scholarly journals Using Artificial Intelligence Reading Label System in Diabetic Retinopathy Grading Training of Junior Ophthalmology Residents and Medical Students

2021 ◽  
Author(s):  
Ruoan Han ◽  
Weihong Yu ◽  
Huan Chen ◽  
Youxin Chen
Keyword(s):  
Artificial Intelligence ◽  
Diabetic Retinopathy ◽  
Medical Students ◽  
Kappa Score ◽  
Resident Students ◽  
Label System ◽  
Color Fundus Images ◽  
Group 2 ◽  
Sensitivity Specificity ◽  
Group 1

Abstract Purpose Evaluate the efficiency of using artificial intelligence reading label system in diabetic retinopathy grading training of junior ophthalmology resident doctors and medical students. Methods Loading 520 diabetic retinopathy patients’ color fundus images in the artificial intelligence reading label system. 13 participants (including 6 junior ophthalmology residents and 7 medical students) read the images randomly for 8 rounds. They evaluated the grading of images and labeled the typical lesions. The sensitivity, specificity and kappa score were determined by comparison with the participants’ results and expert golden standards. Results Through 8 round reading, average kappa score was elevated from 0.67 to 0.81. Average kappa score of round 1 to 4 was 0.77, and average kappa score of round 5 to 8 was 0.81. The participant was divided into two groups. Participants in group 1 were junior ophthalmology resident students and participants in group 2 were medical doctors. Average kappa score of group 1 was elevated from 0.71 to 0.76. Average kappa score of group 2 was elevated from 0.63 to 0.84. Conclusion The artificial intelligence reading label system was a useful tool in training resident doctors and medical students in doing diabetic retinopathy grading.

scholarly journals Self-based Learning Compared to Didactic Learning for Basic Medical or Surgical Technical Skills

2017 ◽  
Vol 5 (4) ◽  
pp. 143-159
Author(s):  
Joseph Nassif ◽  
Joe Eid ◽  
Anwar Nassar ◽  
Imad Bou Akl ◽  
Antoine Abu Musa ◽  
...  
Keyword(s):  
Medical Students ◽  
Student Satisfaction ◽  
Learning Task ◽  
Controlled Study ◽  
Line Insertion ◽  
Surgical Technical ◽  
Significant Difference ◽  
Group 2 ◽  
Didactic Lecture ◽  
Group 1

This randomized controlled study compared self-based learning (SBL) to didactic learning (DL) in teaching medical students medical and surgical skills. Self-based learning is at least as good as didactic learning in teaching medical students. The skills used were IV line insertion and simple interrupted suture. Sixty-four consenting second year medical students were randomly divided into two groups of 32 students each. For the IV line insertion activity, Group 1 was given a short didactic lecture on IV line insertion and Group 2 received a self-based learning task. Both groups were assessed a week later for IV line insertion on a dummy. Then the two groups were crossed over, where Group 2 attended a short didactic lecture and Group 1 underwent a self-based learning task on simple interrupted suturing. Both groups were assessed a week later using a suturing pad model. Statistical analysis of data, collected from assessment forms and questionnaires, was done using χ2 test (chi-square test). The results showed that there was no significant difference between the two groups in terms of their performance assessment, for both skills. However, student satisfaction significantly differed between the two groups with the SBL group expressing higher overall satisfaction in both activities. Self-based learning should be integrated in medical curricula since its comparable to didactic learning in terms of students’ performance and leads to higher student satisfaction.

scholarly journals Lesions of paired eyes at the initial stages of non-proliferative diabetic retinopathy in patients with type 2 diabetes mellitus

2021 ◽  
Vol 9 (1) ◽  
pp. 21-27
Author(s):  
S.O. Rykov ◽  
K.V. Korobov ◽  
S.Yu. Mogilevskyy
Keyword(s):  
Diabetes Mellitus ◽  
Type 2 Diabetes ◽  
Type 2 Diabetes Mellitus ◽  
Diabetic Retinopathy ◽  
Vascular Anomalies ◽  
Vascular Changes ◽  
Group 3 ◽  
Group 2 ◽  
Group 1

Background. One of the early microvascular complications of type 2 diabetes mellitus (T2DM) is diabetic retino­pathy (DR). Its main cause is prolonged hyperglycemia, which triggers the development of microangiopathy. In this regard, the issue of damage to paired eyes and the spread of DR in the initial stages has not been fully clarified. The purpose: to study the peculiarities of lesions of paired eyes at the initial stages of non-proliferative diabetic retinopathy in patients with type 2 dia­betes mellitus. Materials and methods. We examined 91 patients with T2DM (182 eyes), who did not have retinopathy according to the International Diabetic Retinopathy Severity Scale of the American Academy of Ophthalmology (2002). Paired eyes were divided into three groups: group 1 included 132 paired eyes (66 patients) with 10 points according to the Early Treatment Diabetic Retinopathy Group Study (ETDRS); group 2 consisted of 25 eyes with 10 points on ETDRS, and group 3 — 25 paired eyes with retinal vascular anomalies (14–15 points on ETDRS). The patients were examined again after 1 year. According to the ETDRS, Airlie House classification, microaneurysms, microhemorrhages, intraretinal microvascular abnormalities, retinal vascular abnormalities, and retinal nonperfusion were detec­ted. Results. The majority (58.3 %) of paired eyes without initial changes (group 1) had no progression of DR within 1 year, 12.9 % had vascular anomalies (14–15 points on ETDRS), 13.6 % deve­loped mild, and 15.2 % — moderate non-proliferative DR. The highest progression of DR (88.0 % of eyes) was observed in eyes without diabetic vascular changes, which were paired to eyes with such changes (group 2) that was 2.1 times (p < 0.001) higher than the indicator of paired eyes without diabetic changes (group 1; 41.7 %). Most eyes that had mild vascular changes (group 3) progressed to moderate non-proliferative DR after 1 year, which was four times more often than in eyes that had no initial changes (60.0 versus 15.2 %; p < 0.001). DR in the eyes of group 3 with progression accounted for 43–47 points on EDTRS; the visual acuity of these eyes, both before and after 1 year, was significantly lower than in other groups, and the level of glycated hemoglobin in the blood of patients with such eyes was significantly higher. Conclusions. This study established the features of the progression of early stages of DR in patients with T2DM, and the significance of primary retinal vascular anomalies in the presence of which the progression of DR was faster.

scholarly journals Diabetic Retinopathy Detection using Retinal Images and Deep Learning Model

2021 ◽  
Vol 10 (9) ◽  
pp. 35-39
Author(s):  
Vani Ashok ◽  
◽  
Navneet Hosmane ◽  
Ganesh Mahagaonkar ◽  
Aditya Gudigar ◽  
...  
Keyword(s):  
Diabetic Retinopathy ◽  
Deep Learning ◽  
Retinal Images ◽  
Kappa Score ◽  
Early Stages ◽  
Proposed Model ◽  
Learning Capabilities ◽  
Working Age ◽  
Color Fundus Images ◽  
Deep Learning Model

Diabetic Retinopathy (DR) is one of the serious problems caused by diabetes and a leading source of blindness in the working-age population of the advanced world. Detecting DR in the early stages is crucial since the disease generally shows few symptoms until it is too late to provide an effective cure. But detecting DR requires a skilled clinician to examine and assess digital color fundus images of the retina. By simplifying the detection process, severe damages to the eyes can be prevented. Many deep learning models particularly Convolutional Neural Networks (CNNs) have been tested in similar fields as well as in the detection of DR in early stages. In this paper, we propose an automatic model for detecting and suggesting different stages of DR. The work has been carried out on APTOS 2019 Blindness Detection Benchmark Dataset which contains around 3600 retinal images graded by clinicians for the severity of diabetic retinopathy on a range of 0 to 4. The proposed method uses ResNet50 (Residual Network that is 50 layers deep) CNN model along with pre-trained weights as the base neural network model. Due to its depth and better transfer learning capabilities, the proposed model with ResNet50 achieved 82% classification accuracy. The classification ability of the model was further analysed with Cohen Kappa score. The optimized validation Cohen Kappa score of 0.827 indicate that the proposed model didn’t predict the outputs by chance.

scholarly journals The Process of Clinical Reasoning among Medical Students

2018 ◽  
Vol 42 (3) ◽  
pp. 194-200
Author(s):  
Djon Machado Lopes ◽  
Gustavo Henrique Bregagnollo ◽  
Bruna Morais Barbosa ◽  
Ana Maria Nunes de Faria Stamm
Keyword(s):  
Internal Medicine ◽  
Medical Students ◽  
Clinical Reasoning ◽  
Semantic Network ◽  
Simple Random Sample ◽  
Reasoning Process ◽  
Analysis Process ◽  
Group 2 ◽  
Shed Light ◽  
Group 1

ABSTRACT Introduction Research in the field of medical reasoning has shed light on the reasoning process used by medical students. The strategies in this process are related to the analytical [hypothetical-deductive (HD)] and nonanalytic [scheme-inductive) (SI)] systems, and pattern recognition (PR)]. Objective To explore the clinical reasoning process of students from the fifth year of medical school at the end of the clinical cycle of medical internship, and to identify the strategies used in preparing diagnostic hypotheses, knowledge organization and content. Method Qualitative research conducted in 2014 at a Brazilian public university with medical interns. Following Stamm’s method, a case in internal medicine (IM) was built based on the theory of prototypes (Group 1 = 47 interns), in which the interns listed, according to their own perceptions, the signs, symptoms, syndromes, and diseases typical of internal medicine. This case was used for evaluating the clinical reasoning process of Group 2 (30 students = simple random sample) obtained with the “think aloud” process. The verbalizations were transcribed and evaluated by Bardin’s thematic analysis. The content analysis were approved by two experts at the beginning and at the end of the analysis process. Results The interns developed 164 primary and secondary hypotheses when solving the case. The SI strategy prevailed with 48.8%, followed by PR (35.4%), HD (12.2%), and mixed (1.8 % each: SI + HD and HD + PR). The students built 146 distinct semantic axes, resulting in an average of 4.8/participant. During the analysis, 438 interpretation processes were executed (average of 14.6/participant), and 124 combination processes (average of 4.1/participant). Conclusions The nonanalytic strategies prevailed with the PR being the most used in the development of primary hypotheses (46.8%) and the SI in secondary hypotheses (93%). The interns showed a strong semantic network and did three and a half times more interpretation than combination processes, reflecting less deep organization and content of knowledge when compared with experienced physicians.

scholarly journals Influences of 27-Gauge Vitrectomy on Corneal Topographic Conditions

2017 ◽  
Vol 2017 ◽  
pp. 1-6 ◽  
Author(s):  
Takafumi Hirashima ◽  
Takao Utsumi ◽  
Miou Hirose ◽  
Hideyasu Oh
Keyword(s):  
Diabetic Retinopathy ◽  
Epiretinal Membrane ◽  
Spherical Aberration ◽  
The Other ◽  
Significant Difference ◽  
The Difference ◽  
Group 2 ◽  
Higher Order Aberration ◽  
Topographic Parameters ◽  
Group 1

Purpose. To evaluate the influences of 27-gauge vitrectomy on corneal topographic conditions. Method. Fifty-six eyes of 56 patients undergoing 27-gauge vitrectomy were retrospectively studied. Twenty-three eyes with epiretinal membrane (ERM), 23 eyes with macular hole (MH), and 10 eyes with proliferative diabetic retinopathy (PDR) were included. Forty-five of the 56 eyes underwent 27-gauge phacovitrectomy (group 1), and the remaining 11 eyes underwent 27-gauge vitrectomy alone (group 2). Corneal topography was obtained with a wave-front analyzer preoperatively and at 1 and 3 months postoperatively. The corneal topographic parameters evaluated were the average corneal power, regular astigmatism, spherical aberration, and higher-order aberration (HOA). Results. In between-group analyses of groups 1 and 2, no significant differences were observed regarding the changes of the 4 parameters from the baseline to 1 and 3 months postoperatively. No significant differences in the changes of all parameters from the baseline to 1 and 3 months postoperatively were also observed between MH group and the other two groups. A significant difference in the change of HOA from the baseline to 1 month postoperatively was observed between ERM and PDR group however, the difference disappeared at 3 months. Conclusion. 27-gauge vitrectomy did not induce substantial changes in the corneal topographic conditions.

How “Proficient” Is Proficient? Comparison of English and Relative Proficiency Rating as a Predictor of Bilingual Listeners' Word Recognition

2013 ◽  
Vol 22 (1) ◽  
pp. 40-52 ◽  
Author(s):  
Lu-Feng Shi
Keyword(s):  
Word Recognition ◽  
Predictive Power ◽  
English Proficiency ◽  
First Language ◽  
English Word ◽  
Northwestern University ◽  
Test Population ◽  
Group 2 ◽  
Sensitivity Specificity ◽  
Group 1

Purpose The current study attempted to validate that English proficiency self-ratings predict bilinguals' recognition of English words as reported in Shi (2011) and to explore whether relative proficiency ratings (English vs. first language) improve prediction. Method One hundred and twenty-four participants in Shi (2011) and an additional set of 145 participants were included (Groups 1 and 2, respectively) in this study. All listeners rated their proficiency in listening, speaking, and reading (English and first language) on an 11-point scale and listened to a list of words from the Northwestern University Auditory Tests No. 6 (Tillman & Carhart, 1966) at 45 dB HL in quiet. Results English proficiency ratings by Group 2 yielded sensitivity/specificity values comparable to those of Group 1 (Shi, 2011) in predicting word recognition. A cutoff of 8 or 9 in minimum English proficiency rating across listening, speaking, and reading resulted in the best combination of prediction sensitivity/specificity. When relative proficiency was used, prediction of Group 1 performance significantly improved as compared to English proficiency. Improvement was slight for Group 2, mainly due to low specificity. Conclusion Self-rated English proficiency provides clinically acceptable sensitivity/specificity values as a predictor of bilinguals' English word recognition. Relative proficiency has the potential to further improve predictive power, but the size of improvement depends on the characteristics of the test population.

scholarly journals Will mild atypical pneumonia remain underdiagnosed and undertreated in primary care setting?

2020 ◽  
Vol 30 (Supplement_5) ◽  
Author(s):  
S Gatsura ◽  
V Deriushkin ◽  
O Gatsura ◽  
E Ulyanova
Keyword(s):  
Primary Care ◽  
Medical Students ◽  
Primary Care Physicians ◽  
Primary Care Providers ◽  
Atypical Pneumonia ◽  
Care Providers ◽  
Beta Lactams ◽  
Atypical Pathogens ◽  
Group 2 ◽  
Group 1

Abstract Background Non severe community acquired pneumonia (CAP) is a common problem in primary care. So called “walking” CAP is frequently caused by atypical intracellular pathogens Chlamydia pneumoniae and Mycoplasma pneumonia which are resistant to beta-lactams and can be transmitted from an infected person to a healthy one. Taking into account medical and epidemiologic importance of this problem we aimed to estimate appropriateness of the antimicrobial agent (AM) choice for outpatient treatment of mild CAP by current and future primary care providers with regard to atypical pathogens coverage. Methods Total 240 final year medical students of A.I.Yevdokimov Moscow State University of Medicine and Dentistry (Group 1) and 206 Moscow primary care physicians (Group 2) were surveyed in 2019. Respondents were asked to specify in writing what particular AM they would recommend to 35 year old previously healthy male patient with subfebrile body temperature (37.3 °C), non-productive cough and documented CAP. Chi-square test was used to compare the data obtained in both groups. Results Group 1 respondents returned questionnaires with 271 recommendations, Group 2 participants named 230 items. AMs with atypical pathogens coverage (macrolides, fluoroquinolones and doxycycline) accounted for just 33.2% in Group 1 versus 20.0% in Group 2 (p=.0009). Amoxicillin/clavulanate was the leading choice equally popular both in students and physicians (42.1% and 40.9% respectively). The rest of recommendations in both Groups included amoxicillin and various cephalosporins. Conclusions Only one of three students and one of five physicians made the right choice in offered clinical scenario. A majority of respondents in both groups hastily recommended beta-lactams instead of clinical estimation of atypical CAP probability in given situation, but students indicated appropriate AMs more often. This problem obviously persists and requires action from both academics and healthcare managers. Key messages Medical students and primary care physicians’ awareness of atypical CAP presentation and treatment is not quite satisfactory. Resulting undertreatment of atypical CAP may harm the patient and promote further spread of causative pathogen within the community.

A Comparison of Artificial Intelligence and Human Diabetic Retinal Image Interpretation in an Urban Health System

2021 ◽  
pp. 193229682199937
Author(s):  
Nikita Mokhashi ◽  
Julia Grachevskaya ◽  
Lorrie Cheng ◽  
Daohai Yu ◽  
Xiaoning Lu ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Diabetic Retinopathy ◽  
Predictive Value ◽  
Image Interpretation ◽  
High Sensitivity ◽  
Diabetic Patients ◽  
Screening Algorithm ◽  
Significant Difference ◽  
Internal Medicine Clinic ◽  
Sensitivity Specificity

Introduction: Artificial intelligence (AI) diabetic retinopathy (DR) software has the potential to decrease time spent by clinicians on image interpretation and expand the scope of DR screening. We performed a retrospective review to compare Eyenuk’s EyeArt software (Woodland Hills, CA) to Temple Ophthalmology optometry grading using the International Classification of Diabetic Retinopathy scale. Methods: Two hundred and sixty consecutive diabetic patients from the Temple Faculty Practice Internal Medicine clinic underwent 2-field retinal imaging. Classifications of the images by the software and optometrist were analyzed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and McNemar’s test. Ungradable images were analyzed to identify relationships with HbA1c, age, and ethnicity. Disagreements and a sample of 20% of agreements were adjudicated by a retina specialist. Results: On patient level comparison, sensitivity for the software was 100%, while specificity was 77.78%. PPV was 19.15%, and NPV was 100%. The 38 disagreements between software and optometrist occurred when the optometrist classified a patient’s images as non-referable while the software classified them as referable. Of these disagreements, a retina specialist agreed with the optometrist 57.9% the time (22/38). Of the agreements, the retina specialist agreed with both the program and the optometrist 96.7% of the time (28/29). There was a significant difference in numbers of ungradable photos in older patients (≥60) vs younger patients (<60) (p=0.003). Conclusions: The AI program showed high sensitivity with acceptable specificity for a screening algorithm. The high NPV indicates that the software is unlikely to miss DR but may refer patients unnecessarily.

scholarly journals Recent Innovations in Automated Detection and Classification of Diabetic Retinopathy

2019 ◽  
Vol 8 (10) ◽  
pp. 1997-2004
Keyword(s):  
Artificial Intelligence ◽  
Diabetic Retinopathy ◽  
Classification Accuracy ◽  
Medical Diagnosis ◽  
Automated Detection ◽  
Huge Number ◽  
Smart System ◽  
Sensitivity Specificity ◽  
Modern Technologies

Due to the increasing prevalence of diabetic retinopathy worldwide, it’s an urgent need to develop smart system that help to detect disease using one of the modern technologies. Artificial intelligence is one of the popular techniques nowadays which has the ability to learn from experience and carry out human-like tasks. Large number of researches have been conducted to find out effective medical diagnosis methods for numerous diseases. Likewise, huge number of researches have been done that discuss automated detection and classification of diabetic retinopathy. This paper reviews the existing methodologies, datasets, sensitivity, specificity and classification accuracy in diabetic retinopathy.

scholarly journals Predictive Factors of Graduation Delay in a Medical Program: a Retrospective Cohort Study in Brazil, 2010-2016

2020 ◽  
Vol 44 (1) ◽  
Author(s):  
Bárbara Aparecida da Silva Rego Rocha ◽  
Antonio Toledo Júnior
Keyword(s):  
Academic Success ◽  
Medical Students ◽  
Predictive Factors ◽  
Academic Failure ◽  
Admission Process ◽  
First Semester ◽  
Medical Program ◽  
Freshman Students ◽  
Group 2 ◽  
Group 1

Abstract: Introduction: Medical training is a long and expensive process. Admission processes are highly competitive all over the world but being accepted is no guarantee of academic success. Medical school is demanding and stressful, and some students are not able to cope with this new scenario successfully. It is estimated that 10-15% of medical students experience difficulties in adapting to the course, which can lead to academic failure. The identification of predictive factors of failure supports the creation of mechanisms and strategies to avoid course dropout or graduation delay. To identify predictive factors of academic failure in a Brazilian medical program. Methods: A retrospective observational study was carried out with all medical students admitted to a private Brazilian medical school in 2010 and 2011. The main outcome was academic success. Academic failure was defined as graduation delay or course dropout (Group 1), and academic success was defined as graduating within 6 years (Group 2). Sociodemographic and academic data were collected, including grades obtained at the admission process and the first-semester courses. Freshman students and students with passing grades in the first semester (passed students) were analyzed separately. Descriptive and comparative analyses, logistic regression and ROC curve analysis were performed. The level of significance was 0.05. Results: A total of 312 students were admitted during the study period, but 10 were excluded due to lack of information. Of the 302 students included in the study, 105 were included in Group 1 and 197 were included in Group 2. Thirty-two students failed the first semester. The 270 students with passing grades in the first semester were divided into Group 1 (n=73) and Group 2 (n=197). Among the freshman students, lower admission grades were associated with a higher chance of failure (padjusted=0.012). Of the 270 students with passing grades, low academic performance (courses’ mean grades) was associated with graduation delay (padjusted<0.001). Conclusions: Low grades at the admission process (freshman students) and low academic performance in the first semester (students with passing grades) were predictive factors of academic failure.

Sign in / Sign up

Export Citation Format

Share Document