Faculty Bibliography

BACKGROUND:Prospective studies have identified obstructive sleep apnea (OSA) as a risk factor for increased overall cancer incidence and mortality. The potential role of OSA in the risk or progression of specific cancers is not well known. We hypothesized that pathological differences in pancreatic cancers from OSA cases compared to non-OSA cases would implicate OSA in pancreatic cancer progression. METHODS:We reviewed the medical records of 1031 patients who underwent surgical resection without neoadjuvant therapy for pancreatic ductal adenocarcinoma (PDAC) at Johns Hopkins Hospital between 2003 and 2014 and compared the TNM classification of their cancer and their overall survival by patient OSA status. RESULTS:OSA cases were significantly more likely than non-OSA cases to have lymph node-negative tumors (37.7% vs. 21.8%, p = 0.004). Differences in the prevalence of nodal involvement of OSA vs. non-OSA cases were not associated with differences in other pathological characteristics such as tumor size, tumor location, resection margin status, vascular or perineural invasion, or other comorbidities more common to OSA cases (BMI, smoking, diabetes). A logistic regression model found that a diagnosis of OSA was an independent predictor of lymph node status (hazard ratio, 0.051, p = 0.038). Patients with OSA had similar overall survival compared to those without OSA (HR, 0.89, (0.65-1.24), p = 0.41). CONCLUSION/CONCLUSIONS:The observed pathological differences between OSA-associated and non-OSA-associated pancreatic cancers supports the hypothesis that OSA can influence the pathologic features of pancreatic ductal adenocarcinoma.

BACKGROUND AND OBJECTIVES/OBJECTIVE:The prognostic impact of primary colorectal cancer (CRC) location following resection of colorectal liver metastasis (CRLM) remains largely unknown. We sought to characterize the prognostic implications of primary tumor location among patients who underwent curative-intent hepatectomy for CRLM. METHODS:Tumors of the cecum, ascending, and transverse colon were defined as right-sided; tumors of the sigmoid flexure, descending, and sigmoid colon were defined as left-sided. Clinicopathologic and long-term survival data were collected and assessed using univariable and multivariable analyses. RESULTS:About 475 patients who underwent CRLM resection at a single institution were included; most patients had left-sided tumors (n = 284). Median and 5-year RFS was 20.2 months and 28.0%, respectively. Patients who had a left-sided primary tumor had a shorter RFS compared with patients who had a right-sided tumor (P = 0.01). Although site of and time to recurrence did not differ between the two groups (P > 0.05), patients with right-sided primary tumors were more likely to recur with advanced disease (i.e., ≥4 recurrent lesions) (P < 0.01). In turn, patients with right-sided tumors had both worse OS (P = 0.03) and worse survival after recurrence (P = 0.01). CONCLUSION/CONCLUSIONS:While patients with right-sided tumors experienced longer RFS, when these patients did recur following CRLM resection, disease extent was more advanced. In turn, OS following recurrence was shorter among patients with right-sided CRC. J. Surg. Oncol. 2016;114:803-809. © 2016 2016 Wiley Periodicals, Inc.

We previously developed a novel tumor subtype classification model for duodenal adenocarcinomas based on a combination of the CpG island methylator phenotype (CIMP) and MLH1 methylation status. Here, we tested the prognostic value of this model in stage II colorectal cancer (CRC) patients. Tumors were assigned to CIMP+/MLH1-unmethylated (MLH1-U), CIMP+/MLH1-methylated (MLH1-M), CIMP-/MLH1-U, or CIMP-/MLH1-M groups. Age, tumor location, lymphovascular invasion, and mucin production differed among the four patient subgroups, and CIMP+/MLH1-U tumors were more likely to have lymphovascular invasion and mucin production. Kaplan-Meier analyses revealed differences in both disease-free survival (DFS) and overall survival (OS) among the four groups. In a multivariate analysis, CIMP/MLH1 methylation status was predictive of both DFS and OS, and DFS and OS were shortest in CIMP+/MLH1-U stage II CRC patients. These results suggest that tumor subtype classification based on the combination of CIMP and MLH1 methylation status is informative in stage II CRC patients, and that CIMP+/MLH1-U tumors exhibit aggressive features and are associated with poor clinical outcomes.

OBJECTIVES/OBJECTIVE:At diagnosis, 30% of patients with pancreatic cancer are unresectable stage 3 locally advanced. The standard treatment for locally advanced pancreatic cancer (LAPC) is not defined. The current study was conducted to assess the roles of chemotherapy and chemoradiation for LAPC treatment. MATERIALS AND METHODS/METHODS:Between June 2006 and March 2011, 100 patients with LAPC were treated at the Johns Hopkins Hospital. Retrospective analysis was performed to compare cumulative incidence of progression (CIP) and overall survival (OS) among different subgroups. RESULTS:For the 100 patients, the median OS was 15.8 months and the median CIP was 8.4 months. The combination of chemotherapy and chemoradiation before disease progression was significantly associated with improved CIP (P=0.001) and improved OS when compared with chemoradiation alone (median OS: 16.4 vs. 11.1 mo, P=0.03). Among patients receiving combination treatment, patients who received chemotherapy first followed by chemoradiation had a trend toward lower CIP (P=0.09) and improved OS (median OS: 18.1 vs. 11.0 mo, P=0.09). Patients who received >2 cycles of chemotherapy before chemoradiation had a significantly decreased CIP (P=0.008) and a trend toward better OS (median OS: 19.4 vs. 15.7 mo, P=0.10). On multivariate analysis, receiving >2 cycles of chemotherapy before chemoradiation was associated with improved CIP. CONCLUSIONS:Although combination chemotherapy and chemoradiation is favored in the treatment of LAPC, longer induction chemotherapy may play a more important role in sensitization of tumors to subsequent chemoradiation. Our results support treating patients with induction chemotherapy for at least 3 cycles followed by consolidative chemoradiation. These results merit further validation by a prospective study.

BACKGROUND:There are no established guidelines for pathologic diagnosis/reporting of intraductal papillary mucinous neoplasms (IPMNs). DESIGN/METHODS:An international multidisciplinary group, brought together by the Verona Pancreas Group in Italy-2013, was tasked to devise recommendations. RESULTS:(1) Crucial to rule out invasive carcinoma with extensive (if not complete) sampling. (2) Invasive component is to be documented in a full synoptic report including its size, type, grade, and stage. (3) The term "minimally invasive" should be avoided; instead, invasion size with stage and substaging of T1 (1a, b, c; ≤ 0.5, > 0.5-≤ 1, > 1 cm) is to be documented. (4) Largest diameter of the invasion, not the distance from the nearest duct, is to be used. (5) A category of "indeterminate/(suspicious) for invasion" is acceptable for rare cases. (6) The term "malignant" IPMN should be avoided. (7) The highest grade of dysplasia in the non-invasive component is to be documented separately. (8) Lesion size is to be correlated with imaging findings in cysts with rupture. (9) The main duct diameter and, if possible, its involvement are to be documented; however, it is not required to provide main versus branch duct classification in the resected tumor. (10) Subtyping as gastric/intestinal/pancreatobiliary/oncocytic/mixed is of value. (11) Frozen section is to be performed highly selectively, with appreciation of its shortcomings. (12) These principles also apply to other similar tumoral intraepithelial neoplasms (mucinous cystic neoplasms, intra-ampullary, and intra-biliary/cholecystic). CONCLUSIONS:These recommendations will ensure proper communication of salient tumor characteristics to the management teams, accurate comparison of data between analyses, and development of more effective management algorithms.

OBJECTIVES/OBJECTIVE:To evaluate the influence of IGFBP-3 methylation on recurrence in patients with stage II colorectal cancer (CRC) from 2 independent cohorts. BACKGROUND:The relationship between IGFBP-3 methylation in primary tumors (PTs) or lymph nodes (LNs) and risk of recurrence in patients with stage II CRC treated with surgery alone is unknown. METHODS:IGFBP-3 methylation of DNA from 115 PTs and 1641 LNs in patients with stage II CRC from 2 independent cohorts was analyzed. Forty patients developed recurrence, whereas 75 matched patients remained recurrence free for more than 2 years after surgery. Cox proportional hazard models were used to calculate hazard ratios (HRs) of recurrence, adjusted for patient and tumor characteristics. RESULTS:Methylation of IGFBP-3 in PTs was identified to be significantly associated with risk of recurrence in the training set. The signature was tested in a validation set and classified 40.7% of patients as high risk. Five-year recurrence-free survival rates were 76.4% and 58.3% for low- and high-risk patients, respectively, with an HR of 2.21 (95% confidence interval, 1.04-4.68; P = 0.039). In multivariate analysis, the signature remained the most significant prognostic factor, with an HR of 2.40 (95% confidence interval, 1.10-5.25; P = 0.029). A combined analysis of 1641 LNs from the 2 sets identified IGFBP-3 methylation in LNs was not associated with risk of recurrence. CONCLUSIONS:Detection of IGFBP-3 methylation in PTs, but not in LNs, provides a powerful tool for the identification of patients with stage II CRC at high risk of recurrence.

OBJECTIVE:To determine the feasibility of genotyping pancreatic tumors via fine needle aspirates (FNAs). BACKGROUND:FNA is a common method of diagnosis for pancreatic cancer, yet it has traditionally been considered inadequate for molecular studies due to the limited quantity of DNA derived from FNA specimens and tumor heterogeneity. METHODS:In vitro mixing studies were performed to deduce the minimum cellularity needed for genetic analysis. DNA from both simulated FNAs and clinical FNAs was sequenced. Mutational concordance was determined between simulated FNAs and that of the resected specimen. RESULTS:Limiting dilution studies indicated that mutations present at allele frequencies as low as 0.12% are detectable. Comparison of simulated FNAs and matched tumor tissue exhibited a concordance frequency of 100% for all driver genes present. In FNAs obtained from 17 patients with unresectable disease, we identified at least 1 driver gene mutation in all patients including actionable somatic mutations in ATM and MTOR. The constellation of mutations identified in these patients was different than that reported for resectable pancreatic cancers, implying a biologic basis for presentation with locally advanced pancreatic cancer. CONCLUSIONS:FNA sequencing is feasible and subsets of patients may harbor actionable mutations that could potentially impact therapy. Moreover, preoperative FNA sequencing has the potential to influence the timing of surgery relative to systemic therapy. FNA sequencing opens the door to clinical trials in which patients undergo neoadjuvant or a surgery-first approach based on their tumor genetics with the goal of utilizing cancer genomics in the clinical management of pancreatic cancer.

INTRODUCTION/BACKGROUND:International Study Group of Pancreatic Fistula (ISGPF) grade C postoperative pancreatic fistulas (POPF) are the greatest contributor to major morbidity and mortality following pancreatoduodenectomy (PD); however, their infrequent occurrence has hindered deeper analysis. This study sought to develop a predictive algorithm, which could facilitate effective management of this challenging complication. METHODS:Data were accrued from 4301 PDs worldwide. Demographics, postoperative management, and microbiological characteristics of grade C POPFs were evaluated. American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) preoperative and intraoperative variables were compared between grade C POPFs and a 639-case sample of non-grade C POPFs. Risk factors for grade C POPF formation were identified using regression analysis. RESULTS:Grade C POPFs developed in 79 patients (1.8 %). Deaths (90 days) occurred in 2.0 % (N = 88) of the overall series, with 35 % (N = 25) occurring in the presence of a grade C POPF. Reoperations occurred 72.2 % of the time. The rates of single- and multi-system organ failure were 28.2 and 39.7 %, respectively. Mortality rates escalated with pulmonary, renal, and neurologic organ failure, but they were unaffected by reoperation(s). The median number of complications incurred was four (IQR: 2-5), and the median duration of hospital stay was 32 (IQR: 21-54) days. Warning signs for impending grade C POPFs most often presented on postoperative day (POD) 6. Adjuvant chemotherapy might have benefited 55.7 % of grade C POPF patients, yet it was delayed in 25.6 % and never delivered in 67.4 % of these patients. Predictive models for grade C POPF occurrence based on preoperative factors alone and preoperative and intraoperative factors yielded areas under the receiver operating characteristic curve of 0.73 and 0.84 (both P < 0.000001), respectively. CONCLUSION/CONCLUSIONS:This global study represents the largest analysis of grade C POPFs following PD. It describes the severe burden that grade C POPFs incur on patients, with high rates of reoperation and infection, while also potentially worsening overall survival by causing death and delay/omission of adjuvant therapy. Additionally, aggressive clinical management for these POPFs did not improve or worsen 90-day mortality. Predictive tools developed through these data may provide value in managing this difficult complication.

BACKGROUND:Little is known regarding the effects of caseload volume of other relevant members of the "surgical team." The present study sought to report variations in health care utilization and outcomes relative to surgeon and anesthesiologist volume among patients undergoing pancreatic surgery. METHODS:A total of 969 patients undergoing pancreatic surgery from 2011-2013 were identified at a large, tertiary care center. Multivariable regression analyses explored the effects of provider volume on crystalloid administration, blood transfusions, mortality, length of stay, and hospital charges. RESULTS:A total of 11 surgeons were identified while 100 anesthesiologists were involved in providing care to all patients. Annual case volume for surgeons ranged from 5-101 pancreatic resections per year; each anesthesiologist was involved in a fewer number of cases per year with a maximum of 15 patients treated by the same anesthesiologist. Higher volume surgeons had higher transfusions (odds ratio [OR], 1.85; 95% confidence interval [CI], 1.38-2.47; P < 0.001), greater crystalloid administration (OR, 1.62; 95% CI, 1.24-2.12; P < 0.001), and longer length of stay (OR, 1.74; 95% CI, 1.20-2.53; P = 0.003). In contrast, 30-d readmission was lower among higher volume surgeons (low volume versus high volume; 23.1% versus 11.6%; P < 0.001). Variations in patient-related outcomes were not associated with anesthesia provider volume (all P > 0.05). Similarly, total hospital charges and mortality were not associated with provider volumes (both P > 0.05). CONCLUSIONS:Although variability exists in health care practices among providers at the surgeon level, less is observed among anesthesiologists. Although a proportion of this variability can be explained by provider volumes, a significant proportion remains unexplained possibly due to nonmodifiable factors such as patient case mix.

OBJECTIVE:Arterial resection (AR) during pancreatic tumor resection is controversial. We examined the safety and efficacy of AR during pancreatectomy. METHODS:We used a prospective institutional database that includes 6522 patients who underwent pancreatectomy from 1970 to 2014; 35 had AR. We performed a 2:1 propensity match for patients without and with AR on the basis of preoperative patient and tumor variables. We then compared operative and postoperative outcomes between matched groups. RESULTS:AR included 18 hepatic, 8 celiac, 3 splenic, 3 middle colic, 2 superior mesenteric, and 1 left renal artery. There were 20 primary, 4 vein, and 2 graft reconstructions; 11 were emergent and 24 elective. Before matching, patients with AR were younger (58 ± 2 vs 63 ± 0.2 years old; P = .05), more likely to be of black race (26% vs 9%; P = .003), to have received preoperative chemotherapy (17% vs 2%; P < .001), have a later stage and larger tumor (4 ± 0.8 vs 3 ± 0.04 cm; P = .05), more resections that included removal of all macroscopic disease, but microscopic residual tumor remained (31% vs 14%; P = .02), greater blood loss (1285 ± 276 vs 822 ± 16 mL; P = .02), and more frequent cardiac complications (11% vs 4%; P = .03) compared with patients without AR. After propensity matching, baseline patient characteristics were similar between groups. For perioperative outcomes, the groups did not differ in surgical time, blood loss, length of stay, or complications including anastomotic leaks, bleeding, cardiac, infectious complications, or liver infarct or failure (all; P = not significant). Patency was 97% at a mean follow-up of 510 ± 184 days with 1 hepatic artery AR thrombosis. Long-term outcomes were significantly different: patients with AR had a lower rate of local tumor recurrence (20% vs 47%; P = .007) but also lower 1-year (50% vs 87%; P = .002) and median survival (22 ± 18 vs 49 ± 7 months; P = .002). CONCLUSIONS:AR during pancreatectomy is safe and not associated with increased complications. Although it significantly reduces the risk of local tumor recurrence, AR is associated with worse survival compared with patients who do not undergo AR.