AI and Cancer Research ── Cancer biology ── 2026-11-12
Cancer molecular biology and AI
- Abstract
Whole-slide image (WSI) survival prediction is complicated by heterogeneous tumor morphology, where prognostic evidence is distributed across tumor, stroma, immune infiltration, necrosis, and residual tissue regions and often appears as mixed patch-level tissue evidence. Existing models commonly aggregate patch embeddings directly or convert tissue cues into hard pseudo-labels, which can discard uncertainty at tumor-microenvironment interfaces. We propose a prior-guided WSI survival framework that uses soft histopathologic priors derived from the CONCH vision-language foundation model. Prompt-similarity scores are normalized into patch-level tissue distributions and used to guide probability-weighted tissue-prototype initialization, entropy-weighted prior-consistency learning, and prototype-conditioned graph-context aggregation over a spatial k-nearest-neighbor graph. The residual Others class is retained for representation learning but excluded from prognostic prior alignment. In patient-level 5-fold cross-validation on TCGA KIRC, KIRP, LUAD, STAD, and UCEC cohorts, the model achieved C-index values of 0.7307, 0.8153, 0.6375, 0.6744, and 0.7029, respectively, matching or exceeding the strongest baseline in every cohort. Kaplan-Meier analysis showed significant high-versus low-risk separation across all cohorts, and ablation studies confirmed the contributions of soft priors, entropy weighting, the Others channel, and graph-context aggregation. These results show that retaining patch-level histopathologic uncertainty provides a consistent and interpretable signal for WSI-based survival risk stratification.
Journal IF-equivalent: 2.0 (OpenAlex 2-year mean citedness, value as of 2026-11-09, retrieved 2026-11-11; not the official Clarivate IF)Reference: Wang S, Chen Y, Zhang L, Li X, Huang Y, Li X, et al. Prior-guided whole-slide survival prediction with soft histopathologic priors from vision-language foundation models. Biomed Phys Eng Express. 2026 Nov 9 [Epub ahead of print]. doi:10.1088/2057-1976/aebfa5.Checked: Abstract only - Abstract
Cancer biomarkers should be specific, reproducible, and accessible from practical biological samples. Although omics-based approaches have substantially advanced biomarker discovery, many strategies still rely on individual molecules or limited molecular panels, which may not fully capture the complexity of tumor biology. Whether tumor-associated chemical signatures can improve cancer detection remains an open question. Here, we investigated whether a breast tumor-associated multi-ion signature could be detected in mammary gland tissue from the MMTV-PyMT breast cancer mouse model, as well as in liver, serum, and fur collected from the same animals. Low-mass ions ( m / z 50-250) were analyzed using time-of-flight secondary ion mass spectrometry (ToF-SIMS) combined with multivariate analysis and machine learning. A robust mammary tumor multi-ion signature reliably distinguished tumor-bearing from healthy mice, with both partial least squares-discriminant analysis (PLS-DA) and random forest achieving an area under the receiver operating characteristic curve (AUC) of 0.997 under leave-one-animal-out cross-validation. Animal-label permutation testing confirmed that classification performance was significantly greater than expected by chance, and 98 of 313 ions differed significantly at the animal level. While liver and serum samples showed little evidence of tumor-associated discrimination, fur showed the highest classification performance and the greatest overlap in discriminative ions with mammary tissue. These findings demonstrate that low-mass ToF-SIMS enables robust identification of mammary tumor-associated multi-ion signatures, but that their detectability varies across biological matrices. Although the evidence in fur remained exploratory, it suggests that some tumor-associated chemical information may extend beyond the primary tumor and warrants further investigation of hair and related minimally invasive samples for cancer detection.
Journal IF-equivalent: 3.7 (OpenAlex 2-year mean citedness, value as of 2026-11-09, retrieved 2026-11-11; not the official Clarivate IF)Reference: Manaprasertsak A, Tang W, Lööf C, Azemovic L, Mohlin S, Hagerling C, et al. Integrating ToF-SIMS and machine learning reveals a mammary tumor-associated multi-ion signature. Analyst. 2026 [Epub ahead of print]. doi:10.1039/d6an01090c.Checked: Abstract only