AI & Computational Pathology Nature Medicine · 2025

A Multimodal Whole-Slide Foundation Model for Pathology (TITAN)

Pretrained on 335,645 whole-slide images with vision-language alignment, TITAN generates slide-level pathology reports and performs rare cancer retrieval — without fine-tuning — outperforming all prior slide-level foundation models.

Helmholtz Munich / international consortium · Nature Medicine, 2025

TITAN was pretrained on 335,645 whole-slide images using a combination of visual self-supervised learning and vision-language alignment — the latter powered by 423,122 synthetic captions generated from paired pathology reports. The dual training objective gives the model both strong morphological representations and the ability to reason about histological content in natural language.

In zero-shot evaluation, TITAN generates coherent slide-level pathology reports and retrieves visually similar cases for rare cancer types — tasks that prior region-of-interest models could not perform at the slide level. The model outperformed previous slide-level foundation models across classification, survival prediction, and retrieval benchmarks.

The work represents a key step toward slide-level pathology AI that pathologists can interrogate in natural language, rather than only receiving a binary classification output.

Read in Nature Medicine →
AI & Computational  Nature Medicine

UNI: Towards a General-Purpose Foundation Model for Computational Pathology

Trained on 100M+ tissue patches from 100,000+ WSIs via self-supervised learning, UNI outperforms all prior computational pathology models across 34 downstream tasks — cancer classification, subtyping, and transplant assessment.

Mahmood Lab, Harvard / Mass General Brigham ·Nature Medicine, 2024
Read paper →
AI & Computational  Nature Medicine

CONCH: A Vision-Language Foundation Model for Computational Pathology

Contrastive vision-language model pretrained on 1.17 million histopathology image-caption pairs — the largest such dataset at publication — enabling zero-shot pathology report retrieval and cross-modal search beyond vision-only models.

Mahmood Lab, Brigham & Women's Hospital ·Nature Medicine, 2024
Read paper →
AI & Computational

AI Infers Oncotype DX Recurrence Score Directly From H&E — Eliminating Molecular Testing

Review of AI tools in breast pathology highlights Orpheus, a multimodal deep learning model that predicts Oncotype DX Recurrence Score from digitised H&E slides, with implications for cost and access in resource-limited settings.

Surgical Pathology Clinics, 2025
Read paper →
Molecular & Liquid Biopsy

ctDNA Achieves 71% Sensitivity and 98.6% Specificity for Early Esophageal Cancer Detection

Pooled analysis validates ctDNA-based detection viability in a GI malignancy with notoriously difficult early diagnosis. Combining ctDNA with methylation markers further improves sensitivity over single-analyte approaches.

Frontiers in Oncology, 2025
Read paper →
Molecular & Liquid Biopsy

ctDNA and Liquid Biopsy in Cancer Diagnosis, Screening, and Treatment Monitoring — 2025 Review

Comprehensive review of ctDNA detection technologies (NGS, ddPCR), positioning ctDNA as the most analytically reliable liquid biopsy analyte and charting the trajectory toward multi-cancer early detection (MCED) assays.

Parums DV ·Medical Science Monitor, 2025 · 31:e949300
Read paper →
Molecular & Liquid Biopsy  ASCO 2025

Multi-Analyte Liquid Biopsy Frameworks for Cancer Characterisation and Therapy Monitoring

ASCO Educational Book covers multi-analyte frameworks integrating ctDNA with CTCs and extracellular vesicles, with clinical evidence for ctDNA-guided monitoring in NSCLC, breast, and colorectal cancer.

ASCO Educational Book, 2025
Read paper →
IHC & Biomarkers

TMB as Companion Diagnostic for PD-(L)1 Blockade: ORR 29% in TMB-High vs 6% TMB-Low Tumours

Rigorous evaluation of TMB across solid tumour types confirms FDA approval rationale for pembrolizumab TMB-H indication. Outlines assay standardisation challenges across NGS platforms as the critical barrier to broader deployment.

MD Anderson Cancer Center ·Annals of Oncology, 2024
Read paper →
IHC & Biomarkers

Recommendations for Tumour Mutational Burden Assay Harmonisation Across Laboratories

Consensus paper addressing pre-analytical, analytical, and post-analytical standardisation for TMB to function reliably as a companion diagnostic, emphasising comprehensive methodological reporting for cross-assay comparability.

Journal of Molecular Diagnostics, August 2024
Read paper →
IHC & Biomarkers

Immunohistochemistry of Lung Cancer Biomarkers: PD-L1, ALK, ROS1, MET as Surrogates for Targeted Therapy

Reviews current IHC biomarkers in lung cancer for targeted therapy selection, positioning IHC as an essential surrogate molecular method where NGS is unavailable — particularly relevant to resource-limited settings.

Advances in Anatomic Pathology, September 2024
Read paper →
Haematopathology  Modern Pathology

WHO-HEM5: Fifth Edition Classification of Myeloid Neoplasms — Definitive Reference Paper

Implements WHO-HEM5 myeloid classification, expanding categories within MPN, MDS, AML, MDS/MPN, and M/LN-E-TK based on updated genomic knowledge. New molecular-genetic entities and refined NGS-driven diagnostic criteria.

WHO Classification Working Group ·Modern Pathology, Vol. 37, Issue 2, 2024
Read paper →
Haematopathology

Practical Laboratory Guide to Genetic Studies in Clonal Haematopoiesis Under WHO-HEM5

Implementation guide for applying WHO-HEM5 genomic criteria to MDS and AML diagnosis, covering when to order cytogenetics, FISH, and NGS panels, with special attention to differentiating CHIP from early myeloid neoplasia.

Blood Cancer Journal, 2024
Read paper →
Haematopathology

WHO-HEM5 vs ICC: Navigating Dual Classification Systems for Myeloid Neoplasms in Practice

2025 comparative review of WHO-HEM5 and the International Consensus Classification for myeloid neoplasms, highlighting divergences and providing practical guidance for pathologists operating in dual-classification reporting environments.

Journal of Pathology and Translational Medicine, 2025
Read paper →
Surgical & Breast

AI-Assisted Sentinel Lymph Node Detection Cuts IHC Use and Saves ~€40,000 Per Year in Real-World Deployment

Real-world clinical implementation of AI-assisted SLN metastasis detection demonstrates reduced IHC use per detected case and approximately €40,000 annual cost savings — with a pathway to treatment personalisation via digital pathology.

ESMO Real World Data & Digital Oncology, 2025
Read paper →
Surgical & Breast

Digital Pathology and AI in Breast Pathology: Subtyping, Grading, and Biomarker Quantification

Reviews AI tools for automated breast cancer subtype classification, tumour grading, and quantification of ER, PR, HER2, Ki-67, and PD-L1 from digitised H&E and IHC slides — covering clinical validation status of each modality.

Surgical Pathology Clinics, 2025
Read paper →
India & South Asia

Distinct Genomic and Pathological Differences in Breast Cancer Among South Asian Women Compared to European Cohorts

One of the largest molecular characterisations of breast cancer in South Asian and African ancestry women, with direct implications for biomarker testing strategies and therapeutic trial design in underrepresented populations.

medRxiv preprint, 2024 · under review
Read paper →
India & South Asia  JCO Global Oncology

Comprehensive Genomic Profiling of 3,000+ Indian Patients With Lung Cancer — India's Largest Dataset

Molecular analysis of EGFR, ALK, ROS1, and actionable alterations in over 3,000 Indian NSCLC patients reveals distinct mutation frequencies versus Western and East Asian cohorts, with direct implications for companion diagnostic use in Indian clinical practice.

JCO Global Oncology, 2024–25
Read paper →
India & South Asia

WHO Molecular Classification of Endometrial Cancer Is Feasible in Indian Tertiary-Care Settings — Prospective Cohort

South Indian prospective cohort (2016–2024) shows WHO molecular subtypes (POLE, MMR-deficient, p53-abnormal, NSMP) via IHC and molecular testing correlate with survival outcomes consistent with global data — validating feasibility in Indian labs.

Kuriakose et al., Kerala, India ·International Journal of Gynaecology & Obstetrics, 2025
Read paper →
No papers match your search.