Effective tumor kinetics inferred from single routine H&E biopsies enable counterfactual virtual radiotherapy trials
Schlicke, P.; Ercan, C.; Ranjbar, S.; Seldomridge, A. N.; Aggarwal, S.; Mehta, A.; Sainz, T. P.; Zahid, M. U.; Fang, P.; Hofmann, V. A.; Pasetto, S.; Cisneros Napravnik, T.; Puebla-Osorio, N.; Kuttler, C.; Klopp, A.; Colbert, L.; Holder, A. M.; Weiser, R.; Lin, S. H.; Torres-Cabala, C. A.; Vega, F.; Yuan, Y.; Enderling, H.
Show abstract
Routine H&E-histopathology captures spatial snapshots of tumor ecosystems, yet computational pathology largely treats them as textures without explicit links to underlying kinetics. We show that patient-specific parameters of a mechanistic reaction-diffusion model can be inferred from a single biopsy. Cell-nucleus point patterns are fitted jointly to the analytical two-point correlation function and power spectral density, recovering patient-specific mechanistic proliferation and diffusion rates across eleven multicenter cohorts. Derived dispersion length and front velocity define mechanistic phenotypes stratifying progression-free and overall survival and adding information independent of routine covariates. We illustrate counterfactual radiotherapy simulations in glioblastoma and show that a virtual clinical trial deescalating histology-calibrated faster-growing tumors to hypofractionation while dose-intensifying slower-growing tumors adds 96.7 days of restricted mean progression-free in silico survival over standard of care (permutation p<0.001 against random allocation to the same arms). This provides a low-barrier transition from routine histopathology to mechanistically interpretable model-based forward simulations.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Initial Feasibility and Clinical Implementation of Daily MR-guided Adaptive Head and Neck Cancer Radiotherapy on a 1.5T MR-Linac System: Prospective R-IDEAL 2a/2b Systematic Clinical Evaluation of Technical Innovation 91%
- Bayesian Learning to Reduce Cardiac Risk for Locally Advanced NSCLC Patients Based on Personalized Radiotherapy Prescription 91%
- ClonoScreen3D: a novel three-dimensional clonogenic screening platform for identification of radiosensitizers for glioblastoma 90%
Similar papers in this journal
- Synergistic Effects of the DRD2/3 Antagonist ONC201 and Radiation in Glioblastoma 91%
- Deep learning NTCP model for late dysphagia after radiotherapy for head and neck cancer patients based on 3D dose, CT and segmentations 91%
- Precision association of lymphatic disease spread with radiation-associated toxicity in oropharyngeal squamous carcinomas 90%
Similar papers in this journal
- Integration of machine learning and genome-scale metabolic modeling identifies multi-omics biomarkers for radiation resistance 94%
- Priming versus propagating: distinct immune effects of an alpha- versus beta-particle emitting radiopharmaceutical when combined with immune checkpoint inhibition 92%
- Integrated radiogenomics models predict response to neoadjuvant chemotherapy in high grade serous ovarian cancer 91%
Similar papers in this journal
- Intermittent radiotherapy as alternative treatment for recurrent high grade glioma: A modelling study based on longitudinal tumor measurements 93%
- New Approaches for Quantitative Reconstruction of Radiation Dose in Human Blood Cells 92%
- Dual Adversarial Deconfounding Autoencoder for joint batch-effects removal from multi-center and multi-scanner radiomics data 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.