Patient characteristics
In total 595 endometrial cancer cases were analysed, of which 294 (49.4%) were a training case and 301 (50.6%) were not. 502 cases (84.8%) were performed at the Royal Marsden Hospital and 92 cases (15.5%) were performed at other hospitals within the Southwest Thames Gynaecological Cancer Centre. Eighteen gynaecological oncology trainees were trained in RAL surgery for endometrial cancer in a PBP manner.
Table 1 shows the baseline characteristics for training and non-training cases. Training and non-training cases were similar in age (66 versus 67 years, p = 0.0946), ASA score (2 versus 2, p = 0.680) and Carlson Comorbidity Index 10-year median survival estimates (21.6% versus 21.4%, p = 0.259). However, training cases had a lower BMI (30 versus 32 kg/m2, p = 0.016) than non-training cases and there was a difference in distribution of WHO performance status (p = 0.026), but post-hoc testing according to Bonferroni24rendered no significant differences.
No differences were found in median FIGO stage between training and non-training cases (1 versus 1, p = 0.168), but training cases had patients with a higher median histopathological grade (3 versus 2, p = 0.010) and a lower rate of endometrioid tumours (56.8% versus 69.1%, p = 0.002*). Post-hoc testing showed a lower rate of grade 1 (33.9% versus 45.6%, p = 0.004*) and a higher rate of grade 3 tumours (50.3% versus 39.3%, p = 0.007*) in training cases. This resulted in a higher percentage of adjuvant treatment in training cases (69.9% versus 59.6%, p = 0.034). The distribution of ESGO risk scores did not differ between training and non-training cases (p = 0.078). More sentinel lymph node dissections (77.2% versus 67.8%, p = 0.010) were performed in training cases, which is possibly associated with the gradual increase in training cases over time (37.2% in 2015 versus 55.3% in 2022) accompanied with the simultaneous increase in sentinel lymph node procedures (39.5% in 2015 versus 72.4% in 2022). The median number of harvested lymph nodes (3 versus 3, p = 0.325), and rates of pelvic (25.3% versus 29.2%, p = 0.276) and para-aortic lymphadenectomies (0.3% versus 1.3%, p = 0.188) were comparable. Median follow-up in months was comparable (25 versus 28 months, p = 0.140) between training and non-training cases.