Radiomics and machine learning analysis of liver magnetic resonance imaging for prediction and early detection of tumor response in colorectal liver metastases

Purpose@#The aim of this study was to demonstrate the effectiveness of a machine learning-based radiomics model for distinguishing tumor response and overall survival in patients with unresectable colorectal liver metastases (CRLM) treated with targeted biological therapy. @*Methods@#We prospectively recruited 17 patients with unresectable liver metastases of colorectal cancer, who had been given targeted biological therapy as the first line of treatment. All patients underwent liver magnetic resonance imaging (MRI) three times up until 8 weeks after chemotherapy. We evaluated the diagnostic performance of machine learning-based radiomics model in tumor response of liver MRI compared with the guidelines for the Response Evaluation Criteria in Solid Tumors. Overall survival was evaluated using the Kaplan-Meier analysis and compared to the Cox proportional hazard ratios following univariate and multivariate analyses. @*Results@#Performance measurement of the trained model through metrics showed the accuracy of the machine learning model to be 76.5%, and the area under the receiver operating characteristic curve was 0.857 (95% confidence interval [CI], 0.605–0.976; P < 0.001). For the patients classified as non-progressing or progressing by the radiomics model, the median overall survival was 17.5 months (95% CI, 12.8–22.2), and 14.8 months (95% CI, 14.2–15.4), respectively (P = 0.431, log-rank test). @*Conclusion@#Machine learning-based radiomics models could have the potential to predict tumor response in patients with unresectable CRLM treated with biologic therapy.

Automatic Segmentation and Radiologic Measurement of Distal Radius Fractures Using Deep Learning

Background@#Recently, deep learning techniques have been used in medical imaging studies. We present an algorithm that measures radiologic parameters of distal radius fractures using a deep learning technique and compares the predicted parameters with those measured by an orthopedic hand surgeon. @*Methods@#We collected anteroposterior (AP) and lateral X-ray images of 634 wrists in 624 patients with distal radius fractures treated conservatively with a follow-up of at least 2 months. We allocated 507 AP and 507 lateral images to the training set (80% of the images were used to train the model, and 20% were utilized for validation) and 127 AP and 127 lateral images to the test set. The margins of the radius and ulna were annotated for ground truth, and the scaphoid in the lateral views was annotated in the box configuration to determine the volar side of the images. Radius segmentation was performed using attention U-Net, and the volar/dorsal side was identified using a detection and classification model based on RetinaNet. The proposed algorithm measures the radial inclination, dorsal or volar tilt, and radial height by index axes and points from the segmented radius and ulna. @*Results@#The segmentation model for the radius exhibited an accuracy of 99.98% and a Dice similarity coefficient (DSC) of 98.07% for AP images, and an accuracy of 99.75% and a DSC of 94.84% for lateral images. The segmentation model for the ulna showed an accuracy of 99.84% and a DSC of 96.48%. Based on the comparison of the radial inclinations measured by the algorithm and the manual method, the Pearson correlation coefficient was 0.952, and the intraclass correlation coefficient was 0.975. For dorsal/ volar tilt, the correlation coefficient was 0.940, and the intraclass correlation coefficient was 0.968. For radial height, it was 0.768 and 0.868, respectively. @*Conclusions@#The deep learning-based algorithm demonstrated excellent segmentation of the distal radius and ulna in AP and lateral radiographs of the wrist with distal radius fractures and afforded automatic measurements of radiologic parameters.

Automatically Diagnosing Skull Fractures Using an Object Detection Method and Deep Learning Algorithm in Plain Radiography Images

Objective@#: Deep learning is a machine learning approach based on artificial neural network training, and object detection algorithm using deep learning is used as the most powerful tool in image analysis. We analyzed and evaluated the diagnostic performance of a deep learning algorithm to identify skull fractures in plain radiographic images and investigated its clinical applicability. @*Methods@#: A total of 2026 plain radiographic images of the skull (fracture, 991; normal, 1035) were obtained from 741 patients. The RetinaNet architecture was used as a deep learning model. Precision, recall, and average precision were measured to evaluate the deep learning algorithm’s diagnostic performance. @*Results@#: In ResNet-152, the average precision for intersection over union (IOU) 0.1, 0.3, and 0.5, were 0.7240, 0.6698, and 0.3687, respectively. When the intersection over union (IOU) and confidence threshold were 0.1, the precision was 0.7292, and the recall was 0.7650. When the IOU threshold was 0.1, and the confidence threshold was 0.6, the true and false rates were 82.9% and 17.1%, respectively. There were significant differences in the true/false and false-positive/false-negative ratios between the anteriorposterior, towne, and both lateral views (p=0.032 and p=0.003). Objects detected in false positives had vascular grooves and suture lines. In false negatives, the detection performance of the diastatic fractures, fractures crossing the suture line, and fractures around the vascular grooves and orbit was poor. @*Conclusion@#: The object detection algorithm applied with deep learning is expected to be a valuable tool in diagnosing skull fractures.

Excited Skin Syndrome after Patch Testing in a Patient with Atopic Dermatitis Treated with Dupilumab / 대한피부과학회지

The effect of dupilumab on allergic contact dermatitis or patch testing is unknown. Here, we present a case of excited skin syndrome that appeared after patch testing during dupilumab treatment. A 21-year-old woman presented with atopic dermatitis that showed only a partial response to cyclosporine and dupilumab. A patch test was performed to check for underlying allergic contact dermatitis. The result was consistent with excited skin syndrome.Some studies argue that dupilumab suppresses allergic reactions triggered by allergens that activate the Th2 pathway. Others suggest that it does not affect the result of patch testing, regardless of the type of allergen tested. Even if dupilumab suppresses a certain allergic or immunologic pathway, this case shows that it cannot mask excited skin syndrome on patch testing.

Clinical Implications of Expression of Endoglin and Survivin in Patients with Primary Cutaneous Angiosarcoma in an Asian Population / 대한피부과학회지

Background@#Targeted therapy and immunotherapy such as programmed death-1 (PD-1) targeting have been introduced for treating many types of cancers, including primary cutaneous angiosarcoma (CA). However, studies that examined other targeted molecules in CA are scarce. @*Objective@#We aim to declare the expression of endoglin and survivin in addition to PD-1 and assess the clinical correlation between the expression of these molecules and clinical variables, overall survival (OS), and progressionfree survival (PFS) in CA. @*Methods@#We identified 51 patients diagnosed with CA at Asan Medical Center over the last 14 years, based on the staining results of paraffin sections of tissue samples for endoglin, survivin, and PD-1 that were reviewed by two dermatologists. @*Results@#Statistical analysis for the correlation between results and clinical data of CA revealed that whereas 35 (63.6%) and 30 samples (54.5%) were positive for endoglin and survivin respectively, only nine samples were positive for PD-1 (16.4%). Co-expression of endoglin and survivin was detected in 24 lesions (p=0.013) and was significantly correlated to head, neck, face, and scalp (HNFS) lesions in CA (p=0.005, p=0.038, respectively). However, the expression of these target molecules did not correlate with the OS or PFS of CA. @*Conclusion@#Considering that HNFS type CA is associated with unfavorable clinical outcomes in similar populations, our findings can be helpful in matching patients with CA with effective targeted therapy.

Trends of Dental Treatment under General Anesthesia and Patterns of Repeated General Anesthesia in Children and Adolescents with Severe Disabilities / 대한소아치과학회지

The aim of this retrospective study was to evaluate the demographic characteristics of pediatric dental patients who underwent dental treatment under general anesthesia (DTGA) at the Seoul National University Dental Hospital from January 2011 through December 2020 and compare the patterns of repeated DTGA between dental patients with severe disabilities (DSD) and non-DSD (healthy or medically compromised patients without DSD). There were 1,857 DTGAs among 1,719 patients (mean age = 5.1 years; males = 59.3%; ASA 2 or above = 52.9%; DSD = 26.8%). Overall, 6.6% of patients underwent repeated DTGA, and the rate of repeated DTGA over a 10-year period was 7.4%. ASA 2 or above (p < 0.0001) and DSD (p < 0.0001) were more likely to undergo repeated DRGA compared to ASA 1 and non-DSD. At both GA1 and GA2, DSD received significantly more restorative treatment on permanent teeth than non-DSD (p = 0.002, p < 0.0001, respectively). There has been an increasing demand for DTGA in pediatric dentistry over the last 10 years. Regular check-ups and preventive oral health care are necessary for pediatric dental patients with severe disabilities to reduce the possibility of repeated DTGA.

Detection and Weak Segmentation of Masses in Gray-Scale Breast Mammogram Images Using Deep Learning

Purpose@#In this paper, we propose deep-learning methodology with which to enhance the mass differentiation performance of convolutional neural network (CNN)-based architecture. @*Materials and Methods@#We differentiated breast mass lesions from gray-scale X-ray mammography images based on regions of interest (ROIs). Our dataset comprised breast mammogram images for 150 cases of malignant masses from which we extracted the mass ROI, and we composed a CNN-based deep learning model trained on this dataset to identify ROI mass lesions. The test dataset was created by shifting some of the training data images. Thus, although both datasets were different, they retained a deep structural similarity. We then applied our trained deep-learning model to detect masses on 8-bit mammogram images containing malignant masses. The input images were preprocessed by applying a scaling parameter of intensity before being used to train the CNN model for mass differentiation. @*Results@#The highest area under the receiver operating characteristic curve was 0.897 (Î 20). @*Conclusion@#Our results indicated that the proposed patch-wise detection method can be utilized as a mass detection and segmentation tool.

Clinicopathologic Characteristics of Nipple Adenoma: Comparison with Malignant Proliferative Diseases / 대한피부과학회지

Background@#Nipple adenoma (NA) is a rare benign tumor arising in the lactiferous ducts of the nipple. It typically presents as a palpable nodule, erosion, or discharge with erythema of the nipple. NA is different from other mammary proliferative diseases of the nipple; however, its clinicopathologic characteristics have been scarcely elucidated. @*Objective@#In this study, we aimed to assess the clinical and histopathological characteristics of NA and compare them with mammary Paget’s diseases and breast carcinomas of the nipple. @*Methods@#We retrospectively reviewed fifteen patients with NA. Furthermore, we reviewed fifteen patients with nipple Paget’s diseases and five patients with breast carcinomas (ductal carcinoma in situ and invasive ductal carcinoma). Skin lesions’ clinical characteristics and general histopathological findings were investigated. @*Results@#NA showed significantly early onset (p=0.014), delayed time for onset to diagnosis (p=0.026), and smaller lesion than other nipple malignant diseases (p＜0.001). NA was predominantly localized on the right side and exhibited as more palpable mass and less nipple discharge as initial symptoms. Estimated prevalence of Korean cases (0.026%) was twice higher than Western countries (0.012%). p16 immunostaining in NA and other malignant diseases did not differ. @*Conclusion@#NA is a benign neoplasm arising on the nipple. NA showed earlier onset with smaller size at initial presentation than other malignant diseases which presented more crusts. Unnecessary surgical procedures for NA should be avoided with preceding clinical differential diagnosis.

Detection of Proximal Caries Lesions with Deep Learning Algorithm / 대한소아치과학회지

This study aimed to evaluate the effectiveness of deep convolutional neural networks (CNNs) for diagnosis of interproximal caries in pediatric intraoral radiographs. A total of 500 intraoral radiographic images of first and second primary molars were used for the study. A CNN model (Resnet 50) was applied for the detection of proximal caries. The diagnostic accuracy, sensitivity, specificity, receiver operating characteristic (ROC) curve, and area under ROC curve (AUC) were calculated on the test dataset. The diagnostic accuracy was 0.84, sensitivity was 0.74, and specificity was 0.94. The trained CNN algorithm achieved AUC of 0.86. The diagnostic CNN model for pediatric intraoral radiographs showed good performance with high accuracy. Deep learning can assist dentists in diagnosis of proximal caries lesions in pediatric intraoral radiographs.

Retrospective study on three-dimensional characteristics of mesiodens using CBCT in pediatric dentistry / 대한소아치과학회지

The purpose of this study was to investigate the three-dimensional characteristics of mesiodens using Cone-beam Computed Tomography(CBCT) and analyze the factors affecting complications and anesthetic methods of extraction. This study evaluated 602 mesiodens of 452 patients who underwent extraction of mesiodens at the department of Pediatric Dentistry in Seoul National University Dental Hospital between 2017 and 2019.The ratio of mesiodens patients over total patient per year was gradually increased over the past 20 years. Mesiodens with labio-palatally horizontal direction while root directing labial were the most common among the mesiodens with horizontal direction. Mesiodens were the most common at the cervical side of the adjacent teeth(37.0%) and mesiodens located in the near-palatal side were observed about 3.83 times higher than the far-palatal side. Most of the mesiodens(82.1%) were in contact with adjacent permanent teeth on all three sides of the CBCT and 46.2% of mesiodens had curved roots. The patient’s age, vertical position, presence of complications, and proximity showed a significant difference in the selection of general anesthesia among anesthetic methods. The direction and vertical position of mesiodens had a significant effect on complications.These results provide a better understanding of mesiodens for establishing an accurate diagnosis and treatment plan.

Dental Characteristics of Microcephalic Osteodysplastic Primordial Dwarfism Type II / 대한소아치과학회지

Microcephalic osteodysplastic primordial dwarfism type II (MOPD II) is an autosomal recessive inherited disorder form of primordial dwarfism, caused by mutations in the pericentrin gene. The purpose of the study was to examine the clinical and radiological features, physicochemical properties and microstructures of the tooth affected with MOPD II.The mandibular 2nd molar was collected from the MOPD II patient. Micro-computerized tomography, scanning electron microscopy, energy dispersive spectrometry and Vickers microhardness analysis were performed on the MOPD II and the normal control.The morphology of the MOPD II tooth appeared to have malformed pulp and root and showed a small size. The mineral density measurement showed that the MOPD II tooth had similar scores in the enamel, but lower scores in the root 1/2 and apical dentin compared to the normal control. The microhardness values were smaller in the cusp enamel, root 1/2 dentin and apical dentin of the MOPD II compared to the normal control.In this study, the dental characteristics and the physicochemical properties of a tooth affected with MOPD II were analyzed to improve understanding of the oral manifestations of the disease and to assist in proper dental treatment by identifying precautions.

Distribution of Pediatric Malocclusion Patients in Seoul National University Dental Hospital / 대한소아치과학회지

A total of 580 patients, who visited and received an orthodontic diagnosis in the Department of Pediatric Dentistry, Seoul National University Dental Hospital from 2017 to 2019, were investigated in this study. The aim of this study was to evaluate skeletal patterns of pediatric orthodontic patients determined with lateral cephalometric analysis and to analyze the relationship between skeletal pattern and probable associated clinical features. Also, the modality of orthodontic treatment for each skeletal classification was investigated to aid in therapeutic decisions.Patients aged 7 year accounted for the largest age group; 54.2% of patients showed a skeletal class I pattern, 22.2% showed a skeletal class II pattern, and 23.6% showed a skeletal class III pattern. Bi-maxillary retrusion for skeletal class I, retruded mandible with normal positioning of the maxilla for skeletal class II, and retrusion of the maxilla with protrusion of the mandible for skeletal class III were the largest subgroups by skeletal pattern. Brachyfacial type accounted for 55.0% of patients, followed by 31.9% of mesofacial type and 13.1% of dolichofacial type. The prevalence of anterior crossbite in the study was 43.3%, higher than that in previous studies.

Clinical Factors Affecting the Results of Regenerative Endodontic Procedure / 대한소아치과학회지

This study retrospectively analyzed the effect of clinical factors on the outcomes of REP(regenerative endodontic procedure). Patients who received the REP using triple antibiotic paste due to trauma or fracture of dens evaginatus from February, 2011 to January, 2020 were included in the study. Finally, 57 teeth in 54 patients were selected.Investigated clinical factors were as follows: intentional bleeding, etiology, and root development stage. Treatment outcomes evaluated were as follows: improvement of subjective symptoms, changes in the periapical lesion, and the amounts of root development after REP. To compensate for differences in angulation and position between repeated radiographic examinations, images were aligned by Turboreg plugin. To evaluate the amounts of root development, apical diameter, root area, and root length were measured by ImageJ software.Among the aforementioned factors, intentional bleeding had no significant effect on treatment results. Regarding the etiology, the increase in the root area and the root length was significantly less in trauma cases than in dens evaginatus fracture cases. Considering root development stage, more immature teeth presented more increase in the root area.

Root and Canal Morphology of Maxillary Primary Molar using CBCT and 3D CT / 대한소아치과학회지

The purpose of this study is to analyze morphological characteristics of maxillary primary molar’s root and root canal. 268 children aged 3 - 7 years (175 boys, 93 girls) who had CBCT (152 children) and 3D CT (116 children) taken in Seoul National University Dental Hospital from January 2006 to April 2020 were included. The number of roots and root canals were analyzed in 1002 teeth without any root resorption or periapical pathologies. Curvature, angulation, length of root and root canal, as well as cross-sectional shapes of the root canal were analyzed in 218 teeth. By using Mimics and 3-Matics software, volume, surface area, and volume ratio of root canal was analyzed in 48 teeth.More than half of maxillary primary molars have 3 roots and 3 root canals. The degree of symmetry of root canal type was about 0.63 (Cohen’s kappa coefficient). The most frequent shape of roots and canals was linear in 1st primary molars and curved in 2nd primary molars. Angulation, length of root and root canals was the largest on palatal roots. Most teeth showed ovoid or round shapes at apex. The largest root canal volume, surface area, volume ratio was found in the palatal roots.

A 20-year Retrospective Study of Pediatric Oral Lesion Biopsy / 대한소아치과학회지

The aim of this study was to provide pathological information of pediatric oral lesions by retrospectively analyzing oral biopsy results from pediatric patients at the Seoul National University Dental Hospital.Biopsy results of all oral lesions from pediatric patients, aged 0 - 16 years, were collected from the files of the Department of Oral Pathology, Seoul National University Dental Hospital from January 2000 to April 2020. The patients were divided into 3 age groups: 0 - 5, 6 - 11 and 12 - 16 years. All oral lesions were classified into three main categories: inflammatory and reactive, tumor or tumor-like and cystic lesions.Among the total of 2928 biopsy specimens, tumor or tumor-like lesions(35.66%) were the most common, followed by inflammatory and reactive lesions(34.29%) and cystic lesions(30.05%). Regardless of the categories used in this study, odontoma was the most frequently found lesion, mucocele and dentigerous cyst being the next common. This study was the first retrospective review of pediatric oral pathology in Korea, and the results from this study may assist in providing informative insight into the pediatric oral pathology for pediatric dentists.

Detecting of Proximal Caries in Primary Molars using Pen-type QLF Device / 대한소아치과학회지

The purpose of this in vivo study was to assess the clinical screening performance of a quantitative light-induced fluorescence (QLF) device in detecting proximal caries in primary molars. Fluorescence loss, red autofluorescence and a simplified QLF score for proximal caries (QS-proximal) were evaluated for their validity in detecting proximal caries in primary molars compared to bitewing radiography.Three hundred and forty-four primary molar surfaces were included in the study. Carious lesions were scored according to lesion severity assessed by visual-tactile and radiographic examinations. The QLF images were analyzed for two quantitative parameters, fluorescence loss and red autofluorescence, as well as for QS-proximal. For both quantitative parameters and QS-proximal, the sensitivity, specificity and area under receiver operating curve (AUROC) were calculated as a function of the radiographic scoring index at enamel and dentin caries levels.Both quantitative parameters showed fair AUROC values for detecting dentine level caries (△F = 0.794, △R = 0.750). QS-proximal showed higher AUROC values (0.757 - 0.769) than that of visual-tactile scores (0.653) in detecting dentine level caries.The QLF device showed fair screening performance in detecting proximal caries in primary molars compared to bitewing radiography.

Development of an Optimized Deep Learning Model for Medical Imaging

Deep learning has recently become one of the most actively researched technologies in the field of medical imaging. The availability of sufficient data and the latest advances in algorithms are important factors that influence the development of deep learning models. However, several other factors should be considered in developing an optimal generalized deep learning model. All the steps, including data collection, labeling, and pre-processing and model training, validation, and complexity can affect the performance of deep learning models. Therefore, appropriate optimization methods should be considered for each step during the development of a deep learning model. In this review, we discuss the important factors to be considered for the optimal development of deep learning models.

Development of an Optimized Deep Learning Model for Medical Imaging

Deep learning has recently become one of the most actively researched technologies in the field of medical imaging. The availability of sufficient data and the latest advances in algorithms are important factors that influence the development of deep learning models. However, several other factors should be considered in developing an optimal generalized deep learning model. All the steps, including data collection, labeling, and pre-processing and model training, validation, and complexity can affect the performance of deep learning models. Therefore, appropriate optimization methods should be considered for each step during the development of a deep learning model. In this review, we discuss the important factors to be considered for the optimal development of deep learning models.

Web-Based Spine Segmentation Using Deep Learning in Computed Tomography Images / 대한의료정보학회지

OBJECTIVES@#Back pain, especially lower back pain, is experienced in 60% to 80% of adults at some points during their lives. Various studies have found that lower back pain is a very common problem among adolescents, and the highest incidence rates are for adults in their 30s. There has been a remarkable increase in using computer-aided diagnosis to assist doctors in the interpretation of medical images. Spine segmentation in computed tomography (CT) scans using algorithmic methods allows improved diagnosis of back pain.@*METHODS@#In this study, we developed a web-based automatic spine segmentation method using deep learning and obtained the dice coefficient by comparison with the predicted image. Our method is based on convolutional neural networks for segmentation. More specifically, we train a hierarchical data format file using U-Net architecture and then insert the test data label to perform segmentation. Thus, we obtained more specific and detailed results. A total of 344 CT images were used in the experiment. Of these, 330 were used for learning, and the remaining 14 for testing.@*RESULTS@#Our method achieved an average dice coefficient of 90.4%, a precision of 96.81%, and an F1-score of 91.64%.@*CONCLUSIONS@#The proposed web-based deep learning approach can be very practical and accurate for spine segmentation as a diagnostic method.

Competencies of Dental Hygienists for Oral Care Service for People with Disability / 치위생과학회지

Background@#Dental treatment has shifted to the center of the community, and the public policy of the country has expanded to support the vulnerable classes such as the disabled. The dental profession needs education regarding oral health services for persons with disabilities, and it is necessary to derive the competencies for this. Therefore, we conducted this study to derive the normative ability to understand the role of a dental hygienist in the oral health service for persons with disabilities and improvement plans for education. @*Methods@#We conducted a qualitative analysis for deriving competencies by analyzing the data collected through in-depth interviews with experts in order to obtain abilities through practical experience. Based on the competency criterion, relevant competency in the interview response was derived using the priori method, and it was confirmed whether the derived ability matched the ability determined by the respondent. @*Results@#The professional conduct competencies of dental hygienists, devised by the Korean Association of Dental Hygiene, consists of professional behavior, ethical decision-making, self-assessment skills, lifelong learning, and accumulated evidence. Also, core competencies of the American Dental Education Association competencies for dental hygienist classification such as ethics, responsibility for professional actions, and critical thinking skills were used as the criterion. The dental hygienist's abilities needed for oral health care for people with disabilities, especially in the detailed abilities to fulfill these social needs, were clarified. @*Conclusion@#To activate oral health care for people with disabilities, it is necessary for dental hygienists to fulfill their appropriate roles, and for this purpose, competency-based curriculum restructuring is indispensable. A social safety net for improving the oral health of people with disabilities can be secured by improving the required skills-based education system of dental hygienists and strengthening the related infrastructure.