Hand X-ray Research Articles

A Global-Local Feature Fusion Convolutional Neural Network for Bone Age Assessment of Hand X-ray Images

Bone age assessment plays a critical role in the investigation of endocrine, genetic, and growth disorders in children. This process is usually conducted manually, with some drawbacks, such as reliance on the pediatrician’s experience and extensive labor, as well as high variations among methods. Most deep learning models use one neural network to extract the global information from the whole input image, ignoring the local details that doctors care about. In this paper, we propose a global-local feature fusion convolutional neural network, including a global pathway to capture the global contextual information and a local pathway to extract the fine-grained information from local patches. The fine-grained information is integrated into the global context information layer-by-layer to assist in predicting bone age. We evaluated the proposed method on a dataset with 11,209 X-ray images with an age range of 4–18 years. Compared with other state-of-the-art methods, the proposed global-local network reduces the mean absolute error of the estimated ages to 0.427 years for males and 0.455 years for females; the average accuracy rate is within 6 months and 12 months, reaching 70% and 91%, respectively. In addition, the effectiveness and rationality of the model were verified on a public dataset.

Serum 14-3-3η as predictor of clinical remission and progression of structural damage in early rheumatoid arthritis following a treat-to-target strategy in a randomized controlled trial

Objective 14-3-3η is a proinflammatory mediator critical to joint destruction in rheumatoid arthritis (RA). We aimed to evaluate serum 14-3-3η for predicting disease activity and radiographic progression in patients with early RA in the double-blinded, randomized OPERA trial. Method 180 patients with early RA were randomized to receive methotrexate (MTX) + adalimumab or MTX + placebo in combination with glucocorticoid injections into swollen joints. Disease activity was measured using the 28-joint Disease Activity Score–C-reactive protein (DAS28-CRP). Clinical remission was defined as DAS28-CRP < 2.6. X-rays of hands and feet were evaluated by the Total Sharp van der Heijde score (TSS). Radiographic progression was defined as exceeding the smallest detectable change (1.8 TSS-units). Serum 14-3-3η was determined by enzyme-linked immunosorbent assay. Multivariate logistic regression models were used to identify predictors of DAS28-CRP remission at 6 months and radiographic progression at 12 months. Results Baseline 14-3-3η was a borderline significant independent predictor of radiographic progression at 12 months (odds radio = 1.02, 95% confidence interval 1.00–1.03, p = 0.05). In anti-cyclic citrullinated peptide antibody (ACPA)-negative patients, a moderate/high baseline 14-3-3η concentration increased the risk of radiographic progression at 12 months [4/51 (8%) vs 3/9 (33%), χ2 = 4.823, p = 0.028]. No value of 14-3-3η for predicting achievement of clinical remission was found. Conclusion Serum 14-3-3η was a borderline significant predictor of radiographic progression, particularly in ACPA-negative patients, but not of predicting achievement of clinical remission. Optimal cut-off levels of 14-3-3η for predicting radiographic progression in RA need further clarification.

Automated evaluation of rheumatoid arthritis from hand radiographs using Machine Learning and deep learning techniques

The aim and objectives of the study are as follows: (i) to implement automated patch-based classification of hand X-ray images using modified pre-trained convolutional neural network (CNN) models; (ii) to develop a customized CNN model for automated feature extraction and classification of hand X-ray images and to compare the performance of customized CNN models with non-linear and linear kernels; (iii) to construct the hand crafted feature fusion (SIFT+ Customized CNN features) and categorize the normal and RA using Machine Learning classifiers. The model was trained on 75 images (10,000 patches) of hand radiographs and tested using 25 images (500 patches) that were not included in the training set. The accuracy of the modified pre-trained model GoogLeNet was 89% and the proposed custom model three achieved an accuracy of 95%. The sensitivity and specificity of GoogLeNet were 84% and 90% respectively. The custom model three attained the sensitivity and specificity as 95% and 94% respectively. Furthermore, when compared to the features extracted (SIFT + CNN) from the customized models, the custom3 model outperformed well for the classification of RA compared to ML classifiers. Thus a custom CNN-based computer-aided diagnostic tool can be used as an effective method for the detection of RA.

Folate deficiency is associated with increased radiographic severity of osteoarthritis in knee joints but not in hand joints.

No previous studies have explored the effect of folate deficiency on the severity of osteoarthritis (OA). Therefore, we investigated the relationship between folate level and features on knee and hand radiographs in a large, population-based OA cohort. Among 9,260 subjects enrolled in the Dong-gu study, 2,489 who had knee and hand joint radiographs were included. Of these, subjects with a history of amputation or total knee replacement were excluded. Serum folate levels were measured using blood samples collected at the time of enrolment and stored. A semi-quantitative system was used to grade the severity of hand and knee x-ray changes. Linear regression was performed to assess relationships between serum folate levels and knee and hand radiographic scores after adjusting for age, sex, body mass index, smoking, alcohol consumption, education, physical activity, occupation, vitamin D, and ferritin. A total of 2,322 subjects were recruited. After adjusting for confounders, participants with folate deficiency (<4 ng/mL) had higher total (p<0.001), osteophyte (p<0.001), joint space narrowing (p=0.002), tibial attrition (p<0.001), and sclerosis (p=0.005) scores for knee joint radiographs compared to participants with a normal folate level. After adjusting for confounders, the radiographic scores for hand joints did not differ between the groups. Folate deficiency is associated with increased radiographic severity of OA in knee joints, but not in hand joints. Further studies are needed to explore the differential effects of folate on the severity of knee and hand OA.

Reliability of biological maturation analyses performed by equations predicting skeletal age and peak height velocity with hand and wrist X-ray results.

Hand-wrist radiography is the gold standard for the assessment of skeletal age (SA) and the assignment of biological maturation (BM) status. However, its practical use in the field is questioned, prompting the development of anthropometric equations to estimate BM. However, there is no consensus on the reliability of these equations in relation to hand-wrist X-rays. To test the reliability of the BM estimated from predictive equations with the results of hand-wrist X-rays. One-hundred forty-four young athletes (38%-males, 62%-females; 12.3 ± 1.4 years) were recruited. SA was assessed by analyzing X-rays of the hand-wrist using the Greulich-Pyle protocol. SA was predicted using the anthropometric equations of Cabral and Macêdo. By subtracting chronological age from SA, a measure of BM or biological age (BA) was computed. In addition, a BA of years from peak height velocity (PHV) was predicted using the anthropometric equations of Mirwald and Moore. Relative to X-ray, Cabral's equations showed high reliability in estimating BM in both sexes (p < .0001) while Macêdo's equations showed substantial reliability or males (p < .0001) only. Mirwald and Moore's equations showed weak reliability (p < .05 and p=.043, respectively). The prediction of SA using anthropometric equations was found to be reliable in categorizing individuals as late, synchronized, and/or early maturers in both sexes. However, the predictive equations of years from PHV only reliably categorized late and synchronized maturation stages in males and the synchronized maturers in females.

Attention mechanism-based deep learning method for hairline fracture detection in hand X-rays.

Wrist and finger fractures detection is always the weak point of associate study, because there are small targets in X-rays, such as hairline fractures. In this paper, a dataset, consisting of 4346 anteroposterior, lateral and oblique hand X-rays, is built from many orthopedic cases. Specifically, it contains a lot of hairline fractures. An automatic preprocessing based on generative adversative network (GAN) and a detection network, called WrisNet, are designed to improve the detection performance of wrist and finger fractures. In the preprocessing, an attention mechanism-based GAN is proposed for obtaining the approximation of manual windowing enhancement. A multiscale attention-module-based generator of the GAN is proposed to increase continuity between pixels. The discriminator and the generator can achieve 93% structural similarity (SSIM) as manual windowing enhancement without manual parameter adjustment. The designed WrisNet is composed of two components: a feature extraction module and a detection module. A group convolution and a lightweight but efficient triplet attention mechanism are elaborately embedded into the feature extraction module, resulting in richer representations of hairline fractures. To obtain more accurate locating information in this condition, the soft non-maximum suppression algorithm is employed as the post-processing method of the detection module. As shown in experimental results, the designed method can have obvious average precision (AP) improvement up to 7% or more than other mainstream frameworks. The automatic preprocessing and the detection net can greatly reduce the degree of artificial intervention, so it is easy to be implemented in real clinical environment.

Association of Serum Osteoprotegerin Level With Myocardial Injury and Cardiovascular Calcification in Chronic Kidney Disease Patients.

BackgroundChronic kidney disease has emerged as a significant independent risk factor for cardiovascular disease. Cardiovascular calcification is an active process involving a complex interaction of inducers and inhibitors. High sensitivity cardiac troponin T assay detects troponin T with higher sensitivity and precision at an earlier point of time than the conventional assays, and is associated with poor outcomes. Serum osteoprotegerin is classed as an inhibitory factor for cardiovascular calcification. It is involved in the pathological processes of vascular damage and linked to the excess cardiovascular morbidity. The aim of the present study was to evaluate the extent of cardiovascular calcification and serum high sensitivity cardiac troponin T level, and their association with serum osteoprotegerin level in patients with chronic kidney disease stages 3–5.Methods90 chronic kidney disease patients were enrolled in this study, and they were divided into two groups: group (1) included 45 non-dialysis-dependent chronic kidney disease patients (stages 3–5) and group (2) included 45 chronic hemodialysis patients. Each group further subdivided according to the presence of cardiovascular calcification into subgroup A and B. Vascular calcifications were assessed by lateral lumbar, pelvis and hands X-ray radiographs. Valvular calcification was assessed by echocardiography. Serum cardiac troponin T was measured by high sensitivity assay and serum osteoprotegerin was measured by ELISA.ResultsCardiovascular calcification distribution was 22.2% in group (1) and 33.3% in group (2). Serum osteoprotegerin and troponin T in calcification groups (1A and 2A) were significantly higher than non-calcification groups (1B and 2B; P < 0.001). Osteoprotegerin correlated positively with high sensitivity cardiac troponin T (rs = 0.72, P < 0.001). cardiovascular calcification correlated positively with osteoprotegerin, troponin T, and phosphorus. osteoprotegerin and phosphorus were significant independent predictors of cardiovascular calcification at cut-off values ≥4.6 ng/L and ≥6.95 mg/dl, respectively (P < 0.001). Serum phosphorus and creatinine were independent predictors of osteoprotegerin (P < 0.001 and 0.048, respectively).ConclusionOsteoprotegerin is strongly associated with cardiovascular calcification and high sensitivity cardiac troponin T. In addition, there is a positive association between calcification and troponin T. This suggests a role for osteoprotegerin in the pathogenesis and risk stratification of cardiovascular calcification and myocardial injury in chronic kidney disease patients with a potential role as a therapeutic target.

Use of primary health care and radiological imaging preceding a diagnosis of rheumatoid arthritis: a nationwide cohort study in Denmark.

To explore health-care use in the 12 months preceding a diagnosis of RA in Denmark. We conducted a population-based cohort study using data from national registries. Every patient diagnosed with RA in 2014-18 was matched to 10 reference individuals without RA from the Danish background population. Health-care use was defined as contacts to general practitioners, contacts to private practicing physiotherapists, and X-rays of hands and/or feet performed in primary or secondary care. We estimated the monthly contact rates for patients and references in the 12 months preceding the diagnosis, and we compared incidence rates for health-care use in each month between the two groups while adjusting for sociodemographic characteristics and comorbidity. We included 7427 patients with RA and 74 270 references. Patients with RA had increasing contact rates with general practitioners and physiotherapists from 6 to 8 months before the diagnosis. Compared with references, women with RA had statistically significantly more contacts to general practitioners and physiotherapists during all 12 months. A similar contact pattern was seen in men, albeit less distinct. The number of X-rays increased slightly from 8 months before the diagnosis, with a steep increase in the last 3 months. Increased contacts to general practitioners and physiotherapists were seen in all 12 months preceding the RA diagnosis, intensifying in the last 6 to 8 months. Imaging increased from 3 months before the diagnosis. This indicates an opportunity to expedite referral to specialist care and ensure earlier diagnosis of RA.

Deep Learning-Based Computer-Aided Diagnosis of Rheumatoid Arthritis with Hand X-ray Images Conforming to Modified Total Sharp/van der Heijde Score

Introduction: Rheumatoid arthritis (RA) is a systemic autoimmune disease; early diagnosis and treatment are crucial for its management. Currently, the modified total Sharp score (mTSS) is widely used as a scoring system for RA. The standard screening process for assessing mTSS is tedious and time-consuming. Therefore, developing an efficient mTSS automatic localization and classification system is of urgent need for RA diagnosis. Current research mostly focuses on the classification of finger joints. Due to the insufficient detection ability of the carpal part, these methods cannot cover all the diagnostic needs of mTSS. Method: We propose not only an automatic label system leveraging the You Only Look Once (YOLO) model to detect the regions of joints of the two hands in hand X-ray images for preprocessing of joint space narrowing in mTSS, but also a joint classification model depending on the severity of the mTSS-based disease. In the image processing of the data, the window level is used to simulate the processing method of the clinician, the training data of the different carpal and finger bones of human vision are separated and integrated, and the resolution is increased or decreased to observe the changes in the accuracy of the model. Results: Integrated data proved to be beneficial. The mean average precision of the proposed model in joint detection of joint space narrowing reached 0.92, and the precision, recall, and F1 score all reached 0.94 to 0.95. For the joint classification, the average accuracy was 0.88, and the accuracy of severe, mild, and healthy reached 0.91, 0.79, and 0.9, respectively. Conclusions: The proposed model is feasible and efficient. It could be helpful for subsequent research on computer-aided diagnosis in RA. We suggest that applying the one-hand X-ray imaging protocol can improve the accuracy of mTSS classification model in determining mild disease if it is used in clinical practice.

OP0263 FAVORABLE BALANCE OF BENEFIT AND HARM OF LONG-TERM, LOW-DOSE PREDNISOLONE ADDED TO STANDARD TREATMENT IN RHEUMATOID ARTHRITIS PATIENTS AGED 65+: THE PRAGMATIC, MULTICENTER, PLACEBO- CONTROLLED GLORIA TRIAL

OP0263 FAVORABLE BALANCE OF BENEFIT AND HARM OF LONG-TERM, LOW-DOSE PREDNISOLONE ADDED TO STANDARD TREATMENT IN RHEUMATOID ARTHRITIS PATIENTS AGED 65+: THE PRAGMATIC, MULTICENTER, PLACEBO- CONTROLLED GLORIA TRIAL

Osteopetrosis: A rare case

Osteopetrosis is a rare inherited metabolic bone disease characterized by failure of osteoclasts to resorb bone leading to impairment of bone modeling and remodeling. The defect in bone turnover results in skeletal fragility despite increased bone mass, hematopoietic insufficiency, disturbed tooth eruption, nerve entrapment syndromes, and growth failure. It has two major clinical forms: an autosomal dominant benign type or an autosomal recessive malignant type. The recessive form is usually fatal whereas the dominant type is more compatible with life. A rare autosomal recessive (intermediate form) is more prevalent and has less severe presentation. A 12-year-old boy presented with short stature, bilateral progressive vision loss, and backache. X-ray of nondominant hand for bone age estimation showed bone in bone formation. Followed which skeletal survey was done which showed characteristic radiological findings suggestive of osteopetrosis. Osteopetrosis is a rare disease transmitted by autosomal dominant or recessive inheritance having variable penetrance. We report here intermediate form of osteopetrosis. Although the genetic test is used to differentiate between the subtypes, diagnosis is mainly radiological.

Accuracy and self-validation of automated bone age determination

The BoneXpert method for automated determination of bone age from hand X-rays was introduced in 2009 and is currently running in over 200 hospitals. The aim of this work is to present version 3 of the method and validate its accuracy and self-validation mechanism that automatically rejects an image if it is at risk of being analysed incorrectly. The training set included 14,036 images from the 2017 Radiological Society of North America (RSNA) Bone Age Challenge, 1642 images of normal Dutch and Californian children, and 8250 images from Tübingen from patients with Short Stature, Congenital Adrenal Hyperplasia and Precocious Puberty. The study resulted in a cross-validated root mean square (RMS) error in the Tübingen images of 0.62 y, compared to 0.72 y in the previous version. The RMS error on the RSNA test set of 200 images was 0.45 y relative to the average of six manual ratings. The self-validation mechanism rejected 0.4% of the RSNA images. 121 outliers among the self-validated images of the Tübingen study were rerated, resulting in 6 cases where BoneXpert deviated more than 1.5 years from the average of the three re-ratings, compared to 72 such cases for the original manual ratings. The accuracy of BoneXpert is clearly better than the accuracy of a single manual rating. The self-validation mechanism rejected very few images, typically with abnormal anatomy, and among the accepted images, there were 12 times fewer severe bone age errors than in manual ratings, suggesting that BoneXpert could be safer than manual rating.

FINGER JOINT SEGMENTATION USING MACHINE LEARNING AND MINIMIZED TRAINING SET

Purpose: Hand Osteoarthritis (HOA) is a common form of OA disease. About half of all women and one-quarter of all men will experience the stiffness and pain of HOA by the time they are 85 years old. The 29 bones of hands and wrists come together to form many small joints that can be affected by HOA. One common HOA evaluation is to measure joint space width (JSW) on hand X-ray for each finger joint. This measurement usually requires manual labeling of finger joint bone edge, which is time-consuming.

The inter-and intra-observer reliability of volar angulation measurements in a fifth metacarpal neck fracture.

The fifth metacarpal neck fracture is the most common metacarpal fracture. The palmar angulation from the fracture displacement is critical for determining treatment, yet there is no consensus regarding the angulation measurement method or the surgical cut-off value. This study aimed to identify a reliable measurement method for assessing palmar angulation. We evaluated inter-observer and intra-observer validation of measuring palmar angulation in oblique plain X-ray and computed tomography (CT) sagittal cuts. We identified surgically treated patients for acute isolated fifth metacarpal neck fracture between January 1, 2008, and December 31, 2020, and obtained preoperative, opposite hand, and final follow-up oblique X-rays and sagittal computed tomography (CT) radiograms. The oblique radiograph was taken with a 45° posteroanterior pronation. The metacarpal neck palmar angulation was measured in the radiograms using the metacarpal neck-shaft center (MNSC) angle and the shaft articular surface (SAS) angle methods by three orthopedic surgeons in two sessions. For the CT radiograms, each measurer selected the sagittal slot at their discretion to measure the angle. The final palmar angulation was the average of six measurements (two sessions, three measurers per session). The study included 51 patients; the average age was 32.5 (range 18-73) years, with 46 men and 5 women. The MNSC angle inter-observer reliability was better than the SAS angle. The MNSC angle inter-observer reliability was better than that of SAS angle. Intraclass coefficients (ICCs) for the MNSC angle demonstrated an excellent inter-observer agreement among the three measurers in the first (0.93) and second (0.88) session compared to ICCs for the SAS angle in the first (0.81) and second (0.87) session. The MNSC angle intra-observer reliability was also better than the SAS angle, with higher ICCs. Preoperative CT radiograms were available for 42 patients. Using CT scans for measurements, in the two sessions, the MNSC angle inter-observer reliability was higher than that of the SAS angle [MNSC: 0.83; SAS: 0.35], second [MSNC: 0.85; SAS: 0.81]. The intra-observer reliability was also better in the MNSC angle. When comparing average value among obtained radiograms, the physiologic angulation of the opposite hand oblique X-ray had the smallest average value, followed by preoperative CT and preoperative oblique radiography. Overall, the SAS angle measurement had a slightly larger angle than the MNSC method in the fractured and non-fractured hand measurements. Finally, a serial comparison of the oblique X-rays (pre-and postoperative, final follow-up, and the opposite hand with closed reduction and internal fixation) indicated that the angulation significantly decreased, and the post-operative values did not differ from the final follow-up X-ray for either method. The palmar angulation measurement in 45° pronated oblique X-ray using the MNSC angle method had good-to-excellent reliability, with superior results to sagittal CT radiograms. Although the angle is likely overestimated, the MNSC method is reliable for judging the fracture degree and reduction adequacy after surgery compared to the non-fractured hand physiologic angulation.

The Verification of the Degree of Concordance of the SMI/CVMS Indexes in Evaluating the Pubertal Growth Stages—Longitudinal Study

The research aims to verify the concordance between the skeletal maturity index (SMI) measured on the hand and wrist X-rays using Fishman method and the cervical vertebral maturation stage (CVMS), measured on the lateral cephalometric X-rays using Baccetti method. The concordance of the two indexes (SMI and CVMS) has been statistically verified with the help of the Cohen’s kappa coefficient, by relating them to the growth stages, within a longitudinal study done upon a group of 38 patients, 22 female and 16 male, aged between 8–18 y, the analyzed investigations being done in series, along the orthodontic treatment. The research showed a strong correlation between the SMI and CVMS indexes within the analyzed group, confirmed by the obtained values (k = 0.84 for female and k = 0.85 for male).

Osteo-Proliferative Lesions of the Phalanges on Radiography: Associations with Sex, Age, and Osteoarthritis.

Objectives: The effects of aging such as osteophyte formation, acral shape changes, cortical tunneling, and bone porosity as well as enthesophytes can be studied in the X-rays of hands. However, during the interpretation of radiographs of the hands, misinterpretation and false-positive findings for psoriatic arthritis often occur because periosteal proliferations of the phalanges are overinterpreted and too little is known about enthesophytes of the phalanges in this area. Method: It included a total of 1153 patients (577 men, 576 women) who presented themselves to the emergency department and received a radiography of their right hand to exclude fractures. The Osseographic Scoring System was used in a modified form to record osteophytes and enthesophytes. A linear regression model for periosteal lesions was computed with age, sex, osteophytes, and global diagnosis as covariables. The inter-reader agreement was assessed using ICC (two-way mixed model) on the sum scores of osteophytes and periosteal lesions. Results: Overall, men exhibited more periosteal lesions, demonstrated by a higher mean sum score of 4.14 vs. 3.21 in women (p = 0.008). In both sexes, the second and third proximal phalanx were most frequently affected by periosteal lesions, but the frequencies were significantly higher in men. The female sex was negatively associated with an extent of periosteal lesions with a standardized beta of −0.082 (p = 0.003), while age and osteophytes were positively associated with betas of 0.347 (p < 0.001) and 0.156 (p < 0.001), respectively. The distribution of osteophytes per location did not differ between men and women (p > 0.05). The inter-reader agreement was excellent for periosteal lesions with ICC of 0.982 (95%CI 0.973–0.989, p < 0.001). Conclusions: Special care should be taken not to confuse normal periosteal changes in aging with periosteal apposition in psoriatic arthritis.

An Efficient CNN for Hand X-Ray Overall Scoring of Rheumatoid Arthritis

Rheumatoid arthritis (RA) is a progressive systemic autoimmune disease characterized by inflammation of the joints and surrounding tissues, which seriously affects the life of patients. The Sharp/van der Heijde method has been widely used in clinical evaluation for the RA disease. However, this manual method is time-consuming and laborious. Even if two radiologists evaluate a specific location, their subjective evaluation may lead to low inter-rater reliability. Here, we developed an efficient model powered by deep convolutional neural networks to solve these problems and automated the overall scoring on hand X-rays. The depthwise separable (Dwise) convolution technique is used based on ResNet-50 due to the high resolution of hand X-rays. An inverted residual block is introduced to devise a ResNet-Dwise50 model to enhance the efficiency of the model. The model was trained and tested using bilateral posteroanterior (two-handed, side by side) images of 3818 patients. The experiment results show the ResNet-Dwise50 model achieved an MAE of 14.90 and RMSE of 22.01 while ensuring high efficiency. There was no statistical difference between the average scores given by two experienced radiologists and predicted scores from our model.

Subcutaneous and visceral emphysema due to industrial air gun injury

A 21-year-old man presented to our emergency department suffering for right arm tightness and stiffness with mild pain following a high-pressure air working injury. On examination, the right hand palm showed a small gunshot wound (Figure 1), a swelling and crepitus was palpable from the tips of the fingers through to the whole upper limb extending up to the lateral right cervical region. No neurological impairments or fever were recorded. His vital signs were stable and laboratory tests were within a normal range (WBC 8.39 109/L mmc, CRP 0.10 mg/dL, fibrinogen 245 mg/dL). The wound was irrigated, cleaned and sterile dressing was applied. The right hand x-ray resulted negative for acute fracture.

Detection of turner syndrome using hand X-ray using anchor based links segmentation method.

Turner Syndrome (TS) is a chromosomal disorder, wherein the female's growth is impacted. Immature ovaries, low stature, and heart abnormalities are a range of developmental and medical issues due to TS. The condition of TS might be detected prior to birth, throughout infancy or in the early years of life. The diagnosis of TS in girls with modest symptoms and indications is sometimes deferred until they reach adolescence or become young adults. This study presents an algorithm to segment the hand digital X-ray image in children with TS. In medical image and computer vision examination, image segmentation is demanding, and very crucial. Prevailing segmentation algorithms even now suffer from common segmentation issues including under-segmentation, over-segmentation, and spurious or non-closed edges, regardless of the several years of studies. In this paper, Anchor Based Link (ABL) segmentation approach is proposed to detect TS based on fourth Metacarpal bone from left hand X-ray images. The detection of TS is demonstrated based upon the comparison of proposed approach with existing watershed segmentation and Gaussian-Mixture-Model-based Hidden-Markov-Random-Field (GMM-HMRF) method. The proposed approach attains better segmentation based on the ratio of height and width of left fourth finger that is analyzed for normal children and children having TS with the help of edge pixel present in the metacarpal bone that has been segmented. The suggested method is verified on fifty (50) sample X-ray hand images of carpal bones, providing 0.60 ± 0.02 as an average Dice coefficient.

Automatic Carpal Site Detection Method for Evaluation of Rheumatoid Arthritis Using Deep Learning

Approximately 600,000 to 1,000,000 patients are diagnosed with rheumatoid arthritis (RA) in Japan. To provide appropriate treatment, it is necessary to accurately measure the progression of RA by diagnosing the disease several times a year. The modified total sharp score (mTSS) calculated from hand X-ray images is a standard diagnostic method for RA progression. However, this diagnostic method is time-consuming as the scores are rated at as many as 16 points per hand. Accordingly, in order to shorten the diagnosis time of RA patients and improve the quality of diagnosis, the development of computer-aided diagnosis (CAD) systems is expected. We have previously proposed a CAD system that can detect finger joint positions using a support vector machine and can estimate the mTSS using ridge regression. In this study, we propose a fully automatic detection method of RA score evaluation points in the carpal site from simple hand X-ray images using deep learning. The proposed method first segments the carpal site using deep learning. Next, the RA evaluation points are automatically determined from each segment based on prior knowledge. Experimental results on X-ray images of the hands of 140 patients with RA showed that the mTSS evaluation point at the carpal site could be detected with an average error of 25 pixels. This study enables the automatic detection of RA score evaluation points in the carpal site. In the diagnosis of RA, the time required for diagnosis can be reduced by automating the determination of diagnostic points by physician.