Vertebral fracture severity assessment on anteroposterior radiographs with a new semi-quantitative technique

We developed a new tool to assess the severity of osteoporotic vertebral fracture using radiographs of the spine. Our technique can be used in patient care by helping to stratify patients with osteoporotic vertebral fractures into appropriate treatment pathways. It can also be used for research purposes.

PURPOSE

The aim of our study was to propose a semi-quantitative (SQ) grading scheme for osteoporotic vertebral fracture (OVF) on anteroposterior (AP) radiographs.

METHODS

On AP radiographs, the vertebrae are divided into right and left halves, which are graded (A) vertical rectangle, (B) square, (C) traverse rectangle, and (D) trapezoid; whole vertebrae are graded (E) transverse band or (F) bow-tie. Type A and B were compared with normal and Genant SQ grade 1 OVF, Type C and D with grade 2 OVF, and Type E and F with grade 3 OVF. Spine AP radiographs and lateral radiographs of 50 females were assessed by AP radiographs SQ grading. After training, an experienced board-certified radiologist and a radiology trainee assessed the 50 AP radiographs.

RESULTS

The height-to-width ratio of the half vertebrae varied 1.32-1.48. On lateral radiographs, 84 vertebrae of the 50 patients had OVFs (38 grade 1, 24 grade 2, and 22 grade 3). On AP radiographs, the radiologist correctly assigned 84.2%, 91.7%, and 77.2% and the trainee correctly assigned 68.4%, 79.2%, and 81.8% of grade 1, 2, and 3 OVFs, respectively. Compared with lateral radiographs, the radiologist had a weighted Kappa of 0.944 including normal vertebrae and 0.883 not including normal vertebrae, while the corresponding Kappa values for the trainee were 0.891 and 0.830, respectively.

CONCLUSION

We propose a new semi-quantitative grading system for vertebral fracture severity assessment on AP spine radiographs.

Radiological diagnosis of prevalent osteoporotic vertebral fracture on radiographs: an interim consensus from a group of experts of the ESSR osteoporosis and metabolism subcommittee

When a low-energy trauma induces an acute vertebral fracture (VF) with clinical symptoms, a definitive diagnosis of osteoporotic vertebral fracture (OVF) can be made. Beyond that, a "gold" radiographic standard to distinguish osteoporotic from non-osteoporotic VFs does not exist. Fracture-shaped vertebral deformity (FSVD) is defined as a deformity radiographically indistinguishable from vertebral fracture according to the best of the reading radiologist's knowledge. FSVD is not uncommon among young populations with normal bone strength. FSVD among an older population is called osteoporotic-like vertebral fracture (OLVF) when the FSVD is likely to be associated with compromised bone strength. In more severe grade deformities or when a vertebra is collapsed, OVF diagnosis can be made with a relatively high degree of certainty by experienced readers. In "milder" cases, OVF is often diagnosed based on a high probability rather than an absolute diagnosis. After excluding known mimickers, singular vertebral wedging in older women is statistically most likely an OLVF. For older women, three non-adjacent minimal grade OLVF (< 20% height loss), one minimal grade OLVF and one mild OLVF (20-25% height loss), or one OLVF with ≥ 25% height loss, meet the diagnosis of osteoporosis. For older men, a single OLVF with < 40% height loss may be insufficient to suggest the subject is osteoporotic. Common OLVF differential diagnoses include X-ray projection artifacts and scoliosis, acquired and developmental short vertebrae, osteoarthritic wedging, oncological deformities, deformity due to high-energy trauma VF, lateral hyperosteogeny of a vertebral body, Cupid's bow, and expansive endplate, among others.

Radiographic findings in patients suspected of subacromial impingement syndrome: prevalence and reliability

OBJECTIVE

Aims were to (i) report prevalence and (ii) evaluate reliability of the radiographic findings in examinations of patients suspected of subacromial impingement syndrome (SIS), performed before a patient's first consultation at orthopaedic department.

MATERIALS AND METHODS

This cross-sectional study examined radiographs from 850 patients, age 18 to 63 years, referred to orthopaedic clinic on suspicion of SIS. Prevalence (%) of radiographic findings were registered. Inter- and intrarater reliability was analysed using expected and observed agreement (%), kappa coefficients, Bland-Altman plots, or intraclass coefficients.

RESULTS

A total of 850 patients with a mean age of 48.2 years (SD = 8.8) were included. Prevalence of the radiographic findings was as follows: calcification 24.4%, Bigliani type III (hooked) acromion 15.8%, lateral/medial acromial spurs 11.1%/6.6%, acromioclavicular osteoarthritis 12.0%, and Bankart/Hill-Sachs lesions 7.1%. Inter- and intrarater Kappa values for most radiographic findings ranged between 0.40 and 0.89; highest values for the presence of calcification (0.85 and 0.89) and acromion type (0.63 and 0.66). The inter- and intrarater intraclass coefficients ranged between 0.41 and 0.83; highest values for acromial tilt (0.79 and 0.83) and calcification area (0.69 and 0.81).

CONCLUSION

Calcification, Bigliani type III (hooked) acromion, and acromioclavicular osteoarthritis were prevalent findings among patients seen in orthopaedic departments on suspicion of SIS. Spurs and Bankart/Hill-Sachs lesions were less common. Optimal reliabilities were found for the presence of calcification, calcification area, and acromial tilt. Calcification qualities, acromion type, lateral spur, and acromioclavicular osteoarthritis showed suboptimal reliabilities. Newer architectural measures (acromion index and lateral acromial angle) performed well with respect to reliability.

Temporal and geographical variations in diagnostic imaging in Norway

BACKGROUND

Unwarranted temporal and geographical variations are acknowledged as a profound problem for equal access and justice in the provision of health services. Even more, they challenge the quality, safety, and efficiency of such services. This is highly relevant for imaging services.

OBJECTIVE

To analyse the temporal and geographical variation in the number of diagnostic images in Norway from 2013 to 2021.

METHODS

Data on outpatient imaging provided by the Norwegian Health Economics Administration (HELFO) and inpatient data afforded by fourteen hospital trusts and hospitals in Norway. Data include the total number of imaging examinations according to the Norwegian Classification of Radiological Procedures (NCRP). Analyses were performed with descriptive statistics.

RESULTS

More than 37 million examinations were performed in Norway during 2013-2021 giving an average of 4.2 million examinations per year. In 2021 there was performed and average of 0.8 examinations per person and 2.2 examinations per person for the age group > 80. There was a 9% increase in the total number of examinations from 2013 to 2015 and a small and stable decrease of 0.5% per year from 2015 to 2021 (with the exception of 2020 due to the pandemic). On average 71% of all examinations were outpatient examinations and 32% were conducted at private imaging centres. There were substantial variations between the health regions, with Region South-East having 53.1% more examinations per inhabitant than Region West. The geographical variation was even more outspoken when comparing catchment areas, where Oslo University Hospital Trust had twice as many examinations per inhabitant than Finnmark Hospital Trust.

CONCLUSION

As the population in Norway is homogeneous it is difficult to attribute the variations to socio-economic or demographic factors. Unwarranted and supply-sensitive variations are challenging for healthcare systems where equal access and justice traditionally are core values.

Prediction of postoperative curve correction based on the supine radiographs for adult idiopathic scoliosis (AdIS) patients

BACKGROUND CONTEXT

Preoperative supine radiographs are mandatory in the new adult idiopathic scoliosis (AdIS) classification. Supine radiographs are easily reproducible and highly predictive of side bending radiographs. However, few studies evaluated the use of supine radiographs in predicting postoperative curve correction after posterior spinal fusion (PSF) in AdIS.

PURPOSE

To investigate the use of supine and side bending (SB) radiographs in predicting postoperative curve correction in AdIS patients who underwent PSF.

STUDY DESIGN

Retrospective study.

PATIENT SAMPLE

93 AdIS patients who underwent PSF between 2022 and 2023 were included.

OUTCOME MEASURES

Demographic data were age, gender, height, weight, body mass index (BMI), Risser grade, Lenke curve types and Cobb angles. Main outcome measures were preoperative and immediate postoperative Cobb angle (proximal thoracic [PT], main thoracic [MT] and thoracolumbar/lumbar [TL/L] curves), Supine Cobb angle and Flexibility rate (PT, MT and TL/L), and Correction rate (PT, MT and TL/L).

METHODS

Correlation study was performed between Supine Cobb angle vs. postoperative Cobb angle for PT, MT and TL/L curves. A predictive formula was derived from the correlation plots.

RESULTS

A total of 93 subjects were included in our study with a median age of 24.7 years and comprised of 80 females (86.0%). Preoperative Supine Cobb angle (r=0.835, r=0.881, r=0.767, p<.001) and preoperative SB Cobb angle (r=0.815, r=0.872, r=0.801, p<.001) showed similar strong positive correlation with postoperative PT, MT and TL/L Cobb angle, but preoperative Supine Cobb angle had slightly stronger correlation in PT and MT, whereas preoperative SB Cobb angle had stronger correlation in TL/L curve. Using the derived predictive formulae, there was a significant, strong, positive correlation between the predicted value and actual value of postoperative standing Cobb angle, (r=0.852, p<.001), with 71.0% of the patients had predicted postoperative Cobb angle from the supine radiographs within 5 degrees of the actual value.

CONCLUSION

Both supine radiographs and side bending radiographs had strong predictability of the postoperative Cobb angle for PT, MT and TL/L curves. In 71.0% of patients, the actual postoperative Cobb angle was within 5 degrees of the predicted postoperative Cobb angle using the predictive formulae.

Radiographic Imaging of Community-Acquired Pneumonia: A Case-Based Review

The chest radiograph is the most common imaging examination performed in most radiology departments, and one of the more common indications for these studies is suspected infection. Radiologists must therefore be aware of less common radiographic patterns of pulmonary infection if they are to add value in the interpretation of chest radiographs for this indication. This review uses a case-based format to illustrate a range of imaging findings that can be associated with acute pulmonary infection and highlight findings that should prompt investigation for diseases other than community-acquired pneumonia to prevent misdiagnosis and delays in appropriate management.

The diagnostic performance of impacted third molars in the mandible: A review of deep learning on panoramic radiographs

Background

Mandibular third molar is prone to impaction, resulting in its inability to erupt into the oral cavity. The radiographic examination is required to support the odontectomy of impacted teeth. The use of computer-aided diagnosis based on deep learning is emerging in the field of medical and dentistry with the advancement of artificial intelligence (AI) technology. This review describes the performance and prospects of deep learning for the detection, classification, and evaluation of third molar-mandibular canal relationships on panoramic radiographs.

Methods

This work was conducted using three databases: PubMed, Google Scholar, and Science Direct. Following the literature selection, 49 articles were reviewed, with the 12 main articles discussed in this review.

Results

Several models of deep learning are currently used for segmentation and classification of third molar impaction with or without the combination of other techniques. Deep learning has demonstrated significant diagnostic performance in identifying mandibular impacted third molars (ITM) on panoramic radiographs, with an accuracy range of 78.91% to 90.23%. Meanwhile, the accuracy of deep learning in determining the relationship between ITM and the mandibular canal (MC) ranges from 72.32% to 99%.

Conclusion

Deep learning-based AI with high performance for the detection, classification, and evaluation of the relationship of ITM to the MC using panoramic radiographs has been developed over the past decade. However, deep learning must be improved using large datasets, and the evaluation of diagnostic performance for deep learning models should be aligned with medical diagnostic test protocols. Future studies involving collaboration among oral radiologists, clinicians, and computer scientists are required to identify appropriate AI development models that are accurate, efficient, and applicable to clinical services.

Accuracy and tracing time of cephalometric analyses on a tablet or desktop computer : A prospective study

BACKGROUND

This prospective study aimed to evaluate the influence of the computer type (tablet or desktop) on accuracy and tracing time of cephalometric analyses.

METHODS

Dental students used a web-based application specifically developed for this purpose to perform cephalometric analyses on tablet and desktop computers. Landmark locations and timestamps were exported to measure the accuracy, successful detection rate and tracing time. Reference landmarks were established by six experienced orthodontists. Statistical analysis included reliability assessment, descriptive statistics, and linear mixed effect models.

RESULTS

Over a period of 8 semesters a total of 277 cephalometric analyses by 161 students were included. The interrater reliability of the orthodontists establishing the reference coordinates was excellent (ICC > 0.9). For the students, the mean landmark deviation was 2.05 mm and the successful detection rate for the clinically acceptable threshold of 2 mm suggested in the literature was 68.6%, with large variations among landmarks. No effect of the computer type on accuracy and tracing time of the cephalometric analyses could be found.

CONCLUSION

The use of tablet computers for cephalometric analyses can be recommended.

Validity and reliability of the Group for Learning Useful and Performant Swallowing (GLUPS) tool

INTRODUCTION

To validate the Group for Learning Useful and Performant Swallowing (GLUPS), a clinical tool dedicated to videofluoroscopy swallowing study (VFSS).

METHODS

Forty-five individuals were recruited from January 2022 to March 2023 from the Department of Otolaryngology Head and Neck Surgery of University Hospital Saint-Pierre (Brussels, Belgium). Subjects underwent VFSS, which was rated with GLUPS tool by two blinded otolaryngologists and one speech-therapist. VFSS were rated twice with GLUPS within a 7-day period to assess test-retest reliability.

RESULTS

Twenty-four patients and twenty-one controls completed the evaluations. The internal consistency (α = 0.745) and the test-retest reliability (rs = 0.941; p = 0.001) were adequate. GLUPS reported a high external validity regarding the significant correlation with the Penetration-Aspiration Scale (rs = 0.551; p = 0.001). Internal validity was adequate, because GLUPS score was significant higher in patients compared to controls (6.21 ± 4.42 versus 2.09 ± 2.00; p = 0.001). Interrater reliability did not report significant differences in the GLUPS sub- and total score among the independent judges. The mean GLUPS score of individuals without any evidence of VFSS abnormalities was 2.09/23 (95% CI  1.23-2.95), which supported that a GLUPS score ≥ 3.0 is suggestive of pathological VFSS.

CONCLUSIONS

GLUPS is a clinical instrument documenting the abnormal findings of oral and pharyngeal phases at the VFSS. GLUPS demonstrated high reliability and excellent criterion-based validity. GLUPS may be used in clinical practice for the swallowing evaluation at the VFSS.

AI-based hip prosthesis failure prediction through evolutional radiological indices

INTRODUCTION

This study aimed to develop artificial intelligence models for predicting hip implant failure from radiological features. Analyzing the evolution of the periprosthetic bone and implant's position throughout the entire follow-up period has shown the potential to be more relevant in outcome prediction than simply considering the latest radiographic images. Thus, we investigated an AI-based model employing a small set of evolutional parameters derived from conventional radiological features to predict hip prosthesis failure.

MATERIALS AND METHODS

One hundred sixty-nine radiological features were annotated from historical anteroposterior and lateral radiographs for 162 total hip arthroplasty patients, 32 of which later underwent implant failure. Linear regression on each patient's chronologically sorted radiological features was employed to derive 169 corresponding evolutional parameters per image. Three sets of machine learning predictors were developed: one employing the original features (standard model), one the evolutional ones (evolutional model), and the last their union (hybrid model). Each set included a model employing all the available features (full model) and a model employing the few most predictive ones according to Gini importance (minimal model).

RESULTS

The evolutional and hybrid predictors resulted highly effective (area under the ROC curve (AUC) of full models = 0.94), outperforming the standard one, whose AUC was only 0.82. The minimal hybrid model, employing just four features, three of which evolutional, scored an AUC of 0.95, proving even more accurate than the full one, exploiting 173 features. This tool could be shaped to be either highly specific (sensitivity: 80%, specificity: 98.6%) or highly sensitive (sensitivity: 90%, specificity: 92.4%).

CONCLUSION

The proposed predictor may represent a highly sensitive screening tool for clinicians, capable to predict THA failure with an advance between a few months and more than a year through only four radiological parameters, considering either their value at the latest visit or their evolution through time.

Artificial intelligence-based detection of paediatric appendicular skeletal fractures: performance and limitations for common fracture types and locations

BACKGROUND

Research into artificial intelligence (AI)-based fracture detection in children is scarce and has disregarded the detection of indirect fracture signs and dislocations.

OBJECTIVE

To assess the diagnostic accuracy of an existing AI-tool for the detection of fractures, indirect fracture signs, and dislocations.

MATERIALS AND METHODS

An AI software, BoneView (Gleamer, Paris, France), was assessed for diagnostic accuracy of fracture detection using paediatric radiology consensus diagnoses as reference. Radiographs from a single emergency department were enrolled retrospectively going back from December 2021, limited to 1,000 radiographs per body part. Enrolment criteria were as follows: suspected fractures of the forearm, lower leg, or elbow; age 0-18 years; and radiographs in at least two projections.

RESULTS

Lower leg radiographs showed 607 fractures. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were high (87.5%, 87.5%, 98.3%, 98.3%, respectively). Detection rate was low for toddler's fractures, trampoline fractures, and proximal tibial Salter-Harris-II fractures. Forearm radiographs showed 1,137 fractures. Sensitivity, specificity, PPV, and NPV were high (92.9%, 98.1%, 98.4%, 91.7%, respectively). Radial and ulnar bowing fractures were not reliably detected (one out of 11 radial bowing fractures and zero out of seven ulnar bowing fractures were correctly detected). Detection rate was low for styloid process avulsions, proximal radial buckle, and complete olecranon fractures. Elbow radiographs showed 517 fractures. Sensitivity and NPV were moderate (80.5%, 84.7%, respectively). Specificity and PPV were high (94.9%, 93.3%, respectively). For joint effusion, sensitivity, specificity, PPV, and NPV were moderate (85.1%, 85.7%, 89.5%, 80%, respectively). For elbow dislocations, sensitivity and PPV were low (65.8%, 50%, respectively). Specificity and NPV were high (97.7%, 98.8%, respectively).

CONCLUSIONS

The diagnostic performance of BoneView is promising for forearm and lower leg fractures. However, improvement is mandatory before clinicians can rely solely on AI-based paediatric fracture detection using this software.

Self-supervised multi-modal training from uncurated images and reports enables monitoring AI in radiology

The escalating demand for artificial intelligence (AI) systems that can monitor and supervise human errors and abnormalities in healthcare presents unique challenges. Recent advances in vision-language models reveal the challenges of monitoring AI by understanding both visual and textual concepts and their semantic correspondences. However, there has been limited success in the application of vision-language models in the medical domain. Current vision-language models and learning strategies for photographic images and captions call for a web-scale data corpus of image and text pairs which is not often feasible in the medical domain. To address this, we present a model named medical cross-attention vision-language model (Medical X-VL), which leverages key components to be tailored for the medical domain. The model is based on the following components: self-supervised unimodal models in medical domain and a fusion encoder to bridge them, momentum distillation, sentencewise contrastive learning for medical reports, and sentence similarity-adjusted hard negative mining. We experimentally demonstrated that our model enables various zero-shot tasks for monitoring AI, ranging from the zero-shot classification to zero-shot error correction. Our model outperformed current state-of-the-art models in two medical image datasets, suggesting a novel clinical application of our monitoring AI model to alleviate human errors. Our method demonstrates a more specialized capacity for fine-grained understanding, which presents a distinct advantage particularly applicable to the medical domain.

Intraoperative Imaging Challenges During Pelvic Ring Disruptions and Acetabular Fracture Surgery

Achieving high-quality intraoperative imaging is crucial for successful pelvic ring and acetabular fracture surgery, yet it remains clinically challenging. Due to the complex anatomy of the pelvic ring and acetabulum, it is necessary to obtain multiple images oriented in different planes to reliably confirm reduction accuracy and implant positioning. Intraoperative image quality can be compromised by factors such as patient body habitus, bowel gas, abdominal packing, contrast dye, and nonstandardized language between surgeon and radiology technician. This article reviews common intraoperative imaging challenges encountered during pelvic ring and acetabular fracture surgery, while providing practical and evidence-based solutions and prevention strategies.

Patterns of Ankle Fractures Based on Radiographs and CT Images of 1000 Consecutive Patients

Objectives

The knowledge of different types of ankle fractures based on plain radiographs and computed tomography (CT) images can help improve patients' management.

Methods

This cross-sectional study assessed the plain radiographs and CT images of 1,000 consecutive patients observed in an emergency department between March 2015 and March 2020. Fractures were labeled as uni-, bi-, or trimalleolar. Malleolar fractures were classified into medial, lateral, and posterior ones based on Herscovici, Danis-Weber, and Mason and Molloy classifications, respectively. Bi- and trimalleolar fractures, on the other hand, were categorized according to the Lauge-Hansen classification.

Results

This study included 1,000 patients with 1,003 ankle fractures. Of them, 901 were adults (mean±SD age: 41.6±16.7, male: 567 [62.9%]) with 904 fractures. In total, 53% of adult patients were 18 to 39 years old. Considering unilateral ankle fractures, the medial malleolar fracture was the most common unimalleolar fracture (62.6%), with Herscovici C being the most frequent subtype (65.3%). On the other hand, the most common type of lateral malleolar fracture was Danis-Weber type B (65.5%). There were also 209 (23.3%) bimalleolar and 114 (12.7%) trimalleolar fractures, 5.8% (16 fractures) of which could not be classified based on the Lauge-Hansen classification. Unimalleolar fractures were also observed in 87 (87.9%) children, with the medial malleolar fracture being the most common type (89.7%).

Conclusion

Medial malleolar fractures were the most frequent malleoli in patients observed in the emergency department under study. Among bi- and trimalleolar ankle fractures, supination-external rotation and pronation-external rotation injuries were the most common patterns. The Lauge-Hansen classification was not applicable in 5.8% of bi- and trimalleolar fractures.

Dental age estimation methods applied to Indian children and adolescents: A systematic review and meta-analysis

Country-specific systematic reviews and meta-analyses have been proposed to compile the available literature and rank methods based on their performance for a target population. India is a country with a vast scientific literature on dental age estimation. This systematic review aimed to provide evidence to help the decision of experts regarding the method of choice for dental age estimation in India. The research protocol was registered in Open Science Framework. Literature Search was performed in Embase, LILACS, MedLine (via PubMed), SciELO, Scopus and Web of Science. Grey Literature was searched in Google Scholar and ProQuest. Observational cross-sectional studies that compared chronological and estimated ages using Demirjian (original [DEM] and Chaillet's modification [modified-DEM]) and Acharya (ACH) methods were included. JBI tool was used to assess the risk of bias. The search detected 9799 studies, out of which 56 were eligible (n=13,107 panoramic radiographs of Indian individuals). Low risk of bias was registered for 48 studies, while 8 presented a moderate risk of bias. The meta-analysis showed a standardized mean difference between chronological and estimated ages of -0.11 (95%CI: -0.29; 0.07), 0.74 (95%CI: 0.39; 1.09), and -0.01 (95%CI: -0.23; 0.22) years for DEM, modified-DEM and ACH, respectively. High heterogeneity (I2=88-93%) was observed across studies for all the methods, including subgroup analyses based on sex. This study ranked ACH, DEM and modified-DEM (from the best to the worse) performances in the Indian population.

Variations in wait times for imaging services: a register-based study of self-reported wait times for specific examinations in Norway

BACKGROUND

While the number of medical images has increased substantially, the demand has outpaced access, resulting in long wait times in many countries. Long wait times are a key problem for patient safety and quality of care as they can result in prolonged suffering, delayed diagnosis and treatment, as well as poorer prognosis and loss of lives. Surprisingly, little is known about wait times for imaging services.

OBJECTIVE

Investigate wait times for specific imaging services in Norway and to compare wait times with the total number of examinations and their development over time.

METHODS

Data from the wait time registry at the Norwegian Directorate of Health from 2018 to 2021 as well as data on outpatient imaging provided by the Norwegian Health Economics Administration (HELFO) and in-patient data afforded by fourteen hospital trusts and hospitals in Norway were analysed. Data include the total number of imaging examinations according to the Norwegian Classification of Radiological Procedures (NCRP). Analyses were performed with descriptive statistics.

RESULTS

Wait times vary through the months of the year. Conventional X-ray (XR) had the shortest wait times (3.0-4.4 weeks), and Magnetic Resonance Imaging (MRI) and ultrasound (US) had the lengthiest (8.7-12.0 and 7.9-11.4 weeks respectively). The wait times were lengthiest during the summer and winter holidays. Variations in wait times were also found for specific examination types between Norway's four public health regions. In addition, there was variation over time within the health regions. The wait times with the private health providers were substantially lower than with the public health providers. From 2018 to 2021, the wait time for MRIs increased by 6.6%, while the number of examinations (per 10,000) increased by 8.6%. Those regions with the highest number of examinations per 1,000 inhabitants per year had the lowest wait times.

CONCLUSION

Wait times for diagnostic imaging procedures varied with time, region, and modality in Norway from 2018 to 2021. Long wait times may entail many negative consequences for patients, professionals, and the healthcare system. Reducing long wait times is an obvious way to improve the quality, safety, and efficiency of care.

Role of a radiopaque agent and surveillance radiographs for peripherally inserted central catheters in newborn infants

BACKGROUND

Controversy exists regarding the use of a radiopaque agent to identify peripherally inserted central catheter (PICC) tip positions in newborn infants and of serial radiography to monitor PICC tip migration.

OBJECTIVE

To investigate the roles of (1) the injection of a radiopaque agent to identify PICC tip position and (2) the performance of weekly radiography to monitor PICC migration.

MATERIALS AND METHODS

This retrospective single-centre cohort study included newborn infants who received a PICC between 1 January 2016 and 31 December 2020. A radiopaque agent was injected to identify PICC tip position and radiographs were performed weekly to detect PICC migration.

RESULTS

We identified 676 PICC episodes in 601 infants. A radiopaque agent was used for 590 of these episodes. There was no difference in the proportion of central PICC tip positions based on radiopaque agent use status (490/590, 83% for the radiopaque agent used group versus 73/85, 85.8% for the radiopaque agent not used group, P=0.51). Irrespective of the site of PICC insertion, outward migration was observed for most centrally placed PICCs over their entire in situ duration. Inward migration was identified in 23 out of 643 PICC episodes (3.6%) only on radiographs obtained on or before day 7. Based on serial radiographs, the odds for PICC tips remaining in a central position were lower the longer the PICC remained in situ (adjusted odds ratio-OR 0.93; 95% confidence interval 0.92-0.95). There was no difference in PICC migration between side and limb of insertion.

CONCLUSION

PICC tips can be identified without injection of a radiopaque agent. Serial radiographs identified PICC migration over the in situ duration. This study has implications for reducing exposure to a radiopaque agent and ongoing migration surveillance practices.

Defining Haglund's deformity radiographically by incorporating the biomechanics of ankle motion

BACKGROUND

Haglund's deformity is clinically defined by the presence of retrocalcaneal tenderness, with previous radiographic parameters relying on calcaneal anatomic parameters that did not considering the influence of ankle motion on posterior calcaneal-Achilles impingement METHODS: Standing foot radiographs from 55 patients with clinically defined Haglund's deformity were compared to 50 control patients using previously described measurements and 2 new angular measurements based on the linkage between ankle rotation and posterior calcaneal-Achilles impingement. The ability of each measure to distinguish between Haglund's and control patients was assessed.

RESULTS

The combination of the angles to account for both increased calcaneal tubercle height and increased posterior calcaneal prominence could distinguish between the two patient groups (p = .018, Area under the curve = 63.2%). None of the previously published radiographic criteria were different between the two patient groups.

CONCLUSIONS

The proposed radiographic criteria were more predictive than previous criteria that did not address the role of ankle motion.

When MRI would be useful in patients without evidence of sacroiliitis on radiographs?

We aimed to identify when magnetic resonance imaging (MRI) would be useful to diagnose patients with suspected axial spondyloarthropathy (AxSpA) without evidence of sacroiliitis on radiographs. We retrospectively reviewed electronic medical records of patients who underwent pelvis MRI after radiographs at the rheumatology clinic in a single tertiary center in Korea. Patients underwent imaging from January 2020 to July 2022. We collected data including complete blood count, erythrocyte sedimentation rate, C-reactive protein (CRP), human leukocyte antigen (HLA)-B27, history of acute anterior uveitis (AAU), peripheral arthritis, dactylitis, inflammatory bowel disease (IBD), enthesopathy, and psoriasis. A total of 105 patients who showed no evidence of sacroiliitis on radiographs were included. The median age of patients was 41.0 years, and 44.8% were male. Of them, 34 showed sacroiliitis on MRI (group 1), and 71 showed no evidence of sacroiliitis even on MRI (group 2). Known AxSpA-related clinical features including AAU, peripheral arthritis, dactylitis, IBD, enthesopathy, and psoriasis were not different between the two groups. HLA-B27 positivity (79.4% vs. 40.0%, p < 0.001), median white blood cell count (7700 vs. 6300, p = 0.007), mean platelet count (307.7 ± 69.7 vs. 265.3 ± 68.9 × 103/µL, p = 0.005), and median CRP level (0.38 vs. 0.10, p = 0.001) showed significant differences between the two groups. In a multivariate analysis, HLA-B27 positivity and platelet count were significantly associated with sacroiliitis on MRI. In our cohort, sacroiliitis was observed on MRI in one-third of patients without radiographic evidence. MRI could be recommended to evaluate sacroiliitis in patients with positive HLA-B27 and a high platelet count.

Patellar medial-lateral position can be used to correct the effect of leg rotation on preoperative planning in total knee arthroplasty for varus knees

BACKGROUND

Lower limb malrotations can be observed in long leg radiographs, affecting the measurement of the angle between the mechanical and anatomical axes. The purposes were to analyze the effect of limb rotation and to evaluate the accuracy of the corrected angle between the mechanical and anatomical axes based on the patellar ML position.

HYPOTHESIS

The hypothesis was that the correction of the angle between the mechanical and anatomical axes according to the patellar ML position can reduce the error from the angle in the true AP view in most of the knees.

PATIENTS AND METHODS

A total of 100 consecutive knees with varus deformity undergoing primary total knee arthroplasty were included. Computed tomography images were digitally reconstructed in the neutral position, and internally and externally rotated at 10° and 20°, respectively. The patellar ML position relative to the medial (0%) and lateral (100%) epicondyles and the angle between the mechanical and anatomical axes of the femur were measured. The corrected angle between the mechanical and anatomical axes was calculated using the averaged translational ratio.

RESULTS

In the neutral position, the patellar center position was 56.1% (standard deviation [SD]=4.7%), which was 31.4% (SD=7.2%) and 80.2% (SD=5.4%) in the 20° internal and external rotation, respectively. The angle between the mechanical and anatomical axes was 2.6° (SD=2.0°) and 8.1° (SD=2.1°) in the 20° internal and external rotation, respectively. On average, if the patellar center shifted 10%, the change of the angle between the mechanical and anatomical axes of the femur was 1.13°. Applying the corrected angle, a discrepancy from the neutral position decreased.

CONCLUSION

The method to correct the angle between the mechanical and anatomical axes according to the patellar ML position can be used to reduce the measurement error for preoperative planning using a long leg radiograph.

LEVEL OF EVIDENCE

III.

Does radiolucency really predict loose components in revision shoulder arthroplasty?

OBJECTIVE

The number of shoulder arthroplasties is increasing along with the need for revision surgeries. Determining the stability of the implant is crucial in preoperative planning. This study aims to investigate whether radiolucent lines (RLL) in preoperative radiographs predict component loosening.

MATERIALS AND METHODS

Preoperative radiographs of 93 cases in 88 patients who underwent shoulder arthroplasty revision were evaluated regarding the presence of RLL. Correlation analyses were performed for radiographic findings and demographic factors (age, gender, BMI, prior surgeries) compared to intraoperative findings.

RESULTS

The presence of RLL around the humeral component correlated with loosening (p < 0.001, Phi 0.511), and the distal zones 3 and 5 showed the strongest correlation (Phi 0.536). While RLL in only one zone did not predict loosening (p = 0.337), RLL present in two or more zones showed correlation with loosening (p < 0.001). Risk factors associated with loosening were a higher age at the time of revision surgery (p = 0.030) and the number of zones with RLL (p < 0.001). The glenoid component was loose in 39.0% of the cases; 5.5% of the glenoid components with RLL were stable. Nevertheless, the presence of RLL was highly associated with loosening (p < 0.001, Phi 0.603). A longer time between implantation and revision correlated with loosening of the glenoid component (p = 0.046).

CONCLUSION

While RLL do not predict loosening of the implant in general, occurrence in more than one zone correlates with loosening. If located in distal zones and with increasing number of zones with RLL, the correlation becomes even stronger and loosening is more likely.

Evaluating cone cut in rectangular collimation in intraoral radiography: application of ALADA and radiation stewardship

OBJECTIVE

Rectangular collimation is a popular method used in intraoral radiography to reduce patient exposure to ionizing radiation. One of the perceived drawbacks of rectangular collimation is the possibility of an increase in cone cut errors ultimately impacting the diagnostic value of the radiographs. Thus, the aim of this study was to explore the frequency of cone cut errors in radiographs taken using a rectangular collimator.

MATERIALS AND METHODS

Radiographs taken using PSP plates at Academic Center for Dentistry Amsterdam in the Netherlands by staff and students from January to December 2015 were assessed for cone cut errors. The radiographs were grouped as bitewings, front teeth, inferior premolars and molars, and superior premolars and molars and categorized as no cone cut, cone cut but diagnostically usable, and cone cut but diagnostically not usable. The results were entered into Microsoft Excel and analyzed thereafter.

RESULTS

A total of 53,684 radiographs were assessed, 79% had no cone cut errors and consequently 21% had some degree of cone cut. However, the diagnostic value was unaffected in 18% of the radiographs with cone cut. Only 3% of the radiographs were deemed diagnostically unusable due to cone cut. The most common area of cone cut was in the premolar and molar areas while cone cut in the front teeth was least likely to be diagnostically unusable.

CONCLUSION

Cone cut from the use of a rectangular collimator does not seem to result in an increase of diagnostically unusable radiographs. Thus, rectangular collimation should be preferred as it decreases the amount of radiation exposure to the patient while producing diagnostically usable radiographs and thus allowing the dental professional to adhere to the ALADA principle and practice radiation stewardship.

CLINICAL RELEVANCE

Scientific rationale for the study: rectangular collimation is a method used to reduce patient exposure to ionizing radiation; however, this benefit is negligible if radiographs must be retaken due to cone cut errors that make the radiograph diagnostically unusable. Therefore, the aim of this study was to explore the frequency of cone cut in radiographs taken using a rectangular collimator.

PRINCIPAL FINDINGS

cone cut was observed in 21% of the radiographs; however, only 3% of the radiographs were considered diagnostically unusable.

PRACTICAL IMPLICATIONS

rectangular collimation does not result in a high number of diagnostically unusable radiographs and should be used to reduce patient exposure to ionizing radiation.

The use of robot-assisted surgery for the unstable traumatic spine: A retrospective cohort study

Background

Robotic assistance has been shown to increase instrumentation placement accuracy in open and minimally invasive spinal fusion. These gains have been achieved without increases in operative times, blood loss, or hospitalization duration. However, most work has been done in the degenerative population and little is known of the utility of robotic assistance when applied to spinal trauma. This is largely due to the uncertainty stemming from the disruption of normal anatomy by the traumatic injury. Since the robot depends upon registration for instrumentation guidance according to the fiducials it uses, trauma can introduce unique challenges. The present study sought to evaluate the safety and efficacy of robotic assistance in a consecutive cohort of spine trauma patients.

Methods

All patients with Thoracolumbar Injury Classification and Severity Scale (TLICS) >4 who underwent robot-assisted spinal fusion using the Globus ExcelsiusGPS at a single tertiary care center for trauma between 2020 and 2022 were identified. Demographic, clinical, and surgical data were collected and analyzed; the primary endpoints were operative time, fluoroscopy time, estimated blood loss, postoperative complications, admission time, and 90-day readmission rate. The paired t-test was used to compare differences between mean values when looking at the number of surgical levels.

Results

Forty-two patients undergoing robot-assisted spinal surgery were included (mean age 61.3±17.1 year; 47% female. Patients were stratified by the number of operative levels, 2 (n = 10), 3-4 (n = 11), 5 to 6 (n = 13), or >6 (n = 8). There appeared to be a positive correlation between number of levels instrumented and odds of postoperative complications, admission duration, fluoroscopy time, and estimated blood loss. There were no instances of screw malposition or breach.

Conclusions

This initial experience suggests robotic assistance can be safely employed in the spine trauma population. Additional experiences in larger patient populations are necessary to delineate those traumatic pathologies most amenable to robotic assistance.

A retrospective evaluation of supine pelvic radiography image quality using centring points and anatomical axial rotation, including reliability of measurements (ARLEX-P STUDY)

INTRODUCTION

Pelvic radiographs are commonly used for the investigation of a variety of conditions. Comparison between examinations requires a consistent radiographic technique but variations in image quality and radiographic centring points are frequently reported in the literature. The aim of this study was to establish the amount of variation in the radiographic centring point (RCP) and pelvic axial rotation (PAR), with a secondary aim of reporting the reliability of such measures.

METHODS

Using a previously acquired imaging archive, 633 adult pelvis/hip radiographs were identified on a Picture Archiving and Communication System (PACS). Radiographs with bilateral prostheses, evidence of acute pelvic trauma, projections acquired on a stretcher/trolley and those demonstrating large discontinuity between the detector and X-ray field centre were excluded. To determine centring point variation (+ values denote superior variations) and axial rotation multiple measurements were obtained from each radiograph. A video was used to train five observers and each of these reviewed ten random cases to determine inter- and intra-rater reliability. One of the five observers then performed the measurements on all remaining radiographs.

RESULTS

Following exclusions 380 radiographs were evaluated. The median (IQR) RCP deviation from the inter-acetabular line was +22 (+2 to +43) mm where both iliac crests were present and -29 (-45 to -12) mm where they were not. Eleven (3%) cases demonstrate RCP variation from the midline of greater than 25 mm (no bias towards the left or right side). The median (IQR) PAR was 0.0 (-1.5 to 1.4) degrees with greater variance in PAR for male participants (p = 0.004). Almost 60% of inter-rater ICC measurements were categorised as excellent, good or moderate.

CONCLUSION

Variations in RCP and PAR exist when evaluating a sample of routinely acquired pelvis radiographs. Some initial factors, such as sex and sub-examination type (full pelvis [XPEL] or low centred pelvis [XHIPB]) have been identified as having a statistical affect on variability. Further research and methods to standardise radiographic techniques is required and must be multidimensional in nature.

IMPLICATIONS FOR PRACTICE

Selection of radiographic technique, including RCP, appears to influence components of the pelvis radiograph. Given the increasing clinical requirements for pelvic radiography further standardisation alongside individual optimisation is warranted.

Evolving role of chest radiographs for diagnosis of pediatric pulmonary tuberculosis

Chest radiographs (CXR) have played an important and evolving role in diagnosis, classification and management of pediatric pulmonary tuberculosis (TB). During the pre-chemotherapy era, CXR aided in determining infectiousness, mainly to guide isolation practices, by detecting calcified and non-calcified lymphadenopathy. The availability of TB chemotherapy from the mid-1900s increased the urgency to find accurate diagnostic tools for what had become a treatable disease. Chest radiographs provided the mainstay of diagnosis in children, despite high inter-reader variability limiting its accuracy. The development of cross-sectional imaging modalities, such as computed tomography, provided more accurate intra-thoracic lymph node assessment, but these modalities have major availability, cost and radiation exposure disadvantages. As a consequence, CXR remains the most widely used modality for childhood  pulmonary TB diagnosis, given its relatively low cost and accessibility. Publication of the revised 2022 World Health Organization Consolidated TB guidelines added practical value to CXR interpretation in children, by allowing the selection of children for shorter TB treatment using radiological signs of severity of disease, that have high reliability. This article provides a review of the historical journey and evolving role of CXR in pediatric pulmonary TB.

Imaging recommendations and algorithms for pediatric tuberculosis: part 1-thoracic tuberculosis

Tuberculosis (TB) remains a global health problem and is the second leading cause of death from a single infectious agent, behind the novel coronavirus disease of 2019. Children are amongst the most vulnerable groups affected by TB, and imaging manifestations are different in children when compared to adults. TB primarily involves the lungs and mediastinal lymph nodes. Clinical history, physical examination, laboratory examinations and various medical imaging tools are combined to establish the diagnosis. Even though chest radiography is the accepted initial radiological imaging modality for the evaluation of children with TB, this paper, the first of two parts, aims to discuss the advantages and limitations of the various medical imaging modalities and to provide recommendations on which is most appropriate for the initial diagnosis and assessment of possible complications of pulmonary TB in children. Practical, evidence-based imaging algorithms are also presented.

Detection of Pacemaker and Identification of MRI-conditional Pacemaker Based on Deep-learning Convolutional Neural Networks to Improve Patient Safety

With the increased availability of magnetic resonance imaging (MRI) and a progressive rise in the frequency of cardiac device implantation, there is an increased chance that patients with implanted cardiac devices require MRI examination during their lifetime. Though MRI is generally contraindicated in patients who have undergone pacemaker implantation with electronic circuits, the recent introduction of MR Conditional pacemaker allows physicians to take advantage of MRI to assess these patients during diagnosis and treatment. When MRI examinations of patients with pacemaker are requested, physicians must confirm whether the device is a conventional pacemaker or an MR Conditional pacemaker by reviewing chest radiographs or the electronic medical records (EMRs). The purpose of this study was to evaluate the utility of a deep convolutional neural network (DCNN) trained to detect pacemakers on chest radiographs and to determine the device's subclassification. The DCNN perfectly detected pacemakers on chest radiographs and the accuracy of the subclassification of pacemakers using the internal and external test datasets were 100.0% (n = 106/106) and 90.1% (n = 279/308). The DCNN can be applied to the radiologic workflow for double-checking purposes, thereby improving patient safety during MRI and preventing busy physicians from making errors.

How to place a cephalomedullary screw when visualization is obscured by the jig in peritrochanteric hip fractures using "peek radiographs."

Lag screw positioning can be difficult to discern intraoperatively on lateral fluoroscopic imaging during intramedullary fixation of proximal femur fractures in some nailing systems due to the drill guide handle obstructing the view. We have described a method of obtaining non-obstructed lateral-oblique "peek" views that reliably assist in obtaining adequate tip-apex distance (TAD) measurements when using intramedullary fixation for these fractures. The purpose of this study was (1) to describe an intraoperative radiographic technique to obtain non-obscured views for appropriate center-center placement of the lag screw(s) within the femoral head during jig-aided cephalomedullary nailing of peritrochanteric hip fractures and (2) to present a case series detailing the radiographic results using this technique. This clinical series of sixty-five patients with intertrochanteric or subtrochanteric proximal femur fractures stabilized with a cephalomedullary nail had an average TAD of 15.1 ± 3.3mm (range: 9.4mm-26.2mm). This suggests that our technical trick of obtaining "peek" radiographs intraoperatively may aid in a precise lag-screw placement.

The tangential view described by Moneim to demonstrate scapholunate dissociation: an update

PURPOSE

Scapholunate dissociation is a common and significant injury to the wrist. Radiographs are important in the diagnosis of this injury and in the planning of treatment. The tangential radiograph view was described almost 40 years ago as a method for accurately measuring scapholunate gaps. The hand is positioned on a 20° foam rubber block and the thumb on the cassette, which positions the scaphoid and lunate articular surfaces parallel, without patient discomfort or effort. The goal of this study was to review this method with more recent data and in a larger group of patients.

METHODS

Radiographs of 31 patients who had scapholunate interosseous ligament tears and surgical repair over a 9 year period were retrospectively evaluated. Each of the four authors independently measured scapholunate gaps for posteroanterior and tangential views.

RESULTS

The tangential view gaps were significantly greater than the posteroanterior gaps overall. Similar results were found for borderline cases where the posteroanterior gap was less than 3 mm. Every tangential view gap measurement was greater than its respective posteroanterior gap with good inter-rater reliability.

CONCLUSION

The tangential view is a reliable radiographic method to identify scapholunate gaps. It should be obtained when there is clinical concern for scapholunate dissociation, especially in patients with borderline posteroanterior gaps.

Understanding transfer learning for chest radiograph clinical report generation with modified transformer architectures

The image captioning task is increasingly prevalent in artificial intelligence applications for medicine. One important application is clinical report generation from chest radiographs. The clinical writing of unstructured reports is time consuming and error-prone. An automated system would improve standardization, error reduction, time consumption, and medical accessibility. In this paper we demonstrate the importance of domain specific pre-training and propose a modified transformer architecture for the medical image captioning task. To accomplish this, we train a series of modified transformers to generate clinical reports from chest radiograph image input. These modified transformers include: a meshed-memory augmented transformer architecture with visual extractor using ImageNet pre-trained weights, a meshed-memory augmented transformer architecture with visual extractor using CheXpert pre-trained weights, and a meshed-memory augmented transformer whose encoder is passed the concatenated embeddings using both ImageNet pre-trained weights and CheXpert pre-trained weights. We use BLEU(1-4), ROUGE-L, CIDEr, and the clinical CheXbert F1 scores to validate our models and demonstrate competitive scores with state of the art models. We provide evidence that ImageNet pre-training is ill-suited for the medical image captioning task, especially for less frequent conditions (e.g.: enlarged cardiomediastinum, lung lesion, pneumothorax). Furthermore, we demonstrate that the double feature model improves performance for specific medical conditions (edema, consolidation, pneumothorax, support devices) and overall CheXbert F1 score, and should be further developed in future work. Such a double feature model, including both ImageNet pre-training as well as domain specific pre-training, could be used in a wide range of image captioning models in medicine.

Effect of humeral rotation on the reliability of radiographic measurements for proximal humerus fractures

BACKGROUND

There are concerns as to the reliability of proximal humerus radiographic measurements, particularly regarding the rotational position of the humerus when obtaining radiographs.

METHODS

Twenty-four patients with proximal humerus fractures fixed surgically with locked plates received postoperative anteroposterior radiographs with the humerus in neutral rotation and in 30° of internal and external rotation. Radiographic measurements for head shaft angle, humeral offset and humeral head height were performed in each humeral rotation position. Intra-class correlation coefficient was used to assess inter-rater and intra-rater reliability. Mean differences (md) in measurements between humeral positions was evaluated using one-way ANOVA.

RESULTS

Head shaft angle demonstrated good-to-excellent reliability; the highest estimates for inter-rater reliability (ICC: 0.85; 95% CI: 0.76, 0.94) and intra-rater reliability (ICC: 0.96; 95% CI: 0.93, 0.98) were achieved in neutral rotation. There were significant differences in measurement values between each rotational position, with mean head shaft angle of 133.1° in external rotation, and increasingly valgus measurements in neutral (md: 7.6°; 95% CI: 5.0, 10.3°; p < 0.001) and internal rotation (md: 26.4°; 95% CI: 21.8, 30.9°; p < 0.001). Humeral head height and humeral offset showed good-to-excellent reliability in neutral and external rotation, but poor inter-rater reliability in internal rotation. Humeral head height was significantly greater using internal compared to external rotation (md: 4.5 mm; 95% CI: 1.7, 7.3 mm; p = 0.002). Humeral offset was significantly greater in external compared to internal rotation (md: 4.6 mm; 95% CI: 2.6, 6.6 mm; p < 0.001).

CONCLUSIONS

Views of the humerus in neutral rotation and 30° of external rotation displayed superior reliability. Differences in radiographic measurement values, depending on humeral rotation views, can make for problematic correlations with patient outcome measures. Studies assessing radiographic outcomes following proximal humerus fractures should ensure standardized humeral rotation for obtaining anteroposterior shoulder radiographs, with neutral rotation and external rotation views likely yielding the most reliable results.

LEVEL OF EVIDENCE

Level IV.

The role of imaging in osteoarthritis

Osteoarthritis is a complex whole-organ disorder that involves molecular, anatomic, and physiologic derangement. Advances in imaging techniques have expanded the role of imaging in evaluating osteoarthritis and functional changes. Radiography, magnetic resonance imaging, computed tomography (CT), and ultrasonography are commonly used imaging modalities, each with advantages and limitations in evaluating osteoarthritis. Radiography comprehensively analyses alignment and osseous features, while MRI provides detailed information about cartilage damage, bone marrow edema, synovitis, and soft tissue abnormalities. Compositional imaging derives quantitative data for detecting cartilage and tendon degeneration before structural damage occurs. Ultrasonography permits real-time scanning and dynamic joint evaluation, whereas CT is useful for assessing final osseous detail. Imaging plays an essential role in the diagnosis, management, and research of osteoarthritis. The use of imaging can help differentiate osteoarthritis from other diseases with similar symptoms, and recent advances in deep learning have made the acquisition, management, and interpretation of imaging data more efficient and accurate. Imaging is useful in monitoring and predicting the prognosis of osteoarthritis, expanding our understanding of its pathophysiology. Ultimately, this enables early detection and personalized medicine for patients with osteoarthritis. This article reviews the current state of imaging in osteoarthritis, focusing on the strengths and limitations of various imaging modalities, and introduces advanced techniques, including deep learning, applied in clinical practice.

Opportunistic Screening for Osteoporosis Using Hand Radiographs: A Preliminary Study

Patients with low bone mineral density (BMD) are at risk for fractures however are often undiagnosed. Therefore, there is a need to opportunistically screen for low BMD in patients who present for other studies. This is a retrospective study of 812 patients aged 50 years or older who had dual-energy X-ray absorptiometry (DXA) and radiographs of the hands within 12 months of each other. This dataset was randomly split into training/validation (n=533) and test (n=136) datasets. A deep learning (DL) framework was used to predict osteoporosis/osteopenia. Correlations between the textural analysis of the bones and DXA measurements were obtained. We found that the DL model had an accuracy of 82.00%, sensitivity of 87.03%, specificity of 61.00% and an area under the curve (AUC) of 74.00% to detect osteoporosis/osteopenia. Our findings show that radiographs of the hand can be used to screen for osteoporosis/osteopenia and identify patients who should get formal DXA evaluation.

Two-year imaging outcomes from a phase 3 randomized trial of secukinumab in patients with non-radiographic axial spondyloarthritis

BACKGROUND

Radiographic progression and course of inflammation over 2 years in patients with non-radiographic axial spondyloarthritis (nr-axSpA) from the phase 3, randomized, PREVENT study are reported here.

METHODS

In the PREVENT study, adult patients fulfilling the Assessment of SpondyloArthritis International Society classification criteria for nr-axSpA with elevated CRP and/or MRI inflammation received secukinumab 150 mg or placebo. All patients received open-label secukinumab from week 52 onward. Sacroiliac (SI) joint and spinal radiographs were scored using the modified New York (mNY) grading (total sacroiliitis score; range, 0-8) and modified Stoke Ankylosing Spondylitis Spine Score (mSASSS; range, 0-72), respectively. SI joint bone marrow edema (BME) was assessed using the Berlin Active Inflammatory Lesions Scoring (0-24) and spinal MRI using the Berlin modification of the AS spine MRI (ASspiMRI) scoring (0-69).

RESULTS

Overall, 78.9% (438/555) of patients completed week 104 of the study. Over 2 years, minimal changes were observed in total radiographic SI joint scores (mean [SD] change, - 0.04 [0.49] and 0.04 [0.36]) and mSASSS scores (0.04 [0.47] and 0.07 [0.36]) in the secukinumab and placebo-secukinumab groups. Most of the patients showed no structural progression (increase ≤ smallest detectable change) in SI joint score (87.7% and 85.6%) and mSASSS score (97.5% and 97.1%) in the secukinumab and placebo-secukinumab groups. Only 3.3% (n = 7) and 2.9% (n = 3) of patients in the secukinumab and placebo-secukinumab groups, respectively, who were mNY-negative at baseline were scored as mNY-positive at week 104. Overall, 1.7% and 3.4% of patients with no syndesmophytes at baseline in the secukinumab and placebo-secukinumab group, respectively, developed ≥ 1 new syndesmophyte over 2 years. Reduction in SI joint BME observed at week 16 with secukinumab (mean [SD], - 1.23 [2.81] vs - 0.37 [1.90] with placebo) was sustained through week 104 (- 1.73 [3.49]). Spinal inflammation on MRI was low at baseline (mean score, 0.82 and 1.07 in the secukinumab and placebo groups, respectively) and remained low (mean score, 0.56 at week 104).

CONCLUSION

Structural damage was low at baseline and most patients showed no radiographic progression in SI joints and spine over 2 years in the secukinumab and placebo-secukinumab groups. Secukinumab reduced SI joint inflammation, which was sustained over 2 years.

TRIAL REGISTRATION

ClinicalTrials.gov, NCT02696031.

Evaluation of two methods of bone age assessment in peripubertal children in Zimbabwe

OBJECTIVES

Bone age (BA) measurement in children is used to evaluate skeletal maturity and helps in the diagnosis of growth disorders in children. The two most used methods are Greulich and Pyle (GP), and Tanner and Whitehouse 3 (TW3), both based upon assessment of a hand-wrist radiograph. To our knowledge no study has compared and validated the two methods in sub-Saharan Africa (SSA), and only a few have determined BA despite it being a region where skeletal maturity is often impaired for example by HIV and malnutrition. This study aimed to compare BA as measured by two methods (GP and TW3) against chronological age (CA) and determine which method is most applicable in peripubertal children in Zimbabwe.

METHODS

We conducted a cross-sectional study of boys and girls who tested negative for HIV. Children and adolescents were recruited by stratified random sampling from six schools in Harare, Zimbabwe. Non-dominant hand-wrist radiographs were taken, and BA assessed manually using both GP and TW3. Paired sample Student t-tests were used to calculate the mean differences between BA and chronological age (CA) in boys and girls. Bland-Altman plots compared CA to BA as determined by both methods, and agreement between GP and TW3 BA. All radiographs were graded by a second radiographer and 20 % of participants of each sex were randomly selected and re-graded by the first observer. Intraclass correlation coefficient assessed intra- and inter-rater reliability and coefficient of variation assessed precision.

RESULTS

We recruited 252 children (111 [44 %] girls) aged 8.0-16.5 years. The boys and girls were of similar mean ± SD CA (12.2 ± 2.4 and 11.7 ± 1.9 years) and BA whether assessed by GP (11.5 ± 2.8 and 11.5 ± 2.1 years) or TW3 (11.8 ± 2.5 and 11.8 ± 2.1 years). In boys BA was lower than CA by 0.76 years (95 % CI: -0.95, -0.57) when using GP, and by 0.43 years (95 % CI: -0.61, -0.24) when using TW3. Among the girls there was no difference between BA and CA by either GP [-0.19 years (95 % CI: -0.40, 0.03)] or TW3 [0.07 years (95 % CI: -0.16, 0.29)]. In both boys and girls, there were no systematic differences between CA and TW3 BA across age groups whereas agreement improved between CA and GP BA as children got older. Inter-operator precision was 1.5 % for TW3 and 3.7 % for GP (n = 252) and intra-operator precision was 1.5 % for TW3 and 2.4 % for GP (n = 52).

CONCLUSION

The TW3 BA method had better precision than GP and did not systematically differ from CA, meaning that TW3 is the preferred method of assessment of skeletal maturity in Zimbabwean children and adolescents. TW3 and GP methods do not agree for estimates of BA and therefore cannot be used interchangeably. The systematic differences in GP BA assessments over age means it is not appropriate for use in all age groups or stages of maturity in this population.

The radiographic diversity of dental patterns for human identification - Systematic review and meta-analysis

This study aimed to revisit the scientific literature related to the diversity of dental patterns observed in radiographs. The rationale was to find evidence to support dental human identifications. A systematic review was performed following the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P). Strategic search was accomplished in five electronic data sources (SciELO, Medline/Pubmed, Scopus, Open Grey and OATD) were searched. The study model of choice was observational analytical cross-sectional. The search resulted 4.337 entries. The sequential screening based on title, abstract and full-text reading led to 9 eligible studies (n = 5.700 panoramic radiographs) published between 2004 and 2021. Studies from Asian countries were predominant (e.g., South Korea, China, and India). All the studies showed low risk of bias (measured according to the Johanna Briggs Institute's critical appraisal tool for observational cross-sectional studies). Morphological, therapeutic, and pathological identifiers were charted from radiographs to create dental patterns across studies. Six studies (n = 2.553 individuals) had similar methodology and outcome metrics and were included in the quantitative analysis. A meta-analysis was performed and revealed a pooled diversity of the human dental pattern of 0.979 combining maxillary and mandibular teeth. The additional subgroup analysis with maxillary and mandibular teeth have a diversity rate of 0.897 and 0.924, respectively. The existing literature shows that human dental patterns are highly distinctive, especially if morphological, therapeutic and pathological dental features are combined. The diversity of dental identifiers found in the maxillary, mandibular and combined arches is hereby corroborated by this meta-analyzed systematic review. These outcomes support applications for evidence-based human identification.

Can an artificial intelligence powered software reliably assess pelvic radiographs?

PURPOSE

Despite advances of three-dimensional imaging pelvic radiographs remain the cornerstone in the evaluation of the hip joint. However, large inter- and intra-rater variabilities were reported due to subjective landmark setting. Artificial intelligence (AI)-powered software applications could improve the reproducibility of pelvic radiograph evaluation by providing standardized measurements. The aim of this study was to evaluate the reliability and agreement of a newly developed AI algorithm for the evaluation of pelvic radiographs.

METHODS

Three-hundred pelvic radiographs from 280 patients with different degrees of acetabular coverage and osteoarthritis (Tönnis Grade 0 to 3) were evaluated. Reliability and agreement between manual measurements and the outputs of the AI software were assessed for the lateral-center-edge (LCE) angle, neck-shaft angle, sharp angle, acetabular index, as well as the femoral head extrusion index.

RESULTS

The AI software provided reliable results in 94.3% (283/300). The ICC values ranged between 0.73 for the Acetabular Index to 0.80 for the LCE Angle. Agreement between readers and AI outputs, given by the standard error of measurement (SEM), was good for hips with normal coverage (LCE-SEM: 3.4°) and no osteoarthritis (LCE-SEM: 3.3°) and worse for hips with undercoverage (LCE-SEM: 5.2°) or severe osteoarthritis (LCE-SEM: 5.1°).

CONCLUSION

AI-powered applications are a reliable alternative to manual evaluation of pelvic radiographs. While being accurate for patients with normal acetabular coverage and mild signs of osteoarthritis, it needs improvement in the evaluation of patients with hip dysplasia and severe osteoarthritis.

GUI Enabled Optimized Approach of CNN for Automatic Diagnosis of COVID-19 Using Radiograph Images

World Health Organization (WHO) proclaimed the Corona virus (COVID-19) as a pandemic, since it contaminated billions of individuals and killed lakhs. The spread along with the severity of the disease plays a key role in early detection and classification to reduce the rapid spread as the variants are changing. COVID-19 could be categorized as a pneumonia infection. Bacterial pneumonia, fungal pneumonia, viral pneumonia, etc., are the classifications of several forms of pneumonia, which are subcategorized into more than 20 forms and COVID-19 will come under viral pneumonia. The wrong prediction of any of these can mislead humans into improper treatment, which leads to a matter of life. From the radiograph that is X-ray images, diagnosis of all these forms can be possible. For detecting these disease classes, the proposed method will employ a deep learning (DL) technique. Early detection of the COVID-19 is possible with this model; hence, the spread of the disease is minimized by isolating the patients. For execution, a graphical user interface (GUI) provides more flexibility. The proposed model, which is a GUI approach, is trained with 21 types of pneumonia radiographs by a convolutional neural network (CNN) trained on Image Net and adjusts them to act as feature extractors for the Radiograph images. Next, the CNNs are combined with united AI strategies. For the classification of COVID-19 detection, several approaches are proposed in which those approaches are concerned with COVID-19, pneumonia, and healthy patients only. In classifying more than 20 types of pneumonia infections, the proposed model attained an accuracy of 92%. Likewise, COVID-19 images are effectively distinguished from the other pneumonia images of radiographs.

Timeliness of Lung Cancer Care From the Point of Suspicious Image at an Urban Safety Net Hospital

BACKGROUND

Timeliness of care is an important metric for lung cancer patients, and care delays in the safety-net setting have been described. Timeliness from the point of the suspicious image is not well-studied. Herein, we evaluate time intervals in the workup of lung cancer at an urban, safety net hospital and assess for disparities by demographic and clinical factors.

PATIENTS AND METHODS

We performed a retrospective analysis of lung cancer patients receiving some portion of their care at Boston Medical Center between 2015 and 2020. A total of 687 patients were included in the final analysis. Median times from suspicious image to first treatment (SIT), suspicious image to diagnosis (SID), and diagnosis to treatment (DT) were calculated. Nonparametric tests were applied to assess for intergroup differences in time intervals.

RESULTS

SIT, SID, and DT for the entire cohort was 78, 34, and 32 days, respectively. SIT intervals were 87 days for females and 72 days for males (p < .01). SIT intervals were 106, 110, 81, and 41 days for stages I, II, III, and IV, respectively (p < .01). SID intervals differed between black (40.5) and Hispanic (45) patients compared to white (28) and Asian (23) patients (p < .05).

CONCLUSION

Advanced stage at presentation and male gender were associated with more timely treatment from the point of suspicious imaging while white and Asian were associated with more timely lung cancer diagnosis. Future analyses should seek to elucidate drivers of timeliness differences and assess for the impact of timeliness disparities on patient outcomes in the safety net setting.

Radiographic findings involved in knee osteoarthritis progression are associated with pain symptom frequency and baseline disease severity: a population-level analysis using deep learning

PURPOSE

To (1) develop a deep-learning (DL) algorithm capable of producing limb-length and knee-alignment measurements, and (2) determine the association between limb-length discrepancy (LLD), coronal-plane alignment, osteoarthritis (OA) severity, and patient-reported knee pain.

METHODS

A multicenter, prospective patient cohort from the Osteoarthritis Initiative between 2004 and 2015 with full-limb standing radiographs at 12 month follow-up was included. A convolutional neural network was developed to automate measurements of the hip-knee-ankle (HKA) angle, femur, and tibia lengths, and LLD. At 12 month follow-up, patients reported their frequency of knee pain since enrollment and current level of knee pain.

RESULTS

A total of 1011 patients (2022 knees, 52.3% female) with an average age of 61.2 ± 9.0 years were included. The algorithm performed 12,312 measurements in 5.4 h. ICC values of HKA and LLD ranged between 0.87 and 1.00 when compared against trained radiologist measurements. Knees producing pain most days of the month were significantly more varus (mean HKA:- 3.9° ± 2.8°) or valgus (mean HKA:2.8° ± 2.3°) compared to knees that did not produce any pain (p < 0.05). In varus knees, those producing pain on most days were part of the shorter limb compared to nonpainful knees (p < 0.05). Baseline Kellgren-Lawrence grade was significantly associated with HKA magnitude, LLD, and pain frequency at 12 month follow-up (p < 0.05 all).

CONCLUSION

A higher frequency of knee pain was associated with more severe coronal plane deformity, with valgus deviation being one degree less than varus on average, suggesting that the knee tolerates less valgus deformation before symptoms become more consistent. Knee pain frequency was also associated with greater LLD and baseline KL grade, suggesting an association between radiographically apparent joint degeneration and pain frequency.

LEVEL OF EVIDENCE

IV case series.

Convolutional neural network for detecting rib fractures on chest radiographs: a feasibility study

BACKGROUND

Chest radiography is the standard investigation for identifying rib fractures. The application of artificial intelligence (AI) for detecting rib fractures on chest radiographs is limited by image quality control and multilesion screening. To our knowledge, few studies have developed and verified the performance of an AI model for detecting rib fractures by using multi-center radiographs. And existing studies using chest radiographs for multiple rib fracture detection have used more complex and slower detection algorithms, so we aimed to create a multiple rib fracture detection model by using a convolutional neural network (CNN), based on multi-center and quality-normalised chest radiographs.

METHODS

A total of 1080 radiographs with rib fractures were obtained and randomly divided into the training set (918 radiographs, 85%) and the testing set (162 radiographs, 15%). An object detection CNN, You Only Look Once v3 (YOLOv3), was adopted to build the detection model. Receiver operating characteristic (ROC) and free-response ROC (FROC) were used to evaluate the model's performance. A joint testing group of 162 radiographs with rib fractures and 233 radiographs without rib fractures was used as the internal testing set. Furthermore, an additional 201 radiographs, 121 with rib fractures and 80 without rib fractures, were independently validated to compare the CNN model performance with the diagnostic efficiency of radiologists.

RESULTS

The sensitivity of the model in the training and testing sets was 92.0% and 91.1%, respectively, and the precision was 68.0% and 81.6%, respectively. FROC in the testing set showed that the sensitivity for whole-lesion detection reached 91.3% when the false-positive of each case was 0.56. In the joint testing group, the case-level accuracy, sensitivity, specificity, and area under the curve were 85.1%, 93.2%, 79.4%, and 0.92, respectively. At the fracture level and the case level in the independent validation set, the accuracy and sensitivity of the CNN model were always higher or close to radiologists' readings.

CONCLUSIONS

The CNN model, based on YOLOv3, was sensitive for detecting rib fractures on chest radiographs and showed great potential in the preliminary screening of rib fractures, which indicated that CNN can help reduce missed diagnoses and relieve radiologists' workload. In this study, we developed and verified the performance of a novel CNN model for rib fracture detection by using radiography.

Sagittal mechanical ratio: A novel technique to define sagittal alignment of the femur independent of distal anatomic landmarks

BACKGROUND

Radiographic measurements to study sagittal alignment in the setting of knee are frequently difficult to evaluate due the presence of a prosthesis or implant that obscures traditional radiographic landmarks. In this paper we present a novel method of determining sagittal femoral alignment in the presence of obscuring implants.

METHODS

98 full-length femoral radiographs were reviewed and divided into two groups. In Group 1, the Distal Mechanical Point (DMP) was used to calculate the Distal Mechanical Ratio (DMR), defined as the ratio of the linear distance from the DMP to the anterior cortical axis divided by the distance from the anterior cortical axis to posterior condylar cortex. In group 2, the sagittal mechanical axis was measured using the true DMP (tDMP) and then separately measured using the DMR to find the calculated DMP (cDMP), and the angular variance between the calculated (cSMA) and true (tSMA) sagittal mechanical axis was calculated, as well as the linear distance between the tDMP and cDMP. Twenty additional patients with knee replacements were then selected and two observers used a cSMA to determine a femoral prosthesis flexion angle (FPFA), with intraobserver correlation calculated.

RESULTS

The mean DMR was found to be 0.24, with high intraobserver correlation and normal distribution. Validation of the model demonstrated angular variance between tSMA and cSMA less than 1 degree and linear distance between tDMP and cDMP less than 1 mm. Calculation of cCMA in the presence of total knee arthroplasty revealed very strong intraobserver correlation of 0.89.

CONCLUSION

The Distal Mechanical Ratio reliably predicted the true Sagittal Mechanical Axis within 1 degree and true Distal Mechanical Point within 1 mm, indicating that it may be a valuable tool for evaluating sagittal femoral alignment in cases where anatomic landmarks may be absent or obscured.

Pattern of pterygomaxillary disarticulation associated with Le Fort I maxillary osteotomy

Pterygomaxillary disarticulation (PMD) contributes to surgical complications of Le Fort 1 osteotomy and is associated with undesirable fractures of the pterygoid plates. The aim of this paper was to investigate the patterns of PMD in Le Fort I osteotomies using Rowe's disimpaction forceps, and to evaluate correlations with age and anatomical measurements. Cone-beam computed tomography (CBCT) scans of 70 consecutive orthognathic patients were retrospectively evaluated to study four patterns of PMD: Type 1 - PMD at, or anterior to, the pterygomaxillary junction (PMJ); Type 2 - PMD posterior to the PMJ; Type 3 - PMJ separation with comminuted fracture of the pterygoid plates; Type 4 - disarticulation of the maxilla involving the pterygoid plates above the level of the osteotomy line. The preoperative anteroposterior and mediolateral thicknesses of the PMJ and the length of the medial and lateral pterygoid plates were assessed. Satisfactory PMD was achieved in all cases and no severe complications were reported, including vascular, dental, mucosal, or neural damage. The most common PMD was Type 1 (54.3%), followed by Type 2 (40%). Comminuted fracture of the pterygoid plates was limited to 5.7% of cases, and no Type 4 was detected. A weak correlation was detected between PMJ thickness and PMD pattern (p = 0.04). No statistically significant correlation was detected between patients' age and type of PMD. PMD of Le Fort I maxillary osteotomy using a Smith spreader and Rowe's disimpaction forceps proved safe, with minimal damage to the pterygoid plates.

Ensemble deep learning model for predicting anterior cruciate ligament tear from lateral knee radiograph

OBJECTIVE

To develop an ensemble deep learning model (DLM) predicting anterior cruciate ligament (ACL) tears from lateral knee radiographs and to evaluate its diagnostic performance.

MATERIALS AND METHODS

In this study, 1433 lateral knee radiographs (661 with ACL tear confirmed on MRI, 772 normal) from two medical centers were split into training (n = 1146) and test sets (n = 287). Three single DLMs respectively classifying radiographs with ACL tears, abnormal lateral femoral notches, and joint effusion were developed. An ensemble DLM predicting ACL tears was developed by combining the three DLMs via stacking method. The sensitivities, specificities, and area under the receiver operating characteristic curves (AUCs) of the DLMs and three radiologists were compared using McNemar test and Delong test. Subgroup analysis was performed to identify the radiologic features associated with the sensitivity.

RESULTS

The sensitivity, specificity, and AUC of the ensemble DLM were 86.8% (95% confidence interval [CI], 79.9-92.0%), 89.4% (95% CI, 83.4-93.8%), and 0.927 (95% CI, 0.891-0.954), achieving diagnostic performance comparable with that of a musculoskeletal radiologist (P = 0.193, McNemar test; P = 0.131, Delong test). The AUC of the ensemble DLM was significantly higher than those of non-musculoskeletal radiologists (P = 0.043, P < 0.001). The sensitivity of the DLM was higher than that of the radiologists in the absence of an abnormal lateral femoral notch or joint effusion.

CONCLUSION

The diagnostic performance of the ensemble DLM in predicting lateral knee radiographs with ACL tears was comparable to that of a musculoskeletal radiologist.

Predicting final results of brace treatment of adolescents with idiopathic scoliosis: first out-of-brace radiograph is better than in-brace radiograph-SOSORT 2020 award winner

PURPOSE

In-brace radiograph of adolescents with idiopathic scoliosis (AIS) has been shown to reflect brace efficacy and the possibility of achieving curve correction. Conversely, the first out-of-brace radiograph could demonstrate the patient's ability to maintain the correction. We aimed to determine which of the two radiographs is the best predictor of the Cobb angle at the end of treatment (final radiograph).

DESIGN

Retrospective cohort study of a prospective dataset.

METHODS

The population was selected based on the following inclusion criteria: AIS, age 10-18 years; Risser score 0-2; Cobb angle 25-40°; brace treatment; availability of all radiographs.

STATISTICS

Pearson correlations provide a first exploration of data. The univariate and multivariate logistic regression model tested the predictors. Finally ROC curve provided a check of model accuracy.

RESULTS

A total of 131 patients were included (mean age 13.0 ± 1.3, Cobb angle 33.2 ± 5.5°; 78% females). At the end of treatment, 56% had stabilised, 9% had progressed, and 44% had improved. The difference between the in-brace and final radiographs was 8.0 ± 6.0°, while the difference between the first out-of-brace and final radiographs was 1.8 ± 5.2°. The best predictor of final outcome was the first out-of-brace radiograph (0.80), compared to in-brace (0.68) and baseline (0.59) radiographs. The best cut-offs to predict avoidance of progression were 30% and 10% of the correction rates for the in-brace and first out-of-brace radiographs, respectively.

CONCLUSION

The first out-of-brace radiograph predicts end results better than the in-brace radiograph. It offers an excellent clinical reference for clinicians and patients. The first out-of-brace radiograph should be considered an essential element of future predictive models. LEVEL OF EVIDENCE 1: Diagnostic: individual cross-sectional studies with consistently applied reference standard and blinding.

Can immediate postoperative radiographs predict outcomes in pediatric clubfoot?

BACKGROUND

The goal of treatment for pediatric idiopathic clubfoot is to enable the patient to comfortably walk on his or her soles without pain. However, currently accepted treatment protocols are not always successful. Based on the abnormal bone alignment reported in this disease, some studies have noted a correlation between radiographic characteristics and outcome, but this correlation remains debated.

AIM

To assess the correlation between immediately postoperative radiographic parameters and functional outcomes and to identify which best predicts functional outcome.

METHODS

To predict the outcome and prevent early failure of the Ponseti’s method, we used a simple radiographic method to predict outcome. Our study included newborns with idiopathic clubfoot treated with Ponseti’s protocol from November 2018 to August 2022. After Achilles tenotomy and a long leg cast were applied, the surgeon obtained a single lateral radiograph. Radiographic parameters included the tibiocalcaneal angle (TiCal), talocalcaneal angle (TaCal), talofirst metatarsal angle (Ta1st) and tibiotalar angle (TiTa). During the follow-up period, the Dimeglio score and functional score were examined 1 year after surgery. Additionally, recurring events were reported. The correlation between functional score and radiographic characteristics was analyzed using sample and multiple logistic regression, and the optimal predictor was also identified.

RESULTS

In total, 54 feet received approximately 8 manipulations of casting and Achilles tenotomy at a mean age of 149 days. The average TiCal, TaCal, Ta1st, and TiTa angles were 75.24, 28.96, 7.61, and 107.31 degrees, respectively. After 12 mo of follow up, we found 66% excellent-to-good and 33.3% fair-to-poor functional outcomes. The Dimeglio score significantly worsened in the poor outcome group (P value < 0.001). Tical and TaCal showed significant differences between each functional outcome (P value < 0.05), and the TiCal strongly correlated with outcome, with a smaller angle indicating a better outcome, each 1 degree decrease improved the functional outcome by 10 percent. The diagnostic test revealed that a TiCal angle of 70 degrees predicts an inferior functional outcome.

CONCLUSION

The TiCal, derived from lateral radiographs immediately after Achilles tenotomy, can predict functional outcome at 1 year postoperatively, justifying its use for screening patients who need very close follow-up.

Distribution of frontal sinus pattern amongst Malaysian population: a skull radiograph study

Frontal sinus has unique anatomical features that are distinct to every population. However, the distribution of frontal sinus patterns has yet to be explored in the Malaysian population. This study aimed to describe the distribution of frontal sinus patterns among adult Malaysians. 409 adult Malaysian posteroanterior skull radiographs, consisting of 200 males and 209 females of Malay, Chinese, and Indian races aged between 20–69 years old, were included in the study. The frontal sinus patterns were classified according to total and percentage of presence or absence of frontal sinus, symmetry or asymmetrical (right or left dominant), unilateral absence (right or left), bilateral absence, and lobulation. The findings showed that bilateral presence of frontal sinus is common, in 95.4% of individuals and bilateral absence was noted in 2.7% individuals. Unilateral absence was found in 2.0% of individuals. Asymmetrical frontal sinus was observed in 54.5% of population meanwhile 40.8% showed symmetrical frontal sinus. The majority of individuals, regardless of sex, race, and age, possessed 1 to 3 lobes on both sides of the frontal sinus. The findings suggest that the frontal sinus is highly asymmetric, and the absence of the frontal sinus is rare. This morphological variation provides an insight into the landmarking placement for measurement during forensic application and assists neurosurgeons in surgical procedure to avoid breaching of the frontal sinus.

Measuring the critical shoulder angle on radiographs: an accurate and repeatable deep learning model

PURPOSE

Since the critical shoulder angle (CSA) is considered a risk factor for shoulder pathology and the intra- and inter-rater variabilities in its calculation are not negligible, we developed a deep learning model that calculates it automatically and accurately.

METHODS

We used a dataset of 8467 anteroposterior x-ray images of the shoulder annotated with 3 landmarks of interest. A Convolutional Neural Network model coupled with a spatial to numerical transform (DSNT) layer was used to predict the landmark coordinates from which the CSA was calculated. The performances were evaluated by calculating the Euclidean distance between the ground truth coordinates and the predicted ones normalized with respect to the distance between the first and the second points, and the error between the CSA angle measured by a human observer and the predicted one.

RESULTS

Regarding the normalized point distances, we obtained a median error of 2.9%, 2.5%, and 2% for the three points among the entire set. Considering CSA calculations, the median errors were 1° (standard deviation 1.2°), 0.88° (standard deviation 0.87°), and 0.99° (standard deviation 1°) for angles below 30°, between 30° and 35°, and above 35°, respectively.

DISCUSSION

These results demonstrate that the model has the potential to be used in clinical settings where the replicability is important. The reported standard error of the CSA measurement is greater than 2° that is above the median error of our model, indicating a potential accuracy sufficient to be used in a clinical setting.

The West China Hospital radiographic classification for fibrous dysplasia in femur and adjacent bones: A retrospective analysis of 205 patients

Objective

This study aims to investigate the reliability and clinical outcome of a newly developed classification system for patients with fibrous dysplasia (FD) of the femur and adjacent bones, optimizing its evaluation and management.

Methods

A total of 205 patients (121 female and 84 male) with FD in the femur and adjacent bones were included in this retrospective study. All affected femurs were measured and treated based on this classification at our institution between 2009 and 2019. Based on previous studies and extensive clinical follow‐up, we cautiously proposed the West China Hospital radiographic classification for FD in femur and adjacent bones following corresponding treatment options. There are five types with five radiographic features, including proximal femur bone loss, coxa vara, femoral shaft deformity, genu valgum, and hip arthritis. The intraobserver and interobserver reproducibility of this classification was assessed by four observers using the Cohen kappa statistic. The clinical outcome was evaluated using the criteria of Guille.

Results

At a median follow‐up of 60 months (range 6–120), 205 patients (median 34.8 years old, range 18–73 years old) were categorized into the following five types: Type I 31.7%, Type II 30.2%, Type III 20.5%, Type IV 10.2%, and Type V 7.3%. The mean interobserver and intraobserver kappa scores were 0.85 (range 0.77–0.89) and 0.85 (range 0.79–0.92), respectively. For clinical outcomes, there was no significant difference in the postoperative Guille score for Type I patients (mean 9.01 ± 1.22). There was a significant increase in the postoperative Guille score in Type II, III, IV, and V, compared to the preoperative values (P < 0.01). For complications, two Type III patients reported pain, and one Type III patient had mild‐to‐moderate Trendelenburg gait. One Type IV patient had a mild Trendelenburg gait. And two Type V patients still had mild limping.

Conclusion

This classification is reproducible and serves as a tool for evaluating and treating FD in the femur and adjacent bones. Therefore, we recommend this classification for the diagnosis and treatment of FD‐related deformities in the femur and adjacent bones.