Mean Surface Distance Research Articles

Impact of Artificial Intelligence-Based Autosegmentation of Organs at Risk in Low- and Middle-Income Countries

Radiation therapy (RT) processes require significant human resources and expertise, creating a barrier to rapid RT deployment in low- and middle-income countries (LMICs). Accurate segmentation of tumor targets and organs at risk (OARs) is crucial for optimal RT. This study assessed the impact of artificial intelligence (AI)-based autosegmentation of OARs in 2 LMICs. Ten patients, comprising 5 head and neck (HN) cancer patients and 5 prostate cancer patients, were randomly selected. Planning computed tomography images were subjected to autosegmentation using an Food and Drug Administration-approved AI software tool and manual segmentation by experienced radiation oncologists from 2 LMIC RT clinics. The control data, obtained from a large academic institution in the United States, consisted of contours obtained by an experienced radiation oncologist. The segmentation time, DICE similarity coefficient (DSC), Hausdorff distance, and mean surface distance were evaluated. AI significantly reduced segmentation time, averaging 2 minutes per patient, compared with 57 to 84 minutes for manual contouring in LMICs. Compared with the control data, the AI pelvic contours provided better agreement than did the LMIC manual contours (mean DSC of 0.834 vs 0.807 in LMIC1 and 0.844 vs 0.801 in LMIC2). For HN contours, AI provided better agreement for the majority of OAR contours than manual contours in LMIC1 (mean DSC: 0.823 vs 0.821) or LMIC2 (mean DSC: 0.792 vs 0.748). Neither the AI nor LMIC manual contours had good agreement with the control data (DSC < 0.600) for the optic nerves, chiasm, and cochlea. AI-based autosegmentation generates OAR contours of comparable quality to manual segmentation for both pelvic and HN cancer patients in LMICs, with substantial time savings.

Automated Organ Segmentation for Radiation Therapy: A Comparative Analysis of AI-Based Tools Versus Manual Contouring in Korean Cancer Patients.

Accurate delineation of tumors and organs at risk (OARs) is crucial for intensity-modulated radiation therapy. This study aimed to evaluate the performance of OncoStudio, an AI-based auto-segmentation tool developed for Korean patients, compared with Protégé AI, a globally developed tool that uses data from Korean cancer patients. A retrospective analysis of 1200 Korean cancer patients treated with radiotherapy was conducted. Auto-contours generated via OncoStudio and Protégé AI were compared with manual contours across the head and neck and thoracic, abdominal, and pelvic organs. Accuracy was assessed using the Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff distance (HD). Feedback was obtained from 10 participants, including radiation oncologists, residents, and radiation therapists, via an online survey with a Turing test component. OncoStudio outperformed Protégé AI in 85% of the evaluated OARs (p < 0.001). For head and neck organs, OncoStudio achieved a similar DSC (0.70 vs. 0.70, p = 0.637) but significantly lower MSD and 95% HD values (p < 0.001). In thoracic organs, OncoStudio performed excellently in 90% of cases, with a significantly greater DSC (male: 0.87 vs. 0.82, p < 0.001; female: 0.95 vs. 0.87, p < 0.001). OncoStudio also demonstrated superior accuracy in abdominal (DSC 0.88 vs. 0.81, p < 0.001) and pelvic organs (male: DSC 0.95 vs. 0.85, p < 0.001; female: DSC 0.82 vs. 0.73, p < 0.001). Clinicians favored OncoStudio in 70% of cases, with 90% endorsing its clinical suitability for Korean patients. OncoStudio, which is tailored for Korean patients, demonstrated superior segmentation accuracy across multiple anatomical regions, suggesting its suitability for radiotherapy planning in this population.

Geometric and dosimetric evaluation for breast and regional nodal auto-segmentation structures.

The accuracy of artificial intelligence (AI) generated contours for intact-breast and post-mastectomy radiotherapy plans was evaluated. Geometric and dosimetric comparisons were performed between auto-contours (ACs) and manual-contours (MCs) produced by physicians for target structures. Breast and regional nodal structures were manually delineated on 66 breast cancer patients. ACs were retrospectively generated. The characteristics of the breast/post-mastectomy chestwall (CW) and regional nodal structures (axillary [AxN], supraclavicular [SC], internal mammary [IM]) were geometrically evaluated by Dice similarity coefficient (DSC), mean surface distance, and Hausdorff Distance. The structures were also evaluated dosimetrically by superimposing the MC clinically delivered plans onto the ACs to assess the impact of utilizing ACs with target dose (Vx%) evaluation. Positive geometric correlations between volume and DSC for intact-breast, AxN, and CW were observed. Little or anti correlations between volume and DSC for IM and SC were shown. For intact-breast plans, insignificant dosimetric differences between ACs and MCs were observed for AxNV95% (p=0.17) and SCV95% (p=0.16), while IMNV90% ACs and MCs were significantly different. The average V95% for intact-breast MCs (98.4%) and ACs (97.1%) were comparable but statistically different (p=0.02). For post-mastectomy plans, AxNV95% (p=0.35) and SCV95% (p=0.08) were consistent between ACs and MCs, while IMNV90% was significantly different. Additionally, 94.1% of AC-breasts met ΔV95% variation<5% when DSC>0.7. However, only 62.5% AC-CWs achieved the same metrics, despite AC-CWV95% (p=0.43) being statistically insignificant. The AC intact-breast structure was dosimetrically similar to MCs. The AC AxN and SC may require manual adjustments. Careful review should be performed for AC post-mastectomy CW and IMN before treatment planning. The findings of this study may guide the clinical decision-making process for the utilization of AI-driven ACs for intact-breast and post-mastectomy plans. Before clinical implementation of this auto-segmentation software, an in-depth assessment of agreement with each local facilities MCs is needed.

Head and neck automatic multi-organ segmentation on Dual-Energy Computed Tomography

Background and purposeDeep-learning-based automatic segmentation is widely used in radiation oncology to delineate organs-at-risk. Dual-energy CT (DECT) allows the reconstruction of enhanced contrast images that could help with manual and auto-delineation. This paper presents a performance evaluation of a commercial auto-segmentation software on image series generated by a DECT. Material and methodsDifferent types of DECT images from seventy four head-and-neck (HN) patients were retrieved, including polyenergetic images at different voltages [80 kV reconstructed with a kernel corresponding to the commercial algorithm DirectDensity™ (PEI80-DD), 80 kV (PEI80), 120 kV-mixed (PEI120)] and a virtual-monoenergetic image at 40 keV (VMI40). Delineations used for treatment planning were considered as ground truth (GT) and were compared with the auto-segmentations performed on the 4 DECT images. A blinded qualitative evaluation of 3 structures (thyroid, left parotid, left nodes level II) was carried out. Performance metrics were calculated for thirteen HN structures to evaluate the auto-contours including dice similarity coefficient (DSC), 95th percentile Hausdorff distance (95HD) and mean surface distance (MSD). ResultsWe observed a high rate of low scores for PEI80-DD and VMI40 auto-segmentations on the thyroid and for GT and VMI40 contours on the nodes level II. All images received excellent scores for the parotid glands. The metrics comparison between GT and auto-segmented contours revealed that PEI80-DD had the highest DSC scores, significantly outperforming other reconstructed images for all organs (p < 0.05). ConclusionsThe results indicate that the auto-contouring system cannot generalize to images derived from DECT acquisition. It is therefore crucial to identify which organs benefit from these acquisitions to adapt the training datasets accordingly.

Impact of geometric correction on echo-planar imaging-based apparent diffusion coefficient maps for abdominal radiotherapy

Objective. The apparent diffusion coefficient (ADC) extracted from diffusion-weighted magnetic resonance imaging (DWI) is a potential biomarker in radiotherapy (RT). DWI is often implemented with an echo-planar imaging (EPI) read-out due to speed, but unfortunately low geometric accuracy follows. This study aimed to investigate the influence of geometric distortions on the ADCs extracted from the gross tumor volume (GTV) and on the shape of the GTV in abdominal EPI-DWI. Approach. Twenty-one patients had EPI-DWI scans on a 1.5 T MRI sim before treatment and on a 1.5 T MRI-Linac at one of the first treatment fractions. Off-resonance correction with and without eddy current correction were applied to ADC maps. The clinical GTVs were deformed based on the same (but inverted) corrections to assess the local-regional geometric influence of distortions. Mean surface distance (MSD), Hausdorff distance (HD), and Dice similarity coefficient (DSC) were calculated to compare the original and distorted GTVs, and ADC values were calculated based on a mono-exponential model. Phantom measurements were performed to validate the applied correction method. Main results. The median (range) ADC change within the GTV after full distortion correction was 1.3% (0.02%–6.9%) for MRI-Sim and 1.5% (0.1%–6.4%) for MRI-Linac. The additional effect of the eddy current correction was small in both systems. The median (range) MSD, HD, and DSC comparing the original and off-resonance distorted GTVs for all patients were 0.43 mm (0.11–0.94 mm), 4.00 mm (1.00–7.81 mm) and 0.93 (0.82–0.99), respectively. Significance. Overall effect of distortion correction was small in terms of derived ADC values, indicating that distortion correction is unimportant for prediction of outcomes based on ADC. However, large local geometric changes occurred after off-resonance distortion correction for some patients, suggesting that if the spatial information from ADC maps is to be used for dose painting strategies, corrections should be applied.

Deep-learning-based segmentation using individual patient data on prostate cancer radiation therapy.

Organ-at-risk segmentation is essential in adaptive radiotherapy (ART). Learning-based automatic segmentation can reduce committed labor and accelerate the ART process. In this study, an auto-segmentation model was developed by employing individual patient datasets and a deep-learning-based augmentation method for tailoring radiation therapy according to the changes in the target and organ of interest in patients with prostate cancer. Two computed tomography (CT) datasets with well-defined labels, including contoured prostate, bladder, and rectum, were obtained from 18 patients. The labels of the CT images captured during radiation therapy (CT2nd) were predicted using CT images scanned before radiation therapy (CT1st). From the deformable vector fields (DVFs) created by using the VoxelMorph method, 10 DVFs were extracted when each of the modified CT and CT2nd images were deformed and registered to the fixed CT1st image. Augmented images were acquired by utilizing 110 extracted DVFs and spatially transforming the CT1st images and labels. An nnU-net autosegmentation network was trained by using the augmented images, and the CT2nd label was predicted. A patient-specific model was created for 18 patients, and the performances of the individual models were evaluated. The results were evaluated by employing the Dice similarity coefficient (DSC), average Hausdorff distance, and mean surface distance. The accuracy of the proposed model was compared with those of models trained with large datasets. Patient-specific models were developed successfully. For the proposed method, the DSC values of the actual and predicted labels for the bladder, prostate, and rectum were 0.94 ± 0.03, 0.84 ± 0.07, and 0.83 ± 0.04, respectively. We demonstrated the feasibility of automatic segmentation by employing individual patient datasets and image augmentation techniques. The proposed method has potential for clinical application in automatic prostate segmentation for ART.

Landmark-based auto-contouring of clinical target volumes for radiotherapy of nasopharyngeal cancer.

The delineation of clinical target volumes (CTVs) for radiotherapy for nasopharyngeal cancer is complex and varies based on the location and extent of disease. The current study aimed to develop an auto-contouring solution following one protocol guidelines (NRG-HN001) that can be adjusted to meet other guidelines, such as RTOG-0225 and the 2018 International guidelines. The study used 2-channel 3-dimensional U-Net and nnU-Net framework to auto-contour 27 normal structures in the head and neck (H&N) region that are used to define CTVs in the protocol. To define the CTV-Expansion (CTV1 and CTV2) and CTV-Overall (the outer envelope of all the CTV contours), we used adjustable morphological geometric landmarks and mimicked physician interpretation of the protocol rules by partially or fully including select anatomic structures. The results were evaluated quantitatively using the dice similarity coefficient (DSC) and mean surface distance (MSD) and qualitatively by independent reviews by two H&N radiation oncologists. The auto-contouring tool showed high accuracy for nasopharyngeal CTVs. Comparison between auto-contours and clinical contours for 19 patients with cancers of various stages showed a DSC of 0.94±0.02 and MSD of 0.4±0.4mm for CTV-Expansion and a DSC of 0.83±0.02 and MSD of 2.4±0.5mm for CTV-Overall. Upon independent review, two H&N physicians found the auto-contours to be usable without edits in 85% and 75% of cases. In 15% of cases, minor edits were required by both physicians. Thus, one physician rated 100% of the auto-contours as usable (use as is, or after minor edits), while the other physician rated 90% as usable. The second physician required major edits in 10% of cases. The study demonstrates the ability of an auto-contouring tool to reliably delineate nasopharyngeal CTVs based on protocol guidelines. The tool was found to be clinically acceptable by two H&N radiation oncology physicians in at least 90% of the cases.

Personalized Deep Learning Model for Clinical Target Volume on Daily Cone Beam Computed Tomography in Breast Cancer Patients

PurposeHerein, we developed a deep learning algorithm to improve the segmentation of the clinical target volume (CTV) on daily cone-beam computed tomography (CBCT) scans in breast cancer radiotherapy. By leveraging the Intentional Deep Overfit Learning (IDOL) framework, we aimed to enhance personalized image-guided radiotherapy (IGRT) based on patient-specific learning. MethodsWe utilized 240 CBCT scans from 100 breast cancer patients and employed a two-stage training approach. The first stage involved training a novel general deep learning model (Swin UNETR, UNET, and SegResNET) on 90 patients. The second stage utilized intentional overfitting on the remaining 10 patients for patient-specific CBCT outputs. Quantitative evaluation was conducted using the Dice Similarity Coefficient (DSC), Hausdorff Distance (HD), mean surface distance (MSD), and independent samples t-test with expert contours on CBCT scans from the 1st to 15th fractions. ResultsIDOL integration significantly improved CTV segmentation, particularly with the Swin UNETR model (p-values <0.05). Using patient-specific data, IDOL enhanced the DSC, HD, and MSD metrics. The average DSC for the 15th fraction improved from 0.9611 to 0.9819, the average HD decreased from 4.0118 mm to 1.3935 mm, and the average MSD decreased from 0.8723 to 0.4603. Incorporating CBCT scans from the initial treatments and 1st to 3rd fractions further improved results, with an average DSC of 0.9850, an average HD of 1.2707 mm, and an average MSD of 0.4076 for the 15th fraction, closely aligned with physician-drawn contours. ConclusionCompared with a general model, our patient-specific deep learning-based training algorithm significantly improved CTV segmentation accuracy of CBCT scans in patients with breast cancer. This approach, coupled with continuous deep learning training using daily CBCT scans, demonstrated enhanced CTV delineation accuracy and efficiency. Future studies should explore the adaptability of the IDOL framework to diverse deep learning models, datasets, and cancer sites.

Autocontouring of primary lung lesions and nodal disease for radiotherapy based only on computed tomography images

Background and purposeIn many clinics, positron-emission tomography is unavailable and clinician time extremely limited. Here we describe a deep-learning model for autocontouring gross disease for patients undergoing palliative radiotherapy for primary lung lesions and/or hilar/mediastinal nodal disease, based only on computed tomography (CT) images. Materials and methodsAn autocontouring model (nnU-Net) was trained to contour gross disease in 379 cases (352 training, 27 test); 11 further test cases from an external centre were also included. Anchor-point-based post-processing was applied to remove extraneous autocontoured regions. The autocontours were evaluated quantitatively in terms of volume similarity (Dice similarity coefficient [DSC], surface Dice coefficient, 95th percentile Hausdorff distance [HD95], and mean surface distance), and scored for usability by two consultant oncologists. The magnitude of treatment margin needed to account for geometric discrepancies was also assessed. ResultsThe anchor point process successfully removed all erroneous regions from the autocontoured disease, and identified two cases to be excluded from further analysis due to ‘missed’ disease. The average DSC and HD95 were 0.8 ± 0.1 and 10.5 ± 7.3 mm, respectively. A 10-mm uniform margin-distance applied to the autocontoured region was found to yield “full coverage” (sensitivity > 0.99) of the clinical contour for 64 % of cases. Ninety-seven percent of evaluated autocontours were scored by both clinicians as requiring no or minor edits. ConclusionsOur autocontouring model was shown to produce clinically usable disease outlines, based on CT alone, for approximately two-thirds of patients undergoing lung radiotherapy. Further work is necessary to improve this before clinical implementation.

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

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. 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. 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. 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±2.1mm, and HD95 of 13.9±7.6mm. When the PET images were included, the performance of our model was improved to DSC of 0.62±0.14, MSD of 2.8±1.7mm, and HD95 of 10.5±6.5mm. 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. 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.

Diffeomorphic transformer-based abdomen MRI-CT deformable image registration.

Stereotactic body radiotherapy (SBRT) is a well-established treatment modality for liver metastases in patients unsuitable for surgery. Both CT and MRI are useful during treatment planning for accurate target delineation and to reduce potential organs-at-risk (OAR) toxicity from radiation. MRI-CT deformable image registration (DIR) is required to propagate the contours defined on high-contrast MRI to CT images. An accurate DIR method could lead to more precisely defined treatment volumes and superior OAR sparing on the treatment plan. Therefore, it is beneficial to develop an accurate MRI-CT DIR for liver SBRT. To create a new deep learning model that can estimate the deformation vector field (DVF) for directly registering abdominal MRI-CT images. The proposed method assumed a diffeomorphic deformation. By using topology-preserved deformation features extracted from the probabilistic diffeomorphic registration model, abdominal motion can be accurately obtained and utilized for DVF estimation. The model integrated Swin transformers, which have demonstrated superior performance in motion tracking, into the convolutional neural network (CNN) for deformation feature extraction. The model was optimized using a cross-modality image similarity loss and a surface matching loss. To compute the image loss, a modality-independent neighborhood descriptor (MIND) was used between the deformed MRI and CT images. The surface matching loss was determined by measuring the distance between the warped coordinates of the surfaces of contoured structures on the MRI and CT images. To evaluate the performance of the model, a retrospective study was carried out on a group of 50 liver cases that underwent rigid registration of MRI and CT scans. The deformed MRI image was assessed against the CT image using the target registration error (TRE), Dice similarity coefficient (DSC), and mean surface distance (MSD) between the deformed contours of the MRI image and manual contours of the CT image. When compared to only rigid registration, DIR with the proposed method resulted in an increase of the mean DSC values of the liver and portal vein from 0.850±0.102 and 0.628±0.129 to 0.903±0.044 and 0.763±0.073, a decrease of the mean MSD of the liver from 7.216±4.513mm to 3.232±1.483mm, and a decrease of the TRE from 26.238±2.769mm to 8.492±1.058mm. The proposed DIR method based on a diffeomorphic transformer provides an effective and efficient way to generate an accurate DVF from an MRI-CT image pair of the abdomen. It could be utilized in the current treatment planning workflow for liver SBRT.

Cardiac MRI segmentation using shifted-window multilayer perceptron mixer networks

Objectives. In this work, we proposed a deep-learning segmentation algorithm for cardiac magnetic resonance imaging to aid in contouring of the left ventricle, right ventricle, and Myocardium (Myo). Approach. We proposed a shifted window multilayer perceptron (Swin-MLP) mixer network which is built upon a 3D U-shaped symmetric encoder-decoder structure. We evaluated our proposed network using public data from 100 individuals. The network performance was quantitatively evaluated using 3D volume similarity between the ground truth contours and the predictions using Dice score coefficient, sensitivity, and precision as well as 2D surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMSD). We benchmarked the performance against two other current leading edge networks known as Dynamic UNet and Swin-UNetr on the same public dataset. Results. The proposed network achieved the following volume similarity metrics when averaged over three cardiac segments: Dice = 0.952 ± 0.017, precision = 0.948 ± 0.016, sensitivity = 0.956 ± 0.022. The average surface similarities were HD = 1.521 ± 0.121 mm, MSD = 0.266 ± 0.075 mm, and RMSD = 0.668 ± 0.288 mm. The network shows statistically significant improvement in comparison to the Dynamic UNet and Swin-UNetr algorithms for most volumetric and surface metrics with p-value less than 0.05. Overall, the proposed Swin-MLP mixer network demonstrates better or comparable performance than competing methods. Significance. The proposed Swin-MLP mixer network demonstrates more accurate segmentation performance compared to current leading edge methods. This robust method demonstrates the potential to streamline clinical workflows for multiple applications.

A transformer-guided cross-modality adaptive feature fusion framework for esophageal gross tumor volume segmentation

Background and ObjectiveAccurate segmentation of esophageal gross tumor volume (GTV) indirectly enhances the efficacy of radiotherapy for patients with esophagus cancer. In this domain, learning-based methods have been employed to fuse cross-modality positron emission tomography (PET) and computed tomography (CT) images, aiming to improve segmentation accuracy. This fusion is essential as it combines functional metabolic information from PET with anatomical information from CT, providing complementary information. While the existing three-dimensional (3D) segmentation method has achieved state-of-the-art (SOTA) performance, it typically relies on pure-convolution architectures, limiting its ability to capture long-range spatial dependencies due to convolution's confinement to a local receptive field. To address this limitation and further enhance esophageal GTV segmentation performance, this work proposes a transformer-guided cross-modality adaptive feature fusion network, referred to as TransAttPSNN, which is based on cross-modality PET/CT scans. MethodsSpecifically, we establish an attention progressive semantically-nested network (AttPSNN) by incorporating the convolutional attention mechanism into the progressive semantically-nested network (PSNN). Subsequently, we devise a plug-and-play transformer-guided cross-modality adaptive feature fusion model, which is inserted between the multi-scale feature counterparts of a two-stream AttPSNN backbone (one for the PET modality flow and another for the CT modality flow), resulting in the proposed TransAttPSNN architecture. ResultsThrough extensive four-fold cross-validation experiments on the clinical PET/CT cohort. The proposed approach acquires a Dice similarity coefficient (DSC) of 0.76 ± 0.13, a Hausdorff distance (HD) of 9.38 ± 8.76 mm, and a Mean surface distance (MSD) of 1.13 ± 0.94 mm, outperforming the SOTA competing methods. The qualitative results show a satisfying consistency with the lesion areas. ConclusionsThe devised transformer-guided cross-modality adaptive feature fusion module integrates the strengths of PET and CT, effectively enhancing the segmentation performance of esophageal GTV. The proposed TransAttPSNN has further advanced the research of esophageal GTV segmentation.

Automatic segmentation of tumour and organs at risk in 3D MRI for cervical cancer radiation therapy with anatomical variations.

Cervical cancer is a common cancer in women globally, with treatment usually involving radiation therapy (RT). Accurate segmentation for the tumour site and organ-at-risks (OARs) could assist in the reduction of treatment side effects and improve treatment planning efficiency. Cervical cancer Magnetic Resonance Imaging (MRI) segmentation is challenging due to a limited amount of training data available and large inter- and intra- patient shape variation for OARs. The proposed Masked-Net consists of a masked encoder within the 3D U-Net to account for the large shape variation within the dataset, with additional dilated layers added to improve segmentation performance. A new loss function was introduced to consider the bounding box loss during training with the proposed Masked-Net. Transfer learning from a male pelvis MRI data with a similar field of view was included. The approaches were compared to the 3D U-Net which was widely used in MRI image segmentation. The data used consisted of 52 volumes obtained from 23 patients with stage IB to IVB cervical cancer across a maximum of 7 weeks of RT with manually contoured labels including the bladder, cervix, gross tumour volume, uterus and rectum. The model was trained and tested with a 5-fold cross validation. Outcomes were evaluated based on the Dice Similarity Coefficients (DSC), the Hausdorff Distance (HD) and the Mean Surface Distance (MSD). The proposed method accounted for the small dataset, large variations in OAR shape and tumour sizes with an average DSC, HD and MSD for all anatomical structures of 0.790, 30.19mm and 3.15mm respectively.

Fully semantic segmentation for rectal cancer based on post-nCRT MRl modality and deep learning framework

PurposeRectal tumor segmentation on post neoadjuvant chemoradiotherapy (nCRT) magnetic resonance imaging (MRI) has great significance for tumor measurement, radiomics analysis, treatment planning, and operative strategy. In this study, we developed and evaluated segmentation potential exclusively on post-chemoradiation T2-weighted MRI using convolutional neural networks, with the aim of reducing the detection workload for radiologists and clinicians.MethodsA total of 372 consecutive patients with LARC were retrospectively enrolled from October 2015 to December 2017. The standard-of-care neoadjuvant process included 22-fraction intensity-modulated radiation therapy and oral capecitabine. Further, 243 patients (3061 slices) were grouped into training and validation datasets with a random 80:20 split, and 41 patients (408 slices) were used as the test dataset. A symmetric eight-layer deep network was developed using the nnU-Net Framework, which outputs the segmentation result with the same size. The trained deep learning (DL) network was examined using fivefold cross-validation and tumor lesions with different TRGs.ResultsAt the stage of testing, the Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), and mean surface distance (MSD) were applied to quantitatively evaluate the performance of generalization. Considering the test dataset (41 patients, 408 slices), the average DSC, HD95, and MSD were 0.700 (95% CI: 0.680–0.720), 17.73 mm (95% CI: 16.08–19.39), and 3.11 mm (95% CI: 2.67–3.56), respectively. Eighty-two percent of the MSD values were less than 5 mm, and fifty-five percent were less than 2 mm (median 1.62 mm, minimum 0.07 mm).ConclusionsThe experimental results indicated that the constructed pipeline could achieve relatively high accuracy. Future work will focus on assessing the performances with multicentre external validation.

Deep learning for synovial volume segmentation of the first carpometacarpal joint in osteoarthritis patients

ObjectiveThe objective of this study was to develop a deep-learning-based approach to automatically segment 3D ultrasound images of the synovial tissue in osteoarthritis of the first carpometacarpal (CMC1 OA). DesignDeep learning predictions on 2D ultrasound slices sampled in the transverse plane were used to view the synovial tissue of the first carpometacarpal in patients with OA, followed by reconstruction into 3D surfaces. A modified 2D U-Net was trained using a dataset of 832 2D US images resliced from 89 3D US images. Segmentation accuracy was evaluated using a testing dataset of 208 2D US images resliced from 15 3D US images. Absolute and signed performance metrics were computed, and segmentation performance was compared between the manual segmentations of raters 1 and 2. ResultsResults of the U-Net-based run were mean 3D DSC 86.9 ± 4.8%, recall 93.7 ± 3.6%, precision 81.1 ± 6.9%, volume percent difference 16.9 ± 10.2%, mean surface distance 0.18 ± 0.04 mm, and Hausdorff distance 1.8 ± 0.8 mm. The algorithm demonstrated an overall increase in performance after 3D segmentation reconstruction compared to 2D predictions, but the difference was not statistically significant. ConclusionThis study investigated the use of a modified U-Net algorithm to automatically segment the synovial tissue volume (STV) of CMC1 OA patients and demonstrated that the addition of this deep learning method increases the efficiency of STV estimations in clinical trial settings.

Morphometric and clinical comparison of MRI-based synthetic CT to conventional CT of the hip in children.

MRI-based synthetic CT (sCT) generates CT-like images from MRI data. To evaluate equivalence, inter- and intraobserver reliability, and image quality of sCT compared to conventional (cCT) for assessing hip morphology and maturity in pediatric patients. We prospectively enrolled patients <21 years old with cCT and 3T MRI of the hips/pelvis. A dual-echo gradient-echo sequence was used to generate sCT via a commercially available post-processing software (BoneMRI v1.5 research version, MRIguidance BV, Utrecht, NL). Two pediatric musculoskeletal radiologists measured seven morphologic hip parameters. 3D surface distances between cCT and sCT were computed. Physeal status was established at seven locations with cCT as reference standard. Images were qualitatively scored on a 5-point Likert scale regarding diagnostic quality, signal-to-noise ratio, clarity of bony margin, corticomedullary differentiation, and presence and severity of artifacts. Quantitative evaluation of Hounsfield units (HU) was performed in bone, muscle, and fat tissue.Inter- and intraobserver reliability were measured by intraclass correlation coefficients. The cCT-to-sCT intermodal agreement was assessed via Bland-Altman analysis. The equivalence between modalities was tested using paired two one-sided tests. The quality parameter scores of each imaging modality were compared via Wilcoxon signed-rank test. For tissue-specific HU measurements, mean absolute error and mean percentage error values were calculated using the cCT as the reference standard. Thirty-eight hips in 19 patients were included (16.6 ± 3 years, range 9.9-20.9; male = 5). cCT- and sCT-based morphologic measurements demonstrated good to excellent inter- and intraobserver correlation (0.77<ICC<0.98). Measurements were statistically equivalent (P< 0.05), with average intermodal differences <2°. Mean surface distance between cCT and sCT was 0.23 ± 0.18mm. Accuracy to determine physeal maturity status by sCT was 98.4% (242/246, 95% CI 96.0-99.6). No significant differences were found on overall diagnostic quality, bony margin clarity, corticomedullary differentiation, and artifacts between sCT and cCT; signal-to-noise ratio was higher on sCT (P< 0.001). HU were within reference values on both sCT and cCT. sCT is equivalent to cCT for the assessment of hip morphology, physeal status, and radiodensity assessment in pediatric patients.

Transfer learning for auto-segmentation of 17 organs-at-risk in the head and neck: Bridging the gap between institutional and public datasets.

Auto-segmentation of organs-at-risk (OARs) in the head and neck (HN) on computed tomography (CT) images is a time-consuming component of the radiation therapy pipeline that suffers from inter-observer variability. Deep learning (DL) has shown state-of-the-art results in CT auto-segmentation, with larger and more diverse datasets showing better segmentation performance. Institutional CT auto-segmentation datasets have been small historically (n<50) due to the time required for manual curation of images and anatomical labels. Recently, large public CT auto-segmentation datasets (n>1000 aggregated) have become available through online repositories such as The Cancer Imaging Archive. Transfer learning is a technique applied when training samples are scarce, but a large dataset from a closely related domain is available. The purpose of this study was to investigate whether a large public dataset could be used in place of an institutional dataset (n>500), or to augment performance via transfer learning, when building HN OAR auto-segmentation models for institutional use. Auto-segmentation models were trained on a large public dataset (public models) and a smaller institutional dataset (institutional models). The public models were fine-tuned on the institutional dataset using transfer learning (transfer models). We assessed both public model generalizability and transfer model performance by comparison with institutional models. Additionally, the effect of institutional dataset size on both transfer and institutional models was investigated. All DL models used a high-resolution, two-stage architecture based on the popular 3D U-Net. Model performance was evaluated using five geometric measures: the dice similarity coefficient (DSC), surface DSC, 95th percentile Hausdorff distance, mean surface distance (MSD), and added path length. For a small subset of OARs (left/right optic nerve, spinal cord, left submandibular), the public models performed significantly better (p<0.05) than, or showed no significant difference to, the institutional models under most of the metrics examined. For the remaining OARs, the public models were inferior to the institutional models, although performance differences were small (DSC≤0.03, MSD<0.5mm) for seven OARs (brainstem, left/right lens, left/right parotid, mandible, right submandibular). The transfer models performed significantly better than the institutional models for seven OARs (brainstem, right lens, left/right optic nerve, left/right parotid, spinal cord) with a small margin of improvement (DSC≤0.02, MSD<0.4mm). When numbers of institutional training samples were limited, public and transfer models outperformed the institutional models for most OARs (brainstem, left/right lens, left/right optic nerve, left/right parotid, spinal cord, and left/right submandibular). Training auto-segmentation models with public data alone was suitable for a small number of OARs. Using only public data incurred a small performance deficit for most other OARs, when compared with institutional data alone, but may be preferable over time-consuming curation of a large institutional dataset. When a large institutional dataset was available, transfer learning with models pretrained on a large public dataset provided a modest performance improvement for several OARs. When numbers of institutional samples were limited, using the public dataset alone, or as a pretrained model, was beneficial for most OARs.

Brain Extraction Methods in Neonatal Brain MRI and Their Effects on Intracranial Volumes

Magnetic resonance imaging (MRI) plays an important role in assessing early brain development and injury in neonates. When using an automated volumetric analysis, brain tissue segmentation is necessary, preceded by brain extraction (BE) to remove non-brain tissue. BE remains challenging in neonatal brain MRI, and despite the existence of several methods, manual segmentation is still considered the gold standard. Therefore, the purpose of this study was to assess different BE methods in the MRI of preterm neonates and their effects on the estimation of intracranial volumes (ICVs). This study included twenty-two premature neonates (mean gestational age ± standard deviation: 28.4 ± 2.1 weeks) with MRI brain scans acquired at term, without detectable lesions or congenital conditions. Manual segmentation was performed for T2-weighted scans to establish reference brain masks. Four automated BE methods were used: Brain Extraction Tool (BET2); Simple Watershed Scalping (SWS); HD Brain Extraction Tool (HD-BET); and SynthStrip. Regarding segmentation metrics, HD-BET outperformed the other methods with median improvements of +0.031 (BET2), +0.002 (SWS), and +0.011 (SynthStrip) points for the dice coefficient; and −0.786 (BET2), −0.055 (SWS), and −0.124 (SynthStrip) mm for the mean surface distance. Regarding ICVs, SWS and HD-BET provided acceptable levels of agreement with manual segmentation, with mean differences of −1.42% and 2.59%, respectively.

Custom-Trained Deep Learning-Based Auto-Segmentation for Male Pelvic Iterative CBCT on C-Arm Linear Accelerators

Purpose: To evaluate the clinical applicability of a commercial artificial intelligence (AI)-driven deep learning auto-segmentation (DLAS) tool on enhanced iterative cone-beam CT (iCBCT) acquisitions for intact prostate and prostate bed treatments. Methods and MaterialsDLAS models were trained using 116 iCBCT datasets with manually delineated organs-at-risk (OARs – bladder, femoral heads, and rectum) and target volumes (intact prostate and prostate bed) adhering to institution-specific contouring guidelines. An additional 25 intact prostate and prostate bed iCBCT datasets were utilized for model testing. Segmentation accuracy relative to a reference structure set was quantified using various geometric comparison metrics and qualitatively evaluated by trained physicists and physicians. These results were compared to those obtained for an additional DLAS-based model trained on planning CT (pCT) datasets and for a deformable image registration (DIR)-based automatic contour propagation method. Results: In most instances, statistically significant differences in the Dice similarity coefficient (DSC), 95% directed Hausdorff distance, and mean surface distance metrics were observed between the models, as the iCBCT-trained DLAS model outperformed the pCT-trained DLAS model and DIR-based method for all OARs and the intact prostate target volume. Mean DSC values for the proposed method were ≥0.90 for these volumes-of-interest. The iCBCT-trained DLAS model demonstrated a relatively suboptimal performance for the prostate bed segmentation, as the mean DSC value was <0.75 for this target contour. Overall, 90% of bladder, 93% of femoral head, 67% of rectum, and 92% of intact prostate contours generated by the proposed method were deemed clinically acceptable based on qualitative scoring, while approximately 63% of prostate bed contours required moderate or major manual editing to adhere to institutional contouring guidelines. Conclusion: The proposed method presents the potential for improved segmentation accuracy and efficiency when compared to the DIR-based automatic contour propagation method as commonly applied in CBCT-based dose evaluation and calculation studies.