Clear Cell Renal Cell Carcinoma, Deep Learning, Tumor Grading
Conditions
Brief summary
This study aims to establish an effective deep learning model to extract relevant information about renal tumors and kidneys from computed tomography (CT) images and predict the pathological grades of clear cell renal cell carcinoma (ccRCC). Retrospective data were collected from 483 ccRCC patients across three medical centers. Arterial phase and portal venous phase CT images from the dataset were segmented for renal tumors and kidneys. Three convolutional neural networks (CNNs) were employed to extract features from the regions of interest (ROI) in the CT images across multiple dimensions including 3D, 2.5D, and 2D. Least absolute shrinkage and selection (LASSO) regression was used for feature selection. The models were evaluated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).
Interventions
None listed
Sponsors
Ting Huang
Study design
Observational model
COHORT
Time perspective
RETROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
30 Years to 88 Years
Healthy volunteers
No
Inclusion criteria
* Patients with a single kidney tumor have complete imaging and clinical data * Contrast-enhanced CT scan within 30 days before surgery * No treatment was performed before CT examination
Exclusion criteria
* Patients with tumor recurrence * Obvious artifacts on CT images * The tumor is cystic * Multiple cysts on the affected kidney affect the delineation of renal parenchyma
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| predict the pathological grades of clear cell renal cell carcinoma (ccRCC) | 2019-2024 | AUC curve |
Countries
China
Outcome results
None listed