A dual-labeled dataset and fusion model for automatic teeth segmentation, numbering, and state assessment on panoramic radiographs.

BACKGROUND: Recently, deep learning has been increasingly applied in the field of dentistry. The aim of this study is to develop a model for the automatic segmentation, numbering, and state assessment of teeth on panoramic radiographs. METHODS: We created a dual-labeled dataset on panoramic radiographs for training, incorporating both numbering and state labels. We then developed a fusion model that combines a YOLOv9-e instance segmentation model with an EfficientNetv2-l classification model. The instance segmentation model is used for tooth segmentation and numbering, whereas the classification model is used for state evaluation. The final prediction results integrate tooth position, numbering, and state information. The model's output includes result visualization and automatic report generation. RESULTS: Precision, Recall, mAP50 (mean Average Precision), and mAP50-95 for the tooth instance segmentation task are 0.989, 0.955, 0.975, and 0.840, respectively. Precision, Recall, Specificity, and F1 Score for the tooth classification task are 0.943, 0.933, 0.985, and 0.936, respectively. CONCLUSIONS: This fusion model is the first to integrate automatic dental segmentation, numbering, and state assessment. It provides highly accurate results, including detailed visualizations and automated report generation.

The Application of Magnetic Resonance Imaging in the Early and Accurate Diagnosis of Hip Joint Avascular Necrosis.

Introduction Avascular necrosis (AVN) is characterized by the death (necrosis) of cellular bone components in the subchondral bone or epiphysis due to a lack of or an interruption of the blood supply. In routine practice, AVN is most frequently encountered in the femoral head. In this study, we aim to evaluate the application of magnetic resonance imaging (MRI) in the early and accurate diagnosis of hip joint AVN. Materials and methods This was a retrospective, cross-sectional study conducted in the Department of Radiodiagnosis of Dr. D. Y. Patil tertiary care hospital, Pimpri, Pune, India. We studied 30 patients with complaints of pain and associated limping who underwent primary radiograph analysis of the hip joint, followed by MRI. Results We assessed 30 patients (45 hip joints) using plain radiography and MRI. Of the 45 hips, we could diagnose AVN in 28 hips (62.2%) using plain radiography, but we could not diagnose it in 17 hips (37.8%), whereas we were able to diagnose AVN in all hips (100%) using MRI. Forty percent of the patients (n = 12) were on steroids, 26.7% (n = 8) were chronic alcoholics, and 16.7% (n = 5) were idiopathic. The other less common causes were a history of trauma or fracture of the neck of the femur (n = 3) and sickle cell disease (n = 2). Of the 45 hips of the 30 patients studied, 15 patients had bilateral disease affecting a total of 30 hips (66.7%), and 15 patients had unilateral disease affecting a total of 15 hips (33.4%). Of the 30 hips (bilateral disease), five (13.3%) contralateral hips were clinically occult and were incidentally diagnosed with AVN. Conclusion The assessment of AVN based solely on plain radiography can miss vital information in stages II and III (Ficat and Arlet classification). Due to its multiplanar capability, superior spatial resolution, and better tissue characterization, MRI is very sensitive and able to detect femoral head AVN early and promptly in cases that are radiograph-negative or otherwise clinically unsuspected.

Evaluation of X-ray Beam Collimation in Adult Chest Radiography.

BACKGROUND:  Restricting the irradiated volume can reduce X-ray scattering incidents on the image receptor. Proper X-ray collimation during medical imaging reduces a patient's dose while improving image quality. Even though the patient radiation dose due to chest X-ray imaging is low, the 'as low as reasonably achievable' (ALARA) principle should be satisfied, especially for young patients. PURPOSE: To evaluate the accuracy of collimation in digital chest radiography. MATERIALS AND METHODS: Ninety-eight chest radiographs were studied retrospectively from February 2021 to December 2021. Chest images were collected from three main centers in the Madinah region of Saudi Arabia. The ratio of the field of interest area to the field of view (FOV) was measured and calculated to determine the accuracy of X-ray beam collimation. RESULTS: Out of 98 chest radiographs enrolled in the study, 87.8% (n=80) were adequately collimated, while 12.2% (n=18) were rejected due to inadequate collimation. The ratio of the field of interest collimated area of chest radiographs was 0.547, which indicated an acceptable value. Among the three centers, Center 2 showed higher, significant, adequate collimation than Center 1 (P<0.001) and Center 3 (P=0.007). There was a significant gender difference in collimation levels as the level of collimation of female chest radiographs is inferior to that of males (P>0.001). CONCLUSIONS: The collimation of chest radiographs among the three centers was adequate. Based on the study findings, the X-ray beam collimation was sufficient, indicating good optimization and no unnecessary radiation exposure to patients and staff. The collimation of chest radiographs in females was significantly inferior to that of males.

A multi-task learning model for clinically interpretable sesamoiditis grading.

Sesamoiditis is a common equine disease with varying severity, leading to increased injury risks and performance degradation in horses. Accurate grading of sesamoiditis is crucial for effective treatment. Although deep learning-based approaches for grading sesamoiditis show promise, they remain underexplored and often lack clinical interpretability. To address this issue, we propose a novel, clinically interpretable multi-task learning model that integrates clinical knowledge with machine learning. The proposed model employs a dual-branch decoder to simultaneously perform sesamoiditis grading and vascular channel segmentation. Feature fusion is utilized to transfer knowledge between these tasks, enabling the identification of subtle radiographic variations. Additionally, our model generates a diagnostic report that, along with the vascular channel mask, serves as an explanation of the model's grading decisions, thereby increasing the transparency of the decision-making process. We validate our model on two datasets, demonstrating its superior performance compared to state-of-the-art models in terms of accuracy and generalization. This study provides a foundational framework for the interpretable grading of similar diseases.

Periodontitis diagnosis: A review of current and future trends in artificial intelligence.

BACKGROUND: Artificial intelligence (AI) acts as the state-of-the-art in periodontitis diagnosis in dentistry. Current diagnostic challenges include errors due to a lack of experienced dentists, limited time for radiograph analysis, and mandatory reporting, impacting care quality, cost, and efficiency. OBJECTIVE: This review aims to evaluate the current and future trends in AI for diagnosing periodontitis. METHODS: A thorough literature review was conducted following PRISMA guidelines. We searched databases including PubMed, Scopus, Wiley Online Library, and ScienceDirect for studies published between January 2018 and December 2023. Keywords used in the search included "artificial intelligence," "panoramic radiograph," "periodontitis," "periodontal disease," and "diagnosis." RESULTS: The review included 12 studies from an initial 211 records. These studies used advanced models, particularly convolutional neural networks (CNNs), demonstrating accuracy rates for periodontal bone loss detection ranging from 0.76 to 0.98. Methodologies included deep learning hybrid methods, automated identification systems, and machine learning classifiers, enhancing diagnostic precision and efficiency. CONCLUSIONS: Integrating AI innovations in periodontitis diagnosis enhances diagnostic accuracy and efficiency, providing a robust alternative to conventional methods. These technologies offer quicker, less labor-intensive, and more precise alternatives to classical approaches. Future research should focus on improving AI model reliability and generalizability to ensure widespread clinical adoption.

Artificial Intelligence-Based Surgery Support Model Using Intraoperative Radiographs for Assessing the Acetabular Component Angle.

BACKGROUND: This study aimed to develop an artificial intelligence-based surgical support model for assessing the acetabular component angle using intraoperative radiographs during total hip arthroplasty and verify its accuracy. METHODS: A total of 268 hips were analyzed. At first, 268 preoperative and intraoperative anteroposterior pelvic radiographs were amplified to 536. These radiographs were used to create a learning model to estimate the acetabular component angle from the radiographs intraoperatively. The ground truth was the anteversion and inclination angles obtained from the computed tomography-based navigation system intraoperatively. Bone landmarks on the preoperative and intraoperative radiographs were manually annotated. The distances and angles between each landmark were used as predictor variables. The estimation accuracy was assessed for internal and external test datasets. Mean absolute error (MAE) and R2 values were used as accuracy measures. RESULTS: The MAE and R2 for the internal test set showed 2.19 and 0.850 for anteversion, and 1.18 and 0.805 for inclination, respectively. The MAE and R2 for the external test set showed 2.78 and 0.789 for anteversion, and 1.56 and 0.744 for inclination, respectively. CONCLUSIONS: We developed an artificial intelligence-based surgical support model for accurately assessing the acetabular component angle using intraoperative radiographs. Excellent estimation accuracy was confirmed for the external test set. In the future, the model may help to reduce the risk of adverse postoperative events.

Validation of full-length radiograph based musculoskeletal modeling method to estimate medial and lateral knee contact forces.

BACKGROUND: Anatomical parameters of the pelvis, femur, and tibia derived from the full-length radiograph can be used to create a more accurate musculoskeletal model compared to marker-based linear scaling method. However, whether this model leads to more accurate estimations of medial knee contact force (MCF) and lateral knee contact force (LCF) than marker-based linear scaling method is still unknown. RESEARCH QUESTION: This main purpose of this study was to determine whether musculoskeletal model generated from full-length radiograph improves the estimations of MCF and LCF. METHODS: An open-source dataset including marker trajectories, ground reaction forces, in vivo knee contact forces, and full-length radiograph was used to evaluate the accuracy of full-length radiograph musculoskeletal modeling method. Subject-specific musculoskeletal models were created using anatomical parameters derived from the full-length radiograph or marker-based linear scaling methods. MCF and LCF were estimated using musculoskeletal simulations of normal walking trails. The accuracy of modeling methods was determined by comparing the estimated and in vivo measured MCF and LCF. RESULTS: Compared to the marker-based linear scaling approach, the full-length radiograph musculoskeletal modeling method exhibited decreases of 38.3 % and 41.3 % in root mean square error for MCF and LCF respectively, as well as reductions of 50.0 % and 49.3 % in mean peak errors for MCF and LCF respectively. SIGNIFICANCE: The full-length radiograph musculoskeletal modeling method provides a more accurate way to estimate MCF and LCF compared to the traditional maker-based linear scaling approach, which may contribute to understand the initiation, progression, and treatment of OA.

The radiographic diversity of dental patterns among 7219 young individuals-a contribution to disaster victim identification.

The diversity of dental patterns is a fundamental topic in disaster victim identification. The current scientific literature, however, is scarce of data regarding young individuals. This study aimed to assess the radiographic diversity of dental patterns, considering missing, unrestored, and filled teeth in young individuals. The sample consisted of 7219 panoramic radiographs of individuals between 12 and 22.9 years. The permanent teeth, except third molars, were coded as missing, unrestored, or filled and odds ratios (OR) were calculated based on sex, dental arch, and age. The sex-combined sample had 1.116 distinctive dental patterns. "All unrestored" teeth was the most common pattern (OR: 0.437) followed by the sequence of unrestored teeth except restored mandibular first molars (OR: 0.021). Females had more distinctive dental patterns than males (p < .001), while males had more unrestored teeth (p < .001). In the age category of 12-12.9 years, the OR for finding a distinctive dental pattern was 11%, while in the age category of 22-22.9 years it increased to 58%. On the other hand, the OR for "all unrestored" gradually decreased according to age (74% in the younger category, and 23% in the older age category). The distinctiveness of dental patterns among young individuals is affected by the predominance of unrestored teeth. However, registering a single filled tooth in a remaining unrestored dentition can reduce exponentially the probability of finding an identical pattern of missing, unrestored and filled teeth.

Resolution of Chronic Migraine Headaches and Improvement in Cervical Spine Kyphosis Following Chiropractic BioPhysics® (CBP®) Treatment: A Case Report With a Seven-Month Follow-Up.

We present a chronic migraine (CM) patient demonstrating significant improvement in subjective and objective reported outcome measures with deeper cervical lordosis parameters and reduced forward head posture on radiographs. A 29-year-old male suffered from CM reporting significant pain and disability with aural, sensory, and motor disturbances during the migraine headaches. Aura with visual disturbances, abnormal facial and extremity sensation, sporadic motor weakness, and other signs of CM were found in the patient's history since age 10. The patient reported previous physical therapy, manual chiropractic, and over-the-counter medications. Migraine-specific prescriptions without long-term reduction in pain and disability were reported. The pain and suffering had been reported to be worsening, and he sought Chiropractic BioPhysics® (CBP®) spine and postural rehabilitation protocols. These protocols were used to increase cervical lordosis, reduce coronal imbalances, increase mobility, and create better posture. These protocols include specific prescriptions based on radiography for postural exercises, postural mirror image® (MI®)traction, and specific spinal manipulative therapy (SMT) focused on posture. All outcome measures improved with the resolution of all initial symptoms of CM. There was a 16° improvement in cervical lordosis, a 30% decrease in headache disability, and additional improvements. These improvements were maintained at a seven-month follow-up during which the patient received infrequent maintenance treatments. This successful treatment of a patient with CM with long-term follow-up adds to evidence that CBP® spinal structural rehabilitation may prove effective and serve as a possible tool for clinicians, physicians, and therapists to treat CM.

Longitudinal observation of distal tibial degeneration in varus ankle osteoarthritis using plain radiograph.

BACKGROUND: This study aimed to explore the natural history of distal tibial deformity in varus ankle osteoarthritis and to identify risk factors for its progression. METHODS: This retrospective longitudinal observational study included 97 ankles with moderate to severe varus ankle osteoarthritis, monitored for > 2 years. Radiographic assessments were conducted on weightbearing plain radiographs at the initial and final visit. RESULTS: Significant tibial anterior surface angle reduction (p = 0.008), talus center medial migration (p = 0.038), and tibial lateral surface angle increase (p = 0.034) occurred. Multiple logistic analysis revealed an association between progression and coronal talar tilt angle at the first visit (odds ratio 1.17, p = 0.015) CONCLUSION: The progression of radiographic distal tibial deformity in varus ankle osteoarthritis was clarified, identifying the coronal talar tilt angle as a risk factor. LEVEL OF EVIDENCE: Level IV, retrospective longitudinal observation study.

Development and validation of deep learning models for bowel obstruction on plain abdominal radiograph.

OBJECTIVE: Artificial intelligence (AI) could help medical practitioners in analyzing radiological images to determine the presence and site of bowel obstruction. This retrospective diagnostic study proposed a series of deep learning (DL) models for diagnosing bowel obstruction on abdominal radiograph. METHODS: A total of 2082 upright plain abdominal radiographs were retrospectively collected from four hospitals. The images were labeled as normal, small bowel obstruction and large bowel obstruction by three senior radiologists based on comprehensive examinations and interventions within 48 hours after admission. Gradient-weighted class activation mapping was used to visualize the inferential explanation. RESULTS: In the validation set, the Xception-backboned model achieved the highest accuracy (0.863), surpassing the VGG16 (0.847) and ResNet models (0.836). In the test set, the Xception model (accuracy: 0.807) outperformed other models and a junior radiologist (0.780) but not a senior radiologist (0.840). In the AI-aided diagnostic framework, the junior and senior radiologists made their judgements while aware of the Xception model predictions. Their accuracy significantly improved to 0.887 and 0.913, respectively. CONCLUSIONS: We developed and validated DL-based computer vision models for diagnosing bowel obstruction on plain abdominal radiograph. DL-based computer-aided diagnostic systems could reduce medical practitioners' workloads and improve diagnostic accuracy.

Imaging and Surgical Procedures After Surface-Related Collegiate Football Injuries on Artificial Turf Versus Natural Grass: Prevalence and Trends Over 15 Seasons.

Background: Few longitudinal studies exist comparing medical procedures after football injuries on artificial and natural grass surfaces. Purpose/Hypothesis: The purpose of this study was to specifically compare imaging and surgical procedures after surface-related collegiate football injuries on artificial turf versus natural grass. It was hypothesized that there would be no difference in the incidence of imaging and surgical procedures, combined medical procedures, or combined substantial and severe injuries over time between these surfaces. Study Design: Cohort study; Level of evidence, 2. Methods: A total of 39 universities across all Football Bowl Subdivision conferences were evaluated over 15 seasons (2006-2020). Playing surfaces evaluated were either a heavyweight artificial turf infill system (≥9.0 lb infill/ft2) or natural grass. Outcomes of interest included medical procedures across injury category, primary injury type, injury location, and specific procedures. Data involved multivariate analyses of variance (MANOVA) and Wilks λ criteria using general linear model procedures and were expressed as medical procedure incidence rates (IRs) per 10-game season. Results: Overall, 2224 games were documented: 1106 (49.7%) on artificial turf and 1118 (50.3%) on natural grass. Of the 9137 total injuries reported, 4010 (44%) were surface-related cases. MANOVA indicated significant main effects between surfaces according to imaging procedure (F 2,1738 = 4.718; P = .009), surgical procedure (F 1,539 = 5.974; P = .003), and medical diagnosis (F 2,456 = 2.643; P = .040). Post hoc analyses indicated significantly lower (P < .05) incidences on artificial turf versus natural grass, respectively, for the following outcomes: imaging procedures ordered after player-to-surface impact trauma (IR [95% CI], 1.5 [1.3-1.8] vs 2.1 [1.9-2.4]), shoe-to-surface trauma during physical contact (4.4 [4.1-4.7] vs 5.2 [4.9-5.5]), foot injuries (0.6 [0.5-0.7] vs 1.1 [0.9-1.2]), ligament sprains/tears (4.6 [4.3-4.9] vs 5.3 [5.0-5.6]), muscle strains/tears (0.2 [0.2-0.3] vs 0.7 [0.6-0.9]), surgeries performed after shoe-to-surface trauma during physical contact (1.1 [0.9-1.3] vs 1.6 [1.4-1.8]), lower body surgeries (1.8 [1.6-2.0] vs 2.3 [2.1-2.6]), and surgeries involving ligament tears (1.0 [0.9-1.2] vs 1.5 [1.3-1.7]), as well as fewer diagnoses of syndesmosis sprains/tears (0.7 [0.5-0.8] vs 1.0 [0.8-1.2]) and Lisfranc trauma (0.3 [0.2-0.4] vs 0.5 [0.4-0.7]). Trends over the 15 seasons indicated a significant rise in combined medical procedures (P = .005) and combined substantial and severe injuries (P = .0007) irrespective of surface. Conclusion: Results indicated that collegiate football competition on heavyweight artificial turf resulted in lower incidences of imaging and surgical procedures and medical diagnoses compared with natural grass.

Identifying marginal adaptation discrepancies of lithium disilicate crowns using seven different vertical X-ray angulations.

PURPOSE: To study the effect of different vertical angulations on the ability to radiographically assess vertical marginal discrepancies of lithium disilicate crowns. MATERIALS AND METHODS: Twenty-one lithium disilicate crowns were fabricated for three different prepared natural teeth: incisor, canine, and premolar. Vertical marginal discrepancies ranging from 0 to 300 µm were intentionally created. The seated crowns were radiographed using seven different vertical angulations, totaling 147 images. Thirty experienced evaluators scored each image for marginal discrepancy, and values were statistically analyzed. RESULTS: Significant differences in the ability to accurately assess marginal discrepancies from radiographs were observed for the study factors of angulation, tooth type, and degree of marginal discrepancy (p < 0.001). CONCLUSIONS: The radiographic interpretation of the marginal discrepancies of lithium disilicate crowns is significantly affected by the dimension of the marginal discrepancy. Specifically on premolar crowns, it is significantly affected by different vertical angulations of the X-ray beam. When evaluating marginal discrepancy on lithium disilicate crowns radiographically, vertical beam angulation within ±10° to the cemento-enamel junctionCEJ plane is recommended.

Back to basics: lung volumes on pediatric chest X-rays-pitfalls and diagnostic implications.

Evaluating lung "volumes" when interpreting pediatric chest radiographs is useful for supporting the diagnosis of lung pathology in children. Lung volumes can be estimated using semi-objective criteria such as anterior rib counting. Estimating lung volumes is one advantage plain radiographs have over other imaging modalities such as ultrasound or even cross-sectional imaging (Computed Tomography and Magnetic Resonance Imaging, where volumes can be calculated, but remain difficult to quickly be appreciated subjectively or semi-objectively in clinical practice). Lung hyperinflation is often a surrogate of air-trapping and may be suggestive of diseases such as bronchiolitis. Identifying under-inspired chest radiographs is also important to avoid misinterpreting findings related to low lung volumes as pathology, which is a significant diagnostic pitfall in children. This pictorial review will demonstrate the method for evaluating lung volumes, specifically for identifying lung hyperinflation as a surrogate of air-trapping and under-inspiration for avoiding false positive diagnoses.

Posterior Tibial Slope Measured on Plain Radiograph Versus MRI and Its Association With Revision Anterior Cruciate Ligament Reconstruction: A Matched Case-Control Study.

BACKGROUND: Posterior tibial slope (PTS) has been identified as a possible modifiable risk factor for anterior cruciate ligament (ACL) injury and ACL reconstruction (ACLR) failure. However, the findings in the literature are inconsistent. This may be explained by several different reasons, including different measurement methods, differing definitions of ACLR failure, and possible inclusion of suboptimal films. PURPOSE: To compare PTS values obtained using plain radiographs (XR-PTS) in a young (≤21 years of age), skeletally mature patient population with those obtained using magnetic resonance imaging (MRI), as well as to quantify the number of suboptimal lateral knee radiographs obtained across an integrated health care system and determine the potential effect of including these radiographs on summary statistics of XR-PTS. STUDY DESIGN: Case-control study; Level of evidence, 3. METHODS: Skeletally mature adolescent and young adult patients ≤21 years of age were identified from the ACLR registry of Kaiser Permanente. The cases of concern were patients requiring revision ACLR. The controls were patients who had an ACLR and did not require a revision procedure. The XR-PTS measurements were made on plain radiographs by a single blinded reviewer. These results were compared with measurements obtained using MRI. The quality of each plain radiograph was evaluated by measuring posterior/distal femoral condylar overlap and length of tibial diaphysis captured on the radiograph. Summary statistics with and without inclusion of measurements made on suboptimal radiographs were calculated. RESULTS: Of the initial 634 patients with ACLR (317 case-control pairs), 561 (88.5%) had radiographs available and were included for the analysis comparing radiograph to MRI slope measurements. For the evaluation of slope between case and control pairs with radiograph information available, there were 257 case-control pairs; there were 124 pairs when those with suboptimal radiographs were excluded. There was no difference in MRI-measured lateral tibial posterior slope or medial tibial posterior slope for the 257 case-control pairs with XR-PTS information and for the 124 pairs with optimal radiographs. XR-PTS in the revision cohort was significantly steeper than in the control group when suboptimal radiographs were included in the analysis. There was no difference when patients with suboptimal radiographs were excluded. PTS measurements made on plain radiographs were larger than those made on MRI. There was a poor correlation between measurements made using these 2 modalities (r = 0.22 for radiograph and medial PTS). CONCLUSION: This study did not find a significantly steeper XR-PTS in patients who had to undergo revision ACLR when suboptimal radiographs were not included in the analysis. The present study's results confirmed the findings from a previous study of the same patient population that used MRI. However, there was poor correlation between PTS measurements made using plain radiograph and MRI.

Functional preoperative assessment of coronal knee laxity better predicts postoperative patient outcomes than intraoperative surgeon-defined laxity in total knee arthroplasty.

PURPOSE: Intraoperative laxity assessments in total knee arthroplasty (TKA) are subjective, with few studies comparing against standardised preoperative and postoperative assessments. This study compares coronal knee laxity in TKA patients awake and anaesthetised, preprosthesis and postprosthesis implantation, evaluating relationships to patient-reported outcome measures. METHODS: A retrospective analysis of 49 TKA joints included preoperative and postoperative computed tomography scans, stress radiographs and knee injury and osteoarthritis outcome score (KOOS) questionnaire results preoperatively and 12 months postoperatively. The imaging was used to assess functional laxity (FL) in awake patients, whereas computer navigation measured intraoperative surgical laxity (SL) preimplantation and postimplantation, with patients anaesthetised. Varus and valgus stress states and their difference, joint laxity, were measured. RESULTS: SL was greater than FL in both preimplantation [8.1° (interquartile range, IQR 2.0°) and 3.8° (IQR 2.9°), respectively] and postimplantation [3.5° (IQR 2.3°) and 2.5° (IQR 2.7°), respectively]. Preimplantation, SL was more likely than FL to categorise knees as correctable to ±3° of the mechanical axis. Preoperative FL correlated with KOOS Symptoms (r = 0.33, p = .02) and quality of life (QOL) (r = 0.38, p = .01), whereas reducing medial laxity with TKA enhanced postoperative QOL outcomes (p = .02). CONCLUSIONS: Functional coronal knee laxity assessment of awake patients is generally lower than intraoperative surgical assessments of anaesthetised patients. Preoperative SL may result in overcorrection of coronal TKA alignment, whereas preoperative FL better predicts postoperative patient outcomes and reflects the patients' native and tolerable knee laxity. Preoperative FL assessment can be used to guide surgical planning. LEVEL OF EVIDENCE: Level II.

Comparison between artificial intelligence solution and radiologist for the detection of pelvic, hip and extremity fractures on radiographs in adult using CT as standard of reference.

PURPOSE: The purpose of this study was to compare the diagnostic performance of an artificial intelligence (AI) solution for the detection of fractures of pelvic, proximal femur or extremity fractures in adults with radiologist interpretation of radiographs, using standard dose CT examination as the standard of reference. MATERIALS AND METHODS: This retrospective study included 94 adult patients with suspected bone fractures who underwent a standard dose CT examination and radiographs of the pelvis and/or hip and extremities at our institution between January 2022 and August 2023. For all patients, an AI solution was used retrospectively on the radiographs to detect and localize bone fractures of the pelvis and/or hip and extremities. Results of the AI solution were compared to the reading of each radiograph by a radiologist using McNemar test. The results of standard dose CT examination as interpreted by a senior radiologist were used as the standard of reference. RESULT: A total of 94 patients (63 women; mean age, 56.4 ± 22.5 [standard deviation] years) were included. Forty-seven patients had at least one fracture, and a total of 71 fractures were deemed present using the standard of reference (25 hand/wrist, 16 pelvis, 30 foot/ankle). Using the standard of reference, the analysis of radiographs by the AI solution resulted in 58 true positive, 13 false negative, 33 true negative and 15 false positive findings, yielding 82 % sensitivity (58/71; 95 % confidence interval [CI]: 71-89 %), 69 % specificity (33/48; 95 % CI: 55-80 %), and 76 % accuracy (91/119; 95 % CI: 69-84 %). Using the standard of reference, the reading of the radiologist resulted in 65 true positive, 6 false negative, 42 true negative and 6 false positive findings, yielding 92 % sensitivity (65/71; 95 % CI: 82-96 %), 88 % specificity (42/48; 95 % CI: 75-94 %), and 90 % accuracy (107/119; 95 % CI: 85-95 %). The radiologist outperformed the AI solution in terms of sensitivity (P = 0.045), specificity (P = 0.016), and accuracy (P < 0.001). CONCLUSION: In this study, the radiologist outperformed the AI solution for the diagnosis of pelvic, hip and extremity fractures of the using radiographs. This raises the question of whether a strong standard of reference for evaluating AI solutions should be used in future studies comparing AI and human reading in fracture detection using radiographs.

A Comparison of Two Methods of Dental Age Estimation in a Population of Saudi Children and Adolescents.

This study aimed to compare and evaluate the accuracy of the Demirjian (DE) and the London Atlas (LAE) dental age estimation methods in a Saudi population sample. This retrospective cross-sectional study used digital radiographs from electronic health records in three different dental institutes. In total, 357 male and 354 female (ages 5-15 years) digital orthopantomograms were selected for age estimation. The mean difference between the chronological age (CA) and age estimation method among males and females was 0.03 ± 0.34 and 0.00 ± 0.34, respectively, for LAE and 0.55 ± 0.84 and 0.76 ± 0.51, respectively, for DE. The mean difference between the LAE and DE methods among males and females was 0.52 ± 0.89 and -0.76 ± 0.57, respectively. No statistically significant difference between CA and LAE was found in either males (p = 0.079) or females (p = 0.872). A statistically significant difference was found between CA and DE in both genders (p < 0.001). A statistically significant difference was found between the LAE and DE groups (p < 0.001) in both genders. An overestimation of dental age was observed with DE compared with that in CA. LAE showed higher accuracy than CA, with no clinically significant difference. Although the difference between the LAE and DE methods was insignificant, the LAE method proved to be more accurate.