Machine Learning Model Guided by TLS Predicts Survival and Immune Features in Gastric Cancer

NCT06979817 — Completed, 1200 participants, Qun Zhao

ClinicalTrials Daily · Areas · Conditions · Drugs · Sponsors ·

TLS-Informed Machine Learning Model Predicts Survival and Immune Landscape in Locally Advanced Gastric Cancer

Registry	NCT06979817
Status	Completed
Study type	Observational
Enrollment	1200 (Actual)
Sponsor	Qun Zhao (Other)
Conditions	Locally Advanced Gastric Cancer, Tumor Immune Microenvironment, Tertiary Lymphoid Structures (TLS)
Interventions	TLS-Informed Machine Learning Prognostic Model (Other)
Dates	start 2012-01-01, primary completion 2024-01-01, completion 2024-01-01, first posted 2025-05-20, last updated 2025-05-20

-Summary

This study aims to develop and validate a machine learning model that uses information from tertiary lymphoid structures (TLSs)-specialized immune-related cell clusters found near tumors-to predict survival outcomes and immune characteristics in patients with locally advanced gastric cancer. By analyzing clinical data, pathology, and imaging results, the model may help doctors better understand a patient's prognosis and personalize treatment strategies. The study will also explore how TLS-related immune patterns relate to the effectiveness of certain therapies, potentially offering new insights for immune-based treatment planning.

Expand all sections · Collapse all

+Arms and interventions

Arm	Type	Interventions	Description
Locally Advanced Gastric Cancer Patients		Other: TLS-Informed Machine Learning Prognostic Model

TLS-Informed Machine Learning Prognostic Model (Other): This intervention involves the development and application of a machine learning-based prognostic model that integrates features derived from tertiary lymphoid structures (TLSs) identified in tumor pathology slides, along with clinical and immunological data, to predict overall survival and immune landscape in patients with locally advanced gastric cancer. The model utilizes digital pathology, image analysis, and advanced computational algorithms to quantify TLS-related characteristics and correlate them with patient outcomes. It is designed to stratify patients into risk groups and provide insight into the tumor immune microenvironment, aiming to support personalized treatment planning.

+Outcome measures

Primary outcomes

Overall Survival Predicted by TLS-Informed Machine Learning Model [Up to 5 Years Post-Surgery]

+Eligibility

Sex: All, ages: 18 Years to 80 Years

Inclusion Criteria:

Histologically confirmed locally advanced gastric adenocarcinoma (clinical stage cT2-T4 and/or N+)

Underwent curative-intent gastrectomy (with or without neoadjuvant therapy)

Availability of adequate tumor tissue specimens for TLS assessment via digital pathology

Complete baseline clinical, pathological, and follow-up data

Age ≥ 18 years

Written informed consent provided (if prospective study component is included)

Exclusion Criteria:

Distant metastases at the time of diagnosis or surgery (M1 stage)

Prior history of other malignancies within the past 5 years, except for adequately treated in situ carcinoma or non-melanoma skin cancer

Incomplete or missing essential clinical, pathological, or survival data

Poor-quality tissue samples not suitable for TLS quantification or digital analysis

Participation in another clinical trial that may interfere with the study outcomes