Combination of CT and Ultrasound Radiomics Combined With Liquid Biopsy to Predict Neoadjuvant Chemotherapy Response in Patients With Locally Advanced Gastric Cancer: A Prospective Study
试验速览
- 阶段
- 不适用
- 状态
- 尚未招募
- 发起方
- 入组人数
- 300
- 试验地点
- 1
- 主要终点
- Accuracy of pathological response to neoadjuvant chemotherapyin patients with locally advanced gastric cancer models
研究概览
简要总结
This prospective cohort study aims to construct an artificial intelligence (AI)-derived predictive model for neoadjuvant chemotherapy response prediction in patients with locally advanced gastric cancer based on preoperative ultrasound (US), computed tomography (CT) images and liquid biopsy. Additionally, we explore the potential biological mechanisms behind this model.
研究设计
- 研究类型
- Observational
- 观察模型
- Cohort
- 时间视角
- Prospective
入排标准
- 年龄范围
- 18 Years 至 80 Years（Adult, Older Adult）
- 性别
- All
- 接受健康志愿者
- 否
入选标准
- •Capable of understanding the study and voluntarily signing the written informed consent form (ICF) prior to any study-specified research procedures.
- •Aged ≥18 and ≤80 years old at the time of ICF signing.
- •Pathologically confirmed locally advanced gastric cancer (LAGC, cT2NxM0-cT4NxM0) with clinical indications for neoadjuvant chemotherapy.
- •Completion of gastrointestinal contrast-enhanced ultrasound and contrast-enhanced abdominal CT before neoadjuvant chemotherapy.
- •Provision of peripheral blood samples before chemotherapy (for genetic and protein detection).
- •Availability of postoperative pathological specimens for TRG grading after standardized neoadjuvant chemotherapy.
- •Willing and able to comply with all study protocol requirements.
排除标准
- •Diagnosis of non-primary gastric cancer.
- •Incomplete imaging data, failure to collect peripheral blood samples, or substandard sample quality.
- •Discontinued chemotherapy, modified treatment regimen, or lack of complete postoperative pathological assessment.
- •Unavailable follow-up data precluding evaluation of chemotherapy response.
- •Concurrent participation in another clinical trial; or any other conditions judged by investigators to warrant subject withdrawal, including severe comorbidities requiring simultaneous treatment (psychiatric disorders included), alcohol dependence, substance abuse, or familial/social factors that may compromise subject safety or treatment compliance.
研究组 & 干预措施
Good pathological response
Patients with locally advanced gastric cancer achieved TRG grade 0-1 after the neoadjuvant chemotherapy
Poor pathological response
Patients with locally advanced gastric cancer achieved TRG grade 2-3 after the neoadjuvant chemotherapy
结局指标
主要结局
Accuracy of pathological response to neoadjuvant chemotherapyin patients with locally advanced gastric cancer models
时间窗: Immediately evaluated after the pathological response model was built
This prospective study will collect contrast-enhanced abdominal CT and ultrasound images, plus peripheral blood samples, from 300 patients with locally advanced gastric cancer (LAGC) before neoadjuvant chemotherapy. Using deep learning and machine learning, we will build a tumor regression grade (TRG)-based model to predict pathological response. TRG classification follows the NCCN Guidelines (v4, 2021): TRG 0-1 indicates good response; TRG 2-3, poor response. Model performance is assessed for diagnostic accuracy and stability, quantified by AUC and precision-recall curve.
Accuracy of pathological response to neoadjuvant chemotherapyin patients with locally advanced gastric cancer models
时间窗: The pathological response prediction model will be assessed immediately after its development.
This prospective study will collect contrast-enhanced abdominal CT and ultrasound images, as well as peripheral blood samples, from 300 patients with locally advanced gastric cancer (LAGC) prior to neoadjuvant chemotherapy. Using deep learning and machine learning algorithms, we will construct a tumor regression grade (TRG)-oriented model to predict pathological response to treatment. TRG classification is defined in accordance with the NCCN Guidelines (Version 4, 2021): TRG 0-1 indicates favorable response; TRG 2-3 poor response. The diagnostic accuracy and stability of the model will be evaluated, with performance quantified via the AUC and precision-recall curve.
次要结局
未报告次要终点
研究者
Liu Yang
Attending Physician
Qianfoshan Hospital