Dice Similarity Coefficient Scores Research Articles

Artificial Intelligence–Based Autosegmentation: Advantages in Delineation, Absorbed Dose-Distribution, and Logistics

PurposeTo compare the performance of artificial intelligence (AI) in auto-contouring compared to a human practitioner in terms of precision, differences in dose distribution and time consumption. Methods and MaterialsDatasets of previously irradiated patients in three different segments (head and neck- (HNC), breast- (BC) and prostate cancer (PC)) were retrospectively collected. An experienced radiation oncologist (MD) performed organs-at-risk (OARs) and standard clinical target volume (CTV) delineations as baseline structures for comparison. AI-based auto-contours were generated in two additional CT copies; therefore, three groups were assessed: MD alone, AI alone, AI plus MD corrections (AI+C). Differences in Dice similarity coefficient (DSC) and person-hour burden were assessed. Furthermore, changes in clinically relevant dose-volume parameters were evaluated and compared. ResultsSeventy-five previously treated cases were collected (25 per segment) for the analysis. When compared to MD contours, the mean DSC scores were higher than 0.7 for 74% and 80% of AI and AI+C, respectively. After corrections, 17.1% structures presented DSC score deviations higher than 0.1 and 10.4% dose-volume parameters significantly changed in AI-contoured structures. The time consumption assessment yielded mean person-hour reductions of 68%, 51% and 71% for BC, PC and HNC, respectively. ConclusionIn great extent, AI yielded clinically acceptable OARs and certain CTVs in the explored anatomical segments. Sparse correction and assessment requirements place AI+C as a standard workflow. Minimal clinically relevant differences in OAR-exposure were identified. A substantial amount of person-hours could be repurposed with this technology.

MUE-CoT: multi-scale uncertainty entropy-aware co-training framework for left atrial segmentation

Objective. Accurate left atrial segmentation is the basis of the recognition and clinical analysis of atrial fibrillation. Supervised learning has achieved some competitive segmentation results, but the high annotation cost often limits its performance. Semi-supervised learning is implemented from limited labeled data and a large amount of unlabeled data and shows good potential in solving practical medical problems. Approach. In this study, we proposed a collaborative training framework for multi-scale uncertain entropy perception (MUE-CoT) and achieved efficient left atrial segmentation from a small amount of labeled data. Based on the pyramid feature network, learning is implemented from unlabeled data by minimizing the pyramid prediction difference. In addition, novel loss constraints are proposed for co-training in the study. The diversity loss is defined as a soft constraint so as to accelerate the convergence and a novel multi-scale uncertainty entropy calculation method and a consistency regularization term are proposed to measure the consistency between prediction results. The quality of pseudo-labels cannot be guaranteed in the pre-training period, so a confidence-dependent empirical Gaussian function is proposed to weight the pseudo-supervised loss. Main results. The experimental results of a publicly available dataset and an in-house clinical dataset proved that our method outperformed existing semi-supervised methods. For the two datasets with a labeled ratio of 5%, the Dice similarity coefficient scores were 84.94% ± 4.31 and 81.24% ± 2.4, the HD95 values were 4.63 mm ± 2.13 and 3.94 mm ± 2.72, and the Jaccard similarity coefficient scores were 74.00% ± 6.20 and 68.49% ± 3.39, respectively. Significance. The proposed model effectively addresses the challenges of limited data samples and high costs associated with manual annotation in the medical field, leading to enhanced segmentation accuracy.

Clinical Implementation and Evaluation of Auto-Segmentation Tools for Multi-Site Contouring in Radiotherapy

Background and purposeTools for auto-segmentation in radiotherapy are widely available, but guidelines for clinical implementation are missing. The goal was to develop a workflow for performance evaluation of three commercial auto-segmentation tools to select one candidate for clinical implementation. Materials and MethodsOne hundred patients with six treatment sites (brain, head-and-neck, thorax, abdomen, and pelvis) were included. Three sets of AI-based contours for organs-at-risk (OAR) generated by three software tools and manually drawn expert contours were blindly rated for contouring accuracy. The dice similarity coefficient (DSC), the Hausdorff distance, and a dose/volume evaluation based on the recalculation of the original treatment plan were assessed. Statistically significant differences were tested using the Kruskal-Wallis test and the post-hoc Dunn Test with Bonferroni correction. ResultsThe mean DSC scores compared to expert contours for all OARs combined were 0.80 ± 0.10, 0.75 ± 0.10, and 0.74 ± 0.11 for the three software tools. Physicians' rating identified equivalent or superior performance of some AI-based contours in head (eye, lens, optic nerve, brain, chiasm), thorax (e.g., heart and lungs), and pelvis and abdomen (e.g., kidney, femoral head) compared to manual contours. For some OARs, the AI models provided results requiring only minor corrections. Bowel-bag and stomach were not fit for direct use. During the interdisciplinary discussion, the physicians' rating was considered the most relevant. ConclusionA comprehensive method for evaluation and clinical implementation of commercially available auto-segmentation software was developed. The in-depth analysis yielded clear instructions for clinical use within the radiotherapy department.

CAU-Net: An Improved Attention U-Net for CT Head and Neck Organs at Risk Contouring

Segmentation of head and neck (H&N) organs at risk (OARs) is an intricate process, we propose Contest- extractor Attention U-Net (CAU-Net) to improve the segmentation accuracy of OARs contours by addressing the problem of limited accuracy of 2D segmentation algorithms in current radiotherapy techniques. A total of 60 patients from CSTRO2019's H&N dataset containing 22 organs, were available to train and evaluate a prototype deep learning-based normal tissue 2D auto-segmentation algorithm. Our CAU-Net is based on U-Net, by using the edge attention module to enhance the boundary representation, the null convolution block in the context extraction module to encode high-level semantic feature information, and the convolution of the sensory field to assign to different targets. A Dice loss function combine contour loss function was used in training the models. The contour loss function was improved to segment the target regions by the weights of different organ occurrences and the region assignment of false positives and false negatives to accurately predict the boundary structure. The OARs were delineated by a single experienced physician. A subset of 10 cases was withheld from training and used for validation. On those, we set three different deep-learning networks trained with CSTRO and compared them to the gold data: A) CAU-Net, B) nnUNet, and C) UNet++. To test its applicability, we used another public H&N dataset Public Domain Database for Computational Anatomy (PDDCA) containing 8 organs with 47 patients, among which 10 cases were used for validation: D) CAU-Net with the PDDCA, E) UNet2022 with the PDDCA. The Dice similarity coefficient (DSC) was used to measure the overlap between the results of the gold data and the automated segmentations. The average DSC scores for method A, B, and C across all OARs in the 10 evaluation cases were 0.67±0.08, 0.58±0.11 and 0.62±0.12, respectively. The difference in mean DSC scores was significant (p<0.05). The A/B difference was significant in Lens-L, Lens-R and Pituitary. Method A scored the highest DSC in all OARs except for the Spinal Cord, Mandible-L and Mandible-R. 16 OARs showed DSC≥0.6 on CSTRO. Method D, and E achieved 0.84±0.10 and 0.83±0.09 average DSC respectively. All OARs showed DSC≥0.7 on PDDCA. The CAU-Net proposed by us achieved better results than the baseline network for H&N OAR segmentation. This new development will provide the possibility of H&N organ segmentation and rapid diagnosis of radiotherapy. All the networks trained with PDDCA scored higher than CSTRO. Auto segmentation results can differ significantly when the same algorithm is trained on data from different institutions.

Unsupervised Domain Adaptation of Auto-Segmentation on Multi-Source MRIs

Deep learning has achieved great success in medical image segmentation. Most existing deep learning (DL) approaches make no adjustments to the model prior to inference. These models can perform well on the data of the same distribution, but their performance usually degrades when applied to the images from different source, i.e., different scanners. To tackle the problem caused by domain shift, we proposed an unsupervised domain adaptation (UDA) method based on entropy minimization and physical consistency constraints. The proposed method combines feature-level and instance-level domain adaptation techniques to transfer knowledge from the source to the target domain. Specifically, the feature-level adaptation technique uses a graph-based entropy minimization to reduce the discrepancy between the source and target domains. The instance-level adaptation technique employs a novel consistency loss to regularize the physical consistency of the same object, such as volume, length, and centroid, thus improving the segmentation accuracy of the target domain. A collection of 93 abdominal MR images, comprising 45 cases from a 0.35T MRI scanner (TRUFI) and 48 cases from a 1.5T MRI scanner (T2), was utilized to evaluate the effectiveness of the proposed method. The contours of 6 organs-at-risk were delineated by a senior radiation oncologist, serving as the ground truth. Three models, the source model (SRC) trained on the source domain, the target model (TGT) trained on the target domain, and the UDA model adapted from the source domain to the target domain, were compared on the target domain using the Dice Similarity Coefficient (DSC). In the experiment of 0.35T-to-1.5T, the proposed UDA method outperformed the source model, achieving an average DSC score of 0.82 ± 0.11, compared to 0.58 ± 0.23 (SRC) and 0.85 ± 0.09 (TGT), respectively. In the inverse experiment 1.5T-to-0.35T, the UDA model achieved an average DSC score of 0.79±0.13, compared to DSCs of 0.52 ± 0.25 and 0.81 ± 0.11 for the SRC and TGT respectively. The UDA method yielded a significant improvement of 46%, compared to the SRC. Particularly, OARs (organ at risk) with higher deformability such as the stomach and duodenum achieved a 58% and 63% improvement in performance, respectively. This work presents a compelling approach of UDA for auto-segmentation on multi-source MRIs. Experimental results demonstrate that the UDA effectively improve the segmentation performance of the source model in the target domain, resulting in a more robust segmentation model.

Patient-Adaptive Automated Segmentation in Daily kVCT Images for Radiotherapy of Head and Neck and Prostate Cancer

The purpose of this study was to examine the use of transfer learning in deep learning-based auto-segmentation of daily kilovoltage computed tomography (kVCT) images for patient-specific adaptive radiotherapy. Using data from the first cohort of patients treated with the innovative BgRT system, the objective of this study was to evaluate the potential benefits of this approach in facilitating efficient and effective adaptive radiotherapy. For the head and neck (HaN) site and pelvic site, we first trained a deep convolutional segmentation network using a population dataset, consisting of 67 and 56 patient cases, respectively. This population network was then fine-tuned for a specific patient using a transfer learning approach, adapting the network weights. The auto-segmentation network utilized in this study was a 23-layer U-Net with batch normalization, a dropout rate of 0.5, and four skip connections between the encoder and decoder at different levels. We used initial planning CT and 5-26 sets of daily kVCT scans with a total of 8,039 images for patient-specific learning in the 6 HaN cases and 4 pelvic cases, particularly analyzing the relationship between the number of sequential patient-specific training data and the performance of the auto-segmentation. We compared the performance of the patient-specific network with the population network and the clinical rigid registration method, using the Dice similarity coefficient (DSC) as the evaluation metric. Additionally, we investigated the corresponding dosimetric impacts of the different auto-segmentation and registration methods. The patient-specific network showed improved mean DSC scores of 0.88 and 0.90 for three HaN organs at risk (OARs) and eight pelvic targets and OARs, respectively, compared to the population network (0.70 and 0.63) and the registration method (0.72 and 0.72). The DSC of the patient-specific network steadily improved as the number of longitudinal training cases increased, reaching near saturation after 6 training cases. The use of the patient-specific auto-segmentation resulted in a reduction of the mean discrepancy in target and OAR doses between delivery and planning from 5.5% with the clinical rigid registration to 1.1%. The use of patient-specific transfer learning in auto-segmenting kVCT images showed higher accuracy compared to a conventional population network and clinical registration-based method. This approach holds promise for enhancing dose evaluation accuracy in adaptive radiotherapy.

Deep learning-based segmentation of dental implants on cone-beam computed tomography images: A validation study

ObjectivesTo train and validate a cloud-based convolutional neural network (CNN) model for automated segmentation (AS) of dental implant and attached prosthetic crown on cone-beam computed tomography (CBCT) images. MethodsA total dataset of 280 maxillomandibular jawbone CBCT scans was acquired from patients who underwent implant placement with or without coronal restoration. The dataset was randomly divided into three subsets: training set (n = 225), validation set (n = 25) and testing set (n = 30). A CNN model was developed and trained using expert-based semi-automated segmentation (SS) of the implant and attached prosthetic crown as the ground truth. The performance of AS was assessed by comparing with SS and manually corrected automated segmentation referred to as refined-automated segmentation (R-AS). Evaluation metrics included timing, voxel-wise comparison based on confusion matrix and 3D surface differences. ResultsThe average time required for AS was 60 times faster (<30 s) than the SS approach. The CNN model was highly effective in segmenting dental implants both with and without coronal restoration, achieving a high dice similarity coefficient score of 0.92±0.02 and 0.91±0.03, respectively. Moreover, the root mean square deviation values were also found to be low (implant only: 0.08±0.09 mm, implant+restoration: 0.11±0.07 mm) when compared with R-AS, implying high AI segmentation accuracy. ConclusionsThe proposed cloud-based deep learning tool demonstrated high performance and time-efficient segmentation of implants on CBCT images. Clinical significanceAI-based segmentation of implants and prosthetic crowns can minimize the negative impact of artifacts and enhance the generalizability of creating dental virtual models. Furthermore, incorporating the suggested tool into existing CNN models specialized for segmenting anatomical structures can improve pre-surgical planning for implants and post-operative assessment of peri‑implant bone levels.

Prostate cancer of magnetic resonance imaging automatic segmentation and detection of based on 3D-Mask RCNN

BackgroundProstate cancer is a widespread form of cancer that impacts men across the world. MRI plays a pivotal role in the detection and precise localization of cancerous regions, aiding medical professionals in devising effective treatment strategies for patients. As a result, MRI is often used in the diagnosis of prostate cancer. PurposeOur proposed method employs deep learning to achieve automatic segmentation and detection of prostate cancer in MRI single series, particularly T2-weighted imaging (T2WI). Materials and methodsOur study utilized data from 133 patients at a hospital, consisting of 71 cases of prostate cancer and 62 cases of benign prostatic tumors. We employed T2-weighted imaging (T2WI) MRI single series from 93 prostates as the training set for our 3D-Mask RCNN model, while the remaining data from 40 prostates were used for validation. The masks were manually delineated by an experienced radiologist, with pathology serving as the reference standard. Our approach was evaluated using several metrics, such as dice similarity coefficient (DSC), accuracy, sensitivity, specificity, and receiver operating characteristic (ROC) curve analysis. ResultsOur study produced promising results using the 3D Mask R-CNN model. The training set yielded a DSC score of 0.856, sensitivity of 0.921, and specificity of 0.961. The test set was also successful, with a DSC score of 0.849, sensitivity of 0.911, and specificity of 0.931. Furthermore, our model achieved an AUC value of 0.865 and an accuracy, sensitivity, and specificity of 0.866, 0.875, and 0.835, respectively, for the training set. The test set had an AUC value of 0.842 and an accuracy, sensitivity, and specificity of 0.836, 0.847, and 0.819, respectively. These findings demonstrate that our approach is capable of accurately detecting and segmenting prostate cancer in MRI single series - T2WI. ConclusionThe use of the 3D-Mask RCNN model in segmenting prostate tumors and detecting cancer in MRI T2WI has been shown to be highly effective and precise. This approach has the potential to greatly benefit radiologists by improving the accuracy and efficiency of diagnoses, leading to more effective treatment planning for patients with prostate cancer. By automating the segmentation process, this approach can also reduce the workload of radiologists and increase the consistency of diagnoses. The high performance of this model highlights the potential of deep learning techniques in medical imaging and demonstrates the significant impact that these approaches can have on improving patient outcomes.

Deep learning to detect macular atrophy in wet age-related macular degeneration using optical coherence tomography

Here, we have developed a deep learning method to fully automatically detect and quantify six main clinically relevant atrophic features associated with macular atrophy (MA) using optical coherence tomography (OCT) analysis of patients with wet age-related macular degeneration (AMD). The development of MA in patients with AMD results in irreversible blindness, and there is currently no effective method of early diagnosis of this condition, despite the recent development of unique treatments. Using OCT dataset of a total of 2211 B-scans from 45 volumetric scans of 8 patients, a convolutional neural network using one-against-all strategy was trained to present all six atrophic features followed by a validation to evaluate the performance of the models. The model predictive performance has achieved a mean dice similarity coefficient score of 0.706 ± 0.039, a mean Precision score of 0.834 ± 0.048, and a mean Sensitivity score of 0.615 ± 0.051. These results show the unique potential of using artificially intelligence-aided methods for early detection and identification of the progression of MA in wet AMD, which can further support and assist clinical decisions.

Automatic segmentation and quantification of the optic nerve on MRI using a 3D U-Net.

Pathological conditions associated with the optic nerve (ON) can cause structural changes in the nerve. Quantifying these changes could provide further understanding of disease mechanisms. We aim to develop a framework that automatically segments the ON separately from its surrounding cerebrospinal fluid (CSF) on magnetic resonance imaging (MRI) and quantifies the diameter and cross-sectional area along the entire length of the nerve. Multicenter data were obtained from retinoblastoma referral centers, providing a heterogeneous dataset of 40 high-resolution 3D T2-weighted MRI scans with manual ground truth delineations of both ONs. A 3D U-Net was used for ON segmentation, and performance was assessed in a tenfold cross-validation () and on a separate test-set () by measuring spatial, volumetric, and distance agreement with manual ground truths. Segmentations were used to quantify diameter and cross-sectional area along the length of the ON, using centerline extraction of tubular 3D surface models. Absolute agreement between automated and manual measurements was assessed by the intraclass correlation coefficient (ICC). The segmentation network achieved high performance, with a mean Dice similarity coefficient score of 0.84, median Hausdorff distance of 0.64mm, and ICC of 0.95 on the test-set. The quantification method obtained acceptable correspondence to manual reference measurements with mean ICC values of 0.76 for the diameter and 0.71 for the cross-sectional area. Compared with other methods, our method precisely identifies the ON from surrounding CSF and accurately estimates its diameter along the nerve's centerline. Our automated framework provides an objective method for ON assessment in vivo.

Assessment of Variabilities in Lung-Contouring Methods on CBCT Preclinical Radiomics Outputs.

Radiomics image analysis has the potential to uncover disease characteristics for the development of predictive signatures and personalised radiotherapy treatment. Inter-observer and inter-software delineation variabilities are known to have downstream effects on radiomics features, reducing the reliability of the analysis. The purpose of this study was to investigate the impact of these variabilities on radiomics outputs from preclinical cone-beam computed tomography (CBCT) scans. Inter-observer variabilities were assessed using manual and semi-automated contours of mouse lungs (n = 16). Inter-software variabilities were determined between two tools (3D Slicer and ITK-SNAP). The contours were compared using Dice similarity coefficient (DSC) scores and the 95th percentile of the Hausdorff distance (HD95p) metrics. The good reliability of the radiomics outputs was defined using intraclass correlation coefficients (ICC) and their 95% confidence intervals. The median DSC scores were high (0.82-0.94), and the HD95p metrics were within the submillimetre range for all comparisons. the shape and NGTDM features were impacted the most. Manual contours had the most reliable features (73%), followed by semi-automated (66%) and inter-software (51%) variabilities. From a total of 842 features, 314 robust features overlapped across all contouring methodologies. In addition, our results have a 70% overlap with features identified from clinical inter-observer studies.

A dual channel and spatial attention network for automatic spine segmentation of MRI images

AbstractAccurate image segmentation plays an essential role in diagnosing and treating various spinal diseases. However, traditional segmentation methods often consume a lot of time and energy. This research proposes an innovative deep‐learning‐based automatic segmentation method for spine magnetic resonance imaging (MRI) images. The proposed method DAUNet++ is supported by UNet++, which adds residual structure and attention mechanism. Specifically, a residual block is utilized for down‐sampling to construct the RVNet, as a new skeleton structure. Furthermore, two novel types of dual channel and spatial attention modules are proposed to emphasize rich feature regions, enhance useful information, and improve the network performance by recalibrating the characteristic. The published spinesagt2wdataset3 spinal MRI image dataset is adopted in the experiment. The dice similarity coefficient score on the test set is 0.9064. Higher segmentation accuracy and efficiency are achieved, indicating the effectiveness of the proposed method.

Deep Machine Learning Algorithms in Glioblastoma (GBM) MRI evaluation (PL4.005)

<h3>Objective:</h3> To test the reliability of automatic segmentation models on post-operative MRIs in glioblastoma evaluation. <h3>Background:</h3> The extent of tumor resection in glioblastoma is an independent prognostic factor associated with clinical outcomes. Because radiological assessment to identify residual tumoral components can be time-consuming and subject to inter-rater variability, the implementation of artificial intelligence (AI) may improve radiological assessment. Hereby, we evaluate performance of different automatic brain tumor segmentation algorithms trained on pre-operative MRI datasets, on post-operative MRI scans of patients with glioblastoma. <h3>Design/Methods:</h3> We evaluated the performance of ten different deep-learning models designed for automated brain tumor segmentation. Models were trained using a publicly available dataset from the Brain Tumor Segmentation (BRaTS) Challenge 2020 and was tested in our institutional MRI scans including pre-operative and post-operative MRI examinations. Semiautomatic segmentation was performed in the institutional scans and was used as the ground truth. Dice Similarity Coefficient (DSC) was used to compare each model’s performance against the ground truth. <h3>Results:</h3> A total of 329 MR examinations were used for training purposes and a separate 100 scans were used to test the deep-learning algorithms. Overall, models demonstrated better performance on pre-operative scans. Among post-operative MRI studies, Knowledge Distillation algorithm exhibited the best performance with a DSC score of 0.836 for whole tumor segmentation. When evaluating specific regions within the tumor, Knowledge Distillation demonstrated a DSC of 0.82 for the FLAIR hyperintense region, 3D Dilated Multi-Fiber Network achieved a DSC of 0.88 in the contrast-enhancing tumoral region, and 3D U-Net exhibited the highest DCS for the non-enhancing core (0.84). <h3>Conclusions:</h3> Deep machine learning has the potential to quantify enhancing and FLAIR volumes in brain MRI from glioblastoma patients with high accuracy. Current automatic segmentation models can perform automatic segmentation on post-operative MRI scans with an 88% accuracy. Further investigations are needed to improve this promising AI technology. <b>Disclosure:</b> An immediate family member of Dr. Zhu has stock in Abbott lab. The institution of Dr. Zhu has received research support from NRG NIH. The institution of Dr. Zhu has received research support from Novocure, Inc. The institution of Dr. Zhu has received research support from Five Prime pharmaceutical. The institution of Dr. Zhu has received research support from Denovo, inc. The institution of Dr. Zhu has received research support from Boston Biomedical Sumitomo Dainippon Pharma Global Oncology. The institution of Dr. Zhu has received research support from CNS pharmaceutical. Mr. HSIEH has nothing to disclose. Ms. Kabir has nothing to disclose. Dr. Nunez has nothing to disclose. Carlos Rodriguez Quintero has nothing to disclose. Dr. Cai has nothing to disclose. Dr. Hsu has nothing to disclose. Dr. Arevalo has nothing to disclose. Kangyi Zhao has nothing to disclose. Jackie Jiaqi Zhang has nothing to disclose. Roy Riascos-Castaneda has nothing to disclose. The institution of Dr. Jiang has received research support from NIH. The institution of Dr. Jiang has received research support from CPRIT. The institution of Dr. Jiang has received research support from NSF. shayan shams has nothing to disclose.

Abstract LB067: A confident and operator-independent deep segmentation model to measure residual tumor volume in the follow-up MRIs for glioblastoma

Abstract Introduction: Magnetic resonance imaging (MRI) is the most common tool to examine glioblastoma. Preoperative MRI can be used to initial diagnosis, and surgery planning, while follow-up MRIs can be used to evaluate treatment responses, identify recurrency, and detect side effects. The follow-up MRIs are usually taken after the first-line therapy, such as maximal safe resection. In the past, radiologists manually segment the tumor regions from normal brain tissue on follow-up MRIs, which is time-consuming, error-prone, and challenging. Several deep-learning (DL) models have been developed utilizing preoperative images, but their performance has yet to be evaluated on follow-up MRIs. In this research, we built the largest follow-up MRI cohort (311 patients) to assess these DL models and their generalizability and performance on independent preoperative and follow-up images. We also made the first follow-up-based deep learning models for this specific task. Methods: All evaluation deep learning models (10 models) were trained by the Brain Tumor Segmentation challenge 2020 (BraTS’20) and evaluated by fifty pairs of preoperative and follow-up scans from our institution. The segmentation form our institution is evaluated by board certified radiologist. MRIs in the BraTS’20 dataset were all preoperative scans. After the evaluation, we randomly assigned 264 patients’ scans from our institution to the training dataset and 47 patients’ scans to the testing dataset. We compared three types of models in our follow-up deep learning model, including 1) UNet-3D, 2) UNet-3D+transfer-learning, and 3) UNet-3D+transfer-learning+baysian-learning. The benchmark for all models was the Dice similarity coefficient (DSC). DSC can measure the spatial overlap between model prediction and ground truth. The value of DSC is between zero to one. Zero means no overlap and one indicates complete overlap. Results: Our study demonstrates that the BraTS'20 trained models' performance decreased by 13.05% in independent preoperative MRI scans and 19.04% in follow-up MRI scans. The most significant mismatch regions were FLAIR hyperintense regions (3.68% drop in independent preoperative scans and 10.61% drop in independent follow-up scans) and Non-enhancing core (5.20% drop in independent preoperative scans and 11.99% drop in independent follow-up scans). Our best model can achieve the best DSC among three tumor regions compared to all evaluation models (FLAIR hyperintense regions: DSC 0.77 V.S. DSC 0.58; Enhancing tumor region: DSC 0.87 V.S. DSC 0.68; Non-enhancing tumor region: DSC 0.92 V.S. DSC 0.55). Conclusion: Maximal safe resection induced brain structure change, decreasing the performance of the preoperative-based DL model. Implementing a follow-up MRI-based segmentation model is essential to make accurate and generalizable results to address structural changes after maximal safe resection. Our follow-up DL model demonstrates the DSC score can be recovered. We commit to further developing the tool to assist radiologists in handling follow-up MRIs for glioblastoma patients. Citation Format: Kang Lin Hsieh, Tanjida Kabir, Luis Nunez-Rubiano, Yu-Chun Hsu, Yu Cai, Juan Rodriguez Quintero, Octavio Arevalo, Kangyi Zhao, Jackie Zhang, Jiguang Zhu Zhu, Roy Riascos, Xioaqian Jiang, Shayan Shams. A confident and operator-independent deep segmentation model to measure residual tumor volume in the follow-up MRIs for glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB067.

Hepatic vessels segmentation using deep learning and preprocessing enhancement.

Liver hepatic vessels segmentation is a crucial step for the diagnosis process in patients with hepatic diseases. Segmentation of liver vessels helps to study the liver internal segmental anatomy that helps in the preoperative planning of surgical treatment. Recently, the convolutional neural networks (CNN) have been proved to be efficient for the task of medical image segmentation. The paper proposes an automatic deep learning-based system for liver hepatic vessels segmentation of Computed Tomography (CT) datasets from different sources. The proposed work focuses on the combination of different steps; it starts by a preprocessing step to improve the vessels appearance within the liver region of interest in the CT scans. Coherence enhancing diffusion filtering (CED) and vesselness filtering methods are used to improve vessels contrast and intensity homogeneity. The proposed U-net based network architecture is implemented with modified residual block to include concatenation skip connection. The effect of enhancement using filtering step was studied. Also, the effect of data mismatch used in training and validation is studied. The proposed method is evaluated using many CT datasets. Dice similarity coefficient (DSC) is used to evaluate the method. The average DSC score achieved a score 79%. The proposed approach succeeded to segment liver vasculature from the liver envelope accurately, which makes it as potential tool for clinical preoperative planning.

Quantification of liver-Lung shunt fraction on 3D SPECT/CT images for selective internal radiation therapy of liver cancer using CNN-based segmentations and non-rigid registration

Purpose: Selective internal radiation therapy (SIRT) has been proven to be an effective treatment for hepatocellular carcinoma (HCC) patients. In clinical practice, the treatment planning for SIRT using 90Y microspheres requires estimation of the liver-lung shunt fraction (LSF) to avoid radiation pneumonitis. Currently, the manual segmentation method to draw a region of interest (ROI) of the liver and lung in 2D planar imaging of 99mTc-MAA and 3D SPECT/CT images is inconvenient, time-consuming and observer-dependent. In this study, we propose and evaluate a nearly automatic method for LSF quantification using 3D SPECT/CT images, offering improved performance compared with the current manual segmentation method.Methods: We retrospectively acquired 3D SPECT with non-contrast-enhanced CT images (nCECT) of 60 HCC patients from a SPECT/CT scanning machine, along with the corresponding diagnostic contrast-enhanced CT images (CECT). Our approach for LSF quantification is to use CNN-based methods for liver and lung segmentations in the nCECT image. We first apply 3D ResUnet to coarsely segment the liver. If the liver segmentation contains a large error, we dilate the coarse liver segmentation into the liver mask as a ROI in the nCECT image. Subsequently, non-rigid registration is applied to deform the liver in the CECT image to fit that obtained in the nCECT image. The final liver segmentation is obtained by segmenting the liver in the deformed CECT image using nnU-Net. In addition, the lung segmentations are obtained using 2D ResUnet. Finally, LSF quantitation is performed based on the number of counts in the SPECT image inside the segmentations.Evaluations and Results: To evaluate the liver segmentation accuracy, we used Dice similarity coefficient (DSC), asymmetric surface distance (ASSD), and max surface distance (MSD) and compared the proposed method to five well-known CNN-based methods for liver segmentation. Furthermore, the LSF error obtained by the proposed method was compared to a state-of-the-art method, modified Deepmedic, and the LSF quantifications obtained by manual segmentation. The results show that the proposed method achieved a DSC score for the liver segmentation that is comparable to other state-of-the-art methods, with an average of 0.93, and the highest consistency in segmentation accuracy, yielding a standard deviation of the DSC score of 0.01. The proposed method also obtains the lowest ASSD and MSD scores on average (2.6 mm and 31.5 mm, respectively). Moreover, for the proposed method, a median LSF error of 0.14% is obtained, which is a statically significant improvement to the state-of-the-art-method (p=0.004), and is much smaller than the median error in LSF manual determination by the medical experts using 2D planar image (1.74% and p<0.001).Conclusions: A method for LSF quantification using 3D SPECT/CT images based on CNNs and non-rigid registration was proposed, evaluated and compared to state-of-the-art techniques. The proposed method can quantitatively determine the LSF with high accuracy and has the potential to be applied in clinical practice.

Index lesion contouring on prostate MRI for targeted MRI/US fusion biopsy – Evaluation of mismatch between radiologists and urologists

PurposeMistargeting of focal lesions due to inaccurate segmentations can lead to false-negative findings on MRI-guided targeted biopsies. The purpose of this retrospective study was to examine inter-reader agreement of prostate index lesion segmentations from actual biopsy data between urologists and radiologists. MethodConsecutive patients undergoing transperineal MRI-targeted prostate biopsy for PI-RADS 3–5 lesions between January 2020 and December 2021 were included. Agreement between segmentations on T2w-images between urologists and radiologists was assessed with Dice similarity coefficient (DSC) and 95 % Hausdorff distance (95 % HD). Differences in similarity scores were compared using Wilcoxon test. Differences depending on lesion features (size, zonal location, PI-RADS scores, lesion distinctness) were tested with Mann-Whitney U test. Correlation with prostate signal-intensity homogeneity score (PSHS) and lesion size was tested with Spearman’s rank correlation. ResultsNinety-three patients (mean age 64.9 ± 7.1y, median serum PSA 6.5 [4.33–10.00]) were included. Mean similarity scores were statistically significantly lower between urologists and radiologists compared to radiologists only (DSC 0.41 ± 0.24 vs. 0.59 ± 0.23, p < 0.01; 95 %HD 6.38 ± 5.45 mm vs. 4.47 ± 4.12 mm, p < 0.01). There was a moderate and strong positive correlation between DSC scores and lesion size for segmentations from urologists and radiologists (ρ = 0.331, p = 0.002) and radiologists only (ρ = 0.501, p < 0.001). Similarity scores were worse in lesions ≤ 10 mm while other lesion features did not significantly influence similarity scores. ConclusionThere is significant mismatch of prostate index lesion segmentations between urologists and radiologists. Segmentation agreement positively correlates with lesion size. PI-RADS scores, zonal location, lesion distinctness, and PSHS show no significant impact on segmentation agreement. These findings could underpin benefits of perilesional biopsies.

Magnetic resonance imaging organ at risk delineation for nasopharyngeal radiotherapy: Measuring the effectiveness of an educational intervention.

Magnetic resonance imaging (MRI) demonstrates superior soft tissue contrast and is increasingly being used in radiotherapy planning. This study evaluated the impact of an education workshop in minimising inter-observer variation (IOV) for nasopharyngeal organs at risk (OAR) delineation on MRI. Ten observers delineated 14 OARs on 4 retrospective nasopharyngeal MRI data sets. Standard contouring guidelines were provided pre-workshop. Following an education workshop on MRI OAR delineation, observers blinded to their original contours repeated the 14 OAR delineations. For comparison, reference volumes were delineated by two head and neck radiation oncologists. IOV was evaluated using dice similarity coefficient (DSC), Hausdorff distance (HD) and relative volume. Location of largest deviations was evaluated with centroid values. Observer confidence pre- and post-workshop was also recorded using a 6-point Likert scale. The workshop was deemed beneficial for an OAR if ≥50% of observers mean scores improved in any metric and ≥50% of observers' confidence improved. All OARs had ≥50% of observers improve in at least one metric. Base of tongue, larynx, spinal cord and right temporal lobe were the only OARs achieving a mean DSC score of ≥0.7. Base of tongue, left and right lacrimal glands, larynx, left optic nerve and right parotid gland all exhibited statistically significant HD improvements post-workshop (P < 0.05). Brainstem and left and right temporal lobes all had statistically significant relative volume improvements post-workshop (P < 0.05). Post-workshop observer confidence improvement was observed for all OARs (P < 0.001). The educational workshop reduced IOV and improved observers' confidence when delineating nasopharyngeal OARs on MRI.

Pixel-Level Classification of Five Histologic Patterns of Lung Adenocarcinoma.

Lung adenocarcinoma is the most common histologic type of lung cancer. The pixel-level labeling of histologic patterns of lung adenocarcinoma can assist pathologists in determining tumor grading with more details than normal classification. We manually annotated a dataset containing a total of 1000 patches (200 patches for each pattern) of 512 × 512 pixels and 420 patches (contains test sets) of 1024 × 1024 pixels according to the morphological features of the five histologic patterns of lung adenocarcinoma (lepidic, acinar, papillary, micropapillary, and solid). To generate an even large amount of data patches, we developed a data stitching strategy as a data augmentation for classification in model training. Stitched patches improve the Dice similarity coefficient (DSC) scores by 24.06% on the whole-slide image (WSI) with the solid pattern. We propose a WSI analysis framework for lung adenocarcinoma pathology, intelligently labeling lung adenocarcinoma histologic patterns at the pixel level. Our framework contains five branches of deep neural networks for segmenting each histologic pattern. We test our framework with 200 unclassified patches. The DSC scores of our results outpace comparing networks (U-Net, LinkNet, and FPN) by up to 10.78%. We also perform results on four WSIs with an overall accuracy of 99.6%, demonstrating that our network framework exhibits better accuracy and robustness in most cases.

Interpretable machine learning for automated left ventricular scar quantification in hypertrophic cardiomyopathy patients.

Scar quantification on cardiovascular magnetic resonance (CMR) late gadolinium enhancement (LGE) images is important in risk stratifying patients with hypertrophic cardiomyopathy (HCM) due to the importance of scar burden in predicting clinical outcomes. We aimed to develop a machine learning (ML) model that contours left ventricular (LV) endo- and epicardial borders and quantifies CMR LGE images from HCM patients.We retrospectively studied 2557 unprocessed images from 307 HCM patients followed at the University Health Network (Canada) and Tufts Medical Center (USA). LGE images were manually segmented by two experts using two different software packages. Using 6SD LGE intensity cutoff as the gold standard, a 2-dimensional convolutional neural network (CNN) was trained on 80% and tested on the remaining 20% of the data. Model performance was evaluated using the Dice Similarity Coefficient (DSC), Bland-Altman, and Pearson's correlation. The 6SD model DSC scores were good to excellent at 0.91 ± 0.04, 0.83 ± 0.03, and 0.64 ± 0.09 for the LV endocardium, epicardium, and scar segmentation, respectively. The bias and limits of agreement for the percentage of LGE to LV mass were low (-0.53 ± 2.71%), and correlation high (r = 0.92). This fully automated interpretable ML algorithm allows rapid and accurate scar quantification from CMR LGE images. This program does not require manual image pre-processing, and was trained with multiple experts and software, increasing its generalizability.