AI-Based Multi-Omics Approaches in Gynecological Cancers: A Review of Diagnostic, Prognostic, and Therapeutic Applications
Background: Gynecological cancers, particularly ovarian, endometrial, and cervical cancers, have a high disease burden in terms of incidence and mortality, and their diagnosis and treatment remain challenging. AI-based multi-omics approaches integrate multilayer data, enabling comprehensive modeling and clinical decision-making. This narrative review aimed to critically examine the current and prospective applications of these approaches in three common gynecological cancers and to evaluate their clinical potential. Methods: In this narrative review, following searches of PubMed, Web of Science, Scopus, and Google Scholar and screening of 1254 records, 51 original studies were included and synthesized qualitatively and descriptively. Relevant systematic reviews and meta-analyses were consulted only for contextualization and were not included in the primary evidence synthesis. Results: AI-based multi-omics approaches in gynecological cancers may help distinguish cancer from healthy tissue, benign from malignant conditions, and early from advanced stages; identify molecular subgroups; predict survival and recurrence; predict responses to chemotherapy and immunotherapy; and discover therapeutic targets. Minimally invasive biofluids and machine learning models showed better performance than CA125 and HE4 in some studies. Factors such as the tumor microenvironment, immunity, LDHA, and ITGA1 were associated with drug resistance in computational or preclinical studies, whereas HRD and TP53 alterations are clinically validated determinants of platinum response. In cervical cancer, the evidence was more limited and heterogeneous. Overall, the findings are promising but preliminary, and further extensive clinical validation is needed. Conclusions: AI-based multi-omics approaches may contribute to early diagnosis, molecular classification, prognostic assessment, and treatment selection in gynecological cancers; however, they remain at the proof-of-concept and initial validation stages. Prospective, multicenter, standardized studies with external validation are necessary for clinical application. These tools may serve as complementary approaches to current diagnostic pathways.
Authors
- Mai Abdel Haleem Abusalah (ORCID: https://orcid.org/0000-0002-8473-9392)
- Zahra Shahabinia (ORCID: https://orcid.org/0009-0009-9955-4735)
- Do‐Youn Lee (ORCID: https://orcid.org/0000-0003-0886-1713)
- Leila Allahqoli (ORCID: https://orcid.org/0000-0002-9851-6771)
- Afrooz Mazidimoradi (ORCID: https://orcid.org/0000-0002-4068-991X)
- Hamid Salehiniya (ORCID: https://orcid.org/0000-0001-7642-5214)
- Maedeh Mirasheh (ORCID: https://orcid.org/0009-0006-8637-432X)
Institutions
- Al-Ahliyya Amman University (JO)
- Kookmin University (KR)
- Cyprus International University (CY)
- Shiraz University of Medical Sciences (IR)
- Tehran University of Medical Sciences (IR)
- Birjand University of Medical Sciences (IR)
Publication Details
- Journal
- Journal of Clinical Medicine
- Published
- 2026-11-13
- DOI
- https://doi.org/10.3390/jcm15228907
- Primary Topic
- AI in cancer detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00