Letter to the EditorBivariate meta-analysis of sensitivity and specificity with sparse data: a generalized linear mixed model approach
References (5)
- J.B. Reitsma et al.
Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews
J Clin Epidemiol
(2005) - H.C. van Houwelingen et al.
Advanced methods in meta-analysis: multivariate approach and meta-regression
Stat Med
(2002)
Cited by (540)
Systematic review and meta-analysis of diagnostic test accuracy of ST-segment elevation for acute coronary occlusion
2024, International Journal of CardiologyTo evaluate the diagnostic sensitivity and specificity of ST-segment elevation on a 12‑lead ECG in detecting ACO across any coronary artery, challenging the current STEMI-NSTEMI paradigm.
Studies from MEDLINE and Scopus (2012−2023) comparing ECG findings with coronary angiograms were systematically reviewed and analyzed following PRISMA-DTA guidelines. QUADAS-2 assessed the risk of bias.
Studies included focused on AMI patients and provided data enabling the construction of contingency tables for sensitivity and specificity calculation, excluding those with non-ACS conditions, outdated STEMI criteria, or a specific focus on bundle branch blocks or other complex diagnoses. Data were extracted systematically and pooled test accuracy estimates were computed using MetaDTA software, employing bivariate analyses for within- and between-study variation. The primary outcomes measured were the sensitivity and specificity of ST-segment elevation in detecting ACO.
Three studies with 23,704 participants were included. The pooled sensitivity of ST-segment elevation for detecting ACO was 43.6% (95% CI: 34.7%–52.9%), indicating that over half of ACO cases may not exhibit ST-segment elevation. The specificity was 96.5% (95% CI: 91.2%–98.7%). Additional analysis using the OMI-NOMI strategy showed improved sensitivity (78.1%, 95% CI: 62.7%–88.3%) while maintaining similar specificity (94.4%, 95% CI: 88.6%–97.3%).
The findings reveal a significant diagnostic gap in the current STEMI-NSTEMI paradigm, with over half of ACO cases potentially lacking ST-segment elevation. The OMI-NOMI strategy could offer an improved diagnostic approach. The high heterogeneity and limited number of studies necessitate cautious interpretation and further research in diverse settings.
Application of the convolution neural network in determining the depth of invasion of gastrointestinal cancer: a systematic review and meta-analysis
2024, Journal of Gastrointestinal SurgeryWith the development of endoscopic technology, endoscopic submucosal dissection (ESD) has been widely used in the treatment of gastrointestinal tumors. It is necessary to evaluate the depth of tumor invasion before the application of ESD. The convolution neural network (CNN) is a type of artificial intelligence that has the potential to assist in the classification of the depth of invasion in endoscopic images. This meta-analysis aimed to evaluate the performance of CNN in determining the depth of invasion of gastrointestinal tumors.
A search on PubMed, Web of Science, and SinoMed was performed to collect the original publications about the use of CNN in determining the depth of invasion of gastrointestinal neoplasms. Pooled sensitivity and specificity were calculated using an exact binominal rendition of the bivariate mixed-effects regression model. I2 was used for the evaluation of heterogeneity.
A total of 17 articles were included; the pooled sensitivity was 84% (95% CI, 0.81-0.88), specificity was 91% (95% CI, 0.85-0.94), and the area under the curve (AUC) was 0.93 (95% CI, 0.90-0.95). The performance of CNN was significantly better than that of endoscopists (AUC: 0.93 vs 0.83, respectively; P = .0005).
Our review revealed that CNN is one of the most effective methods of endoscopy to evaluate the depth of invasion of early gastrointestinal tumors, which has the potential to work as a remarkable tool for clinical endoscopists to make decisions on whether the lesion is feasible for endoscopic treatment.
Systematic review and meta-analysis of the accuracy of McIsaac and Centor score in patients presenting to secondary care with pharyngitis
2024, Clinical Microbiology and InfectionCentor and McIsaac scores are clinical prediction rules for diagnosing group A streptococcus (GAS) infection in patients with pharyngitis. Their recommended thresholds vary between guidelines.
To estimate the sensitivity and specificity of the McIsaac and Centor scores to diagnose GAS pharyngitis and evaluate their impact on antibiotic prescribing at each threshold in patients presenting to secondary care.
MEDLINE, Embase, and Web of Science were searched from inception to September 2022.
Studies of patients presenting with acute pharyngitis to emergency or outpatient clinics that estimated the accuracy of McIsaac or Centor scores against throat cultures and/or rapid antigen detection tests (RADT) as reference standards.
Centor or McIsaac score.
Throat cultures and/or RADT.
Quality Assessment of Diagnostic Accuracy Studies.
The sensitivities and specificities of the McIsaac and Centor scores were pooled at each threshold using bivariate random effects meta-analysis.
Fourteen studies were included (eight McIsaac and six Centor scores). Eight studies had unclear and six had a high risk of bias. The McIsaac score had higher estimated sensitivity and lower specificity relative to Centor scores at equivalent thresholds but with wide and overlapping confidence regions. Using either score as a triage to RADT to decide antibiotic treatment would reduce antibiotic prescription to patients with non-GAS pharyngitis relative to RADT test for everyone, but also reduce antibiotic prescription to patients with GAS.
Centor and McIsaac scores are equally ineffective at triaging patients who need antibiotics presenting with pharyngitis at hospitals. At high thresholds, too many true positive cases are missed, whereas at low thresholds, too many false positives are treated, leading to the over prescription of antibiotics. The former may be compensated by adequate safety netting by clinicians, ensuring that patients can seek help if symptoms worsen.
Programmed cell death ligand-1 (PD-L1) expression on tumor cells, evaluated by immunohistochemistry, guides the use of immunotherapy in advanced non-small cell lung cancer (NSCLC).
What is the sensitivity and specificity of PD-L1 testing performed in cytologic vs paired histologic specimens in patients with NSCLC?
The MEDLINE, Embase, Web of Science, and Cochrane Library databases were searched through June 1, 2021. The primary outcome was pooled sensitivity and specificity of PD-L1 testing performed on cytologic specimens compared with the reference standard of histologic specimens, analyzed at the PD-L1 expression cutoffs (tumor proportion score) ≥ 1% and ≥ 50%. Pooled sensitivity and specificity, and associated 95% CIs, were estimated using bivariate generalized linear mixed models.
Twenty-six articles were included, encompassing a total of 1,064 pairs of histology specimens and cytology cell blocks, and 267 pairs of histology specimens and direct smears. Among these, 946 paired specimens were acquired without interval treatment between the collection of histology and cytology samples. The pooled sensitivity and specificity of cytology specimens compared with paired histology specimens at the PD-L1 expression cutoff ≥ 1% were 0.84 (95% CI, 0.77-0.89) and 0.88 (95% CI, 0.82-0.93), respectively, whereas the pooled sensitivity and specificity at cutoff ≥ 50% were 0.78 (95% CI, 0.69-0.86) and 0.94 (95% CI, 0.91-0.96), respectively. When only paired specimens acquired without interval treatment were considered, the pooled sensitivity and specificity of cytology specimens at PD-L1 expression cutoff ≥ 1% were 0.84 (95% CI, 0.76-0.90) and 0.89 (95% CI, 0.82-0.94), respectively, whereas the pooled sensitivity and specificity at cutoff ≥ 50% were 0.80 (95% CI, 0.71-0.89) and 0.94 (95% CI, 0.91-0.96), respectively.
Cytologic specimens provide an accurate assessment of PD-L1 expression in most patients with NSCLC, at both ≥ 1% and ≥ 50% cutoffs, when compared with histologic specimens.
PROSPERO; No.: CRD42020153279; URL: https://www.crd.york.ac.uk/prospero/
Clinical effectiveness of point of care tests for diagnosing urinary tract infection: a systematic review
2024, Clinical Microbiology and InfectionPoint of care tests (POCTs) have the potential to improve the urinary tract infection (UTI) diagnostic pathway, as they can provide a diagnosis quickly in near-patient settings, and some also identify causative pathogens/antimicrobial sensitivity.
To assess the clinical impact, accuracy, and technical characteristics of POCT for diagnosing UTI.
Narrative summary and bivariate random effects meta-analyses to estimate summary sensitivity and specificity.
Five electronic databases, two clinical trial registries, study reports and review reference lists, and websites.
Randomized controlled trials/non-randomized studies and diagnostic test accuracy studies published since 2000.
People with suspected UTI.
Rapid tests (results <40 minutes): Astrego PA-100 system, Lodestar DX, Uriscreen, UTRiPLEX. Culture tests (results <24 hours): Flexicult Human, ID Flexicult, Diaslide, Dipstreak, Chromostreak, Uricult, Uricult Trio, Uricult Plus.
Any.
Risk of Bias-2, Quality Assessment of Diagnostic Accuracy Studies-2, Quality Assessment of Diagnostic Accuracy Studies-C.
Two randomized controlled trials evaluated Flexicult Human (one against standard care; one against ID Flexicult). No difference was reported in antibiotic use concordant with culture results (OR 0.84 95% CI 0.58–1.20) or appropriate antibiotic prescribing (OR 1.44 95% CI 1.03–1.99). Initial antibiotic prescribing was lower with Flexicult than standard care (OR 0.56 95% CI 0.35–0.88). No difference for other measures of antibiotic use, symptom duration, patient enablement, or resource use. Fifteen studies reported accuracy data. Limited data were available, with most POCT evaluated in single studies or not evaluated at all. Uriscreen (four studies), Uricult Trio (three studies), Flexicult Human (four studies), and ID Flexicult (two studies) had modest sensitivity and specificity. POCTs were easier to use and interpret than standard culture.
There is currently insufficient evidence to support the use of POCTs in UTI diagnosis. Due to the rapid development of POCT, this review should be updated regularly.
The accuracy of ultrasound scan in diagnosing retained products of conception: a systematic review and meta-analysis
2024, American Journal of Obstetrics and GynecologyThis study aimed to analyze and summarize the evidence on the accuracy of different ultrasound methods in the diagnosis of retained products of conception.
We searched Ovid SP, the Cumulative Register to Nursing & Allied Health Literature, EBSCO, and grey literature including Core, Trip, Networked Digital Library of Theses and Dissertations Global ETD search, BMJ Best Practice, PubMed, GreyLit report website (http://www.greylit.org/), Cochrane Central Register of Controlled Trials, and Google scholar (https://scholar.google.com/).
We included prospective and retrospective cross-sectional or Cohort studies that evaluated both ultrasound findings (before management of retained products of conception) and histopathologic results of retained products of conception at all gestational ages.
We used Covidence for data extraction from the studies and quality assessment. The meta-analysis was performed using RevMan 5.4 (forest plot), MetaDTA version 2.01, and Meta-DiSc 2.0 online software.
In total, 11 studies were eligible for data extraction and meta-analysis. The total number of study participants from these 11 studies were 1567. Of these, 9 studies were included to test the accuracy of an echogenic mass, 4 studies analyzed the accuracy of endometrial thickness, and 5 studies analyzed the accuracy of color Doppler flow to predict retained products of conception. We found that echogenic mass had the highest sensitivity, specificity, and diagnostic odds ratio for predicting retained products of conception. The sensitivity, specificity, and diagnostic odds ratio were 0.915 (95% confidence interval, 0.844–0.955), 0.843 (95% confidence interval, 0.615–0.947), and 57.787 (95% confidence interval, 15.171–220.112), respectively. The diagnostic threshold for endometrial thickness was set at 10 mm with a sensitivity, specificity, and diagnostic odds ratio of 0.667 (95% confidence interval, 0.072–0.981), 0.866 (95% confidence interval, 0.375–0.986), and 12.927 (95% confidence interval, 0.23–726.582). The sensitivity, specificity, and diagnostic odds ratio of color Doppler flow were 0.850 (95% confidence interval, 0.756–0.913), 0.406 (95% confidence interval, 0.198–0.655), and 3.893 (95% confidence interval, 1.005–15.081).
Our review concluded that an echogenic mass is the most sensitive and specific predictor of retained products of conception after any pregnancy event. The most important limitation of our review is that the design of the studies included led to significant statistical heterogeneity.