OSPACS: Ultrasound image management system

Background  

Ultrasound scanning uses the medical imaging format, DICOM, for electronically storing the images and data associated with a particular scan. Large health care facilities typically use a picture archiving and communication system (PACS) for storing and retrieving such images. However, these systems are usually not suitable for managing large collections of anonymized ultrasound images gathered during a clinical screening trial.     

基于Visual C#语言的DICOM标准医学图像的显示和实现

介绍DICOM3.0医学图像文件的格式和C#语言的特点,首次利用Visual C#语言对该标准的图像进行显示和处理,能够直接读取DICOM格式原始图像数据,并可批量转换成BMP等格式进行处理,此项工作可为医学图像处理研究及相关医学图像软件开发奠定基础。     

Digital Imaging and Communications in Medicine (DICOM) information conversion procedure for SUV calculation of PET scanners with different DICOM header information

PurposeIn nuclear medicine, the standardized uptake value (SUV) obtained using positron emission tomography with 2-deoxy-2-fluoro-D-glucose (FDG-PET) is widely used as a semi-quantitative diagnosis factor. We found that the header file of the Philips Allegro PET scanner using the Digital Imaging and Communications in Medicine (DICOM) standard was stored differently than with other scanners. Thus, the purpose of this study was to develop a DICOM header information conversion program to ensure compatibility between Allegro and other equipment.Methods and resultsThe NEMA IEC Body phantom was scanned using the Allegro PET scanner. We conducted measurements and performed calculations by using commercial software and the proposed self-developed program, respectively, to compare the SUVs by using conversion data. The program consists of three parts: an input part that can load data regardless of the number of DICOM images, and conversion and output parts that can be used to convert the DICOM header information and store it in the order of slices. The results of the calculation are in good agreement with the data measured at 12 circular regions of interest. The percent difference was lower than the 20%.ConclusionIn conclusion, this study suggested a simple and convenient method to solve the incompatibility through conversion of the DICOM header information. This study thus provides physicians more accurate information for diagnosis and treatment.     

A methodology to generate structured computational grids from DICOM data: application to a patient-specific abdominal aortic aneurysm (AAA) model

This study presents the generation of a multi-block structured grid on a real abdominal aortic aneurysm (AAA) acquired from Digital Imaging and Communication in Medicine (DICOM) data. With the use of a computed tomography exam (or medical images in standard DICOM format), the shape of a human organ is extracted and a structured computational grid is created. The structured grid generation is done by utilising Floater's and Gopalsamy et al.'s algorithm. The proposed methodology is applied to the AAA case, but it may also be applied to other human organs, enabling the scientist to develop an advanced patient-specific model. More importantly, the proposed methodology provides a precise reconstruction of the human organs, which is required in an AAA, where small variations in the geometry may alter the flow field, the stresses exerted on the walls and finally the rupture risk of the aneurysm.     

心电数据管理系统中标准数据格式的分析

介绍了心电图机中三种标准化的存储格式,SCP,HL7aECG,DICOM,对这三种数据格式结构和主要特点进行了分析比较,对各自的优缺点进行了对比,随着网格化的发展,市场上会有会有越来越多的产家支持这几种通用的心电数据存储方式。     

浅析PACS中的DICOM标准

随着医学影像设备的广泛应用以及PACS的快速发展,为了统一各种数字化影像设备的图像数据格式和数据传输标准而诞生的DICOM标准已经成为医学数字成像和通讯的共同标准。本文简要的介绍了DICOM标准的历史以及DICOM数据集和DICOM文件格式的组织形式。     

Porte sonde motorisé pour une télé-échographie abdominale en temps différé : étude de faisabilité

PurposeDesign a system of delayed telesonography between expert center and isolated site where there is no sonographer.Materials and methodsA motorized probe holder printing a movement of TILT (± 40°) to any 2D echograph probe available in an isolated site, allows a non-sonographer to capture the patients’ whole organ images. This volume of images is sent to the expert center via internet. The video sequence is decomposed into images in the JPEG format by the software VirtualDub. A post-processing software (ECHO-CNES) allows the expert to find the incidences necessary for the diagnosis. This system has been tested on 50 patients at the CHU Trousseau (Tours, France).ResultsOrgans investigated were liver, portal vein, gall-bladder, kidneys (right and left) and spleen. The acquisition time of the volumes of images was 4 min maximum and allowed to reconstruct images necessary for the diagnosis in 80% of the cases. Scanning duration lasted 4 s on average; processing duration of the volume of images lasted approximately 15 min.ConclusionThis system is in validation in real situation in Togo with the aim of its generalization for medical care according to the obtained results.     

An iterative deconvolution algorithm for image recovery in clinical CT: A phantom study

PurposeTo study the feasibility of using an iterative reconstruction algorithm to improve previously reconstructed CT images which are judged to be non-diagnostic on clinical review. A novel rapidly converging, iterative algorithm (RSEMD) to reduce noise as compared with standard filtered back-projection algorithm has been developed.Materials and methodsThe RSEMD method was tested on in-silico, Catphan^®500, and anthropomorphic 4D XCAT phantoms. The method was applied to noisy CT images previously reconstructed with FBP to determine improvements in SNR and CNR. To test the potential improvement in clinically relevant CT images, 4D XCAT phantom images were used to simulate a small, low contrast lesion placed in the liver.ResultsIn all of the phantom studies the images proved to have higher resolution and lower noise as compared with images reconstructed by conventional FBP. In general, the values of SNR and CNR reached a plateau at around 20 iterations with an improvement factor of about 1.5 for in noisy CT images. Improvements in lesion conspicuity after the application of RSEMD have also been demonstrated. The results obtained with the RSEMD method are in agreement with other iterative algorithms employed either in image space or with hybrid reconstruction algorithms.ConclusionsIn this proof of concept work, a rapidly converging, iterative deconvolution algorithm with a novel resolution subsets-based approach that operates on DICOM CT images has been demonstrated. The RSEMD method can be applied to sub-optimal routine-dose clinical CT images to improve image quality to potentially diagnostically acceptable levels.     

A systematic quality assurance framework for the upgrade of radiation oncology information systems

In spite of its importance, no systematic and comprehensive quality assurance (QA) program for radiation oncology information systems (ROIS) to verify clinical and treatment data integrity and mitigate against data errors/corruption and/or data loss risks is available. Based on data organization, format and purpose, data in ROISs falls into five different categories: (1) the ROIS relational database and associated files; (2) the ROIS DICOM data stream; (3) treatment machine beam data and machine configuration data; (4) electronic medical record (EMR) documents; and (5) user-generated clinical and treatment reports from the ROIS. For each data category, this framework proposes a corresponding data QA strategy to very data integrity. This approach verified every bit of data in the ROIS, including billions of data records in the ROIS SQL database, tens of millions of ROIS database-associated files, tens of thousands of DICOM data files for a group of selected patients, almost half a million EMR documents, and tens of thousands of machine configuration files and beam data files. The framework has been validated through intentional modifications with test patient data. Despite the ‘big data’ nature of ROIS, the multiprocess and multithread nature of our QA tools enabled the whole ROIS data QA process to be completed within hours without clinical interruptions. The QA framework suggested in this study proved to be robust, efficient and comprehensive without labor-intensive manual checks and has been implemented for our routine ROIS QA and ROIS upgrades.     

Mining cancer biology through bioinformatic analysis of proteomic data

ABSTRACT

Introduction: Discovery proteomics for cancer research generates complex datasets of diagnostic, prognostic, and therapeutic significance in human cancer. With the advent of high-resolution mass spectrometers, able to identify thousands of proteins in complex biological samples, only the application of bioinformatics can lead to the interpretation of data which can be relevant for cancer research.

Areas covered: Here, we give an overview of the current bioinformatic tools used in cancer proteomics. Moreover, we describe their applications in cancer proteomics studies of cell lines, serum, and tissues, highlighting recent results and critically evaluating their outcomes.

Expert opinion: The use of bioinformatic tools is a fundamental step in order to manage the large amount of proteins (from hundreds to thousands) that can be identified and quantified in a cancer biological samples by proteomics. To handle this challenge and obtain useful data for translational medicine, it is important the combined use of different bioinformatic tools. Moreover, a particular attention to the global experimental design, and the integration of multidisciplinary skills are essential for best setting of tool parameters and best interpretation of bioinformatics output.     

Breast glandularity and mean glandular dose assessment using a deep learning framework: Virtual patients study

PurposeBreast dosimetry in mammography is an important aspect of radioprotection since women are exposed periodically to ionizing radiation due to breast cancer screening programs. Mean glandular dose (MGD) is the standard quantity employed for the establishment of dose reference levels in retrospective population studies. However, MGD calculations requires breast glandularity estimation. This work proposes a deep learning framework for volume glandular fraction (VGF) estimations based on mammography images, which in turn are converted to glandularity values for MGD calculations.Methods208 virtual breast phantoms were generated and compressed computationally. The mammography images were obtained with Monte Carlo simulations (MC-GPU code) and a ray-tracing algorithm was employed for labeling the training data. The architectures of the neural networks are based on the XNet and multilayer perceptron, adapted for each task. The network predictions were compared with the ground truth using the coefficient of determination (r²).ResultsThe results have shown a good agreement for inner breast segmentation (r² = 0.999), breast volume prediction (r² = 0.982) and VGF prediction (r² = 0.935). Moreover, the DgN coefficients using the predicted VGF for the virtual population differ on average 1.3% from the ground truth values. Afterwards with the obtained DgN coefficients, the MGD values were estimated from exposure factors extracted from the DICOM header of a clinical cohort, with median(75 percentile) values of 1.91(2.45) mGy.ConclusionWe successfully implemented a deep learning framework for VGF and MGD calculations for virtual breast phantoms.     

Comparison of three different phantoms used for Winston-Lutz test with Artiscan software

BackgroundOne of the most important test in every quality assurances process of medical linear accelerators is the Winston-Lutz test, allowing an evaluation of the treatment isocentre in the light of uncertainty of the position of the collimator, the gantry and the couch.AimThe purpose of this work was analysis of the results of the Winston-Lutz test performed with three different phantoms for two different accelerators.Materials and methodsMeasurements were performed on two Varian machines: TrueBeam equipped with aS1200 EPID and TrueBeam equipped with aS1000 EPID. During the study three different phantoms dedicated for verification of the radiation isocentre were used: PTW Isoball, AQUILAB Isocentre Phantom and Varian Isocentre Cube. Analysis of the DICOM images was performed in Artiscan software.ResultsFor TrueBeam with as1200 EPID, gantry MV isocentre was about 0.18 mm larger for Varian Isocentre Cube than for two other phantoms used in this study. The largest variability of this parameter was observed for the couch. The results differed to 1.16 mm. For TrueBeam with as1000 EPID, results for collimator isocentre with PTW Isoball phantom were about 0.10 mm larger than for two other phantoms. For the gantry, results obtained with Varian Isocentre Cube were 0.21 mm larger.ConclusionThe obtained results for all three phantoms are within the accepted tolerance range. The largest differences were observed for treatment couch, which may be related to the phantom mobility during couch movement.     

Validation clinique du module Oncoflash® en scintigraphie planaire thyroïdienne

IntroductionThe Oncoflash® adaptive filtering algorithm is poorly validated in clinical practice. The objective of this study was to evaluate this algorithm efficacity in order to reduce injected activity or acquisition time on thyroid scintigraphy.MethodsOne hundred and five patients who received a thyroid scan have been tested. Three sets of images (conventional acquisition, gross half-time acquisition, post-processed half-time acquisition by Oncoflash®) were interpreted as open ultrasound and biological data without any view on their medical file and the patient's identity by 2 nuclear doctors. The concordance with the diagnosis retained in the medical file, the quality of the images and the inter-observer reproducibility were evaluated.ResultsNo significant differences were found in terms of agreement with the final diagnosis between post-processed Oncoflash® half-time images and conventional images (κ at 0.81 and 0.74 respectively). The quality of Oncoflash® half-time images was rated good or excellent in 90% of cases compared to 88.6% of cases for conventional acquisitions (P = 0.16). No significant difference in inter-observer agreement was found between the 3 sets of images compared two by two.ConclusionThese results suggest the possibility of using the Oncoflash® module in thyroid scintigraphy to reduce acquisition time or injected activity without impacting interpretation.     

Meniscal lesion detection and characterization in adult knee MRI: A deep learning model approach with external validation

PurposeEvaluation of a deep learning approach for the detection of meniscal tears and their characterization (presence/absence of migrated meniscal fragment).MethodsA large annotated adult knee MRI database was built combining medical expertise of radiologists and data scientists’ tools. Coronal and sagittal proton density fat suppressed-weighted images of 11,353 knee MRI examinations (10,401 individual patients) paired with their standardized structured reports were retrospectively collected. After database curation, deep learning models were trained and validated on a subset of 8058 examinations. Algorithm performance was evaluated on a test set of 299 examinations reviewed by 5 musculoskeletal specialists and compared to general radiologists’ reports. External validation was performed using the publicly available MRNet database. Receiver Operating Characteristic (ROC) curves results and Area Under the Curve (AUC) values were obtained on internal and external databases.ResultsA combined architecture of meniscal localization and lesion classification 3D convolutional neural networks reached AUC values of 0.93 (95% CI 0.82, 0.95) for medial and 0.84 (95% CI 0.78, 0.89) for lateral meniscal tear detection, and 0.91 (95% CI 0.87, 0.94) for medial and 0.95 (95% CI 0.92, 0.97) for lateral meniscal tear migration detection. External validation of the combined medial and lateral meniscal tear detection models resulted in an AUC of 0.83 (95% CI 0.75, 0.90) without further training and 0.89 (95% CI 0.82, 0.95) with fine tuning.ConclusionOur deep learning algorithm demonstrated high performance in knee menisci lesion detection and characterization, validated on an external database.     

Exploring Radiologic Criteria for Glioma Grade Classification on the BraTS Dataset

ObjectivesGlioma grading using maching learning on magnetic resonance data is a growing topic. According to the World Health Organization (WHO), the classification of glioma discriminates between low grade gliomas (LGG), grades I, II; and high grade gliomas (HGG), grades III, IV, leading to major issues in oncology for therapeutic management of patients. A well-known dataset for machine-based grade prediction is the MICCAI Brain Tumor Segmentation (BraTS) dataset. However this dataset is not divided into WHO-defined LGG and HGG, since it combines grades I, II and III as “lower grades gliomas”, while its HGG category only presents grade IV glioblastoma multiform. In this paper we want to train a binary grade classifier and investigate the consistency of the original BraTS labels with radiologic criteria using machine-aided predictions.Material and methodsUsing WHO-based radiomic features, we trained a SVM classifier on the BraTS dataset, and used the prediction score histogram to investigate the behaviour of our classifier on the lower grade population. We also asked 5 expert radiologists to annotate BraTS images between low (as opposed to lower) grade and high grade glioma classes, resulting in a new groundtruth.ResultsOur first training reached 84.1% accuracy. The prediction score histogram allows us to identify the radiologically high grade patients among the original lower grade population of the BraTS dataset. Training another SVM on our new radiologically WHO-aligned groundtruth shows robust performances despite important class imbalance, reaching 82.4% accuracy.ConclusionOur results highlight the coherence of radiologic criteria for low grade versus high grade classification under WHO terms. We also show how the histogram of prediction scores and crossed prediction scores can be used as tools for data exploration and performance evaluation. Therefore, we propose to use our radiological groundtruth for future development on binary glioma grading.     

Generation of clinical 177Lu SPECT/CT images based on Monte Carlo simulation with GATE

PurposePatient-specific dosimetry in MRT relies on quantitative imaging, pharmacokinetic assessment and absorbed dose calculation. The DosiTest project was initiated to evaluate the uncertainties associated with each step of the clinical dosimetry workflow through a virtual multicentric clinical trial. This work presents the generation of simulated clinical SPECT datasets based on GATE Monte Carlo modelling with its corresponding experimental CT image, which can subsequently be processed by commercial image workstations.MethodsThis study considers a therapy cycle of 6.85 GBq ¹⁷⁷Lu-labelled DOTATATE derived from an IAEA-Coordinated Research Project (E23005) on “Dosimetry in Radiopharmaceutical therapy for personalised patient treatment”. Patient images were acquired on a GE Infinia-Hawkeye 4 gamma camera using a medium energy (ME) collimator. Simulated SPECT projections were generated based on experimental time points and validated against experimental SPECT projections using flattened profiles and gamma index. The simulated projections were then incorporated into the patient SPECT/CT DICOM envelopes for processing and their reconstruction within a commercial image workstation.ResultsGamma index passing rate (2% − 1 pixel criteria) between 95 and 98% and average gamma between 0.28 and 0.35 among different time points revealed high similarity between simulated and experimental images. Image reconstruction of the simulated projections was successful on HERMES and Xeleris workstations, a major step forward for the initiation of a multicentric virtual clinical dosimetry trial based on simulated SPECT/CT images.ConclusionsRealistic ¹⁷⁷Lu patient SPECT projections were generated in GATE. These modelled datasets will be circulated to different clinical departments to perform dosimetry in order to assess the uncertainties in the entire dosimetric chain.     

A cloud-fog software architecture for dental CBCT dose monitoring using the DICOM structured report: Automated establishment of DRL

PurposeThe diagnostic reference level (DRL) has been established to optimize the diagnostic methods and reduce radiation dose during radiographic examinations. The aim of this study was to present a completely new solution based on Cloud-Fog software architecture for automatic establishment of the DRL values during dental cone-beam computed tomography (CBCT) according to digital imaging and communications in medicine (DICOM) structured reports.Methods and MaterialsA Cloud-Fog software architecture was used for automatic data handling. This architecture used the DICOM structured reports as a source for extracting the required information by fog devices in the imaging center. These devices transferred the derived information to the cloud server. The cloud server calculated the value of indication-based DRL in dental CBCT imaging based upon the parameters and adequate quantities of the absorbed dose. The feedback of DRL value was continuously announced to the imaging centers in 6 phases. In each phase, the level of the dose was optimized in imaging centers.ResultsThe DRL value was established for 5-specific indications, including third molar teeth (511 mGy.cm²), implant (719 mGy.cm²), form and position anomalies of the tooth (408 mGy.cm²), dentoalveolar pathologies (612 mGy.cm²), and endodontics (632 mGy.cm²). The determination of the DRL value in each phase revealed a downward trend until stabilization.ConclusionThe new solution presented in this study makes it possible to calculate and update the DRL value nationally and automatically among all centers. Also, the results showed that this approach is successful in establishing stabilized DRL values.     

TomoEQA: Dose verification for patient-specific quality assurance in helical tomotherapy using an exit detector

PurposeExisting phantom-less quality assurance (QA) platforms does not provide patient-specific QA for helical tomotherapy (HT). A new system, called TomoEQA, is presented to facilitate this using the leaf open time (LOT) of a binary multi-leaf collimator, as measured by an exit detector.MethodsTomoEQA was designed to provide measurement-based LOTs based on detector data and to generate a new digital imaging and communication in medicine (DICOM) dataset that includes the measured LOTs for use by secondary check platforms. To evaluate the system, 20 patient-specific QAs were performed using the program in Mobius3D software, and the results were compared to conventional phantom-based QA results.ResultsFrom our assessment, most of the differences between the planned and measured (or calculated) data, excluding one case, were within the acceptance criteria comparing with those of conventional QA. Regarding the gamma analysis, all results considered in this study were within the acceptance criteria. In addition, the developed system was performed for a failed case and showed approximately the same trends as the conventional approach.ConclusionsTomoEQA could perform patient-specific QAs of HT using Mobius3D and provide reliable patient-specific QAs results by evaluating point dose errors and 3D gamma passing rates. TomoEQA could also distinguish whether an intensity-modulated radiation therapy plan failed or not.