Penalty weight tuning in high dose rate brachytherapy using multi-objective Bayesian optimization.

Objective.Treatment plan optimization in high dose rate brachytherapy often requires manual fine-tuning of penalty weights for each objective, which can be time-consuming and dependent on the planner's experience. To automate this process, this study used a multi-criteria approach called multi-objective Bayesian optimization with q-noisy expected hypervolume improvement as its acquisition function (MOBO-qNEHVI).Approach.The treatment plans of 13 prostate cancer patients were retrospectively imported to a research treatment planning system, RapidBrachyMTPS, where fast mixed integer optimization (FMIO) performs dwell time optimization given a set of penalty weights to deliver 15 Gy to the target volume. MOBO-qNEHVI was used to find patient-specific Pareto optimal penalty weight vectors that yield clinically acceptable dose volume histogram metrics. The relationship between the number of MOBO-qNEHVI iterations and the number of clinically acceptable plans per patient (acceptance rate) was investigated. The performance time was obtained for various parameter configurations.Main results.MOBO-qNEHVI found clinically acceptable treatment plans for all patients. With increasing the number of MOBO-qNEHVI iterations, the acceptance rate grew logarithmically while the performance time grew exponentially. Fixing the penalty weight of the tumour volume to maximum value, adding the target dose as a parameter, initiating MOBO-qNEHVI with 25 parallel sampling of FMIO, and running 6 MOBO-qNEHVI iterations found solutions that delivered 15 Gy to the hottest 95% of the clinical target volume while respecting the dose constraints to the organs at risk. The average acceptance rate for each patient was 89.74% ± 8.11%, and performance time was 66.6 ± 12.6 s. The initiation took 22.47 ± 7.57 s, and each iteration took 7.35 ± 2.45 s to find one Pareto solution.Significance.MOBO-qNEHVI combined with FMIO can automatically explore the trade-offs between treatment plan objectives in a patient specific manner within a minute. This approach can reduce the dependency of plan quality on planner's experience and reduce dose to the organs at risk.

Personalized inference for neurostimulation with meta-learning: a case study of vagus nerve stimulation.

Objective. Neurostimulation is emerging as treatment for several diseases of the brain and peripheral organs. Due to variability arising from placement of stimulation devices, underlying neuroanatomy and physiological responses to stimulation, it is essential that neurostimulation protocols are personalized to maximize efficacy and safety. Building such personalized protocols would benefit from accumulated information in increasingly large datasets of other individuals' responses.Approach. To address that need, we propose a meta-learning family of algorithms to conduct few-shot optimization of key fitting parameters of physiological and neural responses in new individuals. While our method is agnostic to neurostimulation setting, here we demonstrate its effectiveness on the problem of physiological modeling of fiber recruitment during vagus nerve stimulation (VNS). Using data from acute VNS experiments, the mapping between amplitudes of stimulus-evoked compound action potentials (eCAPs) and physiological responses, such as heart rate and breathing interval modulation, is inferred.Main results. Using additional synthetic data sets to complement experimental results, we demonstrate that our meta-learning framework is capable of directly modeling the physiology-eCAP relationship for individual subjects with much fewer individually queried data points than standard methods.Significance. Our meta-learning framework is general and can be adapted to many input-response neurostimulation mapping problems. Moreover, this method leverages information from growing data sets of past patients, as a treatment is deployed. It can also be combined with several model types, including regression, Gaussian processes with Bayesian optimization, and beyond.

A Unified, Scalable Framework for Neural Population Decoding.

Our ability to use deep learning approaches to decipher neural activity would likely benefit from greater scale, in terms of both model size and datasets. However, the integration of many neural recordings into one unified model is challenging, as each recording contains the activity of different neurons from different individual animals. In this paper, we introduce a training framework and architecture designed to model the population dynamics of neural activity across diverse, large-scale neural recordings. Our method first tokenizes individual spikes within the dataset to build an efficient representation of neural events that captures the fine temporal structure of neural activity. We then employ cross-attention and a PerceiverIO backbone to further construct a latent tokenization of neural population activities. Utilizing this architecture and training framework, we construct a large-scale multi-session model trained on large datasets from seven nonhuman primates, spanning over 158 different sessions of recording from over 27,373 neural units and over 100 hours of recordings. In a number of different tasks, we demonstrate that our pretrained model can be rapidly adapted to new, unseen sessions with unspecified neuron correspondence, enabling few-shot performance with minimal labels. This work presents a powerful new approach for building deep learning tools to analyze neural data and stakes out a clear path to training at scale.

Approaching automated applicator digitization from a new angle: Using sagittal images to improve deep learning accuracy and robustness in high-dose-rate prostate brachytherapy.

PURPOSE: To automate the segmentation of treatment applicators on computed tomography (CT) images for high-dose-rate (HDR) brachytherapy prostate patients implanted with titanium needles with the goals of improving plan quality and reducing the patient's time under anesthesia. METHODS: The investigation was performed using 57 retrospective, interstitial prostate treatments randomly assigned to training (n = 27), validation (n = 10), and testing (n = 20). Unique to this work, the CT image set was reformatted into 2D sagittal slices instead of the default axial orientation. A deep learning-based segmentation was performed using a 2D U-Net architecture followed by a density-based linkage clustering algorithm to classify individual catheters in 3D. Potential confounders, such as gold seeds and conjoined applicators with intersecting needle geometries, were corrected using a customized polynomial fitting algorithm. The geometric agreement of the automated digitization was evaluated against the clinically treated manual digitization to measure tip and shaft errors in the reconstruction. RESULTS: The proposed algorithm achieved tip and shaft agreements of -0.1 ± 0.6 mm (range -1.8 mm to 1.4 mm) and 0.13 ± 0.09 mm (maximum 0.96 mm), respectively on a data set with 20 patients and 353 total needles. Our method was able to separate all intersecting applicators reliably. The time to generate the automated applicator digitization averaged approximately 1 min. CONCLUSIONS: Using sagittal instead of axial images for 2D segmentation of interstitial brachytherapy applicators produced submillimeter agreement with manual segmentation. The automated digitization of interstitial applicators in prostate brachytherapy has the potential to improve quality and consistency while reducing the patient's time under anesthesia.

Conditional generation of medical images via disentangled adversarial inference.

Synthetic medical image generation has a huge potential for improving healthcare through many applications, from data augmentation for training machine learning systems to preserving patient privacy. Conditional Adversarial Generative Networks (cGANs) use a conditioning factor to generate images and have shown great success in recent years. Intuitively, the information in an image can be divided into two parts: 1) content which is presented through the conditioning vector and 2) style which is the undiscovered information missing from the conditioning vector. Current practices in using cGANs for medical image generation, only use a single variable for image generation (i.e., content) and therefore, do not provide much flexibility nor control over the generated image. In this work we propose DRAI-a dual adversarial inference framework with augmented disentanglement constraints-to learn from the image itself, disentangled representations of style and content, and use this information to impose control over the generation process. In this framework, style is learned in a fully unsupervised manner, while content is learned through both supervised learning (using the conditioning vector) and unsupervised learning (with the inference mechanism). We undergo two novel regularization steps to ensure content-style disentanglement. First, we minimize the shared information between content and style by introducing a novel application of the gradient reverse layer (GRL); second, we introduce a self-supervised regularization method to further separate information in the content and style variables. For evaluation, we consider two types of baselines: single latent variable models that infer a single variable, and double latent variable models that infer two variables (style and content). We conduct extensive qualitative and quantitative assessments on two publicly available medical imaging datasets (LIDC and HAM10000) and test for conditional image generation, image retrieval and style-content disentanglement. We show that in general, two latent variable models achieve better performance and give more control over the generated image. We also show that our proposed model (DRAI) achieves the best disentanglement score and has the best overall performance.

RapidBrachyDL: Rapid Radiation Dose Calculations in Brachytherapy Via Deep Learning.

PURPOSE: Detailed and accurate absorbed dose calculations from radiation interactions with the human body can be obtained with the Monte Carlo (MC) method. However, the MC method can be slow for use in the time-sensitive clinical workflow. The aim of this study was to provide a solution to the accuracy-time trade-off for 192Ir-based high-dose-rate brachytherapy by using deep learning. METHODS AND MATERIALS: RapidBrachyDL, a 3-dimensional deep convolutional neural network (CNN) model, is proposed to predict dose distributions calculated with the MC method given a patient's computed tomography images, contours of clinical target volume (CTV) and organs at risk, and treatment plan. Sixty-one patients with prostate cancer and 10 patients with cervical cancer were included in this study, with data from 47 patients with prostate cancer being used to train the model. RESULTS: Compared with ground truth MC simulations, the predicted dose distributions by RapidBrachyDL showed a consistent shape in the dose-volume histograms (DVHs); comparable DVH dosimetric indices including 0.73% difference for prostate CTV D90, 1.1% for rectum D2cc, 1.45% for urethra D0.1cc, and 1.05% for bladder D2cc; and substantially smaller prediction time, acceleration by a factor of 300. RapidBrachyDL also demonstrated good generalization to cervical data with 1.73%, 2.46%, 1.68%, and 1.74% difference for CTV D90, rectum D2cc, sigmoid D2cc, and bladder D2cc, respectively, which was unseen during the training. CONCLUSION: Deep CNN-based dose estimation is a promising method for patient-specific brachytherapy dosimetry. Desired radiation quantities can be obtained with accuracies arbitrarily close to those of the source MC algorithm, but with much faster computation times. The idea behind deep CNN-based dose estimation can be safely extended to other radiation sources and tumor sites by following a similar training process.