International Journal of Gynecological Cancer最新文献

Objective: Natural language processing is emerging as a key application of artificial intelligence in oncology. This systematic review aims to evaluate the performance and methodological frameworks of natural language processing systems in gynecologic oncology.

Methods: We conducted a systematic review following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis 2020 guidelines. MEDLINE, EMBASE, and Web of Science were searched for studies published between January 2015 and February 2025. Outcomes were synthesized across 3 research questions: the accuracy of natural language processing systems used as consultation service platforms; the accuracy of natural language processing systems used as clinical decision support systems; and the benchmarking methodologies applied, including their associated methodological outcomes. Consultation service platforms deliver general medical information, whereas clinical decision support systems provide recommendations that are integrated into the patient's clinical workflow.

Results: This review analyzed 12 retrospective studies. Consultation service platforms were less accurate than clinicians (60% vs 86.7%) and rated lower in response quality (2.96/5 vs 4.2/5) but outperformed guideline-based answers (1.54/2 vs 1.38/2). In cervical cancer, ChatGPT surpassed experts (7.0 vs 6.1). ChatGPT-4 showed a concordance of 70% with the National Comprehensive Cancer Network and 60% with the European Society of Gynaecological Oncology guidelines in clinical decision support tasks, with an overall recommendation accuracy of 75%. IBM Watson achieved a 72.8% concordance with guidelines. Prompting was applied from 100% to 37.5% across studies. Qualitative benchmarking varied across studies: 83.3% used clinical guidelines and 37.5% of consultation service platforms studies used expert answers. Four- or 5-point scales and binary scoring were used to assess consultation service platforms and clinical decision support systems, respectively.

Conclusions: Clinicians remain superior in complex reasoning, but natural language processing systems demonstrate robust performance in guideline-driven tasks, with advantages in speed, readability, and reproducibility. However, performance declined in nuanced scenarios and among under-represented patient sub-groups. Large language models currently play a supportive rather than substitutive role in gynecologic oncology.

Objective: This study aimed to assess the frequency of MDM4 amplification in endometrial carcinomas and its association with clinicopathological features, particularly relapse risk in early-stage low-grade endometrioid endometrial carcinoma.

Methods: We conducted a case-control study including 110 stage I to II low-grade endometrioid endometrial carcinomas (39 relapsing and 71 non-relapsing tumors) and an additional cohort of 82 high-grade endometrial carcinomas. An independent retrospective cohort of 194 early-stage low-grade endometrioid endometrial carcinomas was used to validate the results. MDM4 amplification was analyzed by fluorescent in situ hybridization. Gene expression of MDM4 and GADD45A was measured by quantitative real-time polymerase chain reaction. Immunohistochemistry assessed hormone receptor status, p53, and mismatch repair proteins. POLE and CTNNB1 mutations were evaluated by Sanger sequencing.

Results: MDM4 amplification was detected in 31 of 186 tumors (16.7%): 14 of 104 low-grade endometrioid endometrial carcinomas (13.5%) and 17 of 82 high-grade endometrial carcinomas (20.7%). Among low-grade endometrioid endometrial carcinomas, amplification was significantly more frequent in relapsing tumors (11/38, 28.9%) than in non-relapsing ones (3/66, 4.5%) (p <.01). Concordant MDM4 amplification status was observed between biopsies and hysterectomy samples. In some relapsing cases, amplification emerged during disease progression. Multi-variable Cox regression identified MDM4 amplification as an independent predictor of relapse, alongside lymphovascular space invasion, tumor necrosis, and myometrial invasion. In the independent cohort, MDM4 amplification was associated with progression-free survival.

Conclusions: Our study suggested that MDM4 amplification is a potential predictor of the risk of recurrence in early-stage low-grade endometrioid endometrial carcinoma that can be easily evaluated in endometrial samples.

Background: The therapeutic value of lymphadenectomy in endometrial cancer remains controversial, particularly in high-risk patients as defined by the European Society of Gynaecological Oncology-European Society for Radiotherapy and Oncology 2021 histologic and molecular classification. While pelvic and para-aortic lymphadenectomy is the standard of care for high-risk endometrial cancer, sentinel lymph node (SLN) mapping has emerged as a promising alternative for low-risk patients with reduced post-operative morbidity. Recently, the advent of molecular classification has significantly influenced the diagnosis and management of endometrial cancer. In particular, p53 mutation has been recognized as an important prognostic marker and has been included in the criteria for high-risk classification.

Primary objective: To evaluate the impact of SLN mapping versus pelvic and para-aortic lymphadenectomy on disease-free survival at 3 years in high-risk patients with endometrial cancer with p53 mutation and a negative pre-operative positron emission tomography computed tomography scan.

Study hypothesis: SLN mapping will provide non-inferior survival outcomes compared with pelvic and para-aortic lymphadenectomy in this population, with reduced post-operative morbidity.

Trial design: This is a prospective, multicentre, non-inferiority, open-label, de-escalation, randomized, controlled trial in patients with International Federation of Gynecology and Obstetrics 2023 stage I and II endometrial cancer with p53 mutation and negative pre-operative positron emission tomography computed tomography scans. Patients will be randomized 1:1 to SLN mapping or pelvic and para-aortic lymphadenectomy.

Major inclusion/exclusion criteria: Eligible participants are aged ≥18 years, with International Federation of Gynecology and Obstetrics 2023 stage I or II endometrial cancer confirmed by magnetic resonance imaging, a p53 mutation verified by biopsy, and planned surgery including total hysterectomy, bilateral salpingo-oophorectomy, and omentectomy if indicated. Exclusion criteria include recurrent endometrial cancer, prior chemotherapy, radiotherapy, endocrine therapy, or contraindications to lymphadenectomy or laparoscopy.

Primary endpoint: Disease-free survival at 36 months.

Sample size: A total of 374 patients.

Estimated dates for completing accrual and presenting results: Estimated completion of recruitment in September 2028, with 36 months of follow-up completed by September 2031.

Trial registration: ClinicalTrials.gov Identifier: NCT06900582.

Objective: To assess clinical utility of pre-operative imaging in cervical cancer beyond pelvic magnetic resonance imaging (MRI) in patients with pre-operative International Federation of Gynecology and Obstetrics (FIGO) stage IB2 or less.

Methods: We retrospectively identified patients who underwent evaluation or received consultation for newly diagnosed cervical squamous cell carcinoma, adenocarcinoma, or adeno-squamous carcinoma at our institution from January 2006 until February 2024. Patients with stage ≤IB2 disease on examination and a pre-operative pelvic MRI demonstrating a tumor ≤4 cm were included. Cases with evidence of gross pelvic nodal involvement or unequivocal parametrial/vaginal extension on MRI were excluded. All patients then underwent surgical treatment. Patients were included if they also underwent chest, abdominal, and pelvic computed tomography with/without positron emission tomography. Additional imaging was performed prior to consultation at our institution or at the treating physician's discretion. We sought to assess the findings of additional imaging in identifying extra-pelvic gross nodal or extra-nodal abdominal and/or chest disease. We did not seek to identify the role of imaging in identifying microscopic disease in normal-sized lymph nodes and such cases were included.

Results: Among 183 patients, the median age at diagnosis was 36 years (range; 18-81); 100 (54.6%) had adenocarcinoma, 78 (42.6%) squamous cell carcinoma, and 5 (2.7%) adeno-squamous carcinoma. The final pathologic FIGO 2018 stages included IA1 (n = 34, 18.6%), IA2 (n = 18, 9.8%), IB1 (n = 96, 52.5%), IB2 (n = 14, 7.7%), IB3 (n = 1, 0.5%), and IIIC1 (n = 20, 10.9%). The median tumor size was 0 mm (range; 0-38) on imaging and 8 mm (range; 0-41) on final pathology. Twenty-eight patients (15.3%) had non-specific/borderline enlarged pelvic lymph nodes on MRI, 6 (21.4%) of whom had final pathologic lymph node involvement. Thirty-four patients (18.6%) had "extra-pelvic" findings on computed tomography with/without positron emission tomography; 21 (61.8%) had non-specific findings, and 13 (38.2%) underwent further diagnostic intervention but none had cervical cancer metastases. There were no cervical cancer-related findings on additional imaging beyond MRI of the pelvis. Additionally, no extra-pelvic disease was encountered intra-operatively. The false-positive rate for imaging to detect extra-pelvic intra-abdominal metastasis of cervical carcinoma was 100% (13 of 13, 95% confidence interval [CI] 75.3% to 100%), with no false negatives (0%, 95% CI 0% to 1.7%).

Conclusions: For patients with cervical carcinoma ≤4 cm and confined to the cervix on pre-operative MRI, additional imaging appears to be of limited utility, leading to unnecessary interventions.

Objective: This study aimed to evaluate the associations of parity, breastfeeding duration, oral contraceptive use, menopausal hormone replacement therapy, age at menarche, age at menopause, and reproductive lifespan with ovarian cancer risk in post-menopausal women using large-scale real-world data from the Korean National Health Insurance Service.

Methods: This nationwide, population-based retrospective cohort study used National Health Insurance Service health screening and claims data. The study included 3,754,906 post-menopausal women aged 40 to 79 years who underwent national health examinations between 2009 and 2012. Reproductive factors were assessed at baseline using standardized self-administered questionnaires, whereas menopausal hormone replacement therapy exposure was identified from prescription records. Incident ovarian cancer was defined as ≥3 claims with an International Classification of Diseases, 10th Revision (ICD-10) C56 diagnosis. Univariable and multi-variable Cox proportional hazards regression analyses were performed to estimate the hazard ratios and 95% confidence intervals.

Results: During follow-up, 7702 women (0.2%) developed ovarian cancer. Compared with nulliparity, having ≥2 births was associated with lower risk. Breastfeeding for ≥12 months and oral contraceptive use for ≥1 year were also associated with reduced risk. By contrast, menopause at ≥55 years and a reproductive lifespan of ≥40 years were associated with increased risk. Menopausal hormone replacement therapy use for ≥5 years was also associated with higher risk, whereas age at menarche showed no association with ovarian cancer.

Conclusions: In this cohort, reproductive factors that reduce lifetime ovulatory exposure were associated with lower ovarian cancer risk, whereas factors that prolong ovulatory lifespan and long-term menopausal hormone replacement therapy were associated with higher risk. These observational findings may help refine ovarian cancer risk assessment and counseling in post-menopausal women, but do not establish causality.