DICOM Datasets Research Articles

Insights into geometric deviations of medical 3d-printing: a phantom study utilizing error propagation analysis

BackgroundThe use of 3D-printing in medicine requires a context-specific quality assurance program to ensure patient safety. The process of medical 3D-printing involves several steps, each of which might be prone to its own set of errors. The segmentation error (SegE), the digital editing error (DEE) and the printing error (PrE) are the most important partial errors. Approaches to evaluate these have not yet been implemented in a joint concept. Consequently, information on the stability of the overall process is often lacking and possible process optimizations are difficult to implement. In this study, SegE, DEE, and PrE are evaluated individually, and error propagation is used to examine the cumulative effect of the partial errors.MethodsThe partial errors were analyzed employing surface deviation analyses. The effects of slice thickness, kernel, threshold, software and printers were investigated. The total error was calculated as the sum of SegE, DEE and PrE.ResultsThe higher the threshold value was chosen, the smaller were the segmentation results. The deviation values varied more when the CT slices were thicker and when the threshold was more distant from a value of around -400 HU. Bone kernel-based segmentations were prone to artifact formation. The relative reduction in STL file size [as a proy for model complexity] was greater for higher levels of smoothing and thinner slice thickness of the DICOM datasets. The slice thickness had a minor effect on the surface deviation caused by smoothing, but it was affected by the level of smoothing. The PrE was mainly influenced by the adhesion of the printed part to the build plate. Based on the experiments, the total error was calculated for an optimal and a worst-case parameter configuration. Deviations of 0.0093 mm ± 0.2265 mm and 0.3494 mm ± 0.8001 mm were calculated for the total error.ConclusionsVarious parameters affecting geometric deviations in medical 3D-printing were analyzed. Especially, soft reconstruction kernels seem to be advantageous for segmentation. The concept of error propagation can contribute to a better understanding of the process specific errors and enable future analytical approaches to calculate the total error based on process parameters.

A Computer-Aided Model for Dental Image Diagnosis Utilizing Convolutional Neural Networks

A Convolutional Neural Network (CNN) is an artificial neural network that is primarily utilized for the purposes of image recognition and processing, owing to its remarkable ability to recognize patterns within images. CNNs have found widespread application in diverse areas of computer vision, including but not limited to object tracking and recognition, security, and military and biomedical image analysis. CNN in orthodontic medical imaging technologies to reduce orthodontic treatment planning time, including automatic landmark search on cephalometric radiographs, cone beam computed tomography (CBCT) tooth segmentation, and CBCT tooth segmentation. This paper describes the strategy and the architecture of deep convolutional neural networks applied to DICOM datasets to distinguish between X-ray pictures with and without teeth. This work focuses on the application of the CNNs to a DICOM dataset in orthodontics as pre-processing for CNNs using fuzzy C-means clustering to construct a reasonable prediction that results in improved accuracy of this system. The aforementioned proposal has shown encouraging outcomes and visual representations, indicating that utilizing methods based on convolutional neural networks can greatly enhance the computational planning of orthodontic treatments by decreasing the time required for analysis. In many cases, this approach's analysis surpasses the accuracy of a manual orthodontist. This model achieved an accuracy exceeding 98% in applying the CNNs to differentiate between X-ray images with teeth and others with no teeth to focus any further work only on useful images which helps in diagnosis planning.

Covid 19 Pozitif Hastaların Akciğer ve Göğüs Hacminin Incelenmesi

Background: We aimed to reveal lung volume, lesion volume and thorax diameters in patients with Covid-19 positive in CT images according to gender and presence of ground glass opacity. Materials and Methods: We included CT images of 113 people (male;58, female;55) diagnosed with Covid-19 positive in our study. CT images of each patient were imported the DICOM datasets and thorax diameter measurements were performed by using MicroDicom. Moreover, lung volume and lesion volume parameters were calculated from the CT images using the three-dimensional imaging method (Vitrea). All measurement parameters were compared according to gender and ground glass opacity appearance. Results: The 113 patients (male;58, female;55) with a mean age of 44.57±14.59 diagnosed with Covid-19 were included in our study. There was a significant difference between the gender in the mean values of lung volume (males; 3530.90±1454.99 ml and females; 3149.92±1335.16 ml) and lesion volume (males; 434.59±664.19 ml and females; 557.10±671.99 ml) parameters. In addition, a significant difference in lung volume (p<0.001) and lesion volume (p<0.001) was obtained between the groups with ground glass opacity and without ground glass opacity. Moreover, A significant difference (p=0.001) in thorax anteroposterior diameter length was obtained between the groups the with/without ground glass opacity. However, there was no significant difference (p=0.542) between genders in thorax diameters. Conclussion: In our study, we found that in the females may have more lung damage from Covid-19. We believe that our work in the light of the available data, we believe that the findings of our study will contribute to early stages of the pandemic, in which vaccination has not yet , and the long-term effects of post-covid.

Functional Load Capacity of Teeth with Reduced Periodontal Support: A Finite Element Analysis.

The purpose of this study was to investigate the functional load capacity of the periodontal ligament (PDL) in a full arch maxilla and mandible model using a numerical simulation. The goal was to determine the functional load pattern in multi- and single-rooted teeth with full and reduced periodontal support. CBCT data were used to create 3D models of a maxilla and mandible. The DICOM dataset was used to create a CAD model. For a precise description of the surfaces of each structure (enamel, dentin, cementum, pulp, PDL, gingiva, bone), each tooth was segmented separately, and the biomechanical characteristics were considered. Finite Element Analysis (FEA) software computed the biomechanical behavior of the stepwise increased force of 700 N in the cranial and 350 N in the ventral direction of the muscle approach of the masseter muscle. The periodontal attachment (cementum-PDL-bone contact) was subsequently reduced in 1 mm increments, and the simulation was repeated. Quantitative (pressure, tension, and deformation) and qualitative (color-coded images) data were recorded and descriptively analyzed. The teeth with the highest load capacities were the upper and lower molars (0.4-0.6 MPa), followed by the premolars (0.4-0.5 MPa) and canines (0.3-0.4 MPa) when vertically loaded. Qualitative data showed that the areas with the highest stress in the PDL were single-rooted teeth in the cervical and apical area and molars in the cervical and apical area in addition to the furcation roof. In both single- and multi-rooted teeth, the gradual reduction in bone levels caused an increase in the load on the remaining PDL. Cervical and apical areas, as well as the furcation roof, are the zones with the highest functional stress. The greater the bone loss, the higher the mechanical load on the residual periodontal supporting structures.

Dose Escalation in Esophageal Cancer: Comparing Pre-Accrual and On-Trial Target Volume Delineation in the UK SCOPE2 Trial

The ongoing UK SCOPE2 trial evaluates radiotherapy (RT) dose escalation and PET-guided systemic therapy in esophageal cancer, and has an accompanying RT trials quality assurance (RTTQA) program, evolved through the preceding SCOPE trials. We compare pre-accrual with on-trial individual case review (ICR) target volume delineation (TVD). Prior to recruitment, centers were required to undertake TVD exercises using 3D/4D DICOM datasets with relevant clinical details and a RT planning guidance document (RPGD) provided. Contours were then compared against the RTTQA team-defined gold standard. Exceptions were those who had satisfied QA requirements for a previous esophageal RT trial (NeoSCOPE). For ICRs, prospective reviews (prior RT start, PRs) were undertaken for each center's first submission, plus high-dose cases submitted pending formal safety review. Additional PRs were undertaken at the RTTQA team's discretion. Timely retrospective reviews (within 2 weeks of RT start, TRR) were also undertaken for a random 10% sample. TVDs were assessed for compliance using predefined criteria and the RPGD. Resubmission was requested at reviewer's discretion, usually due to unacceptable variation (UV) from protocol. Clarification was sought before contour approval/resubmission request if appropriate. Review outcomes were then evaluated. PTV6000 was new to SCOPE2, along with a greater emphasis on use of 4DCT than in prior SCOPE trials. A total of 85 pre-accrual cases from 33 UK centers were reviewed, of which 20 (24%) were resubmissions, and 50 (59%) were accepted. 99 TVD UVs were observed in 49 cases, most commonly in CTVB (42/99, 42%), which included editing for normal structures and elective lymph node regions, followed by ITV (4D cases only, 14/52, 27%) and PTV6000 (13/99, 13%). 121 ICRs from 31 UK centers were available for review. 87 (72%) were PRs and 34 (28%) TRRs. 43 (36%) completed the relevant SCOPE2 exercise. 19 (16%) were resubmissions, and 82 (68%) were accepted. 72 UVs were observed in 45 ICRs; again, most commonly in CTVB (34/72, 48%), PTV6000 (high dose arm only, 11/46, 24%) and ITV (4D only, 5/26, 19%). Of the 45 cases where a UV was recorded, 16 (36%) had completed the relevant SCOPE2 pre-accrual. Comparing area of UV on SCOPE2 pre-accrual cases and ICRs, 3 (19%) contours contained the same (2 = CTVB, 1 = PTV6000), 5 (31%) contained different and 8 (50%) had no UVs at pre-accrual. The rate of UV was significantly lower for ICR than for pre-accrual submissions (0.60 and 1.16 respectively, p = 0.001). Significantly fewer UVs in ICR compared with pre-accrual supports a robust, educational RTTQA program through national collaboration and evolving trial series. CTVB, along with newer volumes of ITV and PTV6000, were recurring UV domains and should inform RPGD development and RTTQA for ongoing recruitment and future trials.

MA-QC: Free online software for prostate MR quality control and PI-QUAL assessment

PurposeTo validate MRI Analyzer Quality Control (MA-QC), a free and open-source online software designed to facilitate MR data acquisition quality control and PI-QUAL score calculation. Material and methodsMA-QC is a web-based software, designed for analysing DICOM data related to MR acquisition parameters. The software allows automatic extraction of 18 technical criteria, and manual input of 12 visual criteria, to calculate the PI-QUAL score. We collected 100 prostate MRI datasets from four MR device manufacturers to test data compatibility, automatic sequence recognition, and robustness of technical criteria extraction from DICOM data. The main issue was to determine the spatial resolution in the phase and frequency directions, due to variable encoding of the DICOM datasets. ResultsAcquisition data could be extracted from all sample examinations (100%), with a median analysis speed of 15.2 ± 4.4 images per second and mean processing time of 96 [11–326] seconds per examination. MA-CQ automatically detected the optimal T2-w, DWI and DCE sequences in 71 out of 100 (71%) cases, and required manual selection of at least one sequence in 29 out of 100 (29%) cases to get the best parameters. Display of technical criteria for the 3 sequences was instantaneous. PI-QUAL score could be calculated in all cases. ConclusionThis software brings substantial help in the quality assessment of prostate MRI examinations, by providing fast extraction of series data and the 18 technical parameters of PI-QUAL. PI-QUAL scoring can be performed in less than two minutes, helping to focus on the visual criteria, allowing use of this software in the clinical workflow in the aim of improving overall image quality in prostate MR imaging.

Classification of Chest X-ray Images using CNN for Medical Decision Support System

X-rays are a crucial tool used by healthcare professionals to diagnose a range of medical conditions. However, it is important to keep in mind that a timely and accurate diagnosis is crucial for effective patient management and treatment. While chest X-rays can provide highly precise anatomical data, manual interpretation of the images can be time-consuming and prone to errors, which can lead to delays or incorrect diagnoses. To address these issues, healthcare systems have taken steps to improve diagnostic imaging services following the impact of the COVID-19 pandemic. While deep learning-based automated systems for classifying chest X-rays have shown promise, there are still several challenges that need to be addressed before they can be widely used in clinical settings, including the lack of comprehensive and high-quality datasets. To overcome these limitations, a real-time DICOM dataset, has been converted to JPEG format to increase processing speed and improve data control. Three pre-trained models and a convolutional neural network (CNN) model with low complexity and three convolutional layers for feature extraction, along with max pooling layers and ReLU and Softmax activation functions have been implemented. With an validation accuracy of 95.05% on their CNN model using the SGD optimizer, the result has been validated using a separate, real-time unlabeled DICOM dataset of 1000 X-ray images.

Auto-segmentation of thoracic organs in CT scans of breast cancer patients using a 3D U-net cascaded into 2D patchGANs

The quality of organ volume delineation significantly influences the efficacy of radiotherapy treatment for breast cancer patients. This study introduces a novel method for auto-segmentation of the breasts, lungs and heart. The proposed pipeline leverages a multi-class 3D U-Net with a pre-trained ResNet(2+1)D-18 encoder branch, cascaded with a 2D PatchGAN mask correction model for each class. This approach requires a single 3D model, providing a relatively efficient solution. The models were trained and evaluated on 70 thoracic DICOM datasets belonging to breast cancer patients. The evaluation demonstrated state-of-the-art segmentation performance, with mean Dice similarity coefficient values ranging from 0.89 to 0.98, Hausdorff distance values ranging from 2.25 to 8.68 mm, and mean surface distance values ranging from 0.62 to 2.79 mm. These results underscore the pipeline’s potential to enhance breast cancer diagnosis and treatment strategies, with possible applications in other medical sectors utilizing auto-segmentation.

3D Printing, Histological, and Radiological Analysis of Nanosilicate-Polysaccharide Composite Hydrogel as a Tissue-Equivalent Material for Complex Biological Bone Phantom.

Nanosilicate-polysaccharide composite hydrogels are a well-studied class of materials in regenerative medicine that combine good 3D printability, staining, and biological properties, making them an excellent candidate material for complex bone scaffolds. The aim of this study was to develop a hydrogel suitable for 3D printing that has biological and radiological properties similar to those of the natural bone and to develop protocols for their histological and radiological analysis. We synthesized a hydrogel based on alginate, methylcellulose, and laponite, then 3D printed it into a series of complex bioscaffolds. The scaffolds were scanned with CT and CBCT scanners and exported as DICOM datasets, then cut into histological slides and stained using standard histological protocols. From the DICOM datasets, the average value of the voxels in Hounsfield Units (HU) was calculated and compared with natural trabecular bone. In the histological sections, we tested the effect of standard histological stains on the hydrogel matrix in the context of future cytological and histological analysis. The results confirmed that an alginate/methylcellulose/laponite-based composite hydrogel can be used for 3D printing of complex high fidelity three-dimensional scaffolds. This opens an avenue for the development of dynamic biological physical phantoms for bone tissue engineering and the development of new CT-based imaging algorithms for the needs of radiology and radiation therapy.

DoseMRT: A Software Package for Individualised Monte Carlo Dose Calculations of Synchrotron-Generated Microbeam Radiation Therapy

This work describes the creation and experimental validation of DoseMRT, a new software package, and its associated workflow for dose calculations in synchrotron-generated broad beam and microbeam radiation treatment fields. The DoseMRT software package allows users to import CT DICOM datasets into Geant4 for Monte Carlo dose calculations. It also provides basic treatment planning capabilities, simplifying the complexity of performing Geant4 simulations and making our Monte Carlo dose calculation algorithm accessible to a broader range of users. To demonstrate the new package, dose calculations are validated against experimental measurements performed in homogeneous water tank phantoms and the anatomically complex Alderson Radiotherapy Phantom for both broad-beam and microbeam configurations. Additionally, DoseMRT is successfully utilised as the primary method for patient-specific treatment prescription in an in vivo experiment involving tumour-bearing rats at the Imaging and Medical Beamline of the Australian Synchrotron.

Accuracy of Cone-Beam Computed Tomography Software in Predicting the Size of Impacted Canine: A Preliminary Study.

To compare and analyze the precision, accuracy, and reliability of commonly used cone-beam computed tomography (CBCT) software in predicting the mesiodistal diameter of impacted canines. This study was conducted on 11 patients (six males and five females, mean age: 17.5±5.5 years) with either unilateral or bilateral impacted canines in the maxilla or mandible. DICOM data sets of the patients obtained from CBCT scans were then loaded and visualized with four selected CBCT software to measure the widest mesiodistal diameter of the impacted teeth. Physical measurements using a digital vernier caliper, kept as a control, were also made on the extracted teeth and orthodontically erupted teeth. The collected data underwent statistical analysis, and the statistical significance level was set at p<0.05. The Bland-Altman analysis was performed to quantify the agreement between different software to the digital caliper, showing a narrow difference for all plots. Kruskal-Wallis ANOVA test followed by a post hoc test was performed to determine whether there was any difference in measuring the mesiodistal diameter of the impacted canine among the five methods, and tend no statistically significant difference was found among the five methods. Intraclass correlation (ICC) was performed, and measurements made with all CBCT software yielded an ICC greater than 0.95, indicating high reliability of the selected software. All the evaluated CBCT imaging software exhibited a high degree of reliability, and accuracy in precise measurement of the mesiodistal diameter of an impacted tooth.

Morphometric analysis of the infratemporal fossa using three-dimensional (3D) digital models.

The infratemporal fossa contains important neurovascular components and is directly related to other anatomical regions and structures. The morphometric distances between the bones forming its borders have not been thoroughly investigated. The aim of this study was to determine the morphometry of the infratemporal fossa. 3D models of the skull of 83 individuals were reconstructed from DICOM datasets, from which length, depth and width measurements were determined and compared between genders and the right and left sides. All measurements obtained were significantly different between males and females. There were also significant differences between the left and right sides for depth and width measurements. This is the first study to determine and investigate measurements of the infratemporal fossa; as such it provides a comprehensive view of the morphology of the fossa. It provides valuable information for surgical interventions and differential diagnoses of pathologies in this region, as well as enhancing its understanding in medical education.

3D-Printed Iodine-Ink CT Phantom for Radiomics Feature Extraction - Advantages and Challenges.

To test and validate novel CT techniques, such as texture analysis in radiomics, repeat measurements are required. Current anthropomorphic phantoms lack fine texture and true anatomic representation. 3D-printing of iodinated ink on paper is a promising phantom manufacturing technique. Previously acquired or artificially created CT data can be used to generate realistic phantoms. To present the design process of an anthropomorphic 3D-printed iodine ink phantom, highlighting the different advantages and pitfalls in its use. To analyze the phantom's x-ray attenuation properties, and the influences of the printing process on the imaging characteristics, by comparing it to the original input dataset. Two patient CT scans and artificially generated test patterns were combined in a single dataset for phantom printing and cropped to a size of 26×19×30 cm3 . This DICOM dataset was printed on paper using iodinated ink. The phantom was CT-scanned and compared to the original image dataset used for printing the phantom. The water-equivalent diameter of the phantom was compared to that of a patient cohort (N = 104). Iodine concentrations in the phantom were measured using dual-energy CT. 86 radiomics features were extracted from 10 repeat phantom scans and the input dataset. Features were compared using a histogram analysis and a PCA individually and overall, respectively. The frequency content was compared using the normalized spectrum modulus. Low density structures are depicted incorrectly, while soft tissue structures show excellent visual accordance with the input dataset. Maximum deviations of around 30 HU between the original dataset and phantom HU values were observed. The phantom has x-ray attenuation properties comparable to a lightweight adult patient (∼54 kg, BMI 19 kg/m2 ). Iodine concentrations in the phantom varied between 0 and 50 mg/ml. PCA of radiomics features shows different tissue types separate in similar areas of PCA representation in the phantom scans as in the input dataset. Individual feature analysis revealed systematic shift of first order radiomics features compared to the original dataset, while some higher order radiomics features did not. The normalized frequency modulus |f(ω)| of the phantom data agrees well with the original data. However, all frequencies systematically occur more frequently in the phantom compared to the maximum of the spectrum modulus than in the original data set, especially for mid-frequencies (e.g. for ω = 0.3942 mm-1 , |f(ω)|original = 0.09 * |fmax |original and |f(ω)|phantom = 0.12 * |fmax |phantom ). 3D-iodine-ink-printing technology can be used to print anthropomorphic phantoms with a water-equivalent diameter of a lightweight adult patient. Challenges include small residual air enclosures and the fidelity of HU values. For soft tissue, there is a good agreement between the HU values of the phantom and input data set. Radiomics texture features of the phantom scans are similar to the input data set, but systematic shifts of radiomics features in first order features, due to differences in HU values, need to be considered. The paper substrate influences the spatial frequency distribution of the phantom scans. This phantom type is of very limited use for dual-energy CT analyses. This article is protected by copyright. All rights reserved.

Multimodal Imaging of a Culotte Bifurcation Procedure in the Left Main Coronary Artery of a Perfusion-Fixed Human Heart: Step-by-Step with Serial Micro-CT Analyses

Percutaneous coronary intervention can be a high-risk procedure that would benefit from optimizing device-tissue interactions between stents and coronary vessels. Using a perfusion-fixed human heart with coronary artery disease, we performed a percutaneous coronary intervention of the left main coronary artery bifurcation. This heart was perfused and multimodal imaging was utilized to view the procedure with direct visualization, fluoroscopy, and optical coherence tomography (OCT). We followed the European Bifurcation Club’s guidelines to perform a single-stent bifurcation before transitioning to a two-stent Culotte technique. After each procedural step, the heart was removed from the perfusion apparatus and transferred to a micro-CT scanner to obtain unique scans. We conducted apposition analyses of the computational 3D models from micro-CT DICOM datasets, and compared them to the results from direct visualization and commercial OCT’s Apposition Indicator software. Additional measurements of resulting coronary anatomic expansions were taken to determine the potential roles of each step in improving procedural outcomes.Graphical Micro-CT images show stent deformation during a percutaneous coronary intervention(provisional to Culotte bifurcation procedure) in an isolated diseased human heart.

The use of a mini-ridge filter with cyclotron-based pencil beam scanning proton therapy.

Pencil beam scanning (PBS) proton therapy allows for far superior dose conformality compared with passive scattering techniques. However, one drawback of PBS is that the beam delivery time can be long, particularly when treating superficial disease. Minimizing beam delivery time is important for patient comfort and precision of treatment delivery. Mini-ridge filters (MRF) have been shown to reduce beam delivery time for synchrotron-based PBS. Given that cyclotron systems are widely used in proton therapy it is necessary to investigate the potential clinical benefit of mini-ridge filters in such systems. To demonstrate the clinical benefit of using a MRF to reduce beam delivery time for patients with large target volumes and superficial disease in cyclotron-based PBS proton therapy. A MRF beam model was generated by simulating the effect of a MRF on our clinical beam data assuming a fixed snout position relative to the isocenter. The beam model was validated with a series of measurements. The model was used to optimize treatment plans in a water phantom and on six patient DICOM datasets to further study the effect of the MRF and for comparison with physician-approved clinical treatment plans. Beam delivery time was measured for six plans with and without the MRF to demonstrate the reduction achievable. Plans with and without MRF were reviewed to confirm clinical acceptability by a radiation oncologist. Patient-specific QA measurements were carried out with a two-dimensional ionization chamber array detector for one representative patient's plan optimized with the MRF beam model. Results show good agreement between the simulated beam model and measurements with mean and maximum deviations of 0.06mm (0.45%) and 0.61mm (4.9%). The increase in Bragg peak width (FWHM) ranged from 2.7mm at 226MeV to 6.1mm at 70 MeV. The mean and maximum reduction in beam delivery time observed per field was 29.1s (32.2%) and 79.7s (55.3%). MRFs can be used to reduce treatment time in cyclotron-based PBS proton therapy without sacrificing plan quality. This is particularly beneficial for patients with large targets and superficial disease such as in breast cancer where treatment times are generally long, as well as patients treated with deep inspiration breath hold (DIBH).

The use of advanced technology for preoperative planning in cranial surgery – A survey by the EANS Young Neurosurgeons Committee

IntroductionTechnological advancements provided several preoperative tools allowing for precise preoperative planning in cranial neurosurgery, aiming to increase the efficacy and safety of surgery. However, little data are available regarding if and how young neurosurgeons are trained in using such technologies, how often they use them in clinical practice, and how valuable they consider these technologies. Research questionHow frequently these technologies are used during training and clinical practice as well as to how their perceived value can be qualitatively assessed. Materials and methodsThe Young Neurosurgeons’ Committee (YNC) of the European Association of Neurosurgical Societies (EANS) distributed a 14-items survey among young neurosurgeons between June 1st and August 31st, 2022. ResultsA total of 441 responses were collected. Most responders (42.34%) received “formal” training during their residency. Planning techniques were used mainly in neuro-oncology (90.86%), and 3D visualization of patients' DICOM dataset using open-source software was the most frequently used (>20 times/month, 20.34% of responders). Software for 3D visualization of patients’ DICOM dataset was the most valuable technology, especially for planning surgical approach (42.03%). Conversely, simulation based on augmented/mixed/virtual reality was considered the less valuable tool, being rated below sufficiency by 39.7% of responders. Discussion and conclusionTraining for using preoperative planning technologies in cranial neurosurgery is provided by neurosurgical residency programs. Software for 3D visualization of DICOM datasets is the most valuable and used tool, especially in neuro-oncology. Interestingly, simulation tools based on augmented/virtual/mixed reality are considered less valuable and, therefore, less used than other technologies.

A Deep Learning Model for Preoperative Differentiation of Glioblastoma, Brain Metastasis, and Primary Central Nervous System Lymphoma: An External Validation Study

(1) Background: Neuroimaging differentiation of glioblastoma, primary central nervous system lymphoma (PCNSL) and solitary brain metastasis (BM) represents a diagnostic and therapeutic challenge in neurosurgical practice, expanding the burden of care and exposing patients to additional risks related to further invasive procedures and treatment delays. In addition, atypical cases and overlapping features have not been entirely addressed by modern diagnostic research. The aim of this study was to validate a previously designed and internally validated ResNet101 deep learning model to differentiate glioblastomas, PCNSLs and BMs. (2) Methods: We enrolled 126 patients (glioblastoma: n = 64; PCNSL: n = 27; BM: n = 35) with preoperative T1Gd-MRI scans and histopathological confirmation. Each lesion was segmented, and all regions of interest were exported in a DICOM dataset. A pre-trained ResNet101 deep neural network model implemented in a previous work on 121 patients was externally validated on the current cohort to differentiate glioblastomas, PCNSLs and BMs on T1Gd-MRI scans. (3) Results: The model achieved optimal classification performance in distinguishing PCNSLs (AUC: 0.73; 95%CI: 0.62–0.85), glioblastomas (AUC: 0.78; 95%CI: 0.71–0.87) and moderate to low ability in differentiating BMs (AUC: 0.63; 95%CI: 0.52–0.76). The performance of expert neuro-radiologists on conventional plus advanced MR imaging, assessed by retrospectively reviewing the diagnostic reports of the selected cohort of patients, was found superior in accuracy for BMs (89.69%) and not inferior for PCNSL (82.90%) and glioblastomas (84.09%). (4) Conclusions: We investigated whether the previously published deep learning model was generalizable to an external population recruited at a different institution—this validation confirmed the consistency of the model and laid the groundwork for future clinical applications in brain tumour classification. This artificial intelligence-based model might represent a valuable educational resource and, if largely replicated on prospective data, help physicians differentiate glioblastomas, PCNSL and solitary BMs, especially in settings with limited resources.

Comparison of 2 open-sourced 3-dimensional modeling techniques for orthopaedic application.

Objectives: Although 3-dimensional (3D) printing is becoming more widely adopted for clinical applications, it is yet to be accepted as part of standard practice. One of the key applications of this technology is orthopaedic surgical planning for urgent trauma cases. Anatomically accurate replicas of patients' fracture models can be produced to guide intervention. These high-quality models facilitate the design and printing of patient-specific implants and surgical devices. Therefore, a fast and accurate workflow will help orthopaedic surgeons to generate high-quality 3D printable models of complex fractures. Currently, there is a lack of access to an uncomplicated and inexpensive workflow. Methods: Using patient DICOM data sets (n = 13), we devised a novel, simple, open-source, and rapid modeling process using Drishti software and compared its efficacy and data storage with the 3D Slicer image computing platform. We imported the computed tomography image directory acquired from patients into the software to isolate the model of bone surface from surrounding soft tissue using the minimum functions. One pelvic fracture case was further integrated into the customized implant design practice to demonstrate the compatibility of the 3D models generated from Drishti. Results: The data sizes of the generated 3D models and the processing files that represent the original DICOM of Drishti are on average 27% and 12% smaller than that of 3D Slicer, respectively (both P < 0.05). The time frame needed to reach the stage of viewing the 3D bone model and the exporting of the data of Drishti is 39% and 38% faster than that of 3D Slicer, respectively (both P < 0.05). We also constructed a virtual model using third-party software to trial the implant design. Conclusions: Drishti is more suitable for urgent trauma cases that require fast and efficient 3D bone reconstruction with less hardware requirement. 3D Slicer performs better at quantitative preoperative planning and multilayer segmentation. Both software platforms are compatible with third-party programs used to produce customized implants that could be useful for surgical training. Level of Evidence: Level V.

Comparing a knowledge-based 3D reconstruction algorithm to TomTec 3D echocardiogram algorithm in measuring left cardiac chamber volumes in the pediatric population.

Three-dimensional echocardiography (3DE) is an emerging method for volumetric cardiac measurements; however, few vendor-neutral analysis packages exist. Ventripoint Medical System Plus (VMS3.0+) proprietary software utilizes a validated Magnetic resonance imaging (MRI) database of normal ventricular and atrial morphologies to calculate chamber volumes. This study aimed to compare left ventricular (LV) and atrial (LA) volumes obtained using VMS3.0+ to Tomtec echocardiography analysis software. Healthy controls (n=98) aged 0-18 years were prospectively recruited and 3D DICOM datasets focused on the LV and LA acquired. LV and LA volumes and ejection fractions were measured using TomTec Image Arena 3D LV analysis package and using VMS3.0+. Pearson correlation coefficients, Bland-Altman's plots, and intraclass coefficients (ICC) were calculated, along with analysis time. There was a very good correlation between Ventripoint Medical System (VMS) and Tomtec LV systolic (r2 =.88, ICC .89 [95% CI .81, .94]), and diastolic (r2 =.88, ICC .90 [95% CI .77, .95]) volumes, and between VMS and Tomtec LA diastolic (r2 =.75, ICC .89 [95% CI .81, .93]) and systolic (r2 =.88, ICC .91 [95% CI .78, .96]) volumes on linear regression models. Natural log transformations eliminated heteroscedasticity, and power transformations provided the best fit. The time (mins) to analyze volumes using VMS were less than using Tomtec (LV VMS 2.3±.5, Tomtec 3.3±.8, p<.001; LA: VMS 1.9±.4, Tomtec 3.4±1.0, p<.001). There was a very good correlation between knowledge-based (VMS3.0+) and 3D (Tomtec) algorithms when measuring 3D echocardiography-derived LA and LV volumes in pediatric patients. VMS was slightly faster than Tomtec in analyzing volumetric measurements.

“Image to patient” equal-resolution surface registration supported by a surface scanner: analysis of algorithm efficiency for computer-aided surgery

PurposeThe “image to patient” registration procedure is crucial for the accuracy of surgical instrument tracking relative to the medical image while computer-aided surgery. The main aim of this work was to create an equal-resolution surface registration algorithm (ERSR) and analyze its efficiency.MethodsThe ERSR algorithm provides two datasets with equal, high resolution and approximately corresponding points. The registered sets are obtained by projection of a user-designed rectangle(s)-shaped uniform clouds of points on DICOM and surface scanner datasets. The tests of the algorithm were performed on a phantom with titanium microscrews. We analyzed the influence of DICOM resolution on the effect of the ERSR algorithm and compared the ERSR to standard paired-points landmark transform registration. The methods of analysis were Target Registration Error, distance maps, and their histogram evaluation.ResultsThe mean TRE in case of ERSR equaled 0.8 ± 0.3 mm (resolution A), 0.8 ± 0.5 mm (resolution B), and 1.0 ± 0.7 mm (resolution C). The mean values were at least 0.4 mm lower than in the case of landmark transform registration. The distance maps between the model achieved from the scanner and the CT-based model were analyzed by histogram. The frequency of the first bin in a histogram of the distance map for ERSR was about 0.6 for all three resolutions of DICOM dataset and three times higher than in the case of landmark transform registration. The results were statistically analyzed using the Wilcoxon signed-rank test (alpha = 0.05).ConclusionThe tests proved a statistically significant higher efficiency of equal resolution surface registration related to the landmark transform algorithm. It was proven that the lower resolution of the CT DICOM dataset did not degrade the efficiency of the ERSR algorithm. We observed a significantly lower response to decreased resolution than in the case of paired-points landmark transform registration.