08.02 Cutaneous Melanoma in Children, Adolescents and Adults:An Outcomes Study Involving 326,278 Patients

Posted on January 25, 2017

C. S. Lau1,2, K. Kulkarni2, R. S. Chamberlain1,2,3  1Saint Barnabas Medical Center,Surgery,Livingston, NJ, USA 2St. George’s University School Of Medicine,St. George’s, St. George’s, Grenada 3Rutgers University,Surgery,Newark, NEW JERSEY, USA

Introduction: The incidence of cutaneous melanoma has steadily risen over the last decade, with over one million people in the United States living with a melanoma diagnosis. Melanoma is the most common form of skin cancer among pediatric patients, accounting for 7% of pediatric cancers. This study examines a large cohort of pediatric melanoma patients to identify demographic, pathologic, and clinical factors which affect clinical outcomes and survival.

Methods: Data on 326,278 patients with malignant cutaneous melanoma was abstracted from the Surveillance, Epidemiology and End Result (SEER) database (1973-2013). Children were defined as age ‘0-12’, adolescents as ‘13-19’ and adults as ‘≥20’. Standard statistical analyses were performed.

Results: Among 326,278 melanoma patients, 431 were children, 2,093 were adolescents and 323,754 were adults. The majority of childhood and adolescent melanomas occurred in females (60.3% and 57.9%), whereas a majority of adult melanoma occurred in males (57.0%), p<0.01. Children most commonly presented with melanomas of the head and neck (28.9%) and lower extremities (29.7%), while adolescents and adults had melanoma of the trunk (40.5% and 34.2%), p<0.01. While a majority of patients were Caucasian, significantly more children were Hispanic (16.1% vs. 5.5% and 2.9%, p<0.01), African American (2.1% vs. 0.8% and 0.5%, p<0.01), and Asian/Pacific Islander (6.1% vs. 1.2% and 0.6%, p<0.01). Children had significantly lower rates of localized disease (72.8% vs. 84.6% and 85.9%, p<0.01), higher rates of regional disease (22.3% vs. 13.2% and 10.0%, p<0.01) and deeper tumors (28.0% vs. 9.7% and 9.4% had tumors>3.00mm, p<0.01). Surgery was the most common treatment modality and was associated with prolonged survival among all age groups. Despite presenting with more advanced disease, children and adolescents experienced lower overall (6.4% and 9.9% vs. 28.4%, p<0.01) and cancer-specific mortality (1.4% and 2.0% vs. 15.4%, p<0.01) compared to adults. Similarly, children and adolescents experienced significantly higher 5-year overall (91% and 84% vs. 53%, p<0.01) and cancer-specific survival (94% and 88% vs. 79%, p<0.01).

Conclusions: Melanoma among children occurs most commonly among females, often affects the head and neck regions, and has a significantly higher incidence rate among ethnic minorities. In contrast, adolescents present similarly to adults with primarily truncal melanoma and lower rates of metastatic and deep tumors. All pediatric age (<20 years) patients experience prolonged overall and cancer-specific survival compared to adults. Understanding disparities in pediatric melanoma presentation, incidence and survival should help to increase disease awareness and prevention as well as facilitating early detection.

08.01 Congenital Portosystemic Shunts: To Close or Not to Close?

Posted on January 25, 2017

N. M. Corbitt1, D. Hanto1, C. Baron1, P. Bream1, L. Gillis1, N. Wallace1  1Vanderbilt University Medical Center,Nashville, TN, USA

Introduction:  Congenital Portosystemic Shunts (CPS) are extremely rare vascular anomalies, often discovered incidentally during imaging necessitated for evaluation of associated congenital anomalies. CPS can be classified according to the absence (Type 1 CPS) or presence (Type 2 CPS) of portal venous perfusion of the liver. They can be further classified according to anatomic variations as follows: Type 2A arise from left or right portal vein (PV), Type 2B arise from the PV, and Type 2C arise from splanchnic veins. CPS are associated with hyperammonemia, hepatic masses, and a variety of congenital anomalies. Closure can result in portal hypertension. We hypothesize that portal venography and temporary balloon occlusion is required to distinguish between type 1 and 2 CPS and temporary balloon occlusion (TBO) testing can determine if CPS are amenable to endovascular closure. 

Methods:  We report five new cases of CPS with a variety of illustrative presentations. All patients were diagnosed with CPS during childhood using ultrasonography (US) and/or contrasted tomography (CT) imaging. Portal venography and TBO was performed to further characterize CPS and assess feasibility of shunt closure. CPS closure was performed endovascularly using Amplatzer plugs when a single CPS was present and the portal venous pressure was < 20 mmHg. 

Results: Associated anomalies were universally present. Five children had hyperammonemia and 4/5 had hyperbilirubinemia. In patient 4, US and CT imaging was suggestive of Type 1 CPS, however venography and TBO demonstrated intrahepatic PV filling. Also, two additional CPS were identified during TBO of the CPS between the PV and hepatic vein confluence, including a shunt between the SMV and the gonadal vein as well as a shunt between the SMV and right renal vein. In patient 5, venography demonstrated a CPS between the SMV-SV confluence and the IVC-right atrium confluence consistent with pre-procedural imaging. TBO of the CPS failed to demonstrate portal venous capillary filling characteristic of Type 1 CPS. Three patients underwent successful endovascular shunt closure using Aplatzer plugs and demonstrated subsequent resolution of hyperbilirubinemia and hyperammonemia. Three patients had hepatic lesions including foci of nodular hyperplasia. Notably, patient 3 had resolution of hepatic lesions following CPS closure. 

Conclusion: TBO of CPS can be used to identify intrahepatic portal venous perfusion. Endovascular CPS closure can successfully reverse hepatic deprivation of portal venous flow in Type 2 CPS with resultant correction of hyperammonemia  and hyperbilirubinemia.  We suspect CPS closure could reduce hepatic malignancy potential by contributing to resolution of associated predisposing hepatic hyperplasia. 

 

07.19 Molecular Profiling and Primary Location of Melanoma: Experience at an NCI-designated Cancer Center

Posted on January 25, 2017

J. Lou1, M. Renzetti1, I. Soliman1, H. Wu1, B. Luo1, A. J. Olszanski1, S. Movva1, M. Lango1, S. Reddy1, F. Zih1, J. M. Farma1  1Fox Chase Cancer Center,Surgical Oncology,Philadelphia, PA, USA

Introduction:  Molecular profiling is becoming an increasingly important aspect in the interrogation of primary cancers. Next generation sequencing (NGS) is being used at our institution to examine hot-spot mutations in 50 cancer-related genes in various tumors. The principle aim of this investigation is to evaluate molecular profiling of malignant melanoma (MM) and to correlate primary location of the melanoma with genetic mutations.

Methods:  Patients with primary or recurrent MM of all stages were included in the study. Using NGS, we analyzed tissue samples for mutations in targeted regions of 50 cancer-related genes.  Clinical and pathologic data were collected.

Results: Specimens from 138 patients with MM were analyzed, excluding 2 from analysis due to insufficient DNA and 1 due to incomplete profiling. Median age at diagnosis was 65 (range 24-90) and 64% were male (n=86). There were 26 patients with Stage I melanoma, 72 with Stage II, 14 with Stage III, and 12 with Stage IV. Primary tumor locations included head and neck (N=23), upper extremity (N=29), trunk (N=31), lower extremity (N=39), and mucosal sites (N=5). In total, 218 mutations were identified, affecting 30 unique genes. Patients with a primary location at their head and neck had the greatest average number of mutations (x=2.04), with the greatest percentage of mutations in BRAF (including BRAF K601E, G466V, V600E, V600K, V600R) and TP53 genes (Figure 1). Lower extremity sites had the next highest mean number of mutations (x=1.87), with the greatest fraction of mutations in BRAF and NRAS genes (Figure 1).  Mucosal tumors had the lowest mean number of mutations (x=0.4), with only mutations in TP53 (Figure 1). Across all sites, a large percentage of mutations were either in BRAF, NRAS, or TP53 genes, with BRAF being the most common in lower extremity sites and NRAS most common in truncal tumors.

Conclusion: While there is no statistical significance comparing primary tumor site and genetic mutations, the descriptive statistics warrants further investigation with a larger sample size.

 

07.18 Defining Treatment Paradigms for High-Grade Gastroenteropancreatic Neuroendocrine Tumors

Posted on January 25, 2017

R. L. Guyton Jr1, M. W. McMullen2, C. S. Lea2, J. Brinkley2,3, C. Mosquera1, T. L. Fitzgerald1  1East Carolina University Brody School Of Medicine,Division Of Surgical Oncology,Greenville, NC, USA 2East Carolina University Brody School Of Medicine,Department Of Public Health,Greenville, NC, USA 3American Institutes For Research,Chapel Hill, NC, USA

Introduction:  The incidence of high-grade gastroenteropancreatic neuroendocrine tumors (HG GEP-NET) is increasing, but a lack of consensus remains regarding optimal treatment regimens.  This study examines whether surgical resection and adjuvant chemotherapy provide a survival advantage for patients with HG GEP-NET.

Methods:  Incident cases of HG GEP-NET from 2004-2013 were obtained from the National Cancer Data Base.  Data examined included clinical and pathologic characteristics at diagnosis, treatment strategies, and survival outcomes.  Univariate and Cox regression analyses were performed to examine survival outcomes.

Results:  In total, 4,630 HG GEP-NET patients were identified.  The median age of diagnosis was 66 years.  A majority of patients were male (55%), white (84%), and had a colorectal primary (41%).  Surgical resection was associated (p < .0001) with survival on univariate analysis. On Cox regression analysis, surgical resection remained a significant predictor of survival (HR = 0.43, 95% CI 0.37–0.49) after adjusting for age, primary site, tumor size, and regional lymph node (RLN) involvement.  A total of 1,903 HG GEP-NET underwent surgical resection of the primary site.  On Cox regression analysis, primary site (colorectal, referent; appendix, HR = 0.51, 95% CI 0.36–0.71; small intestine, HR = 0.63, 95% CI 0.50–0.77; pancreas, HR = 0.80, 95% CI 0.66–0.96) were associated with improved survival after adjusting for age, tumor size, and RLN involvement, however, adjuvant chemotherapy (HR = 0.90, 95% CI 0.79–1.02) was not.  Multiple models were constructed to define the benefit of chemotherapy based on the primary site: Model 1 – colorectal primaries; Model 2, primaries of the ampulla of Vater, esophagus, hepatobiliary, and stomach; and Model 3 – primaries of the appendix, pancreas, and small intestine.  On Cox regression analysis, adjuvant chemotherapy conferred a survival benefit in Model 1 (HR = 0.68, 95% CI 0.58 – 0.81), indicated no significant difference in survival in Model 2 (HR = 0.90, 95% CI 0.68 – 1.20), and a survival detriment in Model 3 (HR = 1.66, 95% CI 1.30 – 2.13). 

Conclusion:  Surgical resection was associated with a survival advantage over all primary sites for HG GEP-NETs.  Adjuvant chemotherapy may be beneficial for colorectal HG GEP-NET compared to surgery alone; however, the role of adjuvant chemotherapy for other primary sites is unclear.

 

07.17 Comparing Perioperative Chemotherapy to Endocrine Treatment in Breast Cancer Patients : A Meta-analysis

Posted on January 25, 2017

H. Foong1, Y. Cheng1, D. Yakoub1  1University Of Miami,Division Of Surgical Oncology At Department Of Surgery,Miami, FL, USA

Introduction: Current treatment regimens for hormone receptor positive breast cancer patients involve the administration of perioperative chemotherapy or hormonal therapy. However, there are no concluding data as to superiority of one or the other in terms of event free survival.

Methods: Online search of PubMed, Scopus, Embase, Google Scholar was done. Studies comparing neoadjuvant and/or adjuvant chemotherapy to hormonal therapy in both pre and postmenopausal patients with estrogen receptor positive breast cancer were reviewed. Key Bibliographies were reviewed for related articles. Study quality was assessed using STROBE checklist. Pooled odds ratio (OR) along with the 95% confidence intervals (CI) for event free survival at 5 and 9 years were calculated.

Results:A total of 236 studies were identified. Seven met our selection criteria. Total study subjects were 2489 and 2454 for the hormonal therapy and chemotherapy arms respectively. All 7 studies reported on event free survival rates at 5 years, only 4 of them reported on survival rates at 9 years. Meta-analysis of included data showed a trend for better survival in chemotherapy group at 5 years yet this did not reach statistical significance (OR: 1.14; CI: 0.89-1.45). At 9 years, endocrine therapy group seem to have relatively better survival, yet again, the difference did not reach statistical significance (OR: 0.93; CI: 0.78-1.12).

Conclusion:In spite of relatively better survival at 5 years and relatively worse survival at 9 years, there is not enough evidence to claim superiority of perioperative chemotherapy on endocrine therapy in hormone receptor positive pre or post-menopausal breast cancer patients. Further prospective large scale studies are needed to further examine the difference.      

 

07.16 Outpatient vs. Inpatient Mastectomy: An Analysis of Patient Factors in U.S. Women with Breast Cancer

Posted on January 25, 2017

J. Yu1, M. Rendulic2, M. A. Olsen2, A. E. Cyr1, J. A. Margenthaler1  1Washington University,Surgery,St. Louis, MO, USA 2Washington University,Medicine,St. Louis, MO, USA

Introduction:
Despite advances in surgical technique and perioperative care, most women who undergo mastectomy for breast cancer are still commonly admitted as inpatients for pain control, perceived patient satisfaction, and closer monitoring for possible complications.  However, even one night of inpatient admission represents a substantial cost burden to the patient and has not been shown to reduce the odds of 30-day postoperative complications.  It is unclear which patients might be most suitable for same-day or outpatient mastectomy.  We sought to assess patient factors in relation to the utilization of outpatient vs. inpatient mastectomy in U.S. women with breast cancer.

Methods:
Using the Healthcare Cost and Utilization Project State Ambulatory Surgery and Services Databases and State Inpatient Databases for California and Florida from 2006-2011, we analyzed clinical and demographic factors in women over age 18 undergoing unilateral mastectomy for invasive breast cancer, breast cancer in situ, or history of breast cancer.  Clinical data assessment was performed using ICD-9 and CPT codes and the Elixhauser comorbidity index.  Descriptive statistics were performed to analyze the relationship between patient factors and admission status after mastectomy.

Results:
Of 71,619 women who underwent unilateral mastectomy, 23,503 (33%) were treated as an outpatient and 48,116 (67%) were admitted as inpatients postoperatively.  Significant geographic and temporal differences were clear: patients in Florida (9,440; 38%) were much more likely to have outpatient procedures compared to patients in California (14,063; 30%), and fewer outpatient mastectomy procedures were performed in 2010-2011 (6,350; 31%) compared to 2006-2009 (17,153; 33%).  Outpatients were more likely to undergo simple mastectomy (OR 1.61 [95% CI 1.56-1.66]) and less likely to undergo modified radical mastectomy (OR 0.65 [95% CI 0.63-0.67]) or to have simultaneous implant or expander reconstruction (OR 0.77 [95% CI 0.75-0.81]).  Women undergoing mastectomy as outpatients were also more likely to be younger (≤50 years), have private insurance, be Caucasian, and have fewer comorbidities (p<0.05). 

Conclusion:
The utilization of outpatient mastectomy varies widely based on clinical and geographic factors.  Patients with private insurance, fewer comorbidities, and limited disease who do not undergo immediate reconstruction are more likely to undergo mastectomy without inpatient admission.  Assessment of short- and long-term patient outcomes may provide additional evidence to support outpatient mastectomy as a patient-centered and cost-effective approach for certain breast cancer patients.
 

07.15 Measuring Complete Response After Neoadjuvant Therapy for Rectal Cancer: Do X's Really Equal 0's?

Posted on January 25, 2017

C. T. Ellis1, K. B. Stitzenberg1  1University Of North Carolina At Chapel Hill,Surgery,Chapel Hill, NC, USA

Introduction: Tumor response to neoadjuvant chemoradiation is increasingly recognized as an important prognostic indicator.  Individuals with a pathological complete response (pCR) to neoadjuvant therapy for rectal cancer have been shown to have greatly improved long-term survival.  Tumor response can be estimated from cancer registry data when both clinical and pathological stage information is reported. Prior studies have presumed ypTXNX is equal to ypT0N0, and thus both are usually grouped together as representing pCR. However, this methodology is unproven. For this study, we sought to determine how pCR-status is best captured for rectal adenocarcinoma patients using cancer registry data.

Methods:   Using the National Cancer Database (NCDB), we included patients with clinical stage II/III rectal adenocarcinoma that underwent neoadjuvant chemoradiation and a proctectomy from 2004 – 2013.  We compared outcomes amongst three goups: those with ypTXNX, those with ypT0N0, and those with ypT>0N>0.  We estimated the difference in OS by treatment received using Kaplan-Meier survival curves and Adjusted Cox proportional hazards models, controlling for patient, tumor, and facility characteristics.

Results:  Overall, 3,700 (13%) and 2,756 (10%) of our cohort (n=27,859) had a pCR as indicated by ypTXNX and ypT0N0, respectively. Over time, there was a decrease in those with ypTXNX and an increase in those with ypT0N0.  In 2004, 28% of our cohort were ypTXNX and 5% were ypT0N0.  In 2013, only 3% were ypTXNX and 14% were ypT0N0.  5-year OS for ypTXNX patients was more similar to that of individuals with an incomplete response than those with ypT0N0; 5-year OS 77%, 73%, and 87%, respectively (Figure).

Conclusion:  Prior studies using the NCDB assumed ypTXNX and ypT0N0 both represented pCR.  This assumption is supported by the increase in ypT0N0 and decrease in ypTXNX over time, consistent with changes in coding practices.  However, survival outcomes suggest otherwise, as long-term survival for the ypTXNX and ypT0N0 groups is different.  Survival for pCR using ypT0N0 alone is most consistent with a true pCR based on previously published survival outcomes.  Registry studies that include ypTXNX in the definition of pCR may not accurately capture the true pCR cases. Additional research is needed to validate this methodology. 

07.14 Impact of Complications on Length of Stay in Elective Laparoscopic Colectomies for Malignancy

Posted on January 25, 2017

M. M. Mrdutt2, C. Shaver2, C. Isbell2, R. Essani2, R. Warrier2, J. S. Thomas2, H. T. Papaconstantinou2  2Scott & White Healthcare,Temple, Texas, USA

Introduction:

Colorectal surgery carries inherent risks and complications.  In contemporary healthcare, quality and cost drive value.  Length of hospital stay (LOS) is an indirect measure of quality and cost as LOS is influenced by post-operative complications and the average hospital cost is >$1600/day. However, the impact of specific complications on LOS is not well defined. The purpose of this study is to determine the contribution of specific complications to the LOS in elective laparoscopic colectomy patients with malignancy.

Methods:

ACS NSQIP database (2011-2014) was queried for patients undergoing elective laparoscopic partial colectomy due to malignancy.  Demographics, 30-day post-operative complications (superficial surgical site infections (superficial SSIs), wound occurrences, return to OR, transfusion requirement, sepsis, pulmonary outcomes, DVT/PE, acute renal failure, urinary tract infection (UTI) and cardiac outcomes) and length of stay were evaluated.  Wound occurrences include deep and organ space SSIs as well as wound disruption or dehiscence.  A logistic regression model based on LOS and demographic variables for each individual complication was performed to determine LOS impact.  Significance set at p<0.05.  

Results:
59,118 patients were evaluated.  Over 70% were age 50-79, 34.6% obese (BMI ±30), mean BMI 28.6, 47.5% male and 5.3% with current immunosuppression.  Overall mean LOS was 5.4 days.  Observed rate of complications and associated average expected LOS are listed for infectious complications (Table 1).  Complications including return to OR, transfusion requirements, failure to wean from ventilator, reintubation, deep vein thrombosis, pulmonary embolism, cardiac complications (MI or cardiac arrest) and acute renal failure also all increased expected LOS by at least two additional days (p<0.0001).  Right parital colectomy (23%) versus left/sigmoidectomy (77%) was not statistically significant in the multi-variant analysis for LOS (p = 0.095).

Conclusion:

Post-operative complications increase LOS in laparoscopic colectomy patients undergoing elective resection of malignancy, with superficial SSIs increasing predicted LOS by 2 days, wound occurrences by 6 days and pneumonia by roughly a week in addition to average baseline LOS of 5.4 days.  By identifying the financial impact of complications, with LOS as a surrogate for cost, future QI efforts can target high yield post-operative complications to improve care and minimize health care cost. 

 

07.13 Response to Neoadjuvant Radiation Does Not Effect Survival of Patients with Esophageal Adenocarcinoma

Posted on January 25, 2017

M. Waldrop1, J. Swords1, R. J. Cerfolio1, C. M. Contreras1, M. J. Heslin1, T. N. Wang1, S. Reddy1  1University Of Alabama at Birmingham,Surgical Oncology,Birmingham, Alabama, USA

Introduction: The Dutch CROSS trial established the role of neoadjuvant chemoradiotherapy for improved survival in the treatment of patients with advanced esophageal adenocarcinomas. Both surgery and radiation therapy represent means of loco-regional control, and esophageal cancer is disease that is associated primarily with systemic failure. The goal of the current study is assess the clinico-pathological features of patients with resected esophageal cancer undergoing neoadjuvant radiation therapy on long-term survival.

 

Methods: All patients at a single institution who underwent esophagectomy for esophageal adenocarcinoma after neoadjuvant radiation therapy from January 2004 to December 2014 were included. Patients were divided into short-term survivors (STS) who died of disease-specific causes within 1 year of resection and long-term survivors (LTS) who lived in excess of 5 years after surgery. Peri-operative deaths were excluded. Logistic regression analysis was performed to identify factors predicting LTS.

 

Results: Of the 334 patients who underwent esophagectomy for adenocarcinoma after neoadjuvant radiation, there were 74 (22.2%) STS and 28 (8.4%) LTS. STS were older (median age 65.5 vs. 56.6 years, P=0.003) and were more likely to have lymph node metastases prior to radiation (93.2% vs. 39.3%, P<0.0001) than LTS. LTS had a trend towards having a complete pathologic response to radiation (46.4% vs. 27.0%, P=0.10), undergoing an open operation (92.9% vs. 77.0%, P=0.09), and having a shorter post-operative length of stay (median 8.5 vs. 11.5 days, P=0.10). Univariable and multivariable analyses are shown in the Table. Only the absence of lymph node metastases on the initial evaluation correlated with LTS (P=0.04), whereas pathologic complete response to neoadjuvant radiation did not (P=0.18).

 

Conclusion: Response to neoadjuvant radiation therapy in patients with esophageal cancer does not correlate to survival. Presence of pre-treatment lymph node metastases is likely a marker of systemic disease and is the only variable that predicts outcome with patients with esophageal adenocarcinoma. Clinical response to radiation therapy should not dictate decision for esophagectomy.

07.12 Factors Associated with Burden of Anal Condyloma and Need for Operative Intervention

Posted on January 25, 2017

H. Foss1, C. Y. Peterson1, K. A. Ludwig1, T. J. Ridolfi1  1Medical College Of Wisconsin,Colorectal Surgery,Milwaukee, WI, USA

Introduction:  Human papillomavirus (HPV) is the most common sexually transmitted infection in the United States. An estimated 79 million Americans are currently infected with HPV, and approximately 14 million new infections occur each year. Known risk factors for contracting HPV include high number of lifetime sexual partners, failure to use condoms consistently, history of sexually transmitted infections, immunosuppression, and younger age. Many treatment modalities exist and are largely based on the extent of condyloma present. These treatments may range from simple office based procedures to major full thickness skin resections in those with large volume disease. The aim of this project is to identify risk factors associated with need for surgical intervention as well as extent of disease.

Methods:  A retrospective chart review was completed for patients seen at the Medical College of Wisconsin Anal Dysplasia Clinic from April 2014 to June 2016. Information including demographic information, smoking status, HPV subtype, immune condition, need for surgical intervention, and surgical specimen size was then extracted from the medical record. Descriptive statistics were employed to evaluate the patient population. Logistic regression was used to evaluate for statistically significant covariates in predicting need for surgical intervention. Multiple linear regression analysis was used to evaluate for statistically significant covariates in predicting a log transformation of surgical specimen size. Significance was determined using a p-value equal to or less than 0.05.

Results: 283 patients met inclusion criteria, of which 252 (89.05%) were male, 179 (63.25%) had a smoking history, 232 (81.98%) were HIV positive, and 108 (38.16%) had undergone surgical intervention for anal condyloma. HPV subtyping was completed in 236 (83.39%) patients, of which 76 (32.20%) tested positive for HPV 16, 25 (10.59%) tested positive for HPV 18, and 149 (63.40%) tested positive for other high risk HPV subtypes. Of six covariates analyzed with logistic regression, only male gender significantly predicted the need for surgical intervention (p-value 0.04). Of seven covariates analyzed with multiple linear regression, HPV 18 status (p-value 0.05), smoking history (p-value 0.005), and older age (p-value 0.04) significantly predicted larger specimen size.

Conclusion: 

Male gender is predictive of needing surgical intervention, while HPV subtype 18, smoking history, and older age are associated with greater burden of disease. The only modifiable risk factor in predicting burden of anal condyloma is smoking status with an average increase of 1 cm3 of condyloma burden in those who smoke. Smoking cessation should be strongly encouraged in those who are HPV positive.

 

 

07.11 The Negative Impact of Understaging Rectal Cancer Patients

Posted on January 25, 2017

A. M. Dinaux1,2, R. Amri1,2, L. Bordeianou1,2, H. Kunitake1,2, D. L. Berger1,2  1Massachusetts General Hospital,Surgery,Boston, MA, USA 2Harvard School Of Medicine,Surgery,Brookline, MA, USA

Introduction:
Neoadjuvant chemoradiation has been shown to reduce local recurrence in rectal cancer. It also can reduce the size of large tumors simplifying surgical resection. However, this is an expensive regimen with long-term side effects. Overstaging patients leads to unnecessary treatment. Understaging leads to delaying systemic chemotherapy as radiation therapy is usually done prior to systemic treatment postoperatively. Complicating this decision making process, is fact that we have now started to observe complete responders who have been shown to have similar outcomes with or without surgery. This abstract analyzes rectal cancer patients who were understaged and their subsequent outcomes.

Methods:
This abstract compared rectal adenocarcinoma patients who underwent surgical resection in a single center from 2004 through 2014, with either clinical stage I disease with tumor positive nodes on postoperative pathology, or neoadjuvantly treated clinical stage III patients. Patients who had a local excision were excluded.

Results:
Thirty-three clinically stage I had nodal disease on postoperative pathology (cN0 pN+). These patients had worse rates of EMVI (33.3% vs. 14.5%; P=0.008), perineural invasion (39.4% vs. 16.9%; P = 0.003), large vessel invasion (30.3% vs. 13.5%; P = 0.014), and small vessel invasion (48.5% vs. 15%; <0.001), than pathological  stage III patients  who underwent neoadjuvant therapy. Adjuvant chemotherapy rates were comparable (clinically stage I: 78.8% vs. clinically stage III: 80.7%; P = 0.800), whereas adjuvant radiotherapy rates were unsurprisingly higher in the clinically stage I group (54.5% vs. 1.4%; P<0.001). Local and distant recurrence rates were not significantly different, while rectal cancer death in clinically stage I patients was 12.1%, compared to 9.2% in the clinically stage III group (P=0.594).

Conclusion:
Clinically stage I patients who were pathologically stage III had more aggressive disease than neoadjuvantly treated clinical stage III patients. The long-term outcomes also point towards a trend with increased mortality in the understaged patients. This underlines the importance of neoadjuvant therapy and accurate staging preoperatively, which is essential to avoid understaging and therefore undertreatment of pathological stage III patients.

07.10 Modified Frailty Index Predicts Postoperative Outcomes in Older Gastrointestinal Cancer Patients

Posted on January 25, 2017

S. Vermillion2, F. Hsu3, R. Dorrell2, C. Clark1  1Wake Forest Baptist Health,Department Of Surgery,Winston Salem, NORTH CAROLINA, USA 2Wake Forest University School Of Medicine,Winston Salem, NORTH CAROLINA, USA 3Wake Forest University School Of Medicine,Public Health Sciences,Winston Salem, NC, USA

Introduction:  

Frailty disproportionately impacts older patients with gastrointestinal cancer, rendering these patients at increased risk for poor outcomes following surgery. A frailty index may aid physicians in preoperative risk stratification. We hypothesized that high modified frailty index (mFI) scores are associated with adverse clinical outcomes after tumor resection in older, gastrointestinal cancer patients.

Methods:

All patients (over 60 years) who underwent tumor resection of the liver and bile duct, pancreas, colon and rectum, esophagus, or stomach were identified in the 2005-2012 NSQIP Participant Use File. Patients who were categorized as ASA 5, diagnosed with preoperative sepsis, or underwent emergency procedures were excluded. mFI was defined by 11 previously described, preoperative variables. Frailty was defined by an mFI score > 0.27. The postoperative course was evaluated using univariate and multivariate analysis. 

Results: 

41,455 patients (mean age 72.4 years, 47.4% female) met the eligibility criteria. Within the study population, the most prevalent form of cancer was colorectal (68.3%, n=28,708), the mean mFI score was 0.11, and 2.8% of patients had a mFI > 0.27 (n=1,164).  Patients with high mFI (>0.27) were significantly more likely to have increased length of stay (12.3 vs 9.0 days), major complications (13.9 vs. 10.5%), and 30-day mortality (5.5 vs. 2.5%), (all p<0.001). Univariate analysis revealed that mFI was associated with significantly increased major complications (OR 1.88, 95% CI 1.75-2.02, p<0.001) and 30-day mortality (OR 2.35, 95% CI 2.03-2.72, p<0.001). After adjusting for age, gender, BMI, ASA, albumin level, mFI was identified as an independent predictor of major complications (OR 1.52, 95% CI 1.39-1.65, p< 0.001) and 30-day mortality (OR 1.48, 95% CI 1.24-1.75, p< 0.001).

Conclusion:

mFI was found to be associated with the incidence of postoperative complications and mortality in older surgical patients with gastrointestinal cancer. mFI serves as a useful tool for preoperative risk stratification in this vulnerable population.

07.09 OMISSION OF RADIOTHERAPY IN THE U.S. AFTER BREAST CONSERVATION IN THE POST-NEOADJUVANT SETTING.

Posted on January 25, 2017

A. C. Esposito1, J. L. Crawford2, E. R. Sigurdson2, E. Handorf4, R. J. Bleicher2  1Lewis Katz School Of Medicine At Temple University,Philadelphia, PA, USA 2Fox Chase Cancer Center,Surgical Oncology,Philadelpha, PA, USA 4Fox Chase Cancer Center,Biostatistics,Philadelphia, PA, USA

Introduction:

Breast conservation has been standard of care for invasive breast cancer, and complete treatment via breast conservation therapy (BCT) is comprised of both breast conservation surgery and radiotherapy (RT). Neoadjuvant chemotherapy (NACT) is given to make tumors having skin or chest wall involvement resectable, and to downstage those too large for local excision to a size amenable to BCT. On reviewing patterns of postneoadjuvant RT administration in the National Cancer Database (NCDB), we found that significant numbers of BCT patients did not receive RT after NACT. This study was performed to determine what factors predicted RT omission, which is nonstandard therapy.

Methods:

NCDB cases were reviewed for women having unilateral, invasive, non-inflammatory, non-metastatic, clinical stage 2-3 breast cancer, treated with NACT and subsequent BCT between 2008-2012. Only those starting NACT <90 days after diagnosis but 80-270 days preoperatively, and RT 0-20 weeks post-op were included. We used Pearson Chi-square and tests for trend to determine the relationship between patient, tumor, and facility factors and receipt of RT.  We then determined simultaneous effects of these factors using multiple logistic regression with robust standard errors to account for within-hospital clustering.

Results:

10,220 patients were identified who received NACT prior to breast conservation with 974 (9.53%) not receiving RT after surgery. The majority of patients were white, female, had invasive ductal carcinoma, and received therapy at comprehensive community cancer centers or academic/research programs. Predictors of failure to receive RT included older age (ORs 1.17; 95% CI: 1.09-1.27, P=<.0001), more recent year of diagnosis (OR 1.06; 95% CI: 1.01-1.12, P=.0267), US region (ORs varied), insurance status (ORs varied), facility type (ORs 1.48; 95% CI: 1.02-2.16, P=.0404), positive margins (ORs 1.67; 95% CI: 1.29-2.16, P=.0001), receptor status unknown (OR 1.68; 95% CI: 1.19-2.37, P=.003), and HER2 status positive or unknown (ORs varied). Factors increasing the likelihood of RT receipt included N3 disease (ORs 0.59; 95% CI: 0.36-0.95, P=.0312), known grade (ORs varied), primary tumor downstaging (ORs 0.84; 95% CI: 0.72-0.98, P=.024), and receptor positivity (ORs 0.85; 95% CI: 0.73-0.99, P=.0391). Factors having no effect on likelihood of RT included race, education, income, and Charlson comorbidity index. When excluding 314 patients in whom RT was recommended, only age, US location, receptor status, and margins remain predictors.

Conclusion:

It is encouraging that racial and socioeconomic disparities were not found among the predictors for lack of RT receipt. Unfortunately however, comorbidities did not explain the difference in treatment. It remains unclear whether some omission of RT is due to lack of physician knowledge. Further efforts may be needed to ensure that physicians and patients recognize that RT is a required part of BCT even after NACT.

07.08 Better Outcomes for Surgical Resection of Pancreatic Cancer in Academic versus Nonacademic Centers

Posted on January 25, 2017

O. Moaven1, J. M. Richman1, S. Reddy1, T. M. Wang1, C. M. Contreras1  1University Of Alabama At Birmingham,Department Of Surgery,Birmingham, AL, USA

Introduction:  Surgical resection is the only intervention known to significantly affect outcomes in patients with pancreatic cancer. It is very important to identify the factors that can improve the surgical outcomes in pancreatic cancer. In this nationwide retrospective study we compared the outcomes of tumor resection in academic vs. nonacademic centers.

Methods:  We analyzed American College of Surgeons National Cancer Database (NCDB) and included patients with pancreatic adenocarcinoma who have undergone pancreatic resection between 1998 and 2012. Unadjusted chi-square and t-tests and multivariate generalized additive models were used to compare the differences between patient characteristics and outcomes of the patients treated in academic centers vs. nonacademic centers.

Results: We identified 62858 patients with surgical resection of their pancreatic cancer. While 34245 (54.5%) had their resection at an academic (A) center versus 28613 (45.5%) at non-academic (nonA) centers (p<0.001), these proportions varied significantly over time.  Before 2007, 52.2% were in academic centers vs. 47.8% in nonacademic; in 2007 and afterwards, the proportions were, respectively, 56.7% and 43.3% (p<0.001). Overall, Patients traveled further to academic centers (mean miles: A: 75.6, nonA: 26.6, p<0.001), and waited longer for resection (mean days: A 29.2, nonA 20.9, p<0.001). Academic centers also had higher volumes with a mean of 30.1 resections per year vs. 6.9 for nonacademic (p<0.001), examined more lymph nodes per resection with an average of 13.5 vs. 11.0 (p<0.001) and were more likely to achieve negative margins 76.7 vs. 72.5% (p<0.001).  In unadjusted analyses there were several significant differences when comparing outcomes between Academic and Nonacademic centers including: overall survival (A vs. NonA: HR 0.83, 95%CI:0.81-0.84, p<0.001), 30-day post-operative mortality (A vs. NonA: OR 0.64, 95%CI:0.61-0.68, p<0.001), 90-day post-operative mortality (A vs. NonA: OR 0.68, 95%CI:0.65-0.72, p<0.001), and 30-day readmission (A vs. NonA: OR 0.94, 95%CI:0.90-0.99, p=0.03). These differences remained significant in models adjusted for patient, tumor, and operative characteristics but were not significant after surgical volume of the center was added to the adjusted model.

Conclusion: Patients with pancreatic cancer who undergo surgical resection in academic centers have better outcomes likely related to higher annual volume. Centralizing tumor resection in higher volume academic centers is an essential step to improve survival of patients with resectable pancreatic cancer.

 

07.07 Clinical Relevance of MicroRNAs in Breast Cancer Using TCGA

Posted on January 25, 2017

S. Y. Kim1, T. Kawaguchi1, L. Yan2, Q. Qi2, S. Liu2, J. Young1, K. Takabe1  1Roswell Park Cancer Institute,Surgical Oncology,Buffalo, NY, USA 2Roswell Park Cancer Institute,Biostatistics And Bioinformatics,Buffalo, NY, USA

Introduction:  MicroRNAs (miRNAs) are noncoding RNAs with 19-25 nucleotides that exert their genetic effect by either mRNA degradation or inhibition of the translation of mRNA. Dysregulations of miRNAs have been identified to play a critical role in carcinogenesis and the development of various types of cancer including breast cancer (BrCa). Some miRNAs are reported as oncogenic miRNAs that are associated with drug resistance in BrCa patients. Some miRNAs, such as miR-18a and miR-205, are reported to have both oncogenic and suppressive roles. MiR-744, on the other hand, has been reported to promote cell proliferation in vitro, but its relevance in patients is unknown. Identification of novel prognostic biomarkers typically requires a big dataset that provides sufficient statistical power for discovery and validation research. In this study we took advantage of the high-throughput data from The Cancer Genome Atlas (TCGA) as a validation cohort to evaluate the clinical relevance of miR-18a, miR-21, miR-155, miR-205,and miR-744.

Methods:  All data was obtained from The Cancer Genome Atlas (TCGA). Expression of four miRNAs, miR-18a, miR-21, miR-155, and miR-205, which have oncogenic and/or suppressive roles in BrCa patients, and another miRNA with unknown clinical relevance (miR-744) were retrieved from the GDC data portal for analyses.  After miRNA-specific thresholds were derived from the data and used to group the patients into either a high expression or low expression group, survival data was calculated using the Cox proportional hazard model.

Results: Among the 1097 breast cancer samples logged in TCGA, 1053 samples were found to contain both clinical data and microRNA-seq datasets on the miRNAs of interest. High expression of miR-18a, the miRNA that is reported to have both oncogenic and suppressive roles in BrCa patients, demonstrated significantly better prognosis (p=0.037), whereas it was not significant with miR-205 expression (p=0.070). Surprisingly, high expression levels of miR-21 and miR-155, the two miRNAs that are well known as “oncogenic” miRNAs, also demonstrated significantly better prognoses (p=0.030 and 0.005, respectively). MiR-744, which currently has unknown clinical relevance, but has been reported to be oncogenic in vitro, was also associated with a significantly better prognosis when highly expressed (p=0.027).

Conclusion: By utilizing a big dataset (TCGA) with sufficient statistical power, we found that high expression of miR-18a, miR-21, miR-155, and miR-744 were all significantly associated with better overall survival. We were able to clarify that miR-18a and miR-205 had a positive impact on survival despite some reports of their oncogenic functions. We conclude that it is necessary to reevaluate the survival impact of each miRNA whose functions are seemingly straightforward.

 

07.06 Directed Shave Margins in Breast-Conserving Surgery: Accuracy of Intraoperative Surgeon Assessment

Posted on January 25, 2017

J. Yu1, J. Yu1, L. C. Elmore1, A. E. Cyr1, J. A. Margenthaler1  1Washington University,Surgery,St. Louis, MO, USA

Introduction:
During breast-conserving surgery (BCS), additional cavity shave margins may be excised after removal of the primary specimen at the discretion of the surgeon to reduce rates of positive margins. We sought to evaluate the concordance of directed shave margins with disease on pathology and to assess the accuracy of surgeon judgment. 

Methods:
Utilizing a prospectively-maintained institutional database, we reviewed all women undergoing re-excision following breast-conserving surgery for invasive breast cancer or ductal carcinoma in situ (DCIS) from 2010-2013.  We then identified all patients who had directed shave margins taken due to clinical or radiographic suspicion during the index procedure.  Surgeon judgment was considered concordant when the shave margin that was taken corresponded to a positive or close margin on the primary tumor specimen. Positive margins were defined as invasive disease or DCIS touching the edge of the specimen, and close margins were defined as disease within 2 mm of the edge. Descriptive statistics were used in data analysis.

Results:
Of the 384 women undergoing re-excision, 99 patients had additional shave margins taken during their index procedures.  18 (18.2%) patients had invasive carcinoma alone, 27 (27.3%) had DCIS alone, and 54 (54.5%) had both. Of 191 total shave margins, an average of 1.9±0.9 shave margins were taken per patient, and the mean shave margin volume was 10.43 cm3.  Ninety-six (50.3%) shave margins were positive for invasive disease or DCIS. However, only 74 (38.7%) shave margins were taken when the corresponding primary tumor margin was positive or close. There was no difference in concordance when the shave was taken for clinical or radiographic suspicion (38.2% vs. 36.6%, p>0.05). Forty-six (24%) shave margins were positive for disease when the corresponding primary tumor margin was negative. On re-excision histology, 66 (66.7%) patients had no disease, 25 (25.2%) had DCIS, 7 (7.1%) had invasive disease and one had both (1.0%). 

Conclusion:
Surgeons are limited in their ability to accurately assess margin status intraoperatively which leads to imprecise use of directed shave margins. Implementation of routine shave margins or alternative margin assessment methods may be more likely to reduce the rates of positive margins following BCS, and further research is necessary to define the best standard of practice. 

07.05 Utility of CT Imaging in a Novel Form of High-Dose-Rate Intraoperative Breast Radiation Therapy

Posted on January 25, 2017

T. Hassinger1, K. Rea1, A. Schroen1, D. Brenin1, A. Berger2, B. Libby3, T. Showalter3, S. Showalter1  1University Of Virginia,Department Of Surgery, Division Of Surgical Oncology,Charlottesville, VIRGINIA, USA 2Thomas Jefferson University,Department Of Surgery,Philadelphia, PENNSYLVANIA, USA 3University Of Virginia,Department Of Radiation Oncology,Charlottesville, VIRGINIA, USA

Introduction:  Intraoperative radiation therapy (IORT) is an increasingly popular approach to breast conservation. A drawback to conventional breast IORT (CB-IORT) is the lack of CT imaging. We pioneered a novel method of IORT that incorporates customized, CT-based treatment planning and high-dose-rate (HDR) brachytherapy; Precision Breast IORT (PB-IORT). Our aim is to report on the unique utility of CT imaging in PB-IORT, both intraoperatively and in dosimetric planning for the first 84 patients treated in two prospective trials. 

Methods:  We retrospectively reviewed the first 84 patients who participated in prospective clinical trials of PB-IORT. The first 28 were enrolled in the phase I trial, and the subsequent 56 are enrolled in the ongoing phase II trial. All patients underwent lumpectomy, multicatheter balloon placement, intraoperative CT scan, and HDR brachytherapy treatment delivery to 12.5 Gy to 1 cm from the balloon surface. This report focuses on the intraoperative CT findings that led to clinical changes and the use of the CT to make adjustments to the dosimetry.

Results: After initial intraoperative CT, 18 patients (21.4%) had findings that prompted surgical adjustment of the balloon applicator to eradicate large air cavities and/or improve tissue conformity between the balloon and the breast tissue before planning and delivering IORT. In 1 patient, an additional intraoperative CT scan was performed to localize a biopsy clip and aid in excision to negative margin. In 66 patients (78.6%), the dosimetry plan was modified based on intraoperative CT findings in order to sculpt the radiation dose off of the chest, ribs, or skin, with 31 patients (36.9%) adjusted for one reason and 35 patients (41.7%) adjusted for more than one reason. 

Conclusion: Intraoperative CT findings were used in a majority of patients treated with PB-IORT in order to enhance tissue conformity between the breast tissue and the balloon and to sculpt the radiation dose away from normal tissues. CT imaging is not available in CB-IORT. These findings suggest the potential for clinical superiority of our technique employing intraoperative CT given its allowance for patient-specific alterations in both the area and dose of HDR brachytherapy. 

 

07.04 National Trends in Surgical Decision-Making for Early-Stage Esophageal Cancer

Posted on January 25, 2017

E. C. Sturm3, W. E. Zahnd3, J. D. Mellinger3, S. Ganai3  1Southern Illinois University,Department Of Surgery,Springfield, IL, USA 2Southern Illinois University,Center For Clinical Research,Springfield, IL, USA 3Southern Illinois University School Of Medicine,Department Of Surgery,Springfield, IL, USA

Introduction:
Despite high morbidity after esophagectomy, improvements in perioperative management and preoperative staging have led to satisfactory outcomes at high-volume centers. While management of early-stage cancers have recently transitioned to include endoscopic interventions for T1a and induction chemotherapy for high-risk T2 adenocarcinomas, esophagectomy has been the mainstay of management of early-stage (I-IIA) esophageal cancer. This study examines whether or not patients with early-stage, resectable cancers are being offered surgery.

Methods:
A retrospective cohort study was conducted on resectable esophagus cancer cases in the National Cancer Database diagnosed between 1998 and 2012, including AJCC clinical T1 or T2 patients with no lymph node or metastatic involvement. We determined frequencies and percentages to describe the demographic characteristics of patients. We also categorized patients by surgical decision: had surgery and reasons why if they did not have surgery (not offered, refused, comorbidities or advanced age, unknown or died before surgery). We performed a chi-square trend test to determine trends in surgery and reasons for no surgery over time. We performed chi square analysis to compare surgery decisions by demographic characteristics.

Results:
A total of 22,994 patients with T1 or T2 esophageal cancer were identified. 10,150 (44%) had surgery while 12,844 (56%) did not have surgery. Of those who did not undergo surgery, reasons given included advanced age or comorbidities (5.8%), refused (2.5%), unknown or died before surgery (6.7%), or that surgery was not offered to the patient (40.9%).  The proportion of patients who had surgery increased from 33% in 1998 to 50% in 2012 while the proportion who were not offered surgery decreased from 44% to 38% (p<0.001). Surgical decision varied significantly by demographic characteristics. Patients who were women, older, non-white minorities (especially blacks), uninsured or Medicaid insurance status, not treated at an academic center, less educated, and low income were all significantly associated with lower rates of surgery (p<0.001).

Conclusion:
A large proportion of patients with potentially resectable esophageal cancer are not being offered surgery. This percentage has decreased over time but is still unacceptably high. Disparities exist in types of patients who are offered surgery. Further study on surgical access disparities in esophageal cancer are warranted.
 

07.03 Surgical Strategy of Hepatic Resection with Inferior Vena Cava Resection for Liver Cancers

Posted on January 25, 2017

T. Ochiai1, D. Asano1, J. Yoshino1, S. Watanabe1, Y. Ishikawa1, N. Chiyonobu1, Y. Mizuno1, T. Sato1, H. Ueda1, Y. Iwao1, H. Ono1, Y. Mitsunori1, S. Matsumura1, D. Ban1, A. Kudo1, S. Tanaka2, M. Tanabe1  1Tokyo Medical And Dental University,Department Of Hepto-Biliary And Pancreatic Surgery,Bunkyo-ku, Tokyo, Japan 2Tokyo Medical And Dental University,Department Of Molecular Oncology,Bunkyo-ku, Tokyo, Japan

Introduction: The prognosis of patients who have liver cancer associated with inferior vena cava tumor thrombus (IVCTT) or inferior vena cava invasion (IVCI) is very poor, and effective treatment modalities are extremely limited. The objective of this study is to determine the efficacy of surgery and appropriate surgical procedures for liver malignancy with IVCTT or IVCI.

Methods: From January 2003 to December 2015, 19 patients with the liver malignancy (eight metastatic tumors, seven hepatocellular carcinomas, three intrahepatic cholangiocarcinomas and one mixed hepatocellular cholangiocarcinoma) underwent hepatectomy with concomitant IVC resection and reconstruction for eight IVCTT and eleven IVCI. We retrospectively analyzed surgical procedures and survival.

Results:Of the 19 patients, 2 underwent trisegmentectomy, 6 underwent bisegmentectomy, 5 underwent segmentectomy and 6 underwent partial hepatectomy. As IVC reconstruction, 13 underwent primary closure, 4 required 20-mm expansive polytetrafluoroethylene (ePTFE) graft and 2 required patch graft using round ligament and epicardium, respectively. During the IVC cross-clamping, 6 required infrarenal abdominal aortic cross-clamping without cardiopulmonary bypass or venous bypass, 3 required extracorporeal circulation assisting device to maintain stable hemodynamics. Median operation time was 515 minutes (range 256 to 918 minutes) and median intraoperative bleeding was 2353 ml (range 740 to 44000 ml). No hospital death was recognized. The median survival time of hepatocellular carcinoma was 27 months (range 4 to 89 months) and of colorectal liver metastases was 22 months (range 2 to 48 months). One out of 3 in intrahepatic cholangiocarcinoma is still alive 60 months after surgery. 

Conclusion:Hepatic resection with IVC resection and reconstruction for liver tumors provided acceptable outcomes in 19 patients. Considering morphology of the tumor and intraoperative hemodynamics, appropriate surgical procedures should be selected.

 

07.02 Young patients with Gastric Cancer: A National Cancer Data Base Study

Posted on January 25, 2017

I. Nassour1, A. A. Mokdad1, M. Khan1, J. C. Mansour1, A. C. Yopp1, R. M. Minter1, P. M. Polanco1, M. M. Augustine1, M. A. Choti1, S. C. Wang1, M. R. Porembka1  1University Of Texas Southwestern Medical Center,Division Of Surgical Oncology,Dallas, TX, USA

Introduction:  About 10% of patients with gastric cancer present at young age. Previous studies have reported conflicting prognosis for young gastric patients.  The aim of this study is to describe the clinicopathological characteristics, prognostic features, and outcomes from gastric cancer among young patients (18-49 years) compared to middle aged (50-69 years) and elderly patients (70 years).

Methods:  We identified patients with gastric cancer in the National Cancer Data Base between 2006 to 2013. We analyzed patient-, tumor-, and treatment-related factors among all three groups using standard statistical methods. Disease stage was divided into early stage (IA), locally advanced (IB-IIIC), and metastatic (IV). Relative survival was calculated as the ratio of observed survival to expected survival for the United States general population matched for age and sex. The log-rank test was used to compare survival rates among the groups. Cox proportional hazard analysis was used to evaluate the impact of age on survival.

Results: We identified 101,481 patients with gastric cancer with 10.5% being young, 41.6% middle aged and 47.9% elderly. Young patients were more likely to be Hispanic compared to middle-aged and elderly patients (24%, 10.8% and 7.9%, respectively; p<0.01) and were more likely to be uninsured (13.3%, 6.5% and 0.9%, respectively; p<0.01). Young patients presented more often with metastatic disease (46.3%, 37.0% and 27.6%, respectively; p<0.01), and high grade histology (79.9%, 67.4% and 61.7%, respectively; p<0.01). Young patients had increased 5-year relative survival compared to middle aged and elderly patients for both early disease (76.8%, 75.0% and 56.1%, respectively; p<0.01) and locally advanced disease (39.1%, 36.7% and 30.7%, respectively; p<0.01). For metastatic disease, young and middle aged patients had prolonged 5-year relative survival compared to elderly patients (5.2%, 5.2% and 3.0%, respectively; p<0.01). When adjusting for relevant patient-related, tumor-related, and treatment-related factors, young patients had favorable prognosis when compared to middle aged (HR=1.07, 95%CI: 1.04-1.1) and elderly patients (HR=1.44, 95%CI: 1.40-1.48).

Conclusion: While young patients were more likely to present with advanced disease, more aggressive histology, and lower socioeconomic status, younger age was associated with prolonged relative survival compared to their elderly counterparts when adjusting by stage and with prolonged overall survival when adjusting by other relevant clinical and treatment factors. Future studies are needed to determine if the differences seen in presentation and outcome between age groups are due to socioeconomic factors, ability to tolerate treatment, and/or inherent biology.

 

« Previous Page
Next Page »