Active Learning in Brain Tumor Segmentation with Uncertainty Sampling and Annotation Redundancy Restriction.

Deep learning models have demonstrated great potential in medical imaging but are limited by the expensive, large volume of annotations required. To address this, we compared different active learning strategies by training models on subsets of the most informative images using real-world clinical datasets for brain tumor segmentation and proposing a framework that minimizes the data needed while maintaining performance. Then, 638 multi-institutional brain tumor magnetic resonance imaging scans were used to train three-dimensional U-net models and compare active learning strategies. Uncertainty estimation techniques including Bayesian estimation with dropout, bootstrapping, and margins sampling were compared to random query. Strategies to avoid annotating similar images were also considered. We determined the minimum data necessary to achieve performance equivalent to the model trained on the full dataset (α = 0.05). Bayesian approximation with dropout at training and testing showed results equivalent to that of the full data model (target) with around 30% of the training data needed by random query to achieve target performance (p = 0.018). Annotation redundancy restriction techniques can reduce the training data needed by random query to achieve target performance by 20%. We investigated various active learning strategies to minimize the annotation burden for three-dimensional brain tumor segmentation. Dropout uncertainty estimation achieved target performance with the least annotated data.

Predicting Visual Acuity Responses to Anti-VEGF Treatment in the Comparison of Age-related Macular Degeneration Treatments Trials Using Machine Learning.

PURPOSE: To evaluate multiple machine learning (ML) models for predicting 2-year visual acuity (VA) responses to anti-vascular endothelial growth factor (anti-VEGF) treatment in the Comparison of Age-related Macular Degeneration (AMD) Treatments Trials (CATT) for patients with neovascular AMD (nAMD). DESIGN: Secondary analysis of public data from a randomized clinical trial. PARTICIPANTS: A total of 1029 CATT participants who completed 2 years of follow-up with untreated active nAMD and baseline VA between 20/25 and 20/320 in the study eye. METHODS: Five ML models (support vector machine, random forest, extreme gradient boosting, multilayer perceptron neural network, and lasso) were applied to clinical and image data from baseline and weeks 4, 8, and 12 for predicting 4 VA outcomes (≥ 15-letter VA gain, ≥ 15-letter VA loss, VA change from baseline, and actual VA) at 2 years. The CATT data from 1029 participants were randomly split for training (n = 717), from which the models were trained using 10-fold cross-validation, and for final validation on a test data set (n = 312). MAIN OUTCOME MEASURES: Performances of ML models were assessed by R2 and mean absolute error (MAE) for predicting VA change from baseline and actual VA at 2 years, by the area under the receiver operating characteristic curve (AUC) for predicting ≥ 15-letter VA gain and loss from baseline. RESULTS: Using training data up to week 12, the ML models from cross-validation achieved mean R2 of 0.24 to 0.29 (MAE = 9.1-9.8 letters) for predicting VA change and 0.37 to 0.41 (MAE = 9.3-10.2 letters) for predicting actual VA at 2 years. The mean AUCs for predicting ≥ 15-letter VA gain and loss at 2 years was 0.84 to 0.85 and 0.58 to 0.73, respectively. In final validation on the test data set up to week 12, the models had an R2 of 0.33 to 0.38 (MAE = 8.9-9.9 letters) for predicting VA change, an R2 of 0.37 to 0.45 (MAE = 8.8-10.2 letters) for predicting actual VA at 2 years, and AUCs of 0.85 to 0.87 and 0.67 to 0.79 for predicting ≥ 15-letter VA gain and loss, respectively. CONCLUSIONS: Machine learning models have the potential to predict 2-year VA response to anti-VEGF treatment using clinical and imaging features from the loading dose phase, which can aid in decision-making around treatment protocols for patients with nAMD. FINANCIAL DISCLOSURE(S): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Brain Tumor Segmentation for Multi-Modal MRI with Missing Information.

Deep convolutional neural networks (DCNNs) have shown promise in brain tumor segmentation from multi-modal MRI sequences, accommodating heterogeneity in tumor shape and appearance. The fusion of multiple MRI sequences allows networks to explore complementary tumor information for segmentation. However, developing a network that maintains clinical relevance in situations where certain MRI sequence(s) might be unavailable or unusual poses a significant challenge. While one solution is to train multiple models with different MRI sequence combinations, it is impractical to train every model from all possible sequence combinations. In this paper, we propose a DCNN-based brain tumor segmentation framework incorporating a novel sequence dropout technique in which networks are trained to be robust to missing MRI sequences while employing all other available sequences. Experiments were performed on the RSNA-ASNR-MICCAI BraTS 2021 Challenge dataset. When all MRI sequences were available, there were no significant differences in performance of the model with and without dropout for enhanced tumor (ET), tumor (TC), and whole tumor (WT) (p-values 1.000, 1.000, 0.799, respectively), demonstrating that the addition of dropout improves robustness without hindering overall performance. When key sequences were unavailable, the network with sequence dropout performed significantly better. For example, when tested on only T1, T2, and FLAIR sequences together, DSC for ET, TC, and WT increased from 0.143 to 0.486, 0.431 to 0.680, and 0.854 to 0.901, respectively. Sequence dropout represents a relatively simple yet effective approach for brain tumor segmentation with missing MRI sequences.

Evaluation of Multiple Machine Learning Models for Predicting Number of Anti-VEGF Injections in the Comparison of AMD Treatment Trials (CATT).

Purpose: To apply machine learning models for predicting the number of pro re nata (PRN) injections of antivascular endothelial growth factor (anti-VEGF) for neovascular age-related macular degeneration (nAMD) in two years in the Comparison of AMD (age-related macular degeneration) Treatments Trials. Methods: The data from 493 eligible participants randomized to PRN treatment of ranibizumab or bevacizumab were used for training (n = 393) machine learning models including support-vector machine (SVM), random forest, and extreme gradient boosting (XGBoost) models. Model performances of prediction using clinical and image data from baseline, weeks 4, 8, and 12 were evaluated by the area under the receiver operating characteristic curve (AUC) for predicting few (≤8) or many (≥19) injections, by R2 and mean absolute error (MAE) for predicting the total number of injections in two years. The best model was selected for final validation on a test dataset (n = 100). Results: Using training data up to week 12, the models achieved AUCs of 0.79-0.82 and 0.79-0.81 for predicting few and many injections, respectively, with R2 of 0.34-0.36 (MAE = 4.45-4.58 injections) for predicting total injections in two years from cross-validation. In final validation on the test dataset, the SVM model had AUCs of 0.77 and 0.82 for predicting few and many injections, respectively, with R2 of 0.44 (MAE = 3.92 injections). Important features included fluid in optical coherence tomography, lesion characteristics, and treatment trajectory in the first three months. Conclusions: Machine learning models using loading dose phase data have the potential to predict two-year anti-VEGF demand for nAMD and quantify feature importance for these predictions. Translational Relevance: Prediction of anti-VEGF injections using machine learning models from readily available data, after further validation on independent datasets, has the potential to help optimize treatment protocols and outcomes for nAMD patients in an individualized manner.