Assessment of laparotomy conversion, unplanned readmission, and perioperative mortality for minimally invasive surgery in endometrial cancer.

Abstract

Conversion to laparotomy during minimally invasive hysterectomy for endometrial cancer has been an area of active research interest. Such data are scarce within a more contemporary study period. The primary objective was to compare laparotomy conversion rates between robotic-assisted and conventional laparoscopic hysterectomy for endometrial cancer; the secondary objective was to assess unplanned readmissions and perioperative mortality per surgical modality. This comparative effectiveness study queried the Commission-on-Cancer's National Cancer Database in the United States. The study population included 210,615 patients with stage I to III endometrial cancer who underwent upfront minimally invasive hysterectomy from 2012 to 2023. Exposure was intent-level surgical modality, comparing robotic-assisted and conventional laparoscopic hysterectomy (n=164,137 and n=46,478, respectively). The primary outcome measure was laparotomy conversion, defined as open surgery at the per-procedure level; the secondary outcome measures included unplanned postdischarge readmission and perioperative mortality within 30 days of the index anticancer surgery. Temporal trends were assessed with linear segment regression model. The exposure-outcome association was assessed in propensity score inverse probability of treatment weighting cohort to reduce the differences in baseline clinico-pathologic characteristics between the 2 surgical modality groups, created with a multivariable generalized linear model with Poisson distribution. A classification tree was constructed to visualize the patterns of clinico-pathologic characteristics associated with laparotomy conversion. Over the 12-year study period, laparotomy conversion rates decreased from 2.1% to 1.0% in the robotic-assisted hysterectomy group (average annual percentage change, -5.1%; 95% confidence interval, -7.6% to -3.1%) and from 9.0% to 4.8% in the conventional laparoscopic hysterectomy group (average annual percentage change, -6.1%; 95% confidence interval, -7.3% to -4.9%). Larger tumor size was associated with a higher laparotomy conversion rate for both surgical modalities: inflection points for increasing laparotomy conversion rates were 5 cm and 4 cm for robotic-assisted and conventional laparoscopic surgery, respectively. In the propensity score weighted model, robotic-assisted hysterectomy was associated with a 78% lower laparotomy conversion rate compared to conventional laparoscopic hysterectomy (1.5% vs 6.3%; incidence rate ratio, 0.22; 95% confidence interval, 0.21 to 0.24). Unplanned readmission (16.9 vs 19.0 per 1000; incidence rate ratio, 0.89; 95% confidence interval, 0.83 to 0.96) and perioperative mortality (2.0 vs 2.6 per 1000; incidence rate ratio, 0.80; 95% confidence interval, 0.65 to 0.99) rates were also lower for robotic-assisted hysterectomy compared to conventional laparoscopic hysterectomy. A classification tree identified 20 unique clinico-pathologic patterns associated with laparotomy conversion, of which 3 patterns exceeded 10%. All 3 patterns had conventional laparoscopic hysterectomy for tumor size of 6 cm or greater, whereas robotic-assisted surgery for tumor size of smaller than 4 cm had a rate of 1% or lower. The results of this study suggest that among the Commission-on-Cancer-affiliated hospitals in the United States, laparotomy conversion during minimally invasive hysterectomy for endometrial cancer has decreased. While limited by potential confounding by surgeon- or patient-related factors, these data suggest that surgeons who preferentially perform robotic-assisted hysterectomy may have lower associated rates of laparotomy conversion, unplanned postdischarge readmission, and perioperative mortality compared to those who perform conventional laparoscopic hysterectomy. Whether perioperative morbidity differs according to a surgeon's preferred practice for minimally invasive surgery warrants further evaluation.