Cardio-pulmonary substructure segmentation of radiotherapy computed tomography images using convolutional neural networks for clinical outcomes analysis

Dosimetric Impact of Delineation and Motion Uncertainties on the Heart and Substructures in Lung Cancer Radiotherapy

AIMS

Delineation variations and organ motion produce difficult-to-quantify uncertainties in planned radiation doses to targets and organs at risk. Similar to manual contouring, most automatic segmentation tools generate single delineations per structure; however, this does not indicate the range of clinically acceptable delineations. This study develops a method to generate a range of automatic cardiac structure segmentations, incorporating motion and delineation uncertainty, and evaluates the dosimetric impact in lung cancer.

MATERIALS AND METHODS

Eighteen cardiac structures were delineated using a locally developed auto-segmentation tool. It was applied to lung cancer planning CTs for 27 curative (planned dose ≥50 Gy) cases, and delineation variations were estimated by using ten mapping-atlases to provide separate substructure segmentations. Motion-related cardiac segmentation variations were estimated by auto-contouring structures on ten respiratory phases for 9/27 cases that had 4D-planning CTs. Dose volume histograms (DVHs) incorporating these variations were generated for comparison.

RESULTS

Variations in mean doses (Dmean), defined as the range in values across ten feasible auto-segmentations, were calculated for each cardiac substructure. Over the study cohort the median variations for delineation uncertainty and motion were 2.20-11.09 Gy and 0.72-4.06 Gy, respectively. As relative values, variations in Dmean were between 18.7%-65.3% and 7.8%-32.5% for delineation uncertainty and motion, respectively. Doses vary depending on the individual planned dose distribution, not simply on segmentation differences, with larger dose variations to cardiac structures lying within areas of steep dose gradient.

CONCLUSION

Radiotherapy dose uncertainties from delineation variations and respiratory-related heart motion were quantified using a cardiac substructure automatic segmentation tool. This predicts the 'dose range' where doses to structures are most likely to fall, rather than single DVH curves. This enables consideration of these uncertainties in cardiotoxicity research and for future plan optimisation. The tool was designed for cardiac structures, but similar methods are potentially applicable to other OARs.

Deep learning-based automatic segmentation of cardiac substructures for lung cancers

PURPOSE

Accurate and comprehensive segmentation of cardiac substructures is crucial for minimizing the risk of radiation-induced heart disease in lung cancer radiotherapy. We sought to develop and validate deep learning-based auto-segmentation models for cardiac substructures.

MATERIALS AND METHODS

Nineteen cardiac substructures (whole heart, 4 heart chambers, 6 great vessels, 4 valves, and 4 coronary arteries) in 100 patients treated for non-small cell lung cancer were manually delineated by two radiation oncologists. The valves and coronary arteries were delineated as planning risk volumes. An nnU-Net auto-segmentation model was trained, validated, and tested on this dataset with a split ratio of 75:5:20. The auto-segmented contours were evaluated by comparing them with manually drawn contours in terms of Dice similarity coefficient (DSC) and dose metrics extracted from clinical plans. An independent dataset of 42 patients was used for subjective evaluation of the auto-segmentation model by 4 physicians.

RESULTS

The average DSCs were 0.95 (+/- 0.01) for the whole heart, 0.91 (+/- 0.02) for 4 chambers, 0.86 (+/- 0.09) for 6 great vessels, 0.81 (+/- 0.09) for 4 valves, and 0.60 (+/- 0.14) for 4 coronary arteries. The average absolute errors in mean/max doses to all substructures were 1.04 (+/- 1.99) Gy and 2.20 (+/- 4.37) Gy. The subjective evaluation revealed that 94% of the auto-segmented contours were clinically acceptable.

CONCLUSION

We demonstrated the effectiveness of our nnU-Net model for delineating cardiac substructures, including coronary arteries. Our results indicate that this model has promise for studies regarding radiation dose to cardiac substructures.

An open source auto-segmentation algorithm for delineating heart and substructures - Development and validation within a multicenter lung cancer cohort

BACKGROUND AND PURPOSE

Irradiation of the heart in thoracic cancers raises toxicity concerns. For accurate dose estimation, automated heart and substructure segmentation is potentially useful. In this study, a hybrid automatic segmentation is developed. The accuracy of delineation and dose predictions were evaluated, testing the method's potential within heart toxicity studies.

MATERIALS AND METHODS

The hybrid segmentation method delineated the heart, four chambers, three large vessels, and the coronary arteries. The method consisted of a nnU-net heart segmentation and partly atlas- and model-based segmentation of the substructures. The nnU-net training and atlas segmentation was based on lung cancer patients and was validated against a national consensus dataset of 12 patients with breast cancer. The accuracy of dose predictions between manual and auto-segmented heart and substructures was evaluated by transferring the dose distribution of 240 previously treated lung cancer patients to the consensus data set.

RESULTS

The hybrid auto-segmentation method performed well with a heart dice similarity coefficient (DSC) of 0.95, with no statistically significant difference between the automatic and manual delineations. The DSC for the chambers varied from 0.78-0.86 for the automatic segmentation and was comparable with the inter-observer variability. Most importantly, the automatic segmentation was as precise as the clinical experts in predicting the dose distribution to the heart and all substructures.

CONCLUSION

The hybrid segmentation method performed well in delineating the heart and substructures. The prediction of dose by the automatic segmentation was aligned with the manual delineations, enabling measurement of heart and substructure dose in large cohorts. The delineation algorithm will be available for download.

The Association of Incidental Radiation Dose to the Heart Base with Overall Survival and Cardiac Events after Curative-intent Radiotherapy for Non-small Cell Lung Cancer: Results from the NI-HEART Study

AIMS

Cardiac disease is a dose-limiting toxicity in non-small cell lung cancer radiotherapy. The dose to the heart base has been associated with poor survival in multiple institutional and clinical trial datasets using unsupervised, voxel-based analysis. Validation has not been undertaken in a cohort with individual patient delineations of the cardiac base or for the endpoint of cardiac events. The purpose of this study was to assess the association of heart base radiation dose with overall survival and the risk of cardiac events with individual heart base contours.

MATERIALS AND METHODS

Patients treated between 2015 and 2020 were reviewed for baseline patient, tumour and cardiac details and both cancer and cardiac outcomes as part of the NI-HEART study. Three cardiologists verified cardiac events including atrial fibrillation, heart failure and acute coronary syndrome. Cardiac substructure delineations were completed using a validated deep learning-based autosegmentation tool and a composite cardiac base structure was generated. Cox and Fine-Gray regressions were undertaken for the risk of death and cardiac events.

RESULTS

Of 478 eligible patients, most received 55 Gy/20 fractions (96%) without chemotherapy (58%), planned with intensity-modulated radiotherapy (71%). Pre-existing cardiovascular morbidity was common (78% two or more risk factors, 46% one or more established disease). The median follow-up was 21.1 months. Dichotomised at the median, a higher heart base Dmax was associated with poorer survival on Kaplan-Meier analysis (20.2 months versus 28.3 months; hazard ratio 1.40, 95% confidence interval 1.14-1.75, P = 0.0017) and statistical significance was retained in multivariate analyses. Furthermore, heart base Dmax was associated with pooled cardiac events in a multivariate analysis (hazard ratio 1.75, 95% confidence interval 1.03-2.97, P = 0.04).

CONCLUSIONS

Heart base Dmax was associated with the rate of death and cardiac events after adjusting for patient, tumour and cardiovascular factors in the NI-HEART study. This validates the findings from previous unsupervised analytical approaches. The heart base could be considered as a potential sub-organ at risk towards reducing radiation cardiotoxicity.

Cardiac radiation dose is associated with inferior survival but not cardiac events in patients with locally advanced non-small cell lung cancer in the era of immune checkpoint inhibitor consolidation

PURPOSE

We assessed the association of cardiac radiation dose with cardiac events and survival post-chemoradiation therapy (CRT) in patients with locally advanced non-small cell lung cancer (LA-NSCLC) after adoption of modern radiation therapy (RT) techniques, stricter cardiac dose constraints, and immune checkpoint inhibitor (ICI) consolidation.

METHODS AND MATERIALS

This single-institution, multi-site retrospective study included 335 patients with LA-NSCLC treated with definitive, concurrent CRT between October 2017 and December 2021. All patients were evaluated for ICI consolidation. Planning dose constraints included heart mean dose < 20 Gy (<10 Gy if feasible) and heart volume receiving ≥ 50 Gy (V50Gy) < 25 %. Twenty-one dosimetric parameters for three different cardiac structures (heart, left anterior descending coronary artery [LAD], and left ventricle) were extracted. Primary endpoint was any major adverse cardiac event (MACE) post-CRT, defined as acute coronary syndrome, heart failure, coronary revascularization, or cardiac-related death. Secondary endpoints were: grade ≥ 3 cardiac events (per CTCAE v5.0), overall survival (OS), lung cancer-specific mortality (LCSM), and other-cause mortality (OCM).

RESULTS

Median age was 68 years, 139 (41 %) had baseline coronary heart disease, and 225 (67 %) received ICI consolidation. Proton therapy was used in 117 (35 %) and intensity-modulated RT in 199 (59 %). Median LAD V15Gy was 1.4 % (IQR 0-22) and median heart mean dose was 8.7 Gy (IQR 4.6-14.4). Median follow-up was 3.3 years. Two-year cumulative incidence of MACE was 9.5 % for all patients and 14.3 % for those with baseline coronary heart disease. Two-year cumulative incidence of grade ≥ 3 cardiac events was 20.4 %. No cardiac dosimetric parameter was associated with an increased risk of MACE or grade ≥ 3 cardiac events. On multivariable analysis, cardiac dose (LAD V15Gy and heart mean dose) was associated with worse OS, driven by an association with LCSM but not OCM.

CONCLUSIONS

With modern RT techniques, stricter cardiac dose constraints, and ICI consolidation, cardiac dose was associated with LCSM but not OCM or cardiac events in patients with LA-NSCLC.

Assessment of heart-substructures auto-contouring accuracy for application in heart-sparing radiotherapy for lung cancer

Objectives

We validated an auto-contouring algorithm for heart substructures in lung cancer patients, aiming to establish its accuracy and reliability for radiotherapy (RT) planning. We focus on contouring an amalgamated set of subregions in the base of the heart considered to be a new organ at risk, the cardiac avoidance area (CAA), to enable maximum dose limit implementation in lung RT planning.

Methods

The study validates a deep-learning model specifically adapted for auto-contouring the CAA (which includes the right atrium, aortic valve root, and proximal segments of the left and right coronary arteries). Geometric, dosimetric, quantitative, and qualitative validation measures are reported. Comparison with manual contours, including assessment of interobserver variability, and robustness testing over 198 cases are also conducted.

Results

Geometric validation shows that auto-contouring performance lies within the expected range of manual observer variability despite being slightly poorer than the average of manual observers (mean surface distance for CAA of 1.6 vs 1.2 mm, dice similarity coefficient of 0.86 vs 0.88). Dosimetric validation demonstrates consistency between plans optimized using auto-contours and manual contours. Robustness testing confirms acceptable contours in all cases, with 80% rated as "Good" and the remaining 20% as "Useful."

Conclusions

The auto-contouring algorithm for heart substructures in lung cancer patients demonstrates acceptable and comparable performance to human observers.

Advances in knowledge

Accurate and reliable auto-contouring results for the CAA facilitate the implementation of a maximum dose limit to this region in lung RT planning, which has now been introduced in the routine setting at our institution.

Interpretable Machine Learning for Choosing Radiation Dose-volume Constraints on Cardio-pulmonary Substructures Associated with Overall Survival in NRG Oncology RTOG 0617

PURPOSE

Our objective was to use interpretable machine learning for choosing dose-volume constraints on cardiopulmonary substructures (CPSs) associated with overall survival (OS) in radiation therapy for locally advanced non-small cell lung cancer.

METHODS AND MATERIALS

A total of 428 patients with non-small cell lung cancer were randomly divided into training/validation/test subsets (n = 230/149/49) in Radiation Therapy Oncology Group 0617. Manual or automated contouring was performed to segment CPSs, including heart, atria, ventricles, aorta, left/right ventricle/atrium (LV+RV+LA+RA), inferior/superior vena cava, pulmonary artery, and pericardium. Peri (pericardium-heart), rest (heart-[LV+RV+LA+RA]), clinical target volume (CTV), and lungs-CTV contours were also obtained. Dose-volume histogram features were extracted, including minimum/mean dose to the hottest x% volume (Dx%[Gy]/MOHx%[Gy]), minimum/mean/maximum dose, percent volume receiving at least xGy (VxGy[%]), and overlapping volume of each CPS with planning target volume (PTV_Voverlap[%]). Clinical parameters were collected from the National Clinical Trials Network/Community oncology research program data archive. Feature selection was performed using a series of multiblock sparse partial least squares regression, stability selection supervised principal component analysis, and Boruta. Explainable boosting machine (EBM) was trained using a conditional survival distribution-based approach for imputing censored data, treating survival analysis as a regression problem. Harrell's C-index was used to evaluate OS discrimination performance of EBM, Cox proportional hazards (CPH), random survival forest, extreme gradient boosting survival embeddings, and CPH deep neural network (DeepSurv) models in the test set. Dose-volume constraints were selected using the binary change point detection algorithm in Shapley additive explanations-based partial dependence functions.

RESULTS

Selected features included LA_V60Gy(%), pericardium_D30%(Gy), lungs-CTV_PTV_Voverlap(%), RA_V55Gy(%), and received_cons_chemo. All models ranked LA_V60Gy(%) as the most important feature. EBM achieved the best performance for predicting OS, followed by extreme gradient boosting survival embeddings, random survival forest, DeepSurv, and CPH (C-index = 0.653, 0.646, 0.642, 0.638, and 0.632). EBM global explanations suggested that LA_V60Gy(%) < 25.6, lungs-CTV_PTV_Voverlap(%) < 1.1, pericardium_D30%(Gy) < 18.9, RA_V55Gy(%) < 19.5, and received_cons_chemo = 'Yes' for improved OS.

CONCLUSIONS

EBM can be used to discriminate OS while also guiding dose-volume constraint selection for optimal management of cardiac toxicity in lung cancer radiation therapy.

Multi-organ segmentation of CT via convolutional neural network: impact of training setting and scanner manufacturer

Objective. Automated organ segmentation on CT images can enable the clinical use of advanced quantitative software devices, but model performance sensitivities must be understood before widespread adoption can occur. The goal of this study was to investigate performance differences between Convolutional Neural Networks (CNNs) trained to segment one (single-class) versus multiple (multi-class) organs, and between CNNs trained on scans from a single manufacturer versus multiple manufacturers.Methods. The multi-class CNN was trained on CT images obtained from 455 whole-body PET/CT scans (413 for training, 42 for testing) taken with Siemens, GE, and Phillips PET/CT scanners where 16 organs were segmented. The multi-class CNN was compared to 16 smaller single-class CNNs trained using the same data, but with segmentations of only one organ per model. In addition, CNNs trained on Siemens-only (N = 186) and GE-only (N = 219) scans (manufacturer-specific) were compared with CNNs trained on data from both Siemens and GE scanners (manufacturer-mixed). Segmentation performance was quantified using five performance metrics, including the Dice Similarity Coefficient (DSC).Results. The multi-class CNN performed well compared to previous studies, even in organs usually considered difficult auto-segmentation targets (e.g., pancreas, bowel). Segmentations from the multi-class CNN were significantly superior to those from smaller single-class CNNs in most organs, and the 16 single-class models took, on average, six times longer to segment all 16 organs compared to the single multi-class model. The manufacturer-mixed approach achieved minimally higher performance over the manufacturer-specific approach.Significance. A CNN trained on contours of multiple organs and CT data from multiple manufacturers yielded high-quality segmentations. Such a model is an essential enabler of image processing in a software device that quantifies and analyzes such data to determine a patient's treatment response. To date, this activity of whole organ segmentation has not been adopted due to the intense manual workload and time required.

Development of a comprehensive cardiac atlas on a 1.5 Tesla Magnetic Resonance Linear Accelerator

Background and purpose

The 1.5 Tesla (T) Magnetic Resonance Linear Accelerator (MRL) provides an innovative modality for improved cardiac imaging when planning radiation treatment. No MRL based cardiac atlases currently exist, thus, we sought to comprehensively characterize cardiac substructures, including the conduction system, from cardiac images acquired using a 1.5 T MRL and provide contouring guidelines.

Materials and methods

Five volunteers were enrolled in a prospective protocol (NCT03500081) and were imaged on the 1.5 T MRL with Half Fourier Single-Shot Turbo Spin-Echo (HASTE) and 3D Balanced Steady-State Free Precession (bSSFP) sequences in axial, short axis, and vertical long axis. Cardiac anatomy was contoured by (AS) and confirmed by a board certified cardiologist (JR) with expertise in cardiac MR imaging.

Results

A total of five volunteers had images acquired with the HASTE sequence, with 21 contours created on each image. One of these volunteers had additional images obtained with 3D bSSFP sequences in the axial plane and additional images obtained with HASTE sequences in the key cardiac planes. Contouring guidelines were created and outlined. 15-16 contours were made for the short axis and vertical long axis. The cardiac conduction system was demonstrated with eleven representative contours. There was reasonable variation of contour volume across volunteers, with structures more clearly delineated on the 3D bSSFP sequence.

Conclusions

We present a comprehensive cardiac atlas using novel images acquired prospectively on a 1.5 T MRL. This cardiac atlas provides a novel resource for radiation oncologists in delineating cardiac structures for treatment with radiotherapy, with special focus on the cardiac conduction system.

Demystifying the Results of RTOG 0617: Identification of Dose Sensitive Cardiac Subregions Associated With Overall Survival

INTRODUCTION

The RTOG 0617 trial presented a worse survival for patients with lung cancer treated in the high-dose (74 Gy) arm. In multivariable models, radiation level and whole-heart volumetric dose parameters were associated with survival. In this work, we consider heart subregions to explain the observed survival difference between radiation levels.

METHODS

Voxel-based analysis identified anatomical regions where the dose was associated with survival. Bootstrapping clinical and dosimetric variables into an elastic net model selected variables associated with survival. Multivariable Cox regression survival models assessed the significance of dose to the heart subregion, compared with whole heart v5 and v30. Finally, the trial outcome was assessed after propensity score matching of patients on lung dose, heart subregion dose, and tumor volume.

RESULTS

A total of 458 patients were eligible for voxel-based analysis. A region of significance (p < 0.001) was identified in the base of the heart. Bootstrapping selected mean lung dose, radiation level, log tumor volume, and heart region dose. The multivariable Cox model exhibited dose to the heart region (p = 0.02), and tumor volume (p = 0.03) were significantly associated with survival, and radiation level was not significant (p = 0.07). The models exhibited that whole heart v5 and v30 were not associated with survival, with radiation level being significant (p < 0.05). In the matched cohort, no significant survival difference was seen between radiation levels.

CONCLUSIONS

Dose to the base of the heart is associated with overall survival, partly removing the radiation level effect, and explaining that worse survival in the high-dose arm is owing, in part, to the heart subregion dose. By defining a heart avoidance region, future dose escalation trials may be feasible.

IA PULMONARY VEIN ATLAS FOR RADIOTHERAPY PLANNING

BACKGROUND AND PURPOSE

Cardiac arrhythmia is a recognised potential complication of thoracic radiotherapy, but the responsible cardiac substructures for arrhythmogenesis have not been identified. Arrhythmogenic tissue is commonly located in the pulmonary veins (PVs) of cardiology patients with arrhythmia, however these structures are not currently considered organs-at-risk during radiotherapy planning. A standardised approach to their delineation was developed and evaluated.

MATERIALS AND METHODS

The gross and radiological anatomy relevant to atrial fibrillation was derived from cardiology and radiology literature by a multidisciplinary team. A region of interest and contouring instructions for radiotherapy computed tomography scans were iteratively developed and subsequently evaluated. Radiation oncologists (n=5) and radiation technologists (n=2) contoured the PVs on the four-dimensional planning datasets of five patients with locally advanced lung cancer treated with 1.8-2.75 Gy fractions. Contours were compared to reference contours agreed by the researchers using geometric and dosimetric parameters.

RESULTS

The mean dose to the PVs was 35% prescription dose. Geometric and dosimetric similarity of the observer contours with reference contours was fair, with an overall mean Dice of 0.80 ± 0.02. The right superior PV (mean DSC 0.83 ± 0.02) had better overlap than the left (mean DSC 0.80 ± 0.03), but the inferior PVs were equivalent (mean DSC of 0.78). The mean difference in mean dose was 0.79 Gy ± 0.71 (1.46% ± 1.25).

CONCLUSION

A PV atlas with multidisciplinary approval led to reproducible delineation for radiotherapy planning, supporting the utility of the atlas in future clinical radiotherapy cardiotoxicity research encompassing arrhythmia endpoints.

Validation of a Fully Automated Hybrid Deep Learning Cardiac Substructure Segmentation Tool for Contouring and Dose Evaluation in Lung Cancer Radiotherapy

BACKGROUND AND PURPOSE

Accurate and consistent delineation of cardiac substructures is challenging. The aim of this work was to validate a novel segmentation tool for automatic delineation of cardiac structures and subsequent dose evaluation, with potential application in clinical settings and large-scale radiation-related cardiotoxicity studies.

MATERIALS AND METHODS

A recently developed hybrid method for automatic segmentation of 18 cardiac structures, combining deep learning, multi-atlas mapping and geometric segmentation of small challenging substructures, was independently validated on 30 lung cancer cases. These included anatomical and imaging variations, such as tumour abutting heart, lung collapse and metal artefacts. Automatic segmentations were compared with manual contours of the 18 structures using quantitative metrics, including Dice similarity coefficient (DSC), mean distance to agreement (MDA) and dose comparisons.

RESULTS

A comparison of manual and automatic contours across all cases showed a median DSC of 0.75-0.93 and a median MDA of 2.09-3.34 mm for whole heart and chambers. The median MDA for great vessels, coronary arteries, cardiac valves, sinoatrial and atrioventricular conduction nodes was 3.01-8.54 mm. For the 27 cases treated with curative intent (planned target volume dose ≥50 Gy), the median dose difference was -1.12 to 0.57 Gy (absolute difference of 1.13-3.25%) for the mean dose to heart and chambers; and -2.25 to 4.45 Gy (absolute difference of 0.94-6.79%) for the mean dose to substructures.

CONCLUSION

The novel hybrid automatic segmentation tool reported high accuracy and consistency over a validation set with challenging anatomical and imaging variations. This has promising applications in substructure dose calculations of large-scale datasets and for future studies on long-term cardiac toxicity.

Open-source, fully-automated hybrid cardiac substructure segmentation: development and optimisation

Radiotherapy for thoracic and breast tumours is associated with a range of cardiotoxicities. Emerging evidence suggests cardiac substructure doses may be more predictive of specific outcomes, however, quantitative data necessary to develop clinical planning constraints is lacking. Retrospective analysis of patient data is required, which relies on accurate segmentation of cardiac substructures. In this study, a novel model was designed to deliver reliable, accurate, and anatomically consistent segmentation of 18 cardiac substructures on computed tomography (CT) scans. Thirty manually contoured CT scans were included. The proposed multi-stage method leverages deep learning (DL), multi-atlas mapping, and geometric modelling to automatically segment the whole heart, cardiac chambers, great vessels, heart valves, coronary arteries, and conduction nodes. Segmentation performance was evaluated using the Dice similarity coefficient (DSC), mean distance to agreement (MDA), Hausdorff distance (HD), and volume ratio. Performance was reliable, with no errors observed and acceptable variation in accuracy between cases, including in challenging cases with imaging artefacts and atypical patient anatomy. The median DSC range was 0.81-0.93 for whole heart and cardiac chambers, 0.43-0.76 for great vessels and conduction nodes, and 0.22-0.53 for heart valves. For all structures the median MDA was below 6 mm, median HD ranged 7.7-19.7 mm, and median volume ratio was close to one (0.95-1.49) for all structures except the left main coronary artery (2.07). The fully automatic algorithm takes between 9 and 23 min per case. The proposed fully-automatic method accurately delineates cardiac substructures on radiotherapy planning CT scans. Robust and anatomically consistent segmentations, particularly for smaller structures, represents a major advantage of the proposed segmentation approach. The open-source software will facilitate more precise evaluation of cardiac doses and risks from available clinical datasets.

Death without progression as an endpoint to describe cardiac radiation effects in locally advanced non-small cell lung cancer

Background and purpose

Prior studies have examined associations of cardiovascular substructure dose with overall survival (OS) or cardiac events after chemoradiotherapy (CRT) for non-small cell lung cancer (NSCLC). Herein, we investigate an alternative endpoint, death without cancer progression (DWP), which is potentially more specific than OS and more sensitive than cardiac events for understanding CRT toxicity.

Materials and methods

We retrospectively reviewed records of 187 patients with locally advanced or oligometastatic NSCLC treated with definitive CRT from 2008 to 2016 at a single institution. Dosimetric parameters to the heart, lung, and ten cardiovascular substructures were extracted. Charlson Comorbidity Index (CCI), excluding NSCLC diagnosis, was used to stratify patients into CCI low (0-2; n = 66), CCI intermediate (3-4; n = 78), and CCI high (≥5; n = 43) groups. Primary endpoint was DWP, modeled with competing risk regression. Secondary endpoints included OS. An external cohort consisted of 140 patients from another institution.

Results

Median follow-up was 7.3 years for survivors. Death occurred in 143 patients (76.5 %), including death after progression in 118 (63.1 %) and DWP in 25 (13.4 %). On multivariable analysis, increasing CCI stratum and mean heart dose were associated with DWP. For mean heart dose ≥ 10 Gy vs < 10 Gy, DWP was higher (5-year rate, 16.9 % vs 6.7 %, p = 0.04) and OS worse (median, 22.9 vs 34.1 months, p < 0.001). Ventricle (left, right, and bilateral) and pericardial but not atrial substructure dose were associated with DWP, whereas all three were inversely associated with OS. Cutpoint analysis identified right ventricle mean dose ≥ 5.5 Gy as a predictor of DWP. In the external cohort, we confirmed an association of ventricle, but not atrial, dose with DWP.

Conclusion

Cardiovascular substructure dose showed distinct associations with DWP. Future cardiotoxicity studies in NSCLC could consider DWP as an endpoint.

Association of cardiac calcium burden with overall survival after radiotherapy for non-small cell lung cancer

Background and purpose

Coronary calcifications are associated with coronary artery disease in patients undergoing radiotherapy (RT) for non-small cell lung cancer (NSCLC). We quantified calcifications in the coronary arteries and aorta and investigated their relationship with overall survival (OS) in patients treated with definitive RT (Def-RT) or post-operative RT (PORT).

Materials and methods

We analyzed 263 NSCLC patients treated from 2004 to 2017. Calcium burden was ascertained with a Hounsfield unit (HU) cutoff of > 130 in addition to a deep learning (DL) plaque estimator. The HU cutoff volumes were defined for coronary arteries (Plaque_Coro) and coronary arteries and aorta combined (Plaque_Coro+Ao), while the DL estimator ranged from 0 (no plaque) to 3 (high plaque). Patient and treatment characteristics were explored for association with OS.

Results

The median Plaque_Coro and Plaque_Coro+Ao was 0.75 cm³ and 0.87 cm³ in the Def-RT group and 0.03 cm³ and 0.52 cm³ in the PORT group. The median DL estimator was 2 in both cohorts. In Def-RT, large Plaque_Coro (HR:1.11 (95%CI:1.04-1.19); p = 0.008), and Plaque_Coro+Ao (HR:1.06 (95%CI:1.02-1.11); p = 0.03), and poor Karnofsky Performance Status (HR: 0.97 (95%CI: 0.94-0.99); p = 0.03) were associated with worse OS. No relationship was identified between the plaque volumes and OS in PORT, or between the DL plaque estimator and OS in either Def-RT or PORT.

Conclusions

Coronary artery calcification assessed from RT planning CT scans was significantly associated with OS in patients who underwent Def-RT for NSCLC. This HU thresholding method can be straightforwardly implemented such that the role of calcifications can be further explored.

Role of Real-World Data in Assessing Cardiac Toxicity After Lung Cancer Radiotherapy

Radiation-induced heart disease (RIHD) is a recent concern in patients with lung cancer after being treated with radiotherapy. Most of information we have in the field of cardiac toxicity comes from studies utilizing real-world data (RWD) as randomized controlled trials (RCTs) are generally not practical in this field. This article is a narrative review of the literature using RWD to study RIHD in patients with lung cancer following radiotherapy, summarizing heart dosimetric factors associated with outcome, strength, and limitations of the RWD studies, and how RWD can be used to assess a change to cardiac dose constraints.

Validation of an established deep learning auto-segmentation tool for cardiac substructures in 4D radiotherapy planning scans

•

Cardiotoxicity is a common complication of lung cancer radiotherapy.

•

Segmentation of cardiac substructures is time-consuming and challenging.

•

Deep learning segmentation tools can perform this task in 3D and 4D scans.

•

Performance is high when assessed geometrically, dosimetrically and clinically.

•

Auto-segmentation tools may accelerate clinical workflows and enable research.

Background

Emerging data suggest that dose-sparing several key cardiac regions is prognostically beneficial in lung cancer radiotherapy. The cardiac substructures are challenging to contour due to their complex geometry, poor soft tissue definition on computed tomography (CT) and cardiorespiratory motion artefact. A neural network was previously trained to generate the cardiac substructures using three-dimensional radiotherapy planning CT scans (3D-CT). In this study, the performance of that tool on the average intensity projection from four-dimensional (4D) CT scans (4D-AVE), now commonly used in lung radiotherapy, was evaluated.

Materials and Methods

The 4D-AVE of n=20 patients completing radiotherapy for lung cancer 2015–2020 underwent manual and automated cardiac substructure segmentation. Manual and automated substructures were compared geometrically and dosimetrically. Two senior clinicians also qualitatively assessed the auto-segmentation tool’s output.

Results

Geometric comparison of the automated and manual segmentations exhibited high levels of similarity across parameters, including volume difference (11.8% overall) and Dice similarity coefficient (0.85 overall), and were consistent with 3D-CT performance. Differences in mean (median 0.2 Gy, range −1.6–0.3 Gy) and maximum (median 0.4 Gy, range −2.2–0.9 Gy) doses to substructures were generally small. Nearly all structures (99.5 %) were deemed to be appropriate for clinical use without further editing.

Conclusions

Cardiac substructure auto-segmentation using a deep learning-based tool trained on a 3D-CT dataset was feasible on the 4D-AVE scan, meaning this tool is suitable for use on 4D-CT radiotherapy planning scans. Application of this tool would increase the practicality of routine clinical cardiac substructure delineation, and enable further cardiac radiation effects research.

A Survey on Deep Learning for Precision Oncology

Precision oncology, which ensures optimized cancer treatment tailored to the unique biology of a patient’s disease, has rapidly developed and is of great clinical importance. Deep learning has become the main method for precision oncology. This paper summarizes the recent deep-learning approaches relevant to precision oncology and reviews over 150 articles within the last six years. First, we survey the deep-learning approaches categorized by various precision oncology tasks, including the estimation of dose distribution for treatment planning, survival analysis and risk estimation after treatment, prediction of treatment response, and patient selection for treatment planning. Secondly, we provide an overview of the studies per anatomical area, including the brain, bladder, breast, bone, cervix, esophagus, gastric, head and neck, kidneys, liver, lung, pancreas, pelvis, prostate, and rectum. Finally, we highlight the challenges and discuss potential solutions for future research directions.

Mutual enhancing learning-based automatic segmentation of CT cardiac substructure

Objective. Current segmentation practice for thoracic cancer RT considers the whole heart as a single organ despite increased risks of cardiac toxicities from irradiation of specific cardiac substructures. Segmenting up to 15 different cardiac substructures can be a very time-intensive process, especially due to their different volume sizes and anatomical variations amongst different patients. In this work, a new deep learning (DL)-based mutual enhancing strategy is introduced for accurate and automatic segmentation, especially of smaller substructures such as coronary arteries. Approach. Our proposed method consists of three subnetworks: retina U-net, classification module, and segmentation module. Retina U-net is used as a backbone network architecture that aims to learn deep features from the whole heart. Whole heart feature maps from retina U-net are then transferred to four different sets of classification modules to generate classification localization maps of coronary arteries, great vessels, chambers of the heart, and valves of the heart. Each classification module is in sync with its corresponding subsequent segmentation module in a bootstrapping manner, allowing them to share their encoding paths to generate a mutual enhancing strategy. We evaluated our method on three different datasets: institutional CT datasets (55 subjects) 2) publicly available Multi-Modality Whole Heart Segmentation (MM-WHS) challenge datasets (120 subjects), and Automated Cardiac Diagnosis Challenge (ACDC) datasets (100 subjects). For institutional datasets, we performed five-fold cross-validation on training data (45 subjects) and performed inference on separate hold-out data (10 subjects). For each subject, 15 cardiac substructures were manually contoured by a resident physician and evaluated by an attending radiation oncologist. For the MM-WHS dataset, we trained the network on 100 datasets and performed an inference on a separate hold-out dataset with 20 subjects, each with 7 cardiac substructures. For ACDC datasets, we performed five-fold cross-validation on 100 datasets, each with 3 cardiac substructures. We compared the proposed method against four different network architectures: 3D U-net, mask R-CNN, mask scoring R-CNN, and proposed network without classification module. Segmentation accuracies were statistically compared through dice similarity coefficient, Jaccard, 95% Hausdorff distance, mean surface distance, root mean square distance, center of mass distance, and volume difference. Main results. The proposed method generated cardiac substructure segmentations with significantly higher accuracy (P < 0.05) for small substructures, especially for coronary arteries such as left anterior descending artery (CA-LADA) and right coronary artery (CA-RCA) in comparison to four competing methods. For large substructures (i.e. chambers of the heart), our method yielded comparable results to mask scoring R-CNN method, resulting in significantly (P < 0.05) improved segmentation accuracy in comparison to 3D U-net and mask R-CNN. Significance. A new DL-based mutual enhancing strategy was introduced for automatic segmentation of cardiac substructures. Overall results of this work demonstrate the ability of the proposed method to improve segmentation accuracies of smaller substructures such as coronary arteries without largely compromising the segmentation accuracies of larger substructures. Fast and accurate segmentations of up to 15 substructures can possibly be used as a tool to rapidly generate substructure segmentations followed by physicians’ reviews to improve clinical workflow.

Automated 3D Segmentation of the Aorta and Pulmonary Artery on Non-Contrast-Enhanced Chest Computed Tomography Images in Lung Cancer Patients

Pulmonary hypertension should be preoperatively evaluated for optimal surgical planning to reduce surgical risk in lung cancer patients. Preoperative measurement of vascular diameter in computed tomography (CT) images is a noninvasive prediction method for pulmonary hypertension. However, the current estimation method, 2D manual arterial diameter measurement, may yield inaccurate results owing to low tissue contrast in non-contrast-enhanced CT (NECT). Furthermore, it provides an incomplete evaluation by measuring only the diameter of the arteries rather than the volume. To provide a more complete and accurate estimation, this study proposed a novel two-stage deep learning (DL) model for 3D aortic and pulmonary artery segmentation in NECT. In the first stage, a DL model was constructed to enhance the contrast of NECT; in the second stage, two DL models then applied the enhanced images for aorta and pulmonary artery segmentation. Overall, 179 patients were divided into contrast enhancement model (n = 59), segmentation model (n = 120), and testing (n = 20) groups. The performance of the proposed model was evaluated using Dice similarity coefficient (DSC). The proposed model could achieve 0.97 ± 0.007 and 0.93 ± 0.002 DSC for aortic and pulmonary artery segmentation, respectively. The proposed model may provide 3D diameter information of the arteries before surgery, facilitating the estimation of pulmonary hypertension and supporting preoperative surgical method selection based on the predicted surgical risks.

Automatic Cardiac Structure Contouring for Small Datasets with Cascaded Deep Learning Models

Cardiac structure contouring is a time consuming and tedious manual activity used for radiotherapeutic dose toxicity planning. We developed an automatic cardiac structure segmentation pipeline for use in low-dose non-contrast planning CT based on deep learning algorithms for small datasets. Fifty CT scans were retrospectively selected and the whole heart, ventricles and atria were contoured. A two stage deep learning pipeline was trained on 41 non contrast planning CTs, tuned with 3 CT scans and validated on 6 CT scans. In the first stage, An InceptionResNetV2 network was used to identify the slices that contained cardiac structures. The second stage consisted of three deep learning models trained on the images containing cardiac structures to segment the structures. The three deep learning models predicted the segmentations/contours on axial, coronal and sagittal images and are combined to create the final prediction. The final accuracy of the pipeline was quantified on 6 volumes by calculating the Dice similarity coefficient (DC), 95% Hausdorff distance (95% HD) and volume ratios between predicted and ground truth volumes. Median DC and 95% HD of 0.96, 0.88, 0.92, 0.80 and 0.82, and 1.86, 2.98, 2.02, 6.16 and 6.46 were achieved for the whole heart, right and left ventricle, and right and left atria respectively. The median differences in volume were -4, -1, + 5, -16 and -20% for the whole heart, right and left ventricle, and right and left atria respectively. The automatic contouring pipeline achieves good results for whole heart and ventricles. Robust automatic contouring with deep learning methods seems viable for local centers with small datasets.

Pre-treatment immune status predicts disease control in NSCLCs treated with chemoradiation and durvalumab

BACKGROUND

The impact of peripheral blood immune measures and radiation-induced lymphopenia on outcomes in non-small cell lung cancer (NSCLC) patients treated with concurrent chemoradiation (cCRT) and immune check point inhibition (ICI) has yet to be fully defined.

METHODS

Stage III NSCLC patients treated with cCRT and ≥1 dose of durvalumab across a cancer center were examined. Peripheral blood counts were assessed pre-cCRT, during cCRT and at the start of ICI. These measures and risk-scores from two published models estimating radiation dose to immune-bearing organs were tested for association with disease control.

RESULTS

We assessed 113 patients treated with cCRT and a median of 8.5 months of durvalumab. Median PFS was 29 months (95% CI 18-35 months). A lower pre-cCRT ALC (HR: 0.51 (95% CI: 0.32-0.82), p = 0.02) and a higher pre-cCRT ANC (HR: 1.14 (1.06-1.23), p = 0.005) were associated with poor PFS. Neither ALC nadir, ALC at ICI start, ANC at ICI start or the normalized change in ALC from pre-cCRT to nadir were significantly associated with PFS (p = 0.07-0.49). Also, risk scores from the two radiation-dose models were not associated with PFS (p = 0.14, p = 0.21) but were so with the ALC Nadir (p = 0.001, p = 0.002). A higher pre-cCRT NLR was the strongest predictor for PFS (HR: 1.09 (1.05-1.14), p = 0.0001). The 12-month PFS in patients with the bottom vs. top NLR tertile was 84% vs 46% (p = 0.000004).

CONCLUSIONS

Baseline differences in peripheral immune cell populations are associated with disease outcomes in NSCLC patients treated with cCRT and ICI.

Prospectively-validated deep learning model for segmenting swallowing and chewing structures in CT

Objective.Delineating swallowing and chewing structures aids in radiotherapy (RT) treatment planning to limit dysphagia, trismus, and speech dysfunction. We aim to develop an accurate and efficient method to automate this process.Approach.CT scans of 242 head and neck (H&N) cancer patients acquired from 2004 to 2009 at our institution were used to develop auto-segmentation models for the masseters, medial pterygoids, larynx, and pharyngeal constrictor muscle using DeepLabV3+. A cascaded framework was used, wherein models were trained sequentially to spatially constrain each structure group based on prior segmentations. Additionally, an ensemble of models, combining contextual information from axial, coronal, and sagittal views was used to improve segmentation accuracy. Prospective evaluation was conducted by measuring the amount of manual editing required in 91 H&N CT scans acquired February-May 2021.Main results. Medians and inter-quartile ranges of Dice similarity coefficients (DSC) computed on the retrospective testing set (N = 24) were 0.87 (0.85-0.89) for the masseters, 0.80 (0.79-0.81) for the medial pterygoids, 0.81 (0.79-0.84) for the larynx, and 0.69 (0.67-0.71) for the constrictor. Auto-segmentations, when compared to two sets of manual segmentations in 10 randomly selected scans, showed better agreement (DSC) with each observer than inter-observer DSC. Prospective analysis showed most manual modifications needed for clinical use were minor, suggesting auto-contouring could increase clinical efficiency. Trained segmentation models are available for research use upon request viahttps://github.com/cerr/CERR/wiki/Auto-Segmentation-models.Significance.We developed deep learning-based auto-segmentation models for swallowing and chewing structures in CT and demonstrated its potential for use in treatment planning to limit complications post-RT. To the best of our knowledge, this is the only prospectively-validated deep learning-based model for segmenting chewing and swallowing structures in CT. Segmentation models have been made open-source to facilitate reproducibility and multi-institutional research.

Implementation of a Knowledge-Based Treatment Planning Model for Cardiac-Sparing Lung Radiation Therapy

PURPOSE

High radiation doses to the heart have been correlated with poor overall survival in patients receiving radiation therapy for stage III non-small cell lung cancer (NSCLC). We built a knowledge-based planning (KBP) tool to limit the dose to the heart during creation of volumetric modulated arc therapy (VMAT) treatment plans for patients being treated to 60 Gy in 30 fractions for stage III NSCLC.

METHODS AND MATERIALS

A previous study at our institution retrospectively delineated intracardiac volumes and optimized VMAT treatment plans to reduce dose to these substructures and to the whole heart. Two RapidPlan (RP) KBP models were built from this cohort, 1 model using the clinical plans and a separate model using the cardiac-optimized plans. Using target volumes and 6 organs at risk (OARs), models were trained to generate treatment plans in a semiautomated process. The cardiac-sparing KBP model was tested in the same cohort used for training, and both models were tested on an external validation cohort of 30 patients.

RESULTS

Both RP models produced clinically acceptable plans in terms of target coverage, dose uniformity, and dose to OARs. Compared with the previously created cardiac-optimized plans, cardiac-sparing RPs showed significant reductions in the mean dose to the esophagus and lungs while performing similarly or better in all evaluated heart dose metrics. When comparing the 2 models, the cardiac-sparing RP showed reduced (P < .05) heart mean and maximum doses as well as volumes receiving 60 Gy, 50 Gy, and 30 Gy.

CONCLUSIONS

By using a set of cardiac-optimized treatment plans for training, the proposed KBP model provided a means to reduce the dose to the heart and its substructures without the need to explicitly delineate cardiac substructures. This tool may offer reduced planning time and improved plan quality and might be used to improve patient outcomes.

Deep learning model for automatic contouring of cardiovascular substructures on radiotherapy planning CT images: Dosimetric validation and reader study based clinical acceptability testing

BACKGROUND AND PURPOSE

Large radiotherapy (RT) planning imaging datasets with consistently contoured cardiovascular structures are essential for robust cardiac radiotoxicity research in thoracic cancers. This study aims to develop and validate a highly accurate automatic contouring model for the heart, cardiac chambers, and great vessels for RT planning computed tomography (CT) images that can be used for dose-volume parameter estimation.

MATERIALS AND METHODS

A neural network model was trained using a dataset of 127 expertly contoured planning CT images from RT treatment of locally advanced non-small-cell lung cancer (NSCLC) patients. Evaluation of geometric accuracy and quality of dosimetric parameter estimation was performed on 50 independent scans with contrast and without contrast enhancement. The model was further evaluated regarding the clinical acceptability of the contours in 99 scans randomly sampled from the RTOG-0617 dataset by three experienced radiation oncologists.

RESULTS

Median surface dice at 3 mm tolerance for all dedicated thoracic structures was 90% in the test set. Median absolute difference between mean dose computed with model contours and expert contours was 0.45 Gy averaged over all structures. The mean clinical acceptability rate by majority vote in the RTOG-0617 scans was 91%.

CONCLUSION

This model can be used to contour the heart, cardiac chambers, and great vessels in large datasets of RT planning thoracic CT images accurately, quickly, and consistently. Additionally, the model can be used as a time-saving tool for contouring in clinic practice.

Assessment of fully automatic segmentation of pulmonary artery and aorta on noncontrast CT with optimal surface graph cuts

Purpose

Accurate segmentation of the pulmonary arteries and aorta is important due to the association of the diameter and the shape of these vessels with several cardiovascular diseases and with the risk of exacerbations and death in patients with chronic obstructive pulmonary disease. We propose a fully automatic method based on an optimal surface graph‐cut algorithm to quantify the full 3D shape and the diameters of the pulmonary arteries and aorta in noncontrast computed tomography (CT) scans.

Methods

The proposed algorithm first extracts seed points in the right and left pulmonary arteries, the pulmonary trunk, and the ascending and descending aorta by using multi‐atlas registration. Subsequently, the centerlines of the pulmonary arteries and aorta are extracted by a minimum cost path tracking between the extracted seed points, with a cost based on a combination of lumen intensity similarity and multiscale medialness in three planes. The centerlines are refined by applying the path tracking algorithm to curved multiplanar reformatted scans and are then smoothed and dilated nonuniformly according to the extracted local vessel radius from the medialness filter. The resulting coarse estimates of the vessels are used as initialization for a graph‐cut segmentation. Once the vessels are segmented, the diameters of the pulmonary artery (PA) and the ascending aorta (AA) and the PA:AA ratio are automatically calculated both in a single axial slice and in a 10 mm volume around the automatically extracted PA bifurcation level. The method is evaluated on noncontrast CT scans from the Danish Lung Cancer Screening Trial (DLCST). Segmentation accuracy is determined by comparing with manual annotations on 25 CT scans. Intraclass correlation (ICC) between manual and automatic diameters, both measured in axial slices at the PA bifurcation level, is computed on an additional 200 CT scans. Repeatability of the automated 3D volumetric diameter and PA:AA ratio calculations (perpendicular to the vessel axis) are evaluated on 118 scan–rescan pairs with an average in‐between time of 3 months.

Results

We obtained a Dice segmentation overlap of 0.94 ± 0.02 for pulmonary arteries and 0.96 ± 0.01 for the aorta, with a mean surface distance of 0.62 ± 0.33 mm and 0.43 ± 0.07 mm, respectively. ICC between manual and automatic in‐slice diameter measures was 0.92 for PA, 0.97 for AA, and 0.90 for the PA:AA ratio, and for automatic diameters in 3D volumes around the PA bifurcation level between scan and rescan was 0.89, 0.95, and 0.86, respectively.

Conclusion

The proposed automatic segmentation method can reliably extract diameters of the large arteries in non‐ECG‐gated noncontrast CT scans such as are acquired in lung cancer screening.

Robustness of deep learning segmentation of cardiac substructures in noncontrast computed tomography for breast cancer radiotherapy

PURPOSE

To develop and evaluate deep learning-based autosegmentation of cardiac substructures from noncontrast planning computed tomography (CT) images in patients undergoing breast cancer radiotherapy and to investigate the algorithm sensitivity to out-of-distribution data such as CT image artifacts.

METHODS

Nine substructures including aortic valve (AV), left anterior descending (LAD), tricuspid valve (TV), mitral valve (MV), pulmonic valve (PV), right atrium (RA), right ventricle (RV), left atrium (LA), and left ventricle (LV) were manually delineated by a radiation oncologist on noncontrast CT images of 129 patients with breast cancer; among them 90 were considered in-distribution data, also named as "clean" data. The image/label pairs of 60 subjects were used to train a 3D deep neural network while the remaining 30 were used for testing. The rest of the 39 patients were considered out-of-distribution ("outlier") data, which were used to test robustness. Random rigid transformations were used to augment the dataset during training. We investigated multiple loss functions, including Dice similarity coefficient (DSC), cross-entropy (CE), Euclidean loss as well as the variation and combinations of these, data augmentation, and network size on overall performance and sensitivity to image artifacts due to infrequent events such as the presence of implanted devices. The predicted label maps were compared to the ground-truth labels via DSC and mean and 90th percentile symmetric surface distance (90th-SSD).

RESULTS

When modified Dice combined with cross-entropy (MD-CE) was used as the loss function, the algorithm achieved a mean DSC = 0.79 ± 0.07 for chambers and  0.39 ± 0.10 for smaller substructures (valves and LAD). The mean and 90th-SSD were 2.7 ± 1.4 and 6.5 ± 2.8 mm for chambers and 4.1 ± 1.7 and 8.6 ± 3.2 mm for smaller substructures. Models with MD-CE, Dice-CE, MD, and weighted CE loss had highest performance, and were statistically similar. Data augmentation did not affect model performances on both clean and outlier data and model robustness was susceptible to network size. For a certain type of outlier data, robustness can be improved via incorporating them into the training process. The execution time for segmenting each patient was on an average 2.1 s.

CONCLUSIONS

A deep neural network provides a fast and accurate segmentation of large cardiac substructures in noncontrast CT images. Model robustness of two types of clinically common outlier data were investigated and potential approaches to improve them were explored. Evaluation of clinical acceptability and integration into clinical workflow are pending.

The impact of organ-at-risk contour variations on automatically generated treatment plans for NSCLC

BACKGROUND AND PURPOSE

Quality of automatic contouring is generally assessed by comparison with manual delineations, but the effect of contour differences on the resulting dose distribution remains unknown. This study evaluated dosimetric differences between treatment plans optimized using various organ-at-risk (OAR) contouring methods.

MATERIALS AND METHODS

OARs of twenty lung cancer patients were manually and automatically contoured, after which user-adjustments were made. For each contour set, an automated treatment plan was generated. The dosimetric effect of intra-observer contour variation and the influence of contour variations on treatment plan evaluation and generation were studied using dose-volume histogram (DVH)-parameters for thoracic OARs.

RESULTS

Dosimetric effect of intra-observer contour variability was highest for Heart D_max (3.4 ± 6.8 Gy) and lowest for Lungs-GTV D_mean (0.3 ± 0.4 Gy). The effect of contour variation on treatment plan evaluation was highest for Heart D_max (6.0 ± 13.4 Gy) and Esophagus D_max (8.7 ± 17.2 Gy). Dose differences for the various treatment plans, evaluated on the reference (manual) contour, were on average below 1 Gy/1%. For Heart D_mean, higher dose differences were found for overlap with PTV (median 0.2 Gy, 95% 1.7 Gy) vs. no PTV overlap (median 0 Gy, 95% 0.5 Gy). For D_max-parameters, largest dose difference was found between 0-1 cm distance to PTV (median 1.5 Gy, 95% 4.7 Gy).

CONCLUSION

Dose differences arising from automatic contour variations were of the same magnitude or lower than intra-observer contour variability. For Heart D_mean, we recommend delineation errors to be corrected when the heart overlaps with the PTV. For D_max-parameters, we recommend checking contours if the distance is close to PTV (<5 cm). For the lungs, only obvious large errors need to be adjusted.

Automatic delineation of cardiac substructures using a region-based fully convolutional network

PURPOSE

Radiation dose to specific cardiac substructures, such as the atria and ventricles, has been linked to post-treatment toxicity and has shown to be more predictive of these toxicities than dose to the whole heart. A deep learning-based algorithm for automatic generation of these contours is proposed to aid in either retrospective or prospective dosimetric studies to better understand the relationship between radiation dose and toxicities.

METHODS

The proposed method uses a mask-scoring regional convolutional neural network (RCNN) which consists of five major subnetworks: backbone, regional proposal network (RPN), RCNN head, mask head, and mask-scoring head. Multiscale feature maps are learned from computed tomography (CT) via the backbone network. The RPN utilizes these feature maps to detect the location and region-of-interest (ROI) of all substructures, and the final three subnetworks work in series to extract structural information from these ROIs. The network is trained using 55 patient CT datasets, with 22 patients having contrast scans. Threefold cross validation (CV) is used for evaluation on 45 datasets, and a separate cohort of 10 patients are used for holdout evaluation. The proposed method is compared to a 3D UNet.

RESULTS

The proposed method produces contours that are qualitatively similar to the ground truth contours. Quantitatively, the proposed method achieved average Dice score coefficients (DSCs) for the whole heart, chambers, great vessels, coronary arteries, the valves of the heart of 0.96, 0.94, 0.93, 0.66, and 0.77 respectively, outperforming the 3D UNet, which achieved DSCs of 0.92, 0.87, 0.88, 0.48, and 0.59 for the corresponding substructure groups. Mean surface distances (MSDs) between substructures segmented by the proposed method and the ground truth were <2 mm except for the left anterior descending coronary artery and the mitral and tricuspid valves, and <5 mm for all substructures. When dividing results into noncontrast and contrast datasets, the model performed statistically significantly better in terms of DSC, MSD, centroid mean distance (CMD), and volume difference for the chambers and whole heart with contrast. Notably, the presence of contrast did not statistically significantly affect coronary artery segmentation DSC or MSD. After network training, all substructures and the whole heart can be segmented on new datasets in less than 5 s.

CONCLUSIONS

A deep learning network was trained for automatic delineation of cardiac substructures based on CT alone. The proposed method can be used as a tool to investigate the relationship between cardiac substructure dose and treatment toxicities.

Using Auto-Segmentation to Reduce Contouring and Dose Inconsistency in Clinical Trials: The Simulated Impact on RTOG 0617

PURPOSE

Contouring inconsistencies are known but understudied in clinical radiation therapy trials. We applied auto-contouring to the Radiation Therapy Oncology Group (RTOG) 0617 dose escalation trial data. We hypothesized that the trial heart doses were higher than reported due to inconsistent and insufficient heart segmentation. We tested our hypothesis by comparing doses between deep-learning (DL) segmented hearts and trial hearts.

METHODS AND MATERIALS

The RTOG 0617 data were downloaded from The Cancer Imaging Archive; the 442 patients with trial hearts and dose distributions were included. All hearts were resegmented using our DL pipeline and quality assured to meet the requirements for clinical implementation. Dose (V5%, V30%, and mean heart dose) was compared between the 2 sets of hearts (Wilcoxon signed-rank test). Each dose metric was associated with overall survival (Cox proportional hazards). Lastly, 18 volume similarity metrics were assessed for the hearts and correlated with |Dose_DL - Dose_RTOG0617| (linear regression; significance: P ≤ .0028; corrected for 18 tests).

RESULTS

Dose metrics were significantly higher for DL hearts compared with trial hearts (eg, mean heart dose: 15 Gy vs 12 Gy; P = 5.8E-16). All 3 DL heart dose metrics were stronger overall survival predictors than those of the trial hearts (median, P = 2.8E-5 vs 2.0E-4). Thirteen similarity metrics explained |Dose_DL - Dose_RTOG0617|; the axial distance between the 2 centers of mass was the strongest predictor (CENT_Axial; median, R² = 0.47; P = 6.1E-62). CENT_Axial agreed with the qualitatively identified inconsistencies in the superior direction. The trial's qualitative heart contouring score was not correlated with |Dose_DL - Dose_RTOG0617| (median, R² = 0.01; P = .02) or with any of the similarity metrics (median, Rs = 0.13 [range, -0.22 to 0.31]).

CONCLUSIONS

Using a coherent heart definition, as enabled through our open-source DL algorithm, the trial heart doses in RTOG 0617 were found to be significantly higher than previously reported, which may have led to an even more rapid mortality accumulation. Auto-segmentation is likely to reduce contouring and dose inconsistencies and increase the quality of clinical RT trials.

Automatic evaluation of contours in radiotherapy planning utilising conformity indices and machine learning

BACKGROUND AND PURPOSE

Peer-review of Target Volume (TV) and Organ at Risk (OAR) contours in radiotherapy planning are typically conducted visually; this can be time consuming and subject to interobserver variation. This study investigated automatic evaluation of contouring using conformity indices and supervised machine learning.

METHODS

A total of 393 contours from 253 Stereotactic Ablative Body Radiotherapy (SABR) benchmark cases (adrenal gland, liver, pelvic lymph node and spine), delineated by 132 clinicians from 25 centres, were visually evaluated for conformity against gold standard contours. Contours were scored as "pass" or "fail" on visual peer review and six Conformity Indices (CIs) were applied. CI values were mapped to pass/fail scores for each contour and used to train supervised machine learning models. A 5-fold cross validation method was employed to determine the predictive accuracies of each model.

RESULTS

The stomach structure produced models with the highest predictive accuracy overall (96% using Support Vector Machine and Ensemble models), whilst the liver GTV produced models with the lowest predictive accuracy (76% using Logistic Regression). Predictive accuracies across all models ranged from 68-96% (68-87% for TV and 71-96% for OARs).

CONCLUSIONS

Although a final visual review by an experienced clinician is still required, the automatic contour evaluation method could reduce the time for benchmark case reviews by identifying gross contouring errors. This method could be successfully implemented to support departmental training and the continuous assessment of outlining for clinical staff in the peer-review process, to reduce interobserver variability in contouring and improve interpretation of radiological anatomy.