Previously submitted to: JMIR Medical Informatics (no longer under consideration since Jan 08, 2026)
Date Submitted: Aug 8, 2025
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Development and Validation of a Prognostic Model for Stage IV Breast Cancer Based on Primary Tumor Resection with Machine Learning Methods：Retrospective Cohort Study
ABSTRACT
Background:
Primary Tumor Resection (PTR) remains controversial among women with stage IV breast cancer.
Objective:
This study aims to adopt a machine learning (ML) approach to clarify the disputed correlation of primary tumor resection on stage IV breast cancer survival, aiming to pinpoint patients who stand to gain from PTR treatment with enhanced clinical applicability and predictive precision.
Methods:
A propensity-score matched analysis of stage IV breast cancer patients in the SEER registry (2000-2020) was conducted, using Cox regression and Kaplan-Meier methods to estimate overall survival (OS) and cancer-specific survival (CSS). Among five ML models, an internally and externally validated ML model was crafted for predicting survival outcomes, with a user-friendly web platform based on the shiny platform for clinical use.
Results:
In a cohort of 13,702 stage IV breast cancer patients, with 5,732 matched subjects, Cox regression analysis showed PTR's positive associations with OS (HR, 0.75; 95% CI, 0.69 to 0.82) and CSS (HR, 0.75; 95% CI, 0.68 to 0.82). Subgroup analysis indicated better survival for patients with tumors ≤5 cm, N2 status (OS, HR, 0.52; 95% CI, 0.44 - 0.63;CSS,HR, 0.52; 95% CI, 0.43 - 0.63), and HER2 overexpression (OS, HR, 0.56; 95% CI, 0.52 - 0.61;CSS,HR, 0.56; 95% CI, 0.46 - 0.67), especially those without systemic treatment. The best outcomes were seen with PTR combined with chemotherapy and radiation (OS, HR, 0.40; 95% CI, 0.37 - 0.44; CSS, HR, 0.03; 95% CI, 0.01 - 0.10). The Support Vector Machine (SVM) model (6-month: AUC=0.935; 1-year: AUC=0.945; 2-year: AUC=0.941; 3-year: AUC=0.921) was identified as the most precise tool for survival prediction, demonstrating high accuracy and consistency across external datasets. Furthermore, a user-friendly web application was created to make the prognostic model more accessible.
Conclusions:
Patients with tumors ≤5 cm, N2 status, and HER2 overexpression experience improved survival with PTR. Our ML model based on eight clinical indicators predicts survival for stage IV cancer patients, guiding surgical candidacy and clinical decisions. The accompanying online platform facilitates patient assessment for healthcare professionals. Clinical Trial: Name of the registry: Chinese Clinical Trial Registry. Unique Identifying number or registration ID: ChiCTR2400089791 Hyperlink to your specific registration (must be publicly accessible and will be checked): https://www.chictr.org.cn/showproj.html?proj=244567
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.