Tomography Datasets Research Articles

A prospective, randomized, single-blind study to compare two methods of treating cystic lesions in the jaw

The study aimed to assess the usefulness of a bovine bone substitute material in treating cystic lesions in the jaw with a maximum diameter of <4 cm. In this prospective, randomized, single-blind intervention study of 116 patients, 61 underwent cystectomy with a subsequent filling of the defect using a bovine xenograft, and 55 underwent cystectomy alone. Volumetric measurement of the cysts was performed preoperatively and 6 and 12 months postoperatively using the available digital volume tomography data sets. Follow-up appointments were made 14 days and 1, 3, 6, and 12 months postoperatively. Almost complete regeneration was seen in both treatment groups within 12 months, with no significant difference in absolute volume loss between the 2 groups (P=.521). Examination 14 days after surgery revealed a tendency for more wound healing disorders with the use of a bone substitute (P=.077). It was no longer possible to detect any further differences in later examinations. Using bovine bone substitute material has no radiologically measurable advantage over cystectomy alone without defect filling regarding bone regeneration. In addition, there was a tendency for more wound-healing disorders to occur in the bone substitute group.

Simulation-based prediction of bone healing and treatment recommendations for lower leg fractures: Effects of motion, weight-bearing and fibular mechanics.

Despite recent experimental and clinical progress in the treatment of tibial and fibular fractures, in clinical practice rates of delayed bone healing and non-union remain high. The aim of this study was to simulate and compare different mechanical conditions after lower leg fractures to assess the effects of postoperative motion, weight-bearing restrictions and fibular mechanics on the strain distribution and the clinical course. Based on the computed tomography (CT) data set of a real clinical case with a distal diaphyseal tibial fracture, a proximal and a distal fibular fracture, finite element simulations were run. Early postoperative motion data, recorded via an inertial measuring unit system and pressure insoles were recorded and processed to study strain. The simulations were used to compute interfragmentary strain and the von Mises stress distribution of the intramedullary nail for different treatments of the fibula, as well as several walking velocities (1.0km/h; 1.5km/h; 2.0km/h) and levels of weight-bearing restriction. The simulation of the real treatment was compared to the clinical course. The results show that a high postoperative walking speed was associated with higher loads in the fracture zone. In addition, a larger number of areas in the fracture gap with forces that exceeded beneficial mechanical properties longer was observed. Moreover, the simulations showed that surgical treatment of the distal fibular fracture had an impact on the healing course, whereas the proximal fibular fracture barely mattered. Weight-bearing restrictions were beneficial in reducing excessive mechanical conditions, while it is known that it is difficult for patients to adhere to partial weight-bearing recommendations. In conclusion, it is likely that motion, weight bearing and fibular mechanics influence the biomechanical milieu in the fracture gap. Simulations may improve decisions on the choice and location of surgical implants, as well as give recommendations for loading in the postoperative course of the individual patient.

Development and In-Silico and Ex-Vivo Validation of a Software for a Semi-Automated Segmentation of the Round Window Niche to Design a Patient Specific Implant to Treat Inner Ear Disorders.

The aim of this study was to develop and validate a semi-automated segmentation approach that identifies the round window niche (RWN) and round window membrane (RWM) for use in the development of patient individualized round window niche implants (RNI) to treat inner ear disorders. Twenty cone beam computed tomography (CBCT) datasets of unilateral temporal bones of patients were included in the study. Defined anatomical landmarks such as the RWM were used to develop a customized 3D Slicer™ plugin for semi-automated segmentation of the RWN. Two otolaryngologists (User 1 and User 2) segmented the datasets manually and semi-automatically using the developed software. Both methods were compared in-silico regarding the resulting RWM area and RWN volume. Finally, the developed software was validated ex-vivo in N = 3 body donor implantation tests with additively manufactured RNI. The independently segmented temporal bones of the different Users showed a strong consistency in the volume of the RWN and the area of the RWM. The volume of the semi-automated RWN segmentations were 48 ± 11% smaller on average than the manual segmentations and the area of the RWM of the semi-automated segmentations was 21 ± 17% smaller on average than the manual segmentation. All additively manufactured implants, based on the semi-automated segmentation method could be implanted successfully in a pressure-tight fit into the RWN. The implants based on the manual segmentations failed to fit into the RWN and this suggests that the larger manual segmentations were over-segmentations. This study presents a semi-automated approach for segmenting the RWN and RWM in temporal bone CBCT scans that is efficient, fast, accurate, and not dependent on trained users. In addition, the manual segmentation, often positioned as the gold-standard, actually failed to pass the implantation validation.

A Comparison of Hydrological and Geophysical Calibration Data in Layered Hydrologic Models of Mountain Hillslopes

AbstractBoth hydrological and geophysical data can be used to calibrate hillslope hydrologic models. However, these data often reflect hydrological dynamics occurring at disparate spatial scales. Their use as sole objectives in model calibrations may thus result in different optimum hydraulic parameters and hydrologic model behavior. This is especially true for mountain hillslopes where the subsurface is often heterogeneous and the representative elementary volume can be on the scale of several m3. This study explores differences in hydraulic parameters and hillslope‐scale storage and flux dynamics of models calibrated with different hydrological and geophysical data. Soil water content, groundwater level, and two time‐lapse electrical resistivity tomography (ERT) data sets (transfer resistance and inverted resistivity) from two mountain hillslopes in Wyoming, USA, are used to calibrate physics‐based surface–subsurface hydrologic models of the hillslopes. Calibrations are performed using each data set independently and all data together resulting in five calibrated parameter sets at each site. Model predicted hillslope runoff and internal hydrological dynamics vary significantly depending on the calibration data set. Results indicate that water content calibration data yield models that overestimate near‐surface water storage in mountain hillslopes. Groundwater level calibration data yield models that more reasonably represent hillslope‐scale storage and flux dynamics. Additionally, ERT calibration data yield models with reasonable hillslope runoff predictions but relatively poor predictions of internal hillslope dynamics. These observations highlight the importance of carefully selecting data for hydrologic model calibration in mountain environments. Poor selection of calibration data may yield models with limited predictive capability depending on modeling goals and model complexity.

Comparison of Six Different Methods for Measuring the Equine Hoof and Recording of its Three-Dimensional Conformation.

Different measuring techniques have been used to objectify the classification of hoof shape. The MicroScribe is a novel tool that might prove useful for measuring hooves without prior reconstruction or compensation of projection artefacts. The aim of this study was to compare biometric data of the equine hoof collected by the MicroScribe tool and measurements collected directly from hooves, scaled photographs and radiographs, from photogrammetry models and computed tomography datasets. The suitability of MicroScribe generated data to differentiate individual hoof conformations was tested. A total of 62 measures were recorded from 16 forehooves. 21 linear and nine angular measures were collected by at least four methods each, and evaluated further by analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA). Ratios and differences of these measures were calculated as suitable for the definition of hoof shapes and analysed as well. Absolute equivalency of methods was detected for five linear and none of the angular measurements. The precision of the tested measurement methods was comparable. In some cases, different methods measure different structures. Radiographs tended to overestimate, while computed tomography slides to underestimate distances. Photogrammetry and scaled photographs were less suitable for measuring hoof angles. The MicroScribe tool can readily be used for hoof measurements. Its values for linear measures showed good equivalency with other methods based on real hooves. For angular measurements, the uneven hoof surface might introduce imprecision. Not all hoof conformations could be detected based on measuring results alone. Diagnosis by a skilled veterinarian is still essential.

3D Electrochemical-Mechanical Battery Simulation Tool: Implementation with Full Cell Simulations and Verification with Operando X-ray Tomography

Most commercially used electrode materials contract and expand upon cycling. This change in volume influences the microstructure of the cell stack, which in turn impacts a range of performance parameters. Since direct observation of these microstructural changes with operando experiments is challenging and time intensive, a simulation tool that takes a real or artificially generated 3D microstructure and captures the volumetric changes in a cell during cycling would be valuable to enable rapid understanding of the impact of material choice, electrode and cell design, and operating conditions on the microstructural changes and identification of sources of mechanically-driven cell aging. Here, we report the development and verification of such a 3D electrochemical-mechanical tool, and provide an example use-case. We validate the tool by simulating the microstructural evolution of a graphite anode and a Li(Ni,Mn,Co)O2 cathode during cycling and comparing the results to X-ray tomography datasets of these electrodes taken during cycling. As an example use case for such a simulation tool, we explore how different volumetric expansion behaviors of the cathode material impact strain in the cell stack, illustrating how the material selection and its operation impact the mechanical behavior inside a cell.

Clinical Validation of Siemens’ Syngo.via Automatic Contouring System

The manual delineation of organs at risk is a process that requires a great deal of time both for the technician and for the physician. Availability of validated software tools assisted by artificial intelligence would be of great benefit, as it would significantly improve the radiation therapy workflow, reducing the time required for segmentation. The purpose of this article is to validate the deep learning-based autocontouring solution integrated in syngo.via RT Image Suite VB40 (Siemens Healthineers, Forchheim, Germany). For this purpose, we have used our own specific qualitative classification system, RANK, to evaluate more than 600 contours corresponding to 18 different automatically delineated organs at risk. Computed tomography data sets of 95 different patients were included: 30 patients with lung, 30 patients with breast, and 35male patients with pelvic cancer. The automatically generated structures were reviewed in the Eclipse Contouring module independently by 3 observers: an expert physician, an expert technician, and a junior physician. There is a statistically significant difference between the Dice coefficient associated with RANK 4 compared with the coefficient associated with RANKs 2 and 3 (P < .001). In total, 64% of the evaluated structures received the maximum score, 4. Only 1% of the structures were classified with the lowest score, 1. The time savings for breast, thorax, and pelvis were 87.6%, 93.5%, and 82.2%, respectively. Siemens' syngo.via RT Image Suite offers good autocontouring results and significant time savings.

Liver Tumors Segmentation Using 3D SegNet Deep Learning Approach

An ultrasonic filter detects signs of malignant tumors by analysing the image’s pixel quality fluctuations caused by a liver ailment. Signs of malignant growth proximity are identified in an ultrasound filter through image pixel quality variations from a liver’s condition. Those changes are more common in alcoholic liver conditions than in other etiologies of cirrhosis, suggesting that the cause may be alcohol instead of liver disease. Existing Two-Dimensional (2D) ultrasound data sets contain an accuracy rate of 85.9% and a 2D Computed Tomography (CT) data set of 91.02%. The most recent work on designing a Three-Dimensional (3D) ultrasound imaging system in or close to real-time is examined. In this article, a Deep Learning (DL) model is implemented and modified to fit liver CT segmentation, and a semantic pixel classification of road scenes is recommended. The architecture is called semantic pixel-wise segmentation and comprises a hierarchical link of encoder-decoder layers. A standard data set was used to test the proposed model for liver CT scans and the tumor accuracy in the training phase. For the normal class, we obtained 100% precision for chronic cirrhosis hepatitis (73%), offset cirrhosis (59.26%), and offensive cirrhosis (91.67%) for chronic hepatitis or cirrhosis (73,0%). The aim is to develop a Computer-Aided Detection (CAD) screening tool to detect steatosis. The results proved 98.33% exactness, 94.59% sensitivity, and 92.11% case with Convolutional Neural Networks (CNN) classification. Although the classifier’s performance did not differentiate so clearly at this level, it was recommended that CNN generally perform better due to the good relationship between Area under the Receiver Operating Characteristics Curve (AUC) and accuracy.

Non-linear fusion of dual-energy projections for XCT imaging of challenging components

The imaging of multi-material components having large changes in attenuation behaviour is widely regarded as a limitation of X-ray computed tomographic techniques. The use of Dual Energy X-ray Computed Tomography (DECT) is common in the field of medical imaging, providing many benefits to imaging and characterisation. However, multienergy methods such as DECT remain underutilised in the industrial sector. In this work, a novel measurement based multi-energy image fusion approach was evaluated using two radiographic image stacks acquired under two distinct polychromatic spectra; this fusion was then compared to the results of the individual high and low-energy datasets as well as their average. The implemented approach was tested using a Titanium (Ti) and Carbon Fibre reinforced Polymer (CFRP) double-step lap joint to evaluate performance. Results indicated improvements to contrast in Dual Energy fused radiographs, aiding the interrogation of complex components in 2D. The translation of these benefits into the Reconstructed CT volume showed promise for isolated material regions. It was established that careful parameter selection was needed to achieve optimum fusion results, as it was found that failure to do so, resulted in contrast suppression in regions of mixed materials. The process proposed here acts as a first foundational step into the fusion of multi-energy industrial X-ray computed tomography (XCT) datasets using standard XCT equipment. Further work will aim at refining and comparing it to other projection weighting functions and measures.

Evaluation of Radiation Treatment Planning Algorithms in IMRT and VMAT: A Comparative Dosimetric Study in Lung Equivalent Heterogeneous Medium.

In Radiotherapy, computation of dose is important since in a small field with heterogeneity, dose is usually computed with discrepancies. The present study was aimed to evaluate the dosimetry of treatment planning algorithms in lung equivalent heterogeneous medium for Volumetric Modulated Arc Therapy (VMAT) with step and shoot Intensity-Modulated Radiation Therapy (ss-IMRT), and dynamic Intensity-Modulated Radiation Therapy (d-IMRT). In this experimental study, Computerized Imaging Reference System (CIRS) phantom was used with an inhomogeneous Racemosa wood cylinder for two types of tumors, namely, Left Lung Central Tumor (LCT) and Left Lung Peripheral Tumor (LPT) in the CIRS left lung cavity. The computed tomography (CT) datasets were employed with the generation of VMAT, d-IMRT and ss-IMRT plans for the LCT and LPT irradiated with 6 MV photon beams. In this study, the accuracy and efficacy of two algorithms: Monte Carlo (MC) and the Pencil Beam (PB), from the Monaco treatment planning system (TPS), were tested by using Gafchromic EBT3 films and CIRS thorax phantom. Regardless of treatment techniques, both algorithms exhibited higher divergence in LPT than LCT. In both LCT and LPT, the highest deviation was near the tumor-lung junction. However, the deviation was higher in the PB algorithm than MC algorithm, with a minimally acceptable variation of -0.8%. The MC algorithm shows more consistency for EBT3 measured dose in lung equivalent heterogeneous medium. However, accurate dose predictions are complicated due to electronic disequilibrium within and at the interface of inhomogeneity. These constraints may cause variations from the anticipated outcomes of the treatments.

Dosimetric comparison of sequential intensity-modulated radiation therapy (IMRT) and simultaneous integrated boost IMRT for lymph node-positive cervical cancer

Abstract Introduction: Nodal boost is being increasingly employed to escalate the dose to involved nodes in node-positive cervical cancer. The study aimed to compare the dosimetric differences between sequential boost intensity-modulated radiation therapy (SeB-IMRT) and simultaneous integrated boost IMRT (SIB-IMRT) in terms of target coverage and organs-at-risk (OARs) with special emphasis on the effect of nodal shrinkage and anatomical change of normal tissues during radiotherapy. Methods: Two computed tomography (CT) datasets (of phase I and phase II) of 40 patients of node-positive cervical cancer treated with SeB-IMRT [planning target volume (PTV) 45/25] followed by SeB to residual nodes (PTV 12·6/7) were utilised. SIB-IMRT1 plan consisted of PTV pelvis and para-aortic nodal region (PTV 45/25) and SIB to gross nodes (PTV 55/25). In order to account for the change in nodal and normal tissue topography during treatment, a third plan (SIB-IMRT2) was generated by utilising the SIB-IMRT1 plan for 44 Gy in 20 fractions and reproducing the plan on the second CT dataset for 11 Gy in 5 fractions. Dosimetric parameters of the three plans were compared using the Friedman test with Bonferroni correction. Results: We observed that the doses to OARs (bowel, rectum and bladder) were significantly higher in SeB-IMRT plan as compared to the SIB-IMRT plans. V40 Gy of bowel for SeB-IMRT, SIB-IMRT1 and SIB-IMRT2 plans were 354·8 cc, 271 cc and 321·8 cc, respectively (p = 0·001), whereas V30 Gy were 687·8 cc, 635·5 cc and 680 cc, respectively (p = 0·001). The target coverage was marginally better in SeB-IMRT plan as compared to SIB-IMRT1 and SIB-IMRT2 plans (V95% = 99·2 versus 97·7 versus 97·9, respectively, p = 0·000) Conclusion: SIB-IMRT led to better sparing of OARs, especially bowel. However, the magnitude of benefit decreases if the change in nodal and normal tissue topography during radiotherapy is not considered implying the need for frequent image guidance when SIB-IMRT is planned for node-positive cervical cancer.

Unified Modulation Pattern Analysis (UMPA) algorithm for 1D sensitive X-ray phase contrast imaging techniques

X-ray phase contrast imaging (XPCI) techniques are sensitive to refraction (differential-phase) and small-angle X-ray scattering (dark-field) signals, not measurable with conventional absorption imaging techniques. Among XPCI techniques, edge illumination (EI), grating interferometry (GI), and speckle-based imaging (SBI) make use of wavefront markers, such as absorbing masks with periodical apertures or random diffusers, to encode refraction and dark-field signals induced by the sample. The Unified Modulated Pattern Analysis (UMPA) provides an algorithmic solution to extract the transmission, refraction, and dark-field images from EI, GI, and SBI datasets where the wavefront marker is directly resolved by the employed detection system. In its original implementation, UMPA has been designed for XPCI techniques sensitive to refractions along two axes. This work presents a modified version of the algorithm to extend its applicability to all the existing XPCI techniques that use wavefront markers with sensitivity to refraction limited along one direction (UMPA-1D). The algorithm, written in C++ and Cython and parallelized with OpenMP, enables fast reconstruction times that are particularly convenient for large tomographic datasets. The validity of the UMPA-1D has been demonstrated using both simulated images and real acquisitions with an EI setup in beam-tracking mode.

Effect of orthodontic retraction force on thick and thin gingival biotypes in different grades of gingival recession and alveolar bone quality: A finite element analysis.

To evaluate the effect of orthodontic retraction force on thick and thin gingival biotypes of anterior teeth having grade I and II gingival recession with D1 and D2 alveolar bone densities by assessing the displacement of teeth and deformation in gingival tissue, along with evaluating the amount of equivalent stress on teeth and gingiva in finite element model. Based on cone-beam computed tomography datasets, eight models (LC01-LC08) of maxilla with appliance assembly were created. A retraction load of 150 gm was applied on each model during the finite element analysis (FEA), and then values of teeth displacement, gingival deformation, and von Mises stress were evaluated. All the models with D1 bone density (LC01-04) were showing gingival deformation, whereas models with D2 bone quality (LC05-08) give no significant outcome. The amount of von Mises stress for teeth and gingival tissue in each model was the same for both central and lateral incisors. All the models having D1 bone quality showed greater gingival deformation after applying retraction force, whereas in models with D2 bone density, deformation was not significant. Changes in gingival tissue are brought about by orthodontic treatment, which also helps to correct the periodontal defects; however, bone density plays a significant role in improving gingival recession.

Automated Brain Hemorrhage Classification and Volume Analysis

Brain hemorrhage is a serious and life-threatening condition. It can cause permanent and lifelong disability even when it is not fatal. The word hemorrhage denotes leakage of blood within the brain and this leakage of blood from capillaries causes stroke and adequate supply of oxygen to the brain is hindered. Modern imaging methods such as computed tomography (CT) and magnetic resonance imaging (MRI) are employed to get an idea regarding the extent of the damage. An early diagnosis and treatment can save lives and limit the adverse effects of a brain hemorrhage. In this case, a deep neural network (DNN) is an effective choice for the early identification and classification of brain hemorrhage for the timely recovery and treatment of an affected person. In this paper, the proposed research work is divided into two novel approaches, where, one for the classification and the other for volume calculation of brain hemorrhage. Two different datasets are used for two different techniques classification and volume. A novel algorithm is proposed to calculate the volume of hemorrhage using CT scan images. In the first approach, the ‘RSNA’ dataset is used to classify the brain hemorrhage types using transfer learning and achieved an accuracy of 93.77%. Furthermore, in the second approach, a novel algorithm has been proposed to calculate the volume of brain hemorrhage and achieved tremendous results as 1035.91 mm3 and 9.25 cm3, using the PhysioNet CT scan tomography dataset.

ANALYSIS OF UPPER RESPIRATORY TRACT SEGMENTATION FEATURES TO DETERMINE NASAL CONDUCTANCE

The paper examines the features of segmentation of the upper respiratory tract to determine nasal air conduction. 2D and 3D illustrations of the segmentation process and the obtained results are given. When forming an analytical model of the aerodynamics of the nasal cavity, the main indicator that characterizes the configuration of the nasal canal is the equivalent diameter, which is determined at each intersection of the nasal cavity. It is calculated based on the area and perimeter of the corresponding section of the nasal canal. When segmenting the nasal cavity, it is first necessary to eliminate air structures that do not affect the aerodynamics of the upper respiratory tract - these are, first of all, intact spaces of the paranasal sinuses, in which diffuse air exchange prevails. In the automatic mode, this is possible by performing the elimination of unconnected isolated areas and finding the difference coefficients of the areas connected by confluences with the nasal canal in the next step. High coefficients of difference of sections between intersections will indicate the presence of separated areas and contribute to their elimination. The complex configuration and high individual variability of the structures of the nasal cavity does not allow segmentation to be fully automated, but this approach contributes to the absence of interactive correction in 80% of tomographic datasets. The proposed method, which takes into account the intensity of the image elements close to the contour ones, allows to reduce the averaging error from tomographic reconstruction up to 2 times due to artificial sub-resolution. The perspective of the work is the development of methods for fully automatic segmentation of the structures of the nasal cavity, taking into account the individual anatomical variability of the upper respiratory tract.

3D deep convolution neural network for radiation pneumonitis prediction following stereotactic body radiotherapy.

In this study, we investigated 3D convolutional neural networks (CNNs) with input from radiographic and dosimetric datasets of primary lung tumors and surrounding lung volumes to predict the likelihood of radiation pneumonitis (RP). Pre-treatment, 3- and 6-month follow-up computed tomography (CT) and 3D dose datasets from one hundred and ninety-three NSCLC patients treated with stereotactic body radiotherapy (SBRT) were retrospectively collected and analyzed for this study. DenseNet-121 and ResNet-50 models were selected for this study as they are deep neural networks and have been proven to have high accuracy for complex image classification tasks. Both were modified with 3D convolution and max pooling layers to accept 3D datasets. We used a minority class oversampling approach and data augmentation to address the challenges of data imbalance and data scarcity. We built two sets of models for classification of three (No RP, Grade 1 RP, Grade 2 RP) and two (No RP, Yes RP) classes as outputs. The 3D DenseNet-121 models performed better (F1 score [0.81], AUC [0.91] [three class]; F1 score [0.77], AUC [0.84] [two class]) than the 3D ResNet-50 models (F1 score [0.54], AUC [0.72] [three-class]; F1 score [0.68], AUC [0.71] [two-class]) (p=0.017 for three class predictions). We also attempted to identify salient regions within the input 3D image dataset via integrated gradient (IG) techniques to assess the relevance of the tumor surrounding volume for RP stratification. These techniques appeared to indicate the significance of the tumor and surrounding regions in the prediction of RP. Overall, 3D CNNs performed well to predict clinical RP in our cohort based on the provided image sets and radiotherapy dose information.

Smart Visualization of Medical Images as a Tool in the Function of Education in Neuroradiology.

The smart visualization of medical images (SVMI) model is based on multi-detector computed tomography (MDCT) data sets and can provide a clearer view of changes in the brain, such as tumors (expansive changes), bleeding, and ischemia on native imaging (i.e., a non-contrast MDCT scan). The new SVMI method provides a more precise representation of the brain image by hiding pixels that are not carrying information and rescaling and coloring the range of pixels essential for detecting and visualizing the disease. In addition, SVMI can be used to avoid the additional exposure of patients to ionizing radiation, which can lead to the occurrence of allergic reactions due to the contrast media administration. Results of the SVMI model were compared with the final diagnosis of the disease after additional diagnostics and confirmation by neuroradiologists, who are highly trained physicians with many years of experience. The application of the realized and presented SVMI model can optimize the engagement of material, medical, and human resources and has the potential for general application in medical training, education, and clinical research.

Technical note: Performance evaluation of volumetric imaging based on motion modeling by principal component analysis.

To quantitatively evaluate the achievable performance of volumetric imaging based on lung motion modeling by principal component analysis (PCA). In volumetric imaging based on PCA, internal deformation was represented as a linear combination of the eigenvectors derived by PCA of the deformation vector fields evaluated from patient-specific four-dimensional-computed tomography (4DCT) datasets. The volumetric image was synthesized by warping the reference CT image with a deformation vector field which was evaluated using optimal principal component coefficients (PCs). Larger PCs were hypothesized to reproduce deformations larger than those included in the original 4DCT dataset. To evaluate the reproducibility of PCA-reconstructed volumetric images synthesized to be close to the ground truth as possible, mean absolute error (MAE), structure similarity index measure (SSIM) and discrepancy of diaphragm position were evaluated using 22 4DCT datasets of nine patients. Mean MAE and SSIM values for the PCA-reconstructed volumetric images were approximately 80HU and 0.88, respectively, regardless of the respiratory phase. In most test cases including the data of which motion range was exceeding that of the modeling data, the positional error of diaphragm was less than 5 mm. The results suggested that large deformations not included in the modeling 4DCT dataset could be reproduced. Furthermore, since the first PC correlated with the displacement of the diaphragm position, the first eigenvector became the dominant factor representing the respiration-associated deformations. However, other PCs did not necessarily change with the same trend as the first PC, and no correlation was observed between the coefficients. Hence, randomly allocating or sampling these PCs in expanded ranges may be applicable to reasonably generate an augmented dataset with various deformations. Reasonable accuracy of image synthesis comparable to those in the previous research were shown by using clinical data. These results indicate the potential of PCA-based volumetric imaging for clinical applications.

Comparison of dosimetries of carbon-ion pencil beam scanning, proton pencil beam scanning and volumetric modulated arc therapy for locally recurrent rectal cancer.

We compared the dose distributions of carbon-ion pencil beam scanning (C-PBS), proton pencil beam scanning (P-PBS) and Volumetric Modulated Arc Therapy (VMAT) for locally recurrent rectal cancer. The C-PBS treatment planning computed tomography (CT) data sets of 10 locally recurrent rectal cancer cases were randomly selected. Three treatment plans were created using identical prescribed doses. The beam angles for C-PBS and P-PBS were identical. Dosimetry, including the dose received by 95% of the planning target volume (PTV) (D95%), dose to the 2cc receiving the maximum dose (D2cc), organ at risk (OAR) volume receiving > 15Gy (V15) and > 30Gy (V30), was evaluated. Statistical significance was assessed using the Wilcoxon signed-rank test. Mean PTV-D95% values were > 95% of the volume for P-PBS and C-PBS, whereas that for VMAT was 94.3%. However, PTV-D95% values in P-PBS and VMAT were < 95% in five and two cases, respectively, due to the OAR dose reduction. V30 and V15 to the rectum/intestine for C-PBS (V30 = 4.2 ± 3.2cc, V15 = 13.8 ± 10.6cc) and P-PBS (V30 = 7.3 ± 5.6cc, V15 = 21.3 ± 13.5cc) were significantly lower than those for VMAT (V30 = 17.1 ± 10.6cc, V15 = 55.2 ± 28.6cc). Bladder-V30 values with P-PBS/C-PBS (3.9 ± 4.8Gy(RBE)/3.0 ± 4.0Gy(RBE)) were significantly lower than those with VMAT (7.9 ± 8.1Gy). C-PBS provided superior dose conformation and lower OAR doses compared with P-PBS and VMAT. C-PBS may be the best choice for cases in which VMAT and P-PBS cannot satisfy dose constraints. C-PBS could be another choice for cases in which VMAT and P-PBS cannot satisfy dose constraints, thereby avoiding surgical resection.

Clinical Evaluation of Deep Learning and Atlas-Based Auto-Contouring for Head and Neck Radiation Therapy

Various commercial auto-contouring solutions have emerged over past few years to address labor-intensiveness, and inter- and intra-operator variabilities issues of traditional manual anatomy contouring for head and neck (H&amp;N) radiation therapy (RT). The purpose of this study is to compare the clinical performances between RaySearch Laboratories deep learning (DL) and atlas-based auto-contouring tools for organs at risk (OARs) segmentation in the H&amp;N RT with the manual contouring as reference. Forty-five H&amp;N computed tomography datasets were used for the DL and atlas-based auto-contouring tools to contour 16 OARs and time required for the segmentation was measured. Dice similarity coefficient (DSC), Hausdorff distance (HD) and HD 95th-percentile (HD95) were used to evaluate geometric accuracy of OARs contoured by the DL and atlas-based auto-contouring tools. Paired sample t-test was employed to compare the mean DSC, HD, HD95, and contouring time values of the two groups. The DL auto-contouring approach achieved more consistent performance in OARs segmentation than its atlas-based approach, resulting in statistically significant time reduction of the whole segmentation process by 40% (p &lt; 0.001). The DL auto-contouring had statistically significantly higher mean DSC and lower HD and HD95 values (p &lt; 0.001–0.009) for 10 out of 16 OARs. This study proves that the RaySearch Laboratories DL auto-contouring tool has significantly better clinical performances than its atlas-based approach.