Dataset Information

Evaluating the utility of deep learning for predicting therapeutic response in diabetic eye disease.

ABSTRACT:

Purpose

Deep learning (DL) is a technique explored within ophthalmology that requires large datasets to distinguish feature representations with high diagnostic performance. There is a need for developing DL approaches to predict therapeutic response, but completed clinical trial datasets are limited in size. Predicting treatment response is more complex than disease diagnosis, where hallmarks of treatment response are subtle. This study seeks to understand the utility of DL for clinical problems in ophthalmology such as predicting treatment response and where large sample sizes for model training are not available.

Materials and methods

Four DL architectures were trained using cross-validated transfer learning to classify ultra-widefield angiograms (UWFA) and fluid-compartmentalized optical coherence tomography (OCT) images from a completed clinical trial (PERMEATE) dataset (n=29) as tolerating or requiring extended interval Anti-VEGF dosing. UWFA images (n=217) from the Anti-VEGF study were divided into five increasingly larger subsets to evaluate the influence of dataset size on performance. Class activation maps (CAMs) were generated to identify regions of model attention.

Results

The best performing DL model had a mean AUC of 0.507 ± 0.042 on UWFA images, and highest observed AUC of 0.503 for fluid-compartmentalized OCT images. DL had a best performing AUC of 0.634 when dataset size was incrementally increased. Resulting CAMs show inconsistent regions of interest.

Conclusions

This study demonstrated the limitations of DL for predicting therapeutic response when large datasets were not available for model training. Our findings suggest the need for hand-crafted approaches for complex and data scarce prediction problems in ophthalmology.

SUBMITTER: Dong V

PROVIDER: S-EPMC9894083 | biostudies-literature | 2022

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Evaluating the utility of deep learning for predicting therapeutic response in diabetic eye disease.

Dong Vincent V Sevgi Duriye Damla DD Kar Sudeshna Sil SS Srivastava Sunil K SK Ehlers Justis P JP Madabhushi Anant A

Frontiers in ophthalmology 20220812

<h4>Purpose</h4>Deep learning (DL) is a technique explored within ophthalmology that requires large datasets to distinguish feature representations with high diagnostic performance. There is a need for developing DL approaches to predict therapeutic response, but completed clinical trial datasets are limited in size. Predicting treatment response is more complex than disease diagnosis, where hallmarks of treatment response are subtle. This study seeks to understand the utility of DL for clinical ...[more]

PMID: 36744216

Similar Datasets

Project description:The complex, unpredictable nature of pathogen occurrence has required substantial efforts to accurately predict infectious diseases (IDs). With rising popularity of Machine Learning (ML) and Deep Learning (DL) techniques combined with their unique ability to uncover connections between large amounts of diverse data, we conducted a PRISMA systematic review to investigate advances in ID prediction for human and animal diseases using ML and DL. This review included the type of IDs modeled, ML and DL techniques utilized, geographical distribution, prediction tasks performed, input features utilized, spatial and temporal scales, error metrics used, computational efficiency, uncertainty quantification, and missing data handling methods. Among 237 relevant articles published between January 2001 and May 2021, highly contagious diseases in humans were most often represented, including COVID-19 (37.1%), influenza/influenza-like illnesses (9.3%), dengue (8.9%), and malaria (5.1%). Out of 37 diseases identified, 51.4% were zoonotic, 37.8% were human-only, and 8.1% were animal-only, with only 1.6% economically significant, non-zoonotic livestock diseases. Despite the number of zoonoses, 86.5% of articles modeled humans whereas only a few articles (5.1%) contained more than one host species. Eastern Asia (32.5%), North America (17.7%), and Southern Asia (13.1%) were the most represented locations. Frequent approaches included tree-based ML (38.4%) and feed-forward neural networks (26.6%). Articles predicted temporal incidence (66.7%), disease risk (38.0%), and/or spatial movement (31.2%). Less than 10% of studies addressed uncertainty quantification, computational efficiency, and missing data, which are essential to operational use and deployment. This study highlights trends and gaps in ML and DL for ID prediction, providing guidelines for future works to better support biopreparedness and response. To fully utilize ML and DL for improved ID forecasting, models should include the full disease ecology in a One-Health context, important food and agricultural diseases, underrepresented hotspots, and important metrics required for operational deployment.

Project description:BackgroundUnfractionated heparin (UFH) is an anticoagulant drug that is considered a high-risk medication because an excessive dose can cause bleeding, whereas an insufficient dose can lead to a recurrent embolic event. Therapeutic response to the initiation of intravenous UFH is monitored using activated partial thromboplastin time (aPTT) as a measure of blood clotting time. Clinicians iteratively adjust the dose of UFH toward a target, indication-defined therapeutic aPTT range using nomograms, but this process can be imprecise and can take ≥36 hours to achieve the target range. Thus, a more efficient approach is required.ObjectiveIn this study, we aimed to develop and validate a machine learning (ML) algorithm to predict aPTT within 12 hours after a specified bolus and maintenance dose of UFH.MethodsThis was a retrospective cohort study of 3019 patient episodes of care from January 2017 to August 2020 using data collected from electronic health records of 5 hospitals in Queensland, Australia. Data from 4 hospitals were used to build and test ensemble models using cross-validation, whereas data from the fifth hospital were used for external validation. We built 2 ML models: a regression model to predict the aPTT value after a UFH bolus dose and a multiclass model to predict the aPTT, classified as subtherapeutic (aPTT <70 seconds), therapeutic (aPTT 70-100 seconds), or supratherapeutic (aPTT >100 seconds). Modeling was performed using Driverless AI (H2O), an automated ML tool, and 17 different experiments were iteratively conducted to optimize model accuracy.ResultsIn predicting aPTT, the best performing model was an ensemble with 4x LightGBM models with a root mean square error of 31.35 (SD 1.37). In predicting the aPTT class using a repurposed data set, the best performing ensemble model achieved an accuracy of 0.599 (SD 0.0289) and an area under the receiver operating characteristic curve of 0.735. External validation yielded similar results: root mean square error of 30.52 (SD 1.29) for the aPTT prediction model, and accuracy of 0.568 (SD 0.0315) and area under the receiver operating characteristic curve of 0.724 for the aPTT multiclassification model.ConclusionsTo the best of our knowledge, this is the first ML model applied to intravenous UFH dosing that has been developed and externally validated in a multisite adult general medical and surgical inpatient setting. We present the processes of data collection, preparation, and feature engineering for replication.

Dataset Information

Evaluating the utility of deep learning for predicting therapeutic response in diabetic eye disease.

Purpose

Materials and methods

Results

Conclusions

Publications

Evaluating the utility of deep learning for predicting therapeutic response in diabetic eye disease.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets