A Comprehensive Analysis of Metastatic Disease following Surgery for Clinically Localized Cutaneous Melanoma.

INTRODUCTION: The NCCN considers "baseline staging" (whole body CT or PET scan +/- brain MRI) for all asymptomatic melanoma patients with a positive sentinel lymph node biopsy. The true yield of these workups is unknown. METHODS: We created cohorts of adult malignant melanoma patients, using the National Cancer Database (2012-2020) to mimic three common scenarios: (1) clinically node negative, with positive sentinel lymph node(s) (SLNB[+]); (2) clinically node negative, with negative sentinel lymph node(s) (SLNB[-]); (3) clinically node positive with confirmed lymph node metastases (cN[+] and pN[+]). Multivariable regression, supervised decision trees, and nomograms were constructed to assess the risk of metastases based on key features. RESULTS: 10,371 patients were SLNB[+], 55,172 were SLNB[-], and 4,012 were cN[+] and pN[+]. The proportion of patients with any metastatic disease (brain metastases) were as follows: SLNB[+]: 1.4% (0.3%); SLNB[-] 0.3% (<0.1%); cN[+] and pN[+] 11.6% (1.6%). On multivariable regression, Breslow depth > 4, ulceration, and lymphovascular invasion were associated with greater risk of metastatic disease. A supervised decision tree for SLNB[+] and SLNB[-] patients found the only groups with >2% risk of metastases were T4 tumors or T2/T3 tumors with ulceration and LVI. Most groups had a negligible risk (<0.1%) of brain metastases. CONCLUSION: This is the first large analysis to guide the use of imaging for cutaneous melanoma. Among clinically node negative patients, metastatic disease is uncommon and brain metastases are exceedingly rare. Further investigation could promote a tailored approach to metastatic workups guided by individual risk factors.

Perfusion Strategies for Cytoreductive Surgery With Heated Intraperitoneal Chemotherapy for Malignant Peritoneal Mesothelioma.

Cytoreductive surgery (CRS) with heated intraoperative intraperitoneal chemotherapy (HIPEC) has been shown to improve survival for patients with malignant peritoneal mesothelioma (MPM). Presently, there is no standardized HIPEC protocol with respect to chemotherapeutic agent, dose, administration temperature, or duration and limited literature comparing outcomes in different regimens. In this study, we analyze common practices and outcomes of published HIPEC regimens to gain insight into current practice to inform future directions of study. We conducted a literature search for investigational studies of CRS and HIPEC for MPM treatment in adults and identified 35 such articles. These studies were analyzed for institution type and location, drug regimens, perfusion temperatures and time, and study outcomes including median survival, complication rates, and perioperative mortality rates. On review, there is significant heterogeneity in HIPEC regimens and outcome reporting metrics, suggesting a need for multi-institutional standardized study protocols to better determine the safest and most efficacious treatment regimen.

Reasons for Surgical Attrition Among Nonmetastatic Upper Gastrointestinal Cancer Patients: A Single Institutional Experience.

INTRODUCTION: Upper gastrointestinal (UGI) cancers require multidisciplinary treatment, but surgery provides the only potentially curative option. We sought to understand reasons for attrition before surgery within our regional hospital network. METHODS: We performed chart reviews of patients (age 18-80) with stage I-III UGI cancers (gastroesophageal junction, gastric, and hepatopancreatobiliary adenocarcinomas) in our multihospital cancer registry from 2015 to 2021. Our primary outcome was reasons for surgical attrition. Univariable analysis identified factors related to surgical attrition and the Kaplan-Meier method estimated overall survival based on surgery receipt. RESULTS: Seven hundred and ninety-two patients were included in our analysis, of whom 107 (13.5%) did not undergo curative surgery. Reasons for not undergoing surgery included medical comorbidities (30.8%), patient preference/nonmedical barriers (24.3%, which included: not interested without further explanation, worried about complications, nonadherence to appointments, insurance issues, did not wish for blood transfusion, lack of social support, preferring home care, and worried about recurrence), psychosocial (5.6%), progression while on neoadjuvant therapy or waiting for transplant (15.0% and 7.5%), poor performance status (3.7%), side effects of neoadjuvant therapy (3.7%), and death unrelated to treatment or unknown cause (9.4%). Nonsurgical management was not associated with race, socioeconomic status, or distance traveled for care. Survival was greatly improved for patients who underwent surgery (158 vs. 63 weeks, p < 0.05). CONCLUSION: Nearly one in seven patients (18-80 years old) with UGI cancers evaluated at our academic cancer center did not undergo surgical resection. Reasons for surgical attrition included potentially modifiable issues, and addressing these barriers could help overcome inequities in cancer treatment and survival.

The Utility of Sentinel Lymph Node Biopsy in Elderly Patients with Melanoma.

BACKGROUND: Sentinel lymph node biopsy (SLNB) is performed less often for older patients with melanoma. We investigated the association of SLNB and melanoma-specific survival (MSS) in the elderly. METHODS: We retrospectively reviewed the Surveillance, Epidemiology, and End Results (SEER: 2010-2019) for patients ≥ 70 years with cT2-4N0M0 melanoma. We used multivariable Cox proportional hazard models to evaluate the impact of SLNB performance and SLN status on MSS at increasing age cutoffs. In addition, we evaluated the association of different factors with SLNB performance using multivariable logistic regression. RESULTS: We identified 11,548 patients. Sentinel lymph node biopsy occurred in 6754 (58.5%) patients, 1050 (15.5%) of whom had a positive SLN. On adjusted SEER analysis, a negative SLN was independently associated with improved MSS (overall hazard ratio [HR] 0.59, 95% confidence interval [CI] 0.63-0.67) for patients up to 87 years old. Positive SLNB was independently associated with inferior MSS (HR 1.71, 95% CI 1.93-1.98). Increasing age groups were significantly associated with decreased SLNB performance. CONCLUSIONS: Sentinel lymph node biopsy is associated with cancer-specific survival and adds prognostic information for elderly patients with melanoma. Sentinel lymph node biopsy performance should not be eliminated in elderly patients based on age alone, unless justified by poor performance status, patient preference, or other surgical contraindications. Decreased SLNB performance with increasing age in our cohort may indicate a missed therapeutic opportunity in the care of elderly patients with melanoma.

An Age-Stratified Analysis of Early Postoperative Mortality Among Medicare-Eligible Patients With Common Gastrointestinal and Hepato-pancreato-biliary Malignancies.

Background: The frequency of major cancer surgery in the elderly (≥80 years) has increased concomitantly with the rise in average age of the population. We assessed early postoperative mortality following hepato-pancreato-biliary (HPB) and gastrointestinal (GI) procedures for common malignancies stratified by age. Methods: The National Cancer Database (2004-2017) was queried for patients who underwent resection for GI (gastroesophageal and colorectal) or HPB (pancreatic adenocarcinoma, biliary tract, and primary liver) cancers. We compared early postoperative mortality (30 d and 90 d) stratified by age (65-79 vs ≥80 years) and procedure, and compared survival outcomes by age and operative vs nonoperative management. Results: A total of 709,358 patients were included. The 30-day mortality ranged from 1.8% to 5.8% among patients 65-79 years and from 3.2% to 12.4% among patients ≥80 years depending on procedure. The 90-day mortality ranged from 3.6% to 10.6% in patients 65-79 years compared to 8.4%-21.0% among patients ≥80 years. The overall 90-day mortality was 5.2% for patients 65-79 years and 12.0% for patients ≥80 years (P < .001). Age ≥80 was associated with worse survival among operatively managed patients with each upper GI, HPB, and lower GI malignancy relative to younger patients on multivariable analysis. However, operative management of patients ≥80 years was associated with improved survival relative to nonoperative management. Discussion: Elderly patients suffer higher postoperative mortality after major GI and HPB cancer surgery, but operative management is associated with improved survival among patients ≥80 years as compared to nonoperative management. These data are important to contextualize when counseling elderly patients on their treatment options for localized GI and HPB cancers.

To Revise or Not Revise? Isolated Margin Positivity in Localized Pancreatic Ductal Adenocarcinoma.

BACKGROUND: The study determined the proportion of patients with pancreatic adenocarcinoma (PDAC) who had margin-positive disease and no other adverse pathologic findings (APF) using institutional and administrative datasets. METHODS: Patients with clinical stage I or II PDAC in the National Cancer Database (NCDB 2010-2020) and those who underwent pancreatectomy at the authors' institution (2010-2021) were identified. Isolated margin positivity (IMP) was defined as a positive surgical margin with no APF (negative nodes, no lymphovascular/perineural invasion). RESULTS: The study included 225 patients from the authors' institution and 23,598 patients from the NCDB. The margin-positive rates were 21.8% and 20.3%, and the IMP rates were 0.4% and 0.5%, respectively. In the institutional cohort, 68.4% of the patients had recurrence, and most of the patients (65.6%) had distant recurrences. The median recurrence-free survival (RFS) was 63.3 months for no APF, not reached for IMP, 14.8 months for negative margins & 1 APF, 20.3 months for positive margins & 2 APFs, and 12.9 months with all APF positive. The patients in the NCDB with IMP had a lower median OS than the patients with no APF (20.5 vs 390 months), but a higher median OS than those with margin positivity plus 1 APF (20.5 vs 18.0 months) or all those with APF positivity (20.5 vs 15.4 months). Based on institutional rates of IMP, any margin positivity, neck margin positivity (NMP), and no APF, the fraction of patients who might benefit from neck margin revision was 1 in 100,000, and those likely to benefit from any margin revision was 1 in 18,500. In the NCDB, those estimated to derive potential benefit from margin revision was 1 in 25,000. CONCLUSIONS: Isolated margin positivity in resected PDAC is rare, and most patients experience distant recurrence. Revision of IMP appears unlikely to confer benefit to most patients.

Comparison of Hospital Volume and Risk-Standardized Mortality Rate as a Proxy for Hospital Quality in Complex Oncologic Hepatopancreatobiliary Surgery.

BACKGROUND: Centralization of hepatopancreatobiliary procedures to more experienced centers has been recommended but remains controversial. Hospital volume and risk-stratified mortality rates (RSMR) are metrics for interhospital comparison. We compared facility operative volume with facility RSMR as a proxy for hospital quality. PATIENTS AND METHODS: Patients who underwent surgery for liver (LC), biliary tract (BTC), and pancreatic (PDAC) cancer were identified in the National Cancer Database (2004-2018). Hierarchical logistic regression was used to create facility-specific models for RSMR. Volume (high versus low) was determined by quintile. Performance (high versus low) was determined by RSMR tercile. Primary outcomes included median facility RSMR and RSMR distributions. Volume- and RSMR-based redistribution was simulated and compared for reductions in 90-day mortality. RESULTS: A total of 106,217 patients treated at 1282 facilities were included; 17,695 had LC, 23,075 had BTC, and 65,447 had PDAC. High-volume centers (HVC) had lower RSMR compared with medium-volume centers and low-volume centers for LC, BTC, and PDAC (all p < 0.001). High-performance centers (HPC) had lower RSMR compared with medium-performance centers and low-performance centers for LC, BTC, and PDAC (all p < 0.001). Volume-based redistribution required 16.0 patients for LC, 11.2 for BTC, and 14.9 for PDAC reassigned to 15, 22, and 20 centers, respectively, per life saved within each US census region. RSMR-based redistribution required 4.7 patients for LC, 4.2 for BTC, and 4.9 for PDAC reassigned to 316, 403, and 418 centers, respectively, per life saved within each US census region. CONCLUSIONS: HVC and HPC have the lowest overall and risk-standardized 90-day mortality after oncologic hepatopancreatobiliary procedures, but RSMR may outperform volume as a measure of hospital quality.

Creation and Implementation of an Online Tool for Feedback on Resident Teaching: A Pilot Study.

OBJECTIVE: There are few assessments of the competence and growth of surgical residents as educators. We developed and piloted an observation-based feedback tool (FT) to provide residents direct feedback during a specific teaching session, as perceived by medical students (MS). We hypothesized that residents' performance would improve with frequent, low stakes, observation-based feedback. SETTING: This prospective study took place at an academic general surgery program. PARTICIPANTS: Focus groups of MS, surgical residents, and faculty informed FT development. MS completed the FT regarding resident teaching. DESIGN: The FT utilized 5 slider-bar ratings (0 to 100) about the teaching encounter and a checklist of 16 desirable teaching behaviors. QR codes and weekly email links were distributed for 12 months (6 clerkship blocks) to promote use. Residents were sent their results after each block. A survey after each block assessed motivation for use and gathered feedback on the FT. Descriptive statistics were used for analysis (medians, IQRs). Primary measures of performance were median of the slider-bar scores and the number of teaching behaviors. RESULTS: The FT was used 111 times; 37 of 46 residents were rated by up to 65 MS. The median rating on the slider-bars was 100 and the median number of desirable teaching behaviors was 12; there were no differences based on gender or PGY level. 10 residents had 5 or more FT observations during the year. Four residents had evaluations completed in 4 or more blocks and 19 residents had evaluations completed in at least 2 blocks. Over time, 13 residents had consistent slider-bar scores, 1 resident had higher scores, and 5 residents had lower scores (defined as a more than 5-point change from initial rating). Frequency of use of the FT decreased over time (38, 32, 9, 21, 7, 5 uses per block). The post-use survey was completed by 24 MS and 19 residents. Most common reasons for usage were interest in improving surgical learning environment, giving positive feedback (MS), and improving teaching skills (residents). Most common reasons for lack of usage from residents were "I did not think I taught enough to ask for feedback," "I forgot it existed," and "I did not know it existed." CONCLUSIONS: The FT did not lead to any meaningful improvement in resident scores over the course of the year. This may be due to overall high scores, suggesting that the components of the FT may require reevaluation. Additionally, decreased utilization of the instrument over time made it challenging to assess change in performance of specific residents, likely due to lack of awareness of the FT despite frequent reminders. Successful implementation of observation-based teaching assessments may require better integration with residency or clerkship objectives.

First-line Immunotherapy for Metastatic Merkel Cell Carcinoma: Analysis of Real-world Survival Data and Practice Patterns.

OBJECTIVES: Immune checkpoint inhibitors are a promising new therapy for advanced Merkel Cell Carcinoma (MCC). We investigated real-world utilization and survival outcomes of first-line immunotherapies in a contemporary cohort. METHODS: Using the National Cancer Database (NCDB), we identified 759 patients with MCC between 2015 and 2020 with stage IV disease and known status of first-line systemic therapy. Univariable and multivariable analyses were used to determine predictors of immunotherapy usage. Overall survival (OS) was compared for patients receiving immunotherapy, chemotherapy, or no systemic therapies. RESULTS: We identified 759 patients meeting our inclusion criteria: 329 patients received immunotherapy, 161 received chemotherapy, and 269 received no systemic therapy. Adjusting for demographic, clinical, and facility factors, high facility volume significantly predicted first-line immunotherapy use (OR 1.99; P =0.017). Median OS was 16.2, 12.3, and 8.7 months, among patients who received immunotherapy, chemotherapy, or no systemic therapy, respectively ( P <0.001). On Cox multivariable survival analysis, first-line immunotherapy treatment (HR=0.79, P =0.041) and treatment at high-volume centers (HR=0.58, P =0.004) were associated with improved OS. CONCLUSIONS: Consistent with clinical trial results, first-line immunotherapy associated with improvement in median overall survival for patients with stage IV MCC, significantly outperforming chemotherapy in this real-world cohort. Treatment at high-volume centers associated with first-line immunotherapy utilization suggesting that familiarity with this rare disease is important to achieving optimal outcomes for metastatic MCC.

Outcomes of Hepatic Artery-Based Therapies and Systemic Multiagent Chemotherapy in Unresectable Colorectal Liver Metastases: A Systematic Review and Meta-analysis.

BACKGROUND: Treatment of unresectable colorectal liver metastases (UCRLM) includes locoregional and systemic therapy. A comprehensive analysis capturing long-term outcomes of these treatment options has not been performed. OBJECTIVE: A systematic review and meta-analysis was performed to calculate pooled outcomes of hepatic artery infusion with systemic chemotherapy (HAI-S), transarterial chemoembolization with systemic chemotherapy (TACE-S), transarterial radioembolization with systemic chemotherapy (TARE-S), doublet (FOLFOX, FOLFIRI), and triplet chemotherapy (FOLFOXIRI). METHODS: Outcomes included overall survival (OS), progression-free survival (PFS), rate of conversion to resection (CTR), and response rate (RR). RESULTS: A total of 32, 7, 9, and 14 publications were included in the HAI-S, TACE-S, and TARE-S chemotherapy arms. The 6/12/24/36-month OS estimates for HAI-S, TACE-S, TARE-S, FOLFOX, FOLFIRI, and FOLFOXIRI were 97%/80%/54%/35%, 100%/83%/40%/14%, 82%/61%/34%/21%, 96%/83%/53%/36%, and 96%/93%/72%/55%. Similarly, the 6/12/24/36-month PFS estimates were 74%/44%/19%/14%, 66%/20%/9%/3%, 57%/23%/10%/3%, 69%/30%/12%/7%, and 88%/55%/18%/11%. The corresponding CTR and RR rates were 31, 20%, unmeasurable (TARE-S), 35, 53; and 49, 45, 45, 50, 80%, respectively. The majority of chemotherapy studies included first-line therapy and liver-only metastases, whereas most HAI-S studies were pretreated. On subgroup analysis in first-line setting with liver-only metastases, the HAI-S arm had comparable outcomes to FOLFOXIRI and outperformed doublet chemotherapy regimens. Although triplet chemotherapy appeared to outperform other arms, high toxicity and inclusion of potentially resectable patients must be considered while interpreting results. CONCLUSIONS: HAI-S and multiagent chemotherapy are effective therapies for UCRLM. To make definitive conclusions, a randomized trial with comparable patient characteristics and line of therapy will be required. The upcoming EA2222 PUMP trial may help to address this question.

Risk factors of developing nonalcoholic fatty liver disease after pancreatic resection: a systematic review and meta-analysis.

BACKGROUND: Nonalcoholic fatty liver disease (NAFLD) occurs in 10% to 40% of patients after pancreatic resection. Pancreatic exocrine insufficiency (PEI) is thought to be closely associated with NAFLD; however, the mechanism of NAFLD is not clearly understood. We perform a systematic review and meta-analysis to better understand the risk factors of NAFLD. METHODS: A systematic literature search was performed in the MEDLINE database. Studies focused on the risk factors associated with NAFLD in patients undergoing pancreatectomy. The odds ratios (ORs) denoting the association of risk factors with NAFLD after resection were curated. RESULTS: Of 814 published articles, 26 studies met the inclusion criteria. Combined, these studies included clinical data on 4055 patients. The pooled incidence of NAFLD was 29% (23%-35%). Among the various risk factors analyzed, the following had a significant likelihood of NAFLD on forest plot analysis: female gender (OR, 2.44), pancreatic ductal adenocarcinoma (OR, 2.11), portal vein or superior mesenteric vein resection (OR, 1.99), dissection of nerve plexus around the superior mesenteric artery (OR, 1.93), and adjuvant chemotherapy (OR, 1.58). Only 2 studies investigated 2 different measurements of quantitative PEI, which could not be used for analysis. Owing to heterogeneity of studies, pancreatic remanent volume, which is considered a marker for PEI could not be evaluated. Pancreatic enzyme replacement therapy (PERT) was not associated with NAFLD. CONCLUSION: Numerous factors are associated with NAFLD after pancreatectomy. Previous research shows that PEI may be associated with NAFLD; however, this could not be compared in our meta-analysis. Further research is required to study the role of PERT in NAFLD.

Preoperative therapy in melanoma: Evolving perspectives in clinical trials.

We reviewed phase II and III trials beginning after 2010 studying preoperative therapy in melanoma (61 trials). Compared to standard adjuvant treatment, neoadjuvant immune checkpoint inhibitors (ICIs) show improved outcomes with approximately 70-80% recurrence free survival at 2 years. Several biomarkers demonstrate predictive value for pathological response (higher PD-L1 expression) and survival (IFN-γ signatures, CD8 + cell density). A number of 'non-standard' treatment mechanisms are being studied in combination with ICI therapies such as TLR-9 agonists, and anti-LAG3 checkpoint inhibitors, which show promise for alternative therapy options in the neoadjuvant setting. Finally, trials for advanced unresectable melanomas show improved survival compared to definitive systemic treatment when upfront systemic therapies lead to resectability. To conclude, in the preoperative setting for melanoma, ICIs have potential to improve outcomes for patients, and will likely change the standard treatment approach for advanced resectable disease.

Understanding surgical attrition for "resectable" pancreatic cancer.

OBJECTIVES: We used a novel combined analysis to evaluate various factors associated with failure to undergo surgery in non-metastatic pancreatic cancer. METHODS: We identified rates of surgery and reasons for surgical attrition from clinical trials, which studied neoadjuvant therapy in resectable pancreatic cancer. Next, we queried the National Cancer Database (NCDB) for Stage I-III, T1-3 pancreatic adenocarcinoma patients. We investigated the rates and factors associated with the receipt of surgery. Finally, we evaluated variable importance predicting the receipt of surgery. RESULTS: In clinical trials, 25-30 % of patients did not undergo surgery, mostly due to disease progression. In the NCDB, the overall surgical rate was only 49 %, but increased to 67 % in a curated cohort meant to mirror clinical trial patients. Patients treated at low-volume institutions (OR = 0.64, 95 % CI: 0.61-0.67) and who were uninsured (OR = 0.56, 95 % CI: 0.52-0.62) and Medicaid-insured (OR = 0.67, 95 % CI: 0.64-0.71) were less likely to receive potentially curative surgery. CONCLUSION: We have identified a realistic target surgery rate of 70%-75 % in potentially-resectable pancreatic cancer. While attrition to pancreatic cancer surgery is mostly due to tumor biology, our study identified the most important non-medical barriers, such as facility volume and insurance, affecting pancreatic cancer surgery.

Extended antibiotic therapy is associated with a lower rate of clinically relevant postoperative pancreatic fistula after pancreatoduodenectomy in intermediate- and high-risk patients: A single-institution analysis.

BACKGROUND: Clinically relevant postoperative pancreatic fistula remains a common complication after pancreatoduodenectomy. The fistula risk score is a validated tool to predict the risk of clinically relevant postoperative pancreatic fistula. To mitigate complications, we have implemented an extended antibiotic pathway for patients at increased risk of clinically relevant postoperative pancreatic fistula (fistula risk score ≥3). We report outcomes after pancreatoduodenectomy in patients at increased risk for clinically relevant postoperative pancreatic fistula who received extended antibiotic therapy compared to those who received standard perioperative antibiotics (single dose before incision). METHODS: Single-institution analysis of 87 patients who underwent elective pancreatoduodenectomy (2018-2022) with soft gland texture and fistula risk score ≥3 and were treated with (n = 34) or without (n = 53) 10 days of broad-spectrum antibiotics (piperacillin/tazobactam converted to amoxicillin/clavulanic acid at discharge) after surgery. Associations between extended antibiotics and postoperative outcomes were analyzed. RESULTS: Baseline clinicodemographic factors were similar between cohorts. Patients who received extended antibiotics had shorter index days (6 vs 8 days, P = .004) and 90-day composite length of stay (8.5 vs 12 days, P = .018). Patients who received extended antibiotics had lower rates of clinically relevant postoperative pancreatic fistula (11.8% vs 37.7%; odds ratio = 0.17, 95% confidence interval: 0.04-0.68), wound infections (8.8% vs 30.2%; odds ratio = 0.08, 95% confidence interval: 0.01-0.50), organ space infections (14.7% vs 43.4%; odds ratio = 0.15, 95% confidence interval: 0.04-0.52), and image-guided drain placement (8.8% vs 34.0%; odds ratio = 0.15, 95% confidence interval: 0.04-0.62). There were no Clostridium difficile infections in the extended antibiotic group. CONCLUSION: Extended antibiotic therapy is associated with a lower rate of clinically relevant postoperative pancreatic fistula and associated complications after pancreatoduodenectomy in patients with a fistula risk score ≥3. These results form the basis of a randomized controlled trial (NCT05753735).

Surgical Resection Alone is Associated With Higher Long-Term Survival Than Multiagent Chemotherapy Alone for Patients With Localized Biliary Tract Cancers.

INTRODUCTION: We compared long-term survival of patients with localized biliary tract cancers (BTCs) treated with either surgical resection or multiagent chemotherapy. METHODS: Patients with localized BTC [gallbladder adenocarcinoma, extrahepatic cholangiocarcinoma, intrahepatic cholangiocarcinoma] were identified within the National Cancer Database (2010-2017). Piecewise-constant hazard modeling was used to estimate hazard ratios (HRs) at prespecified intervals: 0-30 d, 31-60 d, 61-90 d, and >90 d post-treatment. RESULTS: A total of 5988 patients with localized BTC were identified: 2697 (45.0%) received multiagent chemotherapy and 3291 (55.0%) underwent surgical resection. Patients with gallbladder adenocarcinoma or extrahepatic cholangiocarcinoma who were treated with surgical resection had an associated decline in overall survival (OS) as compared to those treated with multiagent chemotherapy within 0-30 d of treatment initiation (gallbladder adenocarcinoma [adjusted HR = 3.94, 95% confidence interval [CI]: 1.77-8.80]; extrahepatic cholangiocarcinoma [adjusted HR = 4.88, 95% CI: 2.76-8.61]). However, there was an associated improvement in OS for patients treated with surgical resection after 90 d from treatment initiation (gallbladder adenocarcinoma [adjusted HR = 0.36, 95% CI: 0.28-0.46]; extrahepatic cholangiocarcinoma [adjusted HR = 0.27, 95% CI: 0.24-0.32]). Among patients with intrahepatic cholangiocarcinoma, those who underwent surgical resection had an associated improvement in OS at 31-60 d (adjusted HR = 0.63, 95% CI: 0.40-0.99) and a further associated increase in OS at 61-90 d (adjusted HR = 0.34, 95% CI: 0.21-0.54) and after 90 d (HR = 0.23, 95% CI: 0.21-0.27) of treatment initiation. CONCLUSIONS: For patients with localized BTC, surgical resection alone is associated with improved long-term survival outcomes compared to multiagent chemotherapy alone.