Impact of postoperative morbidity on outcomes in patients with advanced epithelial ovarian cancer undergoing intestinal surgery at the time of primary or interval cytoreductive surgery: A Memorial Sloan Kettering Cancer Center Team Ovary study.

A retrospective chart review was performed for patients with ovarian cancer who underwent PCS or ICS with at least one intestinal resection at our institution from 1/1/2015 to 12/31/2020. Patients who underwent surgery for recurrent disease or palliative surgery for malignant bowel obstructions at the time of recurrence were excluded. Appendectomy alone or superficial excision of serosal lesions were not considered intestinal resection. Patient demographics, clinical characteristics, short-term postoperative complications within 30 days of surgery, and time to initiation of postoperative chemotherapy were noted. Short-term complications (surgical secondary event [SSE]) within 30 days of surgery were graded by a validated SSE system [ 23 ]. The Memorial Sloan Kettering Cancer Center (MSK) SSE grading system is based on the established Clavien-Dindo classification system, but further classifies over 220 distinct SSEs into 14 physiologic categories [ 23 , 24 ]. Grade 1 events require bedside care or oral medications; grade 2 require intravenous medications or transfusion; grade 3 require radiologic, endoscopic, or operative intervention; grade 4 result in chronic disability or organ resection; and grade 5 is death [ 23 ]. Patients were categorized by no SSE, grade 1 and 2 SSE, or grade 3 and 4 SSE. AL was defined as a complication requiring reoperation for suspected leak, with confirmed leakage of bowel contents per operative findings. Time to initiation of standard postoperative chemotherapy was calculated as the number of days from surgery to the first day of the first cycle of chemotherapy. Patients were separated by surgical group—PCS or ICS. These two groups were analyzed separately with regard to oncologic outcomes. Patient clinical features and treatments were compared between those who experienced any SSE and those who did not using the Fisher exact test for categorical variables and Wilcoxon-rank sum test for continuous variables. Progression-free survival (PFS) was defined as time from date of surgery to date of progression or date of death or last follow-up. Overall survival (OS) was defined as time from date of surgery to date of death or last follow-up. The Kaplan-Meier method was applied to estimate the median survival time and the survival rate at 3 years. P values for survival outcomes were obtained using log-rank tests for categorical variables and Wald tests based on CoxPH (proportional hazard) models for continuous variables. Landmark analyses, with 30 days as landmark time, were applied to examine the effect of time-dependent variables such as SSE, transfusion etc. on survival outcome. All variables that showed significance on univariate analysis were entered into the multivariable model with SSE. All analyses were performed with R 4.2.2 ( https://www.R-project.org/ ). All p values were 2-sided, and a p value <0.05 was considered significant.

Demographic, clinical, and intraoperative characteristics, as well as short-term postoperative complications are presented in Tables 1 and 2 for the PCS and ICS groups, respectively. Of 437 patients included in the analyses, 289 underwent intestinal resection during PCS and 148 during ICS. In the entire cohort, 6 patients (1.4%) were diagnosed with an anastomotic leak within 30 days of surgery—5 (1.7%) in the PCS group and 1 (0.7%) in the ICS group. Overall, 85 patients (19%) presented to the emergency room within 30 days of surgery, 70 (16%) required an interventional radiology procedure, 16 (3.7%) required re-operation, and 1 (0.2%) was admitted to the intensive care unit. Overall, 291 patients (67%) experienced at least one SSE (grade 1, 35 [8%]; grade 2, 203 [46%]; grade 3, 52 [12%]; and grade 4, 1 [0.2%]). There were no perioperative deaths (grade 5 complications). The PCS and ICS cohorts had similar use of bevacizumab or a poly (ADP-ribose) polymerase inhibitor (PARPi) postoperatively. Of the 289 patients in the PCS group, 110 (38%) underwent one, 124 (43%) two, and 55 (19%) three intestinal resections ( Table 1 ). Of the 55 patients who underwent three intestinal resections, 46 (84%) experienced a postoperative SSE and 9 (16%) did not (p=0.014). Most anastomoses were performed using a stapling technique (98%); in 5 patients (1.9%), a hand-sewn anastomosis was performed. An ostomy was created in 31 patients (11%) and reversed in 26 of them (84%) within a median of 6.22 months (range, 2.4–32.7 months) after surgery. The median number of days to start of postoperative chemotherapy was 36 (range, 13–110 days). There was a statistically significant longer time to initiation of postoperative chemotherapy in patients with any SSE versus no SSE (37 vs. 34 days, respectively; p=0.032). When comparing patients initiating postoperative chemotherapy within 42 days to those initiating chemotherapy after 42 days, there was no difference between those with no SSE vs. any SSE (p=0.15; Supplementary Table 1 ). SSE was associated with longer operating room time (p=0.032), higher estimated blood loss (p<0.001), greater intraoperative crystalloid (p=0.004) and albumin use (p=0.008), transfusion of packed red blood cells (p<0.001), and length of stay (p<0.001). CGR was achieved in 222 patients (77%), with no difference between those with any SSE versus no SSE (p=0.88). Eight patients (2.8%) had a suboptimal cytoreduction (postoperative residual disease >1 cm). Within 30 days of surgery, 62 (21%) of these 289 patients were seen in the urgent care center (UCC), 51 (18%) needed an interventional radiology (IR) procedure, 33 (11%) were re-admitted, and 13 (4.5%) needed re-operation. Twenty-six patients (9.0%) experienced a grade 1, 129 (45%) a grade 2, 43 (15%) a grade 3, and 1 (0.3%) a grade 4 SSE. One patient (0.3%) experienced a postoperative bowel perforation within 30 days of surgery. Five patients (1.7%) experienced an anastomotic leak within 30 days of surgery. Table 1 further details postoperative morbidity in the PCS cohort between those who experienced any SSE and those who did not. On univariable analysis, there was no difference in median PFS for patients with no SSE vs. grade 1/2 SSE vs. grade 3/4 SSE (p=0.672; Supplemental Table 2 , Figure 1A ). On multivariable analysis, any residual disease was associated with worse PFS (HR, 1.39; 95% CI: 1.00, 1.93; p=0.047) and upper abdominal surgery was associated with worse outcomes (HR, 1.80; 95% CI: 1.17, 2.77; p=0.008). One unit increase in preoperative albumin was associated with improved PFS (HR, 0.67; 95% CI: 0.48, 0.94; p=0.019). There was no difference in PFS between patients who experienced an SSE and those who did not within 30 days of surgery (HR, 1.05; 95% CI: 0.75, 1.47; (p=0.75; Table 3 ). Also, there was no difference in PFS between patients who initiated chemotherapy within vs. after 42 days of surgery (p=0.804; Supplementary Table 3 ). On univariable analysis, there was no difference in median OS for patients who underwent PCS with no SSE vs. grade 1/2 SSE vs. grade 3/4 SSE (p=0.685; Supplemental Table 4 , Figure 1B ). On multivariable analysis for OS, one unit increase in preoperative albumin was associated with improved OS (HR, 0.62; 95% CI: 0.39, 0.98; p=0.043). There was no difference in OS for patients who underwent PCS with any SSE vs. no SSE within 30 days of surgery (HR, 0.79; 95% CI: 0.49, 1.26; p=0.32; Table 4 ). Also, there was no difference in OS between patients who initiated chemotherapy within vs. after 42 days of surgery (p=0.678, Supplementary Table 4 ). Of the 148 patients in the ICS group, 73 (49%) underwent one, 56 (38%) two, and 19 (13%) three intestinal resections ( Table 2 ). There was no difference in rate of any SSE for patients who underwent ICS with one, two, or three intestinal resections (p=0.87). Most anastomoses were performed using the stapling technique (97%); in 3 patients, a hand-sewn anastomosis was performed (2.9%). An ostomy was created in 18 patients (12%) and reversed in 13 of them (72%) within a median of 4.41 months (range, 1.61–8.98 months) after surgery. The median number of days to re-start postoperative chemotherapy was 31 (range, 14–91 days). There was a statistically significant longer time to initiation of postoperative chemotherapy in patients with any SSE versus no SSE (34 vs. 29 days, respectively; p=0.005). When comparing patients initiating postoperative chemotherapy within 42 vs. after 42 days of surgery, there were more patients with any SSE starting chemotherapy 42 days after ICS (p=0.047; Supplementary Table 3 ). Any SSE was associated with longer operating room time (p=0.042), estimated blood loss (p=0.009), greater intraoperative albumin use (p<0.001), transfusion of packed red blood cells (p<0.001), and length of stay (p<0.001). There was a statistically significant greater number of median preoperative cycles of chemotherapy in the any SSE versus no SSE groups (4 vs. 3, respectively; p=0.022). A CGR was achieved in 97 patients (66%), with no difference between those with no SSE versus any SSE (p=0.21). Twelve patients (8%) underwent suboptimal cytoreduction. Within 30 days of surgery, 23 patients (16%) were seen in the UCC, 19 (13%) underwent an IR procedure, 11 (7.4%) were re-admitted, 3 (2.0%) required re-operation, and no patients were admitted to the intensive care unit. Nine patients (6.1%) experienced a grade 1, 74 (50%) a grade 2, and 9 (6.1%) a grade 3 SSE. No patient experienced a grade 4 SSE. Two patients (1.4%) experienced a small bowel obstruction and 1 (0.7%) experienced a postoperative bowel perforation. Anastomotic leak within 30 days of ICS was diagnosed in 1 patient (0.7%). Table 2 further details postoperative morbidity in the ICS cohort between those who experienced any SSE and those who did not. On univariable analysis, there was no difference in median PFS for patients who underwent ICS with no SSE vs. grade 1/2 vs. grade 3/4 SSE (p=0.445; Supplemental Table 1 , Figure 1C ). On multivariable analysis for PFS, any residual disease after surgery was associated with worse outcomes (HR, 1.74; 95% CI: 1.19, 2.54; p=0.004). There was no difference in PFS between patients who experienced an SSE and those who did not within 30 days of surgery (HR, 1.43; 95% CI: 0.99, 2.07; p=0.055; Table 3 ). Also, there was no difference in PFS between patients who initiated chemotherapy within vs. after 42 days of surgery (p=0.896; Supplementary Table 4 ). On univariable analysis, there was no difference in median OS for patients who underwent ICS with no SSE vs. grade 1/2 vs. grade 3/4 SSE (p=0.645; Supplemental Table 2 , Figure 1D ). On multivariable analysis for OS, non-white race was associated with worse outcomes (HR, 2.13; 95% CI: 1.20, 3.79; p=0.01). Positive BRCA mutation status was associated with improved OS (HR, 0.16; 95% CI: 0.05, 0.50; p=0.019). There was no difference in OS between patients who experienced an SSE and those who did not within 30 days of surgery (HR, 1.18; 95% CI: 0.72, 1.93; p=0.52; Table 4 ). Also, there was no difference in OS between patients who initiated chemotherapy within vs. after 42 days of surgery (p=0.289; Supplementary Table 4 ).

In this large retrospective analysis, short-term postoperative complications for patients who underwent intestinal surgery during primary surgical management for advanced ovarian cancer did not impact oncologic outcomes. Overall, grade 3 and higher SSEs occurred in 12% of patients within 30 days of PCS or ICS that included intestinal resection. Intestinal-related complications within 30 days of PCS or ICS were <10% in our study. The most common complication within 30 days of surgery was ileus (9.3% in the PCS cohort and 6.8% in the ICS cohort). The anastomotic leak rate was 1.7% in the PCS cohort and 0.7% in the ICS cohort—6 patients in total. These data are comparable and even compare favorably to other large studies assessing short-term postoperative morbidity and complications after PCS and ICS [ 9 , 10 , 22 , 25 – 27 ]. For example, Grimm et al. reported a 6.9% anastomotic leak rate among patients who underwent bowel resection during debulking surgery. In a meta-analysis of 13 studies, Valenti et al. reported that pre-albumin level <3.0 g/dL, multiple bowel resections, and PCS were associated with a significantly higher risk of anastomotic leak (OR=5.29 [95% CI: 1.51, 18.59]; OR=4.4 [95% CI: 1.19, 16.66]; and OR=1.71 [95% CI: 1.05, 2.77], respectively) [ 22 ]. In our study, we did not find a difference in the univariable PFS and OS analyses with regard to number of intestinal resections (one, two, or three). Higher preoperative albumin was significantly associated with improved PFS and OS on multivariable analyses for the PCS group. Patient selection for PCS and ICS remains heavily debated, but our results continue to affirm that patients with higher preoperative albumin levels demonstrate better survival outcomes [ 28 – 32 ]. Of note, 13 patients with stage I/II disease underwent intestinal resection at the time of PCS; their tumors were largely of endometrioid or clear cell histology, endometriosis-associated cancers, with intestinal involvement. Beyond anastomotic leak rates, our analysis provides helpful information regarding ostomy creation during PCS and ICS. We report ostomy rates of 11% in the PCS cohort and 12% in the ICS cohort, with reversal in 84% at a median of 6.2 months and 72% at a median of 4.4 months, respectively. Gockley et al. demonstrated slightly lower rates of ostomy creation after PCS and ICS, but with similar rates between the cohorts [ 26 ]. In a prospective quality-improvement project for patients who underwent rectosigmoid resection during debulking surgery, Kalogera et al. found a non-significant decrease in anastomotic leak rate (7.8% vs. 2.6%) compared to a historical cohort when adding a temporary diverting ostomy based on pre- and intraoperative risk factors. The diverting ostomy rate increased from 8% in the historical cohort to 35% in the quality-improvement cohort, and ostomies were reversed in 88% of patients in the latter cohort [ 18 ]. Additionally, Costantini et al. performed a retrospective analysis of 515 patients who underwent colorectal resection and anastomosis during PCS or ICS, with a protective ileostomy rate of 45% and no association between protective ostomy creation and anastomotic leak rates postoperatively [ 19 ]. Our data, along with data from these prior analyses, provide key information for counseling patients with advanced-stage ovarian cancer undergoing either PCS or ICS, specifically regarding rate of ostomy formation, rate of reversal, and median time with ostomy. Moreover, in line with our results, it is difficult to determine specific pre- and intraoperative risk factors for anastomotic leak that would indicate the use of a diverting ostomy [ 17 , 18 ]. Kalogera et al. reported on a standardized protocol for temporary bowel diversion after rectosigmoid resection: patients with preoperative albumin <3.0g/dL, prior pelvic radiation, rectosigmoid resection plus any additional bowel resection, anastomosis <6 cm from anal verge, or failed leak test or contamination of the pelvis with stool would undergo temporary diversion [ 18 ]. While they report these criteria are helpful for temporary bowel diversion, it is important to continue to mitigate the consequences of anastomotic leaks while also trying to avoid unnecessary ostomies, which are associated with complications and reduced quality of life. Prior large retrospective studies have attempted to assess the association of postoperative complications and morbidity after PCS and ICS with oncologic outcomes [ 8 , 33 – 37 ]. Grimm et al. demonstrated an independent association of anastomotic leak with shortened OS (HR, 1.9; 95% CI: 1.2, 3.4; p=0.01) [ 10 ]. Their analysis did not demonstrate any predictive pre- and/or intraoperative risk factors associated with anastomotic leak. Similarly, Bartl et al. reported a low rate of anastomotic leak after debulking surgery for advanced-stage ovarian cancer, with associated worse OS on univariate (p=0.04) but not multivariable analyses [ 25 ]. In our analysis, the anastomotic leak rate was only 1.7% (n=6), and comparisons to patients without anastomotic leak were not statistically meaningful. However, a descriptive analysis of those 6 patients reveals that one patient was not able to initiate postoperative chemotherapy and died 32 days after surgery. And the remaining 5 patients started postoperative chemotherapy after 40 days, highlighting the severity of anastomotic leak and the potential impact on survival. In our analysis, short-term postoperative complications did not result in worse PFS and OS when controlling for known prognostic factors. Very few patients were left with residual disease >1 cm. This is likely the result of patient selection. Intestinal resections are typically not performed if a CGR or near-complete resection of visible disease is not possible. As such, our data demonstrate a favorable median PFS of 24 months, even in patients with high-volume disease requiring upper abdominal surgery. Time to initiation of postoperative chemotherapy has also been linked to oncologic outcomes in patients undergoing PCS or ICS [ 38 , 39 ]. Our results demonstrate a statistically significant median delay of 3 and 5 days for patients in the PCS and ICS groups, respectively, between patients who experienced an SSE and those who did not ( Tables 1 and 2 ). Even though statistically significant, these results are likely not clinically meaningful since the delay in initiation of chemotherapy is less than 1 week in both groups. Moreover, our exploratory univariate analysis ( Supplementary Table 4 ) demonstrates no difference in PFS and OS for patients initiating chemotherapy within vs. after 42 days of PCS or ICS. Prior studies have demonstrated that delays in initiation of postoperative chemotherapy >6 weeks impact oncologic outcomes [ 39 ]. As stated above, patients who experience anastomotic leak are at highest risk for significant delays in time to initiation of postoperative chemotherapy. Additionally, Castro et al. demonstrated independent factors associated with adjuvant chemotherapy delay, including hypertension, body mass index greater than 30 kg/m 2 , and re-operation and fever within 30 days of surgery [ 40 ]. As such, in line with prior analyses, our results demonstrate that adverse events after PCS or ICS with intestinal resection may delay time to postoperative chemotherapy but are unlikely to impact survival outcomes. This analysis is strengthened by the large cohort of patients who underwent PCS or ICS with at least one intestinal resection. The granular data included in the analysis allows for greater understanding regarding postoperative morbidity related to these surgeries. The analysis is limited by the retrospective nature of the project and data collection. Also, while the survival outcomes differ between the PCS and ICS cohorts, there was no direct comparison of the PCS and ICS groups. Both groups used date of surgery as the starting date, since the focus of the analysis was postoperative morbidity within 30 days of surgery. Furthermore, we assessed only patients who had surgery at our institution, which may limit the generalizability of our findings. Grade 3 or higher SSEs occurred in 12% of patients within 30 days of PCS or ICS that included intestinal resection. Short-term postoperative complications for patients who underwent intestinal surgery during primary surgical management for advanced ovarian cancer did not impact oncologic outcomes. The anastomotic leak rate in our study was very low, and a comparison with survival was not statistically meaningful. A potential negative impact of anastomotic leak on survival outcomes cannot be ruled out. In conclusion, comprehensive surgical cytoreduction that includes intestinal resection is safe and should be considered for patients with advanced-stage ovarian cancer when a CGR or near-complete cytoreduction is feasible.

In 2023, an estimated 19,710 women will be diagnosed with ovarian cancer and 13,270 patients will die from this disease in the United States [ 1 ]. Due to a lack of effective screening modalities, patients often present with advanced stage at the time of diagnosis. The upfront management of advanced-stage disease involves cytoreductive surgery and chemotherapy [ 2 – 4 ]. Primary cytoreductive surgery (PCS) and interval cytoreductive surgery (ICS) are often complex and lengthy and can involve multiple organ systems and bowel resections [ 5 – 12 ]. Intestinal resections are frequently necessary to achieve a complete gross resection (CGR) or near-complete resection of visible disease at the time of PCS [ 13 ]. Although less often needed after neoadjuvant chemotherapy, intestinal resections are also performed at the time of ICS [ 14 – 16 ]. Anastomotic leak (AL) is the most concerning and morbid complication of intestinal resection [ 10 , 12 , 13 ]. A retrospective study by Grimm et al. demonstrated an independent association of AL with worse survival outcomes [ 10 ]. In theory, the morbidity and complications resulting from AL could delay initiation of postoperative chemotherapy and portend worse oncologic outcomes. There are limited studies regarding the impact of postoperative morbidity on patients with advanced ovarian cancer undergoing PCS or ICS with intestinal surgery [ 11 , 17 – 22 ]. Our objective was to assess the impact of short-term postoperative complications and morbidity on oncologic outcomes for patients with ovarian cancer undergoing PCS or ICS with at least one intestinal resection.