Error detection and classification in patient‐specific IMRT QA with dual neural networks

Abstract

Purpose: Despite being the standard metric in patient-specific quality assurance (QA) for intensitymodulated radiotherapy (IMRT), gamma analysis has two shortcomings: (a) it lacks sensitivity to small but clinically relevant errors (b) it does not provide efficient means to classify the error sources. The purpose of this work is to propose a dual neural network method to achieve simultaneous error detection and classification in patient-specific IMRT QA. Methods: For a pair of dose distributions, we extracted the dose difference histogram (DDH) for the low dose gradient region and two signed distance-to-agreement (sDTA) maps (one in x direction and one in y direction) for the high dose gradient region. An artificial neural network (ANN) and a convolutional neural network (CNN) were designed to analyze the DDH and the two sDTA maps, respectively. The ANN was trained to detect and classify six classes of dosimetric errors: incorrect multileaf collimator (MLC) transmission (AE1%) and four types of monitor unit (MU) scaling errors (AE1% and AE2%). The CNN was trained to detect and classify seven classes of spatial errors: incorrect effective source size, 1 mm MLC leaf bank overtravel or undertravel, 2 mm single MLC leaf overtravel or undertravel, and device misalignment errors (1 mm in x-or y direction). An in-house planar dose calculation software was used to simulate measurements with errors and noise introduced. Both networks were trained and validated with 13 IMRT plans (totaling 88 fields). A fivefold cross-validation technique was used to evaluate their accuracy. Results: Distinct features were found in the DDH and the sDTA maps. The ANN perfectly identified all four types of MU scaling errors and the specific accuracies for the classes of no error, MLC transmission increase, MLC transmission decrease were 98.9%, 96.6%, and 94.3%, respectively. For the CNN, the largest confusion occurred between the 1-mm-MLC bank overtravel class and the 1-mmdevice alignment error in x-direction class, which brought the specific accuracies down to 90.9% and 92.0%, respectively. The specific accuracy for the 2-mm-single MLC leaf undertravel class was 93.2% as it misclassified 5.7% of the class as being error free (false negative). Otherwise, the specific accuracy was above 95%. The overall accuracies across the fivefold were 98.3 AE 0.7% and 95.6% AE 1.5% for the ANN and the CNN, respectively. Conclusions: Both the DDH and the sDTA maps are suitable features for error classification in IMRT QA. The proposed dual neural network method achieved simultaneous error detection and classification with excellent accuracy. It could be used in complement with the gamma analysis to potentially shift the IMRT QA paradigm from passive pass/fail analysis to active error detection and root cause identification.

The structural similarity index for IMRT quality assurance: radiomics‐based error classification

“…For example, positioning errors (i.e., MLC errors, EPID misalignment), can be clearly reflected in contrast maps or DTA maps. 34 In this work, the measured PD images were directly compared to the calculated PD images for difference map calculation, accounting for the existing discrepancy between the PDIPmodeled images and acquired PD images. The obtained results in the current work further proved the feasibility of such error detection methods in clinical environments.…”

“…Comparing to the DD‐DTA‐blended gamma maps, the four difference maps could reflect different errors patterns resulted from various beam delivery errors separately hence more informative features could be preserved. For example, positioning errors (i.e., MLC errors, EPID misalignment), can be clearly reflected in contrast maps or DTA maps 34 . In this work, the measured PD images were directly compared to the calculated PD images for difference map calculation, accounting for the existing discrepancy between the PDIP‐modeled images and acquired PD images.…”

Error detection model developed using a multi‐task convolutional neural network in patient‐specific quality assurance for volumetric‐modulated arc therapy

“…Several reports have proposed machine learning‐based dose analysis in static beam IMRT QA dosimetry. Potter et al analyzed DTA maps and DD histograms obtained using the approach of a 2‐D diode array with two types of deep learning 29 . They performed error simulation for MLC mispositioning, MLC transmission factor, MU, source size, and misalignment of measuring device.…”

“…Recently, error detection models that focus on errors other than MLC leaf positioning errors have also been investigated. [29][30][31][32] For static beam IMRT QA dosimetry, Potter et al developed an error detection model using a deep learning approach to detect errors of MLC position, MLC transmission, MU, effective source size, and alignment of a measuring device. Sakai et al reported a machine learning approach focusing on MLC modeling error detection, such as dosimetric leaf gap (DLG).…”

Analysis of EPID Transmission Fluence Maps Using Machine Learning Models and CNN for Identifying Position Errors in the Treatment of GO Patients

“…Convolutional neural networks (CNNs) can effectively perform image-related tasks by analyzing images at different scales using convolutional layers to extract useful information and generate final outputs (31). Accordingly, a number of researchers have proposed CNN-based patient-specific dose verification using dose maps as the CNN input (32)(33)(34)(35).…”