Segmentation Of Cone Beam Computed Tomography Images Research Articles

Evaluation of children's oral diagnosis and treatment using imaging examination using AI based Internet of Things.

Although cone beam computed tomography (CBCT) plays an important role in the diagnosis and treatment of oral diseases, its image segmentation method needs to be further improved, and there are still objections about the clinical application effect of general anesthesia (GA) on children's dental fear (CDF). This study aimed to investigate the application value of CBCT based on intelligent computer segmentation model in oral diagnosis and treatment of children in the context of biomedical signals, and to analyze the alleviating effect of GA on CDF. Based on the regional level set (CV) algorithm, the local binary fitting (LBF) model was introduced to optimize it, and the tooth CBCT image segmentation model CV-LBF was established to compare the segmentation accuracy (SA), maximum symmetric surface distance (MSSD), average symmetric surface distance (ASSD), over segmentation rate (OR), and under segmentation rate (UR) between these model and other algorithms. 82 children with CDF were divided into general anesthesia group (GAG) (n= 38) and controls (n= 44) according to the voluntary principle of their families. Children in GAG were treated with GA and controls with protective fixed intervention. Children's fear survey schedule-dental subscale (CFSS-DS) and Venham scores were counted before intervention in the two groups. CFSS-DS scores were recorded at 2 hours after intervention and after recovery in children in GAG. CFSS-DS and Venham scores were performed in all children 1 week after surgery. The results showed that the S⁢A value of CV-LBF algorithm was higher than that of region growing algorithm (P< 0.05). OR, UR, MSSD, and ASSD values of CV-LBF algorithm were evidently lower than those of other algorithms (P< 0.05). CFSS-DS scores were lower in GAG than in controls 2 hours after intervention and at return visits after 1 week of intervention (P< 0.001), and Venham scores were lower in GAG than in controls after intervention (P< 0.001). After intervention, the proportion of children with Venham grade 0, 1, 2, and 3 was obviously higher in GAG than in controls (P< 0.001), while the proportion of children with Venham grade 4 and 5 was clearly higher in controls than in GAG (P< 0.001). The results revealed that the computer intelligent segmentation model CV-LBF has potential application value in CBCT image segmentation of children's teeth, and GA can effectively alleviate anxiety of children with CDF and can be used as biomedical signals.

Evaluation of the accuracy of cone-beam computed tomography image segmentation of isolated tooth roots based on the dynamic threshold method

ObjectiveAccurate quantification of the root surface area (RSA) plays a decisive role in the advancement of periodontal, orthodontic, and restorative treatment modalities. In this study, we aimed to develop a dynamic threshold-based computer-aided system for segmentation and calculation of the RSA of isolated teeth on cone-beam computed tomography (CBCT) and to assess the accuracy of the measured data.MethodWe selected 24 teeth to be extracted, including single-rooted and multi-rooted teeth, from 22 patients who required tooth extraction. In the experimental group, we scanned 24 isolated teeth using CBCT with a voxel size of 0.3 mm. We designed a computer-aided system based on a personalized dynamic threshold algorithm to automatically segment the roots of 24 isolated teeth in CBCT images and calculate the RSA. In the control group, we employed digital intraoral scanner devices to perform optical scanning on 24 isolated teeth and subsequently manually segmented the roots using 3-matic software to calculate the RSA. We used the paired t-test (P < 0.05) and Bland-Altman plots to analyze the consistency of the two measurement methods.ResultsThe results of the paired t-test showed that there was no significant difference in the RSAs obtained using the dynamic threshold method and the optical scanning image reconstruction (t = 1.005, P = 0.325 > 0.05). As per the Bland-Altman plot, the results were evenly distributed within the region of ± 1.96 standard deviations of the mean, with no increasing or decreasing trends and good consistency.ConclusionIn this study, we designed a computer-aided root segmentation system based on a personalized dynamic threshold algorithm to automatically segment the roots of isolated teeth in CBCT images with a voxel size of 0.3 mm. We found that the RSA calculated using this approach was highly accurate, and a voxel of 0.3 mm in size could accurately display the surface area data in CBCT images. Overall, our findings in this study provide a foundation for future work on accurate automatic segmentation of tooth roots in full-mouth CBCT images and the computation of RSA.

Automated Segmentation of Dental CBCT Image Using an Improved U-Net Network

In the field of clinical dental medicine, Cone Beam Computed Tomography (CBCT) is a useful tool for the measurement of various dimensions related to the oral cavity, including height and thickness. This provides invaluable guidance and reference for risk assessment in orthodontic treatment, selection of treatment plans and implant treatment. However, segmentation of the teeth region from CBCT images is a daunting task due to complex root morphology and indistinct boundaries between the root and the alveolar bone. Manual annotation of the teeth area is resource-intensive, and deep learning-based segmentation methods are susceptible to noise, reducing their efficiency. To tackle these complexities, a multi-filter attention module is proposed in this paper, which effectively minimizes the noise in CBCT images through utilization of multiple filters and self-attention techniques. Additionally, an Improved U-Net model is proposed, where the original convolution block in the U-Net is replaced with a Double ConvNeXt block to yield better network performance. Experimentally, the proposed Improved U-Net method showed remarkable progress as it achieved a Dice Similarity Coefficient of 86.95% in oral CBCT image segmentation, surpassing existing models and affirming the effectiveness and advancedness of the proposed model and method.

Does the presence of full-coverage crown interfere with removal of obturation material: A cone-beam computed tomography image segmentation volumetric analysis.

The aim of this study was to assess and compare the influence of the presence of full-coverage crown on the obturation material removal efficiency of different retreatment file systems. Sixty extracted, human, maxillary premolars were accessed, instrumented, and obturated with gutta-percha and AH Plus sealer. The samples were scanned with cone-beam computed tomography and volume of obturating material was measured using ITK-SNAP volumetric software. Samples were divided (n = 30) into with and without all metal full-coverage crown groups. Each group was then subdivided into three subgroups (n = 10) based on retreatment file used (ProTaper URS, HyFlex Remover, and NeoEndo). The volume of remnants of obturating material was remeasured by image segmentation using ITK-SNAP volumetric software. One-way analysis of variance, followed by the Tukey's post hoc test, was used for data analysis, with a significance level of 5%. A significant difference in percentage remnant of obturating material after retreatment file was observed on overall comparison between with (29.84%) and without crown (21.79%) groups (P < 0.05). Each file system removed significantly more obturating material in without crown samples than with crown samples (P < 0.05). The percentage remnant of obturating material after retreatment file use was as follows: ProTaper URS > NeoEndo retreatment file > HyFlex Remover. All file systems showed significantly different (P < 0.05) efficiency in removal of obturating material. None of the retreatment file systems were able to remove obturating material completely. The presence of full-coverage crown negatively impacted the efficiency of retreatment files for the removal of obturating material. HyFlex Remover performed better than NeoEndo which was better than ProTaper URS.

Clinical feasibility of deep learning-based automatic head CBCT image segmentation and landmark detection in computer-aided surgical simulation for orthognathic surgery

The purpose of this ambispective study was to investigate whether deep learning-based automatic segmentation and landmark detection, the SkullEngine, could be used for orthognathic surgical planning. Sixty-one sets of cone beam computed tomography (CBCT) images were automatically inferred for midface, mandible, upper and lower teeth, and 68 landmarks. The experimental group included automatic segmentation and landmarks, while the control group included manual ones that were previously used to plan orthognathic surgery. The qualitative analysis of segmentation showed that all of the automatic results could be used for computer-aided surgical simulation. Among these, 98.4% of midface, 70.5% of mandible, 98.4% of upper teeth, and 93.4% of lower teeth could be directly used without manual revision. The Dice similarity coefficient was 96% and the average symmetric surface distance was 0.1 mm for all four structures. With SkullEngine, it took 4 minutes to complete the automatic segmentation and an additional 10 minutes for a manual touchup. The results also showed the overall mean difference between the two groups was 2.3 mm for the midface and 2.4 mm for the mandible. In summary, the authors believe that automatic segmentation using SkullEngine is ready for daily practice. However, the accuracy of automatic landmark digitization needs to be improved.

Quantitative and qualitative comparisons of pulp cavity volumes produced by cone beam computed tomography and micro-computed tomography through semiautomatic segmentation: An ex vivo investigation

The aim of this study was to measure the volume and visually assess 3-dimensional (3D) virtual models of pulp cavities obtained through semiautomatic segmentation on images from 6 cone beam computed tomography (CBCT) units compared with the reference standard of micro-CT. Fifteen mandibular premolar teeth were scanned with 6 CBCT units: Prexion 3D Elite, i-CAT Next Generation, NewTom 5G, Cranex 3D, 3Shape X1, and Orthophos SL 3D, using the smallest available field of view and highest resolution settings. Pulp cavity volumes were quantitatively assessed by 2 calibrated examiners. The volumes from each CBCT unit were compared with micro-CT. Qualitative assessment of the 3D reconstructions was also performed. Repeated-measures analysis of variance and the Friedman test compared the CBCT reconstructions to micro-CT. Intra- and interexaminer agreements were calculated with the intraclass correlation coefficient and kappa statistic. The CBCT-based volumes were all significantly larger than micro-CT (P ≤ .0061). Prexion, X1, and Orthophos provided the segmentations that most closely resembled the reference standard. Intra- and interexaminer agreements ranged from good to excellent for quantitative measurements. Interexaminer agreement for qualitative evaluation was substantial. Semiautomatic segmentation of CBCT images is a feasible method to produce virtual 3D models of the pulp cavity. Prexion, X1, and Orthophos were the CBCT units that resulted in 3D reconstructions most similar to the reference standard.

Convolutional Neural Network Performance for Sella Turcica Segmentation and Classification Using CBCT Images.

The present study aims to validate the diagnostic performance and evaluate the reliability of an artificial intelligence system based on the convolutional neural network method for the morphological classification of sella turcica in CBCT (cone-beam computed tomography) images. In this retrospective study, sella segmentation and classification models (CranioCatch, Eskisehir, Türkiye) were applied to sagittal slices of CBCT images, using PyTorch supported by U-Net and TensorFlow 1, and we implemented the GoogleNet Inception V3 algorithm. The AI models achieved successful results for sella turcica segmentation of CBCT images based on the deep learning models. The sensitivity, precision, and F-measure values were 1.0, 1.0, and 1.0, respectively, for segmentation of sella turcica in sagittal slices of CBCT images. The sensitivity, precision, accuracy, and F1-score were 1.0, 0.95, 0.98, and 0.84, respectively, for sella-turcica-flattened classification; 0.95, 0.83, 0.92, and 0.88, respectively, for sella-turcica-oval classification; 0.75, 0.94, 0.90, and 0.83, respectively, for sella-turcica-round classification. It is predicted that detecting anatomical landmarks with orthodontic importance, such as the sella point, with artificial intelligence algorithms will save time for orthodontists and facilitate diagnosis.

Automatic Segmentation of Mandibular Ramus and Condyles.

In order to diagnose TMJ pathologies, we developed and tested a novel algorithm, MandSeg, that combines image processing and machine learning approaches for automatically segmenting the mandibular condyles and ramus. A deep neural network based on the U-Net architecture was trained for this task, using 109 cone-beam computed tomography (CBCT) scans. The ground truth label maps were manually segmented by clinicians. The U-Net takes 2D slices extracted from the 3D volumetric images. All the 3D scans were cropped depending on their size in order to keep only the mandibular region of interest. The same anatomic cropping region was used for every scan in the dataset. The scans were acquired at different centers with different resolutions. Therefore, we resized all scans to 512×512 in the pre-processing step where we also performed contrast adjustment as the original scans had low contrast. After the pre-processing, around 350 slices were extracted from each scan, and used to train the U-Net model. For the cross-validation, the dataset was divided into 10 folds. The training was performed with 60 epochs, a batch size of 8 and a learning rate of 2×10-5. The average performance of the models on the test set presented 0.95 ± 0.05 AUC, 0.93 ± 0.06 sensitivity, 0.9998 ± 0.0001 specificity, 0.9996 ± 0.0003 accuracy, and 0.91 ± 0.03 F1 score. This study findings suggest that fast and efficient CBCT image segmentation of the mandibular condyles and ramus from different clinical data sets and centers can be analyzed effectively. Future studies can now extract radiomic and imaging features as potentially relevant objective diagnostic criteria for TMJ pathologies, such as osteoarthritis (OA). The proposed segmentation will allow large datasets to be analyzed more efficiently for disease classification.

Deep Learning-Based Image Segmentation of Cone-Beam Computed Tomography Images for Oral Lesion Detection.

This paper aimed to study the adoption of deep learning (DL) algorithm of oral lesions for segmentation of cone-beam computed tomography (CBCT) images. 90 patients with oral lesions were taken as research subjects, and they were grouped into blank, control, and experimental groups, whose images were treated by the manual segmentation method, threshold segmentation algorithm, and full convolutional neural network (FCNN) DL algorithm, respectively. Then, effects of different methods on oral lesion CBCT image recognition and segmentation were analyzed. The results showed that there was no substantial difference in the number of patients with different types of oral lesions among three groups (P > 0.05). The accuracy of lesion segmentation in the experimental group was as high as 98.3%, while those of the blank group and control group were 78.4% and 62.1%, respectively. The accuracy of segmentation of CBCT images in the blank group and control group was considerably inferior to the experimental group (P < 0.05). The segmentation effect on the lesion and the lesion model in the experimental group and control group was evidently superior to the blank group (P < 0.05). In short, the image segmentation accuracy of the FCNN DL method was better than the traditional manual segmentation and threshold segmentation algorithms. Applying the DL segmentation algorithm to CBCT images of oral lesions can accurately identify and segment the lesions.

Deep learning in cone-beam computed tomography image segmentation for the diagnosis and treatment of acute pulpitis

Deep learning in cone-beam computed tomography image segmentation for the diagnosis and treatment of acute pulpitis

Comparison of convolutional neural network training strategies for cone-beam CT image segmentation

Background and objectiveOver the past decade, convolutional neural networks (CNNs) have revolutionized the field of medical image segmentation. Prompted by the developments in computational resources and the availability of large datasets, a wide variety of different two-dimensional (2D) and three-dimensional (3D) CNN training strategies have been proposed. However, a systematic comparison of the impact of these strategies on the image segmentation performance is still lacking. Therefore, this study aimed to compare eight different CNN training strategies, namely 2D (axial, sagittal and coronal slices), 2.5D (3 and 5 adjacent slices), majority voting, randomly oriented 2D cross-sections and 3D patches. MethodsThese eight strategies were used to train a U-Net and an MS-D network for the segmentation of simulated cone-beam computed tomography (CBCT) images comprising randomly-placed non-overlapping cylinders and experimental CBCT images of anthropomorphic phantom heads. The resulting segmentation performances were quantitatively compared by calculating Dice similarity coefficients. In addition, all segmented and gold standard experimental CBCT images were converted into virtual 3D models and compared using orientation-based surface comparisons. ResultsThe CNN training strategy that generally resulted in the best performances on both simulated and experimental CBCT images was majority voting. When employing 2D training strategies, the segmentation performance can be optimized by training on image slices that are perpendicular to the predominant orientation of the anatomical structure of interest. Such spatial features should be taken into account when choosing or developing novel CNN training strategies for medical image segmentation. ConclusionsThe results of this study will help clinicians and engineers to choose the most-suited CNN training strategy for CBCT image segmentation.

Multiclass CBCT Image Segmentation for Orthodontics with Deep Learning

Accurate segmentation of the jaw (i.e., mandible and maxilla) and the teeth in cone beam computed tomography (CBCT) scans is essential for orthodontic diagnosis and treatment planning. Although various (semi)automated methods have been proposed to segment the jaw or the teeth, there is still a lack of fully automated segmentation methods that can simultaneously segment both anatomic structures in CBCT scans (i.e., multiclass segmentation). In this study, we aimed to train and validate a mixed-scale dense (MS-D) convolutional neural network for multiclass segmentation of the jaw, the teeth, and the background in CBCT scans. Thirty CBCT scans were obtained from patients who had undergone orthodontic treatment. Gold standard segmentation labels were manually created by 4 dentists. As a benchmark, we also evaluated MS-D networks that segmented the jaw or the teeth (i.e., binary segmentation). All segmented CBCT scans were converted to virtual 3-dimensional (3D) models. The segmentation performance of all trained MS-D networks was assessed by the Dice similarity coefficient and surface deviation. The CBCT scans segmented by the MS-D network demonstrated a large overlap with the gold standard segmentations (Dice similarity coefficient: 0.934 ± 0.019, jaw; 0.945 ± 0.021, teeth). The MS-D network–based 3D models of the jaw and the teeth showed minor surface deviations when compared with the corresponding gold standard 3D models (0.390 ± 0.093 mm, jaw; 0.204 ± 0.061 mm, teeth). The MS-D network took approximately 25 s to segment 1 CBCT scan, whereas manual segmentation took about 5 h. This study showed that multiclass segmentation of jaw and teeth was accurate and its performance was comparable to binary segmentation. The MS-D network trained for multiclass segmentation would therefore make patient-specific orthodontic treatment more feasible by strongly reducing the time required to segment multiple anatomic structures in CBCT scans.

3D morphometric quantification of maxillae and defects for patients with unilateral cleft palate via deep learning-based CBCT image auto-segmentation.

This study aimed to quantify the 3D asymmetry of the maxilla in patients with unilateral cleft lip and palate (UCP) and investigate the defect factors responsible for the variability of the maxilla on the cleft side using a deep-learning-based CBCT image segmentation protocol. Cone beam computed tomography (CBCT) images of 60 patients with UCP were acquired. The samples in this study consisted of 39 males and 21 females, with a mean age of 11.52years (SD=3.27years; range of 8-18years). The deep-learning-based protocol was used to segment the maxilla and defect initially, followed by manual refinement. Paired t-tests were performed to characterize the maxillary asymmetry. A multiple linear regression was carried out to investigate the relationship between the defect parameters and those of the cleft side of the maxilla. The cleft side of the maxilla demonstrated a significant decrease in maxillary volume and length as well as alveolar length, anterior width, posterior width, anterior height and posterior height. A significant increase in maxillary anterior width was demonstrated on the cleft side of the maxilla. There was a close relationship between the defect parameters and those of the cleft side of the maxilla. Based on the 3D volumetric segmentations, significant hypoplasia of the maxilla on the cleft side existed in the pyriform aperture and alveolar crest area near the defect. The defect structures appeared to contribute to the variability of the maxilla on the cleft side.

A User-Friendly Protocol for Mandibular Segmentation of CBCT Images for Superimposition and Internal Structure Analysis.

Background: Since cone-beam computed tomography (CBCT) technology has been widely adopted in orthodontics, multiple attempts have been made to devise techniques for mandibular segmentation and 3D superimposition. Unfortunately, as the software utilized in these methods are not specifically designed for orthodontics, complex procedures are often necessary to analyze each case. Thus, this study aimed to establish an orthodontist-friendly protocol for segmenting the mandible from CBCT images that maintains access to the internal anatomic structures. Methods: The “sculpting tool” in the Dolphin 3D Imaging software was used for segmentation. The segmented mandible images were saved as STL files for volume matching in the 3D Slicer to validate the repeatability of the current protocol and were exported as DICOM files for internal structure analysis and voxel-based superimposition. Results: The mandibles of all tested CBCT datasets were successfully segmented. The volume matching analysis showed high consistency between two independent segmentations for each mandible. The intraclass correlation coefficient (ICC) analysis on 20 additional CBCT mandibular segmentations further demonstrated the high consistency of the current protocol. Moreover, all of the anatomical structures for superimposition identified by the American Board of Orthodontics were found in the voxel-based superimposition, demonstrating the ability to conduct precise internal structure analyses with the segmented images. Conclusion: An efficient and precise protocol to segment the mandible while retaining access to the internal structures was developed on the basis of CBCT images.

Age estimation based on 3D pulp chamber segmentation of first molars from cone-beam-computed tomography by integrated deep learning and level set.

To develop an automatic segmentation method to segment the pulp chamber of first molars from 3D cone-beam-computed tomography (CBCT) images, and to estimate ages by calculated pulp volumes. Patients with CBCT scans were retrospectively identified. The age estimation was formulated as CBCT image segmentation using a coarse-to-fine strategy by integrated deep learning (DL) and level set (LS), followed by establishing a linear regression model. On the training data, DL model was trained for coarse segmentation. The validation set was to determine the optimal DL model, and a LS method established on it was to refine the coarse segmentation. On the testing data, the integrated DL and LS method was applied for pulp chamber segmentation, followed by volume calculation and age estimation. Statistical analysis was performed by Wilcoxon rank sum test to demonstrate gender difference in pulp chamber volume, and volume difference between maxillary and mandibular molars. Wilcoxon signed-rank test was adopted to compare true and estimated ages. A total of 180 CBCT studies were randomly divided into 37/10/133 patients for training, validation, and testing data, respectively. In the training and validation sets, the results showed high spatial overlaps between manual and automatic segmentation (dice = 87.8%). For the testing set, the estimated human ages were not significantly different with true human age (p= 0.57), with a correlation coefficient r= 0.74. An integrated DL and LS method was able to segment pulp chamber of first molars from 3D CBCT images, and the derived pulp chamber volumes could effectively estimate the human ages.

Anatomical and Volumetric Analysis of the Sphenoid Sinus by Semiautomatic Segmentation of Cone Beam Computed Tomography.

This investigation was performed to analyze and evaluate anatomical and volumetric data of the sphenoid sinus by semiautomatic segmentation of cone beam computed tomography (CBCT) datasets.This analysis was carried out on cone beam computed tomography images of 100 patients. By using the discretionary software "ITK-Snap", the volumetric investigation of 197 sphenoid sinuses was conducted. The statistical analyses were carried out by using the paired and independent Student t-test. In addition, the Pearsons chi-square test was performed. P values P < 0.05 were considered significant.The volumetric mean was 4.438 ± 2.434 cm3 for the right and 4.809 ± 3.000 for the left side. Results did not indicate statistically significant differences according to the side, irrespective of gender (P > 0.05). The further investigation of differences between male and female patients did assess significant results (P < 0.05), males (10.477 ± 3.851 cm3) present a larger total sinus volume than females (8.219 ± 3.574 cm3). In accordance to patients' age (median 54 years), smaller volumes were related to younger patients and larger volumes to older ones. But, no statistically significant results in accordance to age could be detected (χ2 = 1.258; P > 0.05).Endoscopic sinus surgery of the vulnerable and intricate anatomy of the sphenoid sinus requires an explicit knowledge and orientation about the anatomical variations such as septums and volumetric capacity. Concerning this matter, semiautomatic segmentation of CBCT images can aid the surgeon in preoperative planning. Our investigation shows that a small sinus volume, that is, affecting female patients and a complex sinus anatomy by the occurrence of a septum, especially located close to neurovascular structures could essentially aggravate endoscopic sphenoid sinus surgery. In this regard, volumetric analysis of the sphenoid sinus by CBCT images can provide presurgical information and facilitates an individualized treatment.

Volumetric analysis of MRONJ lesions by semiautomatic segmentation of CBCT images.

The purpose of this study was to evaluate potential differences in volumes of areas of osteolysis caused by medication-related osteonecrosis of the jaw (MRONJ) between the upper and lower jaw. We aim to analyze the clinical relevance of volumetric measurement of osteolytic lesions for surgical planning of MRONJ patients. Sixty-seven patients who were clinically and histopathologically diagnosed with MRONJ were retrospectively included in this study. Cone beam computed tomography (CBCT) images were evaluated according to localization, affected anatomical structures, and volumetric measurement of osteolytic lesions caused by MRONJ in appliance of CBCT datasets by using ITK-SNAP. The most frequently affected localization is the mandible, whereas female patients show significantly more often lesions of the maxilla. The cortical bone was predominantly affected. Furthermore, the affection of teeth, sinus floor, inferior alveolar nerve canal, or even a pathological fracture of the mandible are infrequently existing. The volumetric measurements revealed a statistically significant greater absolute osteolysis volume in males. Image analysis and volumetric measurements of osteolytic lesions of MRONJ patients is a helpful tool to further understand the clinical appearance and identify compromised anatomic landmarks. Volumetric analysis aids in pre-surgical planning and visualizes the individual extent of the disease for each patient.

Machine learning in orthodontics: Introducing a 3D auto-segmentation and auto-landmark finder of CBCT images to assess maxillary constriction in unilateral impacted canine patients.

To (1) introduce a novel machine learning method and (2) assess maxillary structure variation in unilateral canine impaction for advancing clinically viable information. A machine learning algorithm utilizing Learning-based multi-source IntegratioN frameworK for Segmentation (LINKS) was used with cone-beam computed tomography (CBCT) images to quantify volumetric skeletal maxilla discrepancies of 30 study group (SG) patients with unilaterally impacted maxillary canines and 30 healthy control group (CG) subjects. Fully automatic segmentation was implemented for maxilla isolation, and maxillary volumetric and linear measurements were performed. Analysis of variance was used for statistical evaluation. Maxillary structure was successfully auto-segmented, with an average dice ratio of 0.80 for three-dimensional image segmentations and a minimal mean difference of two voxels on the midsagittal plane for digitized landmarks between the manually identified and the machine learning-based (LINKS) methods. No significant difference in bone volume was found between impaction ([2.37 ± 0.34] [Formula: see text] 104 mm3) and nonimpaction ([2.36 ± 0.35] [Formula: see text] 104 mm3) sides of SG. The SG maxillae had significantly smaller volumes, widths, heights, and depths (P < .05) than CG. The data suggest that palatal expansion could be beneficial for those with unilateral canine impaction, as underdevelopment of the maxilla often accompanies that condition in the early teen years. Fast and efficient CBCT image segmentation will allow large clinical data sets to be analyzed effectively.

Automated segmentation of CBCT image using spiral CT atlases and convex optimization.

Cone-beam computed tomography (CBCT) is an increasingly utilized imaging modality for the diagnosis and treatment planning of the patients with craniomaxillofacial (CMF) deformities. CBCT scans have relatively low cost and low radiation dose in comparison to conventional spiral CT scans. However, a major limitation of CBCT scans is the widespread image artifacts such as noise, beam hardening and inhomogeneity, causing great difficulties for accurate segmentation of bony structures from soft tissues, as well as separating mandible from maxilla. In this paper, we presented a novel fully automated method for CBCT image segmentation. In this method, we first estimated a patient-specific atlas using a sparse label fusion strategy from predefined spiral CT atlases. This patient-specific atlas was then integrated into a convex segmentation framework based on maximum a posteriori probability for accurate segmentation. Finally, the performance of our method was validated via comparisons with manual ground-truth segmentations.

TH‐D‐M100F‐05: Online Region‐Of‐Interest Delineation of Daily Head and Neck Images

Purpose: Online imaging modalities, such as cone beam computed tomography (CBCT) or CT on‐rail provide online volumetric images. A fast, automatic and robust region‐of‐interest (ROI) delineation method is highly desired in image guided radiation therapy (IGRT). We have developed such a method and tested it via segmentation of head and neck (HN) fan beam CT and CBCT images. Material and Methods: ROIs on planning CT images were manually delineated using commercial treatment planning system. A variational‐based deformable image registration algorithm was implemented to register planning CT images to daily CT images. ROIs on planning CT images were automatically mapped to daily images using voxel matching information between planning and daily image datasets. The results were quantitatively and qualitatively validated by comparing to manual delineation. In order to accelerate computing speed, we paralleled the algorithm using message passing interface (MPI) on a Beowulf cluster with 16 processing elements (PE). Speed improvement was benchmarked. Results: The discrepancies between automatically and manually delineated ROIs on fan beam images were mostly within 2mm. Automatic segmentation of CBCT images was acceptable by visual inspection. Benchmark results showed that paralleling efficiencies were above 95% and speedup factors were approximately equal to the number of PE used. With 16 PEs online delineation of HN images took about 1 minute. Conclusion: The online ROI delineation method we have developed is robust, fast and is suitable for HN online adaptive radiation treatments.This research was partially supported by the Department of Defense Prostate Cancer Research Program under award number W81XWH‐07‐0083.