Overall Volume of Upper Gastrointestinal Surgery Positively Impacts Gastric Cancer Outcomes at Centers with Low Gastrectomy Volume.

BACKGROUND: The relationship between hospital volume and surgical mortality is well documented. However, complete centralization of surgical care is not always feasible. The present study investigates how overall volume of upper gastrointestinal surgery at hospitals influences patient outcomes following resection for gastric adenocarcinoma. PATIENTS AND METHODS: National Cancer Database (2010-2019) patients with pathologic stage 1-3 gastric adenocarcinoma who underwent gastrectomy were identified. Three cohorts were created: low-volume hospitals (LVH) for both gastrectomy and overall upper gastrointestinal operations, mixed-volume hospital (MVH) for low-volume gastrectomy but high-volume overall upper gastrointestinal operations, and high-volume gastrectomy hospitals (HVH). Chi-squared tests were used to analyze sociodemographic factors and surgical outcomes and Kaplan-Meier method for survival analysis. RESULTS: In total, 26,398 patients were identified (LVH: 20,099; MVH: 539; HVH: 5,760). The 5-year survival was equivalent between MVH and HVH for all stages of disease (MVH: 56.0%, HVH 55.6%; p = 0.9866) and when stratified into early (MVH: 69.9%, HVH: 65.4%; p = 0.1998) and late stages (MVH: 24.7%, HVH: 32.0%; p = 0.1480), while LVH had worse survival. After matching patients, postoperative outcomes were worse for LVH, but there was no difference between MVH and HVH in terms of adequate lymphadenectomy, margin status, readmission rates, and 90-day mortality rates. CONCLUSIONS: Despite lower gastrectomy volume for cancer, postoperative gastrectomy outcomes at centers that perform a high number of upper gastrointestinal cancer surgeries were similar to hospitals with high gastrectomy volume. These hospitals offer a blueprint for providing equivalent outcomes to high volume centers while enhancing availability of quality cancer care.

Trends in access to minimally invasive pancreaticoduodenectomy for pancreatic cancers.

BACKGROUND: Minimally invasive pancreaticoduodenectomy (MIPD), including robotic (RPD) and laparoscopy (LPD), is becoming more frequently employed in the management of pancreatic ductal adenocarcinoma (PDAC), though the majority of operations are still performed via open approach (OPD). Access to technologic advances often neglect the underserved. Whether disparities in access to MIPD exist, remain unclear. METHODS: The National Cancer Database (NCDB) was queried (2010-2020) for patients who underwent pancreatoduodenectomy for PDAC. Cochran-Armitage tests assessed for trends over time. Social determinants of health (SDH) were compared between approaches. Multinomial logistic models identified predictors of MIPD. RESULTS: Of 16,468 patients, 80.03 % underwent OPD and 19.97 % underwent MIPD (22.60 % robotic; 77.40 % laparoscopic). Black race negatively predicted LPD (vs white (OR 0.822; 95 % CI 0.701-0.964)). Predictors of RPD included Medicare/other government insurance (vs uninsured or Medicaid (OR 1.660; 95 % CI 1.123-2.454)) and private insurance (vs uninsured or Medicaid (OR 1.597; 95 % CI 1.090-2.340)). Early (2010-2014) vs late (2015-2020) diagnosis, stratified by race, demonstrated an increase in Non-White patients undergoing OPD (13.15 % vs 14.63 %; p = 0.016), but not LPD (11.41 % vs 13.57 %;p = 0.125) or RPD (14.15 % vs 15.23 %; p = 0.774). CONCLUSION: SDH predict surgical approach more than clinical stage, facility type, or comorbidity status. Disparities in race and insurance coverage are different between surgical approaches.

The Persistence of Poverty and its Impact on Cancer Diagnosis, Treatment and Survival.

OBJECTIVE: To determine the effect of persistent poverty on the diagnosis, surgical resection and survival of patients with non-small cell lung (NSCLC), breast, and colorectal cancer. BACKGROUND: Disparities in cancer outcomes exist in counties with high levels of poverty, defined as ≥20% of residents below the federal poverty level. Despite this well-established association, little is known about how the duration of poverty impacts cancer care and outcomes. One measure of poverty duration is that of "persistent poverty," defined as counties in high poverty since 1980. METHODS: In this retrospective cohort study, patients with NSCLC, breast and colorectal cancer were identified from SEER (2012-2016). County-level poverty was obtained from the American Community Survey (1980-2015). Outcomes included advanced stage at diagnosis (stage III-IV), resection of localized disease (stage I-II) and cancer-specific survival. Hierarchical generalized linear models and accelerated failure time models with Weibull distribution were used, adjusted for patient-level covariates and region. RESULTS: Overall, 522,514 patients were identified, of which 5.1% were in persistent poverty. Patients in persistent poverty were more likely to present with advanced disease [NSCLC odds ratio (OR): 1.12, 95% confidence interval (CI): 1.06-1.18; breast OR: 1.09, 95% CI: 1.02-1.17; colorectal OR: 1.00, 95% CI: 0.94-1.06], less likely to undergo surgery (NSCLC OR: 0.81, 95% CI: 0.73-0.90; breast OR: 0.82, 95% CI: 0.72-0.94; colorectal OR: 0.84, 95% CI: 0.70-1.00) and had increased cancer-specific mortality (NSCLC HR: 1.09, 95% CI: 1.06-1.13; breast HR: 1.18, 95% CI: 1.05-1.32; colorectal HR: 1.09, 95% CI: 1.03-1.17) as compared with those without poverty. These differences were observed to a lesser magnitude in counties with current, but not persistent, poverty and disappeared in counties no longer in poverty. CONCLUSIONS: The duration of poverty has a direct impact on cancer-specific outcomes, with the greatest effect seen in persistent poverty and resolution of disparities when a county is no longer in poverty. Policy focused on directing resources to communities in persistent poverty may represent a possible strategy to reduce disparities in cancer care and outcomes.

A Paradigm Shifts: Neoadjuvant Therapy for Clearly Resectable Pancreatic Cancer.

Over the past decade, neoadjuvant therapy has become the standard of care for patients with borderline resectable and locally advanced pancreatic cancer. The surgical community remains divided regarding the value of neoadjuvant therapy for patients who present with clearly resectable disease. Thus far, randomized controlled trials comparing neoadjuvant therapy with conventional upfront surgical strategies for patients with clearly resectable pancreatic cancer have been plagued by poor accrual, and are often underpowered. Nonetheless, meta-analyses of the results of these trials suggest that neoadjuvant therapy can be offered as an acceptable standard of care for patients with clearly resectable pancreatic cancer. Previous trials used neoadjuvant gemcitabine, but more recent studies have demonstrated superior survival for patients who were able to tolerate neoadjuvant FOLFIRINOX (leucovorin, 5-fluorouracil, irinotecan hydrochloride, and oxaliplatin). The increased utilization of FOLFIRINOX may be shifting the treatment paradigm in favor of neoadjuvant therapy among patients with clearly resectable disease. Randomized controlled trials assessing the value of neoadjuvant FOLFIRINOX in clearly resectable pancreatic cancer, which are expected to provide more conclusive recommendations, are still ongoing. This review outlines the rationale, considerations, and current level of evidence for the use of neoadjuvant therapy in patients with clearly resectable pancreatic cancer.

Axillary Lymph Node Dissection is Associated with Improved Survival Among Men with Invasive Breast Cancer and Sentinel Node Metastasis.

BACKGROUND: Male breast cancer (MBC) is rare, and management is extrapolated from trials that enroll only women. It is unclear whether contemporary axillary management based on data from landmark trials in women may also apply to men with breast cancer. This study aimed to compare survival in men with positive sentinel lymph nodes after sentinel lymph node biopsy (SLNB) alone versus complete axillary dissection (ALND). PATIENTS AND METHODS: Using the National Cancer Database, men with clinically node-negative, T1 and T2 breast cancer and 1-2 positive sentinel nodes who underwent SLNB or ALND were identified from 2010 to 2020. Both 1:1 propensity score matching and multivariate regression were used to identify patient and disease variables associated with ALND versus SLNB. Survival between ALND and SLNB were compared using Kaplan-Meier methods. RESULTS: A total of 1203 patients were identified: 61.1% underwent SLNB alone and 38.9% underwent ALND. Treatment in academic centers (36.1 vs. 27.7%; p < 0.0001), 2 positive lymph nodes on SLNB (32.9 vs. 17.3%, p < 0.0001) and receipt or recommendation of chemotherapy (66.5 vs. 52.2%, p < 0.0001) were associated with higher likelihood of ALND. After propensity score matching, ALND was associated with superior survival compared with SLNB (5-year overall survival of 83.8 vs. 76.0%; log-rank p = 0.0104). DISCUSSION: The results of this study suggest that among patients with early-stage MBC with limited sentinel lymph node metastasis, ALND is associated with superior survival compared with SLNB alone. These findings indicate that it may be inappropriate to extrapolate the results of the ACOSOG Z0011 and EORTC AMAROS trials to MBC.

Combined High-Volume Common Complex Cancer Operations Safeguard Long-Term Survival in a Low-Volume Individual Cancer Operation Setting.

BACKGROUND: We previously demonstrated the importance of combined complex surgery volume on short-term outcomes of high-risk cancer operations. This study investigates the impact of combined common complex cancer operation volume on long-term outcomes at hospitals with low cancer-specific operation volumes. PATIENTS AND METHODS: A retrospective cohort of National Cancer Data Base (2004-2019) patients undergoing surgery for hepatocellular carcinoma, non-small cell lung cancers, or pancreatic, gastric, esophageal, or rectal adenocarcinomas was utilized. Three separate cohorts were established: low-volume hospitals (LVH), mixed-volume hospitals (MVH) with low-volume individual cancer operations and high-volume total complex operations, and high-volume hospitals (HVH). Survival analyses were performed for overall, early-, and late-stage disease. RESULTS: The 5 year survival was significantly better at MVH and HVH compared with LVH, for all operations except late-stage hepatectomy (HVH survival > LVH and MVH). The 5 year survival probability was similar between MVH and HVH for operations on late-stage cancers. Early and overall survival for gastrectomy, esophagectomy, and proctectomy were equivalent between MVH and HVH. While early and overall survival for pancreatectomy were benefited by HVH over MVH, the opposite was true for lobectomy/pneumonectomy, which were benefited by MVH over HVH; however, none of these differences were likely to have an effect clinically. Only hepatectomy patients demonstrated statistical and clinical significance in 5 year survival at HVH compared with MVH for overall survival. CONCLUSIONS: MVH hospitals performing sufficient complex common cancer operations demonstrate similar long-term survival for specific high-risk cancer operations to HVH. MVH provide an adjunctive model to the centralization of complex cancer surgery, while maintaining quality and access.

The Evaluation of Gallstone Disease in the Year Before Pancreatic Cancer Diagnosis.

INTRODUCTION: Patients with pancreatic cancer can present with a variety of insidious abdominal symptoms, complicating initial diagnosis. Early symptoms of pancreatic cancer often mirror those associated with gallstone disease, which has been demonstrated to be a risk factor for this malignancy. This study aims to compare the incidence of gallstone disease in the year before diagnosis of pancreatic ductal adenocarcinoma (PDAC) as compared to the general population, and evaluate the association of gallstone disease with stage at diagnosis and surgical intervention. METHODS: Patients with PDAC were identified from SEER-Medicare (2008-2015). The incidence of gallstone disease (defined as cholelithiasis, cholecystitis and/or cholecystectomy) in the 1 year before cancer diagnosis was compared to the annual incidence in an age-matched, sex-matched, and race-matched noncancer Medicare cohort. RESULTS: Among 14,654 patients with PDAC, 4.4% had gallstone disease in the year before cancer diagnosis. Among the noncancer controls (n = 14,654), 1.9% had gallstone disease. Both cohorts had similar age, sex and race distributions. PDAC patients with gallstone disease were diagnosed at an earlier stage (stage 0/I-II, 45.8% versus 38.1%, P < 0.0001) and a higher proportion underwent resection (22.7% versus 17.4%, P = 0.0004) compared to patients without gallstone disease. CONCLUSIONS: In the year before PDAC diagnosis, patients present with gallstone disease more often than the general population. Improving follow-up care and differential diagnosis strategies may help combat the high mortality rate in PDAC by providing an opportunity for earlier stage of diagnosis and earlier intervention.

A nationwide propensity score analysis comparing ablation and resection for hepatocellular carcinoma.

BACKGROUND AND OBJECTIVES: Studies have reported ambiguous results regarding the efficacy of ablation for early-stage hepatocellular carcinoma (HCC). Our study compared outcomes of ablation versus resection for HCC ≤50 mm to identify tumor sizes that would most benefit from ablation in terms of long-term survival. METHODS: The National Cancer Database was queried for patients with stage I and II HCC ≤50 mm who underwent ablation or resection (2004-2018). Three cohorts were created based on tumor size: ≤20, 21-30, and 31-50 mm. A propensity score-matched survival analysis was performed using the Kaplan-Meier method. RESULTS: In total, 36.47% (n = 4263) and 63.53% (n = 7425) of patients underwent resection and ablation, respectively. After matching, resection was associated with a significant survival benefit compared to ablation (3-year survival: 78.13% vs. 67.64%; p < 0.0001) in patients with HCC of ≤20 mm. The impact of resection was even more striking among patients with HCC of 21-30 mm (3-year survival: 77.88% vs. 60.53%; p < 0.0001) and 31-50 mm (3-year survival: 67.21% vs. 48.55%; p < 0.0001). CONCLUSIONS: While resection offers a survival benefit over ablation in the treatment of early-stage HCC ≤50 mm, ablation may provide a feasible bridging strategy in patients awaiting transplantation.

Disparities in the Receipt of Recommended Curative Treatment for Patients with Early-Stage Hepatocellular Carcinoma.

BACKGROUND: In early-stage hepatocellular carcinoma (HCC), the receipt of recommended care is critical for long-term survival. Unfortunately, not all patients decide to undergo therapy. We sought to identify factors associated with the decision to decline recommended intervention among patients with early-stage HCC. METHODS: The National Cancer Database was queried for patients diagnosed with clinical stages I and II HCC (2004-2017). Cohorts were created based on the receipt or decline of recommended interventions-hepatectomy, liver transplantation, and ablation. Multivariable logistic regression identified predictors for declining intervention, and propensity score analysis was used to calculate the respective odds. Survival analysis was performed using the Kaplan-Meier method. RESULTS: Of 20,863 patients, 856 (4.1%) declined intervention. Patients who were documented as having declined intervention were more often Black (vs. other: OR, 1.3; 95% CI, 1.1-1.6; p = 0.0038), had Medicaid or no insurance (vs. Private, Medicare, or other government insurance): OR, 1.9; 95% CI, 1.6-2.3; p < 0.0001), lived in a low-income area (vs. other: OR, 1.4; 95% CI, 1.2-1.7; p < 0.0001), and received treatment at a non-academic center (vs. academic: OR, 2.1; 95% CI, 1.9-2.5; p < 0.0001). Patients who declined recommended interventions had worse survival compared to those who received treatment (22.9 vs. 59.2 months; p < 0.0001, respectively). CONCLUSIONS: Racial and socioeconomic disparities persist in the decision to undergo recommended treatment. Underutilization of treatment acts as a barrier to addressing racial and socioeconomic disparities in early-stage HCC outcomes.

Volume of Pancreas-Adjacent Operations Favorably Influences Pancreaticoduodenectomy Outcomes at Lower Volume Pancreas Centers.

OBJECTIVE: This study assesses how the volume of pancreatic-adjacent operations (PAO) impacts the outcomes of pancreaticoduodenectomy (PD). SUMMARY BACKGROUND DATA: It is well-established that regionalization benefits outcomes after PD. However, due to a multitude of factors, including geographic, financial, and personal, not all patients receive their care at high-volume pancreas surgery centers. METHODS: The National Cancer Database was queried for pancreatic cancer patients who underwent PD. Hospital volume was calculated for PD and PAO (defined as gastric, hepatic, complex biliary, or pancreatic operations other than PD) and dichotomized as low- and high-volume centers based on the median. Three study cohorts were created: low-volume hospitals (LVH) for both PD and PAO, mixed-volume hospital (MVH) with low-volume PD but high-volume PAO, and high-volume PD hospital (HVH). RESULTS: In total, 24,572 patients were identified, with 41.5%, 7.2%, and 51.3% patients treated at LVH, MVH, and HVH, respectively. Thirty-day mortality for PD was 5.6% in LVH, 3.2% in MVH, and 2.5% in HVH. On multivariable analyses, LVH was predictive for higher 30-day mortality compared to HVH [odds ratio (OR) 2.068; 95% confidence interval (CI) 1.770-2.418; P< 0.0001]. However, patients at MVH demonstrated similar 30-day mortality to patients treated at HVH (OR 1.258; 95% CI 0.942-1.680; P = 0.1203). CONCLUSIONS: PD outcomes at low-volume centers that have experience with complex cancer operations near the pancreas are similar to PD outcomes at hospitals with high PD volume. MVH provides a model for PD outcomes to improve quality and access for patients who cannot, or choose not to, receive their care at high-volume centers.

Impact of Patient Primary Language upon Immediate Breast Reconstruction After Mastectomy.

BACKGROUND: Preoperative decision-making in patients who speak a primary language other than English is understudied. We investigated whether patient primary language is associated with differences in immediate breast reconstruction (IBR) after mastectomy. PATIENTS AND METHODS: This retrospective observational study analyzed female patients undergoing mastectomy in the New Jersey State Inpatient Database (2009-2014). The primary outcome was the odds of IBR with a prespecified subanalysis of autologous tissue-based IBR. We used multivariable logistic regression and hierarchical generalized linear mixed models to control for patient characteristics and nesting within hospitals. RESULTS: Of 13,846 discharges, 12,924 (93.3%) specified English as the patient's primary language, while 922 (6.7%) specified a language other than English. Among English-speaking patients, 6178 (47.8%) underwent IBR, including 2310 (17.9%) autologous reconstructions. Among patients with a primary language other than English, 339 (36.8%) underwent IBR, including 93 (10.1%) autologous reconstructions. Unadjusted results showed reduced odds of IBR overall [odds ratio (OR) 0.64, 95% CI 0.55-0.73], and autologous reconstruction specifically (OR 0.52, 95% CI 0.41-0.64) among patients with a primary language other than English. After adjustment for patient factors, this difference persisted among the autologous subgroup (OR 0.64, 95% CI 0.51-0.80) but not for IBR overall. A hierarchical model incorporating both patient characteristics and hospital-level effects continued to show a difference among the autologous subgroup (OR 0.75, 95% CI 0.58-0.97). CONCLUSIONS: Primary language other than English was an independent risk factor for lower odds of autologous IBR after adjustments for patient and hospital effects. Focused efforts should be made to ensure that patients who speak a primary language other than English have access to high-quality shared decision-making for postmastectomy IBR.

Surveillance Patterns for Hepatocellular Carcinoma among Screening-Eligible Patients in the Medicare Population.

INTRODUCTION: Routine screening plays a critical role in the diagnosis of hepatocellular carcinoma (HCC), but not all patients undergo consistent surveillance. This study aims to evaluate surveillance patterns and their association with diagnosis stage and survival among Medicare patients at risk for HCC. PATIENTS AND METHODS: Patients with HCC and guideline-based screening eligibility who underwent imaging with ultrasound or abdominal magnetic resonance imaging (MRI) in the 2 years prior to diagnosis were identified from SEER-Medicare (2008-2015). Three surveillance cohorts were created: diagnostic (imaging only within 3 months prior), intermittent (imaging only once within 2 years prior, excluding diagnostic), and routine (at least two imaging encounters within 2 years prior, excluding diagnostic). Multivariable logistic regression was used to predict early-stage diagnosis (stage I-II), and 5-year survival was evaluated using the accelerated failure time method with Weibull distribution. RESULTS: Among 2261 eligible patients, 26.1% were classified as diagnostic, 15.8% as intermittent, and 58.1% as routine surveillance. The median age was 74 years (IQR 70-78 years). The majority of patients had a preexisting cirrhosis diagnosis (81.5%). Routine and intermittent, compared with diagnostic, surveillance were predictive of early-stage disease (routine: OR 2.05, 95% CI 1.64-2.56; intermittent: OR 1.43, 95% CI 1.07-1.90). Patients who underwent routine surveillance had significantly lower risk of mortality (HR 0.84, 95% CI 0.75-0.94) compared with the diagnostic group. CONCLUSIONS: A large proportion of screening-eligible patients do not undergo routine surveillance, which is associated with late-stage diagnosis and higher risk of mortality. These findings demonstrate the impact of timely and consistent healthcare access and can guide interventions for promoting surveillance among these patients.

The Effect of Hospital Versus Surgeon Volume on Short-Term Patient Outcomes After Pancreaticoduodenectomy: a SEER-Medicare Analysis.

BACKGROUND: The volume-outcome relationship has been well-established for pancreaticoduodenectomy (PD). It remains unclear if this is primarily driven by hospital volume or individual surgeon experience. OBJECTIVE: This study aimed to determine the relationship of hospital and surgeon volume on short-term outcomes of patients with pancreatic adenocarcinoma undergoing PD. METHODS: Patients >65 years of age who underwent PD for pancreatic adenocarcinoma were identified from the Surveillance, Epidemiology, and End Results (SEER)-Medicare database (2008-2015). Analyses were stratified by hospital volume and then surgeon volume, creating four volume cohorts: low-low (low hospital, low surgeon), low-high (low hospital, high surgeon), high-low (high hospital, low surgeon), high-high (high hospital, high surgeon). Propensity scores were created for the odds of undergoing surgery with high-volume surgeons. Following matching, multivariable analysis was used to assess the impact of surgeon volume on outcomes within each hospital volume cohort. RESULTS: In total, 2450 patients were identified: 54.3% were treated at high-volume hospitals (27.0% low-volume surgeons, 73.0% high-volume surgeons) and 45.7% were treated at low-volume hospitals (76.9% low-volume surgeons, 23.1% high-volume surgeons). On matched multivariable analysis, there were no significant differences in the risk of major complications, 90-day mortality, and 30-day readmission based on surgeon volume within the low and high hospital volume cohorts. CONCLUSION: Compared with surgeon volume, hospital volume is a more significant factor in predicting short-term outcomes after PD. This suggests that a focus on resources and care pathways, in combination with volume metrics, is more likely to achieve high-quality care for patients undergoing PD across all hospitals.

Impact of fellow compared to resident assistance on outcomes of minimally invasive surgery.

INTRODUCTION: As fellowship training after general surgery residency has become increasingly common, the impact on resident education must be considered. Patient safety and procedure outcomes are often used as justification by attendings who favor fellows over residents in certain minimally invasive surgery (MIS) operations. The aim of the present study was to compare the impact of trainee level on the outcomes of selected MIS operations to determine if giving preference to fellows on grounds of outcomes is warranted. METHODS: Patients who underwent elective laparoscopic hiatal hernia repair (LHHR), laparoscopic sleeve gastrectomy (LSG), laparoscopic Roux-en-Y gastric bypass (LRYGB), laparoscopic splenectomy (LS), laparoscopic cholecystectomy (LC), or laparoscopic ventral hernia repair (LVHR) with assistance of a general surgery chief resident or fellow were identified from the American College of Surgeon's National Surgical Quality Improvement Program database (2007-2012). Patients were matched 1:1 based on propensity score for the odds of undergoing operations assisted by a fellow. RESULTS: 5145 patients underwent LHHR, 1396 LSG, 9656 LRYGB, 863 LS, 13,434 LC, and 3069 LVHR. Fellows assisted in 41.7% of LHHR, 49.2% of LSG, 56.4% of LRYGB, 25.7% of LS, 17.1% of LC, and 27.0% of LVHR cases. After matching, overall and severe complication rates were comparable between cases performed with assistance of a fellow or chief resident. Median operative time was longer for LSG, LRYGB, and LC when a fellow assisted. CONCLUSIONS: Surgical outcomes were similar between fellow and chief resident assistance in MIS operations, arguing that increased resident participation in basic and complex laparoscopic operations is appropriate without compromising patient safety.

Beyond insurance status: the impact of Medicaid expansion on the diagnosis of Hepatocellular Carcinoma.

BACKGROUND: Medicaid expansion has led to earlier stage diagnoses in several cancers but has not been studied in hepatocellular carcinoma (HCC), a disease with complex risk factors. We examined the effect of Medicaid expansion on the diagnosis of HCC and associations with county-level social vulnerability. METHODS: Patients with HCC <65 years of age were identified from the SEER database (2010-2016). County-level social vulnerability factors were obtained from the CDC SVI and BRFSS. A Difference-in-Difference analysis evaluated change in early-stage diagnoses (stage I-II) between expansion and non-expansion states. A Difference-in-Difference-in-Difference analysis evaluated expansion impact among counties with higher proportions of social vulnerability. RESULTS: Of 19,751 patients identified, 81.5% were in expansion states. Uninsured status decreased in expansion states (6.3%-2.4%, p < 0.0001) and remained unchanged in non-expansion states (12.7%-14.8%, p = 0.43). There was no significant difference in the incidence of early-stage diagnoses between expansion states and non-expansion states. Results were consistent when accounting for social vulnerability. CONCLUSION: Medicaid expansion was not associated with earlier stage diagnoses in patients with HCC, including those with higher social vulnerability. Unlike other cancers, expanded access did not translate into higher utilization of care in HCC, suggesting barriers on a multitude of levels.

The impact of upper gastrointestinal surgical volume on short term pancreaticoduodenectomy outcomes for pancreatic adenocarcinoma in the SEER-Medicare population.

BACKGROUND: Patients undergoing pancreaticoduodenectomy (PD) at low volume PD hospitals with high volume for other complex operations have comparable outcomes to high volume PD centers. We evaluated the impact of upper gastrointestinal operations (UGI) hospital volume on the outcomes of elderly, high risk patients undergoing PD. METHODS: Patients >65 years old who underwent PD for pancreatic adenocarcinoma were identified from SEER-Medicare (2008-2015). Four volume cohorts were created using PD tertiles and UGI median: low (1st tertile PD), mixed-low (2nd tertile PD, low UGI), mixed-high (2nd tertile PD, high UGI) and high (3rd tertile PD). Multivariable logistic and negative binomial regression assessed short-term complications. RESULTS: In total, 2717 patients were identified with a median age of 74.5 years. Patients treated at low, mixed-low and mixed-high volume hospitals, versus high volume, had higher risk of short-term complications, including major complications (low: OR 1.441, 95%CI 1.165-1.783; mixed-low: OR 1.374, 95%CI 1.085-1.740; mixed-high: OR 1.418, 95%CI 1.098-1.832) and 90-day mortality (low: OR 2.16, 95%CI 1.454-3.209; mixed-low: OR 2.068, 95%CI 1.347-3.175; mixed-high: OR 1.96, 95%CI 1.245-3.086). CONCLUSION: Patients with pancreatic adenocarcinoma who are older and more medically complex benefit from undergoing surgery at high volume PD centers, independent of the operative experience of that center.

Undertreatment of Gallbladder Cancer: A Nationwide Analysis.

BACKGROUND: Gallbladder cancer has a high mortality rate and an increasing incidence. The current National Comprehensive Cancer Network (NCCN) guidelines recommend resection for all T1b and higher-stage cancers. This study aimed to evaluate re-resection rates and the associated survival impact for patients with gallbladder cancer. METHODS: Patients with gallbladder adenocarcinoma who underwent resection were identified from the National Cancer Database (2004-2015). Re-resection was defined as definitive surgery within 180 days after the first operation. Propensity scores were created for the odds of a patient having a re-resection. Patients were matched 1:2. Survival analyses were performed using the Kaplan-Meier and Cox proportional hazard methods. RESULTS: The study identified 6175 patients, and 466 of these patients (7.6%) underwent re-resection. Re-resection was associated with younger median age (65 vs 72 years; p < 0.0001), private insurance (41.6% vs 27.1%; p < 0.0001), academic centers (50.4% vs 29.7%; p < 0.0001), and treatment location in the Northeast (22.8% vs 20.4%; p = 0.0011). Compared with no re-resection, re-resection was associated with pT stage (pT2: 47.6% vs 42.8%; p = 0.0139) and pN stage (pN1-2: 28.1% vs 20.7%; p < 0.0001), negative margins on final pathology (90.1% vs 72.6%; p < 0.0001), and receipt of chemotherapy (53.7% vs 35.8%; p < 0.0001). The patients who underwent re-resection demonstrated significantly longer overall survival (OS) than the patients who did not undergo re-resection (median OS, 44.0 vs 23.0 months; p < 0.0001). After propensity score-matching, re-resection remained associated with superior survival (median OS, 44.0 vs 31.0 months; p = 0.0004). CONCLUSIONS: Re-resection for gallbladder cancer is associated with improved survival but remains underused, particularly for early-stage disease.

Development and validation of machine learning models to predict gastrointestinal leak and venous thromboembolism after weight loss surgery: an analysis of the MBSAQIP database.

BACKGROUND: Postoperative gastrointestinal leak and venous thromboembolism (VTE) are devastating complications of bariatric surgery. The performance of currently available predictive models for these complications remains wanting, while machine learning has shown promise to improve on traditional modeling approaches. The purpose of this study was to compare the ability of two machine learning strategies, artificial neural networks (ANNs), and gradient boosting machines (XGBs) to conventional models using logistic regression (LR) in predicting leak and VTE after bariatric surgery. METHODS: ANN, XGB, and LR prediction models for leak and VTE among adults undergoing initial elective weight loss surgery were trained and validated using preoperative data from 2015 to 2017 from Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program database. Data were randomly split into training, validation, and testing populations. Model performance was measured by the area under the receiver operating characteristic curve (AUC) on the testing data for each model. RESULTS: The study cohort contained 436,807 patients. The incidences of leak and VTE were 0.70% and 0.46%. ANN (AUC 0.75, 95% CI 0.73-0.78) was the best-performing model for predicting leak, followed by XGB (AUC 0.70, 95% CI 0.68-0.72) and then LR (AUC 0.63, 95% CI 0.61-0.65, p < 0.001 for all comparisons). In detecting VTE, ANN, and XGB, LR achieved similar AUCs of 0.65 (95% CI 0.63-0.68), 0.67 (95% CI 0.64-0.70), and 0.64 (95% CI 0.61-0.66), respectively; the performance difference between XGB and LR was statistically significant (p = 0.001). CONCLUSIONS: ANN and XGB outperformed traditional LR in predicting leak. These results suggest that ML has the potential to improve risk stratification for bariatric surgery, especially as techniques to extract more granular data from medical records improve. Further studies investigating the merits of machine learning to improve patient selection and risk management in bariatric surgery are warranted.

Contemporary Analysis of Senior Level Case Volume Variation between Traditional Vascular Surgery Fellows and Integrated Vascular Surgery Chief Residents.

BACKGROUND: The present study compares the senior level operative experience of graduates from the traditional vascular surgery fellowship (5 + 2) and integrated vascular surgery training programs (0 + 5) using contemporary operative case log data. METHODS: The Accreditation Council for Graduate Medical Education integrated vascular surgery, vascular surgery fellowship, and general surgery case logs for trainees graduating between 2013 and 2018 were queried for vascular surgery procedures. "Senior" cases were categorized as cases logged as "surgeon fellow" by 5 + 2 trainees or "surgeon chief" (post graduate year-4,5) by 0 + 5 trainees. Overall case volume was defined as the combined volume of cases logged as "surgeon junior," "surgeon chief," "surgeon fellow," "teach assist," "first assist," or "secondary procedure." To reflect total vascular experience, all vascular cases done during general surgery residency were combined with cases performed during vascular surgery fellowship. Mean case volumes were compared for all operations/procedures. RESULTS: The 5 + 2 trainees had higher mean volume of open repair of suprarenal aortic aneurysms (2.4 vs. 1.4, P = 0.0026) and open repair of thoracic aortic aneurysms (0.5 vs. 0.3, P = 0.004) at the fellow level compared to 0 + 5 surgeon chief cases. Additionally, 5 + 2 trainees performed more endovascular repair of abdominal aortoiliac aneurysm (44.7 vs. 28.4, P < 0.0001), endovascular repair of iliac artery aneurysm (1.9 vs. 1.2, P = 0.0003), and endovascular repair of thoracic aortic aneurysm (14.9 vs. 8.4, P < 0.0001). The 5 + 2 fellows performed more vein bypasses than 0 + 5 chief residents (femoral-popliteal 9.8 vs. 6.4, P = 0.002; infrapopliteal 13.9 vs. 8.8, P = 0.0490), extra-anatomic bypasses (axillofemoral 4.2 vs. 2.9, P = 0.0004; femoral-femoral 5.6 vs. 3.1, P = 0.034), carotid endarterectomies (47.3 vs. 29.3, P < 0.0001), carotid artery stenting (9.6 vs. 4.5, P = 0.0001), celiac/SMA endarterectomy or bypass (3.7 vs. 1.9, P < 0.0001), renal artery balloon angioplasty/stenting (5.0 vs. 2.5, P = 0.0006), thoracic outlet decompression (5.4 vs. 1.9, P < 0.0001), traumatic repairs [thoracic vessels (0.5 vs. 0.1, P < 0.0001), neck vessels (0.7 vs. 0.3, P = 0.0004), abdominal vessels (3.0 vs. 1.7, P = 0.0005), and peripheral vessels (6.6 vs. 3.1, P = 0.034)], as well as a higher mean volume of arteriovenous (AV) fistulas (30.7 vs. 15.7, P < 0.0001), AV grafts (10.7 vs. 5.1, P < 0.0001), and revision of AV access (16.1 vs. 8.0, P = 0.0003). CONCLUSIONS: Although both pathways graduate trainees with a similar overall surgical experience, 5 + 2 trainees log significantly more "Senior" cases. Further studies investigating potential variation in operative autonomy between both pathways are necessary.

Vascular repairs in gynecologic operations are uncommon but predict major morbidity and mortality.

OBJECTIVE: Gynecologic surgery has potential for adjunct vascular interventions, given the proximity of major intra-abdominal and pelvic blood vessels. Our goal was to determine contemporary incidence, associations, and outcomes of vascular repairs in gynecologic operations. METHODS: The American College of Surgeons National Surgical Quality Improvement Program database (2005-2017) was queried for patients undergoing elective gynecologic operations. Vascular repairs were performed concurrently or during reoperation. Univariable and multivariable analyses evaluated associations with vascular repairs and 30-day morbidity. RESULTS: A total of 201,224 gynecologic operations were identified: hysterectomy (88.3%), myomectomy (5.9%), adnexal surgery (3.5%), vulvovaginectomy/other (1.1%), nonadnexal tumor or cyst excision (0.5%), ectopic pregnancy treatment (0.4%), and pelvic lymphadenectomy (0.3%). There were 187 vascular repairs in 176 (0.09%) patients. Repairs were typically concurrent (89.8%) and most commonly included open abdominal blood vessel repair (51.8%), major abdominal artery ligation (25%), vena cava reconstruction/ligation (6%), common iliac vein ligation (4.2%), and aorta/great vessel repair (4.2%). A minority were performed endovascularly (1.7%). Patients undergoing vascular repairs were older (56 vs 46 years), were more likely to have an open vs minimally invasive/vaginal operation (71.6% vs 28.4%), and were more likely to have a hysterectomy (85.2%; P < .001 for all). In multivariable analysis, vascular repairs were observed more often with hysterectomy (odds ratio [OR]; 7.63, 95% confidence interval [CI], 2.28-25.55; P = .001) and open vs minimally invasive/vaginal operations (OR, 5.24; 95% CI, 2.64-10.42; P < .001). Vascular repairs were also more common for patients with malignant disease (OR, 2.84; 95% CI, 1.78-4.53; P < .001), patients assigned to American Society of Anesthesiologists class 3 or class 4 (OR, 1.85; CI, 1.36-2.53; P = .002), and patients without obesity (OR, 1.45; 95% CI, 1.08-1.96; P = .014). Vascular repairs independently predicted major morbidity and mortality (OR, 7.26; 95% CI, 5.26-10.03; P < .001) after adjustment for open operative approach, American Society of Anesthesiologists class 3 or class 4, and hysterectomy. CONCLUSIONS: Whereas vascular repairs during gynecologic operations are rare, they are associated with morbidity and mortality. These findings provide an evidence base for risk assessment and informed consent.