Selected article for: "data extraction and sample size"
Author: Wewetzer, Larisa; Held, Linda A.; Steinhäuser, Jost
Title: Diagnostic performance of deep-learning-based screening methods for diabetic retinopathy in primary care—A meta-analysis
  • Cord-id: bbg4069x
  • Document date: 2021_8_10
    • ID: bbg4069x
    Snippet: BACKGROUND: Diabetic retinopathy (DR) affects 10–24% of patients with diabetes mellitus type 1 or 2 in the primary care (PC) sector. As early detection is crucial for treatment, deep learning screening methods in PC setting could potentially aid in an accurate and timely diagnosis. PURPOSE: The purpose of this meta-analysis was to determine the current state of knowledge regarding deep learning (DL) screening methods for DR in PC. DATA SOURCES: A systematic literature search was conducted usin
    Document: BACKGROUND: Diabetic retinopathy (DR) affects 10–24% of patients with diabetes mellitus type 1 or 2 in the primary care (PC) sector. As early detection is crucial for treatment, deep learning screening methods in PC setting could potentially aid in an accurate and timely diagnosis. PURPOSE: The purpose of this meta-analysis was to determine the current state of knowledge regarding deep learning (DL) screening methods for DR in PC. DATA SOURCES: A systematic literature search was conducted using Medline, Web of Science, and Scopus to identify suitable studies. STUDY SELECTION: Suitable studies were selected by two researchers independently. Studies assessing DL methods and the suitability of these screening systems (diagnostic parameters such as sensitivity and specificity, information on datasets and setting) in PC were selected. Excluded were studies focusing on lesions, applying conventional diagnostic imaging tools, conducted in secondary or tertiary care, and all publication types other than original research studies on human subjects. DATA EXTRACTION: The following data was extracted from included studies: authors, title, year of publication, objectives, participants, setting, type of intervention/method, reference standard, grading scale, outcome measures, dataset, risk of bias, and performance measures. DATA SYNTHESIS AND CONCLUSION: The summed sensitivity of all included studies was 87% and specificity was 90%. Given a prevalence of DR of 10% in patients with DM Type 2 in PC, the negative predictive value is 98% while the positive predictive value is 49%. LIMITATIONS: Selected studies showed a high variation in sample size and quality and quantity of available data.

    Search related documents:
    Co phrase search for related documents
    • absence presence and accuracy studies: 1, 2, 3, 4
    • absence presence and accurate diagnosis: 1, 2, 3, 4, 5, 6, 7
    • absence presence and accurate evaluation: 1, 2
    • absence presence and additional study: 1, 2, 3, 4
    • absence presence and low prevalence: 1, 2, 3, 4
    • absence presence and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13
    • abstract title and accuracy studies: 1, 2, 3, 4, 5
    • abstract title and additional study: 1, 2
    • abstract title and low prevalence: 1, 2
    • abstract title and machine learning: 1, 2, 3, 4, 5, 6, 7
    • abstract title and machine learning artificial intelligence: 1
    • accuracy studies and additional study: 1, 2, 3
    • accuracy studies and low prevalence: 1, 2, 3
    • accuracy studies and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
    • accuracy studies and machine learning artificial intelligence: 1
    • accuracy studies and macular edema: 1
    • accurate diagnosis and additional study: 1
    • accurate diagnosis and low prevalence: 1, 2
    • accurate diagnosis and machine learning: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43