Feed

Background and Objectives: Previous surgical approach comparisons in endometrial cancer may be confounded by inadequate control for tumour biology-the primary driver of outcomes. This study provides the first surgical approach comparison incorporating molecular classification to control for case selection bias. Materials and Methods: Retrospective analysis of 512 consecutive patients with stage I-II endometrial cancer (FIGO 2009) treated with open (n = 83), laparoscopic (n = 278), or robotic (n = 151) surgery between 2018 and 2024. Molecular classification was available for 219 patients (42.8%) using TCGA criteria and incorporated into analyses to control for case selection bias, with molecular subtype incorporated to control for biological bias rather than as a primary endpoint. Primary outcomes included perioperative metrics and oncological safety. The primary objective was to determine whether apparent surgical outcome differences reflect genuine technique effects or case selection bias based on tumour biology. Results: Molecular subtype distribution varied significantly by surgical approach, with high-risk subtypes concentrated in open surgery, explaining apparent outcome differences. After controlling for molecular subtype and other confounders, minimally invasive approaches demonstrated superior perioperative outcomes: reduced blood loss (laparoscopic 129.8 mL, robotic 157.9 mL vs. open 261.4 mL, p < 0.001), shorter hospital stays (2.4 and 2.2 vs. 5.3 days, p < 0.001), and lower complications (5.7% and 6.6% vs. 21.6%, p < 0.001). In our cohort, recurrence-free survival showed significant differences favouring minimally invasive approaches, with 2-year RFS rates of 92.8%, 96.4%, and 100.0% (p = 0.008) and 3-year RFS rates of 90.4%, 95.0%, and 100.0% (p = 0.003) for open, laparoscopic, and robotic surgery, respectively, although robotic surgery had a shorter follow-up (median 33 vs. 42 months). Within-approach exploratory analyses revealed that p53-abnormal tumours were associated with significantly longer operative times and greater blood loss across all surgical approaches (p < 0.05), although complication rates did not differ significantly by molecular subtype within any approach (open p = 0.124, laparoscopic p = 0.656, robotic p = 0.287). Apparent surgical approach differences largely reflected appropriate case selection based on tumour biology rather than technique superiority. Conclusions: When controlling for tumour biology, minimally invasive approaches offer superior perioperative outcomes with equivalent oncological safety. Higher complication rates in open surgery primarily reflect the inherent morbidity of this approach and appropriate surgeon selection for high-risk cases. Within-approach analyses suggest possible molecular influences on operative parameters that warrant prospective validation. Molecular stratification is essential for fair surgical approach comparison in the contemporary era.