Radiological Technology Research Articles

Aiming for Japanese Society of Radiological Technology (JSRT) That Spreads Its Wings to the World

Aiming for Japanese Society of Radiological Technology (JSRT) That Spreads Its Wings to the World

A deep unsupervised saliency model for lung segmentation in chest X-ray images

Medical image segmentation is a critical and laborious step in the diagnosis workflow by separating regions of organs or anomalies to be analyzed further by a physician, or classified automatically by an intelligent tool. Deep machine learning techniques have achieved impressive results in image segmentation tasks whenever immense amounts of labeled data are provided. However, having big amounts of manually supervised image data is labor intensive, and costly impractical. Unsupervised learning, contrariwise, can provide results using unlabeled images by reaching decisions, clustering, and deriving insights from the data only. We propose a deep unsupervised learning approach to perform lung segmentation in chest X-ray images, which is guided by heuristic saliency maps as a noisy labeling framework, which through a composite loss of reconstruction and regularization reach saliency estimation as final segmentation. Tests were performed using public chest X-ray images from the Japanese Society of Radiological Technology (JSRT), and from Montgomery County (MC) datasets. The proposed model based on saliency scored a dice coefficient of 0.87, against a state-of-the-art unsupervised method, which scored 0.68, on JSRT, and on the MC dataset 0.88, and 0.64, respectively. Qualitatively comparing, the saliency model worked better in reducing background noise. Considering that unsupervised models can be used as robust preparation steps, and as baselines for further different tasks, either guided by constant feedback from physicians, or as input to semi-supervised learning, the results have shown this to be a promise line of research to medical image segmentation problems.

An Institutional Approach to Fight Technologist Burnout: Radiologist Contribution to National Radiologic Technology Week and Other Efforts

Radiologic technologists are an integral part of the function of the radiology department and are at risk of burnout with increasing volumes and demands. Integration of the technologists into departmental wellness efforts may prove to be important to technologist retention and job satisfaction. We describe a single institution's efforts to support the technologists in our department, which started as a spontaneous thank you note drive and has grown into an annual effort to celebrate National Radiologic Technology Week and a broader inclusion of the technologists in departmental wellness.

A multicentric, single arm, prospective, stratified clinical investigation to evaluate MammoWave's ability in breast lesions detection.

Microwave imaging is a safe and promising new technology in breast radiology, avoiding discomfort of breast compression and usage of ionizing radiation. This paper presents the first prospective microwave breast imaging study during which both symptomatic and asymptomatic subjects were recruited. Specifically, a prospective multicentre international clinical trial was performed in 2020-2021, to investigate the capability of a microwave imaging device (MammoWave) in allowing distinction between breasts with no radiological finding (NF) and breasts with radiological findings (WF), i.e., with benign or malignant lesions. Each breast scan was performed with the volunteers lying on a dedicated examination table in a comfortable prone position. MammoWave output was compared to reference standard (i.e., radiologic study obtained within the last month and integrated with histological one if available and deemed necessary by responsible investigator) to classify breasts into NF/WF categories. MammoWave output consists of a selection of microwave images' features (determined prior to trials' start), which allow distinction between NF and WF breasts (using statistical significance p<0.05). 353 women were enrolled in the study (mean age 51 years ± 12 [SD], minimum age 19, maximum age 78); MammoWave data from the first 15 women of each site, all with NF breasts, were used for calibration. Following central assessor evaluation, 111 NF (48 dense) and 272 WF (136 dense) breasts were used for comparison with MammoWave output. 272 WF comprised 182 benign findings and 90 malignant histology-confirmed cancer. A sensitivity of 82.3% was achieved (95%CI: 0.78-0.87); sensitivity is maintained when limiting the investigation to histology-confirmed breasts cancer only (90 histology-confirmed breasts cancer have been included in this analysis, having sizes ranging from 3 mm to 60 mm). Specificity value of approximately 50% was achieved as expected, since thresholds were calculated (for each feature) using median value obtained after recruiting the first 15 women (of each site), all NF. This prospective trial may represent another step for introducing microwave imaging into clinical practice, for helping in breast lesion identification in asymptomatic women.

Speech recognition technology in radiology

Speech recognition devices are promising tools for the healthcare system. Speech recognition technology has had a relatively long history of use in Western healthcare systems since the 1970s. However, it became widely used at the beginning of the 21st century, replacing medical transcriptionists. This technology is relatively new in home healthcare. Its active development began only in the early 2010s, and its implementation in healthcare started in late 2010. This delay is due to the idiosyncrasies of the Russian language and the limited computational power present at the beginning of the 21st century.&#x0D; Currently, complexes of devices and software for speech recognition are used in the voice filling of medical records and can reduce the time for preparing reports for radiological examinations compared with traditional (keyboard) text input.&#x0D; The literature review provides a brief history of speech recognition technology development and application in radiography. Key scientific studies showing its efficacy in Western healthcare systems are reflected. Voice recognition technology in the home is demonstrated, and its effectiveness is evaluated. The prospects for further development of this technology in Russian healthcare are described.

How do Trained and Prospective Physiotherapists and Radiologic Technologists Face Knee Joint Effusion Profession-Specifically and Interdisciplinary?-A Cross-Sectional Study

Objective: Knee osteoarthritis (OA) is often accompanied by joint effusion that requires accurate detecting and diagnosing. In an interprofessional evidence-based practice (EBP) approach to this, physiotherapists (PTs) and radiologic technologists (RTs) should be included. While researchissults regarding the 1st pillar (external clinical evidence) and the 2nd (patients´side) pillar of EBP are available, these are still missing concerning the 3rd pillar (clinical experts´ side). Therefore, the aim of this study is to investigate how PTs and RTs deal with knee joint effusion detection in the course of knee OA patients’ profession-specifically as well as how they rate the other occupational group´s competencies. Methods: An online survey consisting of open and closed questions was conducted in the period from 31st of January 2023 till 6th of March 2023 to collect data. PTs, RTs as well as physiotherapy students (PTSs) and radiologic technology students (RTSs) were defined as target groups. A structuring content analysis was used to evaluate open questions. Descriptive and inferential statistics were used to analyses closed questions. Results: PTs and PTSs determined at least one of the cardinal symptoms of inflammation, inspection, palpation, the patellar tap test, and circumference measurement as knee effusion detection possibilities in patients´ with knee OA. In contrast, the modified stroke test and the bulge sign were hardly mentioned. The majority of RTs and RTSs stated that MRI was the first-choice imaging to detect effusion followed by sonography. Focusing on specific signs in radiologic imaging modalities fluid on T2 sequences in MRI, hypoechoic area on sonography, swelling on conventional radiography, fluid accumulation on CT and radionuclide uptake on scintigraphy were frequently mentioned. Among all 4 participating occupational groups, PTs and RTs were among the top 5 mentioned professions that have knee effusion detection competencies. There is no significant difference regarding the other occupational group´s knee effusion detection competencies. The participants rated them consistently as “rather competent” most frequently. Conclusion: The authors study results indicate that there are some discrepancies among PTs, PTSs, RTs and RTSs in regard to the 1st and 3rd pillar of EBP. All occupational groups show partly contradictory or lack of in-depth knowledge regarding frequently applied and suggested diagnostic procedures in literature. Implementing relevant training formats should be considered in the future.

What is the margin of error of surgeons and radiological imaging in diagnosing acute appendicitis?

Acute appendicitis is the most common emergency requiring surgical intervention in general surgery. Negative appendectomy is defined as the removal of a pathologically normal appendix. In this study, we aimed to show our negative appendectomy rate. This study was carried out among 2990 patients who were operated on for appendicitis between 2015-2020 at the Health Sciences University, İstanbul Kanuni Sultan Suleyman Training, and Research Hospital. Accrual and historical records of the patients were analyzed using NCSS (Number Cruncher Statistical System) 2007 Statistical Software (Utah, USA) package program. The results were evaluated at the significance level of P < 0.05. The mean age of all patients was 33. Of the patients, 1011 were women and 1979 were men. 27 of the women patients were pregnant. We requested a blood test (WBC count) and an abdominal ultrasound for all our patients who came with the complaint of abdominal pain in the right lower quadrant. Negative appendectomy was performed with ultrasonography in 622 patients with pathological diagnoses of lymphoid hyperplasia and fibrous obliteration (20.8%). We had abdominal computerized tomography (CT) for 285 patients and abdominal magnetic resonance imaging (MRI) for 16 of 27 pregnant women (59.25%) due to unclear clinical picture. Diagnostic laparoscopy was performed in 36 of the patients who underwent CT and 4 of the patients who underwent MRI since the diagnosis could not be made. We performed unnecessary appendectomy in 21.2% of the patients. With the increasing clinical follow-up experience of surgeons and developing technology in radiology, our aim is to minimize the negative appendectomy rate as much as possible.

Message from the New Representative Director: Toward Progress and Academic Development of the Radiological Technology

Message from the New Representative Director: Toward Progress and Academic Development of the Radiological Technology

Analysis of association between vitamin D metabolite concentration in blood plasma and pancreatic cancer. Experience of one center

Introduction. An early detection of malignant neoplasms is of great practical importance for improvement of treatment results.&#x0D; Objective. To analyze the levels of dehydrocholecalciferol in patients with different forms of malignant neoplasms of the pancreas and chronic pancrestitis.&#x0D; Materials and methods. The cohort study enclosed the data from 160 patients, who were examined and treated in the Academician A.M. Granov Russian Scientific Center of Radiology and Surgical Technologies for the period from 2018 to 2022. The main group included 89 patients with morphologically confirmed diagnosis of the pancreatic cancer without metastases. The control group included 71 patient with the confirmed diagnosis of chronic pancreatitis. Vitamin D concentration was controlled prior to treatment.&#x0D; Results. The size of the tumor node with the location of pathological process in the head of the pancreas varied from 2.1 to 6.3 cm. In the subgroup of patients with tumor lesion of the head of the pancreas, decompression of biliary ducts due to mechanical jaundice was implemented in 43 (82.6 %) persons. When comparing the indicators of 25(ОН)D concentration, in 36 (69.2 %) patients of the main group, different degrees of fall of vitamin D concentration in blood plasma were found. Most often a pronounced deficit of vitamin D was diagnosed in patients with the tumor localized in the head of the pancreas (8 (15.9 % patients). Out of 37 patients with localization of the tumor node in the body/tail, 25 (67.5 %) persons demonstrated normal values of blood plasma vitamin D and 12 (13.4 %) different degrees of decrease in vitamin D concentration. A marked deficit of vitamin D was diagnosed in 2 (16.6 %) patients, deficit of vitamin D in 7 (58.3 %), insufficiency in 3 (25 %). In the control group, decrease in vitamin D level was determine in 12 (16.9 %) patients. An apparent deficit of vitamin D was diagnosed in 3 (4.2 %) patients, deficit of vitamin D in 4 (5.6 %), and insufficiency of vitamin D in 7 (9.8 %); normal values were registered in 57 (80.2 %) persons.&#x0D; Conclusions. Taking into account the obtained data regarding a reliable association between the deficit and insufficiency of vitamin D and the cancer of the head of the pancreas as well as literature data, it is worthwhile to use this indicator as a marker of early detection of pancreatic cancer in risk group patients.

Prognostic value of anti- and proangiogenic factors in severe preeclampsia

BACKGROUND:Over the years, preeclampsia has remained a common pathology of pregnancy. Its severe form poses a significant threat to the health of the woman and the fetus. To assess the risk of developing preeclampsia, biochemical screening for placental growth factor and soluble tyrosine kinase-1 is carried out. The soluble glycoprotein endoglin, which regulates angiogenesis by influencing cell activation, adhesion, and migration, can supplement the idea of the imbalance of vascular factors in severe preeclampsia. A holistic study of anti- and proangiogenic factors can significantly complement the diagnosis of preeclampsia at the preclinical stage and a look at the course and treatment of this pregnancy complication.&#x0D; AIM:The aim of this work was to evaluate the concentrations of anti- and proangiogenic factors (soluble glycoprotein endoglin, soluble tyrosine kinase-1, placental growth factor) and compare the obtained data with the clinical manifestations of severe preeclampsia.&#x0D; MATERIALS AND METHODS:This case-control study included 81 pregnant women. The main group consisted of patients with severe preeclampsia (n= 41), while the control group comprised individuals with normal pregnancy (n= 40). To determine soluble glycoprotein endoglin levels in biological fluids, we used a new ELISA kit developed in the Laboratory of Hybridoma Technology, Academician A.M. Granov Russian Research Center for Radiology and Surgical Technologies (Saint Petersburg, Russia). placental growth factor and soluble tyrosine kinase-1 concentrations were assessed by electrochemiluminescence immunoassay using commercial Roche Diagnostics kits. Statistical analysis was performed using the StatTech v.3.0.6 program (Stattech Ltd., Russia).&#x0D; RESULTS:When analyzing the concentrations of anti- and proangiogenic factors (soluble glycoprotein endoglin in the blood serum and urine, soluble tyrosine kinase-1, placental growth factor and the ratio of these factors), depending on the study group, statistically significant results were obtained (p 0.05). The diagnostic significance of serum and urine soluble glycoprotein endoglin levels in predicting the probability of severe preeclampsia was assessed by the method of ROC curve analysis. Statistically significant differences (p 0.05) were found in the analysis of the concentrations of anti- and pro-angiogenic factors depending on the clinical manifestations such as systolic blood pressure 160 mm Hg (except for placental growth factor level), a multiple increase in urine soluble glycoprotein endoglin level with massive proteinuria, an increase in soluble glycoprotein endoglin levels in biological fluids with an increase in edematous syndrome. Correlation analysis revealed a significant relationship between anti- and proangiogenic factors and the term of delivery, as well as urine soluble glycoprotein endoglin level and the duration of the treatment.&#x0D; CONCLUSIONS:The data obtained complement the concept of severe preeclampsia. Knowing the concentrations of soluble glycoprotein endoglin in the blood serum and urine allows for predicting the timing of delivery and the possible duration of treatment, which is critical in the management of patients with severe preeclampsia.

Новые подходы к диагностическим информационным системам в радиологии.

The first electronic information systems to help compile radiological reports appeared in the mid-1960s and were introduced primarily in radiology departments of healthcare. Due to the development and achievements in the field of information technology in radiology, a system of archiving and transmission of images (Picture Archiving and Communication System (PACS)) was introduced, this helped to reduce the stages of work processes in the routine practice of a radiologist in comparison with the use of classical systems based on films and paper. Also, the use of PACS has led to an increase in the effective work of the radiation diagnostics department by speeding up the acquisition of images for diagnosticians and clinicians, reducing the costs of the department, speeding up the writing of research protocols, increasing the number of studies to conduct and describe. The use of modern information systems has a number of advantages, storing a large amount of data, combining and storing disparate information systems of a medical organization, obtaining full access to the electronic medical record of the patient, which in turn accelerates work processes and effective interaction of hospital staff. Thus, dependence on these information systems by radiation diagnosticians and doctors of other specializations can affect the diagnosis and improvement of the quality of patient care. This review contains up-to-date information about the work of information systems in radiology departments.

Biomechanical effects of clear aligners with different thicknesses and gingival-margin morphology for appliance design optimization

The objectives of this study were to investigate the biomechanical effects of clear aligners (CAs) with various thermoplastic material thicknesses and gingival-margin designs for space closure in extraction treatment and to propose a computer-aided procedure to optimize CA design. The radiologic and intraoral scanning technology, invitro mechanical experiment, viscoelastic modeling, and finite element analysis (FEA) were integrated to establish an orthodontic simulation model. Twelve FEA models of CA were created, comprising combinations of 2 kinds of thicknesses (0.75 and 0.50 mm), 2 forms of gingival-margin shape (scalloped and straight), and 3 types of margin height (-2, 0, and 2 mm). Invitro testing was carried out to determine the actual properties of material thickness. A 0.75-mm-thick aligner resulted in greater periodontal ligament (PDL) stress than 0.50 mm, and there was no clear correlation between the control ability of tooth movement and the thickness. For different margin designs, PDL stress at -2 mm height was significantly lower than those with a higher border. Aligners with straight margins had higher stress than the scalloped aligners, whereas the differences were unnoticeable at 2 mm height. The optimized aligner with differential margin designs was recommended on the basis of biomechanical calculations, which facilitated the efficiency and control of tooth movement for multiple teeth. The effect of material thickness and margin design of CA on the force and movement differed in different teeth. Preferable CA designs of each tooth during different movement stages should be presented personalized under the guidance of precise biomechanics instead of pure morphologic analysis.

An Invisible Skin Marker for External Beam Radiation Therapy: Utilization of Ultraviolet Fluorescent Marker Pens.

Radiation therapy plays an important role in cancer treatment along with surgery and systemic therapy. The total dose of radiation therapy is divided into small doses, and the treatment is typically delivered once a day. The total treatment period can need several weeks or more, and it is necessary to deliver the radiation dose to the target volume within the patient precisely each time. Therefore, the reproducibility of patient positioning is essential for the precision of the dose delivery. Although radiological technologies such as image-guided radiation therapy have also recently been widely used for positioning patients, skin marking is still widely used in many facilities. Skin marking is an inexpensive and universal positioning technique in patients undergoing radiation therapy; however, it is considered a major source of psychological stress. We proposethe use of fluorescent ink pens, which are invisible in standard room lighting, as skin markers for radiotherapy. The primary technique of fluorescence emission is widely employed in molecular biological experiments and for assessing cleaning protocols for infection control. This technique may reduce the stress induced by skin markings during radiotherapy.

An End-to-End Data-Adaptive Pancreas Segmentation System with an Image Quality Control Toolbox

With the development of radiology and computer technology, diagnosis by medical imaging is heading toward precision and automation. Due to complex anatomy around the pancreatic tissue and high demands for clinical experience, the assisted pancreas segmentation system will greatly promote clinical efficiency. However, the existing segmentation model suffers from poor generalization among images from multiple hospitals. In this paper, we propose an end-to-end data-adaptive pancreas segmentation system to tackle the problems of lack of annotations and model generalizability. The system employs adversarial learning to transfer features from labeled domains to unlabeled domains, seeking a dynamic balance between domain discrimination and unsupervised segmentation. The image quality control toolbox is embedded in the system, which standardizes image quality in terms of intensity, field of view, and so on, to decrease heterogeneity among image domains. In addition, the system implements a data-adaptive process end-to-end without complex operations by doctors. The experiments are conducted on an annotated public dataset and an unannotated in-hospital dataset. The results indicate that after data adaptation, the segmentation performance measured by the dice similarity coefficient on unlabeled images improves from 58.79% to 75.43%, with a gain of 16.64%. Furthermore, the system preserves quantitatively structured information such as the pancreas’ size and volume, as well as objective and accurate visualized images, which assists clinicians in diagnosing and formulating treatment plans in a timely and accurate manner.

Predictors of radiation-induced skin reactions in patients with T-cell skin lymphomas after total skin irradiation: retrospective cohort study

INTRODUCTION: Skin reactions negatively affect a patient’s quality of life and may require radiation therapy premature interruption, what impairs illness control. Numerous agents determine tissue radiosensitivity, but demographic and lifestyle factors remain poorly understood. Those kind of patients herapeutic algorithm demands studies.OBJECTIVE: Radiation-induced skin reactions (RISR) formation and severity prognostic factors determination for primary malignant skin lymphomas patients.MATERIALS AND METHODS: 45 patients malignant T-cell skin lymphoma study, aged 38 to 82 years (29 mycosis fungoides patients, 2 lymphomatoid papulosis, 14 patients with primary anaplastic large cell skin lymphoma) received fast electrons total skin irradiation on the basis of the Federal State Institution «Russian Radiology and Surgical Technologies Scientific Center named after Academician A.M.Granov» of the Ministry of Health of the Russian Federation from 2012 to June 2022. Radiation therapy performed in the usual fractionation mode on linear electron accelerators Elekta Precise with 4–6 MeV bremsstrahlung total doses from 14 to 30 Gy (22±2.4 Gy). In order to identify the role of possible prognostic factors of skin toxicity — gender, age, body mass index (BMI), smoking status, lymphoma body surface area (BSA), the presence of concomitant diseases (diabetes mellitus and arterial hypertension) and previous chemotherapy.Statistics: Statistical analysis was performed using the IBM SPSS Statistics v. 26 program (developed by IBM Corporation). The value of p&lt;0.05 was considered statistically significant. Verification of the sample belonging to the normal general population was carried out using the Shapiro–Wilk criterion. Spearman's rank correlation coefficient was used for correlation analysis. The regression parameters were estimated using the least squares method.RESULTS: The analysis revealed positive associations between the severity of RISR and predictor factors: BMI (r=0.799, p&lt;0.01), smoking status (r=0.655, p&lt;0.01), age (r=0.534, p&lt;0.05), presence of diabetes mellitus (r=0.456, p&lt;0.05), arterial hypertension (r=0.227, p&lt;0.05), previous chemotherapy (r=0.422, p&lt;0.01) and lymphoma body surface area (r=0.378, p&lt;0.01). Severity of RISR scale value increase by those parameters. Negative links found between the severity of RISR and the total focal dose (SOD) of the beginning of their formation (r=–0.109, p&lt;0.05).CONCLUSION: Formation and severity prognostic factors for the RCC upon fast electrons total skin exposure for patients with primary malignant T-cell skin lymphomas (mycosis fungoides, lymphomatoid papulosis, primary anaplastic large cell lymphoma of the skin) were identified.

Radioligand therapy with drugs based on radionuclide &lt;sup&gt;225&lt;/sup&gt;Aс: experience of A.M.Granov Russian scientific center of radiology and surgical technologies

INTRODUCTION: Peptide receptor therapy based on the 225Ac radionuclide, due to the targeted effect on tumor cells, is the most promising method of treatment for patients with generalized forms of neuroendocrine tumors and castration-resistant prostate cancer.OBJECTIVE: The aim of the work was to evaluate the effectiveness of radioligand therapy of medicinal radiopharmaceuticals with 225Ac.MATERIALS AND METHODS: Radioligand therapy based on 225Ac radionuclide was performed in 7 patients with generalized prostate cancer and 3 patients with neuroendocrine tumors.RESULTS: Systemic radiation therapy with 225Ac-PSMA 617 and 225Ac-DOTATE showed its effectiveness in clinical practice — patients have a decrease in the level of accumulation of the radiopharmaceutical (RP) in pathological foci on PET-CT with 68Ga-PSMA and 68Ga-DOTATE; a decrease in the level of prostate-specific antigen (PSA) in the blood (for prostate cancer); reduction of manifestations of asthenic, pain and carcinoid (for neuroendocrine tumors) syndromes. Dosimetric studies make it possible to assess the biodistribution of a radiopharmaceutical drug in the body, accumulation in target organs and lesions.CONCLUSION: The appointment of this type of treatment requires an integrated approach based on a thorough study of medical records, assessment of the receptor status of the tumor tissue, timely detection and prediction of adverse reactions.

Opportunities and challenges of digital world of multimodal interventional breast radiology (review)

The article presents the evolution of the introduction of digital technologies into the oncomammoscreening system, multimodal technologies of interventional radiology, statistics on the growth of malignant breast neoplasms in the world and Russia, shows the advantages of digitalization of radiology, interdisciplinary integration with omix systems biology technologies, which contributes to the development of a new direction of radiogenomics, and also points out the existing problems of introducing the latest technologies of interventional radiology, the need to prepare multimodal specialists. The purpose of the study . To show the possibilities of modern digital radiological non-invasive and invasive technologies, as well as the need for interdisciplinary integration and training of multimodal specialists for the early detection and determination of the phenotype of breast cancer, increasing the effectiveness of choosing adequate treatment tactics aimed at increasing the duration and quality of life of patients. Materials and methods. There are given literary data on development of radiological digital methods of diagnostics of breast diseases, including technologies of interventional radiology. New data are also presented on the direction of radiogenomics, born due to the interdisciplinary integration of radiological digital systems, interventional technologies and high-tech systems biology. Results. The review presents the development of equipment and radiological methods for early recognition of breast diseases, based on technical progress, including the active introduction of digitalization and informatization in medicine. New models of artificial intelligence act as a physician assistant in screening. Advantages of various versions of interventional radiology technologies in oncomammoscreening are shown to increase accuracy of pathomorphological diagnostics and determination of tumor phenotype. The latest data on feasibility of interdisciplinary integration of «computer vision» on signs of medical imaging with capabilities of systems biology in determination of tumor phenotype with correlation validity in 71% are given. Conclusions. The undeniable advantages of digitalization and the new capabilities of interventional radiology in detecting the earliest forms of diseases in an interdisciplinary format open the prospect of high-precision diagnostics and an adequate choice of organ-saving therapeutic tactics. The training of a multimodal specialist – radiologist – radiologist-diagnostics and interventional radiologist, who has a wide range of methods of radiological diagnostics, including invasive interventions, requires correction of organizational forms of work, new educational programs not only in the specialty, but also in the basics of digitalization, which will ensure the rational and effective use of modern achievements of science and technology.

Procurement of artificial intelligence for radiology practice

The development of artificial intelligence (AI) technology for radiology has accelerated in the past decade, but its deployment in radiology practices has been slow. We take a sociotechnical approach and suggest that the limited use of AI in radiology practices can be attributed to a recurring tension between planned and emergent change. The paper contributes with a conceptualization and understanding of the tension during the procurement of AI for radiology. To balance this tension, we suggest that health organizations need to redefine the concept and scope of traditional procurement projects, with well-defined goals and project time. Instead, we propose that health organizations need to conceptualize their procurement and implementation projects of AI technology as evolving change processes. The study is based on an interpretive research approach and informed by the Information Infrastructure framework. Empirically, we study the procurement of AI solutions for radiology at a large health trust in Norway.

Recent Topics on Radiological Physics and Technology

Recent Topics on Radiological Physics and Technology

Suppressing Chest Radiograph Ribs for Improving Lung Nodule Visibility by using Circular Window Adaptive Median Outlier (CWAMO)

Chest radiograph ribs obstruct lung nodules. To see the nodule under the chest radiograph ribs, remove or suppress them. The paper describes a circular median filter approach for finding outliers in chest radiographs. The method uses 147 Japanese Society of Radiological Technology x-ray pictures (JSRT). Pixels with intensities two standard deviations above the median are median outliers. Contrast-Limited Adaptive Histogram Equalization enhances nodule visibility (CLAHE). The method is tested on modest chest radiographs and compared to the Budapest University Bone Shadow Eliminated X-Ray Dataset methodology. The initial test uses 50 modest chest radiographs (Test 1). The proposed approach is applied after active shape modelling (ASM) lung segmentation. True positive nodules are seen on 89% of chest radiographs of various subtleties. Test-2 and Test-3 used 20 subtlety-level photos. In Test-2, the peak signal-to-noise ratio (PSNR), mean-to-standard deviation ratio (MSR), and universal image quality index (IQI) are evaluated for the full image and compared to the existing algorithm. For all three parameters, the suggested technique outperforms the algorithm. Test-3 computes nodule MSR and compares it to Budapest University's Bone Shadow Eliminated Dataset and original chest radiographs. The new algorithm improved nodule area contrast by 3.83% and 23.94% compared to the original chest radiograph. This approach improves chest radiograph nodule visualization.