Average Dice Coefficient Research Articles

TSU-net: Two-stage multi-scale cascade and multi-field fusion U-net for right ventricular segmentation

Accurate segmentation of the right ventricle from cardiac magnetic resonance images (MRI) is a critical step in cardiac function analysis and disease diagnosis. It is still an open problem due to some difficulties, such as a large variety of object sizes and ill-defined borders. In this paper, we present a TSU-net network that grips deeper features and captures targets of different sizes with multi-scale cascade and multi-field fusion in the right ventricle. TSU-net mainly contains two major components: Dilated-Convolution Block (DB) and Multi-Layer-Pool Block (MB). DB extracts and aggregates multi-scale features for the right ventricle. MB mainly relies on multiple effective field-of-views to detect objects at different sizes and fill boundary features. Different from previous networks, we used DB and MB to replace the convolution layer in the encoding layer, thus, we can gather multi-scale information of right ventricle, detect different size targets and fill boundary information in each encoding layer. In addition, in the decoding layer, we used DB to replace the convolution layer, so that we can aggregate the multi-scale features of the right ventricle in each decoding layer. Furthermore, the two-stage U-net structure is used to further improve the utilization of DB and MB through a two-layer encoding/decoding layer. Our method is validated on the RVSC, a public right ventricular data set. The results demonstrated that TSU-net achieved an average Dice coefficient of 0.86 on endocardium and 0.90 on the epicardium, thereby outperforming other models. It effectively assists doctors to diagnose the disease and promotes the development of medical images. In addition, we also provide an intuitive explanation of our network, which fully explain MB and TSU-net's ability to detect targets of different sizes and fill in boundary features.

Precise laminae segmentation based on neural network for robot-assisted decompressive laminectomy

Background and Objective: The decompressive laminectomy is one of the most common operations to treat lumbar spinal stenosis by removing the laminae above the spinal nerve. Recently, an increasing number of robots are deployed during the surgical process to reduce the burden on surgeons and to reduce complications. However, for the robot-assisted decompressive laminectomy, an accurate 3D model of laminae from a CT image is highly desired. The purpose of this paper is to precisely segment the laminae with fewer calculations.Methods: We propose a two-stage neural network SegRe-Net. In the first stage, the entire intraoperative CT image is inputted to acquire the coarse segmentation of vertebrae with low resolution and the probability map of the laminar centers. The second stage is trained to refine the segmentation of laminae.Results: Three public available datasets were used to train and validate the models. The experimental results demonstrated the effectiveness of the proposed network on laminar segmentation with an average Dice coefficient of 96.38% and an average symmetric surface distance of 0.097 mm.Conclusion: The proposed two-stage network can achieve better results than those baseline models in the laminae segmentation task with less calculation amount and learnable parameters. Our methods improve the accuracy of laminar models and reduce the image processing time. It can be used to provide a more precise planning trajectory and may promote the clinical application for the robot-assisted decompression laminectomy surgery.

Volumetric Accuracy Analysis of Virtual Safety Barriers for Cooperative-Control Robotic Mastoidectomy.

Virtual fixtures can be enforced in cooperative-control robotic mastoidectomies with submillimeter accuracy. Otologic procedures are well-suited for robotic assistance due to consistent osseous landmarks. We have previously demonstrated the feasibility of cooperative-control robots (CCRs) for mastoidectomy. CCRs manipulate instruments simultaneously with the surgeon, allowing the surgeon to control instruments with robotic augmentation of motion. CCRs can also enforce virtual fixtures, which are safety barriers that prevent motion into undesired locations. Previous studies have validated the ability of CCRs to allow a novice surgeon to safely complete a cortical mastoidectomy. This study provides objective accuracy data for CCR-imposed safety barriers in cortical mastoidectomies. Temporal bone phantoms were registered to a CCR using preoperative computed tomography (CT) imaging. Virtual fixtures were created using 3D Slicer, with 2D planes placed along the external auditory canal, tegmen, and sigmoid, converging on the antrum. Five mastoidectomies were performed by a novice surgeon, moving the drill to the limit of the barriers. Postoperative CT scans were obtained, and Dice coefficients and Hausdorff distances were calculated. The average modified Hausdorff distance between drilled bone and the preplanned volume was 0.351 ± 0.093 mm. Compared with the preplanned volume of 0.947 cm3, the mean volume of bone removed was 1.045 cm3 (difference of 0.0982 cm3 or 10.36%), with an average Dice coefficient of 0.741 (range, 0.665-0.802). CCR virtual fixtures can be enforced with a high degree of accuracy. Future studies will focus on improving accuracy and developing 3D fixtures around relevant surgical anatomy.

Lesion probability mapping in MS patients using a regression network on MR fingerprinting

BackgroundTo develop a regression neural network for the reconstruction of lesion probability maps on Magnetic Resonance Fingerprinting using echo-planar imaging (MRF-EPI) in addition to T_1, {T_2}^*, NAWM, and GM- probability maps.MethodsWe performed MRF-EPI measurements in 42 patients with multiple sclerosis and 6 healthy volunteers along two sites. A U-net was trained to reconstruct the denoised and distortion corrected T_1 and {T_2}^* maps, and to additionally generate NAWM-, GM-, and WM lesion probability maps.ResultsWM lesions were predicted with a dice coefficient of 0.61pm 0.09 and a lesion detection rate of 0.85pm 0.25 for a threshold of 33%. The network jointly enabled accurate T_1 and {T_2}^* times with relative deviations of 5.2% and 5.1% and average dice coefficients of 0.92pm 0.04 and 0.91pm 0.03 for NAWM and GM after binarizing with a threshold of 80%.ConclusionDL is a promising tool for the prediction of lesion probability maps in a fraction of time. These might be of clinical interest for the WM lesion analysis in MS patients.

Abstract 3071: A method for infratentorial structures segmentation for tumor treating fields treatment planning

Abstract Introduction: Tumor treating fields (TTFields) is an FDA approved treatment for the management of glioblastoma multiform (GBM) and mesothelioma (MPM) and is associated with a significant extension to patients' survival. We have developed software to optimize the location of the TTFields transducer arrays (TAs) on GBM patients' head to maximize the field in the tumor region. To that end, we first segment the patients' head MRIs into five normal and three abnormal tissues. However, the current method for head segmentation was found to be less accurate at the infratentorial regions, mark some sinuses with cerebrospinal fluid (CSF), and does not identify the cerebellum and brain-stem. Therefore, it is assuming the infratentorial structures have the same electrical properties as in the cerebrum. Methods: We have developed a new method for segmentation of the head MRI that overcomes these limitations. As in the previous method, we have incorporated an atlas that is composed of an MRI image and tissue probability maps (TPM). The TPM assign each voxel in the atlas MRI with a value in the range of 0-1. This value reflects the probability that the specific tissue resides in that voxel. We have generated new customized TPMs for the cerebellum, brain stem and sinuses. Moreover, we have carefully revised the other TPMs to ensure best results. Last, we have incorporated these TPMs in a new atlas-based segmentation method. We have validated the method on 10 GBM patients T1w +gad head MRIs. Results: The average Dice coefficient between gold-standard and algorithms' segmentation was 68.4% (SD=11.2%) that is similar to the value observed with the current method implemented in the treatment-planning software (67.8%; SD=10.6%). However, the new method presented here extends the current method with the cerebellum (average Dice = 84%) and brainstem (average Dice = 67%). Moreover, the sinuses are marked with air, and erroneous CSF segments in the sinuses were removed. Conclusions: We have presented a novel method for the segmentation of the head that facilitates accurate infratentorial and supratentorial structures' segmentation. Moreover, the method eliminates erroneous CSF segmentation in the sinuses. These improvements pave the way for TTFields planning for infratentorial tumors. Citation Format: Yana Glozman, Reuben R. Shamir, Zeev Bomzon. A method for infratentorial structures segmentation for tumor treating fields treatment planning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 3071.

Automatic detection of thyroid nodules with ultrasound images: Basing on semi-supervised learning

Ultrasonography is one of the most common method for the diagnosis of thyroid nodules clinically. But the lack of experienced radiologists has become the biggest problem which reduce the accuracy and effectiveness of thyroid nodules’ detection. What’s more, although artificial intelligence (AI) has been widely used to solve the problem, the performance is limited by the size of the database. In this paper, an improved U-Net is designed for the segmentation of thyroid nodules, following a fully convolution network for the classification of thyroid nodules. During the training, pseudo-label is used for semi-supervised learning. The whole work is trained in 2032 ultrasound images (1000 nodules) and tested in 400 ultrasound images (200 nodules). The average dice coefficient is 89.7% in the test set for segmentation, while the precision and recall is 93.2% and 96% separately for classification. These results are higher than the method which only use the images with true labels, proving the effectiveness of pseudo-label.

Diagnosis of obstructive sleep apnea with prediction of flow characteristics according to airway morphology automatically extracted from medical images: Computational fluid dynamics and artificial intelligence approach

BackgroundObstructive sleep apnea syndrome (OSAS) is being observed in an increasing number of cases. It can be diagnosed using several methods such as polysomnography. ObjectivesTo overcome the challenges of time and cost faced by conventional diagnostic methods, this paper proposes computational fluid dynamics (CFD) and machine-learning approaches that are derived from the upper-airway morphology with automatic segmentation using deep learning. MethodWe adopted a 3D UNet deep-learning model to perform medical image segmentation. 3D UNet prevents the feature-extraction loss that may occur by concatenating layers and extracts the anteroposterior coordination and width of the airway morphology. To create flow characteristics of the upper airway training data, we analyzed the changes in flow characteristics according to the upper-airway morphology using CFD. A multivariate Gaussian process regression (MVGPR) model was used to train the flow characteristic values. The trained MVGPR enables the prompt prediction of the aerodynamic features of the upper airway without simulation. Unlike conventional regression methods, MVGPR can be trained by considering the correlation between the flow characteristics. As a diagnostic step, a support vector machine (SVM) with predicted aerodynamic and biometric features was used in this study to classify patients as healthy or suffering from moderate OSAS. SVM is beneficial as it is easy to learn even with a small dataset, and it can diagnose various flow characteristics as factors while enhancing the feature via the kernel function. As the patient dataset is small, the Monte Carlo cross-validation was used to validate the trained model. Furthermore, to overcome the imbalanced data problem, the oversampling method was applied. ResultThe segmented upper-airway results of the high-resolution and low-resolution models present overall average dice coefficients of 0.76±0.041 and 0.74±0.052, respectively. Furthermore, the classification accuracy, sensitivity, specificity, and F1-score of the diagnosis algorithm were 81.5%, 89.3%, 86.2%, and 87.6%, respectively. ConclusionThe convenience and accuracy of sleep apnea diagnosis are improved using deep learning and machine learning. Further, the proposed method can aid clinicians in making appropriate decisions to evaluate the possible applications of OSAS.

Segmentation and recognition of breast ultrasound images based on an expanded U-Net.

This paper establishes a fully automatic real-time image segmentation and recognition system for breast ultrasound intervention robots. It adopts the basic architecture of a U-shaped convolutional network (U-Net), analyses the actual application scenarios of semantic segmentation of breast ultrasound images, and adds dropout layers to the U-Net architecture to reduce the redundancy in texture details and prevent overfitting. The main innovation of this paper is proposing an expanded training approach to obtain an expanded of U-Net. The output map of the expanded U-Net can retain texture details and edge features of breast tumours. Using the grey-level probability labels to train the U-Net is faster than using ordinary labels. The average Dice coefficient (standard deviation) and the average IOU coefficient (standard deviation) are 90.5% (±0.02) and 82.7% (±0.02), respectively, when using the expanded training approach. The Dice coefficient of the expanded U-Net is 7.6 larger than that of a general U-Net, and the IOU coefficient of the expanded U-Net is 11 larger than that of the general U-Net. The context of breast ultrasound images can be extracted, and texture details and edge features of tumours can be retained by the expanded U-Net. Using an expanded U-Net can quickly and automatically achieve precise segmentation and multi-class recognition of breast ultrasound images.

Deep learning based lesion detection from anterior segment optical coherence tomography images and its application in the diagnosis of keratoconus

Objective: To developed an image analysis system of anterior segment optical coherence tomography (AS-OCT) examination results based on deep learning technology, and to evaluate its effect in identifying various types of corneal pathologies and quantified indices. Methods: A total of 4 026 patients (5 617 eyes), including 1 977 males and 2 049 females, aged (45±23) years, were enrolled in Qingdao Eye Hospital from January 2011 to August 2019. The AS-OCT images were used as a training dataset, which were labeled with location information of 16 corneal pathologies (including corneal epithelial defect, corneal epithelial thickening, corneal thinning and so on) by clinical experts, as well as the tissue stratification of the corneal epithelium and stroma. The labeled AS-OCT images were used to train the corneal pathology detection model and corneal stratification model based on deep convolutional neural network algorithm. Then 1 709 AS-OCT images of the affected eyes were collected as a validation dataset. Compared with the artificial labeling results, the accuracy, sensitivity and specificity were evaluated in the corneal pathology detection model, and the overlapping rate (Dice coefficient) between the labeled area of the model and the artificial labeling area was used to evaluate the corneal stratification model. Results: The results of 5 617 training sets showed that there were 1 472 cases of corneal epithelial defect, 2 416 cases of corneal epithelial thickening, 2 001 cases of corneal thinning, 780 cases of corneal lordosis, 2 064 cases of corneal thickening, 358 cases of subepithelial blisters, 486 cases of subepithelial opacity, 1 010 cases of corneal ulcer, 3 635 cases of stromal opacity, 1 060 cases of posterior elastic layer fold, 137 cases of posterior elastic layer detachment, 665 cases of keratic precipitate, 176 cases of corneal perforation, 127 cases of corneal foreign body, 299 cases of after lamellar keratoplasty (LKP) and 234 cases of after penetrating keratoplasty (PKP). Among 1 709 images, 1 596 were manually labeled. The average sensitivity and specificity of the corneal pathology detection model were 96.5% and 96.1% compared with the results of manual labeling. Fifteen samples were missed for detection, and the rate was 0.93%. The average Dice coefficients of the corneal stratification model for the corneal epithelium and stroma were 0.985 and 0.917, respectively. Conclusions: Our artificial intelligence-based diagnosis system with AS-OCT is able to give quantified information and location information of corneal lesions with high accuracy, which can help ophthalmologists improve the efficiency and accuracy of diagnosis. (Chin J Ophthalmol, 2021, 57: 447-453).

Segmentation and Identification of Vertebrae in CT Scans Using CNN, k-Means Clustering and k-NN

The accurate segmentation and identification of vertebrae presents the foundations for spine analysis including fractures, malfunctions and other visual insights. The large-scale vertebrae segmentation challenge (VerSe), organized as a competition at the Medical Image Computing and Computer Assisted Intervention (MICCAI), is aimed at vertebrae segmentation and labeling. In this paper, we propose a framework that addresses the tasks of vertebrae segmentation and identification by exploiting both deep learning and classical machine learning methodologies. The proposed solution comprises two phases: a binary fully automated segmentation of the whole spine, which exploits a 3D convolutional neural network, and a semi-automated procedure that allows locating vertebrae centroids using traditional machine learning algorithms. Unlike other approaches, the proposed method comes with the added advantage of no requirement for single vertebrae-level annotations to be trained. A dataset of 214 CT scans has been extracted from VerSe’20 challenge data, for training, validating and testing the proposed approach. In addition, to evaluate the robustness of the segmentation and labeling algorithms, 12 CT scans from subjects affected by severe, moderate and mild scoliosis have been collected from a local medical clinic. On the designated test set from Verse’20 data, the binary spine segmentation stage allowed to obtain a binary Dice coefficient of 89.17%, whilst the vertebrae identification one reached an average multi-class Dice coefficient of 90.09%. In order to ensure the reproducibility of the algorithms hereby developed, the code has been made publicly available.

Automated tumor segmentation in MR brain image using fuzzy c-means clustering and seeded region methodology

<span id="docs-internal-guid-c8cba487-7fff-2314-f38a-f2936a74e0fd"><span>Automated segmentation of a tumor is still a considerably exciting research topic in the medical imaging processing field, and it plays a considerable role in forming a right diagnosis, to aid effective medical treatment. In this work, a fully automated system for segmentation of the brain tumor in MRI images is introduced. The suggested system consists of three parts: Initially, the image is pre-processed to enhance contrast, eliminate noise, and strip the skull from the image using filtering and morphological operations. Secondly, segmentation of the image happens using two techniques, fuzzy c-means clustering (FCM) and with the application of a seeded region growing algorithm (SGR). Thirdly, this method proposes a post-processing step to smooth segmentation region edges using morphological operations. The testing of the proposed system involved 233 patients, which included 287 MRI images. A comparison of the results ensued, with the manual verification of the traces performed by doctors, which ultimately proved an average Dice Coefficient of 90.13% and an average Jaccard Coefficient of 82.60% also, by comparison with traditional segmentation techniques such as FCM method. The segmentation results and quantitative data analysis demonstrates the effectiveness of the suggested system.</span></span>

Automated atlas-based segmentation for skull base surgical planning.

Computational surgical planning tools could help develop novel skull base surgical approaches that improve safety and patient outcomes. This defines a need for automated skull base segmentation to improve the usability of surgical planning software. The objective of this work was to design and validate an algorithm for atlas-based automated segmentation of skull base structures in individual image sets for skull base surgical planning. Advanced Normalization Tools software was used to construct a synthetic CT template from 6 subjects, and skull base structures were manually segmented to create a reference atlas. Landmark registration followed by Elastix deformable registration was applied to the template to register it to each of the 30 trusted reference image sets. Dice coefficient, average Hausdorff distance, and clinical usability scoring were used to compare the atlas segmentations to those of the trusted reference image sets. The mean for average Hausdorff distance for all structures was less than 2mm (mean for 95th percentile Hausdorff distance was less than 5mm). For structures greater than 2.5mL in volume, the average Dice coefficient was 0.73 (range 0.59-0.82), and for structures less than 2.5mL in volume the Dice coefficient was less than 0.7. The usability scoring survey was completed by three experts, and all structures met the criteria for acceptable effort except for the foramen spinosum, rotundum, and carotid artery, which required more than minor corrections. Currently available open-source algorithms, such as the Elastix deformable algorithm, can be used for automated atlas-based segmentation of skull base structures with acceptable clinical accuracy and minimal corrections with the use of the proposed atlas. The first publicly available CT template and anterior skull base segmentation atlas being released (available at this link: http://hdl.handle.net/1773/46259 ) with this paper will allow for general use of automated atlas-based segmentation of the skull base.

Automated delineation of orbital abscess depicted on CT scan using deep learning.

To develop and validate a deep learning algorithm to automatically detect and segment an orbital abscess depicted on computed tomography (CT). We retrospectively collected orbital CT scans acquired on 67 pediatric subjects with a confirmed orbital abscess in the setting of infectious orbital cellulitis. A context-aware convolutional neural network (CA-CNN) was developed and trained to automatically segment orbital abscess. To reduce the requirement for a large dataset, transfer learning was used by leveraging a pre-trained model for CT-based lung segmentation. An ophthalmologist manually delineated orbital abscesses depicted on the CT images. The classical U-Net and the CA-CNN models with and without transfer learning were trained and tested on the collected dataset using the 10-fold cross-validation method. Dice coefficient, Jaccard index, and Hausdorff distance were used as performance metrics to assess the agreement between the computerized and manual segmentations. The context-aware U-Net with transfer learning achieved an average Dice coefficient and Jaccard index of 0.78±0.12 and 0.65±0.13, which were consistently higher than the classical U-Net or the context-aware U-Net without transfer learning (P<0.01). The average differences of the abscess between the computerized results and the experts in terms of volume and Hausdorff distance were 0.10±0.11mL and 1.94±1.21mm, respectively. The context-aware U-Net detected all orbital abscess without false positives. The deep learning solution demonstrated promising performance in detecting and segmenting orbital abscesses on CT images in strong agreement with a human observer.

Segmentation of Brain Tumors from MRI Images Using Convolutional Autoencoder

The use of machine learning algorithms and modern technologies for automatic segmentation of brain tissue increases in everyday clinical diagnostics. One of the most commonly used machine learning algorithms for image processing is convolutional neural networks. We present a new convolutional neural autoencoder for brain tumor segmentation based on semantic segmentation. The developed architecture is small, and it is tested on the largest online image database. The dataset consists of 3064 T1-weighted contrast-enhanced magnetic resonance images. The proposed architecture’s performance is tested using a combination of two different data division methods, and two different evaluation methods, and by training the network with the original and augmented dataset. Using one of these data division methods, the network’s generalization ability in medical diagnostics was also tested. The best results were obtained for record-wise data division, training the network with the augmented dataset. The average accuracy classification of pixels is 99.23% and 99.28% for 5-fold cross-validation and one test, respectively, and the average dice coefficient is 71.68% and 72.87%. Considering the achieved performance results, execution speed, and subject generalization ability, the developed network has great potential for being a decision support system in everyday clinical practice.

Meta grayscale adaptive network for 3D integrated renal structures segmentation.

Three-dimensional (3D) integrated renal structures (IRS) segmentation targets segmenting the kidneys, renal tumors, arteries, and veins in one inference. Clinicians will benefit from the 3D IRS visual model for accurate preoperative planning and intraoperative guidance of laparoscopic partial nephrectomy (LPN). However, no success has been reported in 3D IRS segmentation due to the inherent challenges in grayscale distribution: low contrast caused by the narrow task-dependent distribution range of regions of interest (ROIs), and the networks representation preferences caused by the distribution variation inter-images. In this paper, we propose the Meta Greyscale Adaptive Network (MGANet), the first deep learning framework to simultaneously segment the kidney, renal tumors, arteries and veins on CTA images in one inference. It makes innovations in two collaborate aspects: 1) The Grayscale Interest Search (GIS) adaptively focuses segmentation networks on task-dependent grayscale distributions via scaling the window width and center with two cross-correlated coefficients for the first time, thus learning the fine-grained representation for fine segmentation. 2) The Meta Grayscale Adaptive (MGA) learning makes an image-level meta-learning strategy. It represents diverse robust features from multiple distributions, perceives the distribution characteristic, and generates the model parameters to fuse features dynamically according to image's distribution, thus adapting the grayscale distribution variation. This study enrolls 123 patients and the average Dice coefficients of the renal structures are up to 87.9%. Fine selection of the task-dependent grayscale distribution ranges and personalized fusion of multiple representations on different distributions will lead to better 3D IRS segmentation quality. Extensive experiments with promising results on renal structures reveal powerful segmentation accuracy and great clinical significance in renal cancer treatment.

Semantic Instance Segmentation of Kidney Cysts in MR Images: A Fully Automated 3D Approach Developed Through Active Learning

Total kidney volume (TKV) is the main imaging biomarker used to monitor disease progression and to classify patients affected by autosomal dominant polycystic kidney disease (ADPKD) for clinical trials. However, patients with similar TKVs may have drastically different cystic presentations and phenotypes. In an effort to quantify these cystic differences, we developed the first 3D semantic instance cyst segmentation algorithm for kidneys in MR images. We have reformulated both the object detection/localization task and the instance-based segmentation task into a semantic segmentation task. This allowed us to solve this unique imaging problem efficiently, even for patients with thousands of cysts. To do this, a convolutional neural network (CNN) was trained to learn cyst edges and cyst cores. Images were converted from instance cyst segmentations to semantic edge-core segmentations by applying a 3D erosion morphology operator to up-sampled versions of the images. The reduced cysts were labeled as core; the eroded areas were dilated in 2D and labeled as edge. The network was trained on 30 MR images and validated on 10 MR images using a fourfold cross-validation procedure. The final ensemble model was tested on 20 MR images not seen during the initial training/validation. The results from the test set were compared to segmentations from two readers. The presented model achieved an averaged R2 value of 0.94 for cyst count, 1.00 for total cyst volume, 0.94 for cystic index, and an averaged Dice coefficient of 0.85. These results demonstrate the feasibility of performing cyst segmentations automatically in ADPKD patients.

CT Guided Diagnosis: Cascaded U-Net for 3D Segmentation of Liver and Tumor

Volumetric estimation of the liver tumor is the first step to identifying the livers critical disorder. The liver and its tumor ratio prerequisite measures to select the therapeutic procedure. 3D printing and virtual reality platform require a segmented liver entity mask to evaluate the pre and post-treatment analysis. A cascaded U-Net model is proposed for automatic segmentation of liver and tumor in CT images. LiTS CT data set utilized for this study. The images were pre-processed using the windowing technique for contrast enhancement. Two U-Net models were modified for liver and tumor segmentation, respectively and connected in a cascaded manner. U-Net decoder end was modified in comparison to the original U-Net. The probability map of the first U-Net fed to the second U-Net and the input image to segment out the liver tumor. Eight subject volumetric CT datasets were utilized to test the cascaded U-Net performance and achieved average Dice coefficient for liver and tumor 0.95 and 0.69, respectively. Liver tumor diagnosis and treatment accuracy depend upon the precision of segmentation algorithms. Designed model segmented liver almost accurately and tumor segmented with limited accuracy. A further modification is required for the tumor segmentation cause of the occurrence of false negative.

A deep cascaded segmentation of obstructive sleep apnea-relevant organs from sagittal spine MRI

PurposeThe main purpose of this work was to develop an efficient approach for segmentation of structures that are relevant for diagnosis and treatment of obstructive sleep apnea syndrome (OSAS), namely pharynx, tongue, and soft palate, from mid-sagittal magnetic resonance imaging (MR) data. This framework will be applied to big data acquired within an on-going epidemiological study from a general population.MethodsA deep cascaded framework for subsequent segmentation of pharynx, tongue, and soft palate is presented. The pharyngeal structure was segmented first, since the airway was clearly visible in the T1-weighted sequence. Thereafter, it was used as an anatomical landmark for tongue location. Finally, the soft palate region was extracted using segmented tongue and pharynx structures and used as input for a deep network. In each segmentation step, a UNet-like architecture was applied.ResultsThe result assessment was performed qualitatively by comparing the region boundaries obtained from the expert to the framework results and quantitatively using the standard Dice coefficient metric. Additionally, cross-validation was applied to ensure that the framework performance did not depend on the specific selection of the validation set. The average Dice coefficients on the test set were 0.89pm 0.03, 0.87pm 0.02, and 0.79pm 0.08 for tongue, pharynx, and soft palate tissues, respectively. The results were similar to other approaches and consistent with expert readings.ConclusionDue to high speed and efficiency, the framework will be applied for big epidemiological data with thousands of participants acquired within the Study of Health in Pomerania as well as other epidemiological studies to provide information on the anatomical structures and aspects that constitute important risk factors to the OSAS development.

Tongue Tumor Detection in Hyperspectral Images Using Deep Learning Semantic Segmentation.

The utilization of hyperspectral imaging (HSI) in real-time tumor segmentation during a surgery have recently received much attention, but it remains a very challenging task. In this work, we propose semantic segmentation methods, and compare them with other relevant deep learning algorithms for tongue tumor segmentation. To the best of our knowledge, this is the first work using deep learning semantic segmentation for tumor detection in HSI data using channel selection, and accounting for more spatial tissue context, and global comparison between the prediction map, and the annotation per sample. Results, and Conclusion: On a clinical data set with tongue squamous cell carcinoma, our best method obtains very strong results of average dice coefficient, and area under the ROC-curve of [Formula: see text], and [Formula: see text], respectively on the original spatial image size. The results show that a very good performance can be achieved even with a limited amount of data. We demonstrate that important information regarding tumor decision is encoded in various channels, but some channel selection, and filtering is beneficial over the full spectra. Moreover, we use both visual (VIS), and near-infrared (NIR) spectrum, rather than commonly used only VIS spectrum; although VIS spectrum is generally of higher significance, we demonstrate NIR spectrum is crucial for tumor capturing in some cases. The HSI technology augmented with accurate deep learning algorithms has a huge potential to be a promising alternative to digital pathology or a doctors' supportive tool in real-time surgeries.

Abstract PO-014: Deep learning-based segmentation accurately captures histological features in cancer-free lymph nodes of breast cancer patients

Abstract Background Histological changes in the cancer-free lymph nodes (LNs) can add to the risk prediction of developing distant metastasis for LN-positive breast cancer patients. We have previously shown that the formation of germinal centres and enlarged sinus are premetastatic features and indicative of superior prognosis. As Whole Slide Images (WSI) demonstrated suitability for detection of LN metastasis with high accuracy based on Convolutional Neural Networks (CNN), we extended the CNN assessments to the detection and segmentation of these premetastatic features. Materials For 136 breast cancer patients, retrospectively collected from Guy’s Hospital, 1,422 Whole Slide Images (WSI) of axillary LNs (containing 1 to 3 LNs per WSI) and primary invasive tumours were 40x scanned by the Hamamatsu Nanozoomer. Two pathologists manually annotated germinal centres and sinuses in 111 of these images facilitating the ground-truth masks for a binary segmentation task using deep learning models. Results Each WSI was split into a series of overlapping patches along with corresponding ground-truth masks at 2.5x, 5x and 10x. Fully Convolutional Neural Networks (FCNs) based on the U-Net architecture were trained on patches from 95 annotated WSI to perform a segmentation for both germinal centres and sinus, with the remaining images used for validation (n= 6) and test sets (n= 10). In addition to a vanilla U-Net encoder-decoder architecture, a model with an attention mechanism to filter feature maps along skip connections was also explored. Additionally, a U-Net with dilated convolutions at each convolution block was used in a multi-scale mechanism that combines features from three different resolutions. To increase the robustness of the FCNs, training data was augmented using random image rotation and different colour perturbation transformations. Predictions were made on regions of WSIs containing whole LNs and were compared with the pathologist annotations on holdout testing images. The Dice coefficient was used to compare the predicted segmentations with the ground-truth annotations. Standard U-Net and attention U-Net architectures achieved an average Dice coefficient of 0.75 across the holdout testing images at 2.5x magnification for germinal centres, with individual images achieving a Dice score above 0.8. Multiscale models achieved the best results at 10x magnification with an average Dice score of 0.85 for germinal centres and 0.75 for sinus. The trained models were then applied to &amp;gt; 1,300 unseen and non-annotated WSIs of axillary LNs. Conclusion Our developed FCN models allow an automated, precise, reproducible and time-efficient approach to annotate germinal centres and sinuses in cancer-free and metastatic LNs. Trained models are currently applied to &amp;gt;1,300 unseen and non-annotated WSIs of axillary LNs, and will demonstrate their comparable performance with manual pathological assessment. Citation Format: Gregory Verghese, Anita Grigoriadis, Amit Sethi, Amit Lohan, Nikhil Cherian, Swati Meena, Harry Chinque, Fangfang Liu, Ed Martin, Tristan Clark, Cheryl Gillet, Julie Owen, Swapnil Rane. Deep learning-based segmentation accurately captures histological features in cancer-free lymph nodes of breast cancer patients [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-014.