RESUMEN
In this work, we study various hybrid models of entropy-based and representativeness sampling techniques in the context of active learning in medical segmentation, in particular examining the role of UMAP (Uniform Manifold Approximation and Projection) as a technique for capturing representativeness. Although UMAP has been shown viable as a general purpose dimension reduction method in diverse areas, its role in deep learning-based medical segmentation has yet been extensively explored. Using the cardiac and prostate datasets in the Medical Segmentation Decathlon for validation, we found that a novel hybrid combination of Entropy-UMAP sampling technique achieved a statistically significant Dice score advantage over the random baseline (3.2% for cardiac, 4.5% for prostate), and attained the highest Dice coefficient among the spectrum of 10 distinct active learning methodologies we examined. This provides preliminary evidence that there is an interesting synergy between entropy-based and UMAP methods when the former precedes the latter in a hybrid model of active learning.
Asunto(s)
Entropía , Humanos , Masculino , Aprendizaje Profundo , Próstata/diagnóstico por imagen , Procesamiento de Imagen Asistido por Computador/métodos , Aprendizaje Automático Supervisado , CorazónRESUMEN
â¢AI dose predictor was fully integrated with treatment planning system and used as a physicain decision support tool to improve uniformity of practice.â¢Model was trained based on our standard of practice, but implemented at the time of expansion with 3 new physicians join the practice.â¢Phase 1 retrospective evaluation demonstrated the non-uniform practice among 3 MDs and only 52.9% frequency planner can achieve physicians' directives.â¢Significant improvement in practice uniformity of practice was observed after utilizing AI as DST and 80.4% frequency clinical plan can achieve AI-guided physician directives.
RESUMEN
Purpose: To present a concept called artificial intelligence-assisted contour editing (AIACE) and demonstrate its feasibility. Materials and Methods: The conceptual workflow of AIACE is as follows: Given an initial contour that requires clinician editing, the clinician indicates where large editing is needed, and a trained deep learning model uses this input to update the contour. This process repeats until a clinically acceptable contour is achieved. In this retrospective, proof-of-concept study, the authors demonstrated the concept on two-dimensional (2D) axial CT images from three head-and-neck cancer datasets by simulating the interaction with the AIACE model to mimic the clinical environment. The input at each iteration was one mouse click on the desired location of the contour segment. Model performance is quantified with the Dice similarity coefficient (DSC) and 95th percentile of Hausdorff distance (HD95) based on three datasets with sample sizes of 10, 28, and 20 patients. Results: The average DSCs and HD95 values of the automatically generated initial contours were 0.82 and 4.3 mm, 0.73 and 5.6 mm, and 0.67 and 11.4 mm for the three datasets, which were improved to 0.91 and 2.1 mm, 0.86 and 2.5 mm, and 0.86 and 3.3 mm, respectively, with three mouse clicks. Each deep learning-based contour update required about 20 msec. Conclusion: The authors proposed the newly developed AIACE concept, which uses deep learning models to assist clinicians in editing contours efficiently and effectively, and demonstrated its feasibility by using 2D axial CT images from three head-and-neck cancer datasets.Keywords: Segmentation, Convolutional Neural Network (CNN), CT, Deep Learning Algorithms Supplemental material is available for this article. © RSNA, 2022.
RESUMEN
Atrial fibrillation (AF) is a common cardiac arrhythmia and affects one to two percent of the population. In this work, we leverage the three-dimensional atrial endocardial unipolar/bipolar voltage map to predict the AF type and recurrence of AF in 1 year. This problem is challenging for two reasons: (1) the unipolar/bipolar voltages are collected at different locations on the endocardium and the shapes of the endocardium vary widely in different patients, and thus the unipolar/bipolar voltage maps need aligning to the same coordinate; (2) the collected dataset size is very limited. To address these issues, we exploit a pretrained 3D point cloud registration approach and finetune it on left atrial voltage maps to learn the geometric feature and align all voltage maps into the same coordinate. After alignment, we feed the unipolar/bipolar voltages from the registered points into a multilayer perceptron (MLP) classifier to predict whether patients have paroxysmal or persistent AF, and the risk of recurrence of AF in 1 year for patients in sinus rhythm. The experiment shows our method classifies the type and recurrence of AF effectively.
Asunto(s)
Fibrilación Atrial , Ablación por Catéter , Fibrilación Atrial/cirugía , Ablación por Catéter/métodos , Atrios Cardíacos/cirugía , Humanos , Redes Neurales de la Computación , Estudios Retrospectivos , Resultado del TratamientoRESUMEN
PURPOSE: Typically, the current dose prediction models are limited to small amounts of data and require retraining for a specific site, often leading to suboptimal performance. We propose a site-agnostic, three-dimensional dose distribution prediction model using deep learning that can leverage data from any treatment site, thus increasing the total data available to train the model. Applying our proposed model to a new target treatment site requires only a brief fine-tuning of the model to the new data and involves no modifications to the model input channels or its parameters. Thus, it can be efficiently adapted to a different treatment site, even with a small training dataset. METHODS: This study uses two separate datasets/treatment sites: data from patients with prostate cancer treated with intensity-modulated radiation therapy (source data), and data from patients with head-and-neck cancer treated with volumetric-modulated arc therapy (target data). We first developed a source model with 3D UNet architecture, trained from random initial weights on the source data. We evaluated the performance of this model on the source data. We then studied the generalizability of the model to the new target dataset via transfer learning. To do this, we built three more models, all with the same 3D UNet architecture: target model, adapted model, and combined model. The source and target models were trained on the source and target data from random initial weights, respectively. The adapted model fine-tuned the source model to the target domain by using the target data. Finally, the combined model was trained from random initial weights on a combined data pool consisting of both target and source datasets. We tested all four models on the target dataset and evaluated quantitative dose-volume histogram metrics for the planning target volume (PTV) and organs at risk (OARs). RESULTS: When tested on the source treatment site, the source model accurately predicted the dose distributions with average (mean, max) absolute dose errors of (0.32%±0.14, 2.37%±0.93) (PTV) relative to the prescription dose, and highest mean dose error of 1.68%±0.76, and highest max dose error of 5.47%± 3.31 for femoral head right. The error in PTV dose coverage prediction is 3.21%±1.51 for D98 , 3.04%±1.69 for D95 , and 1.83%±1.01 for D02 . Averaging across all OARs, the source model predicted the OAR mean dose within 1.38% and the OAR max dose within 3.64%. For the target treatment site, the target model average (mean, max) absolute dose errors relative to the prescription dose for the PTV were (1.08%±0.95, 2.90%±1.35). Left cochlea had the highest mean and max dose errors of 5.37%±5.82 and 8.33%±8.88, respectively. The errors in PTV dose coverage prediction for D98 and D95 were 2.88%±1.59 and 2.55%±1.28, respectively. The target model can predict the OAR mean dose within 2.43% and the OAR max dose within 4.33% on average across all OARs. CONCLUSION: We developed a site-agnostic model for three-dimensional dose prediction and tested its adaptability to a new target treatment site via transfer learning. Our proposed model can make accurate predictions with limited training data.
Asunto(s)
Aprendizaje Profundo , Radioterapia de Intensidad Modulada , Humanos , Masculino , Redes Neurales de la Computación , Órganos en Riesgo , Dosificación Radioterapéutica , Planificación de la Radioterapia Asistida por Computador/métodos , Radioterapia de Intensidad Modulada/métodosRESUMEN
Recently, artificial intelligence technologies and algorithms have become a major focus for advancements in treatment planning for radiation therapy. As these are starting to become incorporated into the clinical workflow, a major concern from clinicians is not whether the model is accurate, but whether the model can express to a human operator when it does not know if its answer is correct. We propose to use Monte Carlo Dropout (MCDO) and the bootstrap aggregation (bagging) technique on deep learning (DL) models to produce uncertainty estimations for radiation therapy dose prediction. We show that both models are capable of generating a reasonable uncertainty map, and, with our proposed scaling technique, creating interpretable uncertainties and bounds on the prediction and any relevant metrics. Performance-wise, bagging provides statistically significant reduced loss value and errors in most of the metrics investigated in this study. The addition of bagging was able to further reduce errors by another 0.34% for [Formula: see text] and 0.19% for [Formula: see text] on average, when compared to the baseline model. Overall, the bagging framework provided significantly lower mean absolute error (MAE) of 2.62, as opposed to the baseline model's MAE of 2.87. The usefulness of bagging, from solely a performance standpoint, does highly depend on the problem and the acceptable predictive error, and its high upfront computational cost during training should be factored in to deciding whether it is advantageous to use it. In terms of deployment with uncertainty estimations turned on, both methods offer the same performance time of about 12 s. As an ensemble-based metaheuristic, bagging can be used with existing machine learning architectures to improve stability and performance, and MCDO can be applied to any DL models that have dropout as part of their architecture.
Asunto(s)
Aprendizaje Profundo , Método de Montecarlo , Dosis de Radiación , Planificación de la Radioterapia Asistida por Computador/métodos , Incertidumbre , Humanos , Dosificación RadioterapéuticaRESUMEN
Brain injury is a significant source of morbidity and mortality for pediatric patients treated with Extracorporeal Membrane Oxygenation (ECMO). Our objective was to utilize neural networks to predict radiographic evidence of brain injury in pediatric ECMO-supported patients and identify specific variables that can be explored for future research. Data from 174 ECMO-supported patients were collected up to 24 h prior to, and for the duration of, the ECMO course. Thirty-five variables were collected, including physiological data, markers of end-organ perfusion, acid-base homeostasis, vasoactive infusions, markers of coagulation, and ECMO-machine factors. The primary outcome was the presence of radiologic evidence of moderate to severe brain injury as established by brain CT or MRI. This information was analyzed by a neural network, and results were compared to a logistic regression model as well as clinician judgement. The neural network model was able to predict brain injury with an Area Under the Curve (AUC) of 0.76, 73% sensitivity, and 80% specificity. Logistic regression had 62% sensitivity and 61% specificity. Clinician judgment had 39% sensitivity and 69% specificity. Sequential feature group masking demonstrated a relatively greater contribution of physiological data and minor contribution of coagulation factors to the model's performance. These findings lay the foundation for further areas of research directions.
RESUMEN
PURPOSE: In the treatment planning process of intensity-modulated radiation therapy (IMRT), a human planner operates the treatment planning system (TPS) to adjust treatment planning parameters, for example, dose volume histogram (DVH) constraints' locations and weights, to achieve a satisfactory plan for each patient. This process is usually time-consuming, and the plan quality depends on planer's experience and available planning time. In this study, we proposed to model the behaviors of human planners in treatment planning by a deep reinforcement learning (DRL)-based virtual treatment planner network (VTPN), such that it can operate the TPS in a human-like manner for treatment planning. METHODS AND MATERIALS: Using prostate cancer IMRT as an example, we established the VTPN using a deep neural network developed. We considered an in-house optimization engine with a weighted quadratic objective function. Virtual treatment planner network was designed to observe an intermediate plan DVHs and decide the action to improve the plan by changing weights and threshold dose in the objective function. We trained the VTPN in an end-to-end DRL process in 10 patient cases. A plan score was used to measure plan quality. We demonstrated the feasibility and effectiveness of the trained VTPN in another 64 patient cases. RESULTS: Virtual treatment planner network was trained to spontaneously learn how to adjust treatment planning parameters to generate high-quality treatment plans. In the 64 testing cases, with initialized parameters, quality score was 4.97 (±2.02), with 9.0 being the highest possible score. Using VTPN to perform treatment planning improved quality score to 8.44 (±0.48). CONCLUSIONS: To our knowledge, this was the first time that intelligent treatment planning behaviors of human planner in external beam IMRT are autonomously encoded in an artificial intelligence system. The trained VTPN is capable of behaving in a human-like way to produce high-quality plans.
Asunto(s)
Neoplasias de la Próstata , Radioterapia de Intensidad Modulada , Inteligencia Artificial , Humanos , Masculino , Neoplasias de la Próstata/radioterapia , Dosificación Radioterapéutica , Planificación de la Radioterapia Asistida por ComputadorRESUMEN
PURPOSE: We propose a novel domain-specific loss, which is a differentiable loss function based on the dose-volume histogram (DVH), and combine it with an adversarial loss for the training of deep neural networks. In this study, we trained a neural network for generating Pareto optimal dose distributions, and evaluate the effects of the domain-specific loss on the model performance. METHODS: In this study, three loss functions - mean squared error (MSE) loss, DVH loss, and adversarial (ADV) loss - were used to train and compare four instances of the neural network model: (a) MSE, (b) MSE + ADV, (c) MSE + DVH, and (d) MSE + DVH+ADV. The data for 70 prostate patients, including the planning target volume (PTV), and the organs at risk (OAR) were acquired as 96 × 96 × 24 dimension arrays at 5 mm3 voxel size. The dose influence arrays were calculated for 70 prostate patients, using a 7 equidistant coplanar beam setup. Using a scalarized multicriteria optimization for intensity-modulated radiation therapy, 1200 Pareto surface plans per patient were generated by pseudo-randomizing the PTV and OAR tradeoff weights. With 70 patients, the total number of plans generated was 84 000 plans. We divided the data into 54 training, 6 validation, and 10 testing patients. Each model was trained for a total of 100,000 iterations, with a batch size of 2. All models used the Adam optimizer, with a learning rate of 1 × 10-3 . RESULTS: Training for 100 000 iterations took 1.5 days (MSE), 3.5 days (MSE+ADV), 2.3 days (MSE+DVH), and 3.8 days (MSE+DVH+ADV). After training, the prediction time of each model is 0.052 s. Quantitatively, the MSE+DVH+ADV model had the lowest prediction error of 0.038 (conformation), 0.026 (homogeneity), 0.298 (R50), 1.65% (D95), 2.14% (D98), and 2.43% (D99). The MSE model had the worst prediction error of 0.134 (conformation), 0.041 (homogeneity), 0.520 (R50), 3.91% (D95), 4.33% (D98), and 4.60% (D99). For both the mean dose PTV error and the max dose PTV, Body, Bladder and rectum error, the MSE+DVH+ADV outperformed all other models. Regardless of model, all predictions have an average mean and max dose error <2.8% and 4.2%, respectively. CONCLUSION: The MSE+DVH+ADV model performed the best in these categories, illustrating the importance of both human and learned domain knowledge. Expert human domain-specific knowledge can be the largest driver in the performance improvement, and adversarial learning can be used to further capture nuanced attributes in the data. The real-time prediction capabilities allow for a physician to quickly navigate the tradeoff space for a patient, and produce a dose distribution as a tangible endpoint for the dosimetrist to use for planning. This is expected to considerably reduce the treatment planning time, allowing for clinicians to focus their efforts on the difficult and demanding cases.
Asunto(s)
Aprendizaje Profundo , Dosis de Radiación , Planificación de la Radioterapia Asistida por Computador/métodos , Humanos , Masculino , Neoplasias de la Próstata/radioterapia , Dosificación RadioterapéuticaRESUMEN
Measurements of the incident fluence of HZE particles, as a function of LET, are used to determine absorbed dose as well as Quality Factors for assigning risk estimates to astronauts during manned space missions. These data are often based on thin solid state detectors that measure energy deposition, dE, and the assumption that the trajectory of the particle, dx, is equivalent to the thickness of the detector. Heavy ions often fragment while penetrating shielding materials in vehicles or habitats. Projectile fragments can be clustered spatially and temporally at the location of the thin detector which are then misclassified as a single particle. Eliminating the confounding effects of coincident events is the first step in extending the reach of flight instruments to identify the charge and velocity of individual particles. Identification of individual particles, in a fragmentation spectrum, will require that detection systems have sufficient segmentation to eliminate coincident events. The objective of this study was to reduce coincident events while avoiding over-design and complexity. Monte Carlo simulations, using Geant4, were performed for 4He, 12C, 28Si and 56Fe ions at energies of 300, 900 and 2400 MeV/n incident upon aluminum shields having areal densities of 5.4, 13.5, and 54 g/cm2. The identity, energy and spatial distribution of all particles downstream from the shielding were analyzed using a novel approach based on proximity distributions. Results indicated that pixel dimensions on the order of 1 mm were sufficient to reduce errors caused by coincident events for active space radiation detectors.
Asunto(s)
Astronautas , Radiación Cósmica , Planeta Tierra , Exposición Profesional/análisis , Monitoreo de Radiación/métodos , Protección Radiológica , Humanos , Método de Montecarlo , Dosis de Radiación , Monitoreo de Radiación/instrumentación , Efectividad Biológica Relativa , Vuelo EspacialRESUMEN
In order to define the ranges of relevant neutron energies for the purposes of measurement and dosimetry in space, we have performed a series of Monte Carlo transport model calculations that predict the neutron field created by Galactic Cosmic Ray interactions inside a variety of simple shielding configurations. These predictions indicate that a significant fraction of the neutron fluence and neutron effective dose lies in the region above 20 MeV up to several hundred MeV. These results are consistent over thicknesses of shielding that range from very thin (2.7 g/cm(2)) to thick (54 g/cm(2)), and over both shielding materials considered (aluminum and water). In addition to these results, we have also investigated whether simplified Galactic Cosmic Ray source terms can yield predictions that are equivalent to simulations run with a full GCR source term. We found that a source using a GCR proton and helium spectrum together with a scaled oxygen spectrum yielded nearly identical results to a full GCR spectrum, and that the scaling factor used for the oxygen spectrum was independent of shielding material and thickness. Good results were also obtained using a GCR proton spectrum together with a scaled helium spectrum, with the helium scaling factor also independent of shielding material and thickness. Using a proton spectrum alone was unable to reproduce the full GCR results.
Asunto(s)
Neutrones , Radiación Cósmica , Modelos Teóricos , Método de Montecarlo , Protones , Dosis de Radiación , Protección Radiológica , Radiometría , Vuelo EspacialRESUMEN
NASA has derived new models for radiological risk assessment based on epidemiological data and radiation biology including differences in Relative Biological Effectiveness for leukemia and solid tumors. Comprehensive approaches were used to develop new risk cross sections and the extension of these into recommendations for risk assessment during space missions. The methodology relies on published data generated and the extensive research initiative managed by the NASA Human Research Program (HRP) and reviewed by the National Academy of Sciences. This resulted in recommendations for revised specifications of quality factors, QNASA(Z,ß) in terms of track structure concepts that extend beyond LET alone. The new paradigm for quality factors placed demands on radiation monitoring procedures that are not satisfied by existing dosimetry systems or particle spectrometers that are practical for space exploration where mass, volume, band width and power consumption are highly constrained. We have proposed a new definition of quality factors that relaxes the requirements for identifying charge, Z, and velocity, ß, of the incident radiation while still preserving the functional form of the inherent risk functions. The departure from the exact description of QNASA(Z,ß) is that the revised values are new functions of LET for solid cancers and leukemia. We present the motivation and process for developing the revised quality factors. We describe results of extensive simulations using GCR distributions in free space as well as the resulting spectra of primary and secondary particles behind aluminum shields and penetration through water. In all cases the revised dose averaged quality factors agreed with those based on the values obtained using QNASA(Z,ß). This provides confidence that emerging technologies for space radiation dosimetry can provide real time measurements of dose and dose equivalent while satisfying constraints on size, mass, power and bandwidth. The revised quality factors are sufficiently generalized to be applicable to radiation protection practices beyond space exploration.