Discussion

An evidence-based approach to test accuracy studies in gynecologic oncology: the ‘STARD’ checklist

Given the wide usage of emergency point-of-care ultrasound (EUS) among emergency physicians (EPs), rigorous study surrounding its accuracy is essential. The Standards for Reporting of Diagnostic Accuracy (STARD) criteria were established to ensure robust reporting methodology for diagnostic studies. Adherence to the STARD criteria among EUS diagnostic studies has yet to be reported.

Our objective was to evaluate a body of EUS literature shortly after STARD publication for its baseline adherence to the STARD criteria.

EUS studies in 5 emergency medicine journals from 2005–2010 were evaluated for their adherence to the STARD criteria. Manuscripts were selected for inclusion if they reported original research and described the use of 1 of 10 diagnostic ultrasound modalities designated as “core emergency ultrasound applications” in the 2008 American College of Emergency Physicians Ultrasound Guidelines. Literature search identified 307 studies; of these, 45 met inclusion criteria for review.

The median STARD score was 15 (interquartile range [IQR] 12–17), representing 60% of the 25 total STARD criteria. The median STARD score among articles that reported diagnostic accuracy was significantly higher than those that did not report accuracy (17 [IQR 15–19] vs. 11 [IQR 9–13], respectively; p < 0.0001). Seventy-one percent of articles met ≥50% of the STARD criteria (56–84%) and 4% met >80% of the STARD criteria.

Significant opportunities exist to improve methodological reporting of EUS research. Increased adherence to the STARD criteria among diagnostic EUS studies will improve reporting and improve our ability to compare outcomes.

In this study, we conducted a systematic review of the literature to re-evaluate the role of C4d in the diagnosis of acute antibody-mediated rejection of kidney allografts. Electronic databases were searched until September 2013. Eligible studies allowed derivation of diagnostic tables for the performance of C4d by immunofluorescence or immunohistochemistry with comparison to histopathological features of acute antibody-mediated rejection and/or donor-specific antibody (DSA) assays. Of 3492 unique abstracts, 29 studies encompassing 3485 indication and 868 surveillance biopsies were identified. Assessment of C4d by immunofluorescence and immunohistochemistry exhibited slight to moderate agreement with glomerulitis, peritubular capillaritis, solid-phase DSA assays, DSA with glomerulitis, and DSA with peritubular capillaritis. The sensitivity and specificity of C4d varied as a function of C4d and comparator test thresholds. Prognostically, the presence of C4d was associated with inferior allograft survival compared with DSA or histopathology alone. Thus, our findings support the presence of complement-dependent and -independent phenotypes of acute antibody-mediated rejection. Whether the presence of C4d in combination with histopathology or DSA should be considered for the diagnosis of acute antibody-mediated rejection warrants further study.

To evaluate the extent to which test accuracy studies published in two leading reproductive medicine journals in the years 1999 and 2004 adhered to the Standards for Reporting of Diagnostic Accuracy (STARD) initiative parameters, and to explore whether the introduction of the STARD statement has led to an improved quality of reporting.

Structured literature search. Articles that reported on the diagnostic performance of a test in comparison with a reference standard were eligible for inclusion. For each article we scored how well the 25 items of the STARD checklist were reported. These items deal with the study question, study participants, study design, test methods, reference standard, statistical methods, reporting of results, and conclusions. We calculated the total number of reported STARD items per article, summary scores for each STARD item, and the average number of reported STARD items per publication year.

Not applicable.

Not applicable.

Not applicable.

Quality of reporting.

We found 24 studies reporting on test accuracy in reproductive medicine in 1999 and 27 studies in 2004. The mean number of reported STARD items for articles published in 1999 was 12.1 ± 3.3 (range 6.5–20) and 12.4 ± 3.2 (range 7–17.5) in 2004, after publication of the STARD statement. Overall, less than half of the studies reported adequately on 50% or more of the STARD items. The reporting of individual items showed a wide variation. There was no significant improvement in mean number of reported items for the articles published after the introduction of the STARD statement.

Authors of test accuracy studies in the two leading fertility journals poorly report the design, conduct, methodology, and statistical analysis of their study. Strict adherence to the STARD guidelines should be encouraged.

The objective of this study was to determine the diagnostic accuracy of magnetic resonance imaging (MRI) in detecting cervical involvement by endometrial cancer.

A retrospective accuracy study of 135 consecutive women who underwent preoperative MRI and surgery for endometrial cancer at a single gynaecological cancer centre between 1st February 2003 and 30th November 2004.

For the detection of any cervical involvement by MRI, the sensitivity was 72%, specificity 93.2%, positive predictive value (PPV) 89.8%, negative predictive value (NPV) 80.2%, positive likelihood ratio (+LR) 10.7 and negative likelihood ratio (−LR) 0.3. When cervical stromal invasion was considered alone, the sensitivity was 84.4%, specificity 87.4%, PPV 67.5%, NPV 94.7%, +LR 6.7 and −LR 0.18.

We believe that MRI is able to accurately predict cervical involvement in endometrial cancer and allows a decision to be made on the type of hysterectomy to be offered.

To determine the accuracy of minimally and non-invasive tests to assess the groin node status in squamous cell vulvar cancer.

A systematic review of published research from 1979 to 2004 that compares the results of tests to determine groin node status with histology at inguinofemoral lymphadenectomy was made. Studies included in the review were those that compared the index test to the standard surgical intervention of inguinofemoral lymphadenectomy and allowed the construction of two-by-two tables. From these tables, sensitivity, specificity, and the likelihood ratios (with 95% confidence intervals) were reported and, where feasible, meta-analysis was used to pool results for each test separately. Sentinel node biopsy using technetium-99m-labelled nanocolloid (^99mTc) had a pooled sensitivity and negative LR of 97% (91–100 95% CI) and 0.12 (0.053–0.28 95% CI), respectively, and was the most accurate test reviewed.

Five diagnostic tests were identified in a total of 29 studies (961 groins). Although the studies were small and the design often poor, this represents the best summary of the data to date. Sentinel node identification using ^99mTc appeared to be the most promising test for accurately excluding lymph node metastases in squamous cell vulvar cancer and potentially reducing the radicality of surgery. Its efficacy as a tool in reducing the need for radical surgery and associated patient morbidity without reducing survival needs further assessment probably in a randomised control trial.