Pulse.

a daily field guide to health research that matters

◆ Console

‹ Fri · 26 Jun 2026
Near-term implementable finding

Deep learning-driven MRI radiomics reveals biological subtypes and predicts recurrence risk in rectal cancer.

Machine learning analysis of routine MRI scans identifies three distinct rectal cancer subtypes with different prognosis, improving prediction of who needs aggressive post-surgery treatment.

A DL MRI radiomics approach on 2060 rectal cancer patients across 4 cohorts identified three biologically distinct subtypes (immune-excluded, lymphocyte-enriched, stromal-dominant) with independent prognostic value confirmed by scRNA-seq and bulk RNA-seq. The integrated recurrence risk score improved 5-year RFS prediction vs clinical model alone (AUC 0.834 vs 0.780 external validation), providing a noninvasive tool for individualized postoperative risk stratification.

What the study was

Study design
Multicenter retrospective study with external validation (4 independent cohorts)
Population
Rectal cancer patients without neoadjuvant therapy (4 cohorts, multicenter China)
Sample size
2060
Category
Diagnostics
Maturity
Validated
Journal
NPJ Precision Oncology

Why it surfaced

Large multicenter study (n=2060) with external validation and multi-omics biological grounding; DL radiomics subtypes map to immune phenotypes relevant for treatment decisions (immunotherapy candidacy); externally validated AUC improvement over clinical staging.

A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.