(P 327) LEARNING CURVE ANALYSIS IN RETROPERITONEAL SARCOMA SURGERY

Objective: In retroperitoneal sarcoma (RPS) surgeons undergo a learning process to acquire the necessary skills for optimal patient outcomes. This study aimed to assess the relationship between surgical experience and operative outcomes and identify distinct phases and factors that influence the learning process in RPS surgery.

Methods: A retrospective analysis was conducted from a prospectively collected database. We analyzed 62 consecutive RPS surgeries performed by the same surgeon (Jan 2014-Jan 2023) at IRCCS Humanitas Research Hospital (Rozzano, Milan). We used the cumulative sum (CUSUM) analysis to assess the learning curve and identify significant changes in surgical performance over time.

Results: The mean age of patients was 61.5 years (range 24-83). Male:female ratio was 27:35. Mean size of tumors was 174 (range 0-617, where we considered as 0 if no macroscopic residual disease was detected at histology). Forty-five cases (72.6%) were primary tumors or residual after incomplete resection, whereas 17 cases (27.4%) were local recurrences. Tumors were located in the left retroperitoneum (RP) in 24 patients (38.7%), in the right RP in 26 patients (42.0%), and in pelvis in 12 patients (19.3%). Histologic subtypes were well-differentiated liposarcoma (WDLPS,19.4%), dedifferentiated liposarcoma (DDLPS,53.2%), leiomyosarcoma (LMS,17.7%), other (9.7%). Forty-four cases (71.0%) were high-grade tumors (G2-G3). Neoadjuvant chemotherapy was performed in 18 cases (29.9%), while neoadjuvant radiotherapy in 10 cases (16.1%). The procedures included 47 (75.8%) multivisceral resections. Pattern of resection can be seen in Table 1. The mean operative time was 358 minutes (range 108-838), with a mean intraoperative blood loss of 590 ml (range 50-4000). The operative time (OT) learning curve was analyzed (Image 1B). The learning curve consisted of three unique phases: phase 1 (the 16 initial cases), phase 2 (the middle 30 cases), and phase 3 (the subsequent 16 cases). Phase 1 was characterized by a negative slope, which represented shorter operating time during this phase. Phase 2 was characterized by a plateau phase. Phase 3 was achieved after 46 cases and showed a positive slope that indicates longer operation times. Comparisons of various parameters between the three phases are presented in Table 2. Statistically significant difference between phase 1 and 2 combined and phase 3 was achieved by size (p=0.003), presentation (p=0.048), pattern of resection (colon+kidney+other, p=0.033), operation time (p=0.006), and length of stay (LOS) (p=0.026).

Conclusion: This study aimed to fill a critical knowledge gap by examining the learning curve in RPS surgery. The findings will contribute to enhance surgical training, optimizing patient outcomes, and to improve decision-making in this challenging surgical field. We selected CUSUM analysis due to the limitations of interpreting raw data plotted chronologically. The CUSUM graph provides insights into the variance from the mean on a case-by-case basis, resulting in a parabolic curve that exhibits three distinct phases. These phases enable the assessment of learning curve correlates. Phase 1 represented the initial learning curve. Phases 2 and 3 showed an increase in operative times that reflect the performance of more complex cases as surgeon competence increases. Understanding the learning curve in RPS surgery will help guide surgical education and refine surgical techniques, ultimately leading to improved patient care and long-term oncological outcomes.