Assessment of safety and procedural learning curve for pulmonary embolism patients undergoing percutaneous mechanical thrombectomy

Abstract

Background: Advancements in percutaneous mechanical thrombectomy (PMT) have emerged as a promising alternative for the treatment of intermediate-risk pulmonary embolism (PE), demonstrating a favorable safety profile and low 30-day mortality. Its adoption and procedural learning curve remain underexplored. This study evaluates the safety, procedural efficiency, and learning curve of PMT over time, using fluoroscopy time, procedure time, and contrast volume as primary metrics.

Methods: This was a single-institution, multicenter retrospective study of 411 patients who underwent PMT for intermediate-risk PE from January 2020 to July 2024 across three medical centers involving 15 vascular surgeons. High-risk patients were excluded from the study. Linear regression analysis assessed trends in procedural efficiency, and segmented regression and cumulative sum analyses identified learning curve plateaus. We used t tests to compare hospital and intensive care unit length of stay, as well as complication rates between the first 50 and most last 50 cases. Additionally, multivariable regression examined the impact of patient factors-including body mass index, sex, age, and comorbidities-on procedural efficiency. All statistical analyses were performed using Python (Python Software Foundation, Wilmington, DE).

Results: Regression analysis revealed statistically significant decreases in fluoroscopy time (R² = 0.103; P < .001), contrast volume (R² = 0.071; P < .001), and procedure time (R² = 0.068; P < .001) over time. Segmented regression identified proficiency plateaus at 55 cases for fluoroscopy time and 138 cases for procedure time. On an individual basis, an estimated 3.4 cases were required to achieve imaging efficiency proficiency and 8.6 cases for procedural proficiency per surgeon. Cumulative sum analysis demonstrated a consistent negative trend in fluoroscopy time, procedure time, and contrast volume, confirming sustained improvements in procedural efficiency over the study period. The t tests comparing the first and last 50 cases showed no statistically significant differences in hospital length of stay (6.23 days vs 5.00 days; P = .406), intensive care unit length of stay (2.31 days vs 1.23 days; P = .256), or complication rate (14% vs 0%; P = .134). The overall mortality rate for the study cohort was 4.49%. No significant trend in 30-day mortality was observed over time, consistent with the low overall mortality rate in this intermediate-risk cohort.

Conclusions: PMT for PE demonstrates a favorable learning curve, with significant improvements in procedural efficiency over time. Fluoroscopy time and procedure time plateau at distinct points, reflecting different aspects of skill acquisition, although other factors like anatomical variability and operating room staff could have had an impact. The procedure maintains low complication and mortality rates, supporting its safety and feasibility across a range of clinical settings. These findings suggest that PMT can be adopted rapidly by vascular surgeons with minimal impact from patient demographics or comorbidities.

Keywords: Catheter-directed; Learning curve; Minimally invasive; Percutaneous mechanical thrombectomy; Procedural efficiency; Pulmonary embolism; Thrombectomy.