Learning Curve in Thoracoscopic Mediastinal Mass Excision: A Multicenter Retrospective Cohort Study Using Cumulative Sum and Multivariable Regression Analysis.

Background: Video-assisted thoracoscopic surgery (VATS) mediastinal mass excision has become the standard approach for anterior mediastinal lesions. Understanding its learning curve is essential for surgical training and quality assurance. The primary objective was to quantify the association between cumulative case experience (case sequence number) and operative time, the primary outcome, using cumulative sum (CUSUM) analysis and multivariable regression modelling. We tested the null hypothesis that operative time does not change with increasing case sequence after adjustment for patient-, tumour- and country-related factors. Secondary objectives were to estimate the number of cases required to achieve proficiency and to compare perioperative outcomes across learning phases and between participating countries. Methods: This multicenter retrospective cohort study was conducted at five academic and tertiary-care centers in Egypt and Saudi Arabia between January 2023 and December 2025. Seventy-seven consecutive patients undergoing VATS excision of mediastinal masses were categorized into three phases: Early (first 20 cases), Intermediate (next 20 cases), and Late (last 37 cases). Cumulative sum (CUSUM) analysis was performed. Operative outcomes, including operative time, blood loss, conversion rate, and postoperative complications, were compared across phases and between countries (Egypt vs. Saudi Arabia). Multivariable linear regression and a linear mixed-effects model with participating centre as a random intercept were used to identify factors independently associated with operative time. Results: Mean operative time significantly decreased across phases (189.5 ± 21.9 vs. 156.5 ± 26.4 vs. 139.7 ± 30.5 minutes; p < 0.001). CUSUM analysis demonstrated a learning curve inflection point between cases 20 and 26. After adjustment for age, sex, body mass index, myasthenia gravis, presence of a discrete mass, thymic neoplasm and country, case sequence number remained independently associated with shorter operative time (β = -1.14 min per case; 95% CI -1.47 to -0.80; p < 0.001; R² = 0.48, adjusted R² = 0.42), with consistent estimates using robust standard errors and a mixed-effects model with centre as a random intercept (β = -1.37; 95% CI -1.80 to -0.95). Patients treated in Saudi Arabia (n = 7) were older and more often had discrete neoplastic masses, but operative time did not differ significantly between countries before (p = 0.128) or after adjustment (p = 0.548). Conversion to open surgery occurred in 3.9% (n = 3), all in the Intermediate and Late phases. Overall mortality was 2.6% (n = 2), both in the Late phase and associated with complex cases. No significant differences were observed regarding ICU stay, hospital stay, or postoperative complications. Conclusion: VATS mediastinal mass excision is safe and feasible, with surgical proficiency achieved after approximately 20-26 cases. Operative time was the most reliable indicator of learning progression, and the learning effect persisted after multivariable adjustment for patient-, tumour- and country-related factors. Structured mentorship and careful case selection are recommended during the early learning phase.