Deep Learning Model for Pure Solid Nodules Classification
- Conditions
- Lung Cancer
- Registration Number
- NCT05542992
- Lead Sponsor
- Chang Chen
- Brief Summary
The purpose of this study is to compare the predictive performance of a CT-based deep learning model for pure-solid nodules classification and compared with the tumor maximum standardized uptake value on PET in a multicenter prospective cohort.
- Detailed Description
Not available
Recruitment & Eligibility
- Status
- RECRUITING
- Sex
- All
- Target Recruitment
- 260
- Participants scheduled for surgery for radiological finding of pulmonary pure-solid lesions from the preoperative thin-section CT scans;
- The maximum short-axis diameter of lymph nodes less than 3 cm on CT scan;
- Age ranging from 18-75 years;
- definied pathological examination report available;
- Obtained written informed consent.
- Multiple lung lesions;
- Poor quality of CT images;
- Participants with incomplete clinical information;
- Participants who have received neoadjuvant therapy before initial CT evaluation.
Study & Design
- Study Type
- OBSERVATIONAL
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method AUC 2022.01-2023.12 Area under the curve of the receiver operating characteristic
- Secondary Outcome Measures
Name Time Method Specificity 2022.01-2023.12 Odds of detecting a negative test in a population judged disease-free (negative) by the gold standard
PPV 2022.01-2023.12 Positive predictive value
NPV 2022.01-2023.12 Negative predictive value
Accuracy 2022.01-2023.12 Ratio of the number of correctly classified samples to the total number of samples
sensitivity 2022.01-2023.12 The probability of detecting a positive test in the population with the gold standard for disease (positive)
Trial Locations
- Locations (5)
Shanghai Pulmonary Hospital
🇨🇳Yangpu, Shanghai, China
Lanzhou
🇨🇳China, Gansu, China
Zunyi
🇨🇳China, Guizhou, China
Nanchang
🇨🇳China, Jiangxi, China
Ningbo
🇨🇳China, Zhejiang, China