Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis

Abstract BackgroundImmune checkpoint inhibitors represent an effective therapeutic approach for advanced gastric cancer. Their efficacy largely depends on the status of tumor biomarkers including human epidermal growth factor receptor 2 (HER2), programmed death-ligand 1 (PD-L1...

Full description

Saved in:
Bibliographic Details
Main Authors: Shiqi Wang, Na Chai, Jingji Xu, Pengfei Yu, Luguang Huang, Quan Wang, Zhifeng Zhao, Bin Yang, Jiangpeng Wei, Xiangjie Wang, Gang Ji, Minwen Zheng
Format: Article
Language:English
Published: JMIR Publications 2025-07-01
Series:JMIR Cancer
Online Access:https://cancer.jmir.org/2025/1/e67379
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849716984183259136
author Shiqi Wang
Na Chai
Jingji Xu
Pengfei Yu
Luguang Huang
Quan Wang
Zhifeng Zhao
Bin Yang
Jiangpeng Wei
Xiangjie Wang
Gang Ji
Minwen Zheng
author_facet Shiqi Wang
Na Chai
Jingji Xu
Pengfei Yu
Luguang Huang
Quan Wang
Zhifeng Zhao
Bin Yang
Jiangpeng Wei
Xiangjie Wang
Gang Ji
Minwen Zheng
author_sort Shiqi Wang
collection DOAJ
description Abstract BackgroundImmune checkpoint inhibitors represent an effective therapeutic approach for advanced gastric cancer. Their efficacy largely depends on the status of tumor biomarkers including human epidermal growth factor receptor 2 (HER2), programmed death-ligand 1 (PD-L1; combined positive score ≥1), and microsatellite instability-high (MSI-H). To noninvasively evaluate these biomarkers, researchers have developed radiomic models for individual biomarker prediction. However, in clinical practice, holistic prediction of these biomarkers as an integrated system is more efficient. Currently, the feasibility of implementing radiomics-based comprehensive biomarker prediction remains unclear, requiring further investigation. ObjectiveThis study aimed to develop a radiomics-based predictive model using multiphase computed tomography (CT) images to holistically evaluate HER2, PD-L1, and MSI-H status in patients with gastric cancer. MethodsA retrospective analysis was conducted on 461 patients with gastric cancer who underwent radical gastrectomy between 2019 and 2022. Clinical data, contrast-enhanced CT images (arterial phase [AP] and portal venous phase [PP]), and pathological results were collected. Patients were categorized into two groups: (1) the programmed cell death protein-1 inhibitor panel-positive group, comprising patients with HER2 overexpression, PD-L1 positive, or MSI-H status; and (2) the negative group, comprising patients without HER2 amplification, PD-L1 negative, or microsatellite instability-low or microsatellite stable condition. Radiomic features (including first-order statistics, shape features, and wavelet-derived textures) were extracted from both AP and PP images, yielding 1834 features per phase. Least absolute shrinkage and selection operator regression was applied to select key features. In total, 3 models were constructed using the Extreme Gradient Boosting algorithm: AP-only (8 features), PP-only (22 features), and a fused model combining AP and PP features (20 features: 6 AP and 14 PP features). Model performance was evaluated using area under the curve (AUC), sensitivity, specificity, and decision curve analysis. ResultsOf the 461 patients, 147 patients (31.9%) were classified into the panel-positive group. The clinical features were similar between the 2 groups. The fused model demonstrated superior performance in the test set (AUC 0.82, 95% CI 0.68‐0.95), significantly outperforming AP-only (AUC 0.61, 95% CI 0.47‐0.74) and PP-only models (AUC 0.70, 95% CI 0.49‐0.91). Sensitivity and specificity for the AP-only, PP-only, and the fused model were 0.33 and 0.85; 0.50 and 0.86; and 0.60 and 0.83, respectively. Decision curve analysis confirmed that the fused model provided higher clinical net benefit across threshold probabilities. ConclusionsThe construction of integrated biomarker prediction models through radiomics demonstrates technical feasibility, offering a promising methodology for comprehensive tumor characterization.
format Article
id doaj-art-bb2c7a5d1b454ccdac210e794040b5f7
institution DOAJ
issn 2369-1999
language English
publishDate 2025-07-01
publisher JMIR Publications
record_format Article
series JMIR Cancer
spelling doaj-art-bb2c7a5d1b454ccdac210e794040b5f72025-08-20T03:12:49ZengJMIR PublicationsJMIR Cancer2369-19992025-07-0111e67379e6737910.2196/67379Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort AnalysisShiqi Wanghttp://orcid.org/0000-0002-2381-044XNa Chaihttp://orcid.org/0009-0003-5691-5912Jingji Xuhttp://orcid.org/0000-0002-9767-0742Pengfei Yuhttp://orcid.org/0000-0003-3825-8798Luguang Huanghttp://orcid.org/0000-0003-3296-4181Quan Wanghttp://orcid.org/0000-0002-2821-5017Zhifeng Zhaohttp://orcid.org/0000-0002-6311-1811Bin Yanghttp://orcid.org/0000-0001-6180-7412Jiangpeng Weihttp://orcid.org/0000-0001-6619-8695Xiangjie Wanghttp://orcid.org/0000-0001-8869-5596Gang Jihttp://orcid.org/0000-0001-9722-872XMinwen Zhenghttp://orcid.org/0000-0002-0862-3208 Abstract BackgroundImmune checkpoint inhibitors represent an effective therapeutic approach for advanced gastric cancer. Their efficacy largely depends on the status of tumor biomarkers including human epidermal growth factor receptor 2 (HER2), programmed death-ligand 1 (PD-L1; combined positive score ≥1), and microsatellite instability-high (MSI-H). To noninvasively evaluate these biomarkers, researchers have developed radiomic models for individual biomarker prediction. However, in clinical practice, holistic prediction of these biomarkers as an integrated system is more efficient. Currently, the feasibility of implementing radiomics-based comprehensive biomarker prediction remains unclear, requiring further investigation. ObjectiveThis study aimed to develop a radiomics-based predictive model using multiphase computed tomography (CT) images to holistically evaluate HER2, PD-L1, and MSI-H status in patients with gastric cancer. MethodsA retrospective analysis was conducted on 461 patients with gastric cancer who underwent radical gastrectomy between 2019 and 2022. Clinical data, contrast-enhanced CT images (arterial phase [AP] and portal venous phase [PP]), and pathological results were collected. Patients were categorized into two groups: (1) the programmed cell death protein-1 inhibitor panel-positive group, comprising patients with HER2 overexpression, PD-L1 positive, or MSI-H status; and (2) the negative group, comprising patients without HER2 amplification, PD-L1 negative, or microsatellite instability-low or microsatellite stable condition. Radiomic features (including first-order statistics, shape features, and wavelet-derived textures) were extracted from both AP and PP images, yielding 1834 features per phase. Least absolute shrinkage and selection operator regression was applied to select key features. In total, 3 models were constructed using the Extreme Gradient Boosting algorithm: AP-only (8 features), PP-only (22 features), and a fused model combining AP and PP features (20 features: 6 AP and 14 PP features). Model performance was evaluated using area under the curve (AUC), sensitivity, specificity, and decision curve analysis. ResultsOf the 461 patients, 147 patients (31.9%) were classified into the panel-positive group. The clinical features were similar between the 2 groups. The fused model demonstrated superior performance in the test set (AUC 0.82, 95% CI 0.68‐0.95), significantly outperforming AP-only (AUC 0.61, 95% CI 0.47‐0.74) and PP-only models (AUC 0.70, 95% CI 0.49‐0.91). Sensitivity and specificity for the AP-only, PP-only, and the fused model were 0.33 and 0.85; 0.50 and 0.86; and 0.60 and 0.83, respectively. Decision curve analysis confirmed that the fused model provided higher clinical net benefit across threshold probabilities. ConclusionsThe construction of integrated biomarker prediction models through radiomics demonstrates technical feasibility, offering a promising methodology for comprehensive tumor characterization.https://cancer.jmir.org/2025/1/e67379
spellingShingle Shiqi Wang
Na Chai
Jingji Xu
Pengfei Yu
Luguang Huang
Quan Wang
Zhifeng Zhao
Bin Yang
Jiangpeng Wei
Xiangjie Wang
Gang Ji
Minwen Zheng
Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
JMIR Cancer
title Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
title_full Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
title_fullStr Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
title_full_unstemmed Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
title_short Prediction of a Panel of Programmed Cell Death Protein-1 (PD-1) Inhibitor–Sensitive Biomarkers Using Multiphase Computed Tomography Imaging Textural Features: Retrospective Cohort Analysis
title_sort prediction of a panel of programmed cell death protein 1 pd 1 inhibitor sensitive biomarkers using multiphase computed tomography imaging textural features retrospective cohort analysis
url https://cancer.jmir.org/2025/1/e67379
work_keys_str_mv AT shiqiwang predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT nachai predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT jingjixu predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT pengfeiyu predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT luguanghuang predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT quanwang predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT zhifengzhao predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT binyang predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT jiangpengwei predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT xiangjiewang predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT gangji predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis
AT minwenzheng predictionofapanelofprogrammedcelldeathprotein1pd1inhibitorsensitivebiomarkersusingmultiphasecomputedtomographyimagingtexturalfeaturesretrospectivecohortanalysis