Error Detection in Intensity-Modulated Radiation Therapy Quality Assurance Using Radiomic Analysis of Gamma Distributions

Wootton, Landon; Nyflot, Matthew J.; Chaovalitwongse, Wanpracha Art; Ford, Eric

doi:10.1016/j.ijrobp.2018.05.033

Cited by 50 publications

(73 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the study that preceded this work, Wootton et al. demonstrated that texture features could be applied to gamma images from IMRT QA measurements with an EPID device, and that gamma‐radiomic models were more accurate at detecting simulated errors than commonly used percent‐passing threshold criteria . Similarly, Valdes et al.…”

Section: Discussionmentioning

confidence: 90%

“…First, it was necessary to generate planar dose maps from clinical radiotherapy plans. The methods for error modeling in the treatment planning system have been previously described . Briefly, the linear accelerator models used were Elekta Synergy linear accelerators with Agility multileaf collimators (MLC) and treatment planning was done in Pinnacle 9.10.…”

Section: Methodsmentioning

confidence: 99%

“…The introduced errors involved the MLC, which shapes the beam throughout IMRT delivery. MLC errors were a practical test case for this study because the impact of MLC misalignment on gamma pass rates has been described in various settings . The two errors were random MLC misalignment, in which each leaf was offset by a distance determined from a Gaussian distribution of width 2 mm, and systematic MLC misalignment, in which all leaves were offset by 2 mm in the same direction.…”

Section: Methodsmentioning

confidence: 99%

“…To provide context for the prediction results, the deep learning approach was compared to an approach using human‐designed image features. These features were chosen based on our prior experience with radiomic analysis of gamma images . Briefly, 17 texture features from intensity histograms, which quantify global trends in the distribution of gamma values, and size zone matrices, which classify spatial groupings of similarly valued pixels, were chosen a priori and extracted from the gamma maps using the PORTS software .…”

Section: Methodsmentioning

confidence: 99%

“…While most of the work in radiomics has been in the field of radiology, radiation oncology is rich in quantitative spatial information, including initial diagnostic imaging and patient simulation, treatment planning systems with detailed image annotation and volumetric dose maps, and sophisticated onboard systems capable of verification imaging in a variety of ways. Most reports to date have used image features such as the so‐called texture features, which consist of features derived from intensity histograms, co‐occurrence matrices, neighborhood differences matrices, zone size matrices, and other prior applications in the literature The first reports of applying radiomics to radiotherapy QA have begun to appear, including analysis of cone‐beam CT images to detect image quality changes and analysis of patient‐specific QA images to predict simulated radiotherapy errors and correlate to gamma pass rates …”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Deep learning for patient‐specific quality assurance: Identifying errors in radiotherapy delivery by radiomic analysis of gamma images with convolutional neural networks

et al. 2018

Self Cite

View full text Add to dashboard Cite

Purpose Patient‐specific quality assurance (QA) for intensity‐modulated radiation therapy (IMRT) is a ubiquitous clinical procedure, but conventional methods have often been criticized as being insensitive to errors or less effective than other common physics checks. Recently, there has been interest in the application of radiomics, quantitative extraction of image features, to radiotherapy QA. In this work, we investigate a deep learning approach to classify the presence or absence of introduced radiotherapy treatment delivery errors from patient‐specific QA. Methods Planar dose maps from 186 IMRT beams from 23 IMRT plans were evaluated. Each plan was transferred to a cylindrical phantom CT geometry. Three sets of planar doses were exported from each plan corresponding to (a) the error‐free case, (b) a random multileaf collimator (MLC) error case, and (c) a systematic MLC error case. Each plan was delivered to the electronic portal imaging device (EPID), and planned and measured doses were used to calculate gamma images in an EPID dosimetry software package (for a total of 558 gamma images). Two radiomic approaches were used. In the first, a convolutional neural network with triplet learning was used to extract image features from the gamma images. In the second, a handcrafted approach using texture features was used. The resulting metrics from both approaches were input into four machine learning classifiers (support vector machines, multilayer perceptrons, decision trees, and k‐nearest‐neighbors) in order to determine whether images contained the introduced errors. Two experiments were considered: the two‐class experiment classified images as error‐free or containing any MLC error, and the three‐class experiment classified images as error‐free, containing a random MLC error, or containing a systematic MLC error. Additionally, threshold‐based passing criteria were calculated for comparison. Results In total, 303 gamma images were used for model training and 255 images were used for model testing. The highest classification accuracy was achieved with the deep learning approach, with a maximum accuracy of 77.3% in the two‐class experiment and 64.3% in the three‐class experiment. The performance of the handcrafted approach with texture features was lower, with a maximum accuracy of 66.3% in the two‐class experiment and 53.7% in the three‐class experiment. Variability between the results of the four machine learning classifiers was lower for the deep learning approach vs the texture feature approach. Both radiomic approaches were superior to threshold‐based passing criteria. Conclusions Deep learning with convolutional neural networks can be used to classify the presence or absence of introduced radiotherapy treatment delivery errors from patient‐specific gamma images. The performance of the deep learning network was superior to a handcrafted approach with texture features, and both radiomic approaches were better than threshold‐based passing criteria. The results suggest that radiomic QA is a promising direction f...

show abstract

Section: Discussionmentioning

confidence: 90%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Deep learning for patient‐specific quality assurance: Identifying errors in radiotherapy delivery by radiomic analysis of gamma images with convolutional neural networks

et al. 2018

Self Cite

View full text Add to dashboard Cite

show abstract

A comprehensive and clinical‐oriented evaluation criteria based on DVH information and gamma passing rates analysis for IMRT plan 3D verification

Dang

et al. 2020

J Applied Clin Med Phys

View full text Add to dashboard Cite

To accomplish the 3D dose verification to IMRT plan by incorporating DVH information and gamma passing rates (GPs) (DVH_GPs) so as to better correlate the patient-specific quality assurance (QA) results with clinically relevant metrics. Materials and methods: DVH_GPs analysis was performed to specific structures of 51 intensity-modulated radiotherapy (IMRT) treatment plans (17 plans each for oropharyngeal neoplasm, esophageal neoplasm, and cervical neoplasm) with Delta4 3D dose verification system. Based on the DVH action levels of 5% and GPs action levels of 90% (3%/2 mm), the evaluation results of DVH_GPs analysis were categorized into four regions as follows: the true positive (TP) (%DE> 5%, GPs < 90%), the false positive (FP) (%DE ≤ 5%, GPs < 90%), the false negative (FN) (%DE> 5%, GPs ≥ 90%), and the true negative (TN) (%DE ≤ 5%, GPs ≥ 90%). Considering the actual situation, the final patient-specific QA determination was made based on the DVH_GPs evaluation results. In order to exclude the impact of Delta4 phantom on the DVH_GPs evaluation results, 5 cm phantom shift verification was carried out to structures with abnormal results (femoral heads, lung, heart). Results: In DVH_GPs evaluation, 58 cases with FN, 5 cases with FP, and 2 cases with TP were observed. After the phantom shift verification, the extremely abnormal FN of both lung (%DE = 21.52%±8.20%) and heart (%DE = 19.76%) in the oropharyngeal neoplasm plans and of the bilateral formal heads (%DE = 26.41%±13.45%) in cervical neoplasm plans disappeared dramatically. DVH_GPs analysis was performed to all evaluation results in combination with clinical treatment criteria. Finally, only one TP case from the oropharyngeal neoplasm plans and one FN case from the esophageal neoplasm plans did not meet the treatment requirements, so they needed to be replanned. Conclusion: The proposed DVH_GPs evaluation method first make up the deficiency of conventional gamma analysis regarding intensity information and space information. Moreover, it improves the correlation between the patient-specific QA results and clinically relevant metrics. Finally, it can distinguish the TP, TN, FP, and

show abstract

Applications of machine and deep learning to patient‐specific IMRT/VMAT quality assurance

Osman

Maalej

2021

J Applied Clin Med Phys

View full text Add to dashboard Cite

In order to deliver accurate and safe treatment to cancer patients in radiation therapy using advanced techniques such as intensity modulated radiation therapy (IMRT) and volumetric‐arc radiation therapy (VMAT), patient specific quality assurance (QA) should be performed before treatment. IMRT/VMAT dose measurements in a phantom using various devices have been clinically adopted as standard method for QA. This approach allows the verification of the accuracy of the dose calculation, data transfer, and the delivery system. However, patient‐specific QA procedures are expensive and require significant time and effort by the physicists. Over the past 5 years, machine learning (ML) and deep learning (DL) algorithms for predictions of IMRT/VMAT QA outcome have been investigated. Various ML and DL models have shown promising prediction accuracy and a high potential as time‐efficient virtual QA tool. In this paper, we review the ML and DL based models that were developed for patient specific IMRT and VMAT QA outcome predictions from algorithmic and clinical applicability perspectives. We focus on comparing the algorithms, the dataset sizes, the input parameters and features, the QA outcome prediction approaches, the validation, the performance, the clinical applicability, and the potential clinical impact. In addition, we discuss the present challenges as well as the future directions in the implementation of these models. To the best of our knowledge, this is the first review on the application of ML and DL based models in IMRT/VMAT QA predictions.

show abstract

Error Detection in Intensity-Modulated Radiation Therapy Quality Assurance Using Radiomic Analysis of Gamma Distributions

Cited by 50 publications

References 39 publications

Deep learning for patient‐specific quality assurance: Identifying errors in radiotherapy delivery by radiomic analysis of gamma images with convolutional neural networks

Deep learning for patient‐specific quality assurance: Identifying errors in radiotherapy delivery by radiomic analysis of gamma images with convolutional neural networks

A comprehensive and clinical‐oriented evaluation criteria based on DVH information and gamma passing rates analysis for IMRT plan 3D verification

Applications of machine and deep learning to patient‐specific IMRT/VMAT quality assurance

Contact Info

Product

Resources

About