict GECs on an independent testing set with 90.48% accuracy. GEC-based subtyping of patients from the I-SPY2 clinical trial revealed an 86% pathologic complete response (pCR), demonstrating how glycogene expression in combination with custom machine learning algorithms could be an approach to molecular cancer diagnosis capable of informing personalised drug treatment.
Journal IF-equivalent: 4.0 (OpenAlex 2-year mean citedness, value as of 2026-11-15, retrieved 2026-11-16; not the official Clarivate IF)Reference: Venter M, Naidoo KJ. Unsupervised machine learning identifies early stage breast cancer subclasses from aberrant glycogene expression and informs personalised drug treatment. Discov Chem. 2026 Nov 14;3(1):630. doi:10.1007/s44371-026-01070-0.Checked: Abstract only - Abstract
To evaluate whether whole-body CT-derived body composition parameters from routine 18 F-FDG PET/CT provide additional prognostic value beyond tumor metabolic activity in patients with stage I-III HER2-negative breast cancer, irrespective of hormone receptor status. In this retrospective study, 178 patients with stage I-III HER2-negative breast cancer who underwent preoperative 18 F-FDG PET/CT were included. Hormone receptor (HR) status was recorded for all patients but was not an eligibility criterion: patients with HER2-negative disease were eligible irrespective of estrogen receptor (ER) and progesterone receptor (PR) status. Whole-body CT images were analyzed using a deep learning-based TotalSegmentator framework to automatically quantify skeletal muscle (SM), visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), and intermuscular adipose tissue (IMAT). Associations between imaging-derived parameters and disease-free survival (DFS) were assessed using Cox proportional hazards models and Kaplan-Meier analysis. The incremental prognostic value of body composition parameters beyond SUVmax was evaluated using Harrell’s C-index, likelihood ratio tests, net reclassification improvement (NRI), and integrated discrimination improvement (IDI). During a median follow-up of 41.0 months, 45 patients (25.3%) experienced DFS events. Compared with the non-DFS event group, patients with DFS events showed significantly higher SUVmax, and increased volumes of IMAT, VAT, and SAT, whereas SM volume and attenuation were significantly lower (all P < 0.05). Multivariable Cox regression analysis demonstrated that higher SUVmax, volumes of IMAT and VAT [HR (95% CI): 4.10 (1.97 ~ 8.55), 3.27 (1.17 ~ 9.15), 3.19 (1.45 ~ 7.00), respectively, p all < 0.05] were associated with increased risk of DFS events, whereas higher SM volume was associated with improved DFS [HR (95% CI): 0.53 (0.29 ~ 0.98), P = 0.045]. The association with IMAT volume was not reproduced when the variable was kept continuous (HR per SD 1.32, 95% CI 0.88 ~ 1.98, P = 0.18). Adding body composition parameters (VAT, IMAT, and SM volumes) to SUVmax improved discrimination, increasing the C-index from 0.71 (95% CI, 0.64 ~ 0.78) to 0.82 (95% CI, 0.77 ~ 0.87; P < 0.001), with improvement in risk reclassification (categorical NRI 0.34, 95% CI 0.07 ~ 0.59, P = 0.008; IDI 0.12, 95% CI 0.04 ~ 0.24, P = 0.004). Among all body composition parameters, VAT volume provided the greatest incremental prognostic value. Furthermore, combined stratification based on SUVmax and VAT volume identified four distinct risk groups, with patients exhibiting both high SUVmax and high VAT volume showing the poorest DFS. In HER2-negative breast cancer patients undergoing preoperative 18 F-FDG PET/CT, higher SUVmax and larger VAT volume were associated with poorer DFS, and larger SM volume with better DFS. Adding body composition parameters to SUVmax improved discrimination modestly, with VAT volume showing the strongest contribution. Patients with both high SUVmax and high VAT volume exhibited the poorest DFS. Further prospective validation in larger, multicentre cohorts is warranted.
Journal IF-equivalent: 4.8 (OpenAlex 2-year mean citedness, value as of 2026-11-12, retrieved 2026-11-13; not the official Clarivate IF)Reference: Yuan G, Sun Z, Li Y, Zhou Y. Deep learning-based body composition analysis from 18F-FDG PET/CT provides incremental prognostic value beyond tumor metabolism in HER2-negative breast cancer. BMC Cancer. 2026 Nov 14 [Epub ahead of print]. doi:10.1186/s12885-026-17368-2.Checked: Abstract only - Abstract
Background/Objectives: Whether Shapley-based explanations of machine learning survival models are transported to new populations is rarely tested. We assessed the geographic and temporal transportability of the performance and time-dependent SurvSHAP(t) explanations of survival models in breast cancer.
Methods: Cox, random survival forest (RSF), gradient-boosted survival (GBSA), DeepSurv and DeepHit models for recurrence-free survival were developed in Rotterdam (n = 2982) and applied without updating to a geographic (German Breast Cancer Study Group [GBSG], n = 686) and a temporal cohort (Duke, 2000–2014, n = 888). Explanation shift was assessed with classifier two-sample tests and density-ratio decomposition; agreement with local reference models refitted in each external cohort was tested with a 500-refit bootstrap null and Holm correction.
Results: Discrimination was preserved (Uno’s C at 0.64–0.68 in GBSG; 0.71–0.73 in Duke), but the three-year risk was overestimated about three-fold in Duke (observed/expected at 0.28–0.35); a true decline in risk could not be separated from differences in event ascertainment. Explanations shifted in both cohorts (classifier area under the curve [AUC] at 0.75–0.90), yet nodal category ranked first in all 15 model–cohort pairs. Disagreement with the local reference model was detected only for RSF in GBSG (Holm-adjusted p = 0.032); elsewhere, and particularly in Duke, non-detection does not establish agreement because power was limited. Local models gave more weight to endocrine therapy in Duke (+0.07 to +0.22; +0.04 to +0.24 across imputations).
Conclusions: Preserved discrimination did not prevent marked miscalibration in a later cohort, and preserved importance rankings did not guarantee that transported explanations matched those of locally trained models. Explanations should be evaluated externally alongside, and separately from, predictive performance.
Journal IF-equivalent: 4.0 (OpenAlex 2-year mean citedness, value as of 2026-10-29, retrieved 2026-10-30; not the official Clarivate IF)Reference: Akaydın Gültürk E. Transportability of Time-Dependent Shapley Explanations of Machine Learning Survival Models in Breast Cancer: A Geographic and Temporal External Validation Study. J Clin Med. 2026 Nov 13;15(22):8900. doi:10.3390/jcm15228900.Checked: Abstract only - Abstract
BACKGROUND: Breast cancer is the most prevalent cancer in women, and postoperative disease-free survival (DFS) is an important indicator of long-term prognosis. However, tumor, node, metastasis staging has limited ability to integrate multidimensional clinical, pathological, and treatment-related factors for individualized DFS prediction. Therefore, machine learning (ML) may provide a more comprehensive approach to prognostic risk assessment. OBJECTIVE: This study aimed to develop and externally validate a ML-based model for predicting postoperative DFS in patients with breast cancer. METHODS: This retrospective study included 2 cohorts: 1231 patients treated at an institutional center between 2012 and 2020 and 304 patients from The Cancer Genome Atlas (TCGA). The institutional cohort was divided into model development (n=861) and internal validation (n=370) sets, while the TCGA cohort served as an external validation cohort. Five ML algorithms were evaluated, with the Cox proportional hazards model additionally fitted as a conventional statistical reference. Model performance was assessed using the concordance index and time-dependent area under the curve (AUC). The net reclassification index and integrated discrimination improvement were used for comparative evaluation. Shapley additive explanations (SHAP) analysis was performed to assess predictor contributions, and patients were stratified into risk groups according to the median predicted risk score. RESULTS: Based on a comprehensive evaluation of model performance, extreme gradient boosting (XGBoost) with an accelerated failure time (AFT) was selected as the final model. The study included 1231 institutional patients, with 861 used for model development and 370 for internal validation, and 304 patients from TCGA for external validation. The 5-year time-dependent AUCs were 0.840, 0.822, and 0.762 in the training, internal validation, and external validation cohorts, respectively. In the external validation cohort, the XGBoost (AFT) model achieved a concordance index of 0.750 (95% CI 0.654-0.843), with 3-, 5-, and 8-year AUCs of 0.791, 0.762, and 0.768, respectively. At 5 years, compared with Cox regression, XGBoost (AFT) showed improved reclassification (net reclassification index=0.311, 95% CI 0.172-0.458; P =.01; integrated discrimination improvement=0.166, 95% CI 0.095-0.235; P <.001). SHAP analysis identified endocrine therapy, pT, pN, lymph node ratio, Ki-67, and chemotherapy among the major contributors to model predictions. Risk stratification based on the median predicted risk score separated patients into groups with distinct DFS outcomes (hazard ratio 5.44, 95% CI 4.04-7.31). CONCLUSIONS: The ML-based model demonstrated favorable performance and external generalizability in predicting postoperative DFS. SHAP analysis provided interpretability for the model, and risk stratification identified patient subgroups with distinct prognostic risks.
Journal IF-equivalent: 6.7 (OpenAlex 2-year mean citedness, value as of 2026-11-04, retrieved 2026-11-05; not the official Clarivate IF)Reference: Chen Z, Wu Y, Tong L, Wu T, Zhao N, Zhu L, et al. Machine Learning–Based Prediction of Disease-Free Survival in Breast Cancer Incorporating Clinical and Treatment Information: Retrospective Cohort Study With External Validation. JMIR Med Inform. 2026 Nov 13;14:e101633-e101633. doi:10.2196/101633. PMID: 43024943.Checked: Abstract only