RESUMO
INTRODUCTION: Unscheduled machine downtime can cause treatment interruptions and adversely impact patient treatment outcomes. Conventional Quality Assurance (QA) programs of a proton Pencil Beam Scanning (PBS) system ensure its operational performance by keeping the beam parameters within clinical tolerances but often do not reveal the underlying issues of the device prior to a machine malfunction event. In this study, we propose a Predictive Maintenance (PdM) approach that leverages an advanced analytical tool built on a deep neural network to detect treatment delivery machine issues early. METHODS: Beam delivery log file data from daily QA performed at the Burr Proton Center of Massachusetts General Hospital were collected. A novel PdM framework consisting of long short-term memory-based autoencoder (LSTM-AE) modeling of the proton PBS delivery system and a Mahalanobis distance-based error metric evaluation was constructed to detect rare anomalous machine events. These included QA beam pauses, clinical operational issues, and treatment interruptions. The model was trained in an unsupervised fashion on the QA data of normal sessions so that the model learned characteristics of normal machine operation. The anomaly is quantified as the multivariate deviation between the model predicted data and the measured data of the day using Mahalanobis distance (M-Score). Two-layer and three-layer Long short-term memory-based stacked autoencoder (LSTM-SAE) models were optimized for exploring model performance improvement. Model validation was performed with two clinical datasets and was analyzed using the area under the precision-recall curve (AUPRC) and the area under the receiver operating characteristic (AUROC). RESULTS: LSTM-SAE models showed strong performance in predicting QA beam pauses for both clinical validation datasets. Despite severe skew in the dataset, the model achieved AUPRC of 0.60 and 0.82 and AUROC of 0.75 and 0.92 in the respective 2018 and 2020 datasets. Moreover, these amount to 2.8-fold and 10.7-fold enhancement compared to the respective baseline event rates. In addition, in terms of treatment interruption events, model prediction enabled 3.88-fold and 51.2-fold detection improvement, while the detection improvement for clinical operational issues was 1.04-fold and 1.37-fold, respectively, in the 2018 and 2020 datasets. CONCLUSION: Our novel deep LSTM-SAE-based framework allows for highly discriminative prediction of anomalous machine events and demonstrates great promise for enabling PdM for proton PBS beam delivery.
Assuntos
Terapia com Prótons , Prótons , Humanos , Redes Neurais de ComputaçãoRESUMO
A spread-out Bragg peak (SOBP) is used in proton beam therapy to create a longitudinal conformality of the required dose to the target. In order to create this effect in a passive beam scattering system, a variety of components must operate in conjunction to produce the desired beam parameters. We will describe how the SOBP is generated and will explore the tolerances of the various components and their subsequent effect on the dose distribution. A specific aspect of this investigation includes a case study involving the use of a beam current modulated system. In such a system, the intensity of the beam current can be varied in synchronization with the revolution of the range-modulator wheel. As a result, the weights of the pulled-back Bragg peaks can be individually controlled to produce uniform dose plateaus for a large range of treatment depths using only a small number of modulator wheels.
Assuntos
Radioterapia/instrumentação , Radioterapia/métodos , Algoritmos , Simulação por Computador , Desenho de Equipamento , Modelos Estatísticos , Aceleradores de Partículas , Prótons , Espalhamento de Radiação , Sensibilidade e Especificidade , Software , Fatores de TempoRESUMO
Proton therapy offers the potential for excellent dose conformity and reduction in integral dose. The superior dose distribution is, however, much more sensitive to changes in radiological depths along the beam path than for photon fields. Respiratory motion can cause such changes for treatments sites like lung, liver, and mediastinum and thus affect the proton dose distribution significantly. We have implemented and commissioned a respiratory-gated system for range-modulated treatment fields. The gating system was designed to ensure that each gate always contains complete modulation cycles so that for any beam segment the delivered dose has the planned depth-dose distribution. Measurements have been made to estimate the time delays for the various components of the system. The total delay between the actual motion and the beam on/off control is in the range of 65-195 ms. Time-resolved dose measurements and film tests were also conducted to examine the overall gating effect.