JMIR Publications

ABSTRACT

Background:

Ovarian cancer is a highly fatal gynecologic malignancy requiring complex clinical decision-making. Large language models (LLMs), such as GPT-based systems, have shown promise in health care tasks including diagnosis, treatment recommendations, and documentation. However, their application in ovarian cancer care has not yet been systematically evaluated.

Objective:

This protocol describes a systematic review and meta-analysis aimed at evaluating the use, performance, and clinical impact of LLMs in ovarian cancer management across various applications (e.g., diagnosis, treatment planning, prognosis, patient communication).

Methods:

The review will follow PRISMA-P guidelines and has been registered with PROSPERO. We will search biomedical and technical databases (PubMed, Embase, IEEE Xplore, CNKI, etc.) from inception to December 31, 2025. We will include clinical evaluations, validation studies, and real-world implementation reports involving LLM applications in ovarian cancer. Two independent reviewers will perform study screening, data extraction, and verification. Outcomes will include diagnostic and predictive performance metrics (e.g., accuracy, sensitivity, specificity, area under the curve [AUC]), impacts on clinical processes, and reported limitations. Where appropriate, we will conduct quantitative meta-analyses using R (meta, metafor, mada packages), including bivariate models for sensitivity and specificity.

Results:

As of November 2025, this review is in the protocol stage. The PROSPERO registration (CRD420251144051) is complete, and the literature search is underway. We anticipate completing study selection and data extraction by mid-2026. No results are available yet, but the final review will provide pooled performance estimates and narrative insights into the strengths and limitations of LLM use in ovarian cancer care.

Conclusions:

This systematic review will be the first to synthesize evidence on LLM applications in ovarian cancer, identifying high-value use cases and areas that require further development. The findings are expected to support evidence-based integration of LLMs into gynecologic oncology workflows and guide future research on the safe and effective deployment of artificial intelligence in ovarian cancer care. Clinical Trial: PROSPERO CRD420251144051