Evaluation of a delineation software for cardiac atlas-based autosegmentation: An example of the use of artificial intelligence in modern radiotherapy

Cardiac substructure delineation in radiation therapy - A state-of-the-art review

Delineation of cardiac substructures is crucial for a better understanding of radiation-related cardiotoxicities and to facilitate accurate and precise cardiac dose calculation for developing and applying risk models. This review examines recent advancements in cardiac substructure delineation in the radiation therapy (RT) context, aiming to provide a comprehensive overview of the current level of knowledge, challenges and future directions in this evolving field. Imaging used for RT planning presents challenges in reliably visualising cardiac anatomy. Although cardiac atlases and contouring guidelines aid in standardisation and reduction of variability, significant uncertainties remain in defining cardiac anatomy. Coupled with the inherent complexity of the heart, this necessitates auto-contouring for consistent large-scale data analysis and improved efficiency in prospective applications. Auto-contouring models, developed primarily for breast and lung cancer RT, have demonstrated performance comparable to manual contouring, marking a significant milestone in the evolution of cardiac delineation practices. Nevertheless, several key concerns require further investigation. There is an unmet need for expanding cardiac auto-contouring models to encompass a broader range of cancer sites. A shift in focus is needed from ensuring accuracy to enhancing the robustness and accessibility of auto-contouring models. Addressing these challenges is paramount for the integration of cardiac substructure delineation and associated risk models into routine clinical practice, thereby improving the safety of RT for future cancer patients.

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.

Radiation Exposure of Cardiac Conduction Nodes During Breast Proton Therapy

Purpose

The exposition of cardiac conduction system during breast radiation therapy has never been studied, despite the increasing use of intensity-modulated radiation therapy, which exposes larger volume to low-dose bath. We evaluated conduction node exposure during breast irradiation with volumetric modulated arc therapy and estimated the potential dosimetric benefit with intensity-modulated proton therapy.

Materials and Methods

Atrioventricular (AVN) and sinoatrial (SAN) nodes were retrospectively delineated according to published guidelines on the simulation computed tomography scans of 12 breast cancer patients having undergone conserving surgery and adjuvant locoregional volumetric modulated arc therapy. Intensity-modulated proton therapy treatment was replanned on the simulation computed tomography scans for all breast cancer patients. Mean and maximum doses delivered to the SAN and the AVN were retrieved and compared. Correlation coefficients were calculated between doses to the SAN or the AVN and the whole heart.

Results

Average mean doses delivered to the SAN and AVN were 2.8 and 2.3 Gy, respectively, for left-sided irradiation and 9.6 and 3.6 Gy, respectively, for right-sided irradiation. Average maximum doses to the SAN and AVN were 3.5 Gy and 2.8 Gy, respectively, for left-sided irradiation and 13.1 and 4.6 Gy, respectively, for right-sided irradiation. Intensity-modulated proton therapy significantly reduced mean and maximum doses to the SAN and AVN. Correlations between doses to the SAN or AVN and whole heart were usually significant.

Conclusion

SAN and AVN can be substantially exposed during breast volumetric modulated arc therapy, especially for right-sided irradiation. Cardiotoxicity studies evaluating conduction node exposure might define dose constraints and criteria for additional cardiac-sparing techniques, such as respiratory techniques or proton therapy, which could benefit patients with underlying rhythmic or conduction disorders.

Dosimetric evaluation of the benefit of deep inspiration breath hold (DIBH) for locoregional irradiation of right breast cancer with volumetric modulated arctherapy (VMAT)

INTRODUCTION

Right-lateralized cardiac substructures can be substantially exposed during right breast cancer (R-BC) radiotherapy. The cardiac benefit of deep inspiration breath hold (DIBH) is established in combination with volumetric modulated arctherapy (VMAT) for left breast cancer with regional node irradiation but is unknown for R-BC. This study evaluated the dosimetric benefit of DIBH for locoregional irradiation of R-BC with VMAT.

MATERIAL AND METHODS

All patients treated for R-BC with adjuvant locoregional DIBH-VMAT in the Department of Radiation Oncology of the Institut Curie (Paris, France) until December 2022 were included, corresponding to 15 patients. FB- and DIBH-VMAT plans were compared both for a normofractionated regimen (50 Gy/25fx) used for treatment and a replanned hypofractionated regimen (40 Gy/15fx). Dose to the heart, cardiac substructures (sinoatrial node (SAN), atrio-ventricular node (AVN), right coronary artery, left anterior descending coronary artery, left ventricle), ipsilateral lung and liver were retrieved and compared.

RESULTS

Mean heart dose (MHD) was 3.33 Gy with FB vs. 3.10 Gy with DIBH on normofractionated plans (p = 0.489), and 2.58 Gy with FB vs. 2.41 Gy with DIBH on hypofractionated plan (p = 0.489). The benefit of DIBH was not significant for any cardiac substructure. The most exposed cardiac substructure were the SAN (mean dose of 6.62 Gy for FB- and 5.64 Gy for DIBH-VMAT on normofractionated plans) and the RCA (mean dose of 4.21 Gy for FB- and 4.06 Gy for DIBH-VMAT on normofractionated plans). The maximum benefit was observed for the RCA with a median individual dose reduction of 0.84 Gy on normofractionated plans (p = 0.599). No significant dosimetric difference were observed for right lung. Liver mean dose was significantly lower with DIBH with median values decreasing from 2.54 Gy to 0.87 Gy (p = 0.01).

CONCLUSION

Adding DIBH to efficient cardiac-sparing radiotherapy techniques, such as VMAT, is not justified in the general case for locoregional R-BC irradiation. Specific R-BC patient subpopulations who could benefit from additional DIBH combination with locoregional VMAT are yet to be identified.

CArdiac and REspiratory adaptive Computed Tomography (CARE-CT): a proof-of-concept digital phantom study

Current respiratory 4DCT imaging for high-dose rate thoracic radiotherapy treatments are negatively affected by the complex interaction of cardiac and respiratory motion. We propose an imaging method to reduce artifacts caused by thoracic motion, CArdiac and REspiratory adaptive CT (CARE-CT), that monitors respiratory motion and ECG signals in real-time, triggering CT acquisition during combined cardiac and respiratory bins. Using a digital phantom, conventional 4DCT and CARE-CT acquisitions for nineteen patient-measured physiological traces were simulated. Ten respiratory bins were acquired for conventional 4DCT scans and ten respiratory bins during cardiac diastole were acquired for CARE-CT scans. Image artifacts were quantified for 10 common thoracic organs at risk (OAR) substructures using the differential normalized cross correlation between axial slices (ΔNCC), mean squared error (MSE) and sensitivity. For all images, on average, CARE-CT improved the ΔNCC for 18/19 and the MSE and sensitivity for all patient traces. The ΔNCC was reduced for all cardiac OARs (mean reduction 21%). The MSE was reduced for all OARs (mean reduction 36%). In the digital phantom study, the average scan time was increased from 1.8 ± 0.4 min to 7.5 ± 2.2 min with a reduction in average beam on time from 98 ± 28 s to 45 s using CARE-CT compared to conventional 4DCT. The proof-of-concept study indicates the potential for CARE-CT to image the thorax in real-time during the cardiac and respiratory cycle simultaneously, to reduce image artifacts for common thoracic OARs.

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.

Cardiotoxicity model-based patient selection for Hodgkin lymphoma proton therapy

INTRODUCTION

Hodgkin lymphoma (HL) is a highly curable hematological malignancy. Consolidation radiation therapy techniques have made significant progresses to improve organ-at-risk sparing in order to reduce late radiation-induced toxicity. Recent technical breakthroughs notably include intensity modulated proton therapy (IMPT), which has demonstrated a major dosimetric benefit at the cardiac level for mediastinal HL patients. However, its implementation in clinical practice is still challenging, notably due to the limited access to proton therapy facilities. In this context, the purpose of this study was to estimate the benefit of IMPT for HL proton therapy for diverse cardiac adverse events and to propose a general frame for mediastinal HL patient selection strategy for IMPT based on cardiotoxicity reduction, patient clinical factors, and IMPT treatment availability.

MATERIAL AND METHODS

This retrospective dosimetric study included 30 mediastinal HL patients treated with VMAT. IMPT plans were generated on the initial simulation scans. Dose to the heart, to the left ventricle and to the valves were retrieved to calculate the relative risk (RR) of ischemic heart disease (IHD), congestive heart failure (CHF) and valvular disease (VD). Composite relative risk reduction (cRRR) of late cardiotoxicity, between VMAT and IMPT, were calculated as the weighted mean of relative risk reduction for IHD, CHF and VD, calculated across a wide range of cardiovascular risk factor combinations. The proportion of mediastinal HL patients who could benefit from IMPT was estimated in European countries, based on the country population and on the number of active gantries, to propose country-specific cRRR thresholds for patient selection.

RESULTS

Compared with VMAT, IMPT significantly reduced average mean doses to the heart (2.36 Gy vs 0.99 Gy, p < 0.01), to the left ventricle (0.67 Gy vs 0.03, p < 0.01) and to the valves (1.29 Gy vs. 0.06, p < 0.01). For a HL patient without cardiovascular risk factor other than anthracycline-based chemotherapy, the relative risks of late cardiovascular complications were significantly lower after IMPT compared with VMAT for ischemic heart disease (1.07 vs 1.17, p < 0.01), for congestive heart failure (2.84 vs. 3.00, p < 0.01), and for valvular disease (1.01 vs. 1.06, p < 0.01). The median cRRR of cardiovascular adverse events with IMPT was 4.8%, ranging between 0.1% and 30.5%, depending on the extent of radiation fields and on the considered cardiovascular risk factors. The estimated proportion of HL patients currently treatable with IMPT in European countries with proton therapy facilities ranged between 8.0% and 100% depending on the country, corresponding to cRRR thresholds ranging from 24.0% to 0.0%.

CONCLUSION

While a statistically significant clinical benefit is theoretically expected for ischemic heart disease, cardiac heart failure and valvular disease for mediastinal HL patients with IMPT, the overall cardiotoxicity risk reduction is notable only for a minority of patients. In the context of limited IMPT availability, this study proposed a general model-based selection approach for mediastinal HL patient based on calculated cardiotoxicity reduction, taking into consideration patient clinical characteristics and IMPT facility availability.

Atlas-based auto-segmentation for delineating the heart and cardiac substructures in breast cancer radiation therapy

BACKGROUND

This study aimed to develop and validate an automatic multi-atlas segmentation method for delineating the heart and substructures in breast cancer radiation therapy (RT).

MATERIAL AND METHODS

The atlas database consisted of non-contrast-enhanced planning CT scans from 42 breast cancer patients, each with one manual delineation of the heart and 22 cardiac substructures. Half of the patients were scanned during free-breathing, the rest were scanned during a deep inspiration breath-hold. The auto-segmentation was developed in the MIM software system and validated geometrically and dosimetrically in two steps: The first validation in a small dataset to ensure consistency of the atlas. This was succeeded by a final test where multiple manual delineations in CT scans of 12 breast cancer patients were compared to the auto-segmentation. For geometric evaluation, the dice similarity coefficient (DSC) and the mean surface distance (MSD) were used. For dosimetric evaluation, the RT doses to each substructure in the manual and the automatic delineations were compared.

RESULTS

In the first validation, a high geometric and dosimetric performance between the automatic and manual delineations was observed for all substructures. The final test confirmed a high agreement between the automatic and manual delineations for the heart (DSC = 0.94) and the cardiac chambers (DSC: 0.75-0.86). The difference in MSD between the automatic and manual delineations was low (<4 mm) in all structures. Finally, a high correlation between mean RT doses for the automatic and the manual delineations was observed for the heart and substructures.

CONCLUSIONS

An automatic segmentation tool for delineation of the heart and substructures in breast cancer RT was developed and validated with a high correlation between the automatic and manual delineations. The atlas is pivotal for large-scale evaluations of radiation-associated heart disease.

Hadrontherapy techniques for breast cancer

Radiotherapy plays a key role in breast cancer treatment, and recent technical advances have been made to improve the therapeutic window by limiting the risk of radiation-induced toxicity or by increasing tumor control. Hadrontherapy is a form a radiotherapy relying on particle beams; compared with photon beams, particle beams have specific physical, radiobiological and immunological properties, which can be valuable in diverse clinical situations. To date, available hadrontherapy techniques for breast cancer irradiation include proton therapy, carbon ion radiation therapy, fast neutron therapy and boron neutron capture therapy. This review analyzes the current rationale and level of evidence for each hadrontherapy technique for breast cancer.

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.

Artificial intelligence in radiation oncology

Artificial intelligence (AI) is a computer science that tries to mimic human-like intelligence in machines that use computer software and algorithms to perform specific tasks without direct human input. Machine learning (ML) is a subunit of AI that uses data-driven algorithms that learn to imitate human behavior based on a previous example or experience. Deep learning is an ML technique that uses deep neural networks to create a model. The growth and sharing of data, increasing computing power, and developments in AI have initiated a transformation in healthcare. Advances in radiation oncology have produced a significant amount of data that must be integrated with computed tomography imaging, dosimetry, and imaging performed before each fraction. Of the many algorithms used in radiation oncology, has advantages and limitations with different computational power requirements. The aim of this review is to summarize the radiotherapy (RT) process in workflow order by identifying specific areas in which quality and efficiency can be improved by ML. The RT stage is divided into seven stages: patient evaluation, simulation, contouring, planning, quality control, treatment application, and patient follow-up. A systematic evaluation of the applicability, limitations, and advantages of AI algorithms has been done for each stage.

Radio-induced cardiotoxicity: From physiopathology and risk factors to adaptation of radiotherapy treatment planning and recommended cardiac follow-up

Cancer and cardiovascular disease (CVD) are the leading cause of mortality worldwide, and breast cancer (BC) the most common malignancy affecting women worldwide. Radiotherapy is an important component of BC treatment and participates in CVD occurrence. It seems, therefore, crucial to gather both radiation oncology and cardiology medical fields to improve the follow-up quality of our BC patients. This review aims at updating our knowledge regarding cardiotoxicities risk factors, and consequently, doses constraints in case of 3D-conformal and IMRT treatment planning. Then we will develop how to reduce cardiac exposure and what kind of cardiac follow-up we could recommend to our breast cancer patients.