Implementation of volumetric-modulated arc therapy for locally advanced breast cancer patients: Dosimetric comparison with deliverability consideration of planning techniques and predictions of patient-specific QA results via supervised machine learning

Deep learning for patient-specific quality assurance of volumetric modulated arc therapy: Prediction accuracy and cost-sensitive classification performance

PURPOSE

To evaluate a deep learning model's performance in predicting and classifying patient-specific quality assurance (PSQA) results for volumetric modulated arc therapy (VMAT), aiming to streamline PSQA workflows and reduce the onsite measurement workload.

METHODS

A total of 761 VMAT plans were analyzed using 3D-MResNet to process multileaf collimator images and monitor unit data, with the gamma passing rate (GPR) as the output. Thresholds for the predicted GPR (Th-p) and measured GPR (Th-m) were established to aid in PSQA decision-making, using cost curves and error rates to assess classification performance.

RESULTS

The mean absolute errors of the model for the test set were 1.63 % and 2.38 % at 3 %/2 mm and 2 %/2 mm, respectively. For the classification of the PSQA results, Th-m was 88.3 % at 2 %/2 mm and 93.3 % at 3 %/2 mm. The lowest cost-sensitive error rates of 0.0127 and 0.0925 were obtained when Th-p was set as 91.2 % at 2 %/2 mm and 96.4 % at 3 %/2 mm, respectively. Additionally, the 2 %/2 mm classifier also achieved a lower total expected cost of 0.069 compared with 0.110 for the 3 %/2 mm classifier. The deep learning classifier under the 2 %/2 mm gamma criterion had a sensitivity and specificity of 100 % (10/10) and 83.5 % (167/200), respectively, for the test set.

CONCLUSIONS

The developed 3D-MResNet model can accurately predict and classify PSQA results based on VMAT plans. The introduction of a deep learning model into the PSQA workflow has considerable potential for improving the VMAT PSQA process and reducing workloads.

Improving the performance of deep learning models in predicting and classifying gamma passing rates with discriminative features and a class balancing technique: a retrospective cohort study

BACKGROUND

The purpose of this study was to improve the deep learning (DL) model performance in predicting and classifying IMRT gamma passing rate (GPR) by using input features related to machine parameters and a class balancing technique.

METHODS

A total of 2348 fields from 204 IMRT plans for patients with nasopharyngeal carcinoma were retrospectively collected to form a dataset. Input feature maps, including fluence, leaf gap, leaf speed of both banks, and corresponding errors, were constructed from the dynamic log files. The SHAP framework was employed to compute the impact of each feature on the model output for recursive feature elimination. A series of UNet++ based models were trained on the obtained eight feature sets with three fine-tuning methods including the standard mean squared error (MSE) loss, a re-sampling technique, and a proposed weighted MSE loss (WMSE). Differences in mean absolute error, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were compared between the different models.

RESULTS

The models trained with feature sets including leaf speed and leaf gap features predicted GPR for failed fields more accurately than the other models (F(7, 147) = 5.378, p < 0.001). The WMSE loss had the highest accuracy in predicting GPR for failed fields among the three fine-tuning methods (F(2, 42) = 14.149, p < 0.001), while an opposite trend was observed in predicting GPR for passed fields (F(2, 730) = 9.907, p < 0.001). The WMSE_FS5 model achieved a superior AUC (0.92) and more balanced sensitivity (0.77) and specificity (0.89) compared to the other models.

CONCLUSIONS

Machine parameters can provide discriminative input features for GPR prediction in DL. The novel weighted loss function demonstrates the ability to balance the prediction and classification accuracy between the passed and failed fields. The proposed approach is able to improve the DL model performance in predicting and classifying GPR, and can potentially be integrated into the plan optimization process to generate higher deliverability plans.

TRIAL REGISTRATION

This clinical trial was registered in the Chinese Clinical Trial Registry on March 26th, 2020 (registration number: ChiCTR2000031276). https://clinicaltrials.gov/ct2/show/ChiCTR2000031276.

A TPS integrated machine learning tool for predicting patient-specific quality assurance outcomes in volumetric-modulated arc therapy

PURPOSE

Machine learning (ML) models have been demonstrated to be beneficial for optimizing the workload of patient-specific quality assurance (PSQA). Implementing them in clinical routine frequently requires third-party applications beyond the treatment planning system (TPS), slowing down the workflow. To address this issue, a PSQA outcomes predictive model was carefully selected and validated before being fully integrated into the TPS.

MATERIALS AND METHODS

Nine ML algorithms were evaluated using cross-validation. The learning database was built by calculating complexity metrics (CM) and binarizing PSQA results into "pass"/"fail" classes for 1767 VMAT arcs. The predictive performance was evaluated using area under the ROC curve (AUROC), sensitivity, and specificity. The ML model was integrated into the TPS via a C# script. Script-guided reoptimization impact on PSQA and dosimetric results was evaluated on ten VMAT plans with "fail"-predicted arcs. Workload reduction potential was also assessed.

RESULTS

The selected model exhibited an AUROC of 0.88, with a sensitivity and specificity exceeding 50 % and 90 %, respectively. The script-guided reoptimization of the ten evaluated plans led to an average improvement of 1.4 ± 0.9 percentage points in PSQA results, while preserving the quality of the dose distribution. A yearly savings of about 140 h with the use of the script was estimated.

CONCLUSIONS

The proposed script is a valuable complementary tool for PSQA measurement. It was efficiently integrated into the clinical workflow to enhance PSQA outcomes and reduce PSQA workload by decreasing the risk of failing QA and thereby, the need for repeated replanning and measurements.

Guaranteed performance of individual control chart used in gamma passing rate-based patient-specific quality assurance

PURPOSE

To assess the effect of sampling variability on the performance of individual charts (I-charts) for PSQA and provide a robust and reliable method for unknown PSQA processes.

MATERIALS AND METHODS

A total of 1327 pretreatment PSQAs were analyzed. Different datasets with samples in the range of 20-1000 were used to estimate the lower control limit (LCL). Based on the iterative "Identify-Eliminate-Recalculate" and direct calculation without any outlier filtering procedures, five I-charts methods, namely the Shewhart, quantile, scaled weighted variance (SWV), weighted standard deviation (WSD), and skewness correction (SC) method, were used to compute the LCL. The average run length (ARL₀) and false alarm rate (FAR₀) were calculated to evaluate the performance of LCL.

RESULTS

The ground truth of the values of LCL, FAR₀, and ARL₀ obtained via in-control PSQAs were 92.31%, 0.135%, and 740.7, respectively. Further, for in-control PSQAs, the width of the 95% confidence interval of LCL values for all methods tended to decrease with the increase in sample size. In all sample ranges of in-control PSQAs, only the median LCL and ARL₀ values obtained via WSD and SWV methods were close to the ground truth. For the actual unknown PSQAs, based on the "Identify-Eliminate-Recalculate" procedure, only the median LCL values obtained by the WSD method were closest to the ground truth.

CONCLUSIONS

Sampling variability seriously affected the I-chart performance in PSQA processes, particularly for small samples. For unknown PSQAs, the WSD method based on the implementation of the iterative "Identify-Eliminate-Recalculate" procedure exhibited sufficient robustness and reliability.

An effective and optimized patient-specific QA workload reduction for VMAT plans after MLC-modelling optimization

INTRODUCTION

Many complexity metrics characterize modulated plans. First, this study aimed at identify the optimal complexity metrics to reduce workload associated to patient-specific quality assurance (PSQA) for our equipment and processes. Second, it intended to optimize our MLC modelling to improve measurement and calculation agreement with expectation of further reducing PSQA workload.

METHODS

Correlation and sensitivity at specificity equals to 1 were evaluated for PSQA results and different complexity metrics. Thresholds to stop PSQA were determined. After validation of the optimal complexity metric and threshold for our equipment and process, the MLC modelling was reviewed with a recently published methodology. This method is based on measurements with a Farmer-type ionization chamber of synchronous and asynchronous sweeping gap plans. Effect on the PSQA results and the identified threshold was investigated.

RESULTS

In our center, the most appropriate complexity metric for reducing our PSQA workload was the Modulation Complexity Score for VMAT (MCSv). The optimization of the MLC modelling significantly reduced the number of controlled plans, specifically for one of our two Varian Clinac. Any plan with a MCSv >= 0.34 is treated without PSQA.

CONCLUSION

This study rationalized and reduced our PSQA workload by approximately 30%. It is a continuing work with new TPS, machine or PSQA equipment. It encourages centers to re-evaluate their MLC modelling as well as assess the benefit of complexity metrics to streamline their PSQA workflow. An easier access, at least for reporting, at best for optimizing plans, into the TPS would be beneficial for the community.

Modelling level I Axillary Lymph Nodes depth for Microwave Imaging

PURPOSE

Patient-specific information on the depth of Axillary Lymph Nodes (ALNs) is important for the development of new diagnostic imaging technologies, e.g. Microwave Imaging (MWI), aiming to assess the diagnosis of ALNs during breast cancer staging. Studies about ALNs depth have been presented for treatment planning, but they lack information on sample size and usability of the data to infer the depth of ALNs. The aim of this study was to create a mathematical model that can be used to predict a depth interval where level I ALNs are likely to be located.

METHODS

We extracted biometric features of 98 patients who underwent breast Magnetic Resonance Imaging (MRI) to train two types of regression models. We then tested different combination of features to predict ALNs depth and found the best predictor. The final prediction models were then implemented in an algorithm used for MWI and tested with anthropomorphic phantoms of the axillary region.

RESULTS

Body Mass Index (BMI) was the feature with best performance to predict ALNs depth with coefficient of determination (R²) ranging from 0.49 to 0.55 and Root Mean Squared Error (RMSE) ranging from 0.68 to 0.91 cm. The proposed model showed satisfactory results in microwave images of patients with different BMIs.

CONCLUSIONS

The presented results contribute to the development of reconstruction algorithms for new imaging technologies and to the assessment of ALNs in other medical applications.