ABSTRACT
PRCIS: In this meta-analysis of 6 studies and 5,269 patients, deep learning algorithms applied to AS-OCT demonstrated excellent diagnostic performance for closed-angle compared to gonioscopy, with a pooled sensitivity and specificity of 94% and 93.6%, respectively. PURPOSE: This study aimed to review the literature and compare the accuracy of deep learning algorithms (DLA) applied to anterior segment optical coherence tomography images (AS-OCT) against gonioscopy in detecting angle-closure in patients with glaucoma. METHODS: We performed a systematic review and meta-analysis evaluating DLA in AS-OCT images for the diagnosis of angle closure compared with gonioscopic evaluation. PubMed, Scopus, Embase, Lilacs, Scielo, and Cochrane Central Register of Controlled Trials were searched. The bivariate model was used to calculate pooled sensitivity and specificity. RESULTS: The initial search identified 214 studies, of which 6 were included for final analysis. The total study population included 5,269 patients. The combined sensitivity of the DLA compared with gonioscopy was 94.0% (95% CI 83.8%-97.9%), whereas the pooled specificity was 93.6% (95% CI 85.7%-97.3%). Sensitivity analyses removing each individual study showed a pooled sensitivity in the range of 90.1% to 95.1%. Similarly, specificity results ranged from 90.3 to 94.5% with the removal of each individual study and recalculation of pooled specificity. CONCLUSION: DLA applied to AS-OCT has excellent sensitivity and specificity in the identification of angle closure. This technology may be a valuable resource in the screening of populations without access to experienced ophthalmologists who perform gonioscopy.
ABSTRACT
OBJECTIVES: To perform a meta-analysis of the diagnostic performance of learning (ML) algorithms (conventional and deep learning algorithms) for the classification of malignant versus benign focal liver lesions (FLLs) on US and CEUS. METHODS: Available databases were searched for relevant published studies through September 2022. Studies met eligibility criteria if they evaluate the diagnostic performance of ML for the classification of malignant and benign focal liver lesions on US and CEUS. The pooled per-lesion sensitivities and specificities for each modality with 95% confidence intervals were calculated. RESULTS: A total of 8 studies on US, 11 on CEUS, and 1 study evaluating both methods met the inclusion criteria with a total of 34,245 FLLs evaluated. The pooled sensitivity and specificity of ML for the malignancy classification of FLLs were 81.7% (95% CI, 77.2-85.4%) and 84.8% (95% CI, 76.0-90.8%) for US, compared to 87.1% (95% CI, 81.8-91.0%) and 87.0% (95% CI, 83.1-90.1%) for CEUS. In the subgroup analysis of studies that evaluated deep learning algorithms, the sensitivity and specificity of CEUS (n = 4) increased to 92.4% (95% CI, 88.5-95.0%) and 88.2% (95% CI, 81.1-92.9%). CONCLUSIONS: The diagnostic performance of ML algorithms for the malignant classification of FLLs was high for both US and CEUS with overall similar sensitivity and specificity. The similar performance of US may be related to the higher prevalence of DL models in that group.
