A 20‐year scoping review of the veterinary interventional radiology and interventional endoscopy literature (2000‐2019)

Abstract Background Interventional radiology (IR) and interventional endoscopy (IE) have broad potential for minimally invasive therapy in veterinary patients, but the scope of original peer‐reviewed veterinary IR/IE research publications has not been described. Objectives Catalogue published applications and indications for noncardiac therapeutic IR/IE in animals and describe type and quality of veterinary IR/IE research over 20 years. Methods Highly‐cited veterinary journals were searched to identify articles published 2000 to 2019 involving therapeutic IR/IE applications for clinical veterinary patients. Articles were assigned a level of evidence (LOE) according to published standards. Authorship, animal data, study design, and interventions were described. Change in publication rate, study size, and LOE of IR/IE articles over time was analyzed. Results One hundred fifty‐nine of 15 512 (1%) articles were eligible, including 2972 animals. All studies were low LOE and 43% were case reports with ≤5 animals. Number of IR/IE articles per year (P < .001), proportion of journals' articles pertaining to IR/IE (P = .02), and study size (P = .04) all increased over time, but LOE (P = .07) did not. Common target body systems were urinary (40%), digestive (23%) respiratory (20%), and vascular (13%). Common indications were nonvascular luminal obstructions (47%), object retrieval (14%), and congenital anomalies (13%). Most procedures involved indwelling medical devices or embolic agents, whereas tissue resection and other procedures were less common. Procedures utilized fluoroscopy (43%), endoscopy (33%), ultrasound (8%), digital radiography (1%), or fluoroscopy in combination with other modalities (16%). Conclusions Treatments involving IR/IE have wide applicability in veterinary medicine but large, rigorous, and comparative studies describing these procedures are lacking.

for treating disease conditions such as small-to moderately sized tumors where tissue sparing is desired (eg, renal, liver, and lung tumors and metastases), 9 certain nonresectable tumors, 10 ureteral obstruction, 11 vascular anomalies, 12 and portal hypertension. 13 In veterinary medicine, IR treatment for congenital and acquired cardiac disease has been widely reported and practiced during the past two decades, 14   screening period, the journal was excluded from the study. If at least one eligible IR/IE article was identified during the 6-year screening period, the journal was included in the study. Eligible articles were original research reports in which a study objective involved describing, comparing, or otherwise evaluating a therapeutic IR or IE technique intended for use in clinical veterinary patients. Table 1 describes the definitions of IR/IE and types of included and excluded interventions used for the purposes of this review. Interventional orthopedic and computer-assisted surgical applications were excluded. Articles relating to treatments applied to the heart muscle, valves, or immediate inflow or outflow vasculature were excluded. Editorials, book reviews, narrative review articles, and studies in which IR/IE occurred but was not central to study objectives were excluded. An example of an article excluded on the latter basis would be a study of urinary cancer chemotherapy outcomes in which some animals were treated with stents but the stent technique or outcomes were not specifically addressed. Articles in which IR/IE techniques were applied with the aid of open surgery were retained.
A single investigator performed an online manual search of titles, abstracts, and full texts to screen each journal issue for potentially eligible articles. For each article identified, two investigators independently read title, abstract and full text to determine whether the article met study inclusion criteria, discussed their results, and resolved discrepancies by joint review and discussion. Data charting was performed by a single investigator using a form jointly developed by two investigators including publication year and journal, first and corresponding author names and institutional affiliations, country where the research was performed, study design, species, number of animals studied, body system to which the intervention was applied, type of intervention, and indication for intervention. Each article was assessed for the presence of specific methodologic features and assigned a LOE according to guidelines published by the Oxford Centre for Evidence-Based Medicine. 15 Single case reports, case series with ≤5 animals, and preclinical studies such as those involving research animals or cadaver tissues were rated level 5 evidence. Levels of evidence rated 1 or 2 were considered "high" level evidence and LOE rated 3, 4, or 5 were considered "low" level evidence. 16 Summary statistics were used to convey descriptive data. Simple linear regression was used to test the null hypotheses of no linear relationship between advancing year and the following predictor variables: annual number of IR/IE articles published, annual proportion of total journal articles pertaining to IR/IE, and number of study subjects (n) per IR/IE article. The nonparametric Cuzick test for trend using ranks was used to evaluate the association between LOE of IR/IE articles (graded 1-5, where 1 is highest LOE and 5 is lowest LOE) and publication year. Tests were 2-sided and P < .05 was considered statistically significant.

| RESULTS
Among 11 journals screened using the 6-year review protocol, 4 did not publish any IR/IE articles during the 6 years and were excluded from the study ( Table 2). Among the remaining 7 journals, the search    The limitations of our study primarily pertain to potential selection bias at the levels of search and review. We limited our review to 11 English language veterinary journals but it is likely we could have identified more articles using a wider manual search strategy of more titles. Alternatively, we might have identified a larger number of sources by entering a limited number of relevant terms in a large database, but the lack of standardized terminology or medical subjects headings and wide variety of procedures and diseases led us to avoid a database strategy. Two investigators reviewed each journal issue manually to identify eligible articles and it is possible that some relevant articles were missed if IR/IE procedures were poorly described, constituted a minor part of the article, or were simply overlooked by both investigators. We did not contact journals to determine if any policies existed during 2000 to 2019 with regard to the publication of IR/IE content, case reports, or study sample sizes. Therefore, it is possible that observed publication patterns could reflect editorial policy or publication bias rather than the generation of content at the investigator level.
Our scoping review highlights for the first time the wide application of IR/IE in veterinary medicine and the thousands of animals that have received these treatments. The body of evidence however is lacking large, rigorous, and comparative studies, and many publications are attributed to a small number of institutions.

ACKNOWLEDGMENT
No funding was received for this study.