Synthetic CT generation from CBCT images via deep learning

PURPOSE

Cone-beam computed tomography (CBCT) scanning is used daily or weekly (i.e., on-treatment CBCT) for accurate patient setup in image-guided radiotherapy. However, inaccuracy of CT numbers prevents CBCT from performing advanced tasks such as dose calculation and treatment planning. Motivated by the promising performance of deep learning in medical imaging, we propose a deep U-net-based approach that synthesizes CT-like images with accurate numbers from planning CT, while keeping the same anatomical structure as on-treatment CBCT.

METHODS

We formulated the CT synthesis problem under a deep learning framework, where a deep U-net architecture was used to take advantage of the anatomical structure of on-treatment CBCT and image intensity information of planning CT. U-net was chosen because it exploits both global and local features in the image spatial domain, matching our task to suppress global scattering artifacts and local artifacts such as noise in CBCT. To train the synthetic CT generation U-net (sCTU-net), we include on-treatment CBCT and initial planning CT of 37 patients (30 for training, seven for validation) as the input. Additional replanning CT images acquired on the same day as CBCT after deformable registration are utilized as the corresponding reference. To demonstrate the effectiveness of the proposed sCTU-net, we use another seven independent patient cases (560 slices) for testing.

RESULTS

We quantitatively compared the resulting synthetic CT (sCT) with the original CBCT image using deformed same-day pCT images as reference. The averaged accuracy measured by mean absolute error (MAE) between sCT and reference CT (rCT) on testing data is 18.98 HU, while MAE between CBCT and rCT is 44.38 HU.

CONCLUSIONS

The proposed sCTU-net can synthesize CT-quality images with accurate CT numbers from on-treatment CBCT and planning CT. This potentially enables advanced CBCT applications for adaptive treatment planning.

Image guidance in radiation therapy for better cure of cancer

The key goal and main challenge of radiation therapy is the elimination of tumors without any concurring damages of the surrounding healthy tissues and organs. Radiation doses required to achieve sufficient cancer-cell kill exceed in most clinical situations the dose that can be tolerated by the healthy tissues, especially when large parts of the affected organ are irradiated. High-precision radiation oncology aims at optimizing tumor coverage, while sparing normal tissues. Medical imaging during the preparation phase, as well as in the treatment room for localization of the tumor and directing the beam, referred to as image-guided radiotherapy (IGRT), is the cornerstone of precision radiation oncology. Sophisticated high-resolution real-time IGRT using X-rays, computer tomography, magnetic resonance imaging, or ultrasound, enables delivery of high radiation doses to tumors without significant damage of healthy organs. IGRT is the most convincing success story of radiation oncology over the last decades, and it remains a major driving force of innovation, contributing to the development of personalized oncology, for example, through the use of real-time imaging biomarkers for individualized dose delivery.

Cone-Beam Computed Tomography in Orthodontics

Unlike patients receiving implants or endodontic treatment, most orthodontic patients are children who are particularly sensitive to ionizing radiation. Cone-beam computed tomography (CBCT) carries risks and benefits in orthodontics. The principal risks and limitations include ionizing radiation, the presence of artifacts, higher cost, limited accessibility, and the need for additional training. However, this imaging modality has several recognized indications in orthodontics, such as the assessment of impacted and ectopic teeth, assessment of pharyngeal airway, assessment of mini-implant sites, evaluation of craniofacial abnormalities, evaluation of sinus anatomy or pathology, evaluation of root resorption, evaluation of the cortical bone plate, and orthognathic surgery planning and evaluation. CBCT is particularly justified when it brings a benefit to the patient or changes the outcome of the treatment when compared with conventional imaging techniques. Therefore, CBCT should be considered for clinical orthodontics for selected patients. Prescription of CBCT requires judicious and sound clinical judgment. The central question of this narrative review article is: when does CBCT add value to the practice of orthodontics? To answer this question, this article presents discussion on radiation dosage of CBCT and other imaging techniques used in orthodontics, limitations of CBCT in orthodontics, justifying the use of CBCT in orthodontics, and the benefits and evidence-based indications of CBCT in orthodontics. This review summarizes the central themes and topics in the literature regarding CBCT in orthodontics and presents ten orthodontic cases in which CBCT proved to be valuable.

Acuros CTS: A fast, linear Boltzmann transport equation solver for computed tomography scatter - Part I: Core algorithms and validation

PURPOSE

To describe Acuros^® CTS, a new software tool for rapidly and accurately estimating scatter in x-ray projection images by deterministically solving the linear Boltzmann transport equation (LBTE).

METHODS

The LBTE describes the behavior of particles as they interact with an object across spatial, energy, and directional (propagation) domains. Acuros CTS deterministically solves the LBTE by modeling photon transport associated with an x-ray projection in three main steps: (a) Ray tracing photons from the x-ray source into the object where they experience their first scattering event and form scattering sources. (b) Propagating photons from their first scattering sources across the object in all directions to form second scattering sources, then repeating this process until all high-order scattering sources are computed using the source iteration method. (c) Ray-tracing photons from scattering sources within the object to the detector, accounting for the detector's energy and anti-scatter grid responses. To make this process computationally tractable, a combination of analytical and discrete methods is applied. The three domains are discretized using the Linear Discontinuous Finite Elements, Multigroup, and Discrete Ordinates methods, respectively, which confer the ability to maintain the accuracy of a continuous solution. Furthermore, through the implementation in CUDA, we sought to exploit the parallel computing capabilities of graphics processing units (GPUs) to achieve the speeds required for clinical utilization. Acuros CTS was validated against Geant4 Monte Carlo simulations using two digital phantoms: (a) a water phantom containing lung, air, and bone inserts (WLAB phantom) and (b) a pelvis phantom derived from a clinical CT dataset. For these studies, we modeled the TrueBeam^® (Varian Medical Systems, Palo Alto, CA) kV imaging system with a source energy of 125 kVp. The imager comprised a 600 μm-thick Cesium Iodide (CsI) scintillator and a 10:1 one-dimensional anti-scatter grid. For the WLAB studies, the full-fan geometry without a bowtie filter was used (with and without the anti-scatter grid). For the pelvis phantom studies, a half-fan geometry with bowtie was used (with the anti-scatter grid). Scattered and primary photon fluences and energies deposited in the detector were recorded.

RESULTS

The Acuros CTS and Monte Carlo results demonstrated excellent agreement. For the WLAB studies, the average percent difference between the Monte Carlo- and Acuros-generated scattered photon fluences at the face of the detector was -0.7%. After including the detector response, the average percent differences between the Monte Carlo- and Acuros-generated scatter fractions (SF) were -0.1% without the grid and 0.6% with the grid. For the digital pelvis simulation, the Monte Carlo- and Acuros-generated SFs agreed to within 0.1% on average, despite the scatter-to-primary ratios (SPRs) being as high as 5.5. The Acuros CTS computation time for each scatter image was ~1 s using a single GPU.

CONCLUSIONS

Acuros CTS enables a fast and accurate calculation of scatter images by deterministically solving the LBTE thus offering a computationally attractive alternative to Monte Carlo methods. Part II describes the application of Acuros CTS to scatter correction of CBCT scans on the TrueBeam system.

Alveolar ridge preservation in the posterior maxilla reduces vertical dimensional change: A randomized controlled clinical trial

OBJECTIVES

To test whether or not alveolar ridge preservation reduces vertical changes in the posterior maxilla compared to spontaneous healing following tooth extraction.

MATERIALS AND METHODS

Forty subjects requiring extraction of maxillary posterior teeth with root apices protruding into the maxillary sinus floor were consecutively enrolled. Patients were randomly assigned to either one of two surgical interventions: an alveolar ridge preservation procedure using collagenated bovine bone mineral and a resorbable collagen membrane (test) or no grafting (control). Cone-beam computed tomographies were taken immediately and at 6 months after surgery, prior to dental implant placement.

RESULTS

Based on radiographic data, the level of the sinus floor remained stable over time (baseline to 6 months) in the test group (-0.14 mm [-0.31, -0.02]). In the control group, the sinus floor level shifted more coronally (-1.16 mm [-1.73, -0.61]) than the test group (p < 0.05). The test group demonstrated a significantly larger residual bone height than the control group at 6 months (7.30 mm [6.36, 8.20] vs. 4.83 mm [3.94, 5.76], respectively, p < 0.05). Implant placement without any additional sinus augmentation procedure was performed in 42.9% of test group cases, whereas in all of the subjects in the control group an additional augmentation procedure was needed (100% of the cases).

CONCLUSION

Alveolar ridge preservation in the posterior maxilla maintained the vertical bone height more efficiently and resulted in less need for sinus augmentation procedures at 6 months compared to spontaneous healing.

Age estimation based on pulp/tooth volume ratio measured on cone-beam CT images

OBJECTIVES

After tooth eruption, the size of the pulp cavity decreases with age owing to deposition of secondary dentine. The aim of this study was to investigate the relation between the chronological age and the ratio of pulp volume (PV) to tooth volume (TV) measurements using CBCT images of single rooted teeth.

METHODS

Maxillary anterior, canine and mandibular canine/premolar CBCT scans of patients older than 15 years of age were collected from the archives between 2013 and 2015 years. Patients with CBCT scans of teeth were seen in detail and patients with known chronological age were included. Teeth with caries, filling or crown restorations, periapical pathologies or pulps that could not be identified were excluded. Consequently, 204 patients with 655 teeth were evaluated. The PV and the TV of each tooth was measured and then the PV/TV ratio was calculated. Simple linear regression analysis was performed in order to predict age estimation by using PV/TV.

RESULTS

The PV/TV of all teeth ranged between 0.01 and 0.08. A negative correlation was found between the PV/TV ratio and age (p < 0.05). The regression analysis showed the highest Pearson correlation (0.532) for the maxillary central incisor tooth. Considering measurements of the PV/TV of all assessed teeth, there was no significant difference in the intercept between both gender (p > 0.05).

CONCLUSIONS

This study revealed that PV/TV ratio was not gender dependent and the strongest correlation was found between the age and PV/TV ratio measured on maxillary central incisors than other teeth.

Comparison of weekly and daily online adaptation for head and neck intensity-modulated proton therapy

The high conformality of intensity-modulated proton therapy (IMPT) dose distributions causes treatment plans to be sensitive to geometrical changes during the course of a fractionated treatment. This can be addressed using adaptive proton therapy (APT). One important question in APT is the frequency of adaptations performed during a fractionated treatment, which is related to the question whether plan adaptation has to be done online or offline. The purpose of this work is to investigate the impact of weekly and daily online IMPT plan adaptation on the treatment quality for head and neck patients. A cohort of ten head and neck patients with daily acquired cone-beam CT (CBCT) images was evaluated retrospectively. Dose tracking of the IMPT treatment was performed for three scenarios: base plan with no adaptation (BP), weekly online adaptation (OA_W), and daily online adaptation (OA_D). Both adaptation schemes used an in-house developed online APT workflow, performing Monte Carlo (MC) dose calculations on scatter-corrected CBCTs. IMPT plan adaptation was achieved by only tuning the weights of a subset of beamlets, based on deformable image registration from the planning CT to each CBCT. Although OA_Dmitigated random delivery errors more effectively than OA_Won a fraction per fraction basis, both OA_Wand OA_Dachieved the clinical goals for all ten patients, while BP failed for six cases. In the high-risk CTV, accumulated values of D_98%ranged between 97.15% and 99.73% of the prescription dose for OA_D, with a median of 98.07%. For OA_W, values between 95.02% and 99.26% were obtained, with a median of 97.61% of the prescription dose. Otherwise, the dose to most organs at risk was similar for all three scenarios. Globally, our results suggest that OA_Wcould be used as an alternative approach to OA_Dfor most patients in order to reduce the clinical workload.

Pelvic multi-organ segmentation on cone-beam CT for prostate adaptive radiotherapy

BACKGROUND AND PURPOSE

The purpose of this study is to develop a deep learning-based approach to simultaneously segment five pelvic organs including prostate, bladder, rectum, left and right femoral heads on cone-beam CT (CBCT), as required elements for prostate adaptive radiotherapy planning.

MATERIALS AND METHODS

We propose to utilize both CBCT and CBCT-based synthetic MRI (sMRI) for the segmentation of soft tissue and bony structures, as they provide complementary information for pelvic organ segmentation. CBCT images have superior bony structure contrast and sMRIs have superior soft tissue contrast. Prior to segmentation, sMRI was generated using a cycle-consistent adversarial networks (CycleGAN), which was trained using paired CBCT-MR images. To combine the advantages of both CBCT and sMRI, we developed a cross-modality attention pyramid network with late feature fusion. Our method processes CBCT and sMRI inputs separately to extract CBCT-specific and sMRI-specific features prior to combining them in a late-fusion network for final segmentation. The network was trained and tested using 100 patients' datasets, with each dataset including the CBCT and manual physician contours. For comparison, we trained another two networks with different network inputs and architectures. The segmentation results were compared to manual contours for evaluations.

RESULTS

For the proposed method, dice similarity coefficients and mean surface distances between the segmentation results and the ground truth were 0.96 ± 0.03, 0.65 ± 0.67 mm; 0.91 ± 0.08, 0.93 ± 0.96 mm; 0.93 ± 0.04, 0.72 ± 0.61 mm; 0.95 ± 0.05, 1.05 ± 1.40 mm; and 0.95 ± 0.05, 1.08 ± 1.48 mm for bladder, prostate, rectum, left and right femoral heads, respectively. As compared to the other two competing methods, our method has shown superior performance in terms of the segmentation accuracy.

CONCLUSION

We developed a deep learning-based segmentation method to rapidly and accurately segment five pelvic organs simultaneously from daily CBCTs. The proposed method could be used in the clinic to support rapid target and organs-at-risk contouring for prostate adaptive radiation therapy.

Ex vivo evaluation of new 2D and 3D dental radiographic technology for detecting caries

OBJECTIVES

Proximal dental caries remains a prevalent disease with only modest detection rates by current diagnostic systems. Many new systems are available without controlled validation of diagnostic efficacy. The objective of this study was to evaluate the diagnostic efficacy of three potentially promising new imaging systems.

METHODS

This study evaluated the caries detection efficacy of Schick 33 (Sirona Dental, Salzburg, Austria) intraoral digital detector images employing an advanced sharpening filter, Planmeca ProMax(®) (Planmeca Inc., Helsinki, Finland) extraoral "panoramic bitewing" images and Sirona Orthophos XG3D (Sirona Dental) CBCT images with advanced artefact reduction. Conventional photostimulable phosphor images served as the control modality. An ex vivo study design using extracted human teeth, ten expert observers and micro-CT ground truth was employed.

RESULTS

Receiver operating characteristic analysis indicated similar diagnostic efficacy of all systems (ANOVA p > 0.05). The sensitivity of the Schick 33 images (0.48) was significantly lower than the other modalities (0.53-0.62). The specificity of the Planmeca images (0.86) was significantly lower than Schick 33 (0.96) and XG3D (0.97). The XG3D showed significantly better cavitation detection sensitivity (0.62) than the other modalities (0.48-0.57).

CONCLUSIONS

The Schick 33 images demonstrated reduced caries sensitivity, whereas the Planmeca panoramic bitewing images demonstrated reduced specificity. XG3D with artefact reduction demonstrated elevated sensitivity and specificity for caries detection, improved depth accuracy and substantially improved cavitation detection. Care must be taken to recognize potential false-positive caries lesions with Planmeca panoramic bitewing images. Use of CBCT for caries detection must be carefully balanced with the presence of metal artefacts, time commitment, financial cost and radiation dose.

Evaluation of peri-implant buccal bone by computed tomography: an experimental study

OBJECTIVES

To evaluate the accuracy of measuring peri-implant buccal bone when using three different computed tomography devices.

MATERIALS AND METHODS

Sixty tissue-level or bone-level dental implants were placed in bovine ribs with either buccal bone full coverage, dehiscence or fenestration. For each site, the distance from the bone defect to the implant neck and the buccal bone thickness 1 mm apical to the crest were measured using a calliper. Subsequently, all sites were scanned in a reproducible position using a multi-slice computed tomography (CT) (Brightspeed, voxel size 0.625 mm) and two cone-beam computed tomography devices (i-CAT NG, voxel size 0.3 mm and Newtom VGi, voxel size 0.2 mm). Bone thickness was measured on images from the three systems similar to direct measurements and differences were evaluated. Factors that could influence the buccal bone identification were assessed by multiple binary logistic regression.

RESULTS

Buccal bone ranged from 0.1 mm to 2.75 mm in thickness and was not visible in 68%, 63% and 60% of cases when using CT, i-CAT and Newtom, respectively. For each mm of bone thickness increment, the odds of radiographic identification increased by 30.6 (P < 0.001). Bone defects negatively affected radiographic visibility (P < 0.05). All devices underestimated bone dimensions although differences among them were not significant.

CONCLUSIONS

Within these experimental conditions, the investigated devices have equivalent low accuracy in diagnosing peri-implant buccal bone. Accuracy was significantly influenced by buccal bone thickness, especially if <1 mm, and in presence of peri-implant marginal defects.

Bisphosphonate-Related Osteonecrosis of the Jaw Bone: Radiological Pattern and the Potential Role of CBCT in Early Diagnosis

Objectives

To systematize the clinico-radiological symptoms and course of bisphosphonate-related osteonecrosis of jaw bone and toevaluate the diagnostic potential of various radiological techniques to detect mild osteonecrosis in each stage of the disease.

Material and Methods

The sample consisted of 22 patients previously diagnosed with extraoral malignant disease. Diagnosis was based on a clinical examination in conjunction to digital panoramic radiography and cone beam computed tomography (CBCT). Two dentomaxillofacial radiologists reviewed all images.

Results

Twenty patients showed mandibular involvement clinically, while two others had a maxillary involvement. Four stages of the disease were proposed based on the clinico-radiological findings. Subclinical cortical and lamina dura thickening was detected with only three-dimensional CBCT and periapical images, while ulceration and cortical bone thickening was detected only by three-dimensional CBCT. Mixed sclerotic, lytic bone destruction involving alveolar and basal bone with or without encroachment on the mandibular canal, pathological mandibular fractures were detected by two-dimensional panoramic and three-dimensional CBCT images. Other findings are non healing extraction sockets, periapical radiolucencies, osteolysis, sequestra, oroantral fistula, and periosteal new bone formation.

Conclusions

The present study showed that bisphosphonate-related osteonecrosis of jaw bone occurs in four distinct clinico-radiological stages. For mild cases, panoramic image diagnosis was much less obvious, whereas cone beam computed tomography was able to fully characterise the bony lesions and describe their extent and involvement of neighbouring structures in all cases. Thus cone beam computed tomography might better contribute to the prevention of bisphosphonate-related osteonecrosis of jaw bone as well to the disease management.

A clinically relevant accuracy study of computer-planned implant placement in the edentulous maxilla using mucosa-supported surgical templates

PURPOSE

The purpose of the study is to determine the clinically relevant accuracy of implant placement in the edentulous maxilla using computer planning and a mucosa-supported surgical template.

MATERIALS AND METHODS

In each of in total 30 consecutive edentulous patients suffering from retention problems of their upper denture, two or four Brånemark MkIII Groovy (Nobel Biocare®, Zürich, Switzerland) implants in the maxilla were installed. Preoperatively, first, a cone-beam computer tomography (cone beam computer tomography) scan was acquired, followed by virtual implant planning. Hereafter, a surgical template was designed to allow flapless implant placement using the template as a guide. To inventory the accuracy of implant placement, a postoperative CBCT scan was obtained and matched to the preoperative scan. The accuracy of implant placement was validated three-dimensionally. The Implant Position Orthogonal Projection validation method was applied to measure the clinically relevant implant deviations (i.e., in both the bucco-lingual and mesio-distal plane). Also, the influence of type of surgery, use of fixation pins, and position on the dental arch were investigated with regard to implant deviations.

RESULTS

In total, 104 implants were installed. In bucco-lingual direction, a mean implant deviation of 0.67 mm was scored at the implant tip, of 0.51 mm at the shoulder, of -0.83 mm in depth, as also a mean deviation of angulation of 1.74°. In mesio-distal direction, a mean implant deviation of 0.75 mm was found at the implant tip, of 0.60 mm at the implant shoulder, of -0.75 mm in depth, and a deviation of angulation of 1.94°. Of all implants, 74% was placed not deep enough compared with the planning. Implant position on the dental arch, the use of fixation pins, and type of surgery showed no significant effect on implant deviations. However, a significant difference for implant deviations in both buccal and mesial direction was observed, explained by a nonoptimal positioning of the surgical template.

CONCLUSIONS

Computer-aided implant planning showed to be a clinically relevant tool for the placement of two or four implants in the maxilla of fully edentulous patients. Exact positioning of the surgical template in anterior/posterior direction is crucial in reducing implant deviations both in buccal and mesial direction.

Synthetic CT generation from CBCT images via unsupervised deep learning

Adaptive-radiation-therapy (ART) is applied to account for anatomical variations observed over the treatment course. Daily or weekly cone-beam computed tomography (CBCT) is commonly used in clinic for patient positioning, but CBCT's inaccuracy in Hounsfield units (HU) prevents its application to dose calculation and treatment planning. Adaptive re-planning can be performed by deformably registering planning CT (pCT) to CBCT. However, scattering artifacts and noise in CBCT decrease the accuracy of deformable registration and induce uncertainty in treatment plan. Hence, generating from CBCT a synthetic CT (sCT) that has the same anatomical structure as CBCT but accurate HU values is desirable for ART. We proposed an unsupervised style-transfer-based approach to generate sCT based on CBCT and pCT. Unsupervised learning was desired because exactly matched CBCT and CT are rarely available, even when they are taken a few minutes apart. In the proposed model, CBCT and pCT are two inputs that provide anatomical structure and accurate HU information, respectively. The training objective function is designed to simultaneously minimize (1) contextual loss between sCT and CBCT to maintain the content and structure of CBCT in sCT and (2) style loss between sCT and pCT to achieve pCT-like image quality in sCT. We used CBCT and pCT images of 114 patients to train and validate the designed model, and another 29 independent patient cases to test the model's effectiveness. We quantitatively compared the resulting sCT with the original CBCT using the deformed same-day pCT as reference. Structure-similarity-index, peak-signal-to-noise-ratio, and mean-absolute-error in HU of sCT were 0.9723, 33.68, and 28.52, respectively, while those of CBCT were 0.9182, 29.67, and 49.90, respectively. We have demonstrated the effectiveness of the proposed model in using CBCT and pCT to synthesize CT-quality images. This model may permit using CBCT for advanced applications such as adaptive treatment planning.

Three-dimensional evaluation of upper airway in patients with different anteroposterior skeletal patterns

OBJECTIVES

To investigate variability in the upper airway of subjects with different anteroposterior skeletal patterns by evaluating the volume and the most constricted cross-sectional area of the pharyngeal airway and defining correlations between the different variables.

MATERIAL AND METHODS

The study sample consisted of 60 patients (29 boys, 31 girls) divided into three groups: Class I (1 ≤ ANB ≤ 3), Class II (ANB>3), and Class III (ANB<1), to evaluate how the jaw relationship affects the airway volume and the most constricted cross-sectional area (Min-CSA). Differences between groups were determined using the Tukey-Kramer test. Correlations between variables were tested using Pearson's correlation coefficient.

RESULTS

The volume and the Min-CSA of the pharyngeal airway (PA) were significantly related to anteroposterior skeletal patterns (p < 0.05). The nasopharyngeal airway (NA) volume of Class I and Class III subjects was significantly larger than that of Class II subjects (p < 0.05). The Min-CSA and the length of PA were significantly related to the volume of PA (p < 0.05). The site and the size of the Min-CSA varied among the three groups.

CONCLUSIONS

The volume and the most constricted cross-sectional area of the airway varied with different anteroposterior skeletal patterns. The NA volume of Class I and Class III subjects was significantly larger than that of patients with a Class II skeletal pattern.

Evaluation of Deformable Image Registration-Based Contour Propagation From Planning CT to Cone-Beam CT

Purpose:

We evaluated the performance of organ contour propagation from a planning computed tomography to cone-beam computed tomography with deformable image registration by comparing contours to manual contouring.

Materials and Methods:

Sixteen patients were retrospectively identified based on showing considerable physical change throughout the course of treatment. Multiple organs in the 3 regions (head and neck, prostate, and pancreas) were evaluated. A cone-beam computed tomography from the end of treatment was registered to the planning computed tomography using rigid registration, followed by deformable image registration. The contours were copied on cone-beam computed tomography image sets using rigid registration and modified by 2 radiation oncologists. Contours were compared using Dice similarity coefficient, mean surface distance, and Hausdorff distance.

Results:

The mean physician-to-physician Dice similarity coefficient for all organs was 0.90. When compared to each physician’s contours, the overall mean for rigid was 0.76 (P < .001), and it was improved to 0.79 (P < .001) for deformable image registration. Comparing deformable image registration to physicians resulted in a mean Dice similarity coefficient of 0.77, 0.74, and 0.84 for head and neck, prostate, and pancreas groups, respectively; whereas, the physician-to-physician mean agreement for these sites was 0.87, 0.90, and 0.93 (P < .001, for all sites). The mean surface distance for physician-to-physician contours was 1.01 mm, compared to 2.58 mm for rigid-to-physician contours and 2.24 mm for deformable image registration-to-physician contours. The mean physician-to-physician Hausdorff distance was 11.32 mm, and when compared to any physician’s contours, the mean for rigid and deformable image registration was 12.1 mm and 12.0 mm (P < .001), respectively.

Conclusion:

The physicians had a high level of agreement via the 3 metrics; however, deformable image registration fell short of this level of agreement. The automatic workflows using deformable image registration to deform contours to cone-beam computed tomography to evaluate the changes during treatment should be used with caution.

Improved Diagnostic Accuracy of Ameloblastoma and Odontogenic Keratocyst on Cone-Beam CT by Artificial Intelligence

OBJECTIVE

The purpose of this study was to utilize a convolutional neural network (CNN) to make preoperative differential diagnoses between ameloblastoma (AME) and odontogenic keratocyst (OKC) on cone-beam CT (CBCT).

METHODS

The CBCT images of 178 AMEs and 172 OKCs were retrospectively retrieved from the Hospital of Stomatology, Wuhan University. The datasets were randomly split into a training dataset of 272 cases and a testing dataset of 78 cases. Slices comprising lesions were retained and then cropped to suitable patches for training. The Inception v3 deep learning algorithm was utilized, and its diagnostic performance was compared with that of oral and maxillofacial surgeons.

RESULTS

The sensitivity, specificity, accuracy, and F1 score were 87.2%, 82.1%, 84.6%, and 85.0%, respectively. Furthermore, the average scores of the same indexes for 7 senior oral and maxillofacial surgeons were 60.0%, 71.4%, 65.7%, and 63.6%, respectively, and those of 30 junior oral and maxillofacial surgeons were 63.9%, 53.2%, 58.5%, and 60.7%, respectively.

CONCLUSION

The deep learning model was able to differentiate these two lesions with better diagnostic accuracy than clinical surgeons. The results indicate that the CNN may provide assistance for clinical diagnosis, especially for inexperienced surgeons.

Initial Clinical Experience With Extremity Cone-Beam CT of the Foot and Ankle in Pediatric Patients

OBJECTIVE

Extremity cone-beam CT (CBCT) scanners have become available for clinical use in the United States. The purpose of this study was to review an initial clinical experience with CBCT of the foot and ankle in pediatric patients.

MATERIALS AND METHODS

A retrospective review was conducted of all foot or ankle CBCT examinations performed on patients 18 years old and younger at one institution from August 1, 2013, through February 28, 2015. A t test was used to compare mean effective dose for CBCT with that for MDCT foot or ankle examinations of age-matched control subjects. To assess changes in utilization, a t test also was used to compare the mean numbers of foot or ankle CT examinations per month before and after installation of the CBCT scanner at the institution.

RESULTS

Thirty-four CBCT examinations were performed. The mean effective dose was 0.013 ± 0.003 mSv compared with 0.023 ± 0.020 mSv for MDCT of age-matched control subjects (p < 0.005). The mean numbers of foot or ankle CT examinations per month were 3.4 in the 18 months before and 3.8 in the 18 months after installation of the CBCT scanner (p = 0.28). The mean number of foot or ankle MDCT examinations per month decreased significantly (3.4 vs 1.9, p = 0.03) over the same period. In 56% of patients, CBCT revealed important findings that were not visible on contemporaneous radiographs. In 68% of patients, the CBCT findings affected clinical management.

CONCLUSION

CBCT of the foot or ankle of pediatric patients is a viable lower-dose alternative to MDCT that provides important information that may affect clinical management.

Radiation dose and image quality comparison during spine surgery with two different, intraoperative 3D imaging navigation systems

Careful protocol selection is required during intraoperative three-dimensional (3D) imaging for spine surgery to manage patient radiation dose and achieve clinical image quality. Radiation dose and image quality of a Medtronic O-arm commonly used during spine surgery, and a Philips hybrid operating room equipped with XperCT C-arm 3D cone-beam CT (hCBCT) are compared. The mobile O-arm (mCBCT) offers three different radiation dose settings (low, standard, and high), for four different patient sizes (small, medium, large, and extra large). The patient's radiation dose rate is constant during the entire 3D scan. In contrast, C-CBCT spine imaging uses three different field of views (27, 37, and 48 cm) using automatic exposure control (AEC) that modulates the patient's radiation dose rate during the 3D scan based on changing patient thickness. hCBCT uses additional x-ray beam filtration. Small, medium, and large trunk phantoms designed to mimic spine and soft tissue were imaged to assess radiation dose and image quality of the two systems. The estimated measured "patient" dose for the small, medium, and large phantoms imaged by the mCBCT considering all the dose settings ranged from 9.4-27.6 mGy, 8.9-33.3 mGy, and 13.8-40.6 mGy, respectively. The "patient" dose values for the same phantoms imaged with hCBCT were 2.8-4.6 mGy, 5.7-10.0 mGy, and 11.0-15.2 mGy. The CNR for the small, medium, and large phantoms was 2.9 to 3.7, 2.0 to 3.0, and 2.5 to 2.6 times higher with the hCBCT system, respectively. Hounsfield unit accuracy, noise, and uniformity of hCBCT exceeded the performance of the mCBCT; spatial resolution was comparable. Added x-ray beam filtration and AEC capability achieved clinical image quality for intraoperative spine surgery at reduced radiation dose to the patient in comparison to a reference O-arm system without these capabilities.

Accuracy of surface-guided patient setup for conventional radiotherapy of brain and nasopharynx cancer

Purpose

To evaluate the accuracy of surface‐guided radiotherapy (SGRT) in cranial patient setup by direct comparison between optical surface imaging (OSI) and cone‐beam computed tomography (CBCT), before applying SGRT‐only setup for conventional radiotherapy of brain and nasopharynx cancer.

Methods and Materials

Using CBCT as reference, SGRT setup accuracy was examined based on 269 patients (415 treatments) treated with frameless cranial stereotactic radiosurgery (SRS) during 2018‐2019. Patients were immobilized in customized head molds and open‐face masks and monitored using OSI during treatment. The facial skin area in planning CT was used as OSI region of interest (ROI) for automatic surface alignment and the skull was used as the landmark for automatic CBCT/CT registration. A 6 degrees of freedom (6DOF) couch was used. Immediately after CBCT setup, an OSI verification image was captured, recording the SGRT setup differences. These differences were analyzed in 6DOFs and as a function of isocenter positions away from the anterior surface to assess OSI‐ROI bias. The SGRT in‐room setup time was estimated and compared with CBCT and orthogonal 2D kilovoltage (2DkV) setups.

Results

The SGRT setup difference (magnitude) is found to be 1.0 ± 2.5 mm and 0.1˚±1.4˚ on average among 415 treatments and within 5 mm/3˚ with greater than 95% confidence level (P < 0.001). Outliers were observed for very‐posterior isocenters: 15 differences (3.6%) are >5.0mm and 9 (2.2%) are >3.0˚. The setup differences show minor correlations (|r| < 0.45) between translational and rotational DOFs and a minor increasing trend (<1.0 mm) in the anterior‐to‐posterior direction. The SGRT setup time is 0.8 ± 0.3 min, much shorter than CBCT (5 ± 2 min) and 2DkV (2 ± 1 min) setups.

Conclusion

This study demonstrates that SGRT has sufficient accuracy for fast in‐room patient setup and allows real‐time motion monitoring for beam holding during treatment, potentially useful to guide radiotherapy of brain and nasopharynx cancer with standard fractionation.

Incorporating Cone-Beam CT Into the Diagnostic Algorithm for Suspected Radiocarpal Fractures: A New Standard of Care?

OBJECTIVE. The purpose of this study was to assess the result of adding cone-beam CT to the standard imaging algorithm for patients with suspected radiographically occult traumatic radiocarpal fractures. SUBJECTS AND METHODS. A prospective review was performed on all patients who had cone-beam CT investigation of acute wrist pain after normal initial radiographs. Patients with no identified fractures were clinically reassessed and referred for MRI if concern for a fracture persisted. RESULTS. In all, 117 patients were assessed; 50.4% had fractures identified with a total of 67 radiographically occult fractures. One fracture was identified on MRI that was not seen on cone-beam CT. Cone-beam CT had sensitivity of 98.3% (95% CI, 91.1-100%), specificity of 100% (95% CI, 93.7-100%), positive predictive value of 100%, and negative predictive value of 98.3% (95% CI, 89.1-100%). Accuracy was 99.1% (95% CI, 95.3-100%). CONCLUSION. Incorporating cone-beam CT into routine clinical practice as part of a standardized diagnostic algorithm yielded a 50% fracture detection rate in patients with negative wrist radiographs but ongoing clinical concern for radiocarpal fracture. Cone-beam CT provides more diagnostic information than radiographs at a lower radiation dose than conventional MDCT. Given the poor accuracy of radiographs for acute radiocarpal fractures and the high fracture prevalence in this cohort, we feel that cone-beam CT should be regarded as the new standard of care in the investigation of these patients.

Deep-learning-based image registration and automatic segmentation of organs-at-risk in cone-beam CT scans from high-dose radiation treatment of pancreatic cancer

PURPOSE

Accurate deformable registration between computed tomography (CT) and cone-beam CT (CBCT) images of pancreatic cancer patients treated with high biologically effective radiation doses is essential to assess changes in organ-at-risk (OAR) locations and shapes and to compute delivered dose. This study describes the development and evaluation of a deep-learning (DL) registration model to predict OAR segmentations on the CBCT derived from segmentations on the planning CT.

METHODS

The DL model is trained with CT-CBCT image pairs of the same patient, on which OAR segmentations of the small bowel, stomach, and duodenum have been manually drawn. A transformation map is obtained, which serves to warp the CT image and segmentations. In addition to a regularity loss and an image similarity loss, an OAR segmentation similarity loss is also used during training, which penalizes the mismatch between warped CT segmentations and manually drawn CBCT segmentations. At test time, CBCT segmentations are not required as they are instead obtained from the warped CT segmentations. In an IRB-approved retrospective study, a dataset consisting of 40 patients, each with one planning CT and two CBCT scans, was used in a fivefold cross-validation to train and evaluate the model, using physician-drawn segmentations as reference. Images were preprocessed to remove gas pockets. Network performance was compared to two intensity-based deformable registration algorithms (large deformation diffeomorphic metric mapping [LDDMM] and multimodality free-form [MMFF]) as baseline. Evaluated metrics were Dice similarity coefficient (DSC), change in OAR volume within a volume of interest (enclosing the low-dose PTV plus 1 cm margin) from planning CT to CBCT, and maximum dose to 5 cm3 of the OAR [D(5cc)].

RESULTS

Processing time for one CT-CBCT registration with the DL model at test time was less than 5 seconds on a GPU-based system, compared to an average of 30 minutes for LDDMM optimization. For both small bowel and stomach/duodenum, the DL model yielded larger median DSC and smaller interquartile variation than either MMFF (paired t-test P < 10-4 for both type of OARs) or LDDMM (P < 10-3 and P = 0.03 respectively). Root-mean-square deviation (RMSD) of DL-predicted change in small bowel volume relative to reference was 22% less than for MMFF (P = 0.007). RMSD of DL-predicted stomach/duodenum volume change was 28% less than for LDDMM (P = 0.0001). RMSD of DL-predicted D(5cc) in small bowel was 39% less than for MMFF (P = 0.001); in stomach/duodenum, RMSD of DL-predicted D(5cc) was 18% less than for LDDMM (P < 10-3 ).

CONCLUSIONS

The proposed deep network CT-to-CBCT deformable registration model shows improved segmentation accuracy compared to intensity-based algorithms and achieves an order-of-magnitude reduction in processing time.

Age and sex estimation based on pulp cavity volume using cone beam computed tomography: development and validation of formulas in a Brazilian sample

OBJECTIVES

To develop and validate formulas for age and sex estimation based on the pulp cavity volume of teeth using cone beam CT.

METHODS

The sample was composed of 116 cone beam CT scans from Brazilian individuals of both sexes, ranging in age from 13 to 70 years. A total of 232 teeth (upper central incisors and canines) were evaluated. Two calibrated examiners determined pulp cavity volumes using the ITK-SNAP software. Pearson's correlation test was used to assess the correlation between chronological age and pulp volume. Linear and logistic regression models were developed for age and sex estimation, respectively, and were validated in another sample of 72 teeth.

RESULTS

Pearson's correlation coefficients between age and pulp volume were negative and significant (p < 0.0001) for both teeth (r = -0.8782 for central incisors and r = -0.8738 for canines). The age estimation formulas showed good determination coefficients (adjusted R² = 0.7614 to 0.8367). For sex estimation, when the age was known, the coefficients were also good (adjusted R² = 0.649 to 0.812). However, when the age was unknown, the coefficients of the sex estimation formulas were low (adjusted R² = 0.047 to 0.393). Validation showed high accuracy of age estimation in individuals older than 35 years, as well as high accuracy of sex estimation when the age was known.

CONCLUSIONS

Our formulas provided excellent results and can be applied to the Brazilian population. The best results were observed for age estimation in females and for sex estimation when the age was known.

Image quality and dose characteristics for an O-arm intraoperative imaging system with model-based image reconstruction

PURPOSE

To assess the imaging performance and radiation dose characteristics of the O-arm CBCT imaging system (Medtronic Inc., Littleton MA) and demonstrate the potential for improved image quality and reduced dose via model-based image reconstruction (MBIR).

METHODS

Two main studies were performed to investigate previously unreported characteristics of the O-arm system. First is an investigation of dose and 3D image quality achieved with filtered back-projection (FBP) - including enhancements in geometric calibration, handling of lateral truncation and detector saturation, and incorporation of an isotropic apodization filter. Second is implementation of an MBIR algorithm based on Huber-penalized likelihood estimation (PLH) and investigation of image quality improvement at reduced dose. Each study involved measurements in quantitative phantoms as a basis for analysis of contrast-to-noise ratio and spatial resolution as well as imaging of a human cadaver to test the findings under realistic imaging conditions.

RESULTS

View-dependent calibration of system geometry improved the accuracy of reconstruction as quantified by the full-width at half maximum of the point-spread function - from 0.80 to 0.65 mm - and yielded subtle but perceptible improvement in high-contrast detail of bone (e.g., temporal bone). Standard technique protocols for the head and body imparted absorbed dose of 16 and 18 mGy, respectively. For low-to-medium contrast (<100 HU) imaging at fixed spatial resolution (1.3 mm edge-spread function) and fixed dose (6.7 mGy), PLH improved CNR over FBP by +48% in the head and +35% in the body. Evaluation at different dose levels demonstrated 30% increase in CNR at 62% of the dose in the head and 90% increase in CNR at 50% dose in the body.

CONCLUSIONS

A variety of improvements in FBP implementation (geometric calibration, truncation and saturation effects, and isotropic apodization) offer the potential for improved image quality and reduced radiation dose on the O-arm system. Further gains are possible with MBIR, including improved soft-tissue visualization, low-dose imaging protocols, and extension to methods that naturally incorporate prior information of patient anatomy and/or surgical instrumentation.

Assessment of the appearance, location and morphology of mandibular lingual foramina using cone beam computed tomography

OBJECTIVES

To investigate the appearance, location and morphology of mandibular lingual foramina (MLF) in the Chinese Han population using cone beam computed tomography (CBCT).

METHODS

CBCT images of the mandibular body in 200 patients (103 female patients and 97 male patients, age range 10-70 years) were retrospectively analysed to identify MLF. The canal number, location and direction were assessed. Additionally, the diameter of the lingual foramen, the distance between the alveolar crest and the lingual foramen, the distance between the tooth apex and the lingual foramen and the distance from the mandibular border to the lingual foramen were examined to describe the MLF characteristics. Gender and age differences with respect to foramina were also studied.

RESULTS

CBCT can be utilized to visualise lingual foramina. In this study, 683 lingual foramina were detected in 200 CBCT scans, with 538 (78.77%) being ≤1 mm in diameter and 145 (21.23%) being >1 mm. In total, 85.07% of MLF are median lingual canals (MLC) and 14.93% are lateral lingual canals (LLC). Two typical types of lingual foramina were identified according to their relationship with the tooth apex. Most lingual foramina (74.08%) were found below the tooth apex, and those above the tooth apex were much smaller in diameter. Male patients had statistically larger lingual foramina. The distance between the lingual foramen and the tooth apex changed with increasing age.

CONCLUSIONS

Determination of the presence, position and size of lingual foramina is important before performing a surgical procedure. Careful implant-prosthetic treatment planning is particularly important in male and/or elderly patients because of the structural characteristics of their lingual foramina.

Cone-beam CT for imaging of the head/brain: Development and assessment of scanner prototype and reconstruction algorithms

PURPOSE

Our aim was to develop a high-quality, mobile cone-beam computed tomography (CBCT) scanner for point-of-care detection and monitoring of low-contrast, soft-tissue abnormalities in the head/brain, such as acute intracranial hemorrhage (ICH). This work presents an integrated framework of hardware and algorithmic advances for improving soft-tissue contrast resolution and evaluation of its technical performance with human subjects.

METHODS

Four configurations of a CBCT scanner prototype were designed and implemented to investigate key aspects of hardware (including system geometry, antiscatter grid, bowtie filter) and technique protocols. An integrated software pipeline (c.f., a serial cascade of algorithms) was developed for artifact correction (image lag, glare, beam hardening and x-ray scatter), motion compensation, and three-dimensional image (3D) reconstruction [penalized weighted least squares (PWLS), with a hardware-specific statistical noise model]. The PWLS method was extended in this work to accommodate multiple, independently moving regions with different resolution (to address both motion compensation and image truncation). Imaging performance was evaluated quantitatively and qualitatively with 41 human subjects in the neurosciences critical care unit (NCCU) at our institution.

RESULTS

The progression of four scanner configurations exhibited systematic improvement in the quality of raw data by variations in system geometry (source-detector distance), antiscatter grid, and bowtie filter. Quantitative assessment of CBCT images in 41 subjects demonstrated: ~70% reduction in image nonuniformity with artifact correction methods (lag, glare, beam hardening, and scatter); ~40% reduction in motion-induced streak artifacts via the multi-motion compensation method; and ~15% improvement in soft-tissue contrast-to-noise ratio (CNR) for PWLS compared to filtered backprojection (FBP) at matched resolution. Each of these components was important to improve contrast resolution for point-of-care cranial imaging.

CONCLUSIONS

This work presents the first application of a high-quality, point-of-care CBCT system for imaging of the head/ brain in a neurological critical care setting. Hardware configuration iterations and an integrated software pipeline for artifacts correction and PWLS reconstruction mitigated artifacts and noise to achieve image quality that could be valuable for point-of-care detection and monitoring of a variety of intracranial abnormalities, including ICH and hydrocephalus.