This repository contains the PyTorch implementation for:
LONGL-Net: Temporal Correlation Structure Guided Deep Learning Model to Predict Longitudinal Age-related Macular Degeneration Severity (to appear in the Proceedings of the National Academy of Sciences (PNAS) Nexus)
Alireza Ganjdanesh, Jipeng Zhang, Emily Chew, Ying Ding, Wei Chen†, Heng Huang†
Abstract: Age-related Macular Degeneration (AMD) is the principal cause of blindness in developed countries, and its prevalence will increase to 288 million people in 2040. Therefore, automated grading and prediction methods can be highly beneficial for recognizing susceptible subjects to late-AMD and enabling clinicians to start preventive actions for them. Clinically, AMD severity is quantified by Color Fundus Photographs (CFP) of the retina, and many machine learning based methods are proposed for grading AMD severity. However, few models were developed to predict the longitudinal progression status, i.e., predicting future late-AMD risk based on the current CFP, which is more clinically interesting. In this paper, we propose a new deep learning based classification model (LONGL-Net) that can simultaneously grade the current CFP and predict the longitudinal outcome, i.e., whether the subject will be in late-AMD in the future time-point. We design a new temporal-correlation-structure-guided Generative Adversarial Network model that learns the interrelations of temporal changes in CFPs in consecutive time-points and provides interpretability for the classifier’s decisions by forecasting AMD symptoms in the future CFPs. We used about 30,000 CFP images from 4,628 participants in the Age-Related Eye Disease Study. Our classifier showed average 0.905 (95% CI: 0.886-0.922) AUC and 0.762 (95% CI: 0.733-0.792) accuracy on the 3-class classification problem of simultaneously grading current time-point’s AMD condition and predicting late AMD progression of subjects in the future time-point. We further validated our model on the UK-Biobank dataset, where our model showed average 0.905 accuracy and 0.797 sensitivity in grading 300 CFP images.
Please refer to Data directory for detailed data preparation steps.
- Python 3.7
- PyTorch + Torchvision
- Pytorch Lightning (for data preparation)
- Tensorflow
- Scikit-learn + Scikit-image
- PIL
main_classifier.py: train the classification model with different time gap values.main_evaluate_classifier.py: code for evaluating the trained classification model's checkpoints and obtaining their performance confidence interval.main_binary_classifier.py: train the a binary classifier to grade whether its input fundus image being in advanced AMD or not.main_UKB.py: evaluate trained classification model on the independent UK Biobank dataset.main_next_step_GAN.py: train the GAN model.main_evaluate_GAN.py: quantitive evaluation of the trained GAN models.main_longitudinal_image_prediction.py: code for longitudinal prediction of fundus images of the subjects using the trained GAN models' checkpoints.main_visualization.py: plotting the saliency maps of the outputs of the classification model.main_resize_images.py: code for preprocessing images in AREDS and UK Biobank datasets.
All of our trained checkpoint models are available here. Please download and put them in their corresponding folder in the checkpoints directory.
If you have any questions, please feel free to contact us through email (alireza.ganjdanesh@pitt.edu).