Prediction of clinical events by deep learning based on CT scan and anapath slide
Principal Investigator
Name
Tanguy Perennec
Degrees
MD
Institution
Institut de Cancérologie de l'Ouest
Position Title
Medical Doctor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-1157
Initial CDAS Request Approval
Nov 13, 2023
Title
Prediction of clinical events by deep learning based on CT scan and anapath slide
Summary
Radiotherapy uses images from several modalities for treatment planning. We are conducting a study to identify predictive models for clinical events (death, recurrence, cardiological events) on these different modalities, with the idea of being able to secondarily adapt treatments to the patient (a patient at cardiovascular risk, for example, should receive less radiation dose). The idea is to be as precise as possible in localizing the risks, so as to be vigilant when planning treatment.
We will analyse as many image as possible (since cardiovascular risk can be predicted on lung cancer patients and healthy patients). Also, training on healthy patients will prevent considering all nodule as cancer.
We will analyse as many image as possible (since cardiovascular risk can be predicted on lung cancer patients and healthy patients). Also, training on healthy patients will prevent considering all nodule as cancer.
Aims
- Predict overall survival on LDCT
- Predict specific survival
- Predict medical complications on LDCT
- Explore the results using GradCam
Collaborators
Loig Vaugier, MD, Institut de Cancerologie de l'Ouest
Alexandra Moignier, PhD, Institut de Cancerologie de l'Ouest