ABSTRACT
PURPOSE: To develop and evaluate an automated, portable algorithm to differentiate active corneal ulcers from healed scars using only external photographs. DESIGN: A convolutional neural network was trained and tested using photographs of corneal ulcers and scars. PARTICIPANTS: De-identified photographs of corneal ulcers were obtained from the Steroids for Corneal Ulcers Trial (SCUT), Mycotic Ulcer Treatment Trial (MUTT), and Byers Eye Institute at Stanford University. METHODS: Photographs of corneal ulcers (n = 1313) and scars (n = 1132) from the SCUT and MUTT were used to train a convolutional neural network (CNN). The CNN was tested on 2 different patient populations from eye clinics in India (n = 200) and the Byers Eye Institute at Stanford University (n = 101). Accuracy was evaluated against gold standard clinical classifications. Feature importances for the trained model were visualized using gradient-weighted class activation mapping. MAIN OUTCOME MEASURES: Accuracy of the CNN was assessed via F1 score. The area under the receiver operating characteristic (ROC) curve (AUC) was used to measure the precision-recall trade-off. RESULTS: The CNN correctly classified 115 of 123 active ulcers and 65 of 77 scars in patients with corneal ulcer from India (F1 score, 92.0% [95% confidence interval (CI), 88.2%-95.8%]; sensitivity, 93.5% [95% CI, 89.1%-97.9%]; specificity, 84.42% [95% CI, 79.42%-89.42%]; ROC: AUC, 0.9731). The CNN correctly classified 43 of 55 active ulcers and 42 of 46 scars in patients with corneal ulcers from Northern California (F1 score, 84.3% [95% CI, 77.2%-91.4%]; sensitivity, 78.2% [95% CI, 67.3%-89.1%]; specificity, 91.3% [95% CI, 85.8%-96.8%]; ROC: AUC, 0.9474). The CNN visualizations correlated with clinically relevant features such as corneal infiltrate, hypopyon, and conjunctival injection. CONCLUSIONS: The CNN classified corneal ulcers and scars with high accuracy and generalized to patient populations outside of its training data. The CNN focused on clinically relevant features when it made a diagnosis. The CNN demonstrated potential as an inexpensive diagnostic approach that may aid triage in communities with limited access to eye care.
Subject(s)
Cicatrix/diagnostic imaging , Corneal Ulcer/diagnostic imaging , Deep Learning , Eye Infections, Bacterial/diagnostic imaging , Eye Infections, Fungal/diagnostic imaging , Photography , Wound Healing/physiology , Algorithms , Area Under Curve , Cicatrix/physiopathology , Corneal Ulcer/classification , Corneal Ulcer/microbiology , Eye Infections, Bacterial/classification , Eye Infections, Bacterial/microbiology , Eye Infections, Fungal/classification , Eye Infections, Fungal/microbiology , False Positive Reactions , Humans , Predictive Value of Tests , ROC Curve , Retrospective Studies , Sensitivity and Specificity , Slit Lamp MicroscopyABSTRACT
COVID-19 forecasting models have been critical in guiding decision-making on surveillance testing, social distancing, and vaccination requirements. Beyond influencing public health policies, an accurate COVID-19 forecasting model can impact community spread by enabling employers and university leaders to adapt worksite policies and practices to contain or mitigate outbreaks. While many such models have been developed for COVID-19 forecasting at the national, state, county, or city level, only a few models have been developed for workplaces and universities. Furthermore, COVID-19 forecasting models have rarely been validated against real COVID-19 case data. Here we present the systematic parameter fitting and validation of an agent-based compartment model for the forecasting of daily COVID-19 cases in single-site workplaces and universities with real-world data. Our approaches include manual fitting, where initial model parameters are chosen based on historical data, and automated fitting, where parameters are chosen based on candidate case trajectory simulations that result in best fit to prevalence estimation data. We use a 14-day fitting window and validate our approaches on 7- and 14-day testing windows with real COVID-19 case data from one employer. Our manual and automated fitting approaches accurately predicted COVID-19 case trends and outperformed the baseline model (no parameter fitting) across multiple scenarios, including a rising case trajectory (RMSLE values: 2.627 for baseline, 0.562 for manual fitting, 0.399 for automated fitting) and a decreasing case trajectory (RMSLE values: 1.155 for baseline, 0.537 for manual fitting, 0.778 for automated fitting). Our COVID-19 case forecasting model allows decision-makers at workplaces and universities to proactively respond to case trend forecasts, mitigate outbreaks, and promote safety.
Subject(s)
COVID-19 , Humans , COVID-19/epidemiology , Universities , Models, Statistical , Disease Outbreaks/prevention & control , Forecasting , Public PolicyABSTRACT
Collaborative learning, which enables collaborative and decentralized training of deep neural networks at multiple institutions in a privacy-preserving manner, is rapidly emerging as a valuable technique in healthcare applications. However, its distributed nature often leads to significant heterogeneity in data distributions across institutions. In this paper, we present a novel generative replay strategy to address the challenge of data heterogeneity in collaborative learning methods. Different from traditional methods that directly aggregating the model parameters, we leverage generative adversarial learning to aggregate the knowledge from all the local institutions. Specifically, instead of directly training a model for task performance, we develop a novel dual model architecture: a primary model learns the desired task, and an auxiliary "generative replay model" allows aggregating knowledge from the heterogenous clients. The auxiliary model is then broadcasted to the central sever, to regulate the training of primary model with an unbiased target distribution. Experimental results demonstrate the capability of the proposed method in handling heterogeneous data across institutions. On highly heterogeneous data partitions, our model achieves â¼4.88% improvement in the prediction accuracy on a diabetic retinopathy classification dataset, and â¼49.8% reduction of mean absolution value on a Bone Age prediction dataset, respectively, compared to the state-of-the art collaborative learning methods.
Subject(s)
Interdisciplinary Placement , Diagnostic Imaging , Humans , Neural Networks, Computer , RadiographyABSTRACT
Zero-shot learning (ZSL) is one of the most promising avenues of annotation-efficient machine learning. In the era of deep learning, ZSL techniques have achieved unprecedented success. However, the developments of ZSL methods have taken place mostly for natural images. ZSL for medical images has remained largely unexplored. We design a novel strategy for generalized zero-shot diagnosis of chest radiographs. In doing so, we leverage the potential of multi-view semantic embedding, a useful yet less-explored direction for ZSL. Our design also incorporates a self-training phase to tackle the problem of noisy labels alongside improving the performance for classes not seen during training. Through rigorous experiments, we show that our model trained on one dataset can produce consistent performance across test datasets from different sources including those with very different quality. Comparisons with a number of state-of-the-art techniques show the superiority of the proposed method for generalized zero-shot chest x-ray diagnosis.
Subject(s)
Machine Learning , Semantics , Phenotype , Radiography , X-RaysABSTRACT
OBJECTIVES: Sharing patient data across institutions to train generalizable deep learning models is challenging due to regulatory and technical hurdles. Distributed learning, where model weights are shared instead of patient data, presents an attractive alternative. Cyclical weight transfer (CWT) has recently been demonstrated as an effective distributed learning method for medical imaging with homogeneous data across institutions. In this study, we optimize CWT to overcome performance losses from variability in training sample sizes and label distributions across institutions. MATERIALS AND METHODS: Optimizations included proportional local training iterations, cyclical learning rate, locally weighted minibatch sampling, and cyclically weighted loss. We evaluated our optimizations on simulated distributed diabetic retinopathy detection and chest radiograph classification. RESULTS: Proportional local training iteration mitigated performance losses from sample size variability, achieving 98.6% of the accuracy attained by centrally hosting in the diabetic retinopathy dataset split with highest sample size variance across institutions. Locally weighted minibatch sampling and cyclically weighted loss both mitigated performance losses from label distribution variability, achieving 98.6% and 99.1%, respectively, of the accuracy attained by centrally hosting in the diabetic retinopathy dataset split with highest label distribution variability across institutions. DISCUSSION: Our optimizations to CWT improve its capability of handling data variability across institutions. Compared to CWT without optimizations, CWT with optimizations achieved performance significantly closer to performance from centrally hosting. CONCLUSION: Our work is the first to identify and address challenges of sample size and label distribution variability in simulated distributed deep learning for medical imaging. Future work is needed to address other sources of real-world data variability.
Subject(s)
Deep Learning , Diagnostic Imaging , Health Information Exchange , Computer Communication Networks , Datasets as Topic , Diabetic Retinopathy/diagnostic imaging , Humans , Pattern Recognition, Automated , Radiography, ThoracicABSTRACT
Objective: Deep learning has become a promising approach for automated support for clinical diagnosis. When medical data samples are limited, collaboration among multiple institutions is necessary to achieve high algorithm performance. However, sharing patient data often has limitations due to technical, legal, or ethical concerns. In this study, we propose methods of distributing deep learning models as an attractive alternative to sharing patient data. Methods: We simulate the distribution of deep learning models across 4 institutions using various training heuristics and compare the results with a deep learning model trained on centrally hosted patient data. The training heuristics investigated include ensembling single institution models, single weight transfer, and cyclical weight transfer. We evaluated these approaches for image classification in 3 independent image collections (retinal fundus photos, mammography, and ImageNet). Results: We find that cyclical weight transfer resulted in a performance that was comparable to that of centrally hosted patient data. We also found that there is an improvement in the performance of cyclical weight transfer heuristic with a high frequency of weight transfer. Conclusions: We show that distributing deep learning models is an effective alternative to sharing patient data. This finding has implications for any collaborative deep learning study.