Patterns of practice of image guided particle therapy for cranio-spinal irradiation: A site specific multi-institutional survey of European Particle Therapy Network.

PURPOSE: To investigate the current practice patterns in image-guided particle therapy (IGPT) for cranio-spinal irradiation (CSI). METHODS: A multi-institutional survey was distributed to European particle therapy centres to analyse all aspects of IGPT. Based on the survey results, a Delphi consensus analysis was developed to define minimum requirements and optimal workflow for clinical practice. The centres participating in the institutional survey were invited to join the Delphi process. RESULTS: Eleven centres participated in the survey. Imaging for treatment planning was rather similar among the centres with Computed Tomography (CT) being the main modality. For positioning verification, 2D IGPT was more commonly used than 3D IGPT. Two centres performed routinely imaging for plan adaptation, by the rest ad hoc. Eight centres participated in the Delphi consensus analysis. The full consensus was reached on the use of CT imaging without contrast for treatment planning and the role of magnetic resonance imaging (MRI) in target and organs-at-risk delineation. There was an agreement on the necessity to perform patient position verification and correction before each isocentre. The most important outcome was the clear need for standardization and harmonization of the workflow. CONCLUSION: There were differences in CSI IGPT clinical practice among the European particle therapy centres. Moreover, the optimal workflow as identified by experts was not yet reached. There is a strong need for consensus guidelines. The state-of-the-art imaging technology and protocols need to be implemented into clinical practice to improve the quality of IGPT for CSI.

Proton PBS Planning Techniques, Robustness Evaluation, and OAR Sparing for the Whole-Brain Part of Craniospinal Axis Irradiation.

Proton therapy is a promising modality for craniospinal irradiation (CSI), offering dosimetric advantages over conventional treatments. While significant attention has been paid to spine fields, for the brain fields, only dose reduction to the lens of the eye has been reported. Hence, the objective of this study is to assess the potential gains and feasibility of adopting different treatment planning techniques for the entire brain within the CSI target. To this end, eight previously treated CSI patients underwent retrospective replanning using various techniques: (1) intensity modulated proton therapy (IMPT) optimization, (2) the modification/addition of field directions, and (3) the pre-optimization removal of superficially placed spots. The target coverage robustness was evaluated and dose comparisons for lenses, cochleae, and scalp were conducted, considering potential biological dose increases. The target coverage robustness was maintained across all plans, with minor reductions when superficial spot removal was utilized. Single- and multifield optimization showed comparable target coverage robustness and organ-at-risk sparing. A significant scalp sparing was achieved in adults but only limited in pediatric cases. Superficial spot removal contributed to scalp V30 Gy reduction at the expense of lower coverage robustness in specific cases. Lens sparing benefits from multiple field directions, while cochlear sparing remains impractical. Based on the results, all investigated plan types are deemed clinically adoptable.

Optimal timing of re-planning for head and neck adaptive radiotherapy.

BACKGROUND AND PURPOSE: Adaptive radiotherapy (ART) relies on re-planning to correct treatment variations, but the optimal timing of re-planning to account for dose changes in head and neck organs at risk (OARs) is still under investigation. We aimed to find out the optimal timing of re-planning in head and neck ART. MATERIALS AND METHODS: A total of 110 head and neck cancer patients were retrospectively enrolled. A semi auto-segmentation method was applied to obtain the weekly mean dose (Dmean) to OARs. The K-nearest-neighbour method was used for missing data imputation of weekly Dmean. A dose deviation map was built using the planning Dmean and weekly Dmean values and then used to simulate different ART scenarios consisting of 1 to 6 re-plannings. The difference between accumulated Dmean and planning Dmean before re-planning (ΔDmean_acc_noART) and after re-planning (ΔDmean_acc_ART) were evaluated and compared. RESULTS: Among all the OARs, supraglottic showed the largest ΔDmean_acc_noART (1.23 ± 3.13 Gy) and most cases of ΔDmean_acc_noART > 3 Gy (26 patients). The 3rd week is suggested in the optimal timing of re-planning for 10 OARs. For all the organs except arytenoid, 2 re-plannings were able to guarantee the ΔDmean_acc_ART below 3 Gy while the average |ΔDmean_acc_ART| was below 1 Gy. ART scenarios of 2_4, 3_4, 3_5 (week of re-planning separated with "_") were able to guarantee ΔDmean_acc_ART of 99 % of patients below 3 Gy simultaneously for 19 OARs. CONCLUSIONS: The optimal timing of re-planning was suggested for different organs at risk in head and neck adaptive radiotherapy. Generic scenarios of timing and frequency for re-planning can be applied to guarantee the increase of accumulated mean dose within 3 Gy simultaneously for multiple organs.

Pretrial Quality Assurance for Hypofractionated Salvage Radiation Therapy After Prostatectomy in the Multi-Institutional PERYTON-trial.

Purpose: The PERYTON trial is a multicenter randomized controlled trial that will investigate whether the treatment outcome of salvage external beam radiation therapy (sEBRT) will be improved with hypofractionated radiation therapy. A pretrial quality assurance (QA) program was undertaken to ensure protocol compliance within the PERYTON trial and to assess variation in sEBRT treatment protocols between the participating centers. Methods and Materials: Completion of the QA program was mandatory for each participating center (N = 8) to start patient inclusion. The pretrial QA program included (1) a questionnaire on the center-specific sEBRT protocol, (2) a delineation exercise of the clinical target volume (CTV) and organs at risk, and (3) a treatment planning exercise. All contours were analyzed using the pairwise dice similarity coefficient (DSC) and the 50th and 95th percentile Hausdorff distance (HD50 and HD95, respectively). The submitted treatment plans were reviewed for protocol compliance. Results: The results of the questionnaire showed that high-quality, state-of-the-art radiation therapy techniques were used in the participating centers and identified variations of the sEBRT protocols used concerning the position verification and preparation techniques. The submitted CTVs showed significant variation, with a range in volume of 29 cm3 to 167 cm3, a mean pairwise DSC of 0.52, and a mean HD50 and HD95 of 2.3 mm and 24.4 mm, respectively. Only in 1 center the treatment plan required adaptation before meeting all constraints of the PERYTON protocol. Conclusions: The pretrial QA of the PERYTON trial demonstrated that high-quality, but variable, radiation techniques were used in the 8 participating centers. The treatment planning exercise confirmed that the dose constraints of the PERYTON protocol were feasible for all participating centers. The observed variation in CTV delineation led to agreement on a new (image-based) delineation guideline to be used by all participating centers within the PERYTON trial.

Prostate-Specific Membrane Antigen Positron Emission Tomography/Computed Tomography-Based Clinical Target Volume Delineation Guideline for Postprostatectomy Salvage Radiation Therapy: The PERYTON Guideline.

PURPOSE: Prostate-specific membrane antigen positron emission tomography/computed tomography (PSMA PET/CT) scan is the standard imaging procedure for biochemical recurrent prostate cancer postprostatectomy because of its high detection rate at low serum prostate-specific antigen levels. However, existing guidelines for clinical target volume (CTV) in prostate bed salvage external beam radiation therapy (sEBRT) are primarily based on experience-based clinical consensus and have been validated using conventional imaging modalities. Therefore, this study aimed to optimize CTV definition in sEBRT by using PSMA PET/CT-detected local recurrences (LRs). METHODS AND MATERIALS: Patients with suspected LR on PSMA PET/CT postprostatectomy were retrospectively enrolled in 9 Dutch centers. Anonymized scans were centrally reviewed by an expert nuclear medicine physician. Each boundary of the CTV guideline from the Groupe Francophone de Radiothérapie en Urologie (GFRU) was evaluated and adapted to improve the accuracy and coverage of the area at risk of LR (CTV) on PSMA PET/CT. The proposed CTV adaptation was discussed with the radiation oncologists of the participating centers, and final consensus was reached. To assess reproducibility, the participating centers were asked to delineate 3 new cases according to the new PERYTON-CTV, and the submitted contours were evaluated using the Dice similarity coefficient (DSC). RESULTS: After central review, 93 LRs were identified on 83 PSMA PET/CTs. The proposed CTV definition improved the coverage of PSMA PET/CT-detected LRs from 67% to 96% compared with the GFRU-CTV, while reducing the GFRU-CTV by 25%. The new CTV was highly reproducible, with a mean DSC of 0.82 (range, 0.81-0.83). CONCLUSIONS: This study contributes to the optimization of CTV definition in postprostatectomy sEBRT by using the pattern of LR detected on PSMA PET/CT. The PERYTON-CTV is highly reproducible across the participating centers and ensures coverage of 96% LRs while reducing the GFRU-CTV by 25%.

Dose evaluation of inter- and intra-fraction prostate motion in extremely hypofractionated intensity-modulated proton therapy for prostate cancer.

Inter- and intra-fractional prostate motion can deteriorate the dose distribution in extremely hypofractionated intensity-modulated proton therapy. We used verification CTs and prostate motion data calculated from 1024 intra-fractional prostate motion records to develop a voxel-wise based 4-dimensional method, which had a time resolution of 1 s, to assess the dose impact of prostate motion. An example of 100 fractional simulations revealed that motion had minimal impact on planning dose, the accumulated dose in 95 % of the scenarios fulfilled the clinical goals for target coverage (D95 > 37.5 Gy). This method can serve as a complementary measure in clinical setting to guarantee plan quality.

Simple immobilization for stereotactic radiotherapy aimed at pelvic metastases.

Stereotactic body radiotherapy (SBRT) is increasingly applied for pelvic oligometastases of prostate cancer, and currently no simple immobilization method is available for cone beam computed tomography (CBCT)-guided treatment. We assessed patient set-up and intrafraction motion using simple immobilization during CBCT-guided pelvic SBRT. Forty patients were immobilized with basic arm- head- and knee support and either a thermoplastic cushion or a foam cushion. Analysis of 454 CBCTs showed mean intrafraction translation <3.0 mm in 94% of fractions and mean intrafraction rotation <1.5° in 95% of fractions. Therefore, simple immobilization ensured stable patient positioning during CBCT-guided pelvic SBRT.

Effects of proton therapy on regional [18F]FDG uptake in non-tumor brain regions of patients treated for head and neck cancer.

Background and purpose: Previous pre-clinical research using [18F]FDG-PET has shown that whole-brain photon-based radiotherapy can affect brain glucose metabolism. This study, aimed to investigate how these findings translate into regional changes in brain [18F]FDG uptake in patients with head and neck cancer treated with intensity-modulated proton therapy (IMPT). Materials and methods: Twenty-three head and neck cancer patients treated with IMPT and available [18F]FDG scans before and at 3 months follow-up were retrospectively evaluated. Regional assessment of the [18F]FDG standardized uptake value (SUV) parameters and radiation dose in the left (L) and right (R) hippocampi, L and R occipital lobes, cerebellum, temporal lobe, L and R parietal lobes and frontal lobe were evaluated to understand the relationship between regional changes in SUV metrics and radiation dose. Results: Three months after IMPT, [18F]FDG brain uptake calculated using SUVmean and SUVmax, was significantly higher than that before IMPT. The absolute SUVmean after IMPT was significantly higher than before IMPT in seven regions of the brain (p ≤ 0.01), except for the R (p = 0.11) and L (p = 0.15) hippocampi. Absolute and relative changes were variably correlated with the regional maximum and mean doses received in most of the brain regions. Conclusion: Our findings suggest that 3 months after completion of IMPT for head and neck cancer, significant increases in the uptake of [18F]FDG (reflected by SUVmean and SUVmax) can be detected in several individual key brain regions, and when evaluated jointly, it shows a negative correlation with the mean dose. Future studies are needed to assess whether and how these results could be used for the early identification of patients at risk for adverse cognitive effects of radiation doses in non-tumor tissues.

An investigation into the risk of population bias in deep learning autocontouring.

BACKGROUND AND PURPOSE: To date, data used in the development of Deep Learning-based automatic contouring (DLC) algorithms have been largely sourced from single geographic populations. This study aimed to evaluate the risk of population-based bias by determining whether the performance of an autocontouring system is impacted by geographic population. MATERIALS AND METHODS: 80 Head Neck CT deidentified scans were collected from four clinics in Europe (n = 2) and Asia (n = 2). A single observer manually delineated 16 organs-at-risk in each. Subsequently, the data was contoured using a DLC solution, and trained using single institution (European) data. Autocontours were compared to manual delineations using quantitative measures. A Kruskal-Wallis test was used to test for any difference between populations. Clinical acceptability of automatic and manual contours to observers from each participating institution was assessed using a blinded subjective evaluation. RESULTS: Seven organs showed a significant difference in volume between groups. Four organs showed statistical differences in quantitative similarity measures. The qualitative test showed greater variation in acceptance of contouring between observers than between data from different origins, with greater acceptance by the South Korean observers. CONCLUSION: Much of the statistical difference in quantitative performance could be explained by the difference in organ volume impacting the contour similarity measures and the small sample size. However, the qualitative assessment suggests that observer perception bias has a greater impact on the apparent clinical acceptability than quantitatively observed differences. This investigation of potential geographic bias should extend to more patients, populations, and anatomical regions in the future.

An efficient strategy to select head and neck cancer patients for adaptive radiotherapy.

BACKGROUND AND PURPOSE: Adaptive radiotherapy (ART) is workload intensive but only benefits a subgroup of patients. We aimed to develop an efficient strategy to select candidates for ART in the first two weeks of head and neck cancer (HNC) radiotherapy. MATERIALS AND METHODS: This study retrospectively enrolled 110 HNC patients who underwent modern photon radiotherapy with at least 5 weekly in-treatment re-scan CTs. A semi auto-segmentation method was applied to obtain the weekly mean dose (Dmean) to OARs. A comprehensive NTCP-profile was applied to obtain NTCP's. The difference between planning and actual values of Dmean (ΔDmean) and dichotomized difference of clinical relevance (BIOΔNTCP) were used for modelling to determine the cut-off maximum ΔDmean of OARs in week 1 and 2 (maxΔDmean_1 and maxΔDmean_2). Four strategies to select candidates for ART, using cut-off maxΔDmean were compared. RESULTS: The Spearman's rank correlation test showed significant positive correlation between maxΔDmean and BIOΔNTCP (p-value <0.001). For major BIOΔNTCP (>5%) of acute and late toxicity, 10.9% and 4.5% of the patients were true candidates for ART. Strategy C using both cut-off maxΔDmean_1 (3.01 and 5.14 Gy) and cut-off maxΔDmean_2 (3.41 and 5.30 Gy) showed the best sensitivity, specificity, positive and negative predictive values (0.92, 0.82, 0.38, 0.99 for acute toxicity and 1.00, 0.92, 0.38, 1.00 for late toxicity, respectively). CONCLUSIONS: We propose an efficient selection strategy for ART that is able to classify the subgroup of patients with >5% BIOΔNTCP for late toxicity using imaging in the first two treatment weeks.

Accurate segmentation of head and neck radiotherapy CT scans with 3D CNNs: consistency is key.

Objective.Automatic segmentation of organs-at-risk in radiotherapy planning computed tomography (CT) scans using convolutional neural networks (CNNs) is an active research area. Very large datasets are usually required to train such CNN models. In radiotherapy, large, high-quality datasets are scarce and combining data from several sources can reduce the consistency of training segmentations. It is therefore important to understand the impact of training data quality on the performance of auto-segmentation models for radiotherapy.Approach.In this study, we took an existing 3D CNN architecture for head and neck CT auto-segmentation and compare the performance of models trained with a small, well-curated dataset (n= 34) and then a far larger dataset (n= 185) containing less consistent training segmentations. We performed 5-fold cross-validations in each dataset and tested segmentation performance using the 95th percentile Hausdorff distance and mean distance-to-agreement metrics. Finally, we validated the generalisability of our models with an external cohort of patient data (n= 12) with five expert annotators.Main results.The models trained with a large dataset were greatly outperformed by models (of identical architecture) trained with a smaller, but higher consistency set of training samples. Our models trained with a small dataset produce segmentations of similar accuracy as expert human observers and generalised well to new data, performing within inter-observer variation.Significance.We empirically demonstrate the importance of highly consistent training samples when training a 3D auto-segmentation model for use in radiotherapy. Crucially, it is the consistency of the training segmentations which had a greater impact on model performance rather than the size of the dataset used.

Quality Assurance for AI-Based Applications in Radiation Therapy.

Recent advancements in artificial intelligence (AI) in the domain of radiation therapy (RT) and their integration into modern software-based systems raise new challenges to the profession of medical physics experts. These AI algorithms are typically data-driven, may be continuously evolving, and their behavior has a degree of (acceptable) uncertainty due to inherent noise in training data and the substantial number of parameters that are used in the algorithms. These characteristics request adaptive, and new comprehensive quality assurance (QA) approaches to guarantee the individual patient treatment quality during AI algorithm development and subsequent deployment in a clinical RT environment. However, the QA for AI-based systems is an emerging area, which has not been intensively explored and requires interactive collaborations between medical doctors, medical physics experts, and commercial/research AI institutions. This article summarizes the current QA methodologies for AI modules of every subdomain in RT with further focus on persistent shortcomings and upcoming key challenges and perspectives.

Clinical relevance of the radiation dose bath in lower grade glioma, a cross-sectional pilot study on neurocognitive and radiological outcome.

Aim: To investigate the clinical relevance of the radiotherapy (RT) dose bath in patients treated for lower grade glioma (LGG). Methods: Patients (n = 17) treated with RT for LGG were assessed with neurocognitive function (NCF) tests and structural Magnetic Resonance Imaging (MRI) and categorized in subgroups based on tumour lateralisation. RT dose, volumetric results and cerebral microbleed (CMB) number were extracted for contralateral cerebrum, contralateral hippocampus, and cerebellum. The RT clinical target volume (CTV) was included in the analysis as a surrogate for focal tumour and other treatment effects. The relationships between RT dose, CTV, NCF and radiological outcome were analysed per subgroup. Results: The subgroup with left-sided tumours (n = 10) performed significantly lower on verbal tests. The RT dose to the right cerebrum, as well as CTV, were related to poorer performance on tests for processing speed, attention, and visuospatial abilities, and more CMB.In the subgroup with right-sided tumours (n = 7), RT dose in the left cerebrum was related to lower verbal memory performance, (immediate and delayed recall, r = -0.821, p = 0.023 and r = -0.937, p = 0.002, respectively), and RT dose to the left hippocampus was related to hippocampal volume (r = -0.857, p = 0.014), without correlation between CTV and NCF. Conclusion: By using a novel approach, we were able to investigate the clinical relevance of the RT dose bath in patients with LGG more specifically. We used combined MRI-derived and NCF outcome measures to assess radiation-induced brain damage, and observed potential RT effects on the left-sided brain resulting in lower verbal memory performance and hippocampus volume.

A novel semi auto-segmentation method for accurate dose and NTCP evaluation in adaptive head and neck radiotherapy.

BACKGROUND AND PURPOSE: Accurate segmentation of organs-at-risk (OARs) is crucial but tedious and time-consuming in adaptive radiotherapy (ART). The purpose of this work was to automate head and neck OAR-segmentation on repeat CT (rCT) by an optimal combination of human and auto-segmentation for accurate prediction of Normal Tissue Complication Probability (NTCP). MATERIALS AND METHODS: Human segmentation (HS) of 3 observers, deformable image registration (DIR) based contour propagation and deep learning contouring (DLC) were carried out to segment 15 OARs on 15 rCTs. The original treatment plan was re-calculated on rCT to obtain mean dose (Dmean) and consequent NTCP-predictions. The average Dmean and NTCP-predictions of the three observers were referred to as the gold standard to calculate the absolute difference of Dmean and NTCP-predictions (|ΔDmean| and |ΔNTCP|). RESULTS: The average |ΔDmean| of parotid glands in HS was 1.40 Gy, lower than that obtained with DIR and DLC (3.64 Gy, p < 0.001 and 3.72 Gy, p < 0.001, respectively). DLC showed the highest |ΔDmean| in middle Pharyngeal Constrictor Muscle (PCM) (5.13 Gy, p = 0.01). DIR showed second highest |ΔDmean| in the cricopharyngeal inlet (2.85 Gy, p = 0.01). The semi auto-segmentation (SAS) adopted HS, DIR and DLC for segmentation of parotid glands, PCM and all other OARs, respectively. The 90th percentile |ΔNTCP|was 2.19%, 2.24%, 1.10% and 1.50% for DIR, DLC, HS and SAS respectively. CONCLUSIONS: Human segmentation of the parotid glands remains necessary for accurate interpretation of mean dose and NTCP during ART. Proposed semi auto-segmentation allows NTCP-predictions within 1.5% accuracy for 90% of the cases.

Validation of separate multi-atlases for auto segmentation of cardiac substructures in CT-scans acquired in deep inspiration breath hold and free breathing.

BACKGROUND AND PURPOSE: Developing NTCP-models for cardiac complications after breast cancer (BC) radiotherapy requires cardiac dose-volume parameters for many patients. These can be obtained by using multi-atlas based automatic segmentation (MABAS) of cardiac structures in planning CT scans. We investigated the relevance of separate multi-atlases for deep inspiration breath hold (DIBH) and free breathing (FB) CT scans. MATERIALS AND METHODS: BC patients scanned in DIBH (n = 10) and in FB (n = 20) were selected to create separate multi-atlases consisting of expert panel delineations of the whole heart, atria and ventricles. The accuracy of atlas-generated contours was validated with expert delineations in independent datasets (n = 10 for DIBH and FB) and reported as Dice coefficients, contour distances and dose-volume differences in relation to interobserver variability of manual contours. Dependency of MABAS contouring accuracy on breathing technique was assessed by validation of a FB atlas in DIBH patients and vice versa (cross-validation). RESULTS: For all structures the FB and DIBH atlases resulted in Dice coefficients with their respective reference contours ≥ 0.8 and average contour distances ≤ 2 mm smaller than slice thickness of (CTs). No significant differences were found for dose-volume parameters in volumes receiving relevant dose levels (WH, LV and RV). Accuracy of the DIBH atlas was at least similar to, and for the ventricles better than, the interobserver variation in manual delineation. Cross-validation between breathing techniques showed a reduced MABAS performance. CONCLUSION: Multi-atlas accuracy was at least similar to interobserver delineation variation. Separate atlases for scans made in DIBH and FB could benefit atlas performance because accuracy depends on breathing technique.

Inter-fraction motion robustness and organ sparing potential of proton therapy for cervical cancer.

PURPOSE: Large-field photon radiotherapy is current standard in the treatment of cervical cancer patients. However, with the increasing availability of Pencil Beam Scanning Proton Therapy (PBS-PT) and robust treatment planning techniques, protons may have significant advantages for cervical cancer patients in the reduction of toxicity. In this study, PBS-PT and photon Volumetric Modulated Arc Therapy (VMAT) were compared, examining target coverage and organ at risk (OAR) dose, taking inter- and intra-fraction motion into account. MATERIALS AND METHODS: Twelve cervical cancer patients were included in this in-silico planning study. In all cases, a planning CT scan, five weekly repeat CT scans (reCTs) and an additional reCT 10 min after the first reCT were available. Two-arc VMAT and robustly optimised two- and four-field (2F and 4F) PBS-PT plans were robustly evaluated on planCTs and reCTs using set-up and range uncertainty. Nominal OAR doses and voxel-wise minimum target coverage robustness were compared. RESULTS: Average voxel-wise minimum accumulated doses for pelvic target structures over all patients were adequate for both photon and proton treatment techniques (D98 > 95%, [91.7-99.3%]). Average accumulated dose of the para-aortic region was lower than the required 95%, D98 > 94.4% [91.1-98.2%]. With PBS-PT 4F, dose to all OARs was significantly lower than with VMAT. Major differences were observed for mean bowel bag V15Gy: 60% [39-70%] for VMAT vs 30% [10-52%] and 32% [9-54%] for PBS-PT 2F and 4F and for mean bone marrow V10Gy: 88% [82-97%] for VMAT vs 66% [60-73%] and 67% [60-75%] for PBS-PT 2F and 4F. CONCLUSION: Robustly optimised PBS-PT for cervical cancer patients shows equivalent target robustness against inter- and intra-fraction variability compared to VMAT, and offers significantly better OAR sparing.

External validation of deep learning-based contouring of head and neck organs at risk.

BACKGROUND AND PURPOSE: Head and neck (HN) radiotherapy can benefit from automatic delineation of tumor and surrounding organs because of the complex anatomy and the regular need for adaptation. The aim of this study was to assess the performance of a commercially available deep learning contouring (DLC) model on an external validation set. MATERIALS AND METHODS: The CT-based DLC model, trained at the University Medical Center Groningen (UMCG), was applied to an independent set of 58 patients from the Radboud University Medical Center (RUMC). DLC results were compared to the RUMC manual reference using the Dice similarity coefficient (DSC) and 95th percentile of Hausdorff distance (HD95). Craniocaudal spatial information was added by calculating binned measures. In addition, a qualitative evaluation compared the acceptance of manual and DLC contours in both groups of observers. RESULTS: Good correspondence was shown for the mandible (DSC 0.90; HD95 3.6 mm). Performance was reasonable for the glandular OARs, brainstem and oral cavity (DSC 0.78-0.85, HD95 3.7-7.3 mm). The other aerodigestive tract OARs showed only moderate agreement (DSC 0.53-0.65, HD95 around 9 mm). The binned measures displayed the largest deviations caudally and/or cranially. CONCLUSIONS: This study demonstrates that the DLC model can provide a reasonable starting point for delineation when applied to an independent patient cohort. The qualitative evaluation did not reveal large differences in the interpretation of contouring guidelines between RUMC and UMCG observers.

Assessment of manual adjustment performed in clinical practice following deep learning contouring for head and neck organs at risk in radiotherapy.

BACKGROUND AND PURPOSE: Auto-contouring performance has been widely studied in development and commissioning studies in radiotherapy, and its impact on clinical workflow assessed in that context. This study aimed to evaluate the manual adjustment of auto-contouring in routine clinical practice and to identify improvements regarding the auto-contouring model and clinical user interaction, to improve the efficiency of auto-contouring. MATERIALS AND METHODS: A total of 103 clinical head and neck cancer cases, contoured using a commercial deep-learning contouring system and subsequently checked and edited for clinical use were retrospectively taken from clinical data over a twelve-month period (April 2019-April 2020). The amount of adjustment performed was calculated, and all cases were registered to a common reference frame for assessment purposes. The median, 10th and 90th percentile of adjustment were calculated and displayed using 3D renderings of structures to visually assess systematic and random adjustment. Results were also compared to inter-observer variation reported previously. Assessment was performed for both the whole structures and for regional sub-structures, and according to the radiation therapy technologist (RTT) who edited the contour. RESULTS: The median amount of adjustment was low for all structures (<2 mm), although large local adjustment was observed for some structures. The median was systematically greater or equal to zero, indicating that the auto-contouring tends to under-segment the desired contour. CONCLUSION: Auto-contouring performance assessment in routine clinical practice has identified systematic improvements required technically, but also highlighted the need for continued RTT training to ensure adherence to guidelines.

Machine learning applications in radiation oncology: Current use and needs to support clinical implementation.

BACKGROUND AND PURPOSE: The use of artificial intelligence (AI)/ machine learning (ML) applications in radiation oncology is emerging, however no clear guidelines on commissioning of ML-based applications exist. The purpose of this study was therefore to investigate the current use and needs to support implementation of ML-based applications in routine clinical practice. MATERIALS AND METHODS: A survey was conducted among medical physicists in radiation oncology, consisting of four parts: clinical applications (1), model training, acceptance and commissioning (2), quality assurance (QA) in clinical practice and General Data Protection Regulation (GDPR) (3), and need for education and guidelines (4). Survey answers of medical physicists of the same radiation oncology centre were treated as a separate unique responder in case reporting on different AI applications. RESULTS: In total, 213 medical physicists from 202 radiation oncology centres were included in the analysis. Sixty-nine percent (1 4 7) was using (37%) or preparing (32%) to use ML in clinic, mostly for contouring and treatment planning. In 86%, human observers were still involved in daily clinical use for quality check of the output of the ML algorithm. Knowledge on ethics, legislation and data sharing was limited and scattered among responders. Besides the need for (implementation) guidelines, training of medical physicists and larger databases containing multicentre data was found to be the top priority to accommodate the further introduction of ML in clinical practice. CONCLUSION: The results of this survey indicated the need for education and guidelines on the implementation and quality assurance of ML-based applications to benefit clinical introduction.

Improving automatic delineation for head and neck organs at risk by Deep Learning Contouring.

INTRODUCTION: Adequate head and neck (HN) organ-at-risk (OAR) delineation is crucial for HN radiotherapy and for investigating the relationships between radiation dose to OARs and radiation-induced side effects. The automatic contouring algorithms that are currently in clinical use, such as atlas-based contouring (ABAS), leave room for improvement. The aim of this study was to use a comprehensive evaluation methodology to investigate the performance of HN OAR auto-contouring when using deep learning contouring (DLC), compared to ABAS. METHODS: The DLC neural network was trained on 589 HN cancer patients. DLC was compared to ABAS by providing each method with an independent validation cohort of 104 patients, which had also been manually contoured. For each of the 22 OAR contours - glandular, upper digestive tract and central nervous system (CNS)-related structures - the dice similarity coefficient (DICE), and absolute mean and max dose differences (|Δmean-dose| and |Δmax-dose|) performance measures were obtained. For a subset of 7 OARs, an evaluation of contouring time, inter-observer variation and subjective judgement was performed. RESULTS: DLC resulted in equal or significantly improved quantitative performance measures in 19 out of 22 OARs, compared to the ABAS (DICE/|Δmean dose|/|Δmax dose|: 0.59/4.2/4.1 Gy (ABAS); 0.74/1.1/0.8 Gy (DLC)). The improvements were mainly for the glandular and upper digestive tract OARs. DLC significantly reduced the delineation time for the inexperienced observer. The subjective evaluation showed that DLC contours were more often preferable to the ABAS contours overall, were considered to be more precise, and more often confused with manual contours. Manual contours still outperformed both DLC and ABAS; however, DLC results were within or bordering the inter-observer variability for the manual edited contours in this cohort. CONCLUSION: The DLC, trained on a large HN cancer patient cohort, outperformed the ABAS for the majority of HN OARs.