Search Results (88)

Background: Necrotizing enterocolitis (NEC) is a severe gastrointestinal disease and a major cause of morbidity and mortality among preterm infants. Traditional diagnostic methods such as abdominal radiography have limited sensitivity in early disease stages, prompting interest in bowel ultrasound (BUS) as a complementary imaging modality. Objective: This scoping review aims to synthesize existing literature on the role of ultra sound in the early detection, diagnosis, and management of NEC, with emphasis on its diagnostic performance, integration into clinical care, and technological innovations. Methods: Following PRISMA-ScR guidelines, a systematic search was conducted across PubMed, Embase, Cochrane Library, and Google Scholar for studies published between January 2000 and December 2025. Inclusion criteria encompassed original research, reviews, and clinical studies evaluating the use of bowel, intestinal, or Doppler ultrasound in neonates with suspected or confirmed NEC. Data were extracted, categorized by study design, population characteristics, ultrasound features, and diagnostic outcomes, and qualitatively synthesized. Results: A total of 101 studies were included. BUS demonstrated superior sensitivity over radiography in detecting early features of NEC, including bowel wall thickening, portal venous gas, and altered peristalsis. Doppler ultrasound, both antenatal and postnatal, was effective in identifying perfusion deficits predictive of NEC onset. Neonatologist-performed ultrasound (NEOBUS) showed high interobserver agreement when standardized protocols were used. Emerging tools such as ultra-high-frequency ultrasound (UHFUS) and artificial intelligence (AI)-enhanced analysis hold potential to improve diagnostic precision. Point-of-care ultrasound (POCUS) appears feasible in resource-limited settings, though implementation barriers remain. Conclusions: Bowel ultrasound is a valuable adjunct to conventional imaging in NEC diagnosis. Standardized protocols, validation of advanced technologies, and out come-based studies are essential to guide its broader clinical adoption. Full article

Two geriatric (>9 years old) dogs presented with vomiting, depression, and anorexia and were diagnosed with extrahepatic biliary obstruction (EHBO) secondary to ruptured gallbladder mucoceles. Diagnosis was based on serum biochemical analysis, abdominal radiography, and ultrasonography, which revealed gallbladder rupture, peritonitis, and common bile duct dilation. Both dogs underwent emergency surgical intervention involving cholecystectomy and choledochal stent placement in the common bile duct without cholecystojejunostomy or cholecystoduodenostomy. Postoperatively, the clinical symptoms and serum chemistry values improved, and both dogs survived without recurrence for over one year. These cases demonstrate that choledochal stenting can be an effective adjunct to cholecystectomy for managing EHBO in dogs with ruptured gallbladder mucoceles, potentially preventing reocclusion and promoting recovery, especially when histopathological evaluation is not feasible in clinical settings. However, persistent elevation of liver enzymes may occur postoperatively, necessitating prolonged monitoring and medical management in some cases. Full article

This study aimed to evaluate the validity of panoramic radiography as an auxiliary method for identifying mandibular bone features consistent with a diagnosis of osteopenia or osteoporosis. Ninety panoramic radiographs were analyzed to assess the quality of the mandibular cortical layer below the mental foramen on both sides of the mandible. Scores C1 (normal), C2 (osteopenia), and C3 (osteoporosis) were attributed according to the cortical morphology. The sample consisted of 78 (86%) women aged 45 years or older and 12 (14%) men older than 60 years old. In 39 (43%) cases, the C1 score was evidenced as the lower mandibular cortical layer was normal on the image. The C2 score was identified in 47 (52%) cases, in which the cortical layer showed semilunar defects. Four (5%) cases presented a C3 score, with the cortical layer showing a clearly porous, thinner bone cortex. The presence of risk behaviors (e.g., smoking and alcoholism) as well as some comorbidities (e.g., systemic arterial hypertension, diabetes mellitus and thyroid disorders) was also observed. Mandibular bone changes were observed in association with a set of risk factors using panoramic radiography. Full article

Medical imaging is crucial for diagnosing, monitoring, and treating medical conditions. The medical reports of radiology images are the primary medium through which medical professionals can attest to their findings, but their writing is time-consuming and requires specialized clinical expertise. Therefore, the automated generation of radiography reports has the potential to improve and standardize patient care and significantly reduce the workload of clinicians. Through our work, we have designed and evaluated an end-to-end transformer-based method to generate accurate and factually complete radiology reports for X-ray images. Additionally, we are the first to introduce curriculum learning for end-to-end transformers in medical imaging and demonstrate its impact in obtaining improved performance. The experiments were conducted using the MIMIC-CXR-JPG database, the largest available chest X-ray dataset. The results obtained are comparable with the current state of the art on the natural language generation (NLG) metrics BLEU and ROUGE-L, while setting new state-of-the-art results on F1 examples-averaged F1-macro and F1-micro metrics for clinical accuracy and on the METEOR metric widely used for NLG. Full article

This study aimed to demonstrate the Convolutional Neural Network (CNN) algorithm’s efficiency in detecting alveolar bone loss using panoramic radiographs. The comparison was evaluated among 1874 pictures retrieved from an institution, from which the training set included 953 showing bone loss and 921 normal cases. A confusion matrix was performed for statistical analysis. The CNN method correctly identified 92 out of 100 bone loss cases and 89 out of 100 healthy cases. The model showed a sensitivity of 0.8327, a specificity of 0.8683, a precision of 0.8918, an accuracy of 0.8927, and an F1 score of 0.8615 in detecting bone loss. This study concluded that a faster CNN model may be used as an adjuvant technique to diagnose periodontal disease and alveolar bone loss using dental panoramic radiography images, thereby minimizing diagnostic effort, and saving assessment time. However, the execution of precisely detecting periodontal cases by fully automated AI models using panoramic radiographs appears imminent and needs clinical periodontal evaluation for definitive diagnosis. The suitability of this approach is supported by the sensitivity, specificity, accuracy, and F-measure, which showed satisfactory performance for classifying cases. Based on population and periodontal disease burden standpoint, the use of AI in diagnosing periodontal diseases may serve as an excellent surveillance method to classify alveolar bone loss. Monitoring a periodontal patient after treatment needs a wide area to cover by AI-based diagnostic modality. With AI as the future of dentistry, performance-based clinical usage of CNN models demands confirmed practical application by dentists. Full article

In this paper, to address the issue of the unknown influence of activation functions on casting defect detection using convolutional neural networks (CNNs), we designed five sets of experiments to investigate how different activation functions affect the performance of casting defect detection. Specifically, the study employs five activation functions—Rectified Linear Unit (ReLU), Exponential Linear Units (ELU), Softplus, Sigmoid Linear Unit (SiLU), and Mish—each with distinct characteristics, based on the YOLOv8 algorithm. The results indicate that the Mish activation function yields the best performance in casting defect detection, achieving an mAP@0.5 value of 90.1%. In contrast, the Softplus activation function performs the worst, with an mAP@0.5 value of only 86.7%. The analysis of the feature maps shows that the Mish activation function enables the output of negative values, thereby enhancing the model’s ability to differentiate features and improving its overall expressive power, which enhances the model’s ability to identify various types of casting defects. Finally, gradient class activation maps (Grad-CAM) are used to visualize the important pixel regions in the casting digital radiography (DR) images processed by the neural network. The results demonstrate that the Mish activation function improves the model’s focus on grayscale-changing regions in the image, thereby enhancing detection accuracy. Full article

X-ray imaging protocols in radiology departments often exhibit variability in exposure parameters and geometric setups, leading to inconsistencies in image quality and potential variations in patient dose. Objectives: This study aimed to harmonise and optimise chest and pelvis X-ray imaging protocols by standardising exposure parameters and geometric setups across departmental systems, minimising radiation dose while ensuring adequate image quality for accurate diagnosis. Methods: The image quality of five pelvic and three chest protocols across different radiographic systems was evaluated both quantitatively and visually. Visual image quality for both chest and pelvis protocols was assessed by radiologists and radiographers using the Visual Grading Analysis (VGA) method. Additionally, the quantitative image quality figure inverse (IQF_inv) metric for all protocols was determined using the CDRAD image quality phantom. Moreover, the patient radiation dose for both chest and pelvis protocols was evaluated using dose area product (DAP) values measured by the systems’ built-in DAP metres. Results: Different quantitative image quality and radiation dose to patients were achieved in various protocol settings for both chest and pelvis examinations, but the visual image quality assessment showed satisfactory image quality for all observers in both the pelvis and chest protocols. The selected protocols for harmonising chest radiography across all imaging systems result in reduced radiation exposure for patients while maintaining adequate image quality compared to the previously used system-specific protocol. Conclusions: The clinical protocol for chest and pelvis radiography has been standardised and optimised in accordance with patient radiation exposure and image quality. This approach aligns with the ALARA (As Low As Reasonably Achievable) principle, ensuring optimal diagnostic information while minimising the radiation risks. Full article

Background/Objectives: This study quantitatively evaluates the novel Intelligent Noise Reduction (INR) software NE 3.10.0.15 across three chest radiography protocols, namely, physical anti-scatter grid, non-grid, and virtual anti-scatter grid, to optimise the patient radiation dose while maintaining sufficient image quality. Methods: Quantitative image quality and radiation dose were evaluated using a CDRAD phantom with 20 cm PMMA to simulate the patient across three chest protocol settings at INR levels of 0, 5, and 8 for both PA and LAT projections. Effective doses were estimated using PCXMC Monte Carlo simulation software 2.0. Results: The findings revealed significant improvements in image quality with increasing INR levels, with INR8 consistently outperforming INR5 and non-INR settings. Protocols employing virtual or no grid achieved substantial radiation dose reductions of 77–82% compared to the physical grid. The virtual grid did enhance the quantitative image quality by 6–9% compared to non-grid configurations. Conclusions: INR software, particularly when combined with virtual anti-scatter grids, offers a promising solution for improving image quality while significantly reducing the patient radiation dose in chest radiography. Future clinical validation, incorporating subjective visual assessments by radiologists, is recommended to confirm these findings and facilitate the integration of INR closer to clinical practice. Full article

Background: Chest radiography remains a cornerstone of thoracic imaging; however, the influence of patient demographics and technical factors on radiologic findings is not yet fully understood. This study investigates how gender, age, and radiographic projection affect the prevalence of three common findings: atelectasis, pleural effusion, and pulmonary nodules. Methods: We analyzed 112,120 frontal chest radiographs from the publicly available NIH ChestX-ray8 dataset. Gender-specific prevalence rates were compared using chi-square tests and unadjusted odds ratios (ORs). Multivariable logistic regression was performed to assess the independent effects of gender, age, and projection (posteroanterior [PA] vs. anteroposterior [AP]), including interaction terms. Results: Atelectasis and nodules were more prevalent in male patients, with unadjusted rates of 10.9% and 5.8% versus 9.5% and 5.4% in females. While the difference in nodules’ prevalence was not statistically significant (OR = 0.96, p = 0.16), multivariable analysis showed a significant association (adjusted OR = 1.378, 95% CI 1.157–1.641; p = 0.0003). Pleural effusion showed no significant gender difference (11.7% vs. 12.1%; OR 0.97, p = 0.10). Increasing age was consistently associated with higher odds of all findings (ORs per year: 1.016–1.012; all p < 0.0001). The PA view reduced the odds of atelectasis (OR 0.59) and effusion (OR 0.60), but increased the odds of detecting nodules (OR 1.31). Interaction terms showed minor but statistically significant gender-related differences in age effects. Conclusions: Gender, age, and radiographic projection substantially affect the radiographic detection of frequently observed thoracic abnormalities. These findings are directly relevant for improving clinical diagnostic accuracy and for reducing demographic and technical biases in AI-based radiograph interpretation, particularly in critical care and screening settings. Full article

Background: The accuracy of fracture diagnosis through radiographic imaging largely depends on image quality and the interpreter’s experience. In resource-limited settings (minimal-resource settings), imaging quality is often lower than in standard-resource facilities, potentially affecting diagnostic accuracy. Objective: This study aims to compare the diagnostic accuracy of plain radiograph interpretations between minimal-resource and standard-resource methods and assess the influence of interpreter experience on diagnostic precision. Methods: This cross-sectional study is based on secondary data from patients’ medical records at the Dr. Zainoel Abidin General Hospital (RSUDZA) Banda Aceh, Indonesia. Comparisons between minimal-resource and standard-resource interpretations were made and validated using a reference standard (gold standard). Statistical analyses included diagnostic testing, Chi-square tests, and ROC curve analysis to evaluate sensitivity, specificity, and accuracy. Results: The findings indicate that standard-resource radiographs have significantly higher accuracy than minimal-resource radiographs (p < 0.05). Radiologists demonstrated the highest diagnostic accuracy compared to general practitioners and radiology residents. Conclusions: The standard-resource method is superior in detecting fractures compared to the minimal-resource method. Enhancing imaging quality and providing additional training for medical personnel are essential to improve diagnostic accuracy in resource-limited settings. Full article

Background/Objectives: Radiological imaging is crucial in intensive care settings, particularly for the differential diagnosis of fever and sepsis. Computed tomography (CT) is the preferred method for detecting infectious foci in critically ill ICU patients. Methods: This study prospectively analyzed non-ECG-gated chest and abdominal CT scans from ICU patients to assess CT’s diagnostic utility. Data from prior imaging modalities (CT, radiography, MRI, ultrasound), microbiological assays (blood cultures, bronchoalveolar lavage, urinalysis), and enzymatic profiles (transaminases, pancreatic enzymes) were included. The predictive value of clinical and laboratory parameters was evaluated via correlation analysis. Results: A total of 112 patients were evaluated, with 99 exhibiting 147 inflammatory foci (92 thoracic, 55 abdominal). Definitive diagnoses were made in 58.5% of cases, while 41.5% remained classified as possible. Prior diagnostic procedures identified inflammatory origins in 57.1% of cases. Fewer CT-detected foci were observed in patients with bronchial asthma or type 2 diabetes mellitus (p = 0.049 and p = 0.006). Conclusions: CT imaging plays a central role in identifying infectious foci in ICU patients with unexplained syndromes, particularly in the thoracic region. CT scanning is recommended for sepsis management when other diagnostic evidence is lacking. Conditions such as bronchial asthma or diabetes mellitus may prompt earlier suspicion of infectious foci due to elevated inflammatory markers. Full article

Chest imaging plays a pivotal role in screening and monitoring patients, and various predictive artificial intelligence (AI) models have been developed in support of this. However, little is known about the effect of decreasing the radiation dose and, thus, image quality on AI performance. This study aims to design a low-dose simulation and evaluate the effect of this simulation on the performance of CNNs in plain chest radiography. Seven pathology labels and corresponding images from Medical Information Mart for Intensive Care datasets were used to train AI models at two spatial resolutions. These 14 models were tested using the original images, 50% and 75% low-dose simulations. We compared the area under the receiver operator characteristic (AUROC) of the original images and both simulations using DeLong testing. The average absolute change in AUROC related to simulated dose reduction for both resolutions was <0.005, and none exceeded a change of 0.014. Of the 28 test sets, 6 were significantly different. An assessment of predictions, performed through the splitting of the data by gender and patient positioning, showed a similar trend. The effect of simulated dose reductions on CNN performance, although significant in 6 of 28 cases, has minimal clinical impact. The effect of patient positioning exceeds that of dose reduction. Full article

Background/Objectives: This study aims to evaluate the accuracy of digital dental radiography in determining working length in root canal treatment via various image modules. Methods: A total of 40 intact single-rooted, single-canal human premolar teeth were examined. Following meticulous cleaning, the teeth were placed in a maxillary premolar socket within a dry human skull. X-ray images were systematically captured via a Carestream RVG digital sensor under standardized conditions. The digital images are processed under five distinct image modules: 1. original module, 2. autoenhancement module, 3. autoenhancement+negative module, 4. negative module, 5. colored module. Three calibrated observers determined the working length of each digital radiograph after the specified image modules were applied. The agreement between the actual working length and the lengths determined by the observers was evaluated via Pearson correlation analysis. A significance level of 0.05 was set for the statistical tests. Results: A high level of intra- and interobserver agreement, with a strong correlation between the actual measurements and all image module groups, was obtained (p < 0.001). The original image module group demonstrated the highest compatibility (ICC = 0.940, r = 0.912), whereas the colored image module group exhibited the lowest compatibility (ICC = 0.924, r = 0.879) with the actual measurement. Conclusions: This study demonstrates the accuracy of digital radiography in determining working length through the application of various image modules, with the original image module exhibiting the closest alignment to actual working lengths. These findings support the continued use and further development of computer-based image processing tools to optimize clinical outcomes in root canal therapy. Full article

Background: The global burden of respiratory diseases such as influenza, tuberculosis, and viral pneumonia necessitates rapid, accurate diagnostic tools to improve healthcare responses. Current methods, including RT-PCR and chest radiography, face limitations in accuracy, speed, accessibility, and cost-effectiveness, especially in resource-constrained settings, often delaying treatment and increasing transmission. Methods: This study introduces an Enhanced Multi-Model Deep Learning (EMDL) approach to address these challenges. EMDL integrates an ensemble of five pre-trained deep learning models (VGG-16, VGG-19, ResNet, AlexNet, and GoogleNet) with advanced image preprocessing (histogram equalization and contrast enhancement) and a novel multi-stage feature selection and optimization pipeline (PCA, SelectKBest, Binary Particle Swarm Optimization (BPSO), and Binary Grey Wolf Optimization (BGWO)). Results: Evaluated on two independent chest X-ray datasets, EMDL achieved high accuracy in the multiclass classification of influenza, pneumonia, and tuberculosis. The combined image enhancement and feature optimization strategies significantly improved diagnostic precision and model robustness. Conclusions: The EMDL framework provides a scalable and efficient solution for accurate and accessible pulmonary disease diagnosis, potentially improving treatment efficacy and patient outcomes, particularly in resource-limited settings. Full article

Objective: Pars fractures are a common cause of lower back pain, especially among young individuals. Although computed tomography (CT) and magnetic resonance imaging (MRI) scanning are commonly used in developed regions, traditional radiography remains the main diagnostic method in many developing countries. This study assessed whether the standard radiographic angles suggested in textbooks are optimal for an Asian population since Asian groups have lower lumbar lordosis. This study found a 35° angle to be the most effective angle for lumbar oblique X-ray imaging. Additionally, the potential for a customized positioning auxiliary device was examined to improve image quality and reduce patient discomfort in cost-sensitive healthcare settings like Taiwan’s single-payer system. Methods: A total of 100 participants underwent lumbar oblique radiography using a specially designed footboard with angle markings. Radiologists evaluated 600 images based on waist-to-hip ratio (WHR) and body mass index to identify the optimal angulation for various body types. Results: For individuals with a WHR of 0.85, a 35° angle provided superior image quality, while 45° was more effective for slimmer patients. This optimized approach indicates the cost-effectiveness and diagnostic value of traditional X-ray imaging. Conclusions: The 35° angulation standardizes lumbar X-ray imaging for an Asian cohort, reducing repeat scans and improving accuracy. Using a positioning device further enhances image quality and patient comfort, supporting the clinical utility of traditional radiography in resource-limited environments. Full article