SIR 2025

General IR

Quality

Session: 2nd Annual Quality Improvement Forum: Improving Outcomes, Safety and Efficiency in IR

Common design and data elements on embolization of pulmonary arteriovenous malformations: A systemic search of literature and report of adherence to quality of reporting among interventional radiology literature.

Monday, March 31, 2025

10:05 AM – 10:14 AM CT

Location: Innovation Theatre

Presenter(s)

Mohammad Mirza-Aghazadeh-Attari, MD, MPH

Postdoctoral Research Fellow
Johns Hopkins University, United States

Author/Co-author(s)

AH

Arezou Hashem Zadeh, MD

Post Doc Research Fellow
Johns Hopkins University School of Medicine, United States
Abby D. Liu, None (she/her/hers)

Student
The Johns Hopkins University School of Medicine, United States
TL

Thalia Liu, None

Student
The Johns Hopkins University School of Medicine, United States
Alan J. Kim, BS (he/him/his)

Medical Student
Johns Hopkins University School of Medicine, United States
SA

Sydney Ashleigh Wade, MPH

Medical student
Johns Hopkins University School of Medicine, United States
CW

Clifford R. Weiss, MD

Professor
Department of Radiology, Radiological Science and Biomedical Engineering, Johns Hopkins University School of Medicine, United States

Purpose: To provide a standardized list of data elements for studies reporting the outcomes of PAVM embolization and determine the degree to which current literature adheres to these standards.

Materials and Methods: A systematic search of the literature was conducted by four independent reviewers using an online repository, following PRISMA guidelines. Covidence, a web-based collaboration platform, was employed to streamline the production of systematic and other types of literature reviews. The reviewers independently searched PubMed, Scopus, Web of Science, Endnote, and ClinicalTrials.gov. Covidence was also used to screen the papers and manage citations for the evidence synthesis process. Additionally, Nested Knowledge (NK) AutoLit and Scite.ai were used to extract common data elements reported in the literature. Based on this extracted information, a rubric was developed and subsequently applied to the selected studies to assess adherence to each data domain.

Results:

A total of 5,511 documents were retrieved, and 3,261 unique documents were included in the initial screening. 340 manuscripts underwent full-text screening. Review articles, commentaries and case reports were excluded. The remaining 271 documents were fed into the large language model capabilities of Scite.ai, and common data points were extracted. Five major themes were detected among the manuscripts, including study design and setting, patient demographics and baseline lesion characteristics, procedural details, and technical and clinical outcome measures. Minor themes could be found on the following link: https://shorturl.at/Gyxzd . The least adherence was seen among the following criteria: Type of PAVM persistence after embolization Quality of life metrics (9.1%), time to occlusion (15.2%), patient Race (19%), accurate reporting of PAVM characteristics (23.2%).

Conclusion:

The development of various embolic devices and agents for the treatment of PAVMs has highlighted the need for studies documenting treatment outcomes and comparing the efficacy of these agents. One approach to generating high-quality evidence is through meta-analyses of existing studies. However, this has not been feasible due to inconsistencies in the reporting of results in the published literature. In this work, we present a series of standardized data elements, derived through text analysis, to facilitate effective meta-analytic processing of the literature.