Platform-agnostic quantification for molecular interception of oral and esophageal malignancies: a global study addressing the diagnostic divide
Abstract
Background: Oral squamous cell carcinoma (OSCC) and esophageal squamous cell carcinoma are commonly diagnosed late because histopathology is subjective and robust molecular tools are lacking. Using interpretable machine learning, we aimed to develop an objective and platform-agnostic molecular measure of malignancy that could overcome the technical and geographic barriers associated with conventional genomic testing.
Methods: Using a discovery-to-generalization framework with 2541 samples, we developed and validated a single-sample transcriptomic scoring system called EsoralPrint. We feature-engineered 34 intrasample ordinal gene pairs (isoPairs) that capture consistent expression-order reversals between malignant and healthy states. The system was trained on oral tissues and externally validated across diverse global cohorts spanning the upper aerodigestive tract, technical platforms, and geographic regions. Independent single-cell RNA sequencing was used to provide post hoc mechanistic explainability.
Results: In the discovery cohort, EsoralPrint distinguished patients histologically graded as having low-risk hyperplasia who nonetheless progressed to cancer (log rank P < .001); we treated this result as exploratory pending independent confirmation. In external validations, EsoralPrint reached 95.0% accuracy and a 96.7% F1 score across OSCC cohorts (460 OSCC, 162 non-OSCC) and 91.6% accuracy and a 91.6% F1 score in esophageal tissue unseen during discovery and training (574 esophageal squamous cell carcinoma, 614 healthy mucosa samples). Further, a malignancy gradient reflecting field cancerization effects in resection boundaries was observed, and a transition zone (scores 0.25-0.75) was defined where molecular alterations preceded morphological changes.
Conclusion: EsoralPrint provides a platform-independent and quantitative complement to histological grading in oral and esophageal squamous malignancy, with the potential to give clinicians lead time for intervention. The tool can be accessed at https://github.com/aiphaqua/EsoralPrint.
Related
- Abdominal Pain as the First Presentation of Severe Heart Failure With Reduced Ejection Fraction and Left Ventricular Thrombus in a Young Asymptomatic Patient: A Diagnostic Interception
- High-Resolution Esophageal Manometry in Mongolia: The First National Experience and Diagnostic Spectrum of Esophageal Motility Disorders
- RUNX1 Familial Platelet Disorder with Associated Myeloid Malignancies Initially Diagnosed as Immune Thrombocytopenia and Reclassified After 10 Years: Diagnostic Pitfall Revealed by a Family History
- Addressing unmet clinical needs in treatment-resistant breast cancer: expanding the diagnostic–prognostic landscape through next-generation inflammatory signatures
- A Key Biomarker lnc-ABCA12-5:2 Compensating Diagnostic Deficiencies for Indeterminate Thyroid Nodules and Genetically Wild-Type Thyroid Malignancies
BibTeX
@article{lai2026platformagnostic,
title = {{Platform-agnostic quantification for molecular interception of oral and esophageal malignancies: a global study addressing the diagnostic divide}},
author = {Hung-Ming Lai and Hua-Ping Liu and Kaia Douglas and Keith Johnson and Xuanhua P Xie},
journal = {JNCI Cancer Spectrum},
year = {2026},
month = oct,
doi = {10.1093/jncics/pkag102},
url = {https://doi.org/10.1093/jncics/pkag102},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - Platform-agnostic quantification for molecular interception of oral and esophageal malignancies: a global study addressing the diagnostic divide
AU - Hung-Ming Lai
AU - Hua-Ping Liu
AU - Kaia Douglas
AU - Keith Johnson
AU - Xuanhua P Xie
PY - 2026
DA - 2026/10/05
JO - JNCI Cancer Spectrum
DO - 10.1093/jncics/pkag102
UR - https://doi.org/10.1093/jncics/pkag102
AB - Background: Oral squamous cell carcinoma (OSCC) and esophageal squamous cell carcinoma are commonly diagnosed late because histopathology is subjective and robust molecular tools are lacking. Using interpretable machine learning, we aimed to develop an objective and platform-agnostic molecular measure of malignancy that could overcome the technical and geographic barriers associated with conventional genomic testing. Methods: Using a discovery-to-generalization framework with 2541 samples, we developed and validated a single-sample transcriptomic scoring system called EsoralPrint. We feature-engineered 34 intrasample ordinal gene pairs (isoPairs) that capture consistent expression-order reversals between malignant and healthy states. The system was trained on oral tissues and externally validated across diverse global cohorts spanning the upper aerodigestive tract, technical platforms, and geographic regions. Independent single-cell RNA sequencing was used to provide post hoc mechanistic explainability. Results: In the discovery cohort, EsoralPrint distinguished patients histologically graded as having low-risk hyperplasia who nonetheless progressed to cancer (log rank P < .001); we treated this result as exploratory pending independent confirmation. In external validations, EsoralPrint reached 95.0% accuracy and a 96.7% F1 score across OSCC cohorts (460 OSCC, 162 non-OSCC) and 91.6% accuracy and a 91.6% F1 score in esophageal tissue unseen during discovery and training (574 esophageal squamous cell carcinoma, 614 healthy mucosa samples). Further, a malignancy gradient reflecting field cancerization effects in resection boundaries was observed, and a transition zone (scores 0.25-0.75) was defined where molecular alterations preceded morphological changes. Conclusion: EsoralPrint provides a platform-independent and quantitative complement to histological grading in oral and esophageal squamous malignancy, with the potential to give clinicians lead time for intervention. The tool can be accessed at https://github.com/aiphaqua/EsoralPrint.
SN - 2515-5091
ER -