Inter-observer variability in the delineation of pharyngo-laryngeal tumor, parotid glands and cervical spinal cord: comparison between CT-scan and MRI

Multi-modal segmentation with missing image data for automatic delineation of gross tumor volumes in head and neck cancers

BACKGROUND

Head and neck (HN) gross tumor volume (GTV) auto-segmentation is challenging due to the morphological complexity and low image contrast of targets. Multi-modality images, including computed tomography (CT) and positron emission tomography (PET), are used in the routine clinic to assist radiation oncologists for accurate GTV delineation. However, the availability of PET imaging may not always be guaranteed.

PURPOSE

To develop a deep learning segmentation framework for automated GTV delineation of HN cancers using a combination of PET/CT images, while addressing the challenge of missing PET data.

METHODS

Two datasets were included for this study: Dataset I: 524 (training) and 359 (testing) oropharyngeal cancer patients from different institutions with their PET/CT pairs provided by the HECKTOR Challenge; Dataset II: 90 HN patients(testing) from a local institution with their planning CT, PET/CT pairs. To handle potentially missing PET images, a model training strategy named the "Blank Channel" method was implemented. To simulate the absence of a PET image, a blank array with the same dimensions as the CT image was generated to meet the dual-channel input requirement of the deep learning model. During the model training process, the model was randomly presented with either a real PET/CT pair or a blank/CT pair. This allowed the model to learn the relationship between the CT image and the corresponding GTV delineation based on available modalities. As a result, our model had the ability to handle flexible inputs during prediction, making it suitable for cases where PET images are missing. To evaluate the performance of our proposed model, we trained it using training patients from Dataset I and tested it with Dataset II. We compared our model (Model 1) with two other models which were trained for specific modality segmentations: Model 2 trained with only CT images, and Model 3 trained with real PET/CT pairs. The performance of the models was evaluated using quantitative metrics, including Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff Distance (HD95). In addition, we evaluated our Model 1 and Model 3 using the 359 test cases in Dataset I.

RESULTS

Our proposed model(Model 1) achieved promising results for GTV auto-segmentation using PET/CT images, with the flexibility of missing PET images. Specifically, when assessed with only CT images in Dataset II, Model 1 achieved DSC of 0.56 ± 0.16, MSD of 3.4 ± 2.1 mm, and HD95 of 13.9 ± 7.6 mm. When the PET images were included, the performance of our model was improved to DSC of 0.62 ± 0.14, MSD of 2.8 ± 1.7 mm, and HD95 of 10.5 ± 6.5 mm. These results are comparable to those achieved by Model 2 and Model 3, illustrating Model 1's effectiveness in utilizing flexible input modalities. Further analysis using the test dataset from Dataset I showed that Model 1 achieved an average DSC of 0.77, surpassing the overall average DSC of 0.72 among all participants in the HECKTOR Challenge.

CONCLUSIONS

We successfully refined a multi-modal segmentation tool for accurate GTV delineation for HN cancer. Our method addressed the issue of missing PET images by allowing flexible data input, thereby providing a practical solution for clinical settings where access to PET imaging may be limited.

Adaptive radiotherapy for head and neck cancer: Pitfalls and possibilities from the radiation oncologist's point of view

BACKGROUND

Patients with head and neck cancer (HNC) may experience substantial anatomical changes during the course of radiotherapy treatment. The implementation of adaptive radiotherapy (ART) proves effective in managing the consequent impact on the planned dose distribution.

METHODS

This narrative literature review comprehensively discusses the diverse strategies of ART in HNC and the documented dosimetric and clinical advantages associated with these approaches, while also addressing the current challenges for integration of ART into clinical practice.

RESULTS AND CONCLUSION

Although based on mainly non-randomized and retrospective trials, there is accumulating evidence that ART has the potential to reduce toxicity and improve quality of life and tumor control in HNC patients treated with RT. However, several questions remain regarding accurate patient selection, the ideal frequency and timing of replanning, and the appropriate way for image registration and dose calculation. Well-designed randomized prospective trials, with a predetermined protocol for both image registration and dose summation, are urgently needed to further investigate the dosimetric and clinical benefits of ART.

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.

Construction and Application of a Nomogram for Predicting Benign and Malignant Parotid Tumors

OBJECTIVE

A prediction model of benign and malignant differentiation was established by magnetic resonance signs of parotid gland tumors to provide an important basis for the preoperative diagnosis and treatment of parotid gland tumor patients.

METHODS

The data from 138 patients (modeling group) who were diagnosed based on a pathologic evaluation in the Department of Stomatology of Jilin University from June 2019 to August 2021 were retrospectively analyzed. The independent factors influencing benign and malignant differentiation of parotid tumors were selected by logistic regression analysis, and a mathematical prediction model for benign and malignant tumors was established. The data from 35 patients (validation group) who were diagnosed based on pathologic evaluation from September 2021 to February 2022 were collected for verification.

RESULTS

Univariate and multivariate logistic regression analysis showed that tumor morphology, tumor boundary, tumor signal, and tumor apparent diffusion coefficient (ADC) were independent risk factors for predicting benign and malignant parotid gland tumors ( P < 0.05). Based on multivariate logistic regression analysis of the modeling group, a mathematical prediction model was established as follows: Y = the ex/(1 + ex) and X = 0.385 + (1.416 × tumor morphology) + (1.473 × tumor border) + (1.306 × tumor signal) + (2.312 × tumor ADC value). The results showed that the area under the receiver operating characteristic curve of the model was 0.832 (95% confidence interval, 0.75-0.91), the sensitivity was 82.6%, and the specificity was 70.65%. The validity of the model was verified using validation group data, for which the sensitivity was 85.71%, the specificity was 96.4%, and the correct rate was 94.3%. The results showed that the area under receiver operating characteristic curve was 0.936 (95% confidence interval, 0.83-0.98).

CONCLUSIONS

Combined with tumor morphology, tumor ADC, tumor boundary, and tumor signal, the established prediction model provides an important reference for preoperative diagnosis of benign and malignant parotid gland tumors.

Impact of bowel space contouring variability on radiation dose and volume assessments in treatment planning for gynaecologic cancers

INTRODUCTION

Correlations between radiation dose/volume measures and small bowel (SB) toxicity are inconsistent in the medical literature. We assessed the impact of inter-provider variation in bowel bag contouring technique on estimates of radiation dose received by the SB during pelvic radiotherapy.

METHODS

Ten radiation oncologists contoured rectum, bladder and bowel bags on treatment planning computed tomography (CT) scans of two patients receiving adjuvant radiation for endometrial cancer. A radiation plan was generated for each patient and used to determine the radiation dose/volume for each organ. Kappa statistics were applied to assess the inter-provider contouring agreement, and Levene test evaluated the homogeneity of variance for radiation dose/volume metrics, including the V45Gy (cm3 ).

RESULTS

The bowel bag showed greater variation in radiation dose/volume estimates compared to the bladder and rectum. The V45Gy ranged from 163 to 384 cm3 for data set A and 109 to 409 cm3 for dataset B. Kappa values were 0.82/0.83, 0.92/0.92 and 0.94/0.86 for the bowel bag, rectum, and bladder on data sets A/B, demonstrating lower inter-provider agreement for bowel bag compared with bladder and rectum.

CONCLUSION

Inter-provider contouring variability is more significant for the bowel bag than the rectum and bladder, with an associated greater variability in dose and volume estimates during radiation planning.

Deep learning algorithm performance in contouring head and neck organs at risk: a systematic review and single-arm meta-analysis

PURPOSE

The contouring of organs at risk (OARs) in head and neck cancer radiation treatment planning is a crucial, yet repetitive and time-consuming process. Recent studies have applied deep learning (DL) algorithms to automatically contour head and neck OARs. This study aims to conduct a systematic review and meta-analysis to summarize and analyze the performance of DL algorithms in contouring head and neck OARs. The objective is to assess the advantages and limitations of DL algorithms in contour planning of head and neck OARs.

METHODS

This study conducted a literature search of Pubmed, Embase and Cochrane Library databases, to include studies related to DL contouring head and neck OARs, and the dice similarity coefficient (DSC) of four categories of OARs from the results of each study are selected as effect sizes for meta-analysis. Furthermore, this study conducted a subgroup analysis of OARs characterized by image modality and image type.

RESULTS

149 articles were retrieved, and 22 studies were included in the meta-analysis after excluding duplicate literature, primary screening, and re-screening. The combined effect sizes of DSC for brainstem, spinal cord, mandible, left eye, right eye, left optic nerve, right optic nerve, optic chiasm, left parotid, right parotid, left submandibular, and right submandibular are 0.87, 0.83, 0.92, 0.90, 0.90, 0.71, 0.74, 0.62, 0.85, 0.85, 0.82, and 0.82, respectively. For subgroup analysis, the combined effect sizes for segmentation of the brainstem, mandible, left optic nerve, and left parotid gland using CT and MRI images are 0.86/0.92, 0.92/0.90, 0.71/0.73, and 0.84/0.87, respectively. Pooled effect sizes using 2D and 3D images of the brainstem, mandible, left optic nerve, and left parotid gland for contouring are 0.88/0.87, 0.92/0.92, 0.75/0.71 and 0.87/0.85.

CONCLUSIONS

The use of automated contouring technology based on DL algorithms is an essential tool for contouring head and neck OARs, achieving high accuracy, reducing the workload of clinical radiation oncologists, and providing individualized, standardized, and refined treatment plans for implementing "precision radiotherapy". Improving DL performance requires the construction of high-quality data sets and enhancing algorithm optimization and innovation.

Impact of bowel space contouring variability on radiation dose and volume assessments in treatment planning for gynaecologic cancers

Introduction

Correlations between radiation dose/volume measures and small bowel (SB) toxicity are inconsistent in the medical literature. We assessed the impact of inter‐provider variation in bowel bag contouring technique on estimates of radiation dose received by the SB during pelvic radiotherapy.

Methods

Ten radiation oncologists contoured rectum, bladder and bowel bags on treatment planning computed tomography (CT) scans of two patients receiving adjuvant radiation for endometrial cancer. A radiation plan was generated for each patient and used to determine the radiation dose/volume for each organ. Kappa statistics were applied to assess the inter‐provider contouring agreement, and Levene test evaluated the homogeneity of variance for radiation dose/volume metrics, including the V45Gy (cm3).

Results

The bowel bag showed greater variation in radiation dose/volume estimates compared to the bladder and rectum. The V45Gy ranged from 163 to 384 cm3 for data set A and 109 to 409 cm3 for dataset B. Kappa values were 0.82/0.83, 0.92/0.92 and 0.94/0.86 for the bowel bag, rectum, and bladder on data sets A/B, demonstrating lower inter‐provider agreement for bowel bag compared with bladder and rectum.

Conclusion

Inter‐provider contouring variability is more significant for the bowel bag than the rectum and bladder, with an associated greater variability in dose and volume estimates during radiation planning.

Spinal Cord Delineation Based on Computed Tomography Myelogram Versus T2 Magnetic Resonance Imaging in Spinal Stereotactic Body Radiation Therapy

Purpose

Spinal cord delineation is critical to the delivery of stereotactic body radiation therapy (SBRT). Although underestimating the spinal cord can lead to irreversible myelopathy, overestimating the spinal cord may compromise the planning target volume coverage. We compare spinal cord contours based on computed tomography (CT) simulation with a myelogram to spinal cord contours based on fused axial T2 magnetic resonance imaging (MRI).

Methods and Materials

Eight patients with 9 spinal metastases treated with spinal SBRT were contoured by 8 radiation oncologists, neurosurgeons, and physicists, with spinal cord definition based on (1) fused axial T2 MRI and (2) CT-myelogram simulation images, yielding 72 sets of spinal cord contours. The spinal cord volume was contoured at the target vertebral body volume based on both images. The mixed-effect model assessed comparisons of T2 MRI- to myelogram-defined spinal cord in centroid deviations (deviations in the center point of the cord) through the vertebral body target volume, spinal cord volumes, and maximum doses (0.035 cc point) to the spinal cord applying the patient's SBRT treatment plan, in addition to in-between and within-subject variabilities.

Results

The estimate for the fixed effect from the mixed model showed that the mean difference between 72 CT volumes and 72 MRI volumes was 0.06 cc and was not statistically significant (95% confidence interval, -0.034, 0.153; P = .1832). The mixed model showed that the mean dose at 0.035 cc for CT-defined spinal cord contours was 1.24 Gy lower than that of MRI-defined spinal cord contours and was statistically significant (95% confidence interval, -2.292, -0.180; P = .0271). Also, the mixed model indicated no statistical significance for deviations in any of the axes between MRI-defined spinal cord contours and CT-defined spinal cord contours.

Conclusions

CT myelogram may not be required when MRI imaging is feasible, although uncertainty at the cord-to-treatment volume interface may result in overcontouring and hence higher estimated cord dose-maximums with axial T2 MRI-based cord definition.

Validation of clinical acceptability of deep-learning-based automated segmentation of organs-at-risk for head-and-neck radiotherapy treatment planning

Introduction

Organ-at-risk segmentation for head and neck cancer radiation therapy is a complex and time-consuming process (requiring up to 42 individual structure, and may delay start of treatment or even limit access to function-preserving care. Feasibility of using a deep learning (DL) based autosegmentation model to reduce contouring time without compromising contour accuracy is assessed through a blinded randomized trial of radiation oncologists (ROs) using retrospective, de-identified patient data.

Methods

Two head and neck expert ROs used dedicated time to create gold standard (GS) contours on computed tomography (CT) images. 445 CTs were used to train a custom 3D U-Net DL model covering 42 organs-at-risk, with an additional 20 CTs were held out for the randomized trial. For each held-out patient dataset, one of the eight participant ROs was randomly allocated to review and revise the contours produced by the DL model, while another reviewed contours produced by a medical dosimetry assistant (MDA), both blinded to their origin. Time required for MDAs and ROs to contour was recorded, and the unrevised DL contours, as well as the RO-revised contours by the MDAs and DL model were compared to the GS for that patient.

Results

Mean time for initial MDA contouring was 2.3 hours (range 1.6-3.8 hours) and RO-revision took 1.1 hours (range, 0.4-4.4 hours), compared to 0.7 hours (range 0.1-2.0 hours) for the RO-revisions to DL contours. Total time reduced by 76% (95%-Confidence Interval: 65%-88%) and RO-revision time reduced by 35% (95%-CI,-39%-91%). All geometric and dosimetric metrics computed, agreement with GS was equivalent or significantly greater (p&lt;0.05) for RO-revised DL contours compared to the RO-revised MDA contours, including volumetric Dice similarity coefficient (VDSC), surface DSC, added path length, and the 95%-Hausdorff distance. 32 OARs (76%) had mean VDSC greater than 0.8 for the RO-revised DL contours, compared to 20 (48%) for RO-revised MDA contours, and 34 (81%) for the unrevised DL OARs.

Conclusion

DL autosegmentation demonstrated significant time-savings for organ-at-risk contouring while improving agreement with the institutional GS, indicating comparable accuracy of DL model. Integration into the clinical practice with a prospective evaluation is currently underway.

Tumor volume definitions in head and neck squamous cell carcinoma - Comparing PET/MRI and histopathology

BACKGROUND AND PURPOSE

In cancer treatment precise definition of the tumor volume is essential, but despite development in imaging modalities, this remains a challenge. Here, pathological tumor volumes from the surgical specimens were obtained and compared to tumor volumes defined from modern PET/MRI hybrid imaging. The purpose is to evaluate mismatch between the volumes defined from imaging and pathology was estimated and potential clinical impact.

METHODS AND MATERIALS

Twenty-five patients with head and neck squamous cell carcinoma were scanned on an integrated PET/MRI system prior to surgery. Three gross tumor volumes (GTVs) from the primary tumor site were delineated defined from MRI (GTV_MRI), PET (GTV_PET) and one by utilizing both anatomical images and clinical information (GTV_ONCO). Twenty-five primary tumor specimens were extracted en bloc, scanned with PET/MRI and co-registered to the patient images. Each specimen was sectioned in blocks, sliced and stained with haematoxylin and eosin. All slices were digitalized and tumor delineated by a head and neck pathologist. The pathological tumor areas in all slices were interpolated yielding a pathological 3D tumor volume (GTV_PATO). GTV_PATOwas compared with the imaging GTV's and potential mismatch was estimated.

RESULTS

Thirteen patients were included. The mean volume of GTV_ONCOwas larger than the GTV's defined from PET or MRI. The mean mismatch of the GTV_PATOcompared to the GTV_PET, GTV_MRIand GTV_ONCOwas 31.9 %, 54.5 % and 27.9 % respectively, and the entire GTV_PATO was only fully encompassed in GTV_ONCO in 1 of 13 patients. However, after the addition of a clinical 5 mm margin the GTV_PATO was fully encompassed in GTV_ONCO in 11 out of 13 patients.

CONCLUSIONS

Despite modern hybrid imaging modalities, a mismatch between imaging and pathological defined tumor volumes was observed in all patients.A 5 mm clinical margin was sufficient to ensure inclusion of the entire pathological volume in 11 out of 13 patients.

Benefits of automated gross tumor volume segmentation in head and neck cancer using multi-modality information

PURPOSE

Gross tumor volume (GTV) delineation for head and neck cancer (HNC) radiation therapy planning is time consuming and prone to interobserver variability (IOV). The aim of this study was (1) to develop an automated GTV delineation approach of primary tumor (GTVp) and pathologic lymph nodes (GTVn) based on a 3D convolutional neural network (CNN) exploiting multi-modality imaging input as required in clinical practice, and (2) to validate its accuracy, efficiency and IOV compared to manual delineation in a clinical setting.

METHODS

Two datasets were retrospectively collected from 150 clinical cases. CNNs were trained for GTV delineation with consensus delineation as ground truth, with either single (CT) or co-registered multi-modal (CT + PET or CT + MRI) imaging data as input. For validation, GTVs were delineated on 20 new cases by two observers, once manually, once by correcting the delineations generated by the CNN.

RESULTS

Both multi-modality CNNs performed better than the single-modality CNN and were selected for clinical validation. Mean Dice Similarity Coefficient (DSC) for (GTVp, GTVn) respectively between automated and manual delineations was (69%, 79%) for CT + PET and (59%,71%) for CT + MRI. Mean DSC between automated and corrected delineations was (81%,89%) for CT + PET and (69%,77%) for CT + MRI. Mean DSC between observers was (76%,86%) for manual delineations and (95%,96%) for corrected delineations, indicating a significant decrease in IOV (p < 10^-5), while efficiency increased significantly (48%, p < 10^-5).

CONCLUSION

Multi-modality automated delineation of GTV of HNC was shown to be more efficient and consistent compared to manual delineation in a clinical setting and beneficial over a single-modality approach.

Emergence of MR-Linac in Radiation Oncology: Successes and Challenges of Riding on the MRgRT Bandwagon

The special issue of JCM on “Advances of MRI in Radiation Oncology” provides a unique forum for scientific literature related to MR imaging in radiation oncology. This issue covered many aspects, such as MR technology, motion management, economics, soft-tissue–air interface issues, and disease sites such as the pancreas, spine, sarcoma, prostate, head and neck, and rectum from both camps—the Unity and MRIdian systems. This paper provides additional information on the success and challenges of the two systems. A challenging aspect of this technology is low throughput and the monumental task of education and training that hinders its use for the majority of therapy centers. Additionally, the cost of this technology is too high for most institutions, and hence widespread use is still limited. This article highlights some of the difficulties and how to resolve them.

Delineation uncertainties of tumour volumes on MRI of head and neck cancer patients

•

Role of target delineation uncertainties in head and neck cancer patients.

•

Knowing contouring variations for MRI allows better adaptation of MRLinac for H&N cancers.

•

An interobserver variation for GTV among 8 observers was below 2 mm using MRI.

•

Variability between observers might improve using other imaging modalities.

Background

During the last decade, radiotherapy using MR Linac has gone from research to clinical implementation for different cancer locations. For head and neck cancer (HNC), target delineation based only on MR images is not yet standard, and the utilisation of MRI instead of PET/CT in radiotherapy planning is not well established. We aimed to analyse the inter-observer variation (IOV) in delineating GTV (gross tumour volume) on MR images only for patients with HNC.

Material/methods

32 HNC patients from two independent departments were included. Four clinical oncologists from Denmark and four radiation oncologists from Australia had independently contoured primary tumour GTVs (GTV-T) and nodal GTVs (GTV-N) on T2-weighted MR images obtained at the time of treatment planning. Observers were provided with sets of images, delineation guidelines and patient synopsis. Simultaneous truth and performance level estimation (STAPLE) reference volumes were generated for each structure using all observer contours. The IOV was assessed using the DICE Similarity Coefficient (DSC) and mean absolute surface distance (MASD).

Results

32 GTV-Ts and 68 GTV-Ns were contoured per observer. The median MASD for GTV-Ts and GTV-Ns across all patients was 0.17 cm (range 0.08–0.39 cm) and 0.07 cm (range 0.04–0.33 cm), respectively. Median DSC relative to a STAPLE volume for GTV-Ts and GTV-Ns across all patients were 0.73 and 0.76, respectively. A significant correlation was seen between median DSCs and median volumes of GTV-Ts (Spearman correlation coefficient 0.76, p < 0.001) and of GTV-Ns (Spearman correlation coefficient 0.55, p < 0.001).

Conclusion

Contouring GTVs in patients with HNC on MRI showed that the median IOV for GTV-T and GTV-N was below 2 mm, based on observes from two separate radiation departments. However, there are still specific regions in tumours that are difficult to resolve as either malignant tissue or oedema that potentially could be improved by further training in MR-only delineation.

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.

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.

Saliency-guided deep learning network for automatic tumor bed volume delineation in post-operative breast irradiation

Efficient, reliable and reproducible target volume delineation is a key step in the effective planning of breast radiotherapy. However, post-operative breast target delineation is challenging as the contrast between the tumor bed volume (TBV) and normal breast tissue is relatively low in CT images. In this study, we propose to mimic the marker-guidance procedure in manual target delineation. We developed a saliency-based deep learning segmentation (SDL-Seg) algorithm for accurate TBV segmentation in post-operative breast irradiation. The SDL-Seg algorithm incorporates saliency information in the form of markers' location cues into a U-Net model. The design forces the model to encode the location-related features, which underscores regions with high saliency levels and suppresses low saliency regions. The saliency maps were generated by identifying markers on CT images. Markers' location were then converted to probability maps using a distance transformation coupled with a Gaussian filter. Subsequently, the CT images and the corresponding saliency maps formed a multi-channel input for the SDL-Seg network. Our in-house dataset was comprised of 145 prone CT images from 29 post-operative breast cancer patients, who received 5-fraction partial breast irradiation (PBI) regimen on GammaPod. The 29 patients were randomly split into training (19), validation (5) and test (5) sets. The performance of the proposed method was compared against basic U-Net. Our model achieved mean (standard deviation) of 76.4(±2.7) %, 6.76(±1.83) mm, and 1.9(±0.66) mm for Dice similarity coefficient, 95 percentile Hausdorff distance, and average symmetric surface distance respectively on the test set with computation time of below 11 seconds per one CT volume. SDL-Seg showed superior performance relative to basic U-Net for all the evaluation metrics while preserving low computation cost. The findings demonstrate that SDL-Seg is a promising approach for improving the efficiency and accuracy of the on-line treatment planning procedure of PBI, such as GammaPod based PBI.

A rapid review of influential factors and appraised solutions on organ delineation uncertainties reduction in radiotherapy

Background and purpose.Accurate volume delineation plays an essential role in radiotherapy. Contouring is a potential source of uncertainties in radiotherapy treatment planning that could affect treatment outcomes. Therefore, reducing the degree of contouring uncertainties is crucial. The role of utilized imaging modality in the organ delineation uncertainties has been investigated. This systematic review explores the influential factors on inter-and intra-observer uncertainties of target volume and organs at risk (OARs) delineation focusing on the used imaging modality for these uncertainties reduction and the reported subsequent histopathology and follow-up assessment.Methods and materials.An inclusive search strategy has been conducted to query the available online databases (Scopus, Google Scholar, PubMed, and Medline). 'Organ at risk', 'target', 'delineation', 'uncertainties', 'radiotherapy' and their relevant terms were utilized using every database searching syntax. Final article extraction was performed following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline. Included studies were limited to the ones published in English between 1995 and 2020 and that just deal with computed tomography (CT) and magnetic resonance imaging (MRI) modalities.Results.A total of 923 studies were screened and 78 were included of which 31 related to the prostate 20 to the breast, 18 to the head and neck, and 9 to the brain tumor site. 98% of the extracted studies performed volumetric analysis. Only 24% of the publications reported the dose deviations resulted from variation in volume delineation Also, heterogeneity in studied populations and reported geometric and volumetric parameters were identified such that quantitative synthesis was not appropriate.Conclusion.This review highlightes the inter- and intra-observer variations that could lead to contouring uncertainties and impede tumor control in radiotherapy. For improving volume delineation and reducing inter-observer variability, the implementation of well structured training programs, homogeneity in following consensus and guidelines, reliable ground truth selection, and proper imaging modality utilization could be clinically beneficial.

Statistical harmonization can improve the development of a multicenter CT-based radiomic model predictive of nonresponse to induction chemotherapy in laryngeal cancers

PURPOSE

To develop a radiomic model predicting nonresponse to induction chemotherapy in laryngeal cancers, from multicenter pretherapeutic contrast-enhanced computed tomography (CE-CT) and evaluate the benefit of feature harmonization in such a context.

METHODS

Patients (n = 104) eligible for laryngeal preservation chemotherapy were included in five centers. Primary tumor was manually delineated on the CE-CT images. The following radiomic features were extracted with an in-house software (MIRAS v1.1, LaTIM UMR 1101): intensity, shape, and textural features derived from Gray-Level Co-occurrence Matrix (GLCM), Neighborhood Gray Tone Difference Matrix (NGTDM), Gray-Level Run Length Matrix (GLRLM), and Gray-Level Size Zone Matrix (GLSZM). Harmonization was performed using ComBat after unsupervised hierarchical clustering, used to determine labels automatically, given the high heterogeneity of imaging characteristics across and within centers. Patients with similar feature distributions were grouped with unsupervised clustering into an optimal number of clusters (2) determined with "silhouette scoring." Statistical harmonization was then carried out with ComBat on these 2 identified clusters. The cohort was split into training/validation (n = 66) and testing (n = 32) sets. Area under the receiver operating characteristics curves (AUC) were used to evaluate the ability of radiomic features (before and after harmonization) to predict nonresponse to chemotherapy, and specificity (Sp) and sensitivity (Se) were used to quantify their performance in the testing set.

RESULTS

Without harmonization, none of the features identified as predictive in the training set remained significant in the testing set. After ComBat, one textural feature identified in the training set keeps a predictive trend in the testing set-Zone Percentage, derived from the GLSZM, was predictive of nonresponse in the training set (AUC = 0.62, Se = 70%, Sp = 64%, P = 0.04) and obtained a satisfactory performance in the testing set (Se = 80%, Sp = 67%, P = 0.03), although significance was limited by the size of the testing set. These results are consistent with previously published findings in head and neck cancers.

CONCLUSIONS

Radiomic features from CE-CT could help in the selection of patients for induction chemotherapy in laryngeal cancers, with relatively good sensitivity and specificity in predicting lack of response. Statistical harmonization with ComBat and unsupervised clustering seems to improve the predictive value of features extracted in such a heterogeneous multicenter setting.

Interobserver variability in organ at risk delineation in head and neck cancer

Background

In radiotherapy inaccuracy in organ at risk (OAR) delineation can impact treatment plan optimisation and treatment plan evaluation. Brouwer et al. showed significant interobserver variability (IOV) in OAR delineation in head and neck cancer (HNC) and published international consensus guidelines (ICG) for OAR delineation in 2015. The aim of our study was to evaluate IOV in the presence of these guidelines.

Methods

HNC radiation oncologists (RO) from each Belgian radiotherapy centre were invited to complete a survey and submit contours for 5 HNC cases. Reference contours (OARref) were obtained by a clinically validated artificial intelligence-tool trained using ICG. Dice similarity coefficients (DSC), mean surface distance (MSD) and 95% Hausdorff distances (HD95) were used for comparison.

Results

Fourteen of twenty-two RO (64%) completed the survey and submitted delineations. Thirteen (93%) confirmed the use of delineation guidelines, of which six (43%) used the ICG. The OARs whose delineations agreed best with the OARref were mandible [median DSC 0.9, range (0.8–0.9); median MSD 1.1 mm, range (0.8–8.3), median HD95 3.4 mm, range (1.5–38.7)], brainstem [median DSC 0.9 (0.6–0.9); median MSD 1.5 mm (1.1–4.0), median HD95 4.0 mm (2.3–15.0)], submandibular glands [median DSC 0.8 (0.5–0.9); median MSD 1.2 mm (0.9–2.5), median HD95 3.1 mm (1.8–12.2)] and parotids [median DSC 0.9 (0.6–0.9); median MSD 1.9 mm (1.2–4.2), median HD95 5.1 mm (3.1–19.2)]. Oral cavity, cochleas, PCMs, supraglottic larynx and glottic area showed more variation. RO who used the consensus guidelines showed significantly less IOV (p = 0.008).

Conclusions

Although ICG for delineation of OARs in HNC exist, they are only implemented by about half of RO participating in this study, which partly explains the delineation variability. However, this study highlights that guidelines alone do not suffice to eliminate IOV and that more effort needs to be done to accomplish further treatment standardisation, for example with artificial intelligence.

Supplementary information

Supplementary information accompanies this paper at 10.1186/s13014-020-01677-2.

Automatic clinical target volume delineation for cervical cancer in CT images using deep learning

PURPOSE

Accurately delineating clinical target volumes (CTV) is essential for completing radiotherapy plans but is time-consuming, labor-intensive, and prone to inter-observer variation. Automating CTV delineation has the benefits of both speeding up contouring process and improving the quality of contours. Recently, auto-segmentation approaches based on deep learning have achieved some improvements. However, unlike organ segmentation, the CTV contains potential tumor spread tissues or subclinical disease tissues, resulting in poorly defined margin interface and irregular shape. It is not reasonable to directly apply the deep learning segmentation algorithms to CTV tasks without considering the unique characteristics of shape and margin. In this work, we propose a novel automatic CTV delineation algorithm based on deep learning addressing the unique shape and margin challenges.

METHODS

Our deep learning method, called RA-CTVNet, segments the CTV from cervical cancer CT images. RA-CTVNet denotes our automatic CTV delineation algorithm based on deep learning with Area-aware reweight strategy and Recursive refinement strategy. (1) In order to process the whole-volume CT images and delineate all CTVs in one shot, our method is built upon the popular 3D Unet architecture. We further extend it with robust residual learning and squeeze-and-excitation blocks for better feature representation. (2) We propose area-aware reweight strategy which assigns different weights for different slices. The core is adjusting model's attention to each slice. (3) In terms of the trade-off between providing performance improvements and meeting the limitations of GPU memory, we exploit a new recursive refinement strategy to address margin challenge.

RESULTS

This retrospective study included 462 patients diagnosed with cervical cancer who received radiotherapy from June 2017 to May 2019. Extensive experiments were conducted to evaluate performance of RA-CTVNet. First, compared to different network architectures, RA-CTVNet achieved improvements in Dice similarity coefficient (DSC). Second, we conducted ablation study. The results showed that compared to the backbone, area-aware reweight strategy increased DSC by 3.3% on average and recursive refinement strategy further increased DSC by 1.6% on average. Then, we compared our method with three human experts. Our RA-CTVNet performed better than two experts while comparably to the third expert. Finally, a multicenter evaluation was conducted to verify the accuracy and generalizability.

CONCLUSIONS

Our findings show that deep learning is able to offer an efficient framework for automatic CTV delineation. The tailored RA-CTVNet can improve the quality of CTV contours, which has great potential for reducing the burden of experts and increasing the accuracy of delineation. In the future, if with more training data, further improvements are possible, bringing this approach closer to real clinical practice.

Prognostic imaging variables for recurrent laryngeal and hypopharyngeal carcinoma treated with primary chemoradiotherapy: A systematic review and meta-analysis

Background

In this systematic review, we aim to identify prognostic imaging variables of recurrent laryngeal or hypopharyngeal carcinoma after chemoradiotherapy.

Methods

A systematic search was performed in PubMed and EMBASE (1990–2020). The crude data and effect estimates were extracted for each imaging variable. The level of evidence of each variable was assessed and pooled risk ratios (RRs) were calculated.

Results

Twenty‐two articles were included in this review, 17 on computed tomography (CT) and 5 on magnetic resonance imaging (MRI) variables. We found strong evidence for the prognostic value of tumor volume at various cut‐off points (pooled RRs ranging from 2.09 to 3.03). Anterior commissure involvement (pooled RR 2.19), posterior commissure involvement (pooled RR 2.44), subglottic extension (pooled RR 2.25), and arytenoid cartilage extension (pooled RR 2.10) were also strong prognostic factors.

Conclusion

Pretreatment tumor volume and involvement of several subsites are prognostic factors for recurrent laryngeal or hypopharyngeal carcinoma after chemoradiotherapy.

Fully convolutional network-based multi-output model for automatic segmentation of organs at risk in thorax

PURPOSE

To propose a multi-output fully convolutional network (MOFCN) to segment bilateral lung, heart and spinal cord in the planning thoracic computed tomography (CT) slices automatically and simultaneously.

METHODS

The MOFCN includes two components: one main backbone and three branches. The main backbone extracts the features about lung, heart and spinal cord. The extracted features are transferred to three branches which correspond to three organs respectively. The longest branch to segment spinal cord is nine layers, including input and output layers. The MOFCN was evaluated on 19,277 CT slices from 966 patients with cancer in the thorax. In these slices, the organs at risk (OARs) were delineated and validated by experienced radiation oncologists, and served as ground truth for training and evaluation. The data from 61 randomly chosen patients were used for training and validation. The remaining 905 patients' slices were used for testing. The metric used to evaluate the similarity between the auto-segmented organs and their ground truth was Dice. Besides, we compared the MOFCN with other published models. To assess the distinct output design and the impact of layer number and dilated convolution, we compared MOFCN with a multi-label learning model and its variants. By analyzing the not good performances, we suggested possible solutions.

RESULTS

MOFCN achieved Dice of 0.95 ± 0.02 for lung, 0.91 ± 0.03 for heart and 0.87 ± 0.06 for spinal cord. Compared to other models, MOFCN could achieve a comparable accuracy with the least time cost.

CONCLUSION

The results demonstrated the MOFCN's effectiveness. It uses less parameters to delineate three OARs simultaneously and automatically, and thus shows a relatively low requirement for hardware and has potential for broad application.

Cross-modality deep learning: Contouring of MRI data from annotated CT data only

PURPOSE

Online adaptive radiotherapy would greatly benefit from the development of reliable auto-segmentation algorithms for organs-at-risk and radiation targets. Current practice of manual segmentation is subjective and time-consuming. While deep learning-based algorithms offer ample opportunities to solve this problem, they typically require large datasets. However, medical imaging data are generally sparse, in particular annotated MR images for radiotherapy. In this study, we developed a method to exploit the wealth of publicly available, annotated CT images to generate synthetic MR images, which could then be used to train a convolutional neural network (CNN) to segment the parotid glands on MR images of head and neck cancer patients.

METHODS

Imaging data comprised 202 annotated CT and 27 annotated MR images. The unpaired CT and MR images were fed into a 2D CycleGAN network to generate synthetic MR images from the CT images. Annotations of axial slices of the synthetic images were generated by propagating the CT contours. These were then used to train a 2D CNN. We assessed the segmentation accuracy using the real MR images as test dataset. The accuracy was quantified with the 3D Dice similarity coefficient (DSC), Hausdorff distance (HD), and mean surface distance (MSD) between manual and auto-generated contours. We benchmarked the approach by a comparison to the interobserver variation determined for the real MR images, as well as to the accuracy when training the 2D CNN to segment the CT images.

RESULTS

The determined accuracy (DSC: 0.77±0.07, HD: 18.04±12.59mm, MSD: 2.51±1.47mm) was close to the interobserver variation (DSC: 0.84±0.06, HD: 10.85±5.74mm, MSD: 1.50±0.77mm), as well as to the accuracy when training the 2D CNN to segment the CT images (DSC: 0.81±0.07, HD: 13.00±7.61mm, MSD: 1.87±0.84mm).

CONCLUSIONS

The introduced cross-modality learning technique can be of great value for segmentation problems with sparse training data. We anticipate using this method with any nonannotated MRI dataset to generate annotated synthetic MR images of the same type via image style transfer from annotated CT images. Furthermore, as this technique allows for fast adaptation of annotated datasets from one imaging modality to another, it could prove useful for translating between large varieties of MRI contrasts due to differences in imaging protocols within and between institutions.

Clinical target volume segmentation for stomach cancer by stochastic width deep neural network

PURPOSE

Precise segmentation of clinical target volume (CTV) is the key to stomach cancer radiotherapy. We proposed a novel stochastic width-deep neural network (SW-DNN) for better automatically contouring stomach CTV.

METHODS

Stochastic width-deep neural network was an end-to-end approach, of which the core component was a novel SW mechanism that employed shortcut connections between the encoder and decoder in a random manner, and thus the width of the SW-DNN was stochastically adjustable to obtain improved segmentation results. In total, 150 stomach cancer patient computed tomography (CT) cases with the corresponding CTV labels were collected and used to train and evaluate the SW-DNN. Three common quantitative measures: true positive volume fraction (TPVF), positive predictive value (PPV), and Dice similarity coefficient (DSC) were used to evaluate the segmentation accuracy.

RESULTS

Clinical target volumes calculated by SW-DNN had significant quantitative advantages over three state-of-the-art methods. The average DSC value of SW-DNN was 2.1%, 2.8%, and 3.6% higher than that of three state-of-the-art methods. The average DSC, TPVF, and PPV values of SW-DNN were 2.1%, 4.0%, and 0.3% higher than that of the corresponding constant width DNN.

CONCLUSIONS

Stochastic width-deep neural network provided better performance for contouring stomach cancer CTV accurately and efficiently. It is a promising solution in clinical radiotherapy planning for stomach cancer.

Influence of inter-observer delineation variability on radiomic features of the parotid gland

PURPOSE

This study aimed to quantify the variability in the values of radiomic features extracted from a right parotid gland (RPG) delineated by a series of independent observers.

METHODS

This was a secondary analysis of anonymous data from a delineation workshop. Inter-observer variability of the RPG from 40 participants was quantified using DICE similarity coefficient (DSC) and Hausdorff distance (HD). An additional contour was generated using Varian SmartSegmentation. Radiomic features extracted include four shape features, six histogram features, and 32 texture features. The absolute mean paired percentage difference (PPD) in feature values from the expert and participants were ranked . Feature robustness was classified using pre- determined thresholds.

RESULTS

63% of participants achieved a DSC > 0.7, the auto- segmentation DSC was 0.76. The average HD for the participants was 16.16 mm ± 0.66 mm, and 15.16 mm for the auto-segmentation. 48% (n = 20) and 33% (n = 14) of features were deemed to be robust with a mean absolute PPD < 5%, for the auto-segmentation and manual delineations respectively; the majority of which were from the grey-run length matrix family. 7% (n = 3) of features from the auto- segmentation and 10% (n = 4) from the manual contours were deemed to be unstable with a mean absolute PPD > 50%. The value of the most robust feature was not related to DSC and HD.

CONCLUSION

Inter-observer delineation variability affects the value of the radiomic features extracted from the RPG. This study identifies the radiomic features least sensitive to these uncertainties. Further investigation of the clinical relevance of these features in prediction of xerostomia is warranted.

Clinical Acceptability of Automated Radiation Treatment Planning for Head and Neck Cancer Using the Radiation Planning Assistant

PURPOSE

Radiation treatment planning for head and neck cancer is a complex process with much variability; automated treatment planning is a promising option to improve plan quality and efficiency. This study compared radiation plans generated from a fully automated radiation treatment planning system to plans generated manually that had been clinically approved and delivered.

METHODS AND MATERIALS

The study cohort consisted of 50 patients treated by a specialized head and neck cancer team at a tertiary care center. An automated radiation treatment planning system, the Radiation Planning Assistant, was used to create autoplans for all patients using their original, approved contours. Common dose-volume histogram (DVH) criteria were used to compare the quality of autoplans to the clinical plans. Fourteen radiation oncologists, each from a different institution, then reviewed and compared the autoplans and clinical plans in a blinded fashion.

RESULTS

Autoplans and clinical plans were very similar with regard to DVH metrics for coverage and critical structure constraints. Physician reviewers found both the clinical plans and autoplans acceptable for use; overall, 78% of the clinical plans and 88% of the autoplans were found to be usable as is (without any edits). When asked to choose which plan would be preferred for approval, 27% of physician reviewers selected the clinical plan, 47% selected the autoplan, 25% said both were equivalent, and 0% said neither. Hence, overall, 72% of physician reviewers believed the autoplan or either the clinical or autoplan was preferable.

CONCLUSIONS

Automated radiation treatment planning creates consistent, clinically acceptable treatment plans that meet DVH criteria and are found to be appropriate on physician review.

Cranial organs at risk delineation: heterogenous practices in radiotherapy planning

BACKGROUND

Segmentation is a crucial step in treatment planning that directly impacts dose distribution and optimization. The aim of this study was to evaluate the inter-individual variability of common cranial organs at risk (OAR) delineation in neurooncology practice.

METHODS

Anonymized simulation contrast-enhanced CT and MR scans of one patient with a solitary brain metastasis was used for delineation and analysis. Expert professionals from 16 radiotherapy centers involved in brain structures delineation were asked to segment 9 OAR on their own treatment planning system. As reference, two experts in neurooncology, produced a unique consensual contour set according to guidelines. Overlap ratio, Kappa index (KI), volumetric ratio, Commonly Contoured Volume, Supplementary Contoured Volume were evaluated using Artiview™ v 2.8.2-according to occupation, seniority and level of expertise of all participants.

RESULTS

For the most frequently delineated and largest OAR, the mean KI are often good (0.8 for the parotid and the brainstem); however, for the smaller OAR, KI degrade (0.3 for the optic chiasm, 0.5% for the cochlea), with a significant discrimination (p < 0.01). The radiation oncologists, members of Association des Neuro-Oncologue d'Expression Française society performed better in all indicators compared to non-members (p < 0.01). Our exercise was effective in separating the different participating centers with 3 of the reported indicators (p < 0.01).

CONCLUSION

Our study illustrates the heterogeneity in normal structures contouring between professionals. We emphasize the need for cerebral OAR delineation harmonization-that is a major determinant of therapeutic ratio and clinical trials evaluation.

Usefulness of dual-energy computed tomography for oral cancer image

OBJECTIVES

We aimed to compare dual-energy computed tomography (DECT) virtual monochromatic imaging (VMI) and iodine density imaging (IDI) of oral cancers in terms of visual scoring and tumour volume estimation.

MATERIALS AND METHODS

Nine patients diagnosed with oral cancer who underwent DECT VMI and IDI were enrolled. One radiation oncologist, one head and neck surgeon and nine oral surgeons evaluated image clarity and quality in each patient in terms of metal artefacts due to dental prosthesis, internal tumour structure, tumour-organ boundary and total quality of images for diagnosis. Tumour volume was estimated using VMI, IDI and magnetic resonance imaging (MRI).

RESULTS

The mean score for image artefact was significantly higher for IDI than for VMI in three observers, the mean score for internal structure was significantly higher for IDI than for VMI in five, the mean score for tumour-organ boundary was significantly higher for IDI than for VMI in two and the mean score for total quality of images for diagnosis was significantly higher for IDI than for VMI in five. Standard deviation of estimated tumour volume was not significantly different between VMI and IDI, but that of MRI was significantly lowest in three images.

CONCLUSIONS

In DECT for oral cancer, IDI has a visual image superior to VMI; thus, we recommend the use of IDI.

TRIAL REGISTRATION

Clinical trial number: UMIN000038994.

MDCC-Net: Multiscale double-channel convolution U-Net framework for colorectal tumor segmentation

PURPOSE

Multiscale feature fusion is a feasible method to improve tumor segmentation accuracy. However, current multiscale networks have two common problems: 1. Some networks only allow feature fusion between encoders and decoders of the same scale. It is obvious that such feature fusion is not sufficient. 2. Some networks have too many dense skip connections and too much nesting between the coding layer and the decoding layer, which causes some features to be lost and means that not enough information will be learned from multiple scales. To overcome these two problems, we propose a multiscale double-channel convolution U-Net (MDCC-Net) framework for colorectal tumor segmentation.

METHODS

In the coding layer, we designed a dual-channel separation and convolution module and then added residual connections to perform multiscale feature fusion on the input image and the feature map after dual-channel separation and convolution. By fusing features at different scales in the same coding layer, the network can fully extract the detailed information of the original image and learn more tumor boundary information.

RESULTS

The segmentation results show that our proposed method has a high accuracy, with a Dice similarity coefficient (DSC) of 83.57%, which is an improvement of 9.59%, 6.42%, and 1.57% compared with nnU-Net, U-Net, and U-Net++, respectively.

CONCLUSION

The experimental results show that our proposed method has good performance in the segmentation of colorectal tumors and is close to the expert level. The proposed method has potential clinical applicability.

Deep learning for elective neck delineation: More consistent and time efficient

BACKGROUND/PURPOSE

Delineation of the lymph node levels of the neck for irradiation of the elective clinical target volume in head and neck cancer (HNC) patients is time consuming and prone to interobserver variability (IOV), although international consensus guidelines exist. The aim of this study was to develop and validate a 3D convolutional neural network (CNN) for semi-automated delineation of all nodal neck levels, focussing on delineation accuracy, efficiency and consistency compared to manual delineation.

MATERIAL/METHODS

The CNN was trained on a clinical dataset of 69 HNC patients. For validation, 17 lymph node levels were manually delineated in 16 new patients by two observers, independently, using international consensus guidelines. Automated delineations were generated by applying the CNN and were subsequently corrected by both observers separately as needed for clinical acceptance. Both delineations were performed two weeks apart and blinded to each other. IOV was quantified using Dice similarity coefficient (DSC), mean surface distance (MSD) and Hausdorff distance (HD). To assess automated delineation accuracy, agreement between automated and corrected delineations were evaluated using the same measures. To assess efficiency, the time taken for manual and corrected delineations were compared. In a second step, only the clinically relevant neck levels were selected and delineated, once again manually and by applying and correcting the network.

RESULTS

When all lymph node levels were delineated, time taken for correcting automated delineations compared to manual delineations was significantly shorter for both observers (mean: 35 vs 52 min, p < 10^-5). Based on DSC, automated delineation agreed best with corrected delineation for lymph node levels Ib, II-IVa, VIa, VIb, VIIa, VIIb (DSC >85%). Manual corrections necessary for clinical acceptance were 1.4 mm MSD on average and were especially low (<1mm) for levels II-IVa, VIa, VIIa and VIIb. IOV was significantly smaller with automated compared to manual delineations (MSD: 1.4 mm vs 2.5 mm, p < 10^-11). When delineating only the clinically relevant neck levels, the correction time was also significantly shorter (mean: 8 vs 15 min, p < 10^-5). Based on DSC, automated delineation agreed very well with corrected delineation (DSC > 87%). Manual corrections necessary for clinical acceptance were 1.3 mm MSD on average. IOV was significantly smaller with automated compared to manual delineations (MSD: 0.8 mm vs 2.3 mm, p < 10^-3).

CONCLUSION

The CNN developed for automated delineation of the elective lymph node levels in the neck in HNC was shown to be more efficient and consistent compared to manual delineation, which justifies its implementation in clinical practice.

Interobserver Variability of Target Volumes Delineated in the Supine and Prone Positions Based on Computed Tomography Images for External-Beam Partial Breast Irradiation After Breast-Conserving Surgery: A Comparative Study

Background: Although the supine position remains the dominant position for external-beam partial breast irradiation (EB-PBI), the advantages of administering EB-PBI in the prone position have been recognized. The interobserver variability between target volumes delineated in the different positions for EB-PBI after breast-conserving surgery needs to be investigated.

Methods: Twenty-seven patients suitable for EB-PBI were enrolled from July 2016 to April 2017. Supine and prone simulation CT images were sequentially acquired for all enrolled patients during free breathing. Five experienced radiotherapists delineated the target volumes for all patients on supine and prone simulation CT images. The selected parameters, including target volumes, the coefficient of variation (COV), the matching degree (MD), and so on, were calculated to analyze the interobserver variability.

Results: Regardless of the patient position, the interobserver variability between tumor bed (TB) and clinical target volume (CTV) measurements in supine and prone positions were statistically significant (F = 31.34, 19.467; 44.000, 41.985; P = 0.000, 0.001; 0.000, 0.001). The interobserver variability of COV_CTV was significantly greater in the supine position than in the prone position (T = 2.64, P = 0.014). Furthermore, the interobserver variabilities of MD_TB and MD_CTV were statistically lower in the supine position than in the prone position (Z = −3.460, −3.195, P = 0.000, 0.001).

Conclusion: When delineating the target volume for EB-PBI, the interobserver variability in the prone position was lower than that in the supine position. Hence, the administration of EB-PBI in the prone position during free breathing is a reasonable option.

Comparison of tumor delineation using dual energy computed tomography versus magnetic resonance imaging in head and neck cancer re-irradiation cases

•

GTVs on the 60 kV and 140 kV from DECT, and the T1c and T2 from MRI were compared.

•

Delineation was the most consistent using T1c (no interobserver difference in DSC).

•

T1c MRI provided higher interobserver agreement for skull base tumors.

•

60 kV DECT provided higher interobserver agreement for non-skull base tumors.

In treatment planning, multiple imaging modalities can be employed to improve the accuracy of tumor delineation but this can be costly. This study aimed to compare the interobserver consistency of using dual energy computed tomography (DECT) versus magnetic resonance imaging (MRI) for delineating tumors in the head and neck cancer (HNC) re-irradiation scenario. Twenty-three patients with recurrent HNC and had planning DECT and MRI were identified. Contoured tumor volumes by seven radiation oncologists were compared. Overall, T1c MRI performed the best with median DSC of 0.58 (0–0.91) for T1c. T1c MRI provided higher interobserver agreement for skull base sites and 60 kV DECT provided higher interobserver agreement for non-skull base sites.

Segmentation of organs-at-risk in cervical cancer CT images with a convolutional neural network

PURPOSE

We introduced and evaluated an end-to-end organs-at-risk (OARs) segmentation model that can provide accurate and consistent OARs segmentation results in much less time.

METHODS

We collected 105 patients' Computed Tomography (CT) scans that diagnosed locally advanced cervical cancer and treated with radiotherapy in one hospital. Seven organs, including the bladder, bone marrow, left femoral head, right femoral head, rectum, small intestine and spinal cord were defined as OARs. The annotated contours of the OARs previously delineated manually by the patient's radiotherapy oncologist and confirmed by the professional committee consisted of eight experienced oncologists before the radiotherapy were used as the ground truth masks. A multi-class segmentation model based on U-Net was designed to fulfil the OARs segmentation task. The Dice Similarity Coefficient (DSC) and 95th Hausdorff Distance (HD) are used as quantitative evaluation metrics to evaluate the proposed method.

RESULTS

The mean DSC values of the proposed method are 0.924, 0.854, 0.906, 0.900, 0.791, 0.833 and 0.827 for the bladder, bone marrow, femoral head left, femoral head right, rectum, small intestine, and spinal cord, respectively. The mean HD values are 5.098, 1.993, 1.390, 1.435, 5.949, 5.281 and 3.269 for the above OARs respectively.

CONCLUSIONS

Our proposed method can help reduce the inter-observer and intra-observer variability of manual OARs delineation and lessen oncologists' efforts. The experimental results demonstrate that our model outperforms the benchmark U-Net model and the oncologists' evaluations show that the segmentation results are highly acceptable to be used in radiation therapy planning.

Inter-observer variability in target delineation increases during adaptive treatment of head-and-neck and lung cancer

Introduction: Inter-observer variability (IOV) in target volume delineation is a well-documented source of geometric uncertainty in radiotherapy. Such variability has not yet been explored in the context of adaptive re-delineation based on imaging data acquired during treatment. We compared IOV in the pre- and mid-treatment setting using expert primary gross tumour volume (GTV) and clinical target volume (CTV) delineations in locoregionally advanced head-and-neck squamous cell carcinoma (HNSCC) and (non-)small cell lung cancer [(N)SCLC]. Material and methods: Five and six observers participated in the HNSCC and (N)SCLC arm, respectively, and provided delineations for five cases each. Imaging data consisted of CT studies partly complemented by FDG-PET and was provided in two separate phases for pre- and mid-treatment. Global delineation compatibility was assessed with a volume overlap metric (the Generalised Conformity Index), while local extremes of IOV were identified through the standard deviation of surface distances from observer delineations to a median consensus delineation. Details of delineation procedures, in particular, GTV to CTV expansion and adaptation strategies, were collected through a questionnaire. Results: Volume overlap analysis revealed a worsening of IOV in all but one case per disease site, which failed to reach significance in this small sample (p-value range .063-.125). Changes in agreement were propagated from GTV to CTV delineations, but correlation could not be formally demonstrated. Surface distance based analysis identified longitudinal target extent as a pervasive source of disagreement for HNSCC. High variability in (N)SCLC was often associated with tumours abutting consolidated lung tissue or potentially invading the mediastinum. Adaptation practices were variable between observers with fewer than half stating that they consistently adapted pre-treatment delineations during treatment. Conclusion: IOV in target volume delineation increases during treatment, where a disparity in institutional adaptation practices adds to the conventional causes of IOV. Consensus guidelines are urgently needed.

Imaging for Target Delineation and Treatment Planning in Radiation Oncology: Current and Emerging Techniques

Imaging in radiation oncology has a wide range of applications. It is necessary not only for tumor staging and treatment response assessment after therapy but also for the treatment planning process, including definition of target and organs at risk, as well as treatment plan calculation. This article provides a comprehensive overview of the main imaging modalities currently used for target delineation and treatment planning and gives insight into new and promising techniques.

Evaluating diffusion-weighted magnetic resonance imaging for target volume delineation in head and neck radiotherapy

INTRODUCTION

Inter-observer variability (IOV) in target volume delineation is a source of error in head and neck radiotherapy. Diffusion-weighted imaging (DWI) has been shown to be useful in detecting recurrent head and neck cancer. This study aims to determine whether DWI improves target volume delineation and IOV.

METHODS

Four radiation oncologists delineated the gross tumour volume (GTV) for ten head and neck cancer patients. Delineation was performed on CT alone as well as fused image sets which incorporated fluorodeoxyglucose (FDG)-positron emission tomography (PET) and magnetic resonance imaging (MRI) in the form of CT/PET, CT/PET/T2W and CT/PET/T2W/DWI image sets. Analysis of the variability of contour volumes was completed by comparison to the simultaneous truth and performance level estimation (STAPLE) volumes. The DICE Similarity Coefficient (DSC) and other IOV metrics for each observer's contour were compared to the STAPLE for each patient and image dataset. A DWI usability scoresheet for delineation was completed.

RESULTS

The CT/PET/T2W/DWI mean GTV volume of 13.37 (10.35-16.39)cm³ was shown to be different to the mean GTV of 10.92 (8.32-13.51)cm³ when using CT alone (P < 0.001). The GTV DSC amongst observers for CT alone was 0.72 (0.65-0.79), CT/PET was 0.73 (0.67-0.80), CT/PET/T2W was 0.71 (0.64-0.77) and CT/PET/T2W/DWI was 0.69 (0.61-0.75).

CONCLUSION

Mean GTVs with the addition of DWI had slightly larger volumes compared to standard CT and CT/PET volumes. DWI may add supplemental visual information for GTV delineation while having a small impact on IOV, therefore potentially improving target volume delineation.

Role and future of MRI in radiation oncology

Technical innovations and developments in areas such as disease localization, dose calculation algorithms, motion management and dose delivery technologies have revolutionized radiation therapy resulting in improved patient care with superior outcomes. A consequence of the ability to design and accurately deliver complex radiation fields is the need for improved target visualization through imaging. While CT imaging has been the standard of care for more than three decades, the superior soft tissue contrast afforded by MR has resulted in the adoption of this technology in radiation therapy. With the development of real time MR imaging techniques, the problem of real time motion management is enticing. Currently, the integration of an MR imaging and megavoltage radiation therapy treatment delivery system (MR-linac or MRL) is a reality that has the potential to provide improved target localization and real time motion management during treatment. Higher magnetic field strengths provide improved image quality potentially providing the backbone for future work related to image texture analysis-a field known as Radiomics-thereby providing meaningful information on the selection of future patients for radiation dose escalation, motion-managed treatment techniques and ultimately better patient care. On-going advances in MRL technologies promise improved real time soft tissue visualization, treatment margin reductions, beam optimization, inhomogeneity corrected dose calculation, fast multileaf collimators and volumetric arc radiation therapy. This review article provides rationale, advantages and disadvantages as well as ideas for future research in MRI related to radiation therapy mainly in adoption of MRL.

Prophylactic enlargement of the thecal sac volume by spinal expansion duroplasty in patients with unresectable malignant intramedullary tumors and metastases prior to radiotherapy

Unresectable malignant intramedullary tumors and metastases usually require radiotherapy which intensifies spinal cord edema and might result in neurological decline. Spinal expansion duroplasty before radiotherapy enlarges the intrathecal volume and might thus prevent neurological deficits. The study aims to evaluate the clinical course of patients undergoing expansion duroplasty. This retrospective analysis (2007-2016) included all patients with unresectable intramedullary tumors who underwent spinal expansion duroplasty. To assess the degree of preoperative cord enlargement, we calculated the "diameter ratio": diameter of the spinal cord below and above the tumor / diameter of the tumor × 2. The presence of perimedullary cerebrospinal fluid (CSF) at the affected levels was analyzed on the preoperative magnetic resonance imaging (MRI). We recorded the occurrence of neurological deficits, wound breakdown, and CSF fistula. We screened 985 patients, 11 of which were included. Eight patients had an intramedullary metastasis, three patients a spinal malignant glioma. A diameter ratio ≤ 0.8 representing a significant preoperative intramedullary enlargement was seen in 10 cases (90.9%). Postoperative imaging was available in 9 patients, demonstrating successful decompression in 8 of the 9 patients (88.9%). The postoperative course was uneventful in 9 patients (81.8%). Mean overall survival was 13.4 (SD 16.2) months. Spinal expansion duroplasty prior to radiotherapy is a previously undescribed concept. Despite neoadjuvant radiation, no wound breakdown or CSF fistula occurred. In unresectable intramedullary tumors and metastases, spinal expansion duroplasty seems to be a safe procedure with the potential to prevent neurological decline due to radiation-induced cord swelling.

Segmentation of parotid glands from registered CT and MR images

PURPOSE

To develop an automatic multimodal method for segmentation of parotid glands (PGs) from pre-registered computed tomography (CT) and magnetic resonance (MR) images and compare its results to the results of an existing state-of-the-art algorithm that segments PGs from CT images only.

METHODS

Magnetic resonance images of head and neck were registered to the accompanying CT images using two different state-of-the-art registration procedures. The reference domains of registered image pairs were divided on the complementary PG regions and backgrounds according to the manual delineation of PGs on CT images, provided by a physician. Patches of intensity values from both image modalities, centered around randomly sampled voxels from the reference domain, served as positive or negative samples in the training of the convolutional neural network (CNN) classifier. The trained CNN accepted a previously unseen (registered) image pair and classified its voxels according to the resemblance of its patches to the patches used for training. The final segmentation was refined using a graph-cut algorithm, followed by the dilate-erode operations.

RESULTS

Using the same image dataset, segmentation of PGs was performed using the proposed multimodal algorithm and an existing monomodal algorithm, which segments PGs from CT images only. The mean value of the achieved Dice overlapping coefficient for the proposed algorithm was 78.8%, while the corresponding mean value for the monomodal algorithm was 76.5%.

CONCLUSIONS

Automatic PG segmentation on the planning CT image can be augmented with the MR image modality, leading to an improved RT planning of head and neck cancer.

Molecular Imaging-Guided Radiotherapy for the Treatment of Head-and-Neck Squamous Cell Carcinoma: Does it Fulfill the Promises?

With the routine use of intensity modulated radiation therapy for the treatment of head-and-neck squamous cell carcinoma allowing highly conformed dose distribution, there is an increasing need for refining both the selection and the delineation of gross tumor volumes (GTV). In this framework, molecular imaging with positron emission tomography and magnetic resonance imaging offers the opportunity to improve diagnostic accuracy and to integrate tumor biology mainly related to the assessment of tumor cell density, tumor hypoxia, and tumor proliferation into the treatment planning equation. Such integration, however, requires a deep comprehension of the technical and methodological issues related to image acquisition, reconstruction, and segmentation. Until now, molecular imaging has had a limited value for the selection of nodal GTV, but there are increasing evidences that both FDG positron emission tomography and diffusion-weighted magnetic resonance imaging has a potential value for the delineation of the primary tumor GTV, effecting on dose distribution. With the apprehension of the heterogeneity in tumor biology through molecular imaging, growing evidences have been collected over the years to support the concept of dose escalation/dose redistribution using a planned heterogeneous dose prescription, the so-called "dose painting" approach. Validation trials are ongoing, and in the coming years, one may expect to position the dose painting approach in the armamentarium for the treatment of patients with head-and-neck squamous cell carcinoma.

Vertebral body segmentation in wide range clinical routine spine MRI data

BACKGROUND AND OBJECTIVE

In this work we propose a 3D vertebral body segmentation approach for clinical magnetic resonance (MR) spine imaging. So far, vertebrae segmentation approaches in MR spine imaging are either limited to particular MR imaging sequences or require minutes to compute, which can be hindering in clinical routine. The major contribution of our work is a reasonably precise segmentation result, within seconds and with minimal user interaction, for spine MR imaging commonly used in clinical routine. Our focus lies on the applicability towards a large variety of clinical MR imaging sequences, dealing with low image quality, high anisotropy and spine pathologies.

METHODS

Our method starts with a intensity correction step to deal with bias field artifacts and a minimal user-assisted initialization. Next, appearance-based vertebral body probability maps guide a subsequent hybrid level-set segmentation.

RESULTS

We tested our method on different MR imaging sequences from 48 subjects. Overall, our evaluation set contains 63 datasets including 419 vertebral bodies, which differ in age, sex and presence of spine pathologies. This is the largest set of reference segmentations of clinical routine spine MR imaging so far. We achieved a Dice coefficient of 86.0%, a mean Euclidean surface distance error of 1.59 ± 0.24 mm and a Hausdorff distance of 6.86 mm.

CONCLUSIONS

These results illustrate the robustness of our segmentation approach towards the variety of MR image data, which is a pivotal aspect for clinical usefulness and reliable diagnosis.

Evaluation of diffusion weighted imaging for tumor delineation in head-and-neck radiotherapy by comparison with automatically segmented 18F-fluorodeoxyglucose positron emission tomography

Background and purpose

Diffusion weighted (DW) MRI may facilitate target volume delineation for head-and-neck (HN) radiation treatment planning. In this study we assessed the use of a dedicated, geometrically accurate, DW-MRI sequence for target volume delineation. The delineations were compared with semi-automatic segmentations on ¹⁸F-fluorodeoxyglucose (FDG) positron emission tomography (PET) images and evaluated for interobserver variation.

Methods and materials

Fifteen HN cancer patients underwent both DW-MRI and FDG-PET for RT treatment planning. Target delineation on DW-MRI was performed by three observers, while for PET a semi-automatic segmentation was performed using a Gaussian mixture model. For interobserver variation and intermodality variation, volumes, overlap metrics and Hausdorff distances were calculated from the delineations.

Results

The median volumes delineated by the three observers on DW-MRI were 10.8, 10.5 and 9.0 cm³ respectively, and was larger than the median PET volume (8.0 cm³). The median conformity index of DW-MRI for interobserver variation was 0.73 (range 0.38–0.80). Compared to PET, the delineations on DW-MRI by the three observers showed a median dice similarity coefficient of 0.71, 0.69 and 0.72 respectively. The mean Hausdorff distance was small with median (range) distances between PET and DW-MRI of 2.3 (1.5–6.8), 2.5 (1.6–6.9) and 2.0 (1.35–7.6) mm respectively. Over all patients, the median 95th percentile distances were 6.0 (3.0–13.4), 6.6 (4.0–24.0) and 5.3 (3.4–26.0) mm.

Conclusion

Using a dedicated DW-MRI sequence, target volumes could be defined with good interobserver agreement and a good overlap with PET. Target volume delineation using DW-MRI is promising in head-and-neck radiotherapy, combined with other modalities, it can lead to more precise target volume delineation.

Atlas ranking and selection for automatic segmentation of the esophagus from CT scans

In radiation treatment planning, the esophagus is an important organ-at-risk that should be spared in patients with head and neck cancer or thoracic cancer who undergo intensity-modulated radiation therapy. However, automatic segmentation of the esophagus from CT scans is extremely challenging because of the structure's inconsistent intensity, low contrast against the surrounding tissues, complex and variable shape and location, and random air bubbles. The goal of this study is to develop an online atlas selection approach to choose a subset of optimal atlases for multi-atlas segmentation to the delineate esophagus automatically. We performed atlas selection in two phases. In the first phase, we used the correlation coefficient of the image content in a cubic region between each atlas and the new image to evaluate their similarity and to rank the atlases in an atlas pool. A subset of atlases based on this ranking was selected, and deformable image registration was performed to generate deformed contours and deformed images in the new image space. In the second phase of atlas selection, we used Kullback-Leibler divergence to measure the similarity of local-intensity histograms between the new image and each of the deformed images, and the measurements were used to rank the previously selected atlases. Deformed contours were overlapped sequentially, from the most to the least similar, and the overlap ratio was examined. We further identified a subset of optimal atlases by analyzing the variation of the overlap ratio versus the number of atlases. The deformed contours from these optimal atlases were fused together using a modified simultaneous truth and performance level estimation algorithm to produce the final segmentation. The approach was validated with promising results using both internal data sets (21 head and neck cancer patients and 15 thoracic cancer patients) and external data sets (30 thoracic patients).

Automatic segmentation of the clinical target volume and organs at risk in the planning CT for rectal cancer using deep dilated convolutional neural networks

PURPOSE

Delineation of the clinical target volume (CTV) and organs at risk (OARs) is very important for radiotherapy but is time-consuming and prone to inter-observer variation. Here, we proposed a novel deep dilated convolutional neural network (DDCNN)-based method for fast and consistent auto-segmentation of these structures.

METHODS

Our DDCNN method was an end-to-end architecture enabling fast training and testing. Specifically, it employed a novel multiple-scale convolutional architecture to extract multiple-scale context features in the early layers, which contain the original information on fine texture and boundaries and which are very useful for accurate auto-segmentation. In addition, it enlarged the receptive fields of dilated convolutions at the end of networks to capture complementary context features. Then, it replaced the fully connected layers with fully convolutional layers to achieve pixel-wise segmentation. We used data from 278 patients with rectal cancer for evaluation. The CTV and OARs were delineated and validated by senior radiation oncologists in the planning computed tomography (CT) images. A total of 218 patients chosen randomly were used for training, and the remaining 60 for validation. The Dice similarity coefficient (DSC) was used to measure segmentation accuracy.

RESULTS

Performance was evaluated on segmentation of the CTV and OARs. In addition, the performance of DDCNN was compared with that of U-Net. The proposed DDCNN method outperformed the U-Net for all segmentations, and the average DSC value of DDCNN was 3.8% higher than that of U-Net. Mean DSC values of DDCNN were 87.7% for the CTV, 93.4% for the bladder, 92.1% for the left femoral head, 92.3% for the right femoral head, 65.3% for the intestine, and 61.8% for the colon. The test time was 45 s per patient for segmentation of all the CTV, bladder, left and right femoral heads, colon, and intestine. We also assessed our approaches and results with those in the literature: our system showed superior performance and faster speed.

CONCLUSIONS

These data suggest that DDCNN can be used to segment the CTV and OARs accurately and efficiently. It was invariant to the body size, body shape, and age of the patients. DDCNN could improve the consistency of contouring and streamline radiotherapy workflows.