Deep Learning Model to Predict the Recurrence of Stage IA Invasive Lung Adenocarcinoma After Sub-lobar Resection

Status

Completed

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT06659601

Acronym

DL-Rec-ILA

Enrollment

Registered

2024-10-26

Start date

2023-06-01

Completion date

2024-10-24

Last updated

2024-10-26

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Focus on Developing a Deep Learning Model to Predict the Recurrence Risk of Stage IA Invasive Lung Adenocarcinoma After Sub-lobar Resection

Keywords

Lung cancer, Sub-lobar resection, Computed tomography, Deep learning

Brief summary

This study aims to develop a deep learning model based on noncontrast CT images to predict the recurrence risk of stage IA invasive lung adenocarcinoma after sub-lobar resection，which can serve as potential tool to assist thoracic surgeons in making optimal treatment decisions.The study will use existing CT data to train and validate the model, without requiring any additional intervention for the participants.

Detailed description

This study is designed to develop a deep learning model to predict the recurrence risk of stage IA invasive lung adenocarcinoma after sub-lobar resection using noncontrast CT images. The best indications for sub-lobar resection in patients with early-stage LADC are still debated, making surgical method selection somewhat difficult. The deep learning model can noninvasively and objectively predict the recurrence risk of patients with stage IA ILADC following sub-lobectomy and are helpful in predicting prognosis of patients with stage IA ILADC after sub-lobectomy and can facilitate the choosing of the optimal surgery mode of these patients. The study will utilize retrospective data from patients with stage IA invasive lung adenocarcinoma after sub-lobar resection . Noncontrast CT images will be collected at admission and used as inputs for the deep learning model. The model will be trained using convolutional neural networks (CNN) to identify patterns associated with recurrence. In addition to model development, the study will also evaluate the model's performance on a separate validation cohort to assess generalizability. Statistical analyses will include performance metrics such as area under the receiver operating characteristic (ROC) curve (AUC) and precision-recall curve. This study aims to provide a valuable tool for clinicians to make timely decisions in choosing the optimal therapeutic approach.

Interventions

None listed

Study design

Observational model

COHORT

Time perspective

RETROSPECTIVE

Eligibility

Sex/Gender

ALL

Healthy volunteers

Inclusion criteria

(i) pathological confirmation of LADC; (ii) undergoing sub-lobar resection (wedge resection or segmentectomy); (iii) CT scanning prior to surgery; (iv) pathological staging of IA; and (v) complete clinical and follow-up data. \-

Exclusion criteria

(i) multiple primary LADC; and (ii) other pulmonary lesions that might interfere with the morphological assessment of tumors. \-

Design outcomes

Primary

Measure	Time frame	Description
Recurrence Prediction Accuracy	October 2024	The primary outcome measure is the accuracy of the 3D deep learning model in predicting the recurrence of stage IA invasive lung adenocarcinoma after sub-lobar resection. Accuracy will be evaluated by comparing the model's predictions with actual patient outcomes using metrics such as sensitivity, specificity, and area under the ROC curve (AUC).

Countries

China

Outcome results