Abstract
Background:
Intravascular imaging (IVI) with percutaneous coronary intervention (PCI) in the United States has historically been low, however, there have been recent increases; IVI use in Medicare increased from 15.3 to 34.6 per 100 PCI between 2019 and 2023. The characteristics of operators who adopted IVI versus those who did not are unknown.
Methods:
Using Medicare Provider Utilization and Payment data, we identified PCI operators in 2019 and 2023 with ≥30 PCI annually. IVI use was categorized as low/none (IVI to PCI ratio [IPR] of 0.0–0.30), intermediate (IPR 0.31–0.70) and high (IPR ≥0.71). Regression models identified provider characteristics associated with IVI adoption.
Results:
In 2019, 3385 operators performed ≥30 PCI. IVI use in 2019 was low/none among 77.8% operators, intermediate in 15.7%, and high in 6.4%. In 2023, 2289 operators performed ≥30 PCI. IVI use in 2023 was low/none among 48.8% operators, intermediate in 21.8%, and high in 29.3%. Among 1671 operators active in both years, 30.9% of prior low/none users adopted IVI. Only 12.2% of intermediate or high IVI users had a reduction in IVI use. Newer high-volume operators showed high IVI use (51.1% high use). Non-adoption of IVI was associated with earlier graduation year, smaller hospitals, Southern region, and higher procedural volume, while gender, practice focus, teaching status of affiliated hospital, and hospital affiliation changes were not significant.
Conclusion:
IVI use has increased nationally but adoption remains uneven, with structural and generational factors driving persistent disparities and highlighting the need for targeted implementation strategies.
INTRODUCTION
Intravascular imaging (IVI) use with percutaneous coronary intervention (PCI) in the US has been historically low. We previously showed at an operator-level that among intermediate- to high-volume operators in 2019, IVI was low (≤30%) among 3 in 4 operators.1 However, recently there has been an increase in data2 supporting IVI use. In parallel, IVI use in the US has increased from 15.3 to 34.6 per 100 PCI from 2019 to 2023 (Medicare data) - an encouraging trend but with room for further improvement. In this context, understanding the differences between operators who adopted IVI and those who continued to have low IVI use can inform future policy and training efforts.
In this study using Medicare Provider Utilization and Payment (MPUP) data we compare IVI use by individual operators in 2019 to that in 2023.
METHODS
Data source and study population
Using MPUP for 2019 and 2023, we identified PCI operators and obtained their PCI and IVI volumes using Current Procedural Terminology (CPT) codes. We focused on operators with annual Medicare Fee-for-Service PCI volume ≥30 - a cohort representative of intermediate to high-volume operators based on prior studies.1 This study was exempt from institutional review board approval.
Outcomes and covariates
IVI use was categorized as low/none (IVI to PCI ratio [IPR] of 0.0–0.30), intermediate (IPR 0.31–0.70) and high (IPR ≥0.71). The lower cutoff was set at 0.30 as the MPUP data suppress CPT codes with <10 billings; thus an operator with ≥30 PCI and no IVI billing has <10 IVI and thus an IPR ≤0.30.
Operator characteristics were obtained by linking National Provider Identification numbers to Medicare Physician Compare data, Doctors and Clinicians National Downloadable Files, and Inpatient Prospective Payment System Impact Files. Operator characteristics included gender, practice focus, PCI volume (≥50 PCI vs <50), US census region, medical school graduation year quartile (Q1: earlier graduation versus Q4: recent graduation), bed size of primary affiliated facility (Q1: smaller hospital versus Q4: larger hospital), teaching status, and whether PCI operator changed their primary affiliated facility i.e primary practice site. The four practice focus categories were defined based on whether the PCI operator performed only PCI or additionally performed peripheral or structural procedures using their billed CPT codes (Supplementary Table 1) - PCI only, PCI with peripheral, PCI with structural, and PCI with both structural and peripheral.
Statistical analysis
Categorical variables were compared using χ2 tests. A generalized linear model with logit link was used to identify independent variables. Significance was set at a 2-sided P value of .05. Analyses were performed using Python and Stata v17.
RESULTS
In 2019, 3385 operators performed ≥30 PCI (intermediate- to high-volume PCI operators). IVI use in 2019 was low/none among 77.8% operators, intermediate in 15.7%, and high in 6.4%. In 2023, 2289 operators performed ≥30 PCI. IVI use in 2023 was low/none among 48.8% operators, intermediate in 21.8%, and high in 29.3%.
Among the 2019 intermediate- to high-volume PCI operators, 1671 (49.3%) continued to perform ≥30 PCIs in Medicare FFS in 2023; 1116 (32.9%) performed between 10–29 PCIs, and 598 (17.6%) performed no or <10 PCI.
Among the 1671 operators who performed ≥30 PCIs in both 2019 and 2023, the proportion of IVI use in 2023 was low/none in 54.6%, intermediate in 21.4%, and high in 24%. Among low/no IVI use operators of 2019 (N=1263), 30.9% adopted IVI (became intermediate or high IVI users) in 2023. Only 12.2% of intermediate or high IVI users had a reduction in IVI use (Figure 1).
Figure 1.

Sankey diagram of IVI use in 2019 compared with 2023 of operators who performed ≥30 PCIs in both 2019 and 2023.
Among operators performing ≥30 PCI in 2023 but none (≤10) in 2019 (presumably new operators as their median medical school graduation year was 2012), IVI use was low/none in 22.9%, intermediate in 26%, and high in 51.1%.
The proportion of 2019 low/no IVI use operators who adopted IVI in 2023 by operator characteristics is shown in Figure 2. In the adjusted model, the characteristics independently associated with not adopting IVI in 2023 were those who graduated earlier [OR 2.6 (1.7–4.0) 1st vs 4th quartile] or practicing at smaller hospitals [OR 2.8 (1.8–4.2)] 1st vs 4th quartile), and those in the South [OR 2.1 (1.5–3.1)] compared to the Northeast and higher volume operators [OR 1.4 (1.1–1.8) ≥50 PCI vs <50 PCI]. Operator gender, practice focus, teaching status, and change in hospital affiliation from 2019 to 2023 had no influence on IVI use. Figure 3.
Figure 2.

Proportion of low/no IVI users in 2019 who adopted IVI in 2023 by operator characteristics. Chi2 *P <0.01, **P<0.001
Figure 3.

Forest plot of a generalized linear model with non-adoption of intravascular imaging as the dependent variable and operator characteristics as independent variables (gender, practice focus, PCI volume, US census region, medical school graduation year quartile, hospital bed size quartile, teaching status, and facility change from 2019 to 2023).
DISCUSSION
In this national analysis of intermediate-high volume PCI operators intravascular imaging (IVI) adoption increased between 2019 and 2023. In 2019, IVI use was predominantly low or absent (77.8% of operators), whereas by 2023 nearly one-third were high users and fewer than half remained low/no users. Adoption was dynamic: nearly one-third of previously low/no users transitioned to intermediate or high use, and more than half of intermediate users escalated to high IVI use.
Adoption was uneven across operator and institutional characteristics. Non-adopters were more likely to have graduated earlier (i.e in practice longer), practice at smaller hospitals, and located in the South – associations persisted after adjustment. In contrast, gender, practice focus, teaching status of affiliated hospital, and hospital affiliation changes were not associated with IVI adoption, indicating that structural and generational factors may be more influential. Interestingly, higher-volume operators were also less likely to adopt IVI, suggesting that established practice patterns among experienced or busy operators may limit incorporation of newer adjunctive technologies. Notably, these factors mirror those previously associated with low IVI use in general1, underscoring the importance of structural and local practice environments in shaping technology diffusion, and thus low IVI users in environments with higher IVI use may be more likely to adopt IVI.
Only a small minority (12.2%) of intermediate or high users de-escalated, suggesting that once incorporated into practice, IVI tends to be sustained. However, attrition in PCI volume was substantial, with only half of 2019 ≥30 PCI operators maintaining this threshold in 2023. Total PCI volume in Medicare reduced from 369,771 in 2019 to 279,842 in 2023. Although there has been a 12% shift in coverage from Medicare FFS to Medicare Advantage3 during this period it does not fully account for this decline and likely reflects broader changes in operator practice patterns over time.
These findings carry important clinical and policy implications. Despite evidence supporting imaging-guided PCI, substantial regional and institutional disparities persist in IVI use. Barriers may include resource constraints, training exposure, reimbursement considerations, and practice culture. Targeted educational efforts, mentorship, and institutional support can narrow these gaps.
Limitations include restriction to intermediate- and high-volume operators due to data suppression, inability to link IVI to specific PCI procedures, inability to attribute PCI practice variables such as complex PCI operators due to data limitations, and generalizability limited to Medicare FFS. Further, the study period does not reflect operator response to upgraded recommendations for IVI in recent guidelines4 or recent trial data showing mixed results for IVI.5
CONCLUSION
IVI use is increasing in the US, but adoption trends are uneven. Lower adoption in smaller institutions, certain geographic regions and by operators longer in practice could improve with targeted efforts.
Funding:
Dr Murugiah received support from the National Heart, Lung, and Blood Institute of the National Institutes of Health (under award K08HL157727).
Footnotes
Disclosures: Dr Krumholz has received options for Element Science and Identifeye and payments from F-Prime for advisory roles; he is a co-founder of and holds equity in Hugo Health, Refactor Health, and ENSIGHT-AI; and he is associated with research contracts through Yale University from Janssen, Kenvue, Novartis, and Pfizer.
REFERENCES
- 1.See C, Sawano M, Nagpal S, Chamie D, Curtis JP, Murugiah K. Intravascular Imaging Use by Intermediate to High-Volume US Operators – A Medicare Data Analysis. Journal of the Society for Cardiovascular Angiography & Interventions. 2023;2. doi: 10.1016/j.jscai.2023.100977 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Zhang J, Gao X, Kan J, Ge Z, Han L, Lu S, Tian N, Lin S, Lu Q, Wu X, et al. Intravascular Ultrasound Versus Angiography-Guided Drug-Eluting Stent Implantation. Journal of the American College of Cardiology. 2018;72:3126–3137. doi: 10.1016/j.jacc.2018.09.013 [DOI] [PubMed] [Google Scholar]
- 3.Kaiser Family Foundation: KFF analysis of CMS Medicare Advantage Enrollment Files, 2010–2025. 2025.
- 4.Rao Sunil V, O’Donoghue Michelle L, Ruel M, Rab T, Tamis-Holland Jaqueline E, Alexander John H, Baber U, Baker H, Cohen Mauricio G, Cruz-Ruiz M, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes. JACC. 2025;85:2135–2237. doi: 10.1016/j.jacc.2024.11.009 [DOI] [PubMed] [Google Scholar]
- 5.Diletti R, Daemen J, Faurie B, Barbierato M, Tchétché D, Hovasse T, Teeuwen K, Cheng Jin M, Landt M, Campo G, et al. Intravascular Ultrasound–Guided or Angiography-Guided Complex High-Risk PCI. New England Journal of Medicine.0. doi: 10.1056/NEJMoa2601521 [DOI] [PubMed] [Google Scholar]
