Endometrial Neoplasms
Conditions
Keywords
Endometrial Neoplasms, machine learning, Radiomics, Pathomics, TCGA classification
Brief summary
Molecular typing provides accurate information for the diagnosis, treatment and prognosis prediction of endometrial cancer, which has important clinical significance. However, due to its high cost and complicated process, it is difficult to be widely used in clinical practice. Based on the artificial intelligence method, this study fused the characteristics of MRI radiomics and pathomics, combined with the clinical pathological information, built a model to predict the molecular typing and prognosis, analyzed the biological characteristics of endometrial cancer from the multi-scale level, guided the personalized and precise diagnosis and treatment, in order to improve the prognosis of patients.
Detailed description
In this project, 150 cases of endometrial cancer were retrospectively collected, and 200 cases of endometrial cancer will be prospectively collected. All patients were pathologically confirmed and underwent Promise molecular typing. Before treatment, all patients completed abdominal MRI. Based on artificial intelligence technology, image features were extracted from magnetic resonance imaging, pathological features were extracted from pathological data, and clinical pathological data were collected at the same time. The treatment effect, recurrence and metastasis of patients were followed up, and the five-year survival rate and five-year progression free survival rate were calculated. It is proposed to focus on the following research: 1. Construction of molecular typing and prognosis prediction model of endometrial cancer based on magnetic resonance imaging Radiomics 2. Construction of molecular typing and prognosis prediction model of endometrial cancer based on pathomics. 3. Construction of a prediction model for molecular typing of endometrial cancer by integrating pathomics and radiomics.
Interventions
DIAGNOSTIC_TESTnext generation sequencing AND Immunohistochemical examination
First, the mismatch repair (MMR) proteins were detected by immunohistochemistry, and the deletion of one or more proteins was classified as d-MMR subtype; Then the POLE gene mutation detection was performed, and the mutation Changes were classified as POLE mutation; Finally, p53 was detected by immunohistochemistry, and p53 mutant (p53 abn) and p53 wild-type (p53wt) were distinguished.
Sponsors
Fujian Provincial Hospital
First Affiliated Hospital of Fujian Medical University
Gutian Hospital
Fujian Cancer Hospital
Study design
Observational model
COHORT
Time perspective
OTHER
Eligibility
Sex/Gender
FEMALE
Age
18 Years to 80 Years
Inclusion criteria
* •Pathologically confirmed as endometrial malignant tumor with complete pathological H＆E stained sections; * Age ≥ 18 years and ≤ 80 years; * No other malignant cancers was found; * The complete immunohistochemical and second-generation sequencing results can be used for the molecular typing of ProMisE; * Magnetic resonance examination was performed within 2 weeks before treatment, and there was at least one measurable lesion according to RECIST 1.1 Criteria.
Exclusion criteria
* • The image quality is poor or the tumor is too small due to serious graphic artifact and degeneration, and the ROI cannot be accurately delineated; * Patients who received any antitumor therapy before surgery; * Diagnostic endometrial biopsy before MRI
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Application of magnetic resonance imaging radiomics and pathomics to construct a model for predicting the molecular classification and prognosis of endometrial cancer | 2026-12-21 | The imaging and pathological features of endometrial cancer patients were extracted by artificial intelligence method. Combined with clinicopathological risk factors and survival time, an imaging nomogram was constructed by lasso regression method to predict the molecular classification and prognosis of endometrial cancer. ROC curve was used to evaluate the test efficiency of the model. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Application of magnetic resonance imaging radiomics to construct a model for predicting the molecular classification and prognosis of endometrial cancer | 2026-12-21 | The imaging features of endometrial cancer patients were extracted by artificial intelligence method. Combined with clinicopathological risk factors and survival time, an imaging nomogram was constructed by lasso regression method to predict the molecular classification and prognosis of endometrial cancer. ROC curve was used to evaluate the test efficiency of the model. |
Other
| Measure | Time frame | Description |
|---|---|---|
| Application of pathomics to construct a model for predicting the molecular classification and prognosis of endometrial cancer | 2026-12-21 | The pathomics features of endometrial cancer patients were extracted by artificial intelligence method. Combined with clinicopathological risk factors and survival time, an imaging nomogram was constructed by lasso regression method to predict the molecular classification and prognosis of endometrial cancer. ROC curve was used to evaluate the test efficiency of the model. |
Countries
China
Contacts
Primary ContactJian Chen, Master
Outcome results
None listed