Reporting of imaging diagnostic accuracy studies with focus on MRI subgroup: Adherence to STARD 2015

Abstract

1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018;47:523-544.

Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies

“…Complete reporting of diagnostic test accuracy systematic reviews may be hindered by incomplete reporting in diagnostic test accuracy primary studies. 40 This challenge makes complete…”

Accuracy of the Liver Imaging Reporting and Data System in Computed Tomography and Magnetic Resonance Image Analysis of Hepatocellular Carcinoma or Overall Malignancy—A Systematic Review

“…This was largely due to incomplete reporting (eg, patient factors including cirrhosis etiology and Child-Pugh score were inconsistently provided) or mixed methods (eg, use of extracellular and hepatobiliary agents) at the primary study level; unfortunately lack of complete reporting at the primary study level in imaging is a common issue. 43 Additional variables that warrant consideration as sources of variability include assessment of the effect of liver observation size, LI-RADS major and ancillary features, and differences between imaging modalities (CT vs MRI). Rather than extracting aggregate study-level data as we have, an individual participant data meta-analysis could allow for more granular subgroup analyses.…”

Best practices for MRI systematic reviews and meta‐analyses

“…For the “Patient” category, relevant details might include: species (human); age (adult); gender (male); and prior testing (prostate cancer detected by transrectal biopsy with Gleason score of >7) . For the “Index Test” category, the general principle is to provide enough detail such that the reader could replicate the test in clinical practice . Relevant “index test” details for the prostate study might include, but not be limited to: field strength (eg, 1.5 vs. 3T), contrast agent used (eg, dynamic contrast enhancement performed with extracellular contrast agent with temporal resolution of at least 10 sec), slice thickness for T 2 ‐weighted imaging (≤4 mm), use of endorectal coil, use of diffusion‐weighted imaging (DWI), and which b‐values were used .…”

“…Preferably, quality assessment should be performed by at least two independent authors. Risk of bias can be designated as either “low,” if all signaling questions within the same domain are answered with “yes”; “high,” if any of the signaling question is answered with “no”; or “unclear,” if risk of bias assessment is hampered by a lack of reported data . Concerns with regard to applicability are not based on signaling questions, but represent an overall judgment for a specific domain, rated as “low,” “high,” or “unclear.”…”

“…21 For the "Index Test" category, the general principle is to provide enough detail such that the reader could replicate the test in clinical practice. 22 Relevant "index test" details for the prostate study might include, but not be limited to: field strength (eg, 1.5 vs. 3T), contrast agent used (eg, dynamic contrast enhancement performed with extracellular contrast agent with temporal resolution of at least 10 sec), slice thickness for T 2 -weighted imaging (4 mm), use of endorectal coil, use of diffusion-weighted imaging (DWI), and which b-values were used. 23,24 For the "target condition" category, readers need to be aware whether the disease or target condition is similar to that encountered in their patient population.…”