Artificial Intelligence, Cervical Cancer Screening, Machine Learning, Risk Assessment
Conditions
Brief summary
By collecting non-image medical data of women undergoing cervical screening in multiple centers in China, including age, HPV infection status, HPV infection type, TCT results, and colposcopy biopsy pathology results, a multi-source heterogeneous cervical lesion collaborative research big data platform was established. Based on artificial intelligence (AI) machine learning, cervical lesion screening features are refined, a multi-modal cervical cancer intelligent screening prediction and risk triage model is constructed, and its clinical application value is preliminarily explored.
Detailed description
By collecting non-image medical data of women undergoing cervical screening in multiple centers in China, including age, HPV infection status, HPV infection type, TCT results, and colposcopy biopsy pathology results, a multi-source heterogeneous cervical lesion collaborative research big data platform was established. Based on artificial intelligence (AI) machine learning, cervical lesion screening features are refined, a multi-modal cervical cancer intelligent screening prediction and risk triage model is constructed, and its clinical application value is preliminarily explored. The effect of clinical application of the model was evaluated by internal data from Fujian Province and external data from several other regions in China.
Interventions
OTHERArtificial intelligence model building
Using non-image medical data of cervical lesions and clinical pathology results in different medical institutions, machine learning is adopted to establish multiple multi-modal cervical cancer intelligent screening prediction models. This method was used to analyze the prediction performance of the multi-modal cervical cancer intelligent screening prediction and risk triage model, and to evaluate and optimize the self-learning ability of the established multi-modal cervical cancer intelligent screening prediction model.
Sponsors
Fujian Maternity and Child Health Hospital
Study design
Observational model
COHORT
Time perspective
RETROSPECTIVE
Eligibility
Sex/Gender
FEMALE
Age
25 Years to 64 Years
Healthy volunteers
Yes
Inclusion criteria
* Age 25-64 years old; * There was no history of precancerous lesions or cervical cancer; * No previous cervical surgery or cervical removal;
Exclusion criteria
* HPV test results are not available; * Pregnant or lactating women; * There is a serious immune system disease, and the disease is active;
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Cervical histopathology | within 8 weeks， | Cervical histopathological diagnosis within 8 weeks |
| colposcopy | Percentage of patients diagnosed with cervical intraepithelial neoplasia of grade 3 (CIN3) or worse by cervical histopathological measurements within 8 weeks | Colposcopists use colposcopic equipment to investigate the occurrence of cervical and vaginal lesions within 8 weeks |
Countries
China
Outcome results
None listed