Clinical and Pathologic Response to Neoadjuvant Immunotherapy in DNA Mismatch Repair Protein-Deficient Gastroesophageal Cancers.

BACKGROUND: Mismatch repair deficient (dMMR) gastroesophageal cancers (GEC) are a distinct subgroup. Among patients with locally advanced disease, previous trial data suggest a good response to neoadjuvant immune checkpoint inhibitors (nICI). PATIENTS AND METHODS: Since 2019, our institution has routinely performed MMR testing for new GEC cases. Patients diagnosed with GEC (2019-2024) were included in the study. Quantitative data are described as the median and interquartile range (IQR); qualitative data are described as quantities and percentages. RESULTS: A total of 24 patients with dMMR GEC were identified following implementation of routine immunohistochemical testing; 14 were potentially resectable with a median follow-up of 14 months (IQR 8-27). All patients underwent pre-treatment positron emission tomography (PET; median SUV 20.9). Among the 14 potentially resectable patients, 4 underwent immediate surgery, 10 were treated with nICI, and 5 underwent surgical resection to date. All regimens included PD-1 inhibitors, with 70% receiving pembrolizumab. Re-staging PET was performed in five patients; the median post-nICI SUV was 5.1 (range 4.7-6.3). All resected specimens had gross ulceration after nICI, but 60% (N = 3) had a pathologic complete response (pCR) following nICI; one patient had a near-complete response (nCR) and one patient had a partial response (pPR). Reduction in SUV was 75% and 82% in the pCR patients, 25% in the nCR patient, and 43% in the pPR patient. CONCLUSIONS: dMMR GECs are responsive to nICI in this limited experience, mirroring early clinical trial data. Given persistent metabolic activity and visible ulceration despite pCR, studies should continue to optimize tools for estimating post-nICI pCR in these patients.

Intra-operative pyloric BOTOX injection versus pyloric surgery for prevention of delayed gastric emptying after esophagectomy.

BACKGROUND: Delayed gastric emptying (DGE) is a common complication after esophagectomy. BOTOX injections and pyloric surgeries (PS), including pyloroplasty (PP) and pyloromyotomy (PM), are performed intraoperatively as prophylaxis against DGE. This study compares the effects of pyloric BOTOX injection and PS for preventing DGE post-esophagectomy. METHODS: We retrospectively reviewed Moffitt's IRB-approved database of 1364 esophagectomies, identifying 475 patients receiving BOTOX or PS during esophageal resection. PS was further divided into PP and PM. Demographics, clinical characteristics, and postoperative outcomes were compared using Chi-Square, Fisher's exact test, Wilcoxon rank-sum, and ANOVA. Propensity-score matching was performed between BOTOX and PP cohorts. RESULTS: 238 patients received BOTOX, 108 received PP, and 129 received PM. Most BOTOX patients underwent fully minimally invasive robotic Ivor-Lewis esophagectomy (81.1% vs 1.7%) while most PS patients underwent hybrid open/Robotic Ivor-Lewis esophagectomy (95.7% vs 13.0%). Anastomotic leak (p = 0.57) and pneumonia (p = 0.75) were comparable between groups. However, PS experienced lower DGE rates (15.9% vs 9.3%; p = 0.04) while BOTOX patients had less postoperative weight loss (9.7 vs 11.45 kg; p = 0.02). After separating PP from PM, leak (p = 0.72) and pneumonia (p = 0.07) rates remained similar. However, PP patients had the lowest DGE incidence (1.9% vs 15.7% vs 15.9%; p = < 0.001) and the highest bile reflux rates (2.8% vs 0% vs 0.4%; p = 0.04). Between matched cohorts of 91 patients, PP had lower DGE rates (18.7% vs 1.1%; p = < 0.001) and less weight loss (9.8 vs 11.4 kg; p = < 0.001). Other complications were comparable (all p > 0.05). BOTOX was consistently associated with shorter LOS compared to PS (all p = < 0.001). CONCLUSION: PP demonstrates lower rates of DGE in unmatched and matched analyses. Compared to BOTOX, PS is linked to reduced DGE rates. While BOTOX is associated with more favorable LOS, this may be attributable to difference in operative approach. PP improves DGE rates after esophagectomy without improving other postoperative complications.

Simultaneous Hepatic and Visceral Resection: Preoperative Risk Stratification and Implications on Return to Intended Oncologic Therapy.

PURPOSE: Sequence of therapies for synchronous liver metastasis (LM) is complex, with data supporting individualized approaches, although no guiding tools are currently available. We assessed the impact of simultaneous hepatic and visceral resections (SHVR) on textbook outcome (TO) and return to intended oncologic therapy (RIOT), and provide risk-stratification tools to guide individualized decision making and counseling. METHODS: Patients with synchronous LM undergoing hepatectomy ± SHVR were included (2015-2021). Primary and secondary outcomes were TO and RIOT (days), respectively. Using multivariable modeling, a risk score for TO was developed. Decision tree analysis using recursive partitioning was performed for hierarchical risk stratification. The associations between SHVR, TO, and RIOT were examined. RESULTS: Among 533 patients identified, 124 underwent SHVR. TO overall was 71.7%; 79.2% in the non-SHVR group and 46.8% in the SHVR group (p < 0.001). SHVR was the strongest predictor of non-TO (right colon/small bowel: odds ratio [OR] 4.63, 95% confidence interval [CI] 2.65-8.08; left colon/rectum: OR 6.09, 95% CI 2.59-14.3; stomach/pancreas: OR 6.69, 95% CI 1.46-30.7; multivisceral: OR 10.9, 95% CI 3.03-39.5). A composite score was developed yielding three risk strata for TO (score 0-2: 89% vs. score 3-5: 67% vs. score ≥ 6: 37%; p < 0.001). Decision tree analysis was congruent, identifying SHVR as the most important determinant of TO. In patients with colorectal LM, SHVR was associated with delayed time to RIOT (p = 0.004); the risk-stratification tool for TO was equally predictive of RIOT (p < 0.01). CONCLUSIONS: SHVR is associated with reduced likelihood of TO and in turn delayed RIOT. As SHVR is increasingly performed in order to consolidate cancer care, patient selection considering these different outcomes is critical.

Tumor Regression Grade and Overall Survival following Gastrectomy with Preoperative Therapy for Gastric Cancer.

BACKGROUND: Pre-/perioperative chemotherapy is well-established for management of locoregional gastric cancer (LRGC). The American Joint Committee on Cancer advocates histopathologic assessment of tumor regression grade (TRG) but does not endorse a specific schema. We sought to examine the prognostic value of the recently revised National Comprehensive Cancer Network (NCCN) definition of TRG specifying TRG0 as no disease in primary tumor or lymph nodes. PATIENTS AND METHODS: Patients with clinical-stage T2+/N+/M0 LRGC receiving preoperative chemotherapy and curative-intent gastrectomy were identified (2000-2020). TRG using the current NCCN definition was retrospectively assigned. Factors associated with TRG were examined using ordinal logistic regression and overall survival (OS) was assessed using the Kaplan-Meier method and Cox regression. RESULTS: Among 117 patients, the most common chemotherapy regimen was epirubicin, cisplatin, plus fluorouracil or capecitabine (ECF/ECX) (n = 48, 41%), followed by folinic acid, fluorouracil, and oxaliplatin (FOLFOX) (n = 30, 26%), and fluorouracil, leucovorin, oxaliplatin, plus docetaxel (FLOT) (n = 13, 11%). TRG3 was the most common histopathologic response (n = 68, 58%), followed by TRG2 (n = 25, 21%), TRG1 (n = 18, 15%), and, lastly, TRG0 (n = 6, 5.1%). The only preoperative factor independently associated with lower TRG was gastroesophageal junction tumor location (OR 0.24, p = 0.012). Higher TRG was independently associated with worse OS in a stepwise fashion (HR 1.49, p = 0.026). Posttreatment pathologic lymph node status was the strongest prognostic factor (HR 1.93, p = 0.026). Independent prognostic value of TRG and ypT stage could not be shown due to substantial overlap. CONCLUSIONS: TRG using the contemporary NCCN definition is associated with OS in LRGC. TRG0 is uncommon but with excellent prognosis. ypN status is the strongest prognostic factor and the revised NCCN definition acknowledging this is appropriate.

Benchmarks and Geographic Differences in Gallbladder Cancer Surgery: An International Multicenter Study.

BACKGROUND: High-quality surgery plays a central role in the delivery of excellent oncologic care. Benchmark values indicate the best achievable results. We aimed to define benchmark values for gallbladder cancer (GBC) surgery across an international population. PATIENTS AND METHODS: This study included consecutive patients with GBC who underwent curative-intent surgery during 2000-2021 at 13 centers, across seven countries and four continents. Patients operated on at high-volume centers without the need for vascular and/or bile duct reconstruction and without significant comorbidities were chosen as the benchmark group. RESULTS: Of 906 patients who underwent curative-intent GBC surgery during the study period, 245 (27%) were included in the benchmark group. These were predominantly women (n = 174, 71%) and had a median age of 64 years (interquartile range 57-70 years). In the benchmark group, 50 patients (20%) experienced complications within 90 days after surgery, with 20 patients (8%) developing major complications (Clavien-Dindo grade ≥ IIIa). Median length of postoperative hospital stay was 6 days (interquartile range 4-8 days). Benchmark values included ≥ 4 lymph nodes retrieved, estimated intraoperative blood loss ≤ 350 mL, perioperative blood transfusion rate ≤ 13%, operative time ≤ 332 min, length of hospital stay ≤ 8 days, R1 margin rate ≤ 7%, complication rate ≤ 22%, and rate of grade ≥ IIIa complications ≤ 11%. CONCLUSIONS: Surgery for GBC remains associated with significant morbidity. The availability of benchmark values may facilitate comparisons in future analyses among GBC patients, GBC surgical approaches, and centers performing GBC surgery.

Sentinel lymph node biopsy in patients with T1a cutaneous malignant melanoma: A multicenter cohort study.

BACKGROUND: Sentinel lymph node biopsy is not routinely recommended for T1a cutaneous melanoma due to the overall low risk of positivity. Prognostic factors for positive sentinel lymph node (SLN+) in this population are poorly characterized. OBJECTIVE: To determine factors associated with SLN+ in patients with T1a melanoma. METHODS: Patients with pathologic T1a (<0.80 mm, nonulcerated) cutaneous melanoma from 5 high-volume melanoma centers from 2001 to 2020 who underwent wide local excision with sentinel lymph node biopsy were included in the study. Patient and tumor characteristics associated with SLN+ were analyzed by univariate and multivariable logistic regression analyses. Age was dichotomized into ≤42 (25% quartile cutoff) and >42 years. RESULTS: Of the 965 patients identified, the overall SLN+ was 4.4% (N = 43). Factors associated with SLN+ were age ≤42 years (7.5% vs 3.7%; odds ratio [OR], 2.14; P = .03), head/neck primary tumor location (9.2% vs 4%; OR, 2.75; P = .04), lymphovascular invasion (21.4% vs 4.2%; OR, 5.64; P = .01), and ≥2 mitoses/mm2 (8.2% vs 3.4%; OR, 2.31; P = .03). Patients <42 years with ≥2 mitoses/mm2 (N = 38) had a SLN+ rate of 18.4%. LIMITATIONS: Retrospective study. CONCLUSION: SLN+ is low in patients with T1a melanomas, but younger age, lymphovascular invasion, mitogenicity, and head/neck primary site appear to confer a higher risk of SLN+.

Optimizing Detection of Lymphatic Invasion in Primary Cutaneous Melanoma With the Use of D2-40 and a Paired Melanocytic Marker.

ABSTRACT: Dual immunohistochemical (IHC) staining with D2-40 and S100 improves detection of lymphatic invasion (LI) in primary cutaneous melanoma. However, limited data exist evaluating this technique using other melanocytic markers, and thus, the optimal marker for detection of LI is unestablished. To address this knowledge gap, a case-control study was performed comparing melanoma specimens from 22 patients with known lymphatic spread (LS) with a control group of 11 patients without LS. Specimens underwent dual IHC staining with D2-40 and MART-1, SOX-10, and S100 to evaluate for LI. Receiver operating characteristic analysis was used to estimate each stain's accuracy for detection of LI. The LS group was more likely to be ≥65 years (P = 0.04), have a tumor thickness of ≥1 mm (P < 0.01), and have ulcerated tumors (P = 0.02). Detection of LI with D2-40/MART-1 significantly correlated with LS (P = 0.03), and the D2-40/MART-1 stain was most accurate for LI based on receiver operating characteristic curve analysis (area under the curve [AUC] 0.705) in comparison with D2-40/SOX-10 (AUC 0.575) and D2-40/S100 (AUC 0.633). These findings suggest that MART-1 may be the optimal melanocytic marker to combine with D2-40 for detection of LI in melanoma. Further studies are needed to determine the utility of routinely performing these stains for histopathologic analysis of melanoma.

Survival Outcomes Following Lymph Node Biopsy in Thin Melanoma-A Propensity-Matched Analysis.

BACKGROUND: The use of sentinel lymph node biopsy in patients with T1 melanoma ≤ 1 mm has not been reported in a prospective clinical trial setting, so these clinical outcomes remain understudied. This study seeks to evaluate overall survival (OS) with and without lymph node biopsy (LNB) in patients with clinical T1N0M0 melanoma (0.5-1.0 mm). PATIENTS AND METHODS: Patients were identified using the National Cancer Data Base (2004-2012). After stratification into 0.5-0.7-mm and 0.8-1.0-mm groups, patients undergoing LNB were propensity score-matched 1:1 to patients not undergoing LNB. OS was compared using the Kaplan-Meier method and the stratified log-rank test. RESULTS: Resection was performed in 28,846 patients, and LNB in 14,028 (49%); 15,194 were included in propensity score-matched analysis. The LNB and no-LNB groups were well balanced on all studied covariates (standardized mean difference < 0.10). Among patients with tumors 0.5-0.7 mm, 5- and 10-year OS were 94.7% and 82.7%, respectively, for the LNB group compared with 94.3% and 84.4% for the no-LNB group (p = 0.35). Among patients with tumors 0.8-1.0 mm in thickness, 5- and 10-year OS were 93.9% and 81.6%, respectively, for the LNB group compared with 90.3% and 74.3% for the no-LNB group (p < 0.0001). CONCLUSIONS: There was no difference in OS by LNB status in patients with lesions 0.5-0.7 mm, consistently with recommendations against its routine use in this group. In lesions 0.8-1.0 mm, receipt of LNB was associated with a clinically small but significant improvement in OS. Further study is warranted to better understand this outcome difference.

Predictors of False Negative Sentinel Lymph Node Biopsy in Clinically Localized Merkel Cell Carcinoma.

BACKGROUND: Sentinel lymph node biopsy (SLNB) is routinely recommended for clinically localized Merkel cell carcinoma (MCC); however, predictors of false negative (FN) SLNB are undefined. METHODS: Patients from six centers undergoing wide excision and SLNB for stage I/II MCC (2005-2020) were identified and were classified as having either a true positive (TP), true negative (TN) or FN SLNB. Predictors of FN SLNB were identified and survival outcomes were estimated. RESULTS: Of 525 patients, 28 (5.4%), 329 (62.7%), and 168 (32%) were classified as FN, TN, and TP, respectively, giving an FN rate of 14.3% and negative predictive value of 92.2% for SLNB. Median follow-up for SLNB-negative patients was 27 months, and median time to nodal recurrence for FN patients was 7 months. Male sex (hazard ratio [HR] 3.15, p = 0.034) and lymphovascular invasion (LVI) (HR 2.22, p = 0.048) significantly correlated with FN, and increasing age trended toward significance (HR 1.04, p = 0.067). The 3-year regional nodal recurrence-free survival for males >75 years with LVI was 78.5% versus 97.4% for females ≤75 years without LVI (p = 0.009). Five-year disease-specific survival (90.9% TN vs. 51.3% FN, p < 0.001) and overall survival (69.9% TN vs. 48.1% FN, p = 0.035) were significantly worse for FN patients. CONCLUSION: Failure to detect regional nodal microscopic disease by SLNB is associated with worse survival in clinically localized MCC. Males, patients >75 years, and those with LVI may be at increased risk for FN SLNB. Consideration of increased nodal surveillance following negative SLNB in these high-risk patients may aid in early identification of regional nodal recurrences.

Grade is a Dominant Risk Factor for Metastasis in Patients with Rectal Neuroendocrine Tumors.

BACKGROUND: Small (< 2 cm) and diminutive (< 1 cm) rectal neuroendocrine tumors (RNETs) are often described as indolent lesions. A large single-center experience was reviewed to determine the incidence of metastasis and the risk factors for its occurrence. METHODS: Cases of RNET between 2010 and 2017 at a single institution were retrospectively reviewed. The rate of metastasis was determined, and outcomes were stratified by tumor size and grade. Uni- and multivariable predictors of metastasis were identified, and a classification and regression tree analysis was used to stratify the risk for distant metastasis. RESULTS: The study identified 98 patients with RNET. The median follow-up period was 28 months. Of the 98 patients, 79 had primary tumors smaller than 1 cm, 8 had tumors 1 to 2 cm in size, and 11 had tumors 2 cm in size or larger. In terms of grade, 86 patients had grade 1 (G1) tumors, 8 patients had grade 2 (G2) tumors, and 4 patients had grade 3 (G3) tumors. Twelve patients developed metastatic disease. Both size and grade were associated with distant metastasis in the uni- and multivariable analyses, but when stratified by grade, size was predictive of metastasis only for G1 tumors (p < 0.001). Among the 12 patients with metastatic disease, 3 (25%) had diminutive primary tumors, and 9 (75%) had primary tumors 2 cm in size or larger. Diminutive tumors that metastasized were all G2. CONCLUSIONS: Patients with diminutive and small RNETs are at risk for metastatic disease. Tumor grade is a dominant predictor of dissemination. More rigorous staging, closer surveillance, or more aggressive initial management may be warranted for patients with G2 tumors, irrespective of size.

The impact of surgery for metastatic pancreatic neuroendocrine tumor: a contemporary evaluation matching for chromogranin a level.

BACKGROUND: Studies supporting surgical management of metastatic pancreatic neuroendocrine tumor (PNET) are limited by selection bias. Chromogranin A (CgA) has been used as a biomarker for PNET and may reflect disease burden or biology. This study aimed to correlate CgA level with overall survival and to use it to match patients selected for different treatment approaches in an analysis of the impact of surgical management. METHODS: 1478 patients diagnosed with PNET in the National Cancer Database (2004-2014) were retrospectively identified, and logistic regression analyses were used to evaluate associations between the presence of metastatic disease and CgA level. After matching patients by CgA level and other factors predictive of surgical management, Kaplan-Meier survival analysis was performed. RESULTS: Median CgA level was significantly higher in metastatic versus localized PNET(169 ng/mL versus 66 ng/mL, p < 0.001). On multivariate logistic regression, CgA level was predictive of metastatic disease(OR 1.002, p < 0.001) and survival in metastatic and non-metastatic patients. After matching for CgA level, surgery was associated with improved overall survival. DISCUSSION: CgA level is predictive of the presence of distant metastatic disease and overall survival in PNET. When matched by CgA and other predictors of treatment approach, patients with metastatic PNET undergoing surgery have improved survival.

Microsatellitosis in Patients with Melanoma.

BACKGROUND: Microsatellitosis (mS) in melanoma has been considered a marker of unfavorable tumor biology, leading to the current American Joint Committee on Cancer staging of IIIB/C/D disease, despite few investigative studies of this entity limited by the small sample sizes and incomplete nodal microstaging. We sought to better characterize outcomes and prognostic factors in a multi-institutional cohort of patients with mS and nodal microstaging. METHODS: The Sentinel Lymph Node Working Group cohort included 414 mS patients who underwent sentinel lymph node (SLN) biopsy. Cox regression analysis was used to evaluate the prognostic significance of established clinicopathologic characteristics. Melanoma-specific survival (MSS) of patients with mS was compared with 3002 similarly staged patients from the Surveillance, Epidemiology, and End Results (SEER) Program registry. RESULTS: The median age of the mS cohort was 64.9 years; 39.6% were female. Median thickness was 3 mm, 40.6% of cases were ulcerated, and the SLN positivity rate was 46.7%. Increasing thickness, male sex, and SLN positivity were significantly associated with poorer MSS. Stage IIIB/C/D 5-year MSS rates were 86.3% (95% confidence interval [CI] 79.4-93.3%), 54.1% (95% CI 45.4-59.7%), and 44.2% (95% CI 25.4-63.0%), respectively. MSS survival for the stage IIIB mS cohort was significantly better than a similarly staged SEER cohort (5-year MSS of 70.1%, 95% CI 66.0-74.2%), while no significant difference was observed for the stage IIIC or D cohorts. CONCLUSIONS: SLN metastases are common and are a significant prognostic factor in patients with mS. Survival in stage IIIB patients with mS was considerably more favorable than their stage would otherwise suggest, which has important implications for decisions regarding adjuvant therapy for patients with mS.

Multispectral photoacoustic imaging for the detection of subclinical melanoma.

BACKGROUND: Photoacoustic imaging (PAI) is a noninvasive technique for melanin detection within melanoma metastases. While ex vivo and early studies suggest promising clinical application, there are no standardized parameters for defining presence of melanoma metastases. METHODS: Following flank/hindlimb melanoma induction in BRaf-PTEN transgenic mice, bilateral inguinal lymph nodes (LN) were imaged in vivo at 4 to 8 weeks using PAI. Fourteen diagnostic parameters for in vivo detection of LN metastases were compared using the receiver operating characteristic and area-under-the-curve (AUC). Limits of detectability were assessed in ex vivo and in vitro phantom studies. RESULTS: Forty-nine LNs were imaged in 25 mice. Among metastatic LNs, tumor size ranged from scattered cells to 2.8 mm. The strongest predictor of LN metastasis was the ratio of peak 10% PA melanin signal in the LN compared with adjacent soft tissue (median 4.22 for positive LNs vs 1.07 for negative LNs, P < 0.0001). The AUC was 0.95 (95% CI, 0.90-1.00). In phantom studies, B16 tumor cells were detectable at a concentration of 10 to 25 cells/µL and at a tissue depth of 2.5-3 cm. CONCLUSIONS: We identified a simple, objective diagnostic parameter for identifying melanoma LN metastases in vivo. These findings may help inform the design of future clinical trials.

Relationship between age and likelihood of lymph node metastases in patients with intermediate thickness melanoma (1.01-4.00 mm): A National Cancer Database study.

BACKGROUND: There is large variability in the risk of sentinel lymph node (SLN) positivity among patients with intermediate thickness melanoma (ITM), with a subgroup of patients exhibiting a low risk of nodal disease. OBJECTIVE: To identify a group of patients with ITM for whom the risk of nodal disease is low. METHODS: A retrospective cohort of patients with ITM who underwent wide excision and nodal evaluation from 2010 to 2013 was identified by using the National Cancer Database and analyzed for the presence of nodal disease. Classification and regression tree analysis identified the most important factors used in a model to identify groups at low risk of SLN positivity. RESULTS: Of 23,440 patients, 14.7% were found to have nodal metastasis. On classification and regression tree analysis, patients older than 55 years without lymphovascular invasion and with a lesion thickness less than 1.7 mm had an SLN positivity rate of 4.9%. A model using age and thickness in nonulcerated patients identified a low-risk subgroup with a corresponding SLN positivity rate of 4.7%. LIMITATIONS: This was a retrospective study, and the model developed requires prospective validation. CONCLUSIONS: Patient age is an important factor in estimating risk of SLN in patients with ITM and may help identify patients without ulceration who may be safely spared an SLN biopsy.

A comprehensive genomic binding map of gene and chromatin regulatory proteins in Saccharomyces.

Hundreds of different proteins regulate and implement transcription in Saccharomyces. Yet their interrelationships have not been investigated on a comprehensive scale. Here we determined the genome-wide binding locations of 200 transcription-related proteins, under normal and acute heat-shock conditions. This study distinguishes binding between distal versus proximal promoter regions as well as the 3' ends of genes for nearly all mRNA and tRNA genes. This study reveals (1) a greater diversity and specialization of regulation associated with the SAGA transcription pathway compared to the TFIID pathway, (2) new regulators enriched at tRNA genes, (3) a global co-occupancy network of >20,000 unique regulator combinations that show a high degree of regulatory interconnections among lowly expressed genes, (4) regulators of the SAGA pathway located largely distal to the core promoter and regulators of the TFIID pathway located proximally, and (5) distinct mobilization of SAGA- versus TFIID-linked regulators during acute heat shock.

Trends in major upper abdominal surgery for cancer in octogenarians: Has there been a change in patient selection?

BACKGROUND: Although there is a general perception that, as the older population grows in number, more are undergoing surgery, there are few data on trends in major resections for cancer and short-term outcomes in this group. METHODS: The Nationwide Inpatient Sample was (NIS) used to estimate the national trends of major upper abdominal resections (esophagus, stomach, liver, pancreas) for cancer in octogenarians (aged ≥80 years) from 2001 to 2011. Resection rates performed per year were incidence-adjusted within this age group for each cancer type as determined by the NIS database. Joinpoint regression was used to calculate average annual percentage changes (AAPC) when evaluating trends over time. RESULTS: During the study period, octogenarians underwent an estimated 30,356 upper abdominal organ resections for cancer in the United States, representing 3.8% of all cancer admissions among octogenarians. Resection rates in octogenarians increased significantly over time (AAPC, 2.54; P < .001) secondary to increasing trends in pancreatic (AAPC, 11.52; P < .001) and hepatic (AAPC, 6.67; P < .001) resections. Elixhauser comorbidity index scores increased from a mean of 3.61 to 4.20 (AAPC, 1.31; P < .001), whereas inpatient mortality during this time decreased from 13.6% to 8.2% (AAPC, 5.58; P < .001). CONCLUSIONS: Overall rates of major upper abdominal cancer resections in octogenarians are increasing over time, driven by increases in liver and pancreatic resections. These increases were observed despite a less favorable patient morbidity profile over time. These patterns may suggest shifting selection criteria for octogenarians undergoing major abdominal surgery over time in the context of diminishing postoperative mortality. Cancer 2018;124:125-35. © 2017 American Cancer Society.