Investigating the effectiveness of monitoring relevant variations during IMRT and VMAT treatments by EPID-based 3D in vivo verification performed using planning CTs

IPEM topical report: guidance for the clinical implementation of online treatment monitoring solutions for IMRT/VMAT

This report provides guidance for the implementation of online treatment monitoring (OTM) solutions in radiotherapy (RT), with a focus on modulated treatments. Support is provided covering the implementation process, from identification of an OTM solution to local implementation strategy. Guidance has been developed by a RT special interest group (RTSIG) working party (WP) on behalf of the Institute of Physics and Engineering in Medicine (IPEM). Recommendations within the report are derived from the experience of the WP members (in consultation with manufacturers, vendors and user groups), existing guidance or legislation and a UK survey conducted in 2020 (Stevenset al2021). OTM is an inclusive term representing any system capable of providing a direct or inferred measurement of the delivered dose to a RT patient. Information on each type of OTM is provided but, commensurate with UK demand, guidance is largely influenced byin vivodosimetry methods utilising the electronic portal imager device (EPID). Sections are included on the choice of OTM solutions, acceptance and commissioning methods with recommendations on routine quality control, analytical methods and tolerance setting, clinical introduction and staffing/resource requirements. The guidance aims to give a practical solution to sensitivity and specificity testing. Functionality is provided for the user to introduce known errors into treatment plans for local testing. Receiver operating characteristic analysis is discussed as a tool to performance assess OTM systems. OTM solutions can help verify the correct delivery of radiotherapy treatment. Furthermore, modern systems are increasingly capable of providing clinical decision-making information which can impact the course of a patient's treatment. However, technical limitations persist. It is not within the scope of this guidance to critique each available solution, but the user is encouraged to carefully consider workflow and engage with manufacturers in resolving compatibility issues.

On the beam hardening correction of the Transit-Guided Radiation Therapy attenuation model

PURPOSE

The Transit-Guided Radiation Therapy (TGRT) technique is a novel technique aimed to quantify the position error of a patient by using the transit portal images (TPI) of the treatment fields. Despite of the promising preliminary results, about 4% of the cases would have led to position overcorrections. In this work, the TGRT formalism is refined to improve its accuracy and, especially, to decrease the risk of overcorrections.

METHODS

A second free parameter accounting for beam hardening has been added to the attenuation model of the TGRT formalism. Five treatment plans combining different delivery techniques and tumour sites have been delivered to an anthropomorphic phantom. TPIs have been obtained under a set of random couch shifts for each field. For each TPI, both the original and the refined TGRT formalism have been used to estimate the underlying true shift.

RESULTS

With respect the original formalism, the refined formalism: (i) decreased both the number (from 5% to 1%) and the magnitude of the overcorrections; (ii) lowered the detection threshold (from approximately 1 mm to <0.3 mm); (iii) largely improved the accuracy in tumour sites with large mass thickness variations; and (iv) largely improved the accuracy for true shifts below 5 mm. For true shifts above 5 mm, the accuracy was slightly impaired.

CONCLUSIONS

The refined TGRT formalism performed globally better than the original TGRT formalism and it largely reduced the risk of overcorrections. Further refinements of the TGRT formalism should focus on true shifts above 5 mm.

Transit-guided radiation therapy: proof of concept of an on-line technique for correcting position errors using transit portal images

Objective. Transit in vivo dosimetry methods monitor that the dose distribution is delivered as planned. However, they have a limited ability to identify and to quantify the cause of a given disagreement, especially those caused by position errors. This paper describes a proof of concept of a simple in vivo technique to infer a position error from a transit portal image (TPI). Approach. For a given treatment field, the impact of a position error is modeled as a perturbation of the corresponding reference (unperturbed) TPI. The perturbation model determines the patient translation, described by a shift vector, by comparing a given in vivo TPI to the corresponding reference TPI. Patient rotations can also be determined by applying this formalism to independent regions of interest over the patient. Eight treatment plans have been delivered to an anthropomorphic phantom under a large set of couch shifts (<15 mm) and rotations (<10°) to experimentally validate this technique, which we have named Transit-Guided Radiation Therapy (TGRT). Main results. The root mean squared error (RMSE) between the determined and the true shift magnitudes was 1.0/2.4/4.9 mm for true shifts ranging between 0–5/5–10/10–15 mm, respectively. The angular accuracy of the determined shift directions was 12° ± 14°. The RMSE between the determined and the true rotations was 0.5°. The TGRT technique decoupled translations and rotations satisfactorily. In 96% of the cases, the TGRT technique decreased the existing position error. The detection threshold of the TGRT technique was around 1 mm and it was nearly independent of the tumor site, delivery technique, beam energy or patient thickness. Significance. TGRT is a promising technique that not only provides reliable determinations of the position errors without increasing the required equipment, acquisition time or patient dose, but it also adds on-line correction capabilities to existing methods currently using TPIs.

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

Purpose

To find a suitable method for analyzing electronic portal imaging device (EPID) transmission fluence maps for the identification of position errors in the in vivo dose monitoring of patients with Graves' ophthalmopathy (GO).

Methods

Position errors combining 0-, 2-, and 4-mm errors in the left-right (LR), anterior-posterior (AP), and superior-inferior (SI) directions in the delivery of 40 GO patient radiotherapy plans to a human head phantom were simulated and EPID transmission fluence maps were acquired. Dose difference (DD) and structural similarity (SSIM) maps were calculated to quantify changes in the fluence maps. Three types of machine learning (ML) models that utilize radiomics features of the DD maps (ML 1 models), features of the SSIM maps (ML 2 models), and features of both DD and SSIM maps (ML 3 models) as inputs were used to perform three types of position error classification, namely a binary classification of the isocenter error (type 1), three binary classifications of LR, SI, and AP direction errors (type 2), and an eight-element classification of the combined LR, SI, and AP direction errors (type 3). Convolutional neural network (CNN) was also used to classify position errors using the DD and SSIM maps as input.

Results

The best-performing ML 1 model was XGBoost, which achieved accuracies of 0.889, 0.755, 0.778, 0.833, and 0.532 in the type 1, type 2-LR, type 2-AP, type 2-SI, and type 3 classification, respectively. The best ML 2 model was XGBoost, which achieved accuracies of 0.856, 0.731, 0.736, 0.949, and 0.491, respectively. The best ML 3 model was linear discriminant classifier (LDC), which achieved accuracies of 0.903, 0.792, 0.870, 0.931, and 0.671, respectively. The CNN achieved classification accuracies of 0.925, 0.833, 0.875, 0.949, and 0.689, respectively.

Conclusion

ML models and CNN using combined DD and SSIM maps can analyze EPID transmission fluence maps to identify position errors in the treatment of GO patients. Further studies with large sample sizes are needed to improve the accuracy of CNN.

Radiomics analysis of EPID measurements for patient positioning error detection in thyroid associated ophthalmopathy radiotherapy

PURPOSE

Electronic portal imaging detector (EPID)-based patient positioning verification is an important component of safe radiotherapy treatment delivery. In computer simulation studies, learning-based approaches have proven to be superior to conventional gamma analysis in the detection of positioning errors. To approximate a clinical scenario, the detectability of positioning errors via EPID measurements was assessed using radiomics analysis for patients with thyroid-associated ophthalmopathy.

METHODS

Treatment plans of 40 patients with thyroid-associated ophthalmopathy were delivered to a solid anthropomorphic head phantom. To simulate positioning errors, combinations of 0-, 2-, and 4-mm translation errors in the left-right (LR), superior-inferior (SI), and anterior-posterior (AP) directions were introduced to the phantom. The positioning errors-induced dose differences between measured portal dose images were used to predict the magnitude and direction of positioning errors. The detectability of positioning errors was assessed via radiomics analysis of the dose differences. Three classification models-support vector machine (SVM), k-nearest neighbors (KNN), and XGBoost-were used for the detection of positioning errors (positioning errors larger or smaller than 3 mm in an arbitrary direction) and direction classification (positioning errors larger or smaller than 3 mm in a specific direction). The receiver operating characteristic curve and the area under the ROC curve (AUC) were used to evaluate the performance of classification models.

RESULTS

For the detection of positioning errors, the AUC values of SVM, KNN, and XGBoost models were all above 0.90. For LR, SI, and AP direction classification, the highest AUC values were 0.76, 0.91, and 0.80, respectively.

CONCLUSIONS

Combined radiomics and machine learning approaches are capable of detecting the magnitude and direction of positioning errors from EPID measurements. This study is a further step toward machine learning-based positioning error detection during treatment delivery with EPID measurements.

Portal dosimetry in radiotherapy repeatability evaluation

The accuracy of radiotherapy is the subject of continuous discussion, and dosimetry methods, particularly in dynamic techniques, are being developed. At the same time, many oncology centers develop quality procedures, including pretreatment and online dose verification and proper patient tracking methods. This work aims to present the possibility of using portal dosimetry in the assessment of radiotherapy repeatability. The analysis was conducted on 74 cases treated with dynamic techniques. Transit dosimetry was made for each collision‐free radiation beam. It allowed the comparison of summary fluence maps, obtained for fractions with the corresponding summary maps from all other treatment fractions. For evaluation of the compatibility in the fluence map pairs (6798), the gamma coefficient was calculated. The results were considered in four groups, depending on the used radiotherapy technique: stereotactic fractionated radiotherapy, breath‐hold, free‐breathing, and conventionally fractionated other cases. The chi² or Fisher's exact test was made depending on the size of the analyzed set and also Mann–Whitney U‐test was used to compare treatment repeatability of different techniques. The aim was to test whether the null hypothesis of error‐free therapy was met. The patient is treated repeatedly if the P‐value in all the fluence maps sets is higher than the level of 0.01. The best compatibility between treatment fractions was obtained for the stereotactic technique. The technique with breath‐holding gave the lowest percentage of compliance of the analyzed fluence pairs. The results indicate that the repeatability of the treatment is associated with the radiotherapy technique. Treated volume location is also an essential factor found in the evaluation of treatment accuracy. The EPID device is a useful tool in assessing the repeatability of radiotherapy. The proposed method of fluence maps comparison also allows us to assess in which therapeutic session the patient was treated differently from the other fractions.

A preliminary study of a photon dose calculation algorithm using a convolutional neural network

The aim of dose calculation algorithm research is to improve the calculation accuracy while maximizing the calculation efficiency. In this study, the three-dimensional distribution of total energy release per unit mass (TERMA) and the electron density (ED) distribution are considered inputs in a method for calculating the three-dimensional dose distribution based on a convolutional neural network (CNN). Attempts are made to improve the efficiency of the collapsed cone convolution/superposition (CCCS) algorithm while providing an approach to improve the efficiency of other traditional dose calculation algorithms. Twelve sets of computed tomography (CT) images were employed for training. Data sets were generated by the CCCS algorithm with a random beam configuration. For each monoenergetic photon model, 7500 samples were generated for the training set, and 1500 samples were generated for the validation set. Training occurred for 0.5 MeV, 1 MeV, 2 MeV, 3 MeV, 4 MeV, 5 MeV, and 6 MeV monoenergetic photon models. To evaluate the usability under linac conditions, a comparison between CCCS and CNN-Dose was performed for the Mohan 6-MV spectrum for 12 additional new sets of CT images with different anatomies. A total of 1512 test samples were generated. For all anatomies, the mean value, 95% lower confidence limit (LCL) and 95% upper confidence limit (UCL) were 99.56%, 99.51% and 99.61%, respectively, at the 3%/2 mm criteria. The mean value, 95% LCL and 95% UCL were 98.57%, 98.46% and 98.67%, respectively, at the 2%/2 mm criteria. The results meet the relevant clinical requirements. In the proposed methods, the dose distribution of clinical energy can be obtained by TERMA, and the electronic density can be obtained with a CNN. This method can also be used for other traditional dose algorithms and displays potential in treatment planning, adaptive radiation therapy, and in vivo verification.

The effect of the choice of patient model on the performance of in vivo 3D EPID dosimetry to detect variations in patient position and anatomy

PURPOSE

In vivo EPID dosimetry is meant to trigger on relevant differences between delivered and planned dose distributions and should therefore be sensitive to changes in patient position and patient anatomy. Three-dimensional (3D) EPID back-projection algorithms can use either the planning computed tomography (CT) or the daily patient anatomy as patient model for dose reconstruction. The purpose of this study is to quantify the effect of the choice of patient model on the performance of in vivo 3D EPID dosimetry to detect patient-related variations.

METHODS

Variations in patient position and patient anatomy were simulated by transforming the reference planning CT images (pCT) into synthetic daily CT images (dCT) representing a variation of a given magnitude in patient position or in patient anatomy. For each variation, synthetic in vivo EPID data were also generated to simulate the reconstruction of in vivo EPID dose distributions. Both the planning CT images and the synthetic daily CT images could be used as patient model in the reconstructions yielding e D pCT and e D dCT EPID reconstructed dose distributions respectively. The accuracy of e D pCT and e D dCT reconstructions was evaluated against absolute dose measurements made in different phantom setups, and against dose distributions calculated by the treatment planning system (TPS). The comparison was performed by γ-analysis (3% local dose/2 mm). The difference in sensitivity between e D pCT and e D dCT reconstructions to detect variations in patient position and in patient anatomy was investigated using receiver operating characteristic analysis and the number of triggered alerts for 100 volumetric modulated arc therapy plans and 12 variations.

RESULTS

e D dCT showed good agreement with both absolute point dose measurements (<0.5%) and TPS data (γ-mean = 0.52 ± 0.11). The agreement degraded with e D pCT , with the magnitude of the deviation varying with each specific case. e D dCT readily detected combined 3 mm translation setup errors in all directions (AUC = 1.0) and combined 3° rotation setup errors around all axes (AUC = 0.86) whereas e D pCT showed good detectability only for 12 mm translations (AUC = 0.85) and 9° rotations (AUC = 0.80). Conversely, e D pCT manifested a higher sensitivity to patient anatomical changes resulting in AUC values of 0.92/0.95 for a 6 mm patient contour expansion/contraction compared to 0.70/0.64 with e D dCT . Using |ΔPTV_D50 | > 3% as clinical tolerance level, the percentage of alerts for 6 mm changes in patient contour were 85%/27% with e D pCT / e D dCT .

CONCLUSIONS

With planning CT images as patient model, EPID dose reconstructions underestimate the dosimetric effects caused by errors in patient positioning and overestimate the dosimetric effects caused by changes in patient anatomy. The use of the daily patient position and anatomy as patient model for in vivo 3D EPID transit dosimetry improves the ability of the system to detect uncorrected errors in patient position and it reduces the likelihood of false positives due to patient anatomical changes.