Abstract
Objective:
To independently validate a published risk-calculator for adverse perioperative outcomes in patients with epithelial ovarian cancer undergoing debulking surgery at a high-volume surgical center.
Methods:
Using our institution’s curated prospective ovarian cancer database, we identified patients with epithelial ovarian cancer who underwent a debulking procedure from 7/2015 to 5/2019, to be used as the validation cohort. Variables used in the published nomogram were collected. These included American Society of Anesthesiology classification, preoperative albumin, history of bleeding disorder, presence of ascites on preoperative imaging, designation of elective or emergent surgery, age of the patient, and a procedure score. Patients were included if they had information available for all the variables used in the nomogram, and 30-day follow-up within our institution. The primary outcome was Clavien-Dindo Class IV with specific conditions (postoperative sepsis, septic shock, cardiac arrest, myocardial infarction, pulmonary embolism, ventilation >48 hours, or unplanned intubation) and 30-day mortality. The combination of these endpoints is called the combined complication rate.
Results:
A total of 700 patients who underwent debulking surgery for epithelial ovarian cancer during the timeframe met inclusion criteria. The combined complication rate was 11.7%; 9.9% of patients were readmitted; 2.7% required reoperation. Sepsis was the most common primary endpoint complication (4.4%), followed by septic shock (1.4%). There was no 30-day mortality in our cohort. The nomogram performed well, with a c index of 0.715 (95% CI 0.66–0.768), which was comparable to the published nomogram.
Conclusions:
We independently validated a complication nomogram at a high-volume surgical center. This nomogram performs well at predicting a lower likelihood of serious postoperative complications. An enhanced nomogram would help identify patients at higher risk for serious complications.
Keywords: Perioperative complications; Cytoreduction, surgical procedures; Ovarian cancer
INTRODUCTION
Management of advanced epithelial ovarian cancer includes a combination of cytoreductive surgery and platinum- and taxane-based chemotherapy. The goal of surgery is to remove as much visible disease as possible, as complete resection of all gross tumor confers a survival benefit1. Complex cytoreductive surgery is often necessary to achieve this goal. This is especially true for disease in the upper abdomen where the patient may require diaphragm peritonectomy, splenectomy, distal pancreatectomy, and dissection of the porta hepatis. Increased radicality is associated with greater perioperative morbidity and mortality 2–4.
Postoperative complications are associated with delay in initiation of chemotherapy, but a delay greater than 35 days is associated with decreased survival5. To balance the survival benefit of complete surgical resection with its associated morbidity, there has been increasing interest in developing predictive models for postoperative complications to help tailor selection of treatment. Previously published studies investigating factors associated with perioperative complications incorporate patient factors such as increasing age, American Society of Anesthesiology (ASA) classification score, surgical complexity score, presence of ascites, preoperative albumin, body mass index (BMI), and CA-125 level6,7. Other factors have been explored, including volume of the surgical center where the procedure is performed, or timing of the procedure8,9 (i.e. whether it was scheduled as the first case, or later in the day). These markers have been investigated in single-institution studies, as well as studies based on national databases.
Cham and colleagues created a nomogram to predict 30-day postoperative complications in women with ovarian cancer who underwent oophorectomy and additional procedures, using data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) national database. The final published nomogram incorporated ASA, age, history of bleeding disorder, preoperative albumin, presence of ascites, undergoing emergent surgery, and a procedure score10, and had an internal discrimination concordance index (c-index) of 0.71. The NSQIP database combines information from medical charts of patients treated at hospitals of varying volume across the United States, from community hospitals to specialized referral centers11. The primary objective of this study was to determine if the published nomogram is applicable to a high-volume specialized cancer center in women undergoing cytoreduction for ovarian cancer.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board at Memorial Sloan Kettering Cancer Center (IRB 18–316). In July of 2015, our institution developed a curated prospective ovarian cancer database. Included in the database are all patients who present to our institution with a chief complaint of a suspected ovarian mass or ovarian cancer; these patients are followed forward in their management. Data is recorded, curated and updated in the database at regular intervals.
For review, the population reported by Cham et al were patients who underwent surgery for primary ovarian, fallopian tube, or peritoneal cancer in the NSQIP database between 2011–2015. Specifically, women who underwent oophorectomy with or without hysterectomy were included. A procedure score was created for patients who underwent additional procedures at the time of cytoreduction. The surgical complexity score was generated by assigning 1 point for each of the following: lymph node dissection, small bowel resection, colon resection, rectosigmoid resection, liver resection, bladder resection, diaphragm surgery, or debulking. Patients were then classified with a score of 0, 1, 2, or ≥310.
Patients treated from July 2015 through May 2019 were identified in our institutional ovarian cancer database. Inclusion criteria were women with ovarian, primary fallopian tube or peritoneal cancers undergoing either primary debulking surgery or interval debulking surgery at our institution, age greater than 18, who had 30-day follow-up within our institution. Exclusion criteria were patients who did not have all variables required for validation of the nomogram by Cham et al, and patients who did not have 30-day follow-up within our institution. To best match the cohort used to create the nomogram, demographic characteristics, procedure scores, and complications were collected in the same manner as reported to ACS NSQIP (ACS NSQIP participant user files (PUF) were used for reference)12.
Our institution uses a preoperative algorithm to determine if a patient is not a candidate for primary debulking surgery. Incorporated in to the decision tree is the criteria described by Aletti et al whereby patients have to meet all three criteria ((1) age > 75 years old, 2) serum albumin < 3.5 g/dL, and 3) American Society of Anesthesiologists (ASA) classification score ≥ 3 or extensive disease) to proceed to NACT13. Additionally, when patients have specific medical conditions precluding them from surgery, such as pulmonary emboli, they proceed to NACT. Stage IVA disease, supradiaphragmatic lymphadenopathy and small subcapsular liver lesions are considered resectable, and these patients may proceed to PDS if medically cleared. We define unresectable Stage IVB disease as brain metastases, multifocal parenchymal liver metastases, bulky thoracic adenopathy (not including cardiophrenic lymphadenopathy), and pulmonary metastases including diffuse pleural involvement14. If a patient is deemed a surgical candidate, they are assigned a radiologic resectability score based upon their disease disruption15,16. This score is used to guide the surgeon in proceeding directly to PDS or triaging the patient to diagnostic laparoscopy14.
The primary outcome predicted by the nomogram was Clavien-Dindo Class IV complications and the following serious complications: postoperative sepsis, septic shock, cardiac arrest, myocardial infarction, pulmonary embolism, ventilation >48 hours, or unplanned intubation or death within 30 days after surgery (defined in the present study as a combined complication rate (CC rate)). Electronic medical records were reviewed for these serious complications as defined by ACS NSQIP PUF, and Clavien-Dindo IV was defined as intensive care unit admission17.
Descriptive statistics were provided. Logistic regression models were applied. Patients were excluded if nomogram-required variables were missing. The ability of a model to separate patients with different outcomes is known as discrimination. A nomogram’s discrimination is measured via a c-index. With a binary outcome, the c-index is identical to the area under the curve (AUC) for a Receiver Operating Characteristic (ROC) curve. A value for of 0.5 indicates random predictions, and a value of 1 indicates perfect prediction. A 95% confidence interval (CI) for the AUC was obtained through bootstrap resampling in the validation cohort with replacement 1000 times and take quantile value 0.025 to 0.975. How far a nomogram’s predictions are from the actual outcomes is referred to as calibration. A calibration curve was provided to show predicted nomogram probabilities versus the actual results18.
RESULTS
We identified 707 patients within the database, treated during the specified time frame, who presented with a new diagnosis of presumed advanced ovarian cancer and underwent debulking surgery. Five patients were excluded from the analysis because they did not have complete variables for the nomogram (the missing variable for each was preoperative albumin). Two additional patients were excluded due to lack of 30-day follow-up at our institution. The remaining 700 patients were included in the analysis.
Age was well represented in the cohort: 15% of women were younger than 50, 23% were 50–59 years, 35% were 60–69 years, and 27% were 70 years or older (Table 1). Most patients were white, had a normal or overweight BMI, underwent elective surgery, were non-smokers and non-diabetic. Most patients did not have ascites, were not anemic, and had a normal albumin preoperatively. In the cohort, 432 patients (62%) underwent primary debulking surgery and 268 patients (38%) underwent interval debulking surgery. Most patients (55%) had a procedure score of 3 or greater, with 49% undergoing a lymph node dissection and 39% undergoing a rectosigmoid resection. Most patients stayed three days or longer in the hospital.
Table 1.
Descriptive statistics of patients’ characteristics for the validation cohort and the cohort reported in Cham et al
| Validation cohort | Cham et al cohort10 | |
|---|---|---|
| N (%) | N (%) | |
| All | 700 (100) | 7029 (100) |
| Year of Operation | ||
| 2015/2011 | 72 (10) | 878 (12.5) |
| 2016/2012 | 187 (27) | 1082 (15.4) |
| 2017/2013 | 202 (29) | 1510 (21.5) |
| 2018/2014 | 188 (27) | 1606 (22.9) |
| 2019/2015 | 51 (7.3) | 1953 (27.8) |
| Age (in years) | ||
| <50 | 107 (15) | 1366 (19.4) |
| 50–59 | 159 (23) | 1864 (26.5) |
| 60–69 | 244 (35) | 2057 (29.3) |
| 70–79 | 149 (21) | 1291 (18.4) |
| >=80 | 41 (5.9) | 451 (6.4) |
| Race/ethnicity | ||
| White | 551 (79) | 5380 (76.5) |
| Black | 39 (5.6) | 445 (6.3) |
| Other | 88 (13) | 384 (5.5) |
| Unknown | 22 (3.1) | 820 (11.7) |
| Elective Surgery | ||
| Yes | 684 (98) | 6370 (90.6) |
| No | 16 (2.3) | 623 (9.0) |
| Unknown | n/a | 820 (11.7) |
| BMI | ||
| Normal (BMI<25) | 333 (48) | 2435 (34.6) |
| Overweight (BMI 25–30) | 198 (28) | 2056 (29.3) |
| Obese (BMI>=30) | 163 (23) | 2495 (35.5) |
| Unknown | 6 (0.9) | 43 (0.6) |
| Diabetes | ||
| Insulin | 8 (1.1) | 221 (3.1) |
| Non-insulin | 55 (7.9) | 555 (7.9) |
| No | 637 (91) | 6253 (89.0) |
| Tobacco use | 30 (4.3) | 928 (13.2) |
| COPD | 21 (3.0) | 190 (2.7) |
| Ascites | 220 (31) | 1323 (18.8) |
| CHF | 2 (0.3) | 19 (0.3) |
| Hypertension | 223 (32) | 2852 (40.6) |
| Bleeding Disorder | 4 (0.6) | 183 (2.6) |
| Albumin (g/dL) | ||
| <3.5 | 54 (7.7) | 1033 (14.7) |
| 3.5–4 | 296 (42) | 1902 (27.1) |
| >4 | 350 (50) | 1974 (28.1) |
| Unknown | n/a | 2120 (30.2) |
| Hematocrit (HCT) | ||
| <36% | 338 (48) | 2780 (39.6) |
| ≥36% | 362 (52) | 4073 (58.0) |
| Unknown | n/a | 176 (2.5) |
| ASA | ||
| ≤1 | 1 (0.1) | 204 (2.9) |
| 2 | 160 (23) | 2959 (41.1) |
| 3 | 527 (75) | 3595 (51.2) |
| ≥4 | 12 (1.7) | 271 (3.9) |
| Procedure Score | ||
| 0 | 16 (2.3) | 1586 (22.6) |
| 1 | 119 (17) | 3493 (49.7) |
| 2 | 181 (26) | 1618 (23.0) |
| ≥3 | 384 (55) | 332 (4.7) |
| Extended Procedures | ||
| LND | 344 (49) | 3047 (43.4) |
| Small Bowel Resection | 78 (11) | 217 (3.1) |
| Colon Reresection | 118 (17) | 272 (3.9) |
| Rectosigmoid Resection | 276 (39) | 475 (6.8) |
| Liver resection | 118 (17) | 123 (1.8) |
| Bladder resection | 21 (3.0) | 21 (0.3) |
| Diaphragm resection | 359 (51) | 154 (2.2) |
| Debulking | 680 (97) | 3503 (49.8) |
| Length of stay | ||
| 0 | 16 (2.3) | 110 (1.6) |
| 1 | 14 (2.0) | 498 (7.1) |
| 2 | 9 (1.3) | 619 (8.8) |
| ≥3 | 661 (94) | 5798 (82.5) |
| Discharge status | ||
| Home | 679 (97) | 6494 (92.4) |
| Dead | n/a | 40 (0.6) |
| Facility | 20 (2.9) | 480 (6.8) |
| Unknown | n/a | 15 (0.2) |
BMI, body mass index; COPD, chronic obstructive pulmonary disease; CHF, congestive heart failure; ASA, American Society of Anesthesiology; LND, lymph node dissection
Overall, 11.7% of patients experienced the primary endpoint of the combined complications, 9.9% were readmitted, and 2.7% required a reoperation (Table 2). In our cohort, we had no 30-day mortality. Sepsis was the most common primary endpoint complication at 4.4%, followed by septic shock at 1.4%. In the patients who underwent primary debulking surgery, the CC rate was 14% (n=61), while patients who underwent interval debulking surgery had a CC rate of 7.8% (n=21) (p = 0.02) (Table 3). The median procedure score for primary debulking surgery was 3 (range 0–7) and for interval debulking surgery was 2 (range 0–5) (Figure 1).
Table 2.
Morbidity and mortality for the validation cohort and rates reported in Cham et al
| Characteristic | Validation Cohort N (%) | Cham et al N (%)10 |
|---|---|---|
| Readmission | 69 (9.9) | 688 (9.8) |
| Reoperation | 19 (2.7) | 214 (3.0) |
| Death within 30 Days | 0 (0) | 64 (0.9) |
| Clavien-Dindo (CD) IV complications | 51 (7.3) | 409 (5.8) |
| Sepsis | 31 (4.4) | 166 (2.4) |
| Shock | 10 (1.4) | 63 (0.9) |
| Cardiac Arrest | 0 (0) | 15 (0.2) |
| Myocardial infarction | 1 (0.1) | 22 (0.3) |
| Pulmonary Embolism | 11 (1.6) | 116 (1.7) |
| Ventilation > 48Hr | 1 (0.1) | 69 (1.0) |
| Unplanned Intubation | 4 (0.6) | 65 (0.9) |
| Death/CD IV Complication | 82 (11.7) | 434 (6.2) |
Table 3.
Characteristics of the validation cohort
| Characteristic | Validation Cohort N (%) |
|---|---|
| Timing of Surgery | |
| PDS | 432 (61.7) |
| IDS | 268 (38.3) |
| Stage | |
| I | 15 (2.1) |
| II | 29 (4.1) |
| III A/B | 53 (7.6) |
| III C | 321 (45.9) |
| IV | 279 (39.9) |
| Unstaged | 3 (0.4) |
PDS, primary debulking surgery; IDS, interval debulking surgery
Figure 1:
Distribution of procedure scores for primary debulking surgery (PDS) and interval debulking surgery (IDS)
When applied to our cohort, the c-index calculated for this 30-day complication model was 0.715 (95% CI 0.66–0.768) (Figure 2). The median predicted probability for having a complication was 0.13; the median predicted probability for not having a complication was 0.09 (Figure S1). In the calibration curve, the predictive probability is well observed in the cohort (Figure 3).
Figure 2.
Receiver operating characteristic curve (ROC curve) and area under the curve (AUC) using the validation cohort
Figure 3: Calibration curve using the validation cohort.
The calibration curve is generated by dividing the whole validation cohort into 4 subgroups based upon their predicted probabilities (represented by the black dots). The actual observed probabilities are plotted versus the averaged predicted probability for each of the 4 subgroups, and the 95% CIs are indicated by the error bars. A perfectly accurate nomogram prediction model would result in plot where the observed and predicted probabilities for given groups fall along the 45-degree line (gray line). The dotted line and the red line represent the fitted and smoothed lines for the 4 groups.
An example of an individual case analysis would be the following: a 72-year-old patient undergoing elective surgery for an ovarian mass, with ascites and carcinomatosis noted on preoperative imaging. She has an ASA of 3, normal preoperative albumin of 4.1, no history of a bleeding disorder, and a planned procedure score of 3 (debulking surgery with colonic resection and diaphragm resection). She would be assigned 39 points for age, 31 points for ASA, 0 points for albumin, 0 points for bleeding disorder, 30 points for ascites, 0 points for elective surgery, and 100 for procedure score. Her total score would be 200 points, correlating with a 17% risk of Clavien-Dindo IV or serious complications or death.
In contrast, another case analysis would be the following: a 64-year-old with multiple comorbidities undergoing elective surgery for an ovarian mass seen on preoperative imaging, with lesions noted along the diaphragm, large bowel, and in the pelvis. She has an ASA of 3, a preoperative albumin of 3.6, no history of a bleeding disorder, and a planned procedure score of 3 (debulking surgery with colonic resection and diaphragm resection). She would be assigned 25 points for age, 31 points for ASA, 25 points for albumin, 0 points for bleeding disorder, 0 points for ascites, 0 points for elective surgery, and 100 for procedure score. Her total score would be 181 points, correlating with a 14% risk of Clavien-Dindo IV or serious complications or death.
DISCUSSION
Overall, the nomogram performed well in this cohort of patients undergoing cytoreduction, and was independently validated with a c-index comparable to that published in the original manuscript (0.715 versus 0.7110). Patients undergoing debulking surgery are at risk for postoperative complications, and this was seen in patients undergoing primary debulking surgery and interval debulking surgery in the current study. Postoperative complication rates for Clavien-Dindo III and higher, have been reported up to 19% of women with ovarian cancer undergoing upper abdominal surgery, with a 30-day mortality of 0.8%19. A recent study using the ACS NSQIP database reported severe postoperative complication rates of 3.8–18% in women with ovarian cancer undergoing cytoreduction. Higher rates of serious complications were seen in cases that had higher procedure scores20. In another study performed at a single institution, the rate of serious complications (Accordion III/IV) was 22.3%7. Our combined complication rate is consistent with the prior published data on ovarian cancer patients.
The nomogram validated in this study performs well in predicting which patients are at lower risk for severe complications. As seen in Figure S1 and Figure 3 and in the example cases, most patients have less than 20% risk for serious complications. One potential reason for this is that the overall complexity of surgery was low in the NSQIP dataset. This is a weakness of the NSQIP dataset, in that the complexity of surgery does not mirror the experience of high-volume centers21. Future nomograms and risk models should be better equipped to control for complexity of surgery. This can be done by applying varying weights to specific surgical procedures based upon their respective risk for complications. For example, Chi et al demonstrated that women with ovarian cancer who underwent extensive upper abdominal procedures had a 22% risk of grade III-V complications where 68% of these patients were managed with percutaneous drainage of fluid collections2. A diaphragm resection may require a higher score than a lymphadenectomy. Future endeavors could investigate weighing procedures and correlation with postoperative complications.
While the risk for having a serious complication is weighed in clinical decision-making, patient preferences regarding the tradeoff between risks of surgery and benefit of survival must also be considered. Many women with ovarian cancer consider a 5–6 month increase in overall survival to be meaningful22,23. Havrilesky and colleagues recently demonstrated that ovarian cancer patients are willing to accept higher rates of complications (upwards of 15%) if it increases their expected overall survival from 3 to 3.5 years24. To assist with challenging clinical decisions, an ideal model would delineate which patients have a significantly elevated risk of complication or death.
The use of readily available preoperative markers allows the patient and physician to make important informed decisions regarding preoperative management. Previous models have reported on the impact of age, ASA, ascites, complexity of the surgical procedure, and preoperative albumin7,25,26 on postoperative complications. The present study confirms that age, ASA, procedure score, ascites, and a history of a bleeding disorder are associated with 30-day postoperative complications. For our cohort, ASA was retrieved from the anesthesia notes. An ASA score of 3 indicates a patient with severe systemic disease. At our institution patients with advanced ovarian cancer are often classified as ASA 3 if they present with symptomatic ascites and/or symptomatic pleural effusions. With respect to the ASA 3 score, differences between our cohort and the NSQIP database may reflect more advanced disease in our cohort. Additional markers, including BMI or stage have also been identified as having significant correlation with postoperative complications27. For future models, we are striving to identify additional preoperative clinical markers to help further stratify patients and their risk for complications.
Preoperative predictive models have variable outcomes as well, including either Clavien-Dindo Class III or IV complications, 30-day, 60-day or 90-day mortality, or Accordion complications. Kumar and colleagues argue that 90-day mortality is an important short-term marker useful in counseling patients, as it not only incorporates postoperative complications but complications associated with chemotherapy and the disease process itself7. A recent publication from that same institution showed that a preoperative triage algorithm for selecting candidates for primary debulking surgery or neoadjuvant chemotherapy/interval debulking surgery reduced 90-day mortality from 8.9% to 2.6% (P=0.002) between the current and historic cohorts. In that cohort, patients were deemed high-risk for morbidity and mortality at primary debulking surgery and were triaged to neoadjuvant chemotherapy if they met one of the following criteria: hypoalbuminemia (<3.5 g/dL), age ≥80, age 75–79 and either ASA 3–4, Stage IV disease, or complete surgery anticipated28. This algorithm must be independently validated, but it highlights the importance of these individual factors and their correlation to risk of serious complications. Future endeavors to develop preoperative nomograms could utilize multiple difference outcomes to include reoperation, readmission, serious complications, and mortality.
The strengths of our study are its large patient cohort at a high-volume single institution. This study represents a more recent, independent cohort that validates the nomogram itself, and continues to support the significance of the individual clinical factors as previously described. Additionally, our institution includes a large group of gynecologic oncology surgeons who share common philosophies about management of patients. However, the outcomes of surgical complications are not based solely on the practice of one surgeon.
One limitation of this study is that it is retrospective. There are notable differences between the original cohort and our population. We felt that this model is most useful and clinically meaningful in women undergoing cytoreduction for ovarian cancer; as such, we selected a population of women undergoing primary or interval cytoreduction from our institutional database. In doing so, we selected for a population of women who were in advanced stages of disease, who would have a procedure score of at least “1” by default because nearly the entire cohort underwent a debulking procedure. The designation of our cytoreduction cohort likely leads to the differences seen in the procedure scores between the two populations and explains why the complication rate was higher in our cohort than in the NSQIP population. It is also challenging to parse the definition of “debulking” in the NSQIP database because administrative billing terms are non-informative. Future models may remove this from the scoring system and, in place, provide a more detailed and complex procedure scoring system. Lastly, our institution follows a prespecified algorithm for the determination of PDS or NACT as primary management. Therefore, the clinical significance of certain variables, like preoperative serum albumin, may be underrepresented in the nomogram.
Several steps have been employed at our institution to reduce postoperative complications in the setting of extensive cytoreductive surgery. This includes an algorithm for the initial management of new patients withr advanced ovarian cancer, to triage patients to primary cytoreduction or neoadjuvant chemotherapy based upon clinical and radiologic factors. Additionally, we implemented our Early Recovery After Surgery (ERAS) protocol, venous thromboembolic prophylaxis, and a surgical site reduction bundle in patients undergoing colorectal surgery29, in an effort to decrease postoperative complications.
In summary, the nomogram performs well in an independent cohort at a high-volume surgical center. The nomogram uses readily available patients’ characteristics in the preoperative setting, allowing for upfront counseling on the risks of surgery. There is the potential to investigate additional preoperative variables such as histology, stage, primary cytoreductive surgery verses interval debulking surgery, or radiologic findings, to improve the stratification of patients at increased risk of complications.
Supplementary Material
Figure S1. Violin plot of predicted probabilty of validation cohort The violin plot shows the distribution of the predicted probability side by side for outcome. The middle line correlates with the median for the groups, and the other lines represent the 1st and 3rd quartiles.
Highlights.
This study describes independent validation of a complication nomogram at a high-volume center.
Preoperative characteristics can be used to predict postoperative complications.
Postoperative risk calculators can aid in patient-centric care.
Acknowledgments
FUNDING: This study was funded in part through the NIH/NCI Support Grant P30 CA008748.
DISCLOSURES: AMS reports patent issued (W02019195097A1) for perineal heating device, outside the submitted work. AI reports personal fees from Mylan, outside the submitted work. JDW has been a consultant for Clovis Oncology, and reports research funding from Merck, outside the submitted work. KLR reports other from Intuitive Surgical Inc., outside the submitted work. DSC reports personal fees from Bovie Medical Co., personal fees from Verthermia Inc. (now Apyx Medical Corp.), personal fees from C Surgeries, other from Intuitive Surgical Inc., other from TransEnterix Inc., personal fees from Biom 'Up, outside the submitted work.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
CONFLICTS OF INTEREST: None declared
REFERENCES
- 1.Sioulas VD, Schiavone MB, Kadouri D, et al. Optimal primary management of bulky stage IIIC ovarian, fallopian tube and peritoneal carcinoma: Are the only options complete gross resection at primary debulking surgery or neoadjuvant chemotherapy? Gynecol Oncol 2017;145(1):15–20. (In eng) DOI: 10.1016/j.ygyno.2017.02.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chi DS, Zivanovic O, Levinson KL, et al. The incidence of major complications after the performance of extensive upper abdominal surgical procedures during primary cytoreduction of advanced ovarian, tubal, and peritoneal carcinomas. Gynecol Oncol 2010;119(1):38–42. DOI: 10.1016/j.ygyno.2010.05.031. [DOI] [PubMed] [Google Scholar]
- 3.Horowitz NS, Miller A, Rungruang B, et al. Does aggressive surgery improve outcomes? Interaction between preoperative disease burden and complex surgery in patients with advanced-stage ovarian cancer: an analysis of GOG 182. J Clin Oncol 2015;33(8):937–943. DOI: 10.1200/JCO.2014.56.3106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gerestein CG, Nieuwenhuyzen-de Boer GM, Eijkemans MJ, Kooi GS, Burger CW. Prediction of 30-day morbidity after primary cytoreductive surgery for advanced stage ovarian cancer. Eur J Cancer 2010;46(1):102–109. DOI: 10.1016/j.ejca.2009.10.017. [DOI] [PubMed] [Google Scholar]
- 5.Seagle BL, Butler SK, Strohl AE, Nieves-Neira W, Shahabi S. Chemotherapy delay after primary debulking surgery for ovarian cancer. Gynecol Oncol 2017;144(2):260–265. DOI: 10.1016/j.ygyno.2016.11.022. [DOI] [PubMed] [Google Scholar]
- 6.AlHilli M, Langstraat C, Tran C, et al. Risk factors and indications for 30-day readmission after primary surgery for epithelial ovarian cancer. Int J Gynecol Cancer 2015;25(2):193–202. DOI: 10.1097/IGC.0000000000000339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kumar A, Janco JM, Mariani A, et al. Risk-prediction model of severe postoperative complications after primary debulking surgery for advanced ovarian cancer. Gynecol Oncol 2016;140(1):15–21. DOI: 10.1016/j.ygyno.2015.10.025. [DOI] [PubMed] [Google Scholar]
- 8.Cowan RA, O’Cearbhaill RE, Gardner GJ, et al. Is It Time to Centralize Ovarian Cancer Care in the United States? Ann Surg Oncol 2016;23(3):989–993. DOI: 10.1245/s10434-015-4938-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tanner EJ, Long KC, Zhou Q, et al. Impact of operative start time on surgical outcomes in patients undergoing primary cytoreduction for advanced ovarian cancer. Gynecol Oncol 2012;126(1):58–63. (In eng) DOI: 10.1016/j.ygyno.2012.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Cham S, Chen L, St Clair CM, et al. Development and validation of a risk-calculator for adverse perioperative outcomes for women with ovarian cancer. Am J Obstet Gynecol 2019;220(6):571 e1–571 e8. DOI: 10.1016/j.ajog.2019.02.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Khuri SF, Daley J, Henderson W, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. National VA Surgical Quality Improvement Program. Ann Surg 1998;228(4):491–507. DOI: 10.1097/00000658-199810000-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.User Guide for the 2015 ACS NSQIP Participant Use Data File (PUF) 2015.
- 13.Aletti GD, Eisenhauer EL, Santillan A, et al. Identification of patient groups at highest risk from traditional approach to ovarian cancer treatment. Gynecol Oncol 2011;120(1):23–28. DOI: 10.1016/j.ygyno.2010.09.010. [DOI] [PubMed] [Google Scholar]
- 14.Straubhar AM, Filippova OT, Cowan RA, et al. A multimodality triage algorithm to improve cytoreductive outcomes in patients undergoing primary debulking surgery for advanced ovarian cancer: A Memorial Sloan Kettering Cancer Center team ovary initiative. Gynecol Oncol 2020;158(3):608–613. DOI: 10.1016/j.ygyno.2020.05.041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Suidan RS, Ramirez PT, Sarasohn DM, et al. A multicenter prospective trial evaluating the ability of preoperative computed tomography scan and serum CA-125 to predict suboptimal cytoreduction at primary debulking surgery for advanced ovarian, fallopian tube, and peritoneal cancer. Gynecol Oncol 2014;134(3):455–461. DOI: 10.1016/j.ygyno.2014.07.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Suidan RS, Ramirez PT, Sarasohn DM, et al. A multicenter assessment of the ability of preoperative computed tomography scan and CA-125 to predict gross residual disease at primary debulking for advanced epithelial ovarian cancer. Gynecol Oncol 2017;145(1):27–31. DOI: 10.1016/j.ygyno.2017.02.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240(2):205–213. DOI: 10.1097/01.sla.0000133083.54934.ae. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Iasonos A, Schrag D, Raj GV, Panageas KS. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol 2008;26(8):1364–1370. DOI: 10.1200/JCO.2007.12.9791. [DOI] [PubMed] [Google Scholar]
- 19.Benedetti Panici P, Di Donato V, Fischetti M, et al. Predictors of postoperative morbidity after cytoreduction for advanced ovarian cancer: Analysis and management of complications in upper abdominal surgery. Gynecol Oncol 2015;137(3):406–411. (In eng) DOI: 10.1016/j.ygyno.2015.03.043. [DOI] [PubMed] [Google Scholar]
- 20.Patankar S, Burke WM, Hou JY, et al. Risk stratification and outcomes of women undergoing surgery for ovarian cancer. Gynecol Oncol 2015;138(1):62–69. (In eng) DOI: 10.1016/j.ygyno.2015.04.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Narasimhulu DM, Kumar A, Weaver AL, McGree ME, Langstraat CL, Cliby WA. Less guessing, more evidence in the decision for neoadjuvant chemotherapy versus surgery: A validated triage algorithm to identify women at high risk of morbidity/mortality associated with ovarian cancer cytoreduction. Gynecol Oncol 2020;159:48 DOI: 10.1016/j.ygyno.2020.06.101. [DOI] [Google Scholar]
- 22.Donovan KA, Greene PG, Shuster JL, Partridge EE, Tucker DC. Treatment preferences in recurrent ovarian cancer. Gynecol Oncol 2002;86(2):200–211. (In eng) DOI: 10.1006/gyno.2002.6748. [DOI] [PubMed] [Google Scholar]
- 23.Minion LE, Coleman RL, Alvarez RD, Herzog TJ. Endpoints in clinical trials: What do patients consider important? A survey of the Ovarian Cancer National Alliance. Gynecol Oncol 2016;140(2):193–198. (In eng) DOI: 10.1016/j.ygyno.2015.11.030. [DOI] [PubMed] [Google Scholar]
- 24.Havrilesky LJ, Yang JC, Lee PS, et al. Patient preferences for attributes of primary surgical debulking versus neoadjuvant chemotherapy for treatment of newly diagnosed ovarian cancer. Cancer 2019;125(24):4399–4406. (In eng) DOI: 10.1002/cncr.32447. [DOI] [PubMed] [Google Scholar]
- 25.Aletti GD, Dowdy SC, Podratz KC, Cliby WA. Relationship among surgical complexity, short-term morbidity, and overall survival in primary surgery for advanced ovarian cancer. Am J Obstet Gynecol 2007;197(6):676 e1–7. DOI: 10.1016/j.ajog.2007.10.495. [DOI] [PubMed] [Google Scholar]
- 26.Uppal S, Al-Niaimi A, Rice LW, et al. Preoperative hypoalbuminemia is an independent predictor of poor perioperative outcomes in women undergoing open surgery for gynecologic malignancies. Gynecol Oncol 2013;131(2):416–422. DOI: 10.1016/j.ygyno.2013.08.011. [DOI] [PubMed] [Google Scholar]
- 27.Kumar A, Bakkum-Gamez JN, Weaver AL, McGree ME, Cliby WA. Impact of obesity on surgical and oncologic outcomes in ovarian cancer. Gynecol Oncol 2014;135(1):19–24. DOI: 10.1016/j.ygyno.2014.07.103. [DOI] [PubMed] [Google Scholar]
- 28.Narasimhulu DM, Kumar A, Weaver AL, McGree ME, Langstraat CL, Cliby WA. Using an evidence-based triage algorithm to reduce 90-day mortality after primary debulking surgery for advanced epithelial ovarian cancer. Gynecol Oncol 2019;155(1):58–62. (In eng) DOI: 10.1016/j.ygyno.2019.08.004. [DOI] [PubMed] [Google Scholar]
- 29.Schiavone MB, Moukarzel L, Leong K, et al. Surgical site infection reduction bundle in patients with gynecologic cancer undergoing colon surgery. Gynecol Oncol 2017;147(1):115–119. DOI: 10.1016/j.ygyno.2017.07.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1. Violin plot of predicted probabilty of validation cohort The violin plot shows the distribution of the predicted probability side by side for outcome. The middle line correlates with the median for the groups, and the other lines represent the 1st and 3rd quartiles.