•Deep learning can be used to contour organs at risk for numerous patients.•Deep learning outperforms atlas based auto-segmentation in head and neck organs at risk contouring.•Subjective analysis ...indicated no need manual corrections in many deep learning generated contours.•Deep learning auto-contouring has the potential to decrease the clinical burden.
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.
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.
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.
The DLC, trained on a large HN cancer patient cohort, outperformed the ABAS for the majority of HN OARs.
Abstract Purpose The objective of this project was to define consensus guidelines for delineating organs at risk (OARs) for head and neck radiotherapy for routine daily practice and for research ...purposes. Methods Consensus guidelines were formulated based on in-depth discussions of a panel of European, North American, Asian and Australian radiation oncologists. Results Twenty-five OARs in the head and neck region were defined with a concise description of their main anatomic boundaries. The Supplemental material provides an atlas of the consensus guidelines, projected on 1 mm axial slices. The atlas can also be obtained in DICOM-RT format on request. Conclusion Consensus guidelines for head and neck OAR delineation were defined, aiming to decrease interobserver variability among clinicians and radiotherapy centers.
Abstract In the last decade, many efforts have been made to characterize anatomic changes of head and neck organs at risk (OARs) and the dosimetric consequences during radiotherapy. This review was ...undertaken to provide an overview of the magnitude and frequency of these effects, and to investigate whether we could find criteria to identify head and neck cancer patients who may benefit from adaptive radiotherapy (ART). Possible relationships between anatomic and dosimetric changes and outcome were explicitly considered. A literature search according to PRISMA guidelines was performed in MEDLINE and EMBASE for studies concerning anatomic or dosimetric changes of head and neck OARs during radiotherapy. Fifty-one eligible studies were found. The majority of papers reported on parotid gland (PG) anatomic and dosimetric changes. In some patients, PG mean dose differences between planning CT and repeat CT scans up to 10 Gy were reported. In other studies, only minor dosimetric effects (i.e. <1 Gy difference in PG mean dose) were observed as a result of significant anatomic changes. Only a few studies reported on the clinical relevance of anatomic and dosimetric changes in terms of complications or quality of life. Numerous potential selection criteria for anatomic and dosimetric changes during radiotherapy were found and listed. The heterogeneity between studies prevented unambiguous conclusions on how to identify patients who may benefit from ART in head and neck cancer. Potential pre-treatment selection criteria identified from this review include tumour location (nasopharyngeal carcinoma), age, body mass index, planned dose to the parotid glands, the initial parotid gland volume, and the overlap volume of the parotid glands with the target volume. These criteria should be further explored in well-designed and well-powered prospective studies, in which possible relationships between anatomic and dosimetric changes and outcome need to be established.
•Optimal timing of re-planning was suggested for a comprehensive set of OARs.•Generic re-planning timing was suggested considering different combinations of OARs.•Dmean increase due to anatomical ...changes were most frequently > 3 Gy in Supraglottic.•The 3rd week is suggested as the optimal timing of re-planning for 10 OARs.•A maximum of 3 re-plannings can guarantee Dmean increase below 3 Gy for all OARs.
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.
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.
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.
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.
•First time comparison of auto- and human segmentation accuracy using dose and NTCP.•DLC performs better than contour warping by DIR for majority of head and neck OARs.•Human segmentation of parotid ...glands is needed during adaptive radiotherapy (ART)•Semi auto-segmentation is feasible for accurate dose and NTCP calculation in ART.
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).
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|).
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.
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.
•Quantitative evaluation showed differences between populations for several organs.•Qualitative evaluation showed that no bias was found regarding patient origin.•Observers of different origins ...perceive deep learning autocontours differently.
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.
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.
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.
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.
Consistent delineation of patient anatomy becomes increasingly important with the growing use of highly conformal and adaptive radiotherapy techniques. This study investigates the magnitude and 3D ...localization of interobserver variability of organs at risk (OARs) in the head and neck area with application of delineation guidelines, to establish measures to reduce current redundant variability in delineation practice.
Interobserver variability among five experienced radiation oncologists was studied in a set of 12 head and neck patient CT scans for the spinal cord, parotid and submandibular glands, thyroid cartilage, and glottic larynx. For all OARs, three endpoints were calculated: the Intraclass Correlation Coefficient (ICC), the Concordance Index (CI) and a 3D measure of variation (3D SD).
All endpoints showed largest interobserver variability for the glottic larynx (ICC = 0.27, mean CI = 0.37 and 3D SD = 3.9 mm). Better agreement in delineations was observed for the other OARs (range, ICC = 0.32-0.83, mean CI = 0.64-0.71 and 3D SD = 0.9-2.6 mm). Cranial, caudal, and medial regions of the OARs showed largest variations. All endpoints provided support for improvement of delineation practice.
Variation in delineation is traced to several regional causes. Measures to reduce this variation can be: (1) guideline development, (2) joint delineation review sessions and (3) application of multimodality imaging. Improvement of delineation practice is needed to standardize patient treatments.
Abstract Background and purpose Current models for the prediction of late patient-rated moderate-to-severe xerostomia (XER12m ) and sticky saliva (STIC12m ) after radiotherapy are based on ...dose-volume parameters and baseline xerostomia (XERbase ) or sticky saliva (STICbase ) scores. The purpose is to improve prediction of XER12m and STIC12m with patient-specific characteristics, based on CT image biomarkers (IBMs). Methods Planning CT-scans and patient-rated outcome measures were prospectively collected for 249 head and neck cancer patients treated with definitive radiotherapy with or without systemic treatment. The potential IBMs represent geometric, CT intensity and textural characteristics of the parotid and submandibular glands. Lasso regularisation was used to create multivariable logistic regression models, which were internally validated by bootstrapping. Results The prediction of XER12m could be improved significantly by adding the IBM “Short Run Emphasis” (SRE), which quantifies heterogeneity of parotid tissue, to a model with mean contra-lateral parotid gland dose and XERbase . For STIC12m , the IBM maximum CT intensity of the submandibular gland was selected in addition to STICbase and mean dose to submandibular glands. Conclusion Prediction of XER12m and STIC12m was improved by including IBMs representing heterogeneity and density of the salivary glands, respectively. These IBMs could guide additional research to the patient-specific response of healthy tissue to radiation dose.
•Head & neck cancer patients can be selected for ART in the first treatment week.•Mean dose change to OARs in week 1 and 2 is predictive for overall BIOΔNTCP.•Comprehensive evaluation of planned vs. ...actual NTCP in 110 patients.•The proposed strategy could classify all true candidates regarding late toxicity.•The strategy could correctly spare 88.2% patients from ART regarding late toxicity.
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.
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.
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).
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.
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.
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.
In total, 213 medical physicists from 202 radiation oncology centres were included in the analysis. Sixty-nine percent (147) 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.
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.