Lung Adenocarcinoma Immune Subtyping Model with TIL Spatial Patterns

A study by Xia Li, last updated in May 2026, developed an interpretable prediction model for lung adenocarcinoma immune subtyping. The model quantifies spatial distribution patterns of tumor-infiltrating lymphocytes in H&E-stained whole-slide images, linking morphological phenotypes to molecular subtypes. It was validated on internal and external cohorts, achieving AUCs of 0.839 and 0.927, respectively.

Use Cases

• Predicting immune subtypes of lung adenocarcinoma based on TIL spatial features mentioned in the description.
• Training or benchmarking deep learning models for tissue and tumor segmentation in histopathology images.
• Correlating morphological image patterns with transcriptomic immune profiles for biomarker discovery.
• Developing cost-effective computational tools for tumor immune microenvironment evaluation to guide immunotherapy decisions.

Strengths

• Model achieved high performance metrics, including a 95.09% Dice score for tissue contour segmentation and an AUC of 0.927 in external validation.
• The dataset underpins a transparent and verifiable analytical workflow, as described in the study's modular design.
• Includes validation across multiple cohorts, linking image-based predictions with immunohistochemical confirmation of immune cell densities.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale model training.
• The dataset's 3.5 MB size suggests it likely contains model parameters or summary results, not the primary image data.

Provenance

Source

figshare, author Xia Li.

Collection Method

Developed from analysis of the TCGA lung adenocarcinoma cohort and external validation cohorts, using transcriptomic analysis and deep learning-based image annotation.

Time Range

null

Freshness

Last updated 2026-05-08 05:40:46; freshness should be verified.

Geography

null