Spatial subsetting enables integrative modeling of oral squamous cell carcinoma multiplex imaging data

Area of Science:

• Oncology
• Immunology
• Computational Pathology

Background:

• Oral squamous cell carcinoma (OSCC) is an aggressive cancer with poor prognosis and few reliable prognostic biomarkers.
• Understanding the tumor immune microenvironment (TIME) is crucial for developing effective treatments.

Purpose of the Study:

• To investigate the TIME in OSCC using highly multiplexed imaging mass cytometry.
• To identify spatial immune cell patterns associated with tumor differentiation and clinical outcomes.
• To develop a machine-learning framework for prognostic biomarker discovery.

Main Methods:

• Highly multiplexed imaging mass cytometry was used on OSCC biopsies.
• A spatial subsetting approach standardized immune cell populations by tissue zone.
• A machine-learning pipeline was employed for feature selection and multivariable modeling.
• Correlation with clinical outcomes was assessed in an independent cohort.

Main Results:

• Accurate histological grade classification of OSCC was achieved (AUC = 0.88).
• Specific TIME features correlated with clinical outcomes: granulocyte MAPKAPK2 signaling, CD4+ memory T cell size, and fibroblast distance from the tumor border.
• TIME patterns associated with loss-of-differentiation were identified.

Conclusions:

• A robust framework for analyzing complex imaging data in OSCC was established.
• Sentinel TIME characteristics were uncovered, offering potential prognostic biomarkers.
• These findings can aid in risk stratification and the development of immunomodulatory therapies for OSCC.