scholarly journals Screening for diabetic retinopathy using an automated diagnostic system based on deep learning: Diagnostic accuracy assessment

2020 ◽  
Author(s):  
Sílvia Rêgo ◽  
Marco Dutra Medeiros ◽  
Filipe Soares ◽  
Matilde Monteiro-Soares
Keyword(s):  
Diabetic Retinopathy ◽  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Positive Likelihood Ratio ◽  
Diagnostic System ◽  
Cross Sectional ◽  
Negative Test ◽  
Negative Test Result ◽  
Test Result ◽  
Test Probability

Purpose: To evaluate the diagnostic accuracy of a diagnostic system software for the automated screening of diabetic retinopathy (DR) on digital colour fundus photographs, the 2019 Convolutional Neural Network (CNN) model with Inception-V3. Methods: In this cross-sectional study 295 fundus images were analysed by the CNN model and compared to a panel of ophthalmologists. Images were obtained from a dataset acquired within a screening programme. Diagnostic accuracy measures and respective 95% confidence intervals (CI) were calculated. Results: The sensitivity and specificity of the CNN model in diagnosing referable DR was 81% [95% confidence interval (CI), 66%−90%] and 97% (95% CI, 95%−99%), respectively. Positive predictive value was 86% (95% CI, 72%−94%) and negative predictive value 96% (95% CI, 93%−98%). The positive likelihood ratio was 33 (95% CI, 15−75) and the negative was 0.20 (95% CI, 0.11−0.35). Its clinical impact is demonstrated by the change observed in the pre-test probability of referable DR (assuming a prevalence of 16%) to a post-test probability for a positive test result of 86% and for a negative test result of 4%. Conclusion: A CNN model negative test result safely excludes DR and its use may significantly reduce the burden of ophthalmologists at reading centres.

scholarly journals Foundational Statistical Principles in Medical Research: Sensitivity, Specificity, Positive Predictive Value, and Negative Predictive Value

Medicina ◽  
2021 ◽  
Vol 57 (5) ◽  
pp. 503
Author(s):  
Thomas F. Monaghan ◽  
Syed N. Rahman ◽  
Christina W. Agudelo ◽  
Alan J. Wein ◽  
Jason M. Lazar ◽  
...  
Keyword(s):  
Positive Predictive Value ◽  
Negative Predictive Value ◽  
Predictive Value ◽  
Medical Literature ◽  
Disease Prevalence ◽  
Positive Test Result ◽  
Predictive Values ◽  
Sensitivity Specificity ◽  
Test Result ◽  
Test Probability

Sensitivity, which denotes the proportion of subjects correctly given a positive assignment out of all subjects who are actually positive for the outcome, indicates how well a test can classify subjects who truly have the outcome of interest. Specificity, which denotes the proportion of subjects correctly given a negative assignment out of all subjects who are actually negative for the outcome, indicates how well a test can classify subjects who truly do not have the outcome of interest. Positive predictive value reflects the proportion of subjects with a positive test result who truly have the outcome of interest. Negative predictive value reflects the proportion of subjects with a negative test result who truly do not have the outcome of interest. Sensitivity and specificity are inversely related, wherein one increases as the other decreases, but are generally considered stable for a given test, whereas positive and negative predictive values do inherently vary with pre-test probability (e.g., changes in population disease prevalence). This article will further detail the concepts of sensitivity, specificity, and predictive values using a recent real-world example from the medical literature.

Diagnostic Accuracy of HRCT Chest in Detection of Covid-19 Infection Taking PCR as Gold Standard

2021 ◽  
Vol 15 (6) ◽  
pp. 1679-1681
Author(s):  
Afaque Ali ◽  
Majid Shaikh ◽  
Ahsanullah . ◽  
Adeel Ahmed ◽  
Abid Ali Sahito ◽  
...  
Keyword(s):  
Computed Tomography ◽  
High Resolution ◽  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Gold Standard ◽  
Cross Sectional Study ◽  
Radiology Department ◽  
High Resolution Computed Tomography ◽  
Cross Sectional ◽  
Head Of Department

Objective: To determine the diagnostic accuracy of High-resolution computed tomography (HRCT) chest in detection of covid-19 infection taking PCR as gold standard. Study Design: Cross-sectional study Setting: Radiology department of Tabba Hospital, Karachi. Duration: From March 2019 to September 2020 Material and Methods: All the clinically suspected patients of covid-19, of any age, both genders and those referred to radiology for High-resolution computed tomography (HRCT) chest to detect the covid-19 infection were included. After two days, patients’ PCR reports were collected from the ward, after taking informed consent and permission from head of department. The diagnostic accuracy of HRCT was established with respect to sensitivity, PPV, NPV, and specificity by taking PCR as gold standard. All the information was collected via study proforma. Results: Total 70 patients suspected for COVID-19 were studied, and the patients’ mean age was 58.23±9.52 years. Males were in majority 54(77.1%). As per HRCT findings, COVID-19 infection was positive in 46 patients, however, 48 patients were detected positive for COVID-19 infection as per PCR findings. In the detection of COVID-19 infection, HRCT chest showed sensitivity of 91%, specificity of 90%, PPV of 83%, NPV of 84% and diagnostic accuracy of 94%; by taking PCR as gold standard. Conclusion: High-resolution computed tomography (HRCT) is a reliable diagnostic approach in promptly detecting the COVID-19; with 91% sensitivity, 90% specificity, 83% positive predictive value, 84% negative predictive value and 94% diagnostic accuracy. Keywords: Accuracy, HRCT, COVID-19

scholarly journals Comparison of Venous and Capillary Sampling in Oral Glucose Testing for the Diagnosis of Gestational Diabetes Mellitus: A Diagnostic Accuracy Cross-Sectional Study Using Accu-Chek Inform II

Diagnostics ◽  
2020 ◽  
Vol 10 (12) ◽  
pp. 1011
Author(s):  
Sofia Nevander ◽  
Eva Landberg ◽  
Marie Blomberg ◽  
Bertil Ekman ◽  
Caroline Lilliecreutz
Keyword(s):  
Diabetes Mellitus ◽  
Gestational Diabetes ◽  
Gestational Diabetes Mellitus ◽  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Cross Sectional Study ◽  
Oral Glucose ◽  
Sectional Study ◽  
Cross Sectional ◽  
Capillary Sampling

Gestational diabetes mellitus (GDM) is a common complication with negative impacts on mother and child. The primary aim of this study was to examine whether plasma glucose cutoffs for GDM diagnosis based on venous sampling can be replaced by cutoffs based on capillary sampling. A prospective cross-sectional study was performed at an antenatal care clinic including 175 pregnant women undergoing an oral glucose tolerance test (OGTT). Duplicate samples were collected by capillary and venous puncture while fasting and 1 h and 2 h after an OGTT. Both samples were analyzed on Accu-Chek Inform II. The cutoffs for a GDM diagnosis using capillary samples were corrected from 5.1 to 5.3 mmol/L for the fasting sample, from 10.0 to 11.1 mmol/L for the 1 h sample, and from 8.5 to 9.4 mmol/L for the 2-h sample using half of the dataset. Applying these cutoffs to the remaining dataset resulted in a sensitivity, specificity, and accuracy of 85.0%, 95.0%, and 90.3%, respectively, with a positive predictive value (PPV) of 83%, an negative predictive value (NPV) of 96%, and a positive negative likelihood ratio (LHR) of 16.4 using capillary sampling for the GDM diagnosis at fasting and 2-h after. Corrected cutoffs and capillary samples can be used for the diagnosis of GDM with maintained diagnostic accuracy using Accu-Chek Inform II.

scholarly journals Diagnostic Accuracy of Sonographic Septations in Tuberculous and Malignant Pleural Effusions

2020 ◽  
Vol 9 (2) ◽  
pp. 95-102
Author(s):  
Abdul Rasheed Qureshi ◽  
Muhammad Irfan ◽  
Zeeshan Ashraf
Keyword(s):  
Pleural Effusion ◽  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Tertiary Care ◽  
Positive Likelihood Ratio ◽  
Fisher Exact Test ◽  
Care Hospital ◽  
Pleural Effusions ◽  
Malignant Pleural Effusions ◽  
Malignant Group

Background: Discrimination between tuberculous (TB) and malignant pleural effusions is a real practical challenge because both exist as exudative lymphocytic type. Transthoracic ultrasonography not only identifies and quantifies pleural effusion but also displays sonographic septations, which are frequently seen in TB pleural effusions and can help in differentiation between tuberculosis and malignancy successfully, without any invasive procedure. We designed this study to determine the diagnostic usefulness of these septations for tuberculous and malignant pleural effusions. Material and Methods: This prospective study was conducted in the OPD of Gulab Devi Chest Hospital Lahore, Pakistan, a 1500 bedded tertiary care hospital, from November 2016 to February 2018. Total of 339 consecutive cases, aged 14-83 years with radiological evidence of pleural effusion were included in the study. After detailed history, thorough physical examination, radiological, haematological and biochemical findings were recorded. Pleural fluid macroscopic, cytological, microbiologic and biochemical analysis results were also recorded. Ultrasonography was done, septated and non-septated pleural effusions identified and findings were noted. SPSS-16 was used for statistical evaluation. Fisher Exact test was utilized for comparison between TB and malignant cases with P-value < 0.05 taken as significant. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), diagnostic accuracy and positive likelihood ratio were calculated. Results: Out of total 339 cases, 49 (14.45%) were malignant and 290 (85.55%) were non-malignant. In the malignant group, only 03 cases (6.12%) showed sonographic septations. In the non-malignant group, 259/290 (89.31%) cases showed tuberculous etiology and 187/259 (72.20%) of these cases displayed sonographic septations. By considering septations as predictor of TB, statistical analysis revealed a sensitivity of 79.23%, specificity of 92.85%, PPV of 98.42%, NPV of 44.31% and diagnostic accuracy of 81.29%, respectively. Conclusions: Sonographic septations can be a valuable predictor of tuberculosis, in a population with high prevalence of the disease. We found it to be a useful feature in differentiating between a malignant and tuberculous etiology, in exudative lymphocytic pleural effusions. It can be used with confidence in patients who are unfit for interventional procedures.

scholarly journals ACCURACY OF THE UPPER LIP BITE TEST;

2012 ◽  
Vol 20 (01) ◽  
pp. 132-138
Author(s):  
MUHAMMAD ATIF ◽  
MUHAMMAD ABDULLAH ◽  
MUHAMMAD JAVAD YOUSAF ◽  
Khalid Buland
Keyword(s):  
Positive Predictive Value ◽  
Diagnostic Accuracy ◽  
Difficult Intubation ◽  
Predictive Value ◽  
Elective Surgery ◽  
Cross Sectional Study ◽  
Military Hospital ◽  
Upper Lip ◽  
Cross Sectional ◽  
Sensitivity Specificity

Objective: To compare the accuracy of Upper lip bite test with modified Mallampati classification for predicting the difficultlaryngoscopic intubation. Study Design: Cross sectional Study. Place and duration of study: The study was carried out at Department ofAnaesthesiology, Intensive Care and Pain management, Combined Military hospital, Rawalpindi from September 2008 to August 2009.Patients and Methods: Four hundred patients undergoing elective surgery meeting the inclusion/exclusion criteria were enrolled afterwritten informed consent. The airways of the patients were evaluated by using the modified Mallampati classification (MMP) and theUpper lip bite test (ULBT). MMP class 3 or 4 and ULBT class 3 were considered as indicators of difficult intubation. The laryngeal view wasgraded by Cormack and Lehane classification (Gold standard). Grade 1 or 2 was considered to represent easy intubation and grade 3 or 4to represent difficult intubation. Sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy werecalculated for both the tests separately by using the 2×2 table. Results: ULBT had a higher accuracy of 94%, specificity of 99.2% andpositive predictive value 70% compared to MMP accuracy of 82.7%, specificity of 84.4% and positive predictive value of 22.7%.Conclusions: The diagnostic accuracy of the Upper lip bite test was more than the modified Mallampati classification. We suggest that itbe compared with the other prevailing tests as well which are often used to assess difficult intubations.

Diagnostic accuracy of point-of-care knee ultrasound for evaluation of meniscus and collateral ligaments pathology in comparison with MRI

2021 ◽  
pp. 028418512110582
Author(s):  
Ahmed Elshimy ◽  
Ahmed M Osman ◽  
Mohamed El Sayed Awad ◽  
Mohamed M Abdel Aziz
Keyword(s):  
Diagnostic Accuracy ◽  
Predictive Value ◽  
Point Of Care ◽  
Operative Procedure ◽  
Point Of Care Ultrasound ◽  
Collateral Ligaments ◽  
Collateral Ligament ◽  
Diagnostic Modality ◽  
Cross Sectional ◽  
Portable Ultrasonography

Background Although magnetic resonance imaging (MRI) is often the “gold standard” for diagnosing knee problems, it has many limitations. Therefore, ultrasonography has been suggested as an effective rapid alternative in many knee abnormalities, especially after injuries of the meniscus and collateral ligaments. Purpose To determine the diagnostic accuracy of point-of-care ultrasound (POCUS) in detecting injuries of the meniscus and collateral ligament compared to MRI. Material and Methods An observational cross-sectional blinded study was conducted of 60 patients with clinically suspicious meniscus and collateral ligament injuries who were planned for an arthroscopy and or operative procedure. These patients underwent both blinded POCUS and MRI of the knees before the intervention procedure and results of both imaging modalities were compared according to the operative and arthroscopic findings. Results The preoperative reliability of POCUS compared to MRI for the assessment of meniscus injuries was sensitivity (92.9% vs. 90.5%), specificity (88.9% vs. 83.3%), positive predictive value (PPV; 95.1% vs. 92.7%), negative predictive value (NPV; 84.2% vs. 79%), and overall accuracy (91.7% vs. 88.3%). However, for diagnosing collateral ligament injures, POCUS versus MRI assessed sensitivity (92.3% vs. 88.5%), specificity (100% vs. 97.1%), PPV (100% vs. 95.8%), NPV (94.4% vs. 91.7%), and overall accuracy (96.7% vs. 93.3%). Conclusion Ultrasonography is a useful screening tool for the initial diagnosis of meniscal and collateral ligament pathology compared to or even with potential advantages over MRI, especially when MRI is unavailable or contraindicated. As newly advanced portable ultrasonography becomes available, it could be considered as a point-of-injury diagnostic modality.

Validity and Diagnostic Accuracy of Clarke's Angle in Determining Paediatric Flexible Flatfoot Using Radiographic Findings as a Criterion Standard Measure: A Cross-Sectional Study

2021 ◽  
Author(s):  
Fatma Hegazy ◽  
Emad Aboelnasr ◽  
Amr A. Abdel-Aziem ◽  
In-Ju Kim
Keyword(s):  
Diagnostic Accuracy ◽  
Likelihood Ratio ◽  
Positive Likelihood Ratio ◽  
Negative Likelihood Ratio ◽  
Cross Sectional Study ◽  
Cross Sectional ◽  
Radiographic Findings ◽  
Flexible Flatfoot ◽  
Foot Posture ◽  
Posttest Probability

Background: Clinical diagnosis of Paediatric flexible flatfoot is still a challenging issue for the healthcare professionals. Clarke's Angle (CA) is frequently used clinically for assessing foot posture; however there is still a debate about its validity and diagnostic accuracy in evaluation of static foot posture especially in Paediatric population with some previous studies supporting, while the others are refuting its validity. The current study aimed to investigate validity and diagnostic accuracy of CA using radiographic findings as a criterion standard measure to determine flexible flatfoot between ages of 6-18 years. Methods: A cross-sectional study conducted on 612 participants (1224 feet) with flexible flatfoot aged 6-18 years (12.36 {plus minus}3.39) was recruited. The clinical measures results were compared to the criterion standard radiographic measures and displayed on the receiver operating characteristic curve (ROC) and the area under ROC curve (AUC) was computed. Intra-rater reliability, sensitivity, specificity, predictive values, and likelihood ratios were calculated for CA. Fagan nomogram was used to detect the posttest probability. Results: Clarke's angle (CA) demonstrated higher intra-rater reliability (ICC=0.99). CA showed higher sensitivity (98.4%), specificity (98.8), positive predictive value (97.3), negative predictive value (99.3), positive likelihood ratio (84) and negative likelihood ratio (0.02). AUC was 0.98[95%CI= 0095-1.00]. The positive likelihood ratio (LR+) yielded a posttest probability of 97%, while the negative likelihood ratio (LR−) yielded a posttest probability of 0.02. Conclusions: CA is a valid with high diagnostic accuracy in diagnosis of flexible flatfoot between the ages of 6-18 years.

scholarly journals Diagnostic accuracy of the risk of malignancy index 1 compared to the more recent IOTA ADNEX model in discriminating benign from malignant adnexal masses: a multi-centric study

2019 ◽  
Vol 8 (3) ◽  
pp. 1001
Author(s):  
Amita Ray ◽  
Divya S. ◽  
B. N. Kumar Guru ◽  
A. S. Ramaswamy ◽  
Bharat Kumar
Keyword(s):  
Diagnostic Accuracy ◽  
Likelihood Ratio ◽  
Predictive Value ◽  
Positive Likelihood Ratio ◽  
Negative Likelihood Ratio ◽  
Adnexal Masses ◽  
Medical College ◽  
Risk Of Malignancy ◽  
Risk Of Malignancy Index ◽  
Almost All

Background: Identification of the nature of an adnexal mass can ensure optimum management. Single parameters as well as diagnostic models using a combination of several parameters are in use. The International Ovarian Tumor Analysis (IOTA) consortium has developed and published the Assessment of Different NEoplasias in the adneXa (ADNEX) model, which differentiates between benign and malignant masses. Authors conducted this study with the aim of finding a cut off value for this model in the study population and comparing the diagnostic accuracy of this model to that of the risk of malignancy (RMI).Methods: Women with adnexal masses admitted to the 3 medical college affiliated hospitals for surgical management were included in this study. Appropriate investigations were done to calculate the RMI-I and ADNEX score for each participant. A cut off score for the ADNEX model was determined and diagnostic accuracy tests were done for comparison.Results: At a cut-off of 29 for the ADNEX model and 200 for RMI model the sensitivity was 75% and 77.8, specificity 100% and 80.6%; Positive Predictive Value (PPV) 100%and 60%; Negative Predictive Value (NPV) 91% and 90.6%; Positive Likelihood ratio of infinity and 4 and a negative Likelihood Ratio of 2.8 and 2.5 respectively.Conclusions: The ADNEX model rates higher than the RMI in almost all tests of diagnostic accuracy and can be used for triaging, framing a referral policy and prioritizing surgery.

scholarly journals Maternal and Perinatal Outcomes in Patients with Suspected COVID-19 and Their Relationship with a Negative RT-PCR Result

2020 ◽  
Vol 9 (11) ◽  
pp. 3552 ◽  
Author(s):  
Yolanda Cuñarro-López ◽  
Óscar Cano-Valderrama ◽  
Pilar Pintado-Recarte ◽  
Ignacio Cueto-Hernández ◽  
Blanca González-Garzón ◽  
...  
Keyword(s):  
Pregnant Women ◽  
Oxygen Therapy ◽  
Perinatal Outcomes ◽  
Tertiary Hospital ◽  
Monocyte Count ◽  
Rt Pcr ◽  
Negative Test ◽  
Negative Test Result ◽  
Test Result ◽  
Pcr Test

This study was designed to examine maternal-perinatal outcomes in pregnant women with suspected coronavirus disease 2019 (COVID-19) according to the result of a real-time reverse transcription polymerase chain reaction (RT-PCR) test and to investigate possible variables that could be useful for predicting a negative RT-PCR result. Participants of this retrospective cohort study were obstetrics patients with suspected COVID-19 who underwent an RT-PCR test in a tertiary hospital in Madrid, Spain. Maternal-perinatal features were analysed according to the results of this test. Clinical, radiological and analytical characteristics that could be associated with a negative result were also explored. In a final subgroup analysis, patients were included if they had pneumonia and a negative test result for the virus. Out of the 111 obstetric patients with suspected COVID-19 that were enrolled, 38.7% returned a negative result. In this RT-PCR-negative group, we recorded lower rates of pneumonia (21.4% vs. 45.6%, p = 0.009), severe or critical clinical features (4.7% vs. 11.8% and 0.0% vs. 5.9%, p = 0.02, respectively), lower lactate dehydrogenase (LDH) levels (168 UI/L vs. 224.5 UI/L, p = 0.003), a greater need for maternal treatment (60.3% vs 24.4%, p < 0.001), a reduced need for oxygen therapy (2.4% vs 28.8%, p < 0.001) and a lower rate of intensive care unit admission (0.0% vs. 3.7%, p = 0.046) than the RT-PCR-positive group. While no differences were found in other variables, the monocyte count was higher (946.2/μL vs. 518.8/μL, p = 0.022) in this group. The predictive model for a negative test result included the monocyte count, LDH level and no need for oxygen therapy. This model was able to identify 73.5% of patients with a negative RT-PCR result. Only 11% of the patients with pneumonia testing negative for the virus had IgG antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The proportion of pregnant women with suspected COVID-19 and a negative RT-PCR result was nearly 39%. In these patients, the symptoms were mild and the systemic severity of the disease was lower. The monocyte count, LDH level and no need for oxygen therapy were the factors that were more related to a negative test result in this group. These variables could be used to guide the management of patients with suspected COVID-19, mainly while waiting for RT-PCR results or in settings where this test is not available.

Interpretations of the dial test should be reconsidered. A diagnostic accuracy study reporting sensitivity, specificity, predictive values and likelihood ratios

2018 ◽  
Vol 3 (4) ◽  
pp. 198-204
Author(s):  
Richard Norris ◽  
Christian Kopkow ◽  
Michael James McNicholas
Keyword(s):  
Diagnostic Accuracy ◽  
Likelihood Ratio ◽  
Predictive Value ◽  
Positive Likelihood Ratio ◽  
Negative Likelihood Ratio ◽  
Knee Ligament ◽  
Predictive Values ◽  
Medial Knee ◽  
Sensitivity Specificity ◽  
Dial Test

ObjectivesTo determine the accuracy of the dial test, used alone and in combination with additional clinical tests, in the diagnosis of an isolated posterolateral corner (PLC) injury, combined PLC-posterior cruciate ligament (PCL) injury or medial knee injury.MethodsA retrospective analysis of consecutive patients who underwent arthroscopic and/or open knee ligament reconstruction surgery was conducted. The dial test was performed in an outpatient’s clinic as part of a routine knee examination. Examination under anaesthetic and intraoperative findings were used as the reference standard test to determine the diagnostic accuracy of the dial test used alone and in combination with other PCL and medial knee tests. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+) and negative likelihood ratio (LR−) were calculated with corresponding 95% CI.ResultsData from 87 patients were available and included in the data analysis. For an isolated PLC injury, the dial test sensitivity and specificity were 0.20 (95% CI 0.08 to 0.39) and 1.00 (95% CI 0.92 to 1.00). The PPV and NPV were 1.00 (95% CI 0.52 to 1.00) and 0.70 (95% CI 0.59 to 0.80). LR+ and LR− of the dial test detecting isolated PLC injury were infinity (95% CI calculation not possible, infinity) and 0.80 (95% CI 0.41 to 1.57). The diagnostic accuracy of the dial test, when used alone and in combination with other PCL and medial knee tests, was also calculated for combined PLC-PCL and medial knee injuries.ConclusionA negative dial test at 30° of knee flexion can rule out a PLC injury, while a test that is positive at 30° and negative at 90° indicates a PLC injury, without concomitant injury to the PCL or medial knee ligaments. A positive test at both 30° and 90° can indicate isolated PLC, combined PLC-PCL or medial ligament injuries, and other knee examination findings are required to differentially diagnose these injury patterns.Level of evidenceII.

Sign in / Sign up

Export Citation Format

Share Document