BACKGROUND AND OBJECTIVES:The proportion of free-standing hospitals is rapidly declining in favor of healthcare systems. We aim to compare outcomes after pancreatoduodenectomy (PD) between hospital volume and surgeon volume within consolidated healthcare systems. METHODS:Utilizing New York State's SPARCS database, we identified adults undergoing PD (2016-2019) at flagship (highest volume within a system) or non-flagship hospitals. Analysis compared 30- and 90-day outcomes by hospital and surgeon annual volume (hospital volume ≥ 20 vs < 20PD/yr and surgical volume ≥ 10 vs < 10PD/yr). We controlled for patient self-selection across surgeons and facilities using propensity-score matching. RESULTS:Among 1633 patients, 68.8% were treated at flagship hospitals. Compared to patients treated at non-flagship hospitals, those at flagship hospitals had lower median postoperative length of stay (8-days vs. 10-days, p < 0.01), fewer transfusions (17% vs 29%, p < 0.01), other post-procedural (8.1% vs 12.6%, p = 0.02), and overall complications (53% vs 59%, p = 0.07). The lowest overall complications and length of stay (p < 0.05) were observed at high-volume hospitals regardless of surgeon volume. CONCLUSIONS:More favorable outcomes among pancreatoduodenectomy patient were observed at high-volume flagship hospitals. When planning a complex surgery, patients should be referred to the highest volume hospital within their healthcare system to optimize post-operative outcomes.

BACKGROUND AND AIMS:While there is low risk of emergent appendectomy patients becoming diagnosed with a neoplasm, the rates are much higher in patients undergoing interval appendectomies. This study compares the incidence rates of new primary right colon neoplasms and postoperative complications between interval and primary appendectomies. METHODS:We performed a retrospective cohort study from 2001 to 2020 involving patients from 92 healthcare organizations within TriNetX. Patients' diagnosis and procedure were extracted from ICD and CPT codes. Primary and interval appendectomy (IA) cohorts were defined as having an appendectomy procedure 0-14 and 15-90 days from the initial diagnosis, respectively. Patients in each cohort were matched 1:1 on demographics via Coarsened Exact Matching (CEM). Postoperative complications and oncological outcomes were defined by specific ICD and/or CPT codes. RESULTS:Post-matching, there were 2803 patients included in each cohort. IA patients were associated with 3.07 times greater incidence rates of malignant neoplasm of the right colon and appendix (p < 0.001). However, IA patients were associated with 2.63 times lower postoperative complication rates (p < 0.001). CONCLUSION:IA patients are more likely to be diagnosed with new right-sided neoplasms postoperatively. Future studies are necessary to determine appropriate cancer screening methods for IA candidates.

e16449 Background: Individual hospitals are rapidly consolidating into large healthcare systems. While evidence exists on volume-outcome relationships for individual hospitals for the Whipple procedure, evidence on healthcare systems is lacking. We aim to compare outcomes for patients undergoing Whipple procedures between flagship and non-flagship hospitals to examine how hospital systems impact surgical quality for highly specialized surgical care. Methods: Utilizing New York State’s (NYS) Statewide Planning and Research Cooperative System, we identified adult patients who underwent Whipple’s procedure between 2016 and 2019. Patients were categorized into two cohorts: those treated in a flagship (highest volume hospital in a system) or non-flagship hospital. Free-standing hospitals and systems with fewer than 5 Whipple procedures annually at their flagship hospital were excluded (n = x). Patients with a non-pancreatic cancer that metastasized to the pancreases and patients with concomitant surgeries performed same-day were excluded (n = x). 1:1 propensity score matching was performed to balance cohorts for analysis. Patient characteristics and outcomes (30-day and 90-day) were analyzed between cohorts in unmatched and matched (on sex, race and insurance) samples using bivariate and multivariate analysis with alpha set at 0.05. Results: Overall, 1633 patients were included: 1126 (68.6%) received care at flagship hospitals and 515 (31.4%) at non-flagship hospitals. Whipple procedures were performed in X% of NYS hospitals including Y systems (Median hospitals per system: X). Patients treated at flagship hospitals were more often White (65% vs. 41%, p < 0.01) and less often Medicaid recipients (17% vs 30%, p < 0.01) versus non-flagship hospitals. After propensity-score-matching, patients at flagship hospitals had lower 30-day transfusion requirements (18% vs. 27%, p < 0.01), and overall fewer postoperative complications at 30-days (47% ) and 90-days (49% vs. 56%, p = 0.03) compared to patients in non-flagship hospitals in the same system; there was no difference in 90-day readmission (18% vs. 17%, p = 0.49) or mortality (2% vs 1%, p = 0.23). Conclusions: When planning a Whipple procedure, patients should be referred to the highest volume Whipple procedure hospital within their healthcare system to optimize post-operative outcomes. To facilitate access to high quality surgical services, providers should screen patients for social determinants or health and help connect patients with appropriate services to minimize patients burden and financial toxicities.

Objective: This study aimed to compare surgical resection versus ablation for managing liver malignancies in patients 65 and older. Material and Methods: Cases with liver tumors were extracted from the NSQIP database for patients aged ≥65 years. Following propensity score matching, multivariate Cox regression was used for 30-day morbidity and mortality for liver resection and ablation. Results: Following a propensity score matching, 1048 patients were 1:1 matched for comorbid conditions. Patients stayed in the hospital three days longer after resection (p<0.001). Mortality was lower after ablation (p= 0.013). This difference was more prominent in patients with primary liver tumors (p= 0.008). Group A had a 10-fold lower risk of developing an abdominal abscess, a fourfold decrease in hospital-associated pneumonia (p= 0.001) and reintubation, a 10-fold reduction in bleeding requiring transfusion (p<0.001), and a three-fold decrease in risk of developing sepsis (p<0.001). Conclusion: Despite being a generally sicker patient population with worse underlying liver function, ablative techniques were associated with a lower risk of adverse outcomes when compared to more aggressive resection of primary malignant tumors of the liver.

Background: Partial nephrectomy is the preferred treatment method for certain kidney tumors owing to its nephron-preserving benefit. We aimed to compare the occurrence and importance of complications after radical (RN) or partial (PN) nephrectomy for localized renal cell carcinoma (RCC) and contribution to patient mortality. Methods: All RCC cases were extracted from the National Surgical Quality Improvement Program between 2005 and 2017. All-cause mortality was the primary endpoint that defined a failure to rescue (FTR) after the occurrences of complications. Propensity score matching (PSM) was performed to adjust for confounding variables between the 2 groups. The exclusion criteria included patients on dialysis, those with distant metastases, and those with concurrent procedures. Null hypotheses were rejected when P-values were < .05. Results: The database included 24,830 patients, with 22,015 in the RN group and 2815 in the PN. After PSM, 2226 patients after PN were matched 1:1 to an equal number of patients who underwent RN. Overall, postoperative complications occurred in 20.5% after RN, more frequent than 15.9% after PN (P < .001). While the mortality rates were similar, patients were more likely to experience blood transfusion and reintubation, as well as longer hospital stay after RN, while they were likely to have more infectious complications, including abdominal abscess, and more likely to return to the operating room after PN. In both groups, cardiac and respiratory complications were associated with FTR, leading to mortality. Excessive bleeding (requiring transfusion) was also a significant cause of death after RN but not after PN. Conclusion: When planning PN or RN, patients should be counseled on the risks of their selective procedure and the potential increased mortality risk with certain complications. These risks should be weighed against the benefit of those cancer surgeries.

Background: Hospitals are rapidly consolidating to create large healthcare systems. Whether outcomes following colorectal cancer resection at flagship hospitals differ from those at non-flagship hospitals is unknown. Methods: A 6-year retrospective analysis of an all-payor New York State (NYS) hospital database was conducted. All adult patients with a colorectal resection for primary resectable colorectal cancer were included. Within each system, the hospital with the most colorectal resections was designated the "flagship" hospital. Thirty-day outcomes at flagship facilities were compared to affiliated, non-flagship hospitals following colorectal resection while matching for patient-level differences. Results: In total, 28 400 patients were included across 31 healthcare systems in NYS. There were no differences in mortality (0.9% vs. 1.1%), 30-day readmissions (10.5% vs. 11.9%), or postoperative outcomes between matched patients treated at flagship versus non-flagship facilities (p > 0.05). Conclusions: There are no differences in perioperative outcomes between flagship and non-flagship hospitals in a given system in NYS. Patients with resectable non-metastatic colorectal cancer may not need to undergo oncologic resection at flagship hospitals to receive high-quality perioperative care.

Background Minimally invasive techniques for pancreaticoduodenectomy (PD) are increasing in practice, however, data remains limited regarding perioperative outcomes. Our study sought to compare patients undergoing open pancreaticoduodenectomy (OPD) with those undergoing laparoscopic (LPD) or robot-assisted pancreaticoduodenectomy (RPD). Methods Patients who underwent PD during 2016-2018 were identified from the New York State Planning and Research Cooperative System database. Results Of the 1954 patients identified, 1708 (87.4%) underwent OPD, 165 (8.4%) underwent LPD, and 81 (4.2%) underwent RPD. The majority of patients were White (63.8%), males (53.3%) with a mean age of 65.4 years. RPD patients had a lower median Charlson Comorbidity Index (2) than OPD (3) or LPD (3, p = 0.01) and had a lower 30-day rate of complications (35.8% vs. 48.3% vs. 43.6% respectively, p = 0.05). After propensity-score matching, however, there were no differences between the groups regarding overall complications, surgical site infections, anastomotic leaks, or mortality (p = NS for all). OPD demonstrated a longer length of stay (median 8 days) compared to LPD (7 days) or RPD (7 days, p < 0.01). Conclusions Patients undergoing LPD and RPD have a shorter length of hospital stay compared to OPD and there was no difference in overall morbidity or mortality when matched to similar patients.

Background. Esophageal squamous cell carcinoma (ESCC) has been poorly studied, approached with therapeutic nihilism, and likely undertreated. We studied the impact of clinical and patient factors on the survival of ESCC in the United States. Methods. We selected patients with stage I to III ESCC from 2004 to 2013, using the National Cancer Database. Patients were categorized into the following treatment modalities: (1) definitive chemoradiation therapy (CR), (2) neoadjuvant therapy followed by esophageal resection (ER), (3) ER alone, and (4) ER followed by adjuvant therapy. Our main outcome measure was overall survival. Results. We identified 11,229 patients with ESCC undergoing definitive CR (78.6%); neoadjuvant therapy followed by ER (8.5%), ER alone (10.1%), and ER followed by adjuvant therapy (2.6%). Compared with neoadjuvant therapy, both ER alone and definitive CR were associated with substantially increased mortality. Patients treated at high-volume centers (>20), regardless of whether they underwent ER, had improved survival compared with facilities that performed 10 to 19, 5 to 9, and less than 5 ERs per year. Conclusions. Patients treated at high-volume facilities were more likely to receive neoadjuvant therapy, and there was a marked inverse relationship between annual surgical volume and long-term survival for both surgically and non-surgically treated patients with stage I to III ESCC. (C) 2019 by The Society of Thoracic Surgeons

Journal of the American College of Surgeons 229(4):p e94, October 2019. | DOI: 10.1016/j.jamcollsurg.2019.08.969

3591 Background: Although preoperative platelet/leukocyte ratio (PLR) is a predictor of postoperative outcomes in various neoplasms, data is lacking for colorectal cancer (CRC). We hypothesized that elevated preoperative PLR would be an independent risk factor for postoperative complications and increased 30-day mortality in patients with surgically resected CRC. Methods: Patients undergoing resections for CRC were identified from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) from 2005 to 2016 dataset. Logistic regression models for 30-day morbidity and mortality for PLR ≤28.4 and >28.4 (calculated from ROC analyses) were completed. Univariate log-rank test, and multivariate Cox proportional hazards regression were used for time-to-event analyses and data are presented as odds/hazard ratio with confidence interval. Results: 98,398 patients were included in the study. Elevated PLR was an independent predictor of 30-day morbidity and mortality. Patients with high PLR were more likely to have a wound infection (1.022, 1.01-1.04), pneumonia (1.2, 1.17-1.23), require reintubation (1.23, 1.19-127), prolonged ventilation (1.31, 1.26-1.36), have renal insufficiency (1.17, 1.11-1.22), and have a postoperative myocardial infarction (1.23, 1.9-127). PLR value > 28.4 was among independent predictors of mortality (1.67, 1.47–1.87), along with the ASA class, age, perioperative use of steroids, serum albumin and INR. (Table). Conclusions: Elevated PLR is a significant preoperative risk factor for 30-day morbidity and mortality. PLR may be considered as a potential risk assessment tool that predicts postoperative outcome following CRC resections. Multivariate analysis. [Table: see text]

Background and ObjectivesTo compare outcomes in patients receiving neoadjuvant stereotactic body radiation therapy (SBRT) with those receiving intensity-modulated radiation therapy (IMRT) for pancreatic adenocarcinoma. MethodsWe analyzed patients receiving neoadjuvant SBRT for borderline resectable (BRPC) and locally advanced pancreatic cancer (LAPC) (2012-2016). Differences in baseline characteristics, perioperative outcomes, progression-free survival (PFS), and overall survival (OS) were compared. ResultsSeventy-five (82.4%) patients received SBRT and 16 (17.6%) received IMRT. There were no differences in surgical resection rates in the SBRT (n=38, 50.7%) and IMRT (n=11, 68.8%) groups (P=0.188). Among resected patients, there was no difference in postoperative outcomes or pathologic outcomes including lymph node status, margin status, lymphovascular and perineural invasion, or pathologic response to neoadjuvant treatment (P>0.05). Among all patients, median PFS and OS were 9.9 and 23.5 months in the SBRT group, respectively, and 15.3 and 21.8 months in the IMRT group, respectively (P>0.05). Similarly, there was no difference in PFS or OS between groups when stratified by BRPC, LAPC, and surgically resected patients (P>0.05). ConclusionsIn the neoadjuvant setting, SBRT and IMRT appear to have similar rates of resection, perioperative outcomes, and survival outcomes, but additional studies with increased sample size and longer follow up are needed.

Planar lymphoscintigraphy (PL) has a lower detection rate of sentinel lymph nodes (SLNs) in head and neck melanoma compared with other sites. We assessed situations when single-photon emission computed tomography/computed tomography (SPECT/CT) identified nodes not seen by PL. We also evaluated the impact of SPECT/CT on surgical approach and oncologic outcomes.

Hyperthermic intraperitoneal chemotherapy (HIPEC) administration can be associated with hyperglycemia during perfusion. Little is known about this effect, and no previous studies have examined patient characteristics associated with perfusion-related hyperglycemia.