3D Variation in delineation of head and neck organs at risk

Evaluation and failure analysis of four commercial deep learning-based autosegmentation software for abdominal organs at risk

PURPOSE

Deep learning-based segmentation of organs-at-risk (OAR) is emerging to become mainstream in clinical practice because of the superior performance over atlas and model-based autocontouring methods. While several commercial deep learning-based autosegmentation solutions are now available, the implementation of these tools is still at such a primitive stage that acceptance criteria are underdeveloped due to a lack of knowledge about the systems' segmentation tendencies and failure modes. As the starting point of the iterative process of clinical implementation, this study focuses on the outlier analysis of four commercial autocontouring tools for the abdominal OARs.

MATERIALS AND METHODS

The autosegmentation software, developed by Limbus AI, MIM Contour ProtégéAI, Radformation AutoContour, and Siemens syngo.via, were used to segment 111 patient cases. Geometric segmentation accuracy was quantitatively compared with clinical contours using the dice similarity coefficient (DSC) and 95% Hausdorff distance (HD95). The outliers from quantitative evaluations of each software were analyzed for the liver, stomach, and kidneys with the possible causes of outliers summarized into six categories: (1) difference in contouring style or guideline, (2) image acquisition and quality, (3) abnormal anatomy of the OAR, (4) abnormal anatomy of abutting organs/tissues, (5) external/internal devices, and (6) other causes.

RESULTS

For the liver segmentation, the most prominent cause of discrepancies for Limbus, which occurred in four of its six outliers, was the existence of biliary stent or internal/external biliary drain as well as the resulting pneumobilia. Siemens included the abutting organs that shared CT numbers similar to those of the liver in 5/8 outliers. 12 of 13 Radformation's liver segmentation outliers included the heart and/or stomach while MIM not only included the stomach in the presence of barium in 5/11 outliers, but also produced fragmented contours in 5/11 other cases. Only Limbus and Radformation provided stomach segmentation, and imaging with barium contrast directly caused incomplete stomach delineation in 10/12 Limbus outliers and 21/25 Radformation outliers. As for the kidneys, Radformation and Siemens consistently followed the RTOG contouring guidelines, whereas the institutional contours excluded the renal pelvis in some cases, resulting in 19/25 Radformation outliers and 18/23 Siemens outliers. By contrast, Limbus contours appeared to follow different contouring guidelines that exclude the renal pelvis. Fragmented kidney contours were found in 10/15 Limbus outliers and 25/26 MIM outliers. The ones in MIM were directly linked to the use of IV contrast in imaging, but there was not enough evidence to identify the origin of Limbus's fragmented contours.

CONCLUSION

The causes of the segmentation outliers of the four commercial deep learning-based autocontouring solutions were summarized for each OAR. This work can help the vendors improve their autosegmentation software and also inform the users of potential modes of failure when using the tools.

Domain-Adaptive and Per-Fraction Guided Deep Learning Framework for Magnetic Resonance Imaging-Based Segmentation of Organs at Risk in Gynecologic Cancers

Purpose

The integration of magnetic resonance imaging into radiation therapy (RT) treatment necessitates automated segmentation algorithms for fast and accurate adaptive interventions, particularly in magnetic resonance imaging-integrated linear accelerator (MR-linac or MRL) treatment systems. However, the scarcity of data hampers the training of these models. This study aimed to address this shortcoming by developing a synthetic MRL-assisted deep learning framework to establish a robust baseline for organ at risk segmentation on MRL images and enable domain adaptation for automatic delineations during adaptive RT treatments.

Methods and Materials

We used a retrospective data set, comprising 158 patients diagnosed with various gynecologic cancers who underwent computed tomography scanning for RT planning and 25 patients with T2-weighted MRL scans for model fine-tuning, adaptation, and evaluation. A patch-based cycle-consistent generative adversarial network was developed to synthesize MRL images from computed tomography data. Subsequently, a domain-adaptive segmentation network was trained to segment the 6 organs at risk on acquired MRL images. In addition, we employed per-fraction adaptation to enhance anatomical conformity guided by prior treatment fractions of individual patients. A quantitative evaluation and blinded human reader assessment were conducted to establish contour acceptance rates.

Results

The synthetic MRL-assisted model improved organ at risk segmentation accuracy on MRL images, with fraction-adapted contours displaying high anatomical fidelity. Two radiation oncologists reported contour acceptance rates of 100% and 98% for treatment planning after adaptation.

Conclusions

This novel framework holds promise to bridge the semantic gap between computed tomography and magnetic resonance imaging databases, potentially facilitating adaptive RT treatments and reducing treatment times as well as clinician burden. The utility of this framework can extend beyond gynecologic and pelvic cancers.

Deep learning for autosegmentation for radiotherapy treatment planning: State-of-the-art and novel perspectives

The rapid development of artificial intelligence (AI) has gained importance, with many tools already entering our daily lives. The medical field of radiation oncology is also subject to this development, with AI entering all steps of the patient journey. In this review article, we summarize contemporary AI techniques and explore the clinical applications of AI-based automated segmentation models in radiotherapy planning, focusing on delineation of organs at risk (OARs), the gross tumor volume (GTV), and the clinical target volume (CTV). Emphasizing the need for precise and individualized plans, we review various commercial and freeware segmentation tools and also state-of-the-art approaches. Through our own findings and based on the literature, we demonstrate improved efficiency and consistency as well as time savings in different clinical scenarios. Despite challenges in clinical implementation such as domain shifts, the potential benefits for personalized treatment planning are substantial. The integration of mathematical tumor growth models and AI-based tumor detection further enhances the possibilities for refining target volumes. As advancements continue, the prospect of one-stop-shop segmentation and radiotherapy planning represents an exciting frontier in radiotherapy, potentially enabling fast treatment with enhanced precision and individualization.

Comparison of Vendor-Pretrained and Custom-Trained Deep Learning Segmentation Models for Head-and-Neck, Breast, and Prostate Cancers

BACKGROUND/OBJECTIVES

We assessed the influence of local patients and clinical characteristics on the performance of commercial deep learning (DL) segmentation models for head-and-neck (HN), breast, and prostate cancers.

METHODS

Clinical computed tomography (CT) scans and clinically approved contours of 210 patients (53 HN, 49 left breast, 55 right breast, and 53 prostate cancer) were used to train and validate segmentation models integrated within a vendor-supplied DL training toolkit and to assess the performance of both vendor-pretrained and custom-trained models. Four custom models (HN, left breast, right breast, and prostate) were trained and validated with 30 (training)/5 (validation) HN, 34/5 left breast, 39/5 right breast, and 30/5 prostate patients to auto-segment a total of 24 organs at risk (OARs). Subsequently, both vendor-pretrained and custom-trained models were tested on the remaining patients from each group. Auto-segmented contours were evaluated by comparing them with clinically approved contours via the Dice similarity coefficient (DSC) and mean surface distance (MSD). The performance of the left and right breast models was assessed jointly according to ipsilateral/contralateral locations.

RESULTS

The average DSCs for all structures in vendor-pretrained and custom-trained models were as follows: 0.81 ± 0.12 and 0.86 ± 0.11 in HN; 0.67 ± 0.16 and 0.80 ± 0.11 in the breast; and 0.87 ± 0.09 and 0.92 ± 0.06 in the prostate. The corresponding average MSDs were 0.81 ± 0.76 mm and 0.76 ± 0.56 mm (HN), 4.85 ± 2.44 mm and 2.42 ± 1.49 mm (breast), and 2.17 ± 1.39 mm and 1.21 ± 1.00 mm (prostate). Notably, custom-trained models showed significant improvements over vendor-pretrained models for 14 of 24 OARs, reflecting the influence of data/contouring variations in segmentation performance.

CONCLUSIONS

These findings underscore the substantial impact of institutional preferences and clinical practices on the implementation of vendor-pretrained models. We also found that a relatively small amount of institutional data was sufficient to train customized segmentation models with sufficient accuracy.

Artificial intelligence contouring in radiotherapy for organs-at-risk and lymph node areas

INTRODUCTION

The delineation of organs-at-risk and lymph node areas is a crucial step in radiotherapy, but it is time-consuming and associated with substantial user-dependent variability in contouring. Artificial intelligence (AI) appears to be the solution to facilitate and standardize this work. The objective of this study is to compare eight available AI software programs in terms of technical aspects and accuracy for contouring organs-at-risk and lymph node areas with current international contouring recommendations.

MATERIAL AND METHODS

From January-July 2023, we performed a blinded study of the contour scoring of the organs-at-risk and lymph node areas by eight self-contouring AI programs by 20 radiation oncologists. It was a single-center study conducted in radiation department at the Lorraine Cancer Institute. A qualitative analysis of technical characteristics of the different AI programs was also performed. Three adults (two women and one man) and three children (one girl and two boys) provided six whole-body anonymized CT scans, along with two other adult brain MRI scans. Using a scoring scale from 1 to 3 (best score), radiation oncologists blindly assessed the quality of contouring of organs-at-risk and lymph node areas of all scans and MRI data by the eight AI programs. We have chosen to define the threshold of an average score equal to or greater than 2 to characterize a high-performing AI software, meaning an AI with minimal to moderate corrections but usable in clinical routine.

RESULTS

For adults CT scans: There were two AI programs for which the overall average quality score (that is, all areas tested for OARs and lymph nodes) was higher than 2.0: Limbus (overall average score = 2.03 (0.16)) and MVision (overall average score = 2.13 (0.19)). If we only consider OARs for adults, only Limbus, Therapanacea, MVision and Radformation have an average score above 2. For children CT scan, MVision was the only program to have a average score higher than 2 with overall average score = 2.07 (0.19). If we only consider OARs for children, only Limbus and MVision have an average score above 2. For brain MRIs: TheraPanacea was the only program with an average score over 2, for both brain delineation (2.75 (0.35)) and OARs (2.09 (0.19)). The comparative analysis of the technical aspects highlights the similarities and differences between the software. There is no difference in between senior radiation oncologist and residents for OARs contouring.

CONCLUSION

For adult CT-scan, two AI programs on the market, MVision and Limbus, delineate most OARs and lymph nodes areas that are useful in clinical routine. For children CT-scan, only one IA, MVision, program is efficient. For adult brain MRI, Therapancea,only one AI program is efficient.

TRIAL REGISTRATION

CNIL-MR0004 Number HDH434.

Impact of annotation imperfections and auto-curation for deep learning-based organ-at-risk segmentation

Background and purpose

Segmentation imperfections (noise) in radiotherapy organ-at-risk segmentation naturally arise from specialist experience and image quality. Using clinical contours can result in sub-optimal convolutional neural network (CNN) training and performance, but manual curation is costly. We address the impact of simulated and clinical segmentation noise on CNN parotid gland (PG) segmentation performance and provide proof-of-concept for an easily implemented auto-curation countermeasure.

Methods and Materials

The impact of segmentation imperfections was investigated by simulating noise in clean, high-quality segmentations. Curation efficacy was tested by removing lowest-scoring Dice similarity coefficient (DSC) cases early during CNN training, both in simulated (5-fold) and clinical (10-fold) settings, using our full radiotherapy clinical cohort (RTCC; N = 1750 individual PGs). Statistical significance was assessed using Bonferroni-corrected Wilcoxon signed-rank tests. Curation efficacies were evaluated using DSC and mean surface distance (MSD) on in-distribution and out-of-distribution data and visual inspection.

Results

The curation step correctly removed median(range) 98(90-100)% of corrupted segmentations and restored the majority (1.2 %/1.3 %) of DSC lost from training with 30 % corrupted segmentations. This effect was masked when using typical (non-curated) validation data. In RTCC, 20 % curation showed improved model generalizability which significantly improved out-of-distribution DSC and MSD (p < 1.0e-12, p < 1.0e-6). Improved consistency was observed in particularly the medial and anterior lobes.

Conclusions

Up to 30% case removal, the curation benefit outweighed the training variance lost through curation. Considering the notable ease of implementation, high sensitivity in simulations and performance gains already at lower curation fractions, as a conservative middle ground, we recommend 15% curation of training cases when training CNNs using clinical PG contours.

EBV-positive glycoproteins associated with nasopharyngeal carcinoma

Nasopharyngeal carcinoma (NPC) is closely related to Epstein-Barr virus (EBV) infection, and glycosylation of proteins is associated with precancerous lesions and carcinogenesis of NPC, and viral glycoproteins mediates the fusion of viruses with B cells or epithelial cells in the infection stage, promoting the conversion of normal epithelial cells into cancer cells. In the process of occurrence and development of NPC, various glycoproteins in the body promote or inhibit the proliferation, invasion, metastasis, and drug resistance of tumor cells, such as the tumor inhibitory effect of NGX6 and inhibin B (INHBB); the cancer-promoting effect of tenascin-C (TNC), fibronectin 1 (FN1), insulin-like growth factor binding protein-3 (IGFBP3), serglycin, and its core protein; and some effects of glycosylation of immune proteins on immunotherapy in NPC. This article provides an overview of the research progress on the interaction of glycoproteins associated with EBV infection with the occurrence and development of NPC.

Clinical validation of commercial deep-learning based auto-segmentation models for organs at risk in the head and neck region: a single institution study

Purpose

To evaluate organ at risk (OAR) auto-segmentation in the head and neck region of computed tomography images using two different commercially available deep-learning-based auto-segmentation (DLAS) tools in a single institutional clinical applications.

Methods

Twenty-two OARs were manually contoured by clinicians according to published guidelines on planning computed tomography (pCT) images for 40 clinical head and neck cancer (HNC) cases. Automatic contours were generated for each patient using two deep-learning-based auto-segmentation models-Manteia AccuContour and MIM ProtégéAI. The accuracy and integrity of autocontours (ACs) were then compared to expert contours (ECs) using the Sørensen-Dice similarity coefficient (DSC) and Mean Distance (MD) metrics.

Results

ACs were generated for 22 OARs using AccuContour and 17 OARs using ProtégéAI with average contour generation time of 1 min/patient and 5 min/patient respectively. EC and AC agreement was highest for the mandible (DSC 0.90 ± 0.16) and (DSC 0.91 ± 0.03), and lowest for the chiasm (DSC 0.28 ± 0.14) and (DSC 0.30 ± 0.14) for AccuContour and ProtégéAI respectively. Using AccuContour, the average MD was<1mm for 10 of the 22 OARs contoured, 1-2mm for 6 OARs, and 2-3mm for 6 OARs. For ProtégéAI, the average mean distance was<1mm for 8 out of 17 OARs, 1-2mm for 6 OARs, and 2-3mm for 3 OARs.

Conclusions

Both DLAS programs were proven to be valuable tools to significantly reduce the time required to generate large amounts of OAR contours in the head and neck region, even though manual editing of ACs is likely needed prior to implementation into treatment planning. The DSCs and MDs achieved were similar to those reported in other studies that evaluated various other DLAS solutions. Still, small volume structures with nonideal contrast in CT images, such as nerves, are very challenging and will require additional solutions to achieve sufficient results.

Evaluation of Deep Learning Clinical Target Volumes Auto-Contouring for Magnetic Resonance Imaging-Guided Online Adaptive Treatment of Rectal Cancer

Purpose

Segmentation of clinical target volumes (CTV) on medical images can be time-consuming and is prone to interobserver variation (IOV). This is a problem for online adaptive radiation therapy, where CTV segmentation must be performed every treatment fraction, leading to longer treatment times and logistic challenges. Deep learning (DL)-based auto-contouring has the potential to speed up CTV contouring, but its current clinical use is limited. One reason for this is that it can be time-consuming to verify the accuracy of CTV contours produced using auto-contouring, and there is a risk of bias being introduced. To be accepted by clinicians, auto-contouring must be trustworthy. Therefore, there is a need for a comprehensive commissioning framework when introducing DL-based auto-contouring in clinical practice. We present such a framework and apply it to an in-house developed DL model for auto-contouring of the CTV in rectal cancer patients treated with MRI-guided online adaptive radiation therapy.

Methods and Materials

The framework for evaluating DL-based auto-contouring consisted of 3 steps: (1) Quantitative evaluation of the model's performance and comparison with IOV; (2) Expert observations and corrections; and (3) Evaluation of the impact on expected volumetric target coverage. These steps were performed on independent data sets. The framework was applied to an in-house trained nnU-Net model, using the data of 44 rectal cancer patients treated at our institution.

Results

The framework established that the model's performance after expert corrections was comparable to IOV, and although the model introduced a bias, this had no relevant impact on clinical practice. Additionally, we found a substantial time gain without reducing quality as determined by volumetric target coverage.

Conclusions

Our framework provides a comprehensive evaluation of the performance and clinical usability of target auto-contouring models. Based on the results, we conclude that the model is eligible for clinical use.

NRG Oncology Assessment of Artificial Intelligence Deep Learning-Based Auto-segmentation for Radiation Therapy: Current Developments, Clinical Considerations, and Future Directions

Deep learning neural networks (DLNN) in Artificial intelligence (AI) have been extensively explored for automatic segmentation in radiotherapy (RT). In contrast to traditional model-based methods, data-driven AI-based models for auto-segmentation have shown high accuracy in early studies in research settings and controlled environment (single institution). Vendor-provided commercial AI models are made available as part of the integrated treatment planning system (TPS) or as a stand-alone tool that provides streamlined workflow interacting with the main TPS. These commercial tools have drawn clinics' attention thanks to their significant benefit in reducing the workload from manual contouring and shortening the duration of treatment planning. However, challenges occur when applying these commercial AI-based segmentation models to diverse clinical scenarios, particularly in uncontrolled environments. Contouring nomenclature and guideline standardization has been the main task undertaken by the NRG Oncology. AI auto-segmentation holds the potential clinical trial participants to reduce interobserver variations, nomenclature non-compliance, and contouring guideline deviations. Meanwhile, trial reviewers could use AI tools to verify contour accuracy and compliance of those submitted datasets. In recognizing the growing clinical utilization and potential of these commercial AI auto-segmentation tools, NRG Oncology has formed a working group to evaluate the clinical utilization and potential of commercial AI auto-segmentation tools. The group will assess in-house and commercially available AI models, evaluation metrics, clinical challenges, and limitations, as well as future developments in addressing these challenges. General recommendations are made in terms of the implementation of these commercial AI models, as well as precautions in recognizing the challenges and limitations.

Large-scale dose evaluation of deep learning organ contours in head-and-neck radiotherapy by leveraging existing plans

Background and purpose

Retrospective dose evaluation for organ-at-risk auto-contours has previously used small cohorts due to additional manual effort required for treatment planning on auto-contours. We aimed to do this at large scale, by a) proposing and assessing an automated plan optimization workflow that used existing clinical plan parameters and b) using it for head-and-neck auto-contour dose evaluation.

Materials and methods

Our automated workflow emulated our clinic's treatment planning protocol and reused existing clinical plan optimization parameters. This workflow recreated the original clinical plan (P OG ) with manual contours (P MC ) and evaluated the dose effect (P OG - P MC ) on 70 photon and 30 proton plans of head-and-neck patients. As a use-case, the same workflow (and parameters) created a plan using auto-contours (P AC ) of eight head-and-neck organs-at-risk from a commercial tool and evaluated their dose effect (P MC - P AC ).

Results

For plan recreation (P OG - P MC ), our workflow had a median impact of 1.0% and 1.5% across dose metrics of auto-contours, for photon and proton respectively. Computer time of automated planning was 25% (photon) and 42% (proton) of manual planning time. For auto-contour evaluation (P MC - P AC ), we noticed an impact of 2.0% and 2.6% for photon and proton radiotherapy. All evaluations had a median Δ NTCP (Normal Tissue Complication Probability) less than 0.3%.

Conclusions

The plan replication capability of our automated program provides a blueprint for other clinics to perform auto-contour dose evaluation with large patient cohorts. Finally, despite geometric differences, auto-contours had a minimal median dose impact, hence inspiring confidence in their utility and facilitating their clinical adoption.

A multi-modal vision-language pipeline strategy for contour quality assurance and adaptive optimization

Objective.Accurate delineation of organs-at-risk (OARs) is a critical step in radiotherapy. The deep learning generated segmentations usually need to be reviewed and corrected by oncologists manually, which is time-consuming and operator-dependent. Therefore, an automated quality assurance (QA) and adaptive optimization correction strategy was proposed to identify and optimize 'incorrect' auto-segmentations.Approach.A total of 586 CT images and labels from nine institutions were used. The OARs included the brainstem, parotid, and mandible. The deep learning generated contours were compared with the manual ground truth delineations. In this study, we proposed a novel contour quality assurance and adaptive optimization (CQA-AO) strategy, which consists of the following three main components: (1) the contour QA module classified the deep learning generated contours as either accepted or unaccepted; (2) the unacceptable contour categories analysis module provided the potential error reasons (five unacceptable category) and locations (attention heatmaps); (3) the adaptive correction of unacceptable contours module integrate vision-language representations and utilize convex optimization algorithms to achieve adaptive correction of 'incorrect' contours.Main results. In the contour QA tasks, the sensitivity (accuracy, precision) of CQA-AO strategy reached 0.940 (0.945, 0.948), 0.962 (0.937, 0.913), and 0.967 (0.962, 0.957) for brainstem, parotid and mandible, respectively. The unacceptable contour category analysis, the(FI,AccI,Fmicro,Fmacro)of CQA-AO strategy reached (0.901, 0.763, 0.862, 0.822), (0.855, 0.737, 0.837, 0.784), and (0.907, 0.762, 0.858, 0.821) for brainstem, parotid and mandible, respectively. After adaptive optimization correction, the DSC values of brainstem, parotid and mandible have been improved by 9.4%, 25.9%, and 13.5%, and Hausdorff distance values decreased by 62%, 70.6%, and 81.6%, respectively.Significance. The proposed CQA-AO strategy, which combines QA of contour and adaptive optimization correction for OARs contouring, demonstrated superior performance compare to conventional methods. This method can be implemented in the clinical contouring procedures and improve the efficiency of delineating and reviewing workflow.

vOARiability: Interobserver and intermodality variability analysis in OAR contouring from head and neck CT and MR images

BACKGROUND

Accurate and consistent contouring of organs-at-risk (OARs) from medical images is a key step of radiotherapy (RT) cancer treatment planning. Most contouring approaches rely on computed tomography (CT) images, but the integration of complementary magnetic resonance (MR) modality is highly recommended, especially from the perspective of OAR contouring, synthetic CT and MR image generation for MR-only RT, and MR-guided RT. Although MR has been recognized as valuable for contouring OARs in the head and neck (HaN) region, the accuracy and consistency of the resulting contours have not been yet objectively evaluated.

PURPOSE

To analyze the interobserver and intermodality variability in contouring OARs in the HaN region, performed by observers with different level of experience from CT and MR images of the same patients.

METHODS

In the final cohort of 27 CT and MR images of the same patients, contours of up to 31 OARs were obtained by a radiation oncology resident (junior observer, JO) and a board-certified radiation oncologist (senior observer, SO). The resulting contours were then evaluated in terms of interobserver variability, characterized as the agreement among different observers (JO and SO) when contouring OARs in a selected modality (CT or MR), and intermodality variability, characterized as the agreement among different modalities (CT and MR) when OARs were contoured by a selected observer (JO or SO), both by the Dice coefficient (DC) and 95-percentile Hausdorff distance (HD95 $_{95}$ ).

RESULTS

The mean (±standard deviation) interobserver variability was 69.0 ± 20.2% and 5.1 ± 4.1 mm, while the mean intermodality variability was 61.6 ± 19.0% and 6.1 ± 4.3 mm in terms of DC and HD95 $_{95}$ , respectively, across all OARs. Statistically significant differences were only found for specific OARs. The performed MR to CT image registration resulted in a mean target registration error of 1.7 ± 0.5 mm, which was considered as valid for the analysis of intermodality variability.

CONCLUSIONS

The contouring variability was, in general, similar for both image modalities, and experience did not considerably affect the contouring performance. However, the results indicate that an OAR is difficult to contour regardless of whether it is contoured in the CT or MR image, and that observer experience may be an important factor for OARs that are deemed difficult to contour. Several of the differences in the resulting variability can be also attributed to adherence to guidelines, especially for OARs with poor visibility or without distinctive boundaries in either CT or MR images. Although considerable contouring differences were observed for specific OARs, it can be concluded that almost all OARs can be contoured with a similar degree of variability in either the CT or MR modality, which works in favor of MR images from the perspective of MR-only and MR-guided RT.

Multi-organ segmentation of organ-at-risk (OAR's) of head and neck site using ensemble learning technique

INTRODUCTION

This paper presents a novel approach to automate the segmentation of Organ-at-Risk (OAR) in Head and Neck cancer patients using Deep Learning models combined with Ensemble Learning techniques. The study aims to improve the accuracy and efficiency of OAR segmentation, essential for radiotherapy treatment planning.

METHODS

The dataset comprised computed tomography (CT) scans of 182 patients in DICOM format, obtained from an institutional image bank. Experienced Radiation Oncologists manually segmented seven OARs for each scan. Two models, 3D U-Net and 3D DenseNet-FCN, were trained on reduced CT scans (192 × 192 x 128) due to memory limitations. Ensemble Learning techniques were employed to enhance accuracy and segmentation metrics. Testing was conducted on 78 patients from the institutional dataset and an open-source dataset (TCGA-HNSC and Head-Neck Cetuximab) consisting of 31 patient scans.

RESULTS

Using the Ensemble Learning technique, the average dice similarity coefficient for OARs ranged from 0.990 to 0.994, indicating high segmentation accuracy. The 95% Hausdorff distance (mm) ranged from 1.3 to 2.1, demonstrating precise segmentation boundaries.

CONCLUSION

The proposed automated segmentation method achieved efficient and accurate OAR segmentation, surpassing human expert performance in terms of time and accuracy.

IMPLICATIONS FOR PRACTICE

This approach has implications for improving treatment planning and patient care in radiotherapy. By reducing manual segmentation reliance, the proposed method offers significant time savings and potential improvements in treatment planning efficiency and precision for head and neck cancer patients.

A cascade transformer-based model for 3D dose distribution prediction in head and neck cancer radiotherapy

Objective. Radiation therapy is one of the primary methods used to treat cancer in the clinic. Its goal is to deliver a precise dose to the planning target volume while protecting the surrounding organs at risk (OARs). However, the traditional workflow used by dosimetrists to plan the treatment is time-consuming and subjective, requiring iterative adjustments based on their experience. Deep learning methods can be used to predict dose distribution maps to address these limitations.Approach. The study proposes a cascade model for OARs segmentation and dose distribution prediction. An encoder-decoder network has been developed for the segmentation task, in which the encoder consists of transformer blocks, and the decoder uses multi-scale convolutional blocks. Another cascade encoder-decoder network has been proposed for dose distribution prediction using a pyramid architecture. The proposed model has been evaluated using an in-house head and neck cancer dataset of 96 patients and OpenKBP, a public head and neck cancer dataset of 340 patients.Main results. The segmentation subnet achieved 0.79 and 2.71 for Dice and HD95 scores, respectively. This subnet outperformed the existing baselines. The dose distribution prediction subnet outperformed the winner of the OpenKBP2020 competition with 2.77 and 1.79 for dose and dose-volume histogram scores, respectively. Besides, the end-to-end model, including both subnets simultaneously, outperformed the related studies.Significance. The predicted dose maps showed good coincidence with ground-truth, with a superiority after linking with the auxiliary segmentation task. The proposed model outperformed state-of-the-art methods, especially in regions with low prescribed doses. The codes are available athttps://github.com/GhTara/Dose_Prediction.

Reference standard for the evaluation of automatic segmentation algorithms: Quantification of inter observer variability of manual delineation of prostate contour on MRI

PURPOSE

The purpose of this study was to investigate the relationship between inter-reader variability in manual prostate contour segmentation on magnetic resonance imaging (MRI) examinations and determine the optimal number of readers required to establish a reliable reference standard.

MATERIALS AND METHODS

Seven radiologists with various experiences independently performed manual segmentation of the prostate contour (whole-gland [WG] and transition zone [TZ]) on 40 prostate MRI examinations obtained in 40 patients. Inter-reader variability in prostate contour delineations was estimated using standard metrics (Dice similarity coefficient [DSC], Hausdorff distance and volume-based metrics). The impact of the number of readers (from two to seven) on segmentation variability was assessed using pairwise metrics (consistency) and metrics with respect to a reference segmentation (conformity), obtained either with majority voting or simultaneous truth and performance level estimation (STAPLE) algorithm.

RESULTS

The average segmentation DSC for two readers in pairwise comparison was 0.919 for WG and 0.876 for TZ. Variability decreased with the number of readers: the interquartile ranges of the DSC were 0.076 (WG) / 0.021 (TZ) for configurations with two readers, 0.005 (WG) / 0.012 (TZ) for configurations with three readers, and 0.002 (WG) / 0.0037 (TZ) for configurations with six readers. The interquartile range decreased slightly faster between two and three readers than between three and six readers. When using consensus methods, variability often reached its minimum with three readers (with STAPLE, DSC = 0.96 [range: 0.945-0.971] for WG and DSC = 0.94 [range: 0.912-0.957] for TZ, and interquartile range was minimal for configurations with three readers.

CONCLUSION

The number of readers affects the inter-reader variability, in terms of inter-reader consistency and conformity to a reference. Variability is minimal for three readers, or three readers represent a tipping point in the variability evolution, with both pairwise-based metrics or metrics with respect to a reference. Accordingly, three readers may represent an optimal number to determine references for artificial intelligence applications.

Contour subregion error detection methodology using deep learning auto-segmentation

BACKGROUND

Inaccurate manual organ delineation is one of the high-risk failure modes in radiation treatment. Numerous automated contour quality assurance (QA) systems have been developed to assess contour acceptability; however, manual inspection of flagged cases is a time-consuming and challenging process, and can lead to users overlooking the exact error location.

PURPOSE

Our aim is to develop and validate a contour QA system that can effectively detect and visualize subregional contour errors, both qualitatively and quantitatively.

METHODS/MATERIALS

A novel contour subregion error detection (CSED) system was developed using subregional surface distance discrepancies between manual and deep learning auto-segmentation (DLAS) contours. A validation study was conducted using a head and neck public dataset containing 339 cases and evaluated according to knowledge-based pass criteria derived from a clinical training dataset of 60 cases. A blind qualitative evaluation was conducted, comparing the results from the CSED system with manual labels. Subsequently, the CSED-flagged cases were re-examined by a radiation oncologist.

RESULTS

The CSED system could visualize the diverse types of subregional contour errors qualitatively and quantitatively. In the validation dataset, the CSED system resulted in true positive rates (TPR) of 0.814, 0.800, and 0.771; false positive rates (FPR) of 0.310, 0.267, and 0.298; and accuracies of 0.735, 0.759, and 0.730, for brainstem and left and right parotid contours, respectively. The CSED-assisted manual review caught 13 brainstem, 19 left parotid, and 21 right parotid contour errors missed by conventional human review. The TPR/FPR/accuracy of the CSED-assisted manual review improved to 0.836/0.253/0.784, 0.831/0.171/0.830, and 0.808/0.193/0.807 for each structure, respectively. Further, the time savings achieved through CSED-assisted review improved by 75%, with the time for review taking 24.81 ± 12.84, 26.75 ± 10.41, and 28.71 ± 13.72 s for each structure, respectively.

CONCLUSIONS

The CSED system enables qualitative and quantitative detection, localization, and visualization of manual segmentation subregional errors utilizing DLAS contours as references. The use of this system has been shown to help reduce the risk of high-risk failure modes resulting from inaccurate organ segmentation.

Evaluation of different algorithms for automatic segmentation of head-and-neck lymph nodes on CT images

PURPOSE

To investigate the performance of 4 atlas-based (multi-ABAS) and 2 deep learning (DL) solutions for head-and-neck (HN) elective nodes (CTVn) automatic segmentation (AS) on CT images.

MATERIAL AND METHODS

Bilateral CTVn levels of 69 HN cancer patients were delineated on contrast-enhanced planning CT. Ten and 49 patients were used for atlas library and for training a mono-centric DL model, respectively. The remaining 20 patients were used for testing. Additionally, three commercial multi-ABAS methods and one commercial multi-centric DL solution were investigated. Quantitative evaluation was assessed using volumetric Dice Similarity Coefficient (DSC) and 95-percentile Hausdorff distance (HD95%). Blind evaluation was performed for 3 solutions by 4 physicians. One recorded the time needed for manual corrections. A dosimetric study was finally conducted using automated planning.

RESULTS

Overall DL solutions had better DSC and HD95% results than multi-ABAS methods. No statistically significant difference was found between the 2 DL solutions. However, the contours provided by multi-centric DL solution were preferred by all physicians and were also faster to correct (1.1 min vs 4.17 min, on average). Manual corrections for multi-ABAS contours took on average 6.52 min Overall, decreased contour accuracy was observed from CTVn2 to CTVn3 and to CTVn4. Using the AS contours in treatment planning resulted in underdosage of the elective target volume.

CONCLUSION

Among all methods, the multi-centric DL method showed the highest delineation accuracy and was better rated by experts. Manual corrections remain necessary to avoid elective target underdosage. Finally, AS contours help reducing the workload of manual delineation task.

Consistency in contouring of organs at risk by artificial intelligence vs oncologists in head and neck cancer patients

BACKGROUND

In the Danish Head and Neck Cancer Group (DAHANCA) 35 trial, patients are selected for proton treatment based on simulated reductions of Normal Tissue Complication Probability (NTCP) for proton compared to photon treatment at the referring departments. After inclusion in the trial, immobilization, scanning, contouring and planning are repeated at the national proton centre. The new contours could result in reduced expected NTCP gain of the proton plan, resulting in a loss of validity in the selection process. The present study evaluates if contour consistency can be improved by having access to AI (Artificial Intelligence) based contours.

MATERIALS AND METHODS

The 63 patients in the DAHANCA 35 pilot trial had a CT from the local DAHANCA centre and one from the proton centre. A nationally validated convolutional neural network, based on nnU-Net, was used to contour OARs on both scans for each patient. Using deformable image registration, local AI and oncologist contours were transferred to the proton centre scans for comparison. Consistency was calculated with the Dice Similarity Coefficient (DSC) and Mean Surface Distance (MSD), comparing contours from AI to AI and oncologist to oncologist, respectively. Two NTCP models were applied to calculate NTCP for xerostomia and dysphagia.

RESULTS

The AI contours showed significantly better consistency than the contours by oncologists. The median and interquartile range of DSC was 0.85 [0.78 - 0.90] and 0.68 [0.51 - 0.80] for AI and oncologist contours, respectively. The median and interquartile range of MSD was 0.9 mm [0.7 - 1.1] mm and 1.9 mm [1.5 - 2.6] mm for AI and oncologist contours, respectively. There was no significant difference in Δ NTCP.

CONCLUSIONS

The study showed that OAR contours made by the AI algorithm were more consistent than those made by oncologists. No significant impact on the Δ NTCP calculations could be discerned.

Automatic Segmentation with Deep Learning in Radiotherapy

This review provides a formal overview of current automatic segmentation studies that use deep learning in radiotherapy. It covers 807 published papers and includes multiple cancer sites, image types (CT/MRI/PET), and segmentation methods. We collect key statistics about the papers to uncover commonalities, trends, and methods, and identify areas where more research might be needed. Moreover, we analyzed the corpus by posing explicit questions aimed at providing high-quality and actionable insights, including: "What should researchers think about when starting a segmentation study?", "How can research practices in medical image segmentation be improved?", "What is missing from the current corpus?", and more. This allowed us to provide practical guidelines on how to conduct a good segmentation study in today's competitive environment that will be useful for future research within the field, regardless of the specific radiotherapeutic subfield. To aid in our analysis, we used the large language model ChatGPT to condense information.

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

BACKGROUND AND PURPOSE

To date, data used in the development of Deep Learning-based automatic contouring (DLC) algorithms have been largely sourced from single geographic populations. This study aimed to evaluate the risk of population-based bias by determining whether the performance of an autocontouring system is impacted by geographic population.

MATERIALS AND METHODS

80 Head Neck CT deidentified scans were collected from four clinics in Europe (n = 2) and Asia (n = 2). A single observer manually delineated 16 organs-at-risk in each. Subsequently, the data was contoured using a DLC solution, and trained using single institution (European) data. Autocontours were compared to manual delineations using quantitative measures. A Kruskal-Wallis test was used to test for any difference between populations. Clinical acceptability of automatic and manual contours to observers from each participating institution was assessed using a blinded subjective evaluation.

RESULTS

Seven organs showed a significant difference in volume between groups. Four organs showed statistical differences in quantitative similarity measures. The qualitative test showed greater variation in acceptance of contouring between observers than between data from different origins, with greater acceptance by the South Korean observers.

CONCLUSION

Much of the statistical difference in quantitative performance could be explained by the difference in organ volume impacting the contour similarity measures and the small sample size. However, the qualitative assessment suggests that observer perception bias has a greater impact on the apparent clinical acceptability than quantitatively observed differences. This investigation of potential geographic bias should extend to more patients, populations, and anatomical regions in the future.

Three-Dimensional Measurement of the Uterus on Magnetic Resonance Images: Development and Performance Analysis of an Automated Deep-Learning Tool

Uterus measurements are useful for assessing both the treatment and follow-ups of gynaecological patients. The aim of our study was to develop a deep learning (DL) tool for fully automated measurement of the three-dimensional size of the uterus on magnetic resonance imaging (MRI). In this single-centre retrospective study, 900 cases were included to train, validate, and test a VGG-16/VGG-11 convolutional neural network (CNN). The ground truth was manual measurement. The performance of the model was evaluated using the objective key point similarity (OKS), the mean difference in millimetres, and coefficient of determination R2. The OKS of our model was 0.92 (validation) and 0.96 (test). The average deviation and R2 coefficient between the AI measurements and the manual ones were, respectively, 3.9 mm and 0.93 for two-point length, 3.7 mm and 0.94 for three-point length, 2.6 mm and 0.93 for width, 4.2 mm and 0.75 for thickness. The inter-radiologist variability was 1.4 mm. A three-dimensional automated measurement was obtained in 1.6 s. In conclusion, our model was able to locate the uterus on MRIs and place measurement points on it to obtain its three-dimensional measurement with a very good correlation compared to manual measurements.

Dosimetric comparison of autocontouring techniques for online adaptive proton therapy

Objective.Anatomical and daily set-up uncertainties impede high precision delivery of proton therapy. With online adaptation, the daily plan is reoptimized on an image taken shortly before the treatment, reducing these uncertainties and, hence, allowing a more accurate delivery. This reoptimization requires target and organs-at-risk (OAR) contours on the daily image, which need to be delineated automatically since manual contouring is too slow. Whereas multiple methods for autocontouring exist, none of them are fully accurate, which affects the daily dose. This work aims to quantify the magnitude of this dosimetric effect for four contouring techniques.Approach.Plans reoptimized on automatic contours are compared with plans reoptimized on manual contours. The methods include rigid and deformable registration (DIR), deep-learning based segmentation and patient-specific segmentation.Main results.It was found that independently of the contouring method, the dosimetric influence of usingautomaticOARcontoursis small (<5% prescribed dose in most cases), with DIR yielding the best results. Contrarily, the dosimetric effect of using theautomatic target contourwas larger (>5% prescribed dose in most cases), indicating that manual verification of that contour remains necessary. However, when compared to non-adaptive therapy, the dose differences caused by automatically contouring the target were small and target coverage was improved, especially for DIR.Significance.The results show that manual adjustment of OARs is rarely necessary and that several autocontouring techniques are directly usable. Contrarily, manual adjustment of the target is important. This allows prioritizing tasks during time-critical online adaptive proton therapy and therefore supports its further clinical implementation.

Contouring quality assurance methodology based on multiple geometric features against deep learning auto-segmentation

BACKGROUND

Contouring error is one of the top failure modes in radiation treatment. Multiple efforts have been made to develop tools to automatically detect segmentation errors. Deep learning-based auto-segmentation (DLAS) has been used as a baseline for flagging manual segmentation errors, but those efforts are limited to using only one or two contour comparison metrics.

PURPOSE

The purpose of this research is to develop an improved contouring quality assurance system to identify and flag manual contouring errors.

METHODS AND MATERIALS

DLAS contours were used as a reference to compare with manually segmented contours. A total of 27 geometric agreement metrics were determined from the comparisons between the two segmentation approaches. Feature selection was performed to optimize the training of a machine learning classification model to identify potential contouring errors. A public dataset with 339 cases was used to train and test the classifier. Four independent classifiers were trained using five-fold cross validation, and the predictions from each classifier were ensembled using soft voting. The trained model was validated on a held-out testing dataset. An additional independent clinical dataset with 60 cases was used to test the generalizability of the model. Model predictions were reviewed by an expert to confirm or reject the findings.

RESULTS

The proposed machine learning multiple features (ML-MF) approach outperformed traditional nonmachine-learning-based approaches that are based on only one or two geometric agreement metrics. The machine learning model achieved recall (precision) values of 0.842 (0.899), 0.762 (0.762), 0.727 (0.842), and 0.773 (0.773) for Brainstem, Parotid_L, Parotid_R, and mandible contours, respectively compared to 0.526 (0.909), 0.619 (0.765), 0.682 (0.882), 0.773 (0.568) for an approach based solely on Dice similarity coefficient values. In the external validation dataset, 66.7, 93.3, 94.1, and 58.8% of flagged cases were confirmed to have contouring errors by an expert for Brainstem, Parotid_L, Parotid_R, and mandible contours, respectively.

CONCLUSIONS

The proposed ML-MF approach, which includes multiple geometric agreement metrics to flag manual contouring errors, demonstrated superior performance in comparison to traditional methods. This method is easy to implement in clinical practice and can help to reduce the significant time and labor costs associated with manual segmentation and review.

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

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

Large scale crowdsourced radiotherapy segmentations across a variety of cancer anatomic sites

Clinician generated segmentation of tumor and healthy tissue regions of interest (ROIs) on medical images is crucial for radiotherapy. However, interobserver segmentation variability has long been considered a significant detriment to the implementation of high-quality and consistent radiotherapy dose delivery. This has prompted the increasing development of automated segmentation approaches. However, extant segmentation datasets typically only provide segmentations generated by a limited number of annotators with varying, and often unspecified, levels of expertise. In this data descriptor, numerous clinician annotators manually generated segmentations for ROIs on computed tomography images across a variety of cancer sites (breast, sarcoma, head and neck, gynecologic, gastrointestinal; one patient per cancer site) for the Contouring Collaborative for Consensus in Radiation Oncology challenge. In total, over 200 annotators (experts and non-experts) contributed using a standardized annotation platform (ProKnow). Subsequently, we converted Digital Imaging and Communications in Medicine data into Neuroimaging Informatics Technology Initiative format with standardized nomenclature for ease of use. In addition, we generated consensus segmentations for experts and non-experts using the Simultaneous Truth and Performance Level Estimation method. These standardized, structured, and easily accessible data are a valuable resource for systematically studying variability in segmentation applications.

E pluribus unum: prospective acceptability benchmarking from the Contouring Collaborative for Consensus in Radiation Oncology crowdsourced initiative for multiobserver segmentation

Purpose

Contouring Collaborative for Consensus in Radiation Oncology (C3RO) is a crowdsourced challenge engaging radiation oncologists across various expertise levels in segmentation. An obstacle to artificial intelligence (AI) development is the paucity of multiexpert datasets; consequently, we sought to characterize whether aggregate segmentations generated from multiple nonexperts could meet or exceed recognized expert agreement.

Approach

Participants who contoured ≥ 1 region of interest (ROI) for the breast, sarcoma, head and neck (H&N), gynecologic (GYN), or gastrointestinal (GI) cases were identified as a nonexpert or recognized expert. Cohort-specific ROIs were combined into single simultaneous truth and performance level estimation (STAPLE) consensus segmentations.STAPLE nonexpert ROIs were evaluated againstSTAPLE expert contours using Dice similarity coefficient (DSC). The expert interobserver DSC (IODSC expert ) was calculated as an acceptability threshold betweenSTAPLE nonexpert andSTAPLE expert . To determine the number of nonexperts required to match theIODSC expert for each ROI, a single consensus contour was generated using variable numbers of nonexperts and then compared to theIODSC expert .

Results

For all cases, the DSC values forSTAPLE nonexpert versusSTAPLE expert were higher than comparator expertIODSC expert for most ROIs. The minimum number of nonexpert segmentations needed for a consensus ROI to achieveIODSC expert acceptability criteria ranged between 2 and 4 for breast, 3 and 5 for sarcoma, 3 and 5 for H&N, 3 and 5 for GYN, and 3 for GI.

Conclusions

Multiple nonexpert-generated consensus ROIs met or exceeded expert-derived acceptability thresholds. Five nonexperts could potentially generate consensus segmentations for most ROIs with performance approximating experts, suggesting nonexpert segmentations as feasible cost-effective AI inputs.

Implementation of a Commercial Deep Learning-Based Auto Segmentation Software in Radiotherapy: Evaluation of Effectiveness and Impact on Workflow

Proper delineation of both target volumes and organs at risk is a crucial step in the radiation therapy workflow. This process is normally carried out manually by medical doctors, hence demanding timewise. To improve efficiency, auto-contouring methods have been proposed. We assessed a specific commercial software to investigate its impact on the radiotherapy workflow on four specific disease sites: head and neck, prostate, breast, and rectum. For the present study, we used a commercial deep learning-based auto-segmentation software, namely Limbus Contour (LC), Version 1.5.0 (Limbus AI Inc., Regina, SK, Canada). The software uses deep convolutional neural network models based on a U-net architecture, specific for each structure. Manual and automatic segmentation were compared on disease-specific organs at risk. Contouring time, geometrical performance (volume variation, Dice Similarity Coefficient-DSC, and center of mass shift), and dosimetric impact (DVH differences) were evaluated. With respect to time savings, the maximum advantage was seen in the setting of head and neck cancer with a 65%-time reduction. The average DSC was 0.72. The best agreement was found for lungs. Good results were highlighted for bladder, heart, and femoral heads. The most relevant dosimetric difference was in the rectal cancer case, where the mean volume covered by the 45 Gy isodose was 10.4 cm³ for manual contouring and 289.4 cm³ for automatic segmentation. Automatic contouring was able to significantly reduce the time required in the procedure, simplifying the workflow, and reducing interobserver variability. Its implementation was able to improve the radiation therapy workflow in our department.

Comparison of atlas-based and deep learning methods for organs at risk delineation on head-and-neck CT images using an automated treatment planning system

BACKGROUND AND PURPOSE

To investigate the performance of head-and-neck (HN) organs-at-risk (OAR) automatic segmentation (AS) using four atlas-based (ABAS) and two deep learning (DL) solutions.

MATERIAL AND METHODS

All patients underwent iodine contrast-enhanced planning CT. Fourteen OAR were manually delineated. DL.1 and DL.2 solutions were trained with 63 mono-centric patients and > 1000 multi-centric patients, respectively. Ten and 15 patients with varied anatomies were selected for the atlas library and for testing, respectively. The evaluation was based on geometric indices (DICE coefficient and 95th percentile-Hausdorff Distance (HD_95%)), time needed for manual corrections and clinical dosimetric endpoints obtained using automated treatment planning.

RESULTS

Both DICE and HD_95% results indicated that DL algorithms generally performed better compared with ABAS algorithms for automatic segmentation of HN OAR. However, the hybrid-ABAS (ABAS.3) algorithm sometimes provided the highest agreement to the reference contours compared with the 2 DL. Compared with DL.2 and ABAS.3, DL.1 contours were the fastest to correct. For the 3 solutions, the differences in dose distributions obtained using AS contours and AS + manually corrected contours were not statistically significant. High dose differences could be observed when OAR contours were at short distances to the targets. However, this was not always interrelated.

CONCLUSION

DL methods generally showed higher delineation accuracy compared with ABAS methods for AS segmentation of HN OAR. Most ABAS contours had high conformity to the reference but were more time consuming than DL algorithms, especially when considering the computing time and the time spent on manual corrections.

Evaluating the Effectiveness of Deep Learning Contouring across Multiple Radiotherapy Centres

Background and purpose

Deep learning contouring (DLC) has the potential to decrease contouring time and variability of organ contours. This work evaluates the effectiveness of DLC for prostate and head and neck across four radiotherapy centres using a commercial system.

Materials and methods

Computed tomography scans of 123 prostate and 310 head and neck patients were evaluated. Besides one head and neck model, generic DLC models were used. Contouring time using centres' existing clinical methods and contour editing time after DLC were compared. Timing was evaluated using paired and non-paired studies. Commercial software or in-house scripts assessed dice similarity coefficient (DSC) and distance to agreement (DTA). One centre assessed head and neck inter-observer variability.

Results

The mean contouring time saved for prostate structures using DLC compared to the existing clinical method was 5.9 ± 3.5 min. The best agreement was shown for the femoral heads (median DSC 0.92 ± 0.03, median DTA 1.5 ± 0.3 mm) and the worst for the rectum (median DSC 0.68 ± 0.04, median DTA 4.6 ± 0.6 mm). The mean contouring time saved for head and neck structures using DLC was 16.2 ± 8.6 min. For one centre there was no DLC time-saving compared to an atlas-based method. DLC contours reduced inter-observer variability compared to manual contours for the brainstem, left parotid gland and left submandibular gland.

Conclusions

Generic prostate and head and neck DLC models can provide time-savings which can be assessed with paired or non-paired studies to integrate with clinical workload. Reducing inter-observer variability potential has been shown.

Clinical evaluation of deep learning and atlas-based auto-segmentation for critical organs at risk in radiation therapy

INTRODUCTION

Contouring organs at risk (OARs) is a time-intensive task that is a critical part of radiation therapy. Atlas-based automatic segmentation has shown some success at reducing this time burden on practitioners; however, this method often requires significant manual editing to reach a clinically accurate standard. Deep learning (DL) auto-segmentation has recently emerged as a promising solution. This study compares the accuracy of DL and atlas-based auto-segmentation in relation to clinical 'gold standard' reference contours.

METHODS

Ninety CT datasets (30 head and neck, 30 thoracic, 30 pelvic) were automatically contoured using both atlas and DL segmentation techniques. Sixteen critical OARs were then quantitatively measured for accuracy using the Dice similarity coefficient (DSC) and Hausdorff distance (HD). Qualitative analysis was performed to visually classify the accuracy of each structure into one of four explicitly defined categories. Additionally, the time to edit atlas and DL contours to a clinically acceptable level was recorded for a subset of 9 OARs.

RESULTS

Of the 16 OARs analysed, DL delivered statistically significant improvements over atlas segmentation in 13 OARs measured with DSC, 12 OARs measured with HD, and 12 OARs measured qualitatively. The mean editing time for the subset of DL contours was 50%, 23% and 61% faster (all P < 0.05) than that of atlas segmentation for the head and neck, thorax, and pelvis respectively.

CONCLUSIONS

Deep learning segmentation comprehensively outperformed atlas-based contouring for the majority of evaluated OARs. Improvements were observed in geometric accuracy and visual acceptability, while editing time was reduced leading to increased workflow efficiency.

Consensus Quality Measures and Dose Constraints for Head and Neck Cancer with an emphasis on Oropharyngeal and Laryngeal Cancer from the Veterans Affairs Radiation Oncology Quality Surveillance Program and American Society for Radiation Oncology Expert Panel

PURPOSE

Safeguarding high-quality care using evidence-based radiation therapy for patients with head and neck cancer is crucial to improving oncologic outcomes, including survival and quality of life.

METHODS AND MATERIALS

The Veterans Administration (VA) National Radiation Oncology Program established the VA Radiation Oncology Quality Surveillance Program (VAROQS) to develop clinical quality measures (QM) in head and neck cancer. As part of the development of QM, the VA commissioned, along with the American Society for Radiation Oncology, a blue-ribbon panel comprising experts in head and neck cancer, to develop QM.

RESULTS

We describe the methods used to develop QM and the final consensus QM, as well as aspirational and surveillance QM, which capture all aspects of the continuum of patient care from initial patient work-up, radiation treatment planning and delivery, and follow-up care, as well as dose volume constraints.

CONCLUSION

These QM are intended for use as part of ongoing quality surveillance for veterans receiving radiation therapy throughout the VA as well as outside the VA. They may also be used by the non-VA community as a basic measure of quality care for head and neck cancer patients receiving radiation.

Clinical Validation of a Deep-Learning Segmentation Software in Head and Neck: An Early Analysis in a Developing Radiation Oncology Center

Background: Organs at risk (OARs) delineation is a crucial step of radiotherapy (RT) treatment planning workflow. Time-consuming and inter-observer variability are main issues in manual OAR delineation, mainly in the head and neck (H & N) district. Deep-learning based auto-segmentation is a promising strategy to improve OARs contouring in radiotherapy departments. A comparison of deep-learning-generated auto-contours (AC) with manual contours (MC) was performed by three expert radiation oncologists from a single center. Methods: Planning computed tomography (CT) scans of patients undergoing RT treatments for H&N cancers were considered. CT scans were processed by Limbus Contour auto-segmentation software, a commercial deep-learning auto-segmentation based software to generate AC. H&N protocol was used to perform AC, with the structure set consisting of bilateral brachial plexus, brain, brainstem, bilateral cochlea, pharyngeal constrictors, eye globes, bilateral lens, mandible, optic chiasm, bilateral optic nerves, oral cavity, bilateral parotids, spinal cord, bilateral submandibular glands, lips and thyroid. Manual revision of OARs was performed according to international consensus guidelines. The AC and MC were compared using the Dice similarity coefficient (DSC) and 95% Hausdorff distance transform (DT). Results: A total of 274 contours obtained by processing CT scans were included in the analysis. The highest values of DSC were obtained for the brain (DSC 1.00), left and right eye globes and the mandible (DSC 0.98). The structures with greater MC editing were optic chiasm, optic nerves and cochleae. Conclusions: In this preliminary analysis, deep-learning auto-segmentation seems to provide acceptable H&N OAR delineations. For less accurate organs, AC could be considered a starting point for review and manual adjustment. Our results suggest that AC could become a useful time-saving tool to optimize workload and resources in RT departments.

Automatic Delineation of Gross Tumor Volume Based on Magnetic Resonance Imaging by Performing a Novel Semisupervised Learning Framework in Nasopharyngeal Carcinoma

PURPOSE

We aimed to validate the accuracy and clinical value of a novel semisupervised learning framework for gross tumor volume (GTV) delineation in nasopharyngeal carcinoma.

METHODS AND MATERIALS

Two hundred fifty-eight patients with magnetic resonance imaging data sets were divided into training (n = 180), validation (n = 20), and testing (n = 58) cohorts. Ground truth contours of nasopharynx GTV (GTVnx) and node GTV (GTVnd) were manually delineated by 2 experienced radiation oncologists. Twenty percent (n = 36) labeled and 80% (n = 144) unlabeled images were used to train the model, producing model-generated contours for patients from the testing cohort. Nine experienced experts were invited to revise model-generated GTV in 20 randomly selected patients from the testing cohort. Six junior oncologists were asked to delineate GTV in 12 randomly selected patients from the testing cohort without and with the assistance of the model, and revision degrees were compared under these 2 modes. The Dice similarity coefficient (DSC) was used to quantify the accuracy of the model.

RESULTS

The model-generated contours showed a high accuracy compared with ground truth contours, with an average DSC score of 0.83 and 0.80 for GTVnx and GTVnd, respectively. There was no significant difference in DSC score between T1-2 and T3-4 patients (0.81 vs 0.83; P = .223), or between N1-2 and N3 patients (0.80 vs 0.79; P = .807). The mean revision degree was lower than 10% in 19 (95%) patients for GTVnx and in 16 (80%) patients for GTVnd. With assistance of the model, the mean revision degree for GTVnx and GTVnd by junior oncologists was reduced from 25.63% to 7.75% and from 21.38% to 14.44%, respectively. Meanwhile, the delineating efficiency was improved by over 60%.

CONCLUSIONS

The proposed semisupervised learning-based model showed a high accuracy for delineating GTV of nasopharyngeal carcinoma. It was clinically applicable and could assist junior oncologists to improve GTV contouring accuracy and save contouring time.

Comparative evaluation of a prototype deep learning algorithm for autosegmentation of normal tissues in head and neck radiotherapy

PURPOSE

To introduce and validate a newly developed deep-learning (DL) auto-segmentation algorithm for head and neck (HN) organs at risk (OARs) and to compare its performance with a published commercial algorithm.

METHODS

A total of 864 HN cancer cases were available to train and evaluate a prototype algorithm. The algorithm is based on a fully convolutional network with combined U-Net and V-net. A Dice loss plus Cross-Entropy Loss function with Adam optimizer was used in training. For 75 validation cases, OAR sets were generated with three DL-based models (A: the prototype model trained with gold data, B: a commercial software trained with the same data, and C: the same software trained with data from another institution). The auto-segmented structures were evaluated with Dice similarity coefficient (DSC), Hausdorff distance (HD), voxel-penalty metric (VPM) and DSC of area under dose-volume histograms. A subjective qualitative evaluation was performed on 20 random cases.

RESULTS

Overall trend was for the prototype algorithm to be the closest to the gold data by all five metrics. The average DSC/VPM/HD for algorithms A, B, and C were 0.81/84.1/1.6 mm, 0.74/62.8/3.2 mm, and 0.66/46.8/3.3 mm, respectively. 93% of model A structures were evaluated to be clinically useful.

CONCLUSION

The superior performance of the prototype was validated, even when trained with the same data. In addition to the challenges of perfecting the algorithms, the auto-segmentation results can differ when the same algorithm is trained at different institutions.

Target Definition in MR-Guided Adaptive Radiotherapy for Head and Neck Cancer

In recent years, MRI-guided radiotherapy (MRgRT) has taken an increasingly important position in image-guided radiotherapy (IGRT). Magnetic resonance imaging (MRI) offers superior soft tissue contrast in anatomical imaging compared to computed tomography (CT), but also provides functional and dynamic information with selected sequences. Due to these benefits, in current clinical practice, MRI is already used for target delineation and response assessment in patients with head and neck squamous cell carcinoma (HNSCC). Because of the close proximity of target areas and radiosensitive organs at risk (OARs) during HNSCC treatment, MRgRT could provide a more accurate treatment in which OARs receive less radiation dose. With the introduction of several new radiotherapy techniques (i.e., adaptive MRgRT, proton therapy, adaptive cone beam computed tomography (CBCT) RT, (daily) adaptive radiotherapy ensures radiation dose is accurately delivered to the target areas. With the integration of a daily adaptive workflow, interfraction changes have become visible, which allows regular and fast adaptation of target areas. In proton therapy, adaptation is even more important in order to obtain high quality dosimetry, due to its susceptibility for density differences in relation to the range uncertainty of the protons. The question is which adaptations during radiotherapy treatment are oncology safe and at the same time provide better sparing of OARs. For an optimal use of all these new tools there is an urgent need for an update of the target definitions in case of adaptive treatment for HNSCC. This review will provide current state of evidence regarding adaptive target definition using MR during radiotherapy for HNSCC. Additionally, future perspectives for adaptive MR-guided radiotherapy will be discussed.

Association between radiation dose to organs at risk and acute patient reported outcome during radiation treatment for head and neck cancers

BACKGROUND

Associations between patient-reported outcomes and dose to organs at risk (OARs) may promote management and guide future investigations.

METHODS

We retrospectively evaluated PROs and OAR dose in head and neck (H&N) cancer.

RESULTS

In 169 patients, we identified weak associations between: "Difficulty swallowing/chewing" and increased mean RT dose to the oral cavity, larynx, pharyngeal constrictor muscles (PCM) and contralateral parotid; "choking/coughing" and larynx mean dose; "problems with mucus in mouth and throat" and oral cavity, contralateral parotid mean dose and parotid V30, contralateral submandibular gland and PCM mean dose; "difficulty with voice/speech" and oral cavity, contralateral parotid, contralateral submandibular gland and larynx mean dose; and "dry mouth" and ipsilateral submandibular gland, oral cavity and PCM mean dose.

CONCLUSION

We identified weak associations between PRO and dose to OARs-these data can guide on treatment management, patient counseling, and serve as a baseline for future investigations.

International assessment of interobserver reproducibility of flap delineation in head and neck carcinoma

Background: Several reports have suggested that radiotherapy after reconstructive surgery for head and neck cancer (HNC), could have deleterious effects on the flaps with respect to functional outcomes. To predict and prevent toxicities, flap delineation should be accurate and reproducible. The objective of the present study was to evaluate the interobserver variability of frequent types of flaps used in HNC, based on the recent GORTEC atlas.Materials and methods: Each member of an international working group (WG) consisting of 14 experts delineated the flaps on a CT set from six patients. Each patient had one of the five most commonly used flaps in HNC: a regional pedicled pectoralis major myocutaneous flap, a local pedicled rotational soft tissue facial artery musculo-mucosal (FAMM) (2 patients), a fasciocutaneous radial forearm free flap, a soft tissue anterolateral thigh (ALT) free flap, or a fibular free flap. The WG's contours were compared to a reference contour, validated by a surgeon and a radiologist specializing in HNC. Contours were considered as reproducible if the median Dice Similarity Coefficient (DSC) was > 0.7.Results: The median volumes of the six flaps delineated by the WG were close to the reference contour value, with approximately 50 cc for the pectoral, fibula, and ALT flaps, 20 cc for the radial forearm, and up to 10 cc for the FAMM. The volumetric ratio was thus close to the optimal value of 100% for all flaps. The median DSC obtained by the WG compared to the reference for the pectoralis flap, the FAMM, the radial forearm flap, ALT flap, and the fibular flap were 0.82, 0.40, 0.76, 0.81, and 0.76, respectively.Conclusions: This study showed that the delineation of four main flaps used for HNC was reproducible. The delineation of the FAMM, however, requires close cooperation between radiologist, surgeon and radiation oncologist because of the poor visibility of this flap on CT and its small size.

Machine learning-based detection of aberrant deep learning segmentations of target and organs at risk for prostate radiotherapy using a secondary segmentation algorithm

Objective.The output of a deep learning (DL) auto-segmentation application should be reviewed, corrected if needed and approved before being used clinically. This verification procedure is labour-intensive, time-consuming and user-dependent, which potentially leads to significant errors with impact on the overall treatment quality. Additionally, when the time needed to correct auto-segmentations approaches the time to delineate target and organs at risk from scratch, the usability of the DL model can be questioned. Therefore, an automated quality assurance framework was developed with the aim to detect in advance aberrant auto-segmentations.Approach. Five organs (prostate, bladder, anorectum, femoral head left and right) were auto-delineated on CT acquisitions for 48 prostate patients by an in-house trained primary DL model. An experienced radiation oncologist assessed the correctness of the model output and categorised the auto-segmentations into two classes whether minor or major adaptations were needed. Subsequently, an independent, secondary DL model was implemented to delineate the same structures as the primary model. Quantitative comparison metrics were calculated using both models' segmentations and used as input features for a machine learning classification model to predict the output quality of the primary model.Main results. For every organ, the approach of independent validation by the secondary model was able to detect primary auto-segmentations that needed major adaptation with high sensitivity (recall = 1) based on the calculated quantitative metrics. The surface DSC and APL were found to be the most indicated parameters in comparison to standard quantitative metrics for the time needed to adapt auto-segmentations.Significance. This proposed method includes a proof of concept for the use of an independent DL segmentation model in combination with a ML classifier to improve time saving during QA of auto-segmentations. The integration of such system into current automatic segmentation pipelines can increase the efficiency of the radiotherapy contouring workflow.

Geometric and Dosimetric Evaluation of the Automatic Delineation of Organs at Risk (OARs) in Non-Small-Cell Lung Cancer Radiotherapy Based on a Modified DenseNet Deep Learning Network

Purpose

To introduce an end-to-end automatic segmentation model for organs at risk (OARs) in thoracic CT images based on modified DenseNet, and reduce the workload of radiation oncologists.

Materials and Methods

The computed tomography (CT) images of 36 lung cancer patients were included in this study, of which 27 patients’ images were randomly selected as the training set, 9 patients’ as the testing set. The validation set was generated by cross validation and 6 patients’ images were randomly selected from the training set during each epoch as the validation set. The autosegmentation task of the left and right lungs, spinal cord, heart, trachea and esophagus was implemented, and the whole training time was approximately 5 hours. Geometric evaluation metrics including the Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95) and average surface distance (ASD), were used to assess the autosegmentation performance of OARs based on the proposed model and were compared with those based on U-Net as benchmarks. Then, two sets of treatment plans were optimized based on the manually contoured targets and OARs (Plan1), as well as the manually contours targets and the automatically contoured OARs (Plan2). Dosimetric parameters, including Dmax, Dmean and Vx, of OARs were obtained and compared.

Results

The DSC, HD95 and ASD of the proposed model were better than those of U-Net. The differences in the DSC of the spinal cord and esophagus, differences in the HD95 of the spinal cord, heart, trachea and esophagus, as well as differences in the ASD of the spinal cord were statistically significant between the two models (P<0.05). The differences in the dose-volume parameters of the two sets of plans were not statistically significant (P>0.05). Moreover, compared with manual segmentation, autosegmentation significantly reduced the contouring time by nearly 40.7% (P<0.05).

Conclusions

The bilateral lungs, spinal cord, heart and trachea could be accurately delineated using the proposed model in this study; however, the automatic segmentation effect of the esophagus must still be further improved. The concept of feature map reuse provides a new idea for automatic medical image segmentation.

Optimising a 3D convolutional neural network for head and neck computed tomography segmentation with limited training data

•

Convolutional neural networks (CNNs) are used for auto-segmentation in radiotherapy.

•

However, CNNs rely on large, high-quality datasets: a scarcity in radiotherapy.

•

We develop a CNN model, trained with limited data, for accurate segmentation.

•

Multiple experiments were performed to optimise key features of our custom model.

•

Our model is competitive with state-of-the-art methods on a public dataset.

Background and purpose

Convolutional neural networks (CNNs) are increasingly used to automate segmentation for radiotherapy planning, where accurate segmentation of organs-at-risk (OARs) is crucial. Training CNNs often requires large amounts of data. However, large, high quality datasets are scarce. The aim of this study was to develop a CNN capable of accurate head and neck (HN) 3D auto-segmentation of planning CT scans using a small training dataset (34 CTs).

Materials and Method

Elements of our custom CNN architecture were varied to optimise segmentation performance. We tested and evaluated the impact of: using multiple contrast channels for the CT scan input at specific soft tissue and bony anatomy windows, resize vs. transpose convolutions, and loss functions based on overlap metrics and cross-entropy in different combinations. Model segmentation performance was compared with the inter-observer deviation of two doctors’ gold standard segmentations using the 95th percentile Hausdorff distance and mean distance-to-agreement (mDTA). The best performing configuration was further validated on a popular public dataset to compare with state-of-the-art (SOTA) auto-segmentation methods.

Results

Our best performing CNN configuration was competitive with current SOTA methods when evaluated on the public dataset with mDTA of (0.81±0.31) mm for the brainstem, (0.20±0.08) mm for the mandible, (0.77±0.14) mm for the left parotid and (0.81±0.28) mm for the right parotid.

Conclusions

Through careful tuning and customisation we trained a 3D CNN with a small dataset to produce segmentations of HN OARs with an accuracy that is comparable with inter-clinician deviations. Our proposed model performed competitively with current SOTA methods.

Robustness and reproducibility of radiomics in T2 weighted images from magnetic resonance image guided linear accelerator in a phantom study

PURPOSE

Quantitative radiomics features extracted from medical images have been shown to provide value in predicting clinical outcomes. The study for robustness and reproducibility of radiomics features obtained with magnetic resonance image guided linear accelerator (MR-Linac) is insufficient. The objective of this work was to investigate the stability of radiomics features extracted from T2-weighted images of MR-Linac for five common effect factors.

MATERIALS AND METHOD

In this work, ten jellies, five fruits/vegetables, and a dynamic phantom were used to evaluate the impact of test-retest, intraobserver, varied thicknesses, radiation, and motion. These phantoms were scanned on a 1.5 T MRI system of MR-Linac. For test-retest data, the phantoms were scanned twice with repositioning within 15 min. To assess for intraobserver comparison, the segmentation of MR images was repeated by one observer in a double-blind manner. Three slice thicknesses (1.2 mm, 2.4 mm, and 4.8 mm) were used to select robust features that were insensitive to different thicknesses. The effect of radiation on features was studied by acquiring images when the beam was on. Common movement images of patients during radiotherapy were simulated by a dynamic phantom with five motion states to study the motion effect. A total of 1409 radiomics features, including shape features, first-order features, and texture features, were extracted from the original, wavelet, square, logarithmic, exponential and gradient images. The robustness and reproducibility features were evaluated using the concordance correlation coefficient (CCC).

RESULT

The intraobserver group had the most robust features (936/1079, 86.7%), while the group of motion effects had the lowest robustness (56/936, 6.0%), followed by the group of different thickness cohorts (374/936, 40.0%). The stability of features in the test-retest and radiation groups was 1072 of 1312 (81.7%) and 810 of 936 (86.5%), respectively. Overall, 25 of 1409 (2.4%) radiomics features remained robust in all five tests, mostly focusing on the image type of the wavelet. The number of stable features extracted from when the beam was on was less than that extracted when the beam was off. Shape features were the most robust of all of the features in all of the groups, excluding the motion group.

CONCLUSION

Compared with other factors fewer features remained robust to the effect of motion. This result emphasizes the need to consider the effect of respiration motion. The study for T2-weighted images from MR-Linac under different conditions will help us to build a robust predictive model applicable for radiotherapy.

Knowledge-based quality control of organ delineations in radiation therapy

PURPOSE

To reduce the likelihood of errors in organ delineations used for radiotherapy treatment planning, a knowledge-based quality control (KBQC) system, which discriminates between valid and anomalous delineations is developed.

METHOD AND MATERIALS

The KBQC is comprised of a group-wise inference system and anomaly detection modules trained using historical priors from 296 locally advanced lung and prostate cancer patient computational tomographies (CTs). The inference system discriminates different organs based on shape, relational, and intensity features. For a given delineated image set, the inference system solves a combinatorial optimization problem that results in an organ group whose relational features follow those of the training set considering the posterior probabilities obtained from support vector machine (SVM), discriminant subspace ensemble (DSE), and artificial neural network (ANN) classifiers. These classifiers are trained on nonrelational features with a 10-fold cross-validation scheme. The anomaly detection module is a bank of ANN autoencoders, each corresponding with an organ, trained on nonrelational features. A heuristic rule detects anomalous organs that exceed predefined organ-specific tolerances for the feature reconstruction error and the classifier's posterior probabilities. Independent data sets with anomalous delineations were used to test the overall performance of the KBQC system. The anomalous delineations were manually manipulated, computer-generated, or propagated based on a transformation obtained by imperfect registrations. Both peer-review-based scoring system and shape similarity coefficient (DSC) were used to label regions of interest (ROIs) as normal or anomalous in two independent test cohorts.

RESULTS

The accuracy of the classifiers was ≥ $\ge$ 99.8%, and the minimum per-class F1-scores were 0.99, 0.99, and 0.98 for SVM, DSE, and ANN, respectively. The group-wise inference system reduced the miss-classification likelihood for the test data set with anomalous delineations compared to each individual classifier and a fused classifier that used the average posterior probability of all classifiers. For 15 independent locally advanced lung patients, the system detected > $>$ 79% of the anomalous ROIs. For 1320 auto-segmented abdominopelvic organs, the anomaly detection system identified anomalous delineations, which also had low Dice similarity coefficient values with respect to manually delineated organs in the training data set.

CONCLUSION

The KBQC system detected anomalous delineations with superior accuracy compared to classification methods that judge only based on posterior probabilities.

Multi-organ auto-delineation in head-and-neck MRI for radiation therapy using regional convolutional neural network

Magnetic resonance imaging (MRI) allows accurate and reliable organ delineation for many disease sites in radiation therapy because MRI is able to offer superb soft-tissue contrast. Manual organ-at-risk delineation is labor-intensive and time-consuming. This study aims to develop a deep-learning-based automated multi-organ segmentation method to release the labor and accelerate the treatment planning process for head-and-neck (HN) cancer radiotherapy. A novel regional convolutional neural network (R-CNN) architecture, namely, mask scoring R-CNN, has been developed in this study. In the proposed model, a deep attention feature pyramid network is used as a backbone to extract the coarse features given by MRI, followed by feature refinement using R-CNN. The final segmentation is obtained through mask and mask scoring networks taking those refined feature maps as input. With the mask scoring mechanism incorporated into conventional mask supervision, the classification error can be highly minimized in conventional mask R-CNN architecture. A cohort of 60 HN cancer patients receiving external beam radiation therapy was used for experimental validation. Five-fold cross-validation was performed for the assessment of our proposed method. The Dice similarity coefficients of brain stem, left/right cochlea, left/right eye, larynx, left/right lens, mandible, optic chiasm, left/right optic nerve, oral cavity, left/right parotid, pharynx, and spinal cord were 0.89 ± 0.06, 0.68 ± 0.14/0.68 ± 0.18, 0.89 ± 0.07/0.89 ± 0.05, 0.90 ± 0.07, 0.67 ± 0.18/0.67 ± 0.10, 0.82 ± 0.10, 0.61 ± 0.14, 0.67 ± 0.11/0.68 ± 0.11, 0.92 ± 0.07, 0.85 ± 0.06/0.86 ± 0.05, 0.80 ± 0.13, and 0.77 ± 0.15, respectively. After the model training, all OARs can be segmented within 1 min.

Evaluating Automatic Segmentation for Swallowing-Related Organs for Head and Neck Cancer

Purpose: To evaluate the accuracy of deep-learning-based auto-segmentation of the superior constrictor, middle constrictor, inferior constrictor, and larynx in comparison with a traditional multi-atlas-based method. Methods and Materials: One hundred and five computed tomography image datasets from 83 head and neck cancer patients were retrospectively collected and the superior constrictor, middle constrictor, inferior constrictor, and larynx were analyzed for deep-learning versus multi-atlas-based segmentation. Eighty-three computed tomography images (40 diagnostic computed tomography and 43 planning computed tomography) were used for training the convolutional neural network, and for atlas-based model training. The remaining 22 computed tomography datasets were used for validation of the atlas-based auto-segmentation versus deep-learning-based auto-segmentation contours, both of which were compared with the corresponding manual contours. Quantitative measures included Dice similarity coefficient, recall, precision, Hausdorff distance, 95th percentile of Hausdorff distance, and mean surface distance. Dosimetric differences between the auto-generated contours and manual contours were evaluated. Subjective evaluation was obtained from 3 clinical observers to blindly score the autosegmented structures based on the percentage of slices that require manual modification. Results: The deep-learning-based auto-segmentation versus atlas-based auto-segmentation results were compared for the superior constrictor, middle constrictor, inferior constrictor, and larynx. The mean Dice similarity coefficient values for the 4 structures were 0.67, 0.60, 0.65, and 0.84 for deep-learning-based auto-segmentation, whereas atlas-based auto-segmentation has Dice similarity coefficient results at 0.45, 0.36, 0.50, and 0.70, respectively. The mean 95th percentile of Hausdorff distance (cm) for the 4 structures were 0.41, 0.57, 0.59, and 0.54 for deep-learning-based auto-segmentation, but 0.78, 0.95, 0.96, and 1.23 for atlas-based auto-segmentation results, respectively. Similar mean dose differences were obtained from the 2 sets of autosegmented contours compared to manual contours. The dose–volume discrepancies and the average modification rates were higher with the atlas-based auto-segmentation contours. Conclusion: Swallowing-related structures are more accurately generated with DL-based versus atlas-based segmentation when compared with manual contours.

Improving radiation physics, tumor visualisation, and treatment quantification in radiotherapy with spectral or dual-energy CT

Over the past decade, spectral or dual‐energy CT has gained relevancy, especially in oncological radiology. Nonetheless, its use in the radiotherapy (RT) clinic remains limited. This review article aims to give an overview of the current state of spectral CT and to explore opportunities for applications in RT.

In this article, three groups of benefits of spectral CT over conventional CT in RT are recognized. Firstly, spectral CT provides more information of physical properties of the body, which can improve dose calculation. Furthermore, it improves the visibility of tumors, for a wide variety of malignancies as well as organs‐at‐risk OARs, which could reduce treatment uncertainty. And finally, spectral CT provides quantitative physiological information, which can be used to personalize and quantify treatment.

Cascaded deep learning-based auto-segmentation for head and neck cancer patients: Organs at risk on T2-weighted magnetic resonance imaging

PURPOSE

To investigate multiple deep learning methods for automated segmentation (auto-segmentation) of the parotid glands, submandibular glands, and level II and level III lymph nodes on magnetic resonance imaging (MRI). Outlining radiosensitive organs on images used to assist radiation therapy (radiotherapy) of patients with head and neck cancer (HNC) is a time-consuming task, in which variability between observers may directly impact on patient treatment outcomes. Auto-segmentation on computed tomography imaging has been shown to result in significant time reductions and more consistent outlines of the organs at risk.

METHODS

Three convolutional neural network (CNN)-based auto-segmentation architectures were developed using manual segmentations and T2-weighted MRI images provided from the American Association of Physicists in Medicine (AAPM) radiotherapy MRI auto-contouring (RT-MAC) challenge dataset (n = 31). Auto-segmentation performance was evaluated with segmentation similarity and surface distance metrics on the RT-MAC dataset with institutional manual segmentations (n = 10). The generalizability of the auto-segmentation methods was assessed on an institutional MRI dataset (n = 10).

RESULTS

Auto-segmentation performance on the RT-MAC images with institutional segmentations was higher than previously reported MRI methods for the parotid glands (Dice: 0.860 ± 0.067, mean surface distance [MSD]: 1.33 ± 0.40 mm) and the first report of MRI performance for submandibular glands (Dice: 0.830 ± 0.032, MSD: 1.16 ± 0.47 mm). We demonstrate that high-resolution auto-segmentations with improved geometric accuracy can be generated for the parotid and submandibular glands by cascading a localizer CNN and a cropped high-resolution CNN. Improved MSDs were observed between automatic and manual segmentations of the submandibular glands when a low-resolution auto-segmentation was used as prior knowledge in the second-stage CNN. Reduced auto-segmentation performance was observed on our institutional MRI dataset when trained on external RT-MAC images; only the parotid gland auto-segmentations were considered clinically feasible for manual correction (Dice: 0.775 ± 0.105, MSD: 1.20 ± 0.60 mm).

CONCLUSIONS

This work demonstrates that CNNs are a suitable method to auto-segment the parotid and submandibular glands on MRI images of patients with HNC, and that cascaded CNNs can generate high-resolution segmentations with improved geometric accuracy. Deep learning methods may be suitable for auto-segmentation of the parotid glands on T2-weighted MRI images from different scanners, but further work is required to improve the performance and generalizability of these methods for auto-segmentation of the submandibular glands and lymph nodes.