Table 1.
Baseline characteristics (study design, data source, selection criteria, time frame, age, and sample size) of 40 included studies on artificial intelligence–based ovarian cancer diagnosis with blood samples.
| Study | Study design | Data source | Selection criteria | Time frame | Age (y), mean or median | Sample size, n |
| Cai et al [24], 2024 | Retrospective | Data from Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, central China; Women’s Hospital, School of Medicine, Zhejiang University, eastern China; and Sun Yat-Sen University Cancer Center, southern China | Patients with a history of other malignant cancers or precancers, pregnancy in the last 6 mo, or affected by HIV; not newly diagnosed in any of the 3 hospitals; individuals without any available laboratory tests were excluded | From January 2012 to April 2021 | Cancer: (53/51/56)a; control: (34/34/48)a | 10,992 (3007/5641/2344)a |
| Abuzinadah et al [28], 2023 | Retrospective | Data from the Third Affiliated Hospital of Soochow University | All patients underwent postoperative case diagnosis, and none of them had received preoperative radiotherapy or chemotherapy | From July 2011 to July 2018 | NRb | 349 (244/105)c |
| Bifarin and Fernandez [29], 2024 | Retrospective | Data from a serum lipidomic analysis of ovarian cancer patients of Korean descent | NR | NR | NR | 325 (227/98)c |
| Cameron et al [30], 2023 | Retrospective | Data from the Welcome Trust Clinical Research Facility at the Western General Hospital, Edinburgh, the Emergency Medicine Research Group at the Edinburgh Royal Infirmary, the Beatson West of Scotland Cancer Centre in Glasgow, the University of Swansea, Royal Preston Hospital, and Manchester Cancer Research Centre | NR | NR | Ovary: 61; NCSd female individuals only: 56 | 385 (NR) |
| Chen et al [31], 2023 | Retrospective | Data from the Department of Gynecology of Harbin Medical University Cancer Hospital, Gene Expression Omnibus database, UCSC Xena | Patients with a primary radiological diagnosis of ovarian tumor; newly diagnosed patients without any significant comorbidities or history of previous malignancies; willingness to participate in the study and provision of written informed consent | From December 2020 to July 2021 | NR | 44 (44)e |
| Dhar et al [8], 2023 | Retrospective | Data from Indivumed (Hamburg, Germany) | NR | NR | NR | 351 (237/114)c |
| Hamidi et al [32], 2023 | Retrospective | Data from the Gene Expression Omnibus database, a Japanese nationwide research project, and patients with cancer who were referred or admitted to the National Cancer Center Hospital | The serum samples for noncancer controls who had no history of cancer and no hospitalization during the previous 3 months were collected; patients with cancer who were treated with preoperative chemotherapy and radiotherapy before serum collection were excluded | NR | Internal set: 52 | 3411 (2156/3079/92/240)f |
| Lai et al [33], 2023 | Retrospective | Data from clinical laboratory examination | NR | From January 2013 to October 2022 | 46.4 | 778 (545/233)g |
| Li et al [34], 2024 | Retrospective | Data from the Cancer Hospital of the University of Chinese Academy of Sciences (Zhejiang Cancer Hospital) | No patients with ovarian cancer received chemotherapy, radiotherapy, or surgery, no healthy donors had a history of cancer before sample collection | NR | Stage 1: 53.6; stage 2: 55.8; stage 3 and 4: 54.9; control: 34.1 | 69 (NR) |
| Reilly et al [35], 2023 | Retrospective and prospective | Data from multiple studies spanning multiple centers | Patient age ≥18 years; informed consent provided by the patient to participate in research; patient agreeable to phlebotomy; patient had a documented adnexal mass | NR | 47.5 | 2186 (NR) |
| Zhang et al [36], 2023 | Retrospective | Data from patients who underwent physical examination at Chinese People’s Armed Police Force and First Medicine Center, People's Liberation Army General Hospital | Discharge diagnosis confirmed by clinical signs, imaging, and pathology for patients with gynecologic tumors and benign gynecologic diseases; if no histopathological examination was available, it was consistently confirmed by ≥2 types of imaging evidence, availability of laboratory test data at the time of first diagnosis, and blood collection before treatment | From January 2010 to June 2019 | Ovarian cancer: (52.11/52.69/55.81)h; NOMGTi: (54.66/52.66/53.95)h; BGDj: (44.30/46.33/39.21)h; Healthy control: (49.96/47.26/49.67)h | 1633(600/301)g |
| Ahamad et al [37], 2022 | Retrospective | Data from Third Affiliated Hospital of Schow University | NR | From July 2017 to July 2018 | NR | 349 (85/21)c |
| Bahado-Singh et al [38], 2022 | Prospective | Data from Oakland University William Beaumont School of Medicine | NR | NR | Cases: 66.2; control: 67.8 | 17 (NR) |
| Gupta et al [39], 2022 | Retrospective | Data from 3 separate commercial biobanks: Dx Biosamples (San Diego, CA), Reprocell USA Inc (Beltsville, MD), and Fidelis Research AD (Sofa, Bulgaria) | NR | NR | NR | 1243 (681)k |
| Hinestrosa et al [40], 2022 | Retrospective | Data from a commercial biorepository (ProteoGenex, Inglewood, CA, United States) | The control group has no known history of cancer, autoimmune diseases, or neurodegenerative disorders, nor any presence of diabetes mellitus | From January 2014 to September 2020 | 60 | 323 (216/107)c |
| Irajizad et al [25], 2022 | Prospective | Data from Anderson Cancer Center and at the Fred Hutchinson Cancer Research Center | NR | NR | NR | 409 (108/118)c |
| Kim et al [41], 2022 | Retrospective | Data from Oakland University William Beaumont School of Medicine | NR | NR | NR | 269 (215)e |
| Li et al [42], 2022 | Prospective | Data from 3 institutions | Women diagnosed with benign, borderline, and malignant ovarian tumors | From December 2018 to January 2020 | NR | 362 (178/184)c |
| Pais et al [43], 2022 | Retrospective | Data from a commercially stored biological sample biobank (Invent diagnostica, Berlin, Germany) | NR | NR | NR | 181 (NR) |
| Jeong et al [44], 2021 | Retrospective | Data from Kangnam Sacred Heart Hospital | NR | From June 2014 to December 2020 | Cancer: 54; control: 49 | 730 (511/219)c |
| Lu et al [45], 2020 | Prospective | Data from the Third Affiliated Hospital of Soochow University | None of the patients with ovarian cancer received preoperative chemotherapy or radiotherapy | From July 2011 to July 2018 | NR | 349(235/114)c |
| Banaei et al [46], 2019 | Prospective | Data from the UMass Memorial Medical Center Chemotherapy Infusion Center and Gastroenterology Clinics and Innovative Research | NR | NR | NR | 20 (40/160)c |
| Whitwell et al [47], 2018 | Retrospective | Data from a synthetic dataset modeled from the United Kingdom Collaborative Trial of Ovarian Cancer Screening | Trial participants at enrollment were postmenopausal women aged 50-74 y who had no family history of ovarian cancer | NR | NR | 89 (NR) |
| Ivanova et al [48], 2016 | Retrospective | Data from the clinical diagnostic laboratory of the LLC LYTECH, the Blokhin Cancer Research Center of Russian Academy of Medical Sciences, the National Research Center of Coloproctology, the Moscow Dermatovenerologic Dispensary, clinical hospitals of Peoples’ Friendship University of Russia | NR | NR | Cancer: 52; control: 49 | 67 (NR) |
| Jiang et al [49], 2013 | Prospective | Data from the affiliated hospital, Sun Yat-Sen University | NR | NR | 61.7 | 87 (51/36)l |
| Yang et al [50], 2013 | Prospective | Data from the Peking University Third Hospital | NR | From January 2003 to December 2009 | Stage I/II: 54.8; stage III: 57.3; stage IV: 58.2; normal: 52.8; carcinoid: 51.6 | 246 (NR) |
| Shan et al [51], 2012 | Prospective | Data from the Tampa, Florida metropolitan area | Women with a prior unilateral or bilateral oophorectomy were ineligible, as were women with a previous history of cancer. All patients underwent preoperative radiologic imaging, either by pelvic ultrasound, CT, or MRI. Only patients who underwent surgery based on clinical suspicion of ovarian cancer were eligible | NR | NR | 423 (NR) |
| Thakur et al [52], 2011 | Prospective | Data from the FDA-NCI clinical proteomics program databank | NR | NR | NR | 216 (173/43)c |
| Donach et al [53], 2010 | Prospective | Data from Padua Hospital (now the Veneto Oncology Institute) | NR | From 1999 to 2005 | 48 | 201 (NR) |
| Ziganshin et al [54], 2008 | Prospective | Data from Byelorussian Oncology Center with patients with ovarian cancer and the Clinical Diagnostic Laboratory with clinically healthy women | NR | NR | Cancer: 51;control: 49 | 118 (NR) |
| Liu et al [55], 2007 | Prospective | Data from Northwestern University, Johns Hopkins University in Baltimore, MD, and the University of Innsbruck, Austria | Normal samples were from patients who had 4-year follow-up examinations to ensure that they did not have cancer at the time the samples were taken | From 1999 to 2002 | NR | 563 (315/78/170)m |
| Zhang et al [56], 2007 | Prospective | Data from the Duke University Medical Center, Durham, NC, St Bartholomew’s Hospital, London, United Kingdom, and the Groningen University Hospital, Groningen, Netherlands | NR | NR | NR | 468 (200/150)c |
| Chatterjee et al [57], 2006 | Retrospective | Data from the Barbara Ann Karmanos Cancer Institute, the MD Anderson Cancer Center, Weill Medical College of Cornell University, Northwestern University Robert H Lurie Comprehensive Cancer Center, and the Gynecologic Oncology Group Tissue Bank | NR | NR | NR | 129 (85/44)c |
| Lin et al [58], 2006 | Prospective | Data from the Tri-Service General Hospital, Taiwan, and Republic of China | Patients with any history of cancer, operations that had removed body organ, or current chronic or acute diseases were excluded | NR | NR | 65 (65)e |
| Liu [59], 2006 | Prospective | Data from Clinical Proteomic Program Databank | NR | NR | NR | 253 (NR) |
| Wu et al [60], 2006 | Prospective | Data from the Tri-Service General Hospital, Taiwan | No history of gynecologic tumors and had a normal pelvic examination and pelvic sonography | NR | NR | 65 (NR) |
| Li and Ramamohanrao [61],2004 | Prospective | Data from a public website | NR | From November 2003 | NR | 253 (215/112)c |
| Li et al [62], 2004 | Retrospective | Data from a public website | NR | From February 2002 | NR | 469 (100/116)c |
| Zhang et al [63], 1999 | Retrospective | Data from an existing data set of clinically diagnosed with pelvic masses and University of Texas MD Anderson Cancer Center | NR | NR | NR | 625 (174/255)c |
| Wilding et al [64], 1994 | Prospective | Data from the Hospital of the University of Pennsylvania | Patients with carcinoma in situ were excluded | NR | NR | 98 (NR) |
aTraining/external validation 1/external validation 2.
bNR: not reported.
cTraining/testing.
dNCS: noncancer symptomatic.
eTraining.
fTraining/internal validation/external validation 1/external validation 2.
gTraining/internal validation.
hTraining/internal validation/external validation.
iNOMGT: nonovarian malignant gynecologic tumor.
jBGD: benign gynecologic disease.
kTesting.
lTraining/prediction.
mTraining/testing/internal validation.