46.03 The Impact of Obesity on Specialty-specific 30-day Complications in Pediatric Surgery

Posted on January 26, 2017

A. T. Train1, S. Cairo1, H. A. Meyers1, C. M. Harmon1,2, D. H. Rothstein1,2  1Women And Children’s Hospital Of Buffalo,Pediatric Surgery,Buffalo, NY, USA 2State University Of New York At Buffalo,Buffalo, NY, USA

Introduction:
Patient obesity has been implicated as a causative factor for surgical site infection and other surgical complications in adult patients. We sought to examine the effects of obesity on specialty-specific surgical outcomes in children.

Methods:
Retrospective cohort study using the Pediatric National Surgical Quality Improvement Program, 2012-2014. Patients included those aged 2-18 years who underwent a surgical procedure in one of the following categories: general surgery, neurosurgery, orthopedic surgery, otolaryngology, plastic surgery or urology. Obesity (body mass index ≥ 95th percentile, adjusted for age and gender) was the primary predictor of interest. Outcomes of interest were post-operative complications (wound complications, non-wound complications, and composite overall morbidity) and operative times. Odds ratios for development of postoperative complications were calculated using stepwise multivariate regression analysis after adjustment for demographics and comborbidities.

Results:
When examining all operated patients, obesity was associated with a significantly greater risk of wound complications (OR 1.24, 95% CI: 1.13-1.36), but decreased risk of non-wound complications (OR 0.68, 95% CI: 0.63–0.73) and overall morbidity (OR 0.79, 95% CI: 0.75–0.84). Obesity was not a significant factor in predicting postoperative complications in patients undergoing neurosurgery, otolaryngology, or plastic surgery procedures (Table). Anesthesia times and operative times were significantly longer for obese patients undergoing most types of pediatric surgical procedures. 

Conclusion:
Obesity confers an increased risk of wound complications in some but not all pediatric surgical specialties, and is associated with overall decreased non-wound complications and morbidity. These findings suggest that the relationship between obesity and postoperative complications is complex and may be more dependent on underlying procedure or specialty-related factors than previously suspected. Future research should attempt to identify these factors and how they may be potentially protective against obesity-related surgical complications.
 

46.02 Firearm Injuries Due to Legal Intervention in Children and Adolescents: A National Analysis

Posted on January 26, 2017

A. E. Wagenaar1, J. Tashiro1, N. Joudi1, E. A. Perez1, J. E. Sola1  1University Of Miami,Surgery,Miami, FL, USA

Introduction:  Firearm injuries related to legal intervention have come under scrutiny due to recent events. We sought to analyze this rare event using a large, population based database.

Methods:  The Kids’ Inpatient Database (1997-2012) was searched for firearm injuries due to legal interventions (ICD-9-CM code E970) requiring inpatient admission in children < 20 years of age. Cases were weighted to provide national estimates. The Brady Campaign Criteria were used to identify lenient vs. strict gun law states.

Results: Overall, a total of 275 cases were identified during the study period, with a 7.5% mortality rate. Case incidence reached a peak at 1.0 per 100,000 admissions in 2006, significantly increased from a low 0.2 per 100,000 admissions in 1997, p<0.001.

Patients were predominantly male (97%). African Americans (44%) represented the largest racial group, followed by Hispanics (30%) and Caucasians (20%). Mean age was 17.5 ± 2.08 years. Patients were most frequently insured by Medicaid (33%) or a private payer (24%), whereas the remainder (22%) were uninsured.

Hospital admissions most frequently occurred at urban teaching hospitals (81%). Cases occurred most frequently in the Southern U.S. (44%), followed by the Western U.S. (35%). Most patients presented to non-children’s hospitals (97%). Mean hospital admission cost was 27,507 ± 40197 USD, whereas mean charges amounted to 75905 ± 116622 USD. Cases of intervention occurred in lenient (56%) gun law states, whereas the remainder occurred in strict (41%) and neutral (3%) states.

When analyzed by race, Caucasians (16%) had a significantly higher mortality rate when compared to African Americans (5%), p=0.03.

Conclusion: An analysis of this very specific injury mechanism demonstrates important findings, which are difficult to collect from conventional data sources. Future research will help in contributing to the objective analysis of this politically charged subject.
 

46.01 Cervical Spine Injuries in Young Children: Pattern and Outcomes in Accidental vs. Inflicted Trauma

Posted on January 26, 2017

J. E. Baerg1, A. Thirumoorthi1, R. Vannix1, R. Hazboun1, P. Krafft1, A. Zouros1  1Loma Linda University School Of Medicine,Pediatric Surgery,Loma LInda, CA, USA

Introduction:
To compare the pattern of cervical spine (c-spine) injuries  and outcomes in children younger than 36 months with  a closed head injury from confirmed inflicted vs. accidental trauma.

Methods:   After IRB approval, a prospective cohort study was performed.  Data were collected between July 2011 and January 2016.  Inclusion criteria were: age below 36 months, loss of consciousness after confirmed inflicted or accidental trauma, and one of the following on initial computed head tomography (CT):  subdural, intra-parenchymal, intraventricular, subarachnoid hemorrhage or cerebral edema.  An imaging protocol of brain and neck magnetic resonance (MR) imaging and MR angiography was obtained within 48 hours of admission.  Variables collected were:  age, gender, initial head CT findings, extremitiy fractures, retinal hemorrhages, brain MRI findings at 48 hours, Injury Severity Score (ISS), Glasgow Coma Score (GCS), and any c-spine injury identifed on imaging protocol.  Outcomes were recorded.  Variables were compared by t-test for continuous and chi-square test for categorical variables. A p<0.05 achieved significance.

Results:  During the study period, 73  were identified, 67% males.  The median age was 11 months (range:  1-35 months).  There were 52 with inflicted (71%) and 21 with accidental trauma (29%).  Ten cervical spine injuries were identified (14%) by imaging protocol, 7 from inflicted (13%)  and 3 with accidental trauma (14%).  There were significantly more subdural hematomas (p= 0.03), brain infarction at 48 hours (p=0.01), retinal hemorrhages (p=0.02) and extremity fractures (p=0.02) in the inflicted group.  The inflicted group had a significantly higher ISS (p= 0.02) and lower GCS (p=0.01) on admission.  No other significant differences were identifed.  C-spine injuries identified in the inflicted group were:  ligamentous injury (2), atlanto-occipital dissociation (AOD) (1), vertebral artery shear injury (1), cord injury with cord epidural hematoma (2) and isolated cord epidural hematoma (1).  C-spine injuries identifed in the accidental group were;  ligamentous injury (1), AOD (1), and cord epidural hematoma (1).  Seven died, all in the inflicted group  (9.5%).  One mortality had an AOD, and 6/7 mortalities had no c-spine injury.  Six children with inflicted trauma and a c-spine injury survived with neurologic devastaion.  Three with accidental trauma and a c-spine injury survived without disability, including one that underwent successful operation for AOD.

Conclusion:  When compared to young children with accidental trauma and a closed head injury, children with inflicted trauma have a similar incidence of c-spine injury, a higher mortality, more extensive initial injuries, more long-term disability and a pattern consistent with shaking.

 

45.20 Is Contralateral Prophylactic Mastectomy an Increasing Topic of Conversation on Online Cancer Communities?

Posted on January 26, 2017

R. A. Marmor1,2, W. Dai2, X. Jiang2, S. L. Blair1, J. Huh2  1University Of California – San Diego,Surgery,San Diego, CA, USA 2University Of California – San Diego,Biomedical Informatics,San Diego, CA, USA

Introduction:  The increased uptake of contralateral prophylactic mastectomy (CPM) among breast cancer patients remains a poorly understood phenomenon. We hypothesized that the increased rate of CPM is represented in conversations on an online breast cancer community and may contribute to patients choosing this operation.

Methods:  We downloaded 328,763 posts and dates of creation from a publicly available online breast cancer community from August 1, 2000- May 22, 2016. We then performed a keyword search to identify posts which mentioned breast cancer surgeries: contralateral prophylactic mastectomy (n=7,095), mastectomy (n=10,889) and lumpectomy (9,694). We graphed the percentage of CPM-related, lumpectomy-related and mastectomy-related conversation over time. We also graphed the frequency of posts which mentioned multiple operations over time. Finally, we performed a qualitative review of 350 posts to identify what types of conversations contributed to the trends we observed.  

Results: Surgically-related posts (e.g., mentioning at least one operation) made up a small percentage (n= 27,678; 8.4%) of all posts on this community.  The percentage of surgically-related posts mentioning CPM was found to increase over time (Figure 1), whereas the percentage of surgically-related posts mentioning mastectomy decreased over time.  Among posts that mentioned more than one operation, mastectomy and lumpectomy were the procedures most commonly mentioned together, followed by mastectomy and CPM. There was no change over time in the frequency of posts that mentioned more than one operation. Our qualitative review found that the majority of posts mentioning a single operation were unrelated to surgical decision-making; rather the operation was mentioned only in the context of the patient’s cancer history. Conversely, the majority of posts mentioning multiple operations centered around the patients’ surgical decision-making process. 

Conclusion: CPM-related conversation is increasing on this online breast cancer community, while mastectomy-related conversation is decreasing. These results appear to be primarily informed by patients reporting the types of operations they have undergone, and thus appear to correspond to the known increased uptake of CPM.

 

45.19 Do Medical Comorbidities Biologically Worsen Breast Cancer?

Posted on January 26, 2017

I. Woelfel1, L. Fernandez4, M. Idowu2, K. Takabe3  1Virginia Commonwealth University,School Of Medicine,Richmond, VA, USA 2Virginia Commonwealth University,Department Of Pathology,Richmond, VA, USA 3Roswell Park Cancer Institute,Buffalo, NY, USA 4Virginia Commonwealth University,Department Of Surgical Oncology,Richmond, VA, USA

Introduction:  Breast cancer is the most prevalent cancer among females and one out of eight US women will be diagnosed during their life time. Despite advances in the diagnosis and treatment of this disease, we still lose close to 40,000 women to breast cancer every year. Given the fact that the risk of developing breast cancer increases with age and that life expectancy is increasing overall, we often see older patients with multiple medical comorbidities with breast cancer in the clinic. Recently it has been proposed that generalized chronic inflammation, such as obesity, biologically worsens breast cancer. Therefore, we hypothesized that the patients with multiple medical comorbidities will have biologically worse cancer, which will be reflected as shorter disease free survival. 

Methods:  A database including study participants from the year 2007 to 2012 was created to include significant health co-morbidities including individual behavioral and biological determinants of health gathered from the electronic medical record. This included smoking history, depression, hypertension, hyperlipidemia, diabetes, coronary artery disease, obesity, endocrine dysfunction and history of drug use. Disease free survival for each patient was determined by calculating the time elapsed between the date of diagnosis and either the patient’s date of relapse or the most recent follow-up without evidence of disease. The determinants of health data points were then used to classify patients into groups according to the number of comorbidities. A one-way analysis of variance was then performed to determine if there was a difference in the means between the groups. 

Results: 177 subjects were included in the analysis with dates of diagnosis ranging from 2007 to 2012. The average age at diagnosis was 56 years. The most prevalent comorbidities were hypertension and obesity (BMI over 30) with 43% of our sample carrying those diagnoses. Others included smoking (21%), diabetes (15%), hyperlipidemia (19%), coronary artery disease (6%), thyroid dysfunction (7%), depression (9%) and history of drug use (4%). Upon analysis of the additive effect of these chronic conditions we found no significant difference in the disease free survival between people with zero comorbidities and those with up to 5 of these comorbidities. (F = 1.40, P = 0.22) The average disease free survival was 1,412 days.

Conclusion: Our results show that the population of women who face breast cancer is heterogeneous with a wide variety of comorbidities. In contrast to our expectation, our data shows that these comorbidities have little impact on the disease free survival. 
 

45.18 Predicting the Invasion Depth of Remnant Gastric Cancer

Posted on January 26, 2017

K. Tokizawa1, H. Takeuchi1, M. Inoue1, M. Takahashi1, K. Fukuda1, R. Nakamura1, K. Suda1, N. Wada1, H. Kawakubo1, Y. Kitagawa1  1Keio University School Of Medicine,Department Of Surgery,Shinjuku-ku, TOKYO, Japan

Introduction: To evaluate the adequacy of endoscopic treatment for remnant gastric cancer (RGC), we investigated the clinicopathological characteristics and the accuracy of preoperative prediction of the invasion depth of RGC.

Methods: We retrospectively surveyed 100 RGC patients who underwent surgical resection or endoscopic treatment in Keio University Hospital between January 2000 and June 2016.

Results: During the mentioned-above period, 75 patients with RGC underwent gastrectomy and 34 patients underwent endoscopic treatment, nine of whom underwent additional surgical resection after non-curative endoscopic treatment. Initial gastric diseases included benign (22 patients, 22.0%) and malignant diseases (78 patients, 78.0%). Reconstructions by initial surgery included Billroth I (69 patients, 69.0%), Billroth II (15 patients, 15.0%), and Roux-en-Y (one patient, 1.0%). Fifteen patients (15.0%) had underwent procedures other than distal gastrectomy. Among 85 cases with RGC following distal gastrectomy, the pathological depth of invasion was confined to the mucosa (M; 30 patients, 35.5%), submucosa (SM; 26 patients, 30.6%), muscularis propia (MP; 7 patients, 8.2%), subserosa (SS; 8 patients, 9.4%), and serosa-exposed (SE; 14 patients, 16.5%). In cases of pathological M, SM and MP, 41 patients (65.1%) had been properly diagnosed before the treatment. Preoperative prediction of the invasion depth was more invasive than pathological diagnosis in 9 patients (14.3%) and less invasive in 13 patients (20.6%). The accuracy rate was apt to be lower in anastomotic sites than in nonanastomotic sites (47.1% vs 71.7%, p = 0.065). Five-year survival rate in pathological stage I patients was not significantly different between patients who underwent endoscopic treatment and those who underwent surgical resection.

Conclusion: Endoscopic treatment may be reasonable option for early RGC. However, it requires careful consideration since there are still difficulties in making accurate preoperative prediction of the invasion depth of RGC.

45.17 Factors Affecting Time From Diagnosis To Surgery In Resectable Pancreatic Adenocarcinoma Patients

Posted on January 26, 2017

C. Ripat1,3, N. Merchant1,2, B. Azab1,2, J. Hu2,3, O. Carrasquillo3, D. Yakoub1,2  1University Of Miami,Department Of Surgery,Miami, FL, USA 2University Of Miami,Sylvester Comprehensive Cancer Center,Miami, FL, USA 3University Of Miami,Department Of Public Health Sciences,Miami, FL, USA

Introduction:  Pancreatic adenocarcinoma is a highly lethal disease with 53,070 patients per year in the U.S., of whom 79% die within 1 year. Early surgical resection or neoadjuvant treatment in borderline resectable patients followed by surgery remains the only possible cure. However, not all patients receive prompt management. Factors associated with increased time from diagnosis to surgical resection (TTS) have yet to be analyzed and correlated with survival time.

Methods: Pancreatic adenocarcinoma patients who underwent surgical resection were identified using the National Cancer Data Base (NCDB).  Regression analysis was performed to identify factors associated with increased TTS with or without neoadjuvant therapy. Cox proportional hazard model was used to determine if time to surgery was associated with increased mortality.

Results: 6359 patients from the NCDB were included. 5340 (Group 1) underwent surgical resection upfront and 1019 (Group 2) underwent neoadjuvant therapy followed by surgical resection. There was no difference in sex distribution in either groups. Median age was 64 years (26-89) in both groups. 87% and 88% of patients were white and 88% and 93% were non-Hispanic in Groups 1 and 2, respectively. In both groups, 48% of the patients were covered by private insurance, 43% by Medicare/Medicaid, while 2% had no insurance. In Group 1, 63% were diagnosed and 48% received care at a comprehensive community cancer program. In contrast, in Group 2, 52% were diagnosed and 68% received care at an academic/ research program. 40% patients in both groups resided in a medium to high income neighborhood whereas 11% resided in a low income neighborhood. 85% resided in an urban/metro setting. In Group 1, specialized cancer centers were predictors of decreased TTS, but academic and research programs were not. Increased age was associated with increased TTS (OR 1.15, 95% CI 1.09-1.23, p<0.01). Sex, race, ethnicity, insurance status and neighborhood income were not associated with increased TTS. In Group 2, black race was associated with increased TTS as were certain Hispanic subgroups. Residence in a metropolitan area was associated with decreased TTS as was shorter distance from hospital (OR 0.98, 95% CI 0.96-0.99, p< 0.01). Patients with adenocarcinoma grade 2 and 3 were associated with decreased TTS. Interestingly, patients with higher income had increased TTS. In subgroup analysis, Hispanic patients with private insurance or Medicare, but not Medicaid, had shorter TTS. Overall, increased TTS was significantly associated with shorter survival.

Conclusion: TTS in pancreatic adenocarcinoma patients is affected by race in certain localities as well as insurance status and accessibility to specific healthcare systems. Increased TTS had a detrimental effect on survival in these patients. Healthcare planning has to accommodate some of these factors to avoid disparity of care delivery to different populations.

45.16 Training Surgeons in Shared Decision-Making with Older Cancer Patients: Shared Benefits within Reach

Posted on January 26, 2017

N. Geessink1, Y. Schoon1, M. Olde Rikkert1, H. Van Goor1  1Radboud University Medical Center,Nijmegen, , Netherlands

Introduction: The number of cancer patients aged 65 years or older presenting for major abdominal surgery such as colorectal (CRC) and pancreatic cancer (PC) resections is rising. In frail older patients such procedures are highly associated with negative outcomes that threaten patients’ quality of life and functioning. Shared decision-making (SDM) and goal-oriented communication are widely recommended to improve treatment decision-making, deliver patient-preferred care, and improve overall outcomes. SDM is particularly applicable for surgical disorders such as rectal and pancreatic cancer where alternatives for a major operation are available. This study aimed to evaluate the EASYcare in Geriatric Onco-surgery (EASY-GO) intervention; an intervention designed to improve the SDM process in older CRC/PC patients.

Methods: The EASY-GO intervention comprised a training for surgeons in frailty assessment and SDM. After training, the EASY-GO working method was implemented by screening all patients on frailty and applying SDM. Adherence to the intervention was stimulated by training-on-the-job: surgeons received feedback post-consultation about the SDM process by a geriatric specialist. Consecutive patients aged ≥65 years with newly diagnosed CRC/PC were included at the surgical department of the Radboud university medical center, the Netherlands. Primary outcomes were patient-reported level of SDM (SDM-Q-9), satisfaction (VAS-S), involvement in decision-making (VAS-I), and decisional regret (DRS). Patient involvement was also rated by surgeons (VAS-I).

Results:Eleven surgeons were trained of whom 4 were eligible for complete evaluation since they consulted patients both before and after implementation in the study’s time frame (11 months). The 4 surgeons consulted 38 patients; 19 (15 PC,4 CRC) before and 19 (13 PC,6 CRC) after implementation. SDM-Q-9 scores increased with 3.9 special symbol2.6 (before 72.8 special symbol11.2,after 76.7 special symbol19.6;p=0.72), VAS-S with 0.8 special symbol1.3 (before 8.0 special symbol0.4,after 8.7 special symbol1.2;p=0.27), and VAS-I with 0.7 special symbol2.6 (before 6.9 special symbol2.8,after 7.6 special symbol1.6;p=0.72). DRS decreased with 7.4 special symbol17.9 (before 27.3 special symbol8.6,after 19.9 special symbol14.0;p=0.47). Surgeons’ VAS-I increased with 0.3 special symbol2.1 (before 7.4 special symbol1.5,after 7.6 special symbol0.7;p=0.47). SDM-Q-9 scores increased both in CRC (before 69.4 special symbol25.8,after 74.7 special symbol18.4;p=0.56) and PC patients (before 76.1 special symbol29.2,after 88.0 special symbol12.3;p=0.52).

Conclusions:Although statistical significance was not realized due to the small sample size, the consistent change in scores in the direction of improved decision-making strongly suggests a positive effect on SDM in this vulnerable onco-surgical patient group. The higher scores of PC patients may be explained by differences in number and duration of consultation and outcome perspective. The promising results suggest that clinically relevant improvements in patient-centeredness of this complex onco-surgery may be realized by ongoing training of surgeons in SDM. The results warrant further study on implementation of the EASY-GO intervention.

45.15 Molecular Profiling in Melanoma: Does Age Make a Difference?

Posted on January 26, 2017

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

Introduction:  Next generation sequencing (NGS) allows us to learn about the genetic components of cancer with the hope to improve detection, diagnosis, treatment, and outcomes. Our tertiary cancer center uses NGS to evaluate 50 targetable cancer-related genes. The goal of our study was to compare how age at diagnosis affects molecular profiling in patients with melanoma. 

Methods:  This was a retrospective review of a prospective database analyzing all patients with melanoma who underwent NGS. Clinical and pathologic data were evaluated.

Results: We analyzed specimens from 135 patients with malignant melanoma. Median age at diagnosis was 66(range 24-89), and 64.4% were male(n=87). At diagnosis, 48.1% of patients were below age 65(n=65) and 51.9% were age 65 or above(n=70). In patients <65, 20.0% were stage I(n=13), 55.4% were stage II(n=36), 13.8% were stage III(n=9), and 10.8% were stage IV(n=7). In patients ≥65, 12.9% were stage I(n=9), 61.4% were stage II(n=43), 10.0% were stage III(n=7), and 15.7% were stage IV(n=11). Patients <65 had an average of 1.55 gene mutations compared to ≥65 had an average of 1.64 mutations. In total, 216 mutations were identified, affecting 30 unique genes. In those <65, 4.6% had no mutations(n=3), 53.8% had 1 mutation(n=35), 26.2% had 2 mutations(n=17), and 15.4% had ≥3 mutations(n=10). In patients ≥65, 15.7% had no mutations(n=11), 45.7% had 1 mutation(n=32), 17.1% had 2 mutations(n=12), and 21.4% had ≥3 mutations(n=15). The most common mutations in those <65 were BRAF(n=29, 44.6%), NRAS(n=23, 35.4%), and TP53(n=12, 18.5%) while the most common mutations in those ≥65 were NRAS(n=22, 31.4%), TP53(n=17, 24.3%), and BRAF(n=13, 18.6%). Patients <65 with 1 or less mutations had an improved disease free survival, compared to patients ≥65 with mutations >1 (p=0.0493). 

Conclusion: Using NGS in patients with malignant melanoma, we see a difference in the most common mutations in the patient group diagnosed <65 and those ≥65. Further studies will identify and correlate specific patterns of mutation based on age with response to therapy, outcomes, and recurrence.
 

45.14 Post Mastectomy Reconstruction Rates:The Effect of Tumor Grade and Radiation Therapy in Younger Women

Posted on January 26, 2017

K. G. Reddy2, P. D. Strassle2,3, K. P. McGuire1,2  1University Of North Carolina At Chapel Hill,Division Of Surgical Oncology,Chapel Hill, NC, USA 2University Of North Carolina At Chapel Hill,Department Of Surgery,Chapel Hill, NC, USA 3University Of North Carolina At Chapel Hill,Department Of Epidemiology,Chapel Hill, NC, USA

Introduction: Despite the well-documented psychological benefits and oncological safety of post mastectomy breast reconstruction (PMBR), most breast cancer patients do not undergo PMBR. To better understand patterns of PMBR use, it is important to examine the interactions between patient/cancer factors and rates of reconstruction in the breast cancer population. Our main objectives were to determine the following: if reconstruction rates vary across age, if cancer grade and radiation therapy are associated with breast reconstruction, and if the effects of cancer grade and radiation therapy vary across age. 

Methods:

Using the Surveillance, Epidemiology, and End Results (SEER) database, we identified women ≥ 18 years old who underwent mastectomy for breast cancer diagnosed between 2000 and 2012. Women with inflammatory BC, Stage IV disease, and bilateral BC were excluded. Multivariable logistic regression, adjusting for clinicopathologic variables, was performed. Age was truncated at 30 (i.e. women ≤30 years old were considered 30 years old) and 90  due to small sample sizes in the tails, and modeled as a quadratic variable as determined by functional form assessment.

Results:

Overall, there were 161,255 women who met the inclusion criteria and 34,432 (21.4%) underwent PMBR.  Age was significantly associated with reconstruction, with younger women more likely to undergo PMBR (p<0.0001). Tumor grade and radiation therapy were both significantly associated with PMBR, p=0.05 and p<0.0001, respectively. The effects of tumor grade (p=0.0006.) and radiation ( p=0.001) were also significantly differential across age. Cancer grade only affected PMBR among women <50 years old, with higher grades being associated with increased PMBR. Similarly, while radiation affected PMBR in women <70, the effect was greatest among younger women and was associated with decreased PMBR.

Conclusion:

When compared to their older counterparts, younger breast cancer patients are more likely to undergo post mastectomy breast reconstruction, and have tumor grade and radiation be associated with reconstruction. Further study would be necessary to better elucidate how these factors affect  the clinical decision-making  process. 

45.13 Feedback for Pre-referral Medical Errors: Attitudes and Practices Among Cancer Specialists

Posted on January 26, 2017

J. S. Lee1, R. M. Kauffman2, M. C. Lee3, G. P. Quinn4, L. A. Dossett1  1University Of Michigan,Department Of Surgery,Ann Arbor, MI, USA 2Vanderbilt University Medical Center,Department Of Surgery,Nashville, TN, USA 3Moffitt Cancer Center And Research Institute,Comprehensive Breast Program,Tampa, FL, USA 4Moffitt Cancer Center And Research Institute,Department Of Health Outcomes And Behavior,Tampa, FL, USA

Introduction:

Physicians are encouraged to disclose their own medical errors and confidential institutional processes facilitate peer review, but no mechanisms exist for providing physician-to-physician feedback when errors are discovered across institutions or health systems. We sought to describe attitudes and practices regarding feedback for medical errors preceding consultation – “pre-referral errors” – that were subsequently discovered by cancer specialists. 

Methods:

We conducted face-to-face interviews with cancer specialists at multiple NCI-designated comprehensive cancer centers using a semi-structured interview guide. Interviews were audiotaped, transcribed verbatim, and independently coded for a priori and emergent themes using the constant comparative method. Open and axial coding were applied using content analysis. 

Results:

All participants were fellowship-trained cancer specialists of various disciplines (n = 30, 40% female, 60% surgeons). The median years of post-graduate training was 8 (range 6-10); median years of independent practice was 9 (range 2-35). Specialists described varying practices, attitudes, and barriers regarding physician-to-physician feedback for pre-referral errors (Table 1). Practices ranged from providing no feedback to emphasizing feedback with a focus on constructive criticism. Event-related thresholds for providing feedback varied, and participants described difficulty providing feedback across specialties. Common barriers to providing feedback included avoiding a superiority image, concern for referral patterns, lack of time, difficulty reaching referring physicians, and the lack of an established system or guidelines.

Conclusion:

All interviewed cancer specialists encountered pre-referral errors, but attitudes regarding physician-to-physician feedback varied, and consensus on practices for feedback or peer review across institutions is lacking.

45.12 Failure to Operate on Resectable Gastric Cancer: Implications for Policy Changes and Regionalization

Posted on January 26, 2017

H. A. Frohman1, J. T. Martin2, A. H. Le1, S. P. Dineen1, C. D. Tzeng1  1University Of Kentucky,Department Of Surgery,Lexington, KY, USA 2Southern Ohio Medical Center,Department Of Cardiothoracic Surgery,Portsmouth, OH, USA

Introduction: Surgical resection is the main component of multimodality therapy for resectable gastric cancer (GC). However, a significant proportion of patients never receive curative-intent surgery. The primary aims of this study were to evaluate national trends of surgical rates for resectable GC and to identify disparities and targetable risk factors associated with failure to operate.

 

Methods: The National Cancer Database was used to identify patients with resectable GC (defined as adenocarcinoma, pre-treatment clinical stage IA-IIIC), diagnosed 2004-2013. Curative-intent surgery included any gastrectomy for stage IA-IIIC and endoscopic mucosal resection for stage IA. Multivariate modeling was used to identify predictors of resection and to analyze the impact of surgery on overall survival (OS).

 

Results: Of 46,970 total patients, 18,085 (39%) with resectable GC did not receive a potentially curative operation. While rate of resection increased from 2004 to 2013 (52% vs. 62%, p<0.001), 69% of unresected patients had no listed co-morbidities. Variables independently associated with no surgery included: insurance other than private or Medicare (odds ratio, OR 1.60, 1.54), non-academic/research hospital (OR 1.16), non-Asian race (OR 1.72), male gender (OR 1.19), older age (OR 1.04), Charlson/Deyo score >1 (OR 1.17), residing in areas with median income <$48,000 (OR 1.23), areas with <13% high school degrees (OR 1.25), small urban populations <20,000 (OR 1.41), and increasing stage (reference IA; OR range 1.36-3.79 from stage IB-IIIC) (all p<0.001). Among all stages, failure to resect was independently associated with reduced median OS (44.4 vs. 11.8mo, hazard ratio, HR 2.09, p<0.001) (Figure). In the multivariate OS analysis, the most critical factors affecting OS were resection (HR 2.09) and stage (reference IA; HR range 1.16-3.50 from stage IB-IIIC). Other independently associated socioeconomic factors, clinical predictors, and demographic variables had HRs ≤1.51.

 

Conclusions: Over one-third of patients with resectable GC fail to receive surgery, which is the greatest determinant of OS besides clinical stage. Patients without access to academic/research hospitals, without adequate insurance, and living in areas with lower socioeconomic status, are most vulnerable to insufficient GC treatment.  Suitable insurance coverage and treatment facility are the most salient (and only modifiable) risk factors for omitting surgery. To mitigate these national disparities in surgical care, policymakers should consider improving insurance coverage in underserved areas and regionalization of GC care. Only by improving these patients’ odds for resection will meaningful improvement in national outcomes be possible.

45.11 Optimal Timing of Surveillance Mammography After Breast Conservation Therapy: 6 vs 12 Months?

Posted on January 26, 2017

M. Witten1, L. Camp1, L. Aguilera1, L. Teng1, F. Philp1, B. Klepchick1, W. Poller1, T. Julian1, M. Cowher1  1Allegheny General Hospital,Breast Surgical Oncology,Pittsburgh, PA, USA

Introduction:  

Our goal is to compare unnecessary downstream work-up for suspicious findings in ipsilateral surveillance mammography at 6 vs 12 months following breast conservation therapy (BCT).

Methods:

A retrospective review of our tumor registry captured 1956 low-risk patients with stage 0-III breast cancer from Jan 2011 to Dec 2014. Patients who had a complete mastectomy or incomplete data (n= 540) were excluded. Of 1416 patients undergoing BCT, 820 had data for analysis and were divided into two groups: follow-up mammograms at 4-9 months (group A;n= 547) and 10-20 months (group B;n=670); 397 patients had data at both timepoints.  The number of callbacks for suspicious findings leading to downstream work-up, as well as recurrence rates, were compared between groups.  Patient radiation exposure and cost of imaging were determined.

Results:

Groups were well-matched for age, tumor size and grade.  A significant number of callbacks was observed in group A compared to group B (40% vs 34%; p≤0.05) leading to additional imaging with non-suspicious findings in more patients at the 6 mo timepoint. Additional views did not result in a significant increase in ipsilateral breast cancer recurrence detection between groups.  Although the majority of patients in both groups received radiation therapy after BCT (73% vs 71%) the callback rates were not affected. The average radiation dose for a two view unilateral diagnostic mammogram (UDM) is approximately 0.4 mSv. Additional imaging adds 0.2-0.4 mSV for compression and/or magnification views resulting in a total of 0.6-0.8 mSv for patients who are called back for additional views. At our institution, patient cost is estimated to be $800 per UDM.

Conclusion:

Surveillance mammography at 6 months leads to a significant number of unnecessary callbacks and downstream imaging when compared to waiting for 10-20 months to image after BCT and eliminating the early UDM did not demonstrate a significant difference in detection of recurrence. Patient radiation dose and costs, including additional imaging views required after callbacks, would be significantly reduced.
 

45.10 Racial and Ethnic Disparities in Presentation, Treatment, and Prognosis of Gastric Cancer

Posted on January 26, 2017

A. A. Mokdad1, I. Nassour1, A. Ali1, 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,Surgery,Dallas, TX, USA

Introduction:  Disparities in cancer care and survival are major obstacles to improved cancer outcomes in the United States. This study identified racial/ethnic differences in presentation, treatment, and survival among a national cohort of patients with gastric cancer. 

Methods:  Adults patients with gastric adenocarcinoma were identified in the National Cancer Data Base between 2006 and 2013. Patient demographics, tumor characteristics, treatment, and overall survival were compared among non-Hispanic whites (NHW), non-Hispanic blacks (NHB), Hispanics (HS), and Asians (AS). Racial/ethnic differences in stage-specific treatment were compared using a multivariable logistic regression model. A Cox proportional hazards model was used to compare risk-adjusted overall survival among race/ethnicity groups. The reference group was NHW unless otherwise noted.

Results: A total of 95,212 patients with gastric adenocarcinoma were identified; 63,511 (67%) patients were NHW, 14,764 (16%) NHB, 10,451 (11%) HS, and 6,486 (7%) AS. Compared to NHW, HS presented at an earlier age with 24% being diagnosed before the age of 50 (8% in NHW, p<0.01) and were more likely to have high grade tumors (grade 3 or 4: 74% vs 64%, p<0.01) and metastatic disease (33% vs 39%, p<0.01). Compared to NHW, AS were more likely to present with early stage disease (16% vs 11%, p<0.01). NHW were more likely to have cardia disease (49%), while NHB, HS, and AS presented most commonly with cancer of the antrum/pylorus (31%, 25%, and 34%, respectively). Overall, NHB were least likely and AS were most likely to receive any treatment (70% vs 78%, p<0.01). Among patients that underwent tumor resection, NHW were more likely to receive neoadjuvant therapy (21% vs 9% (NHB), 12% (HS), and 6%(AS), p<0.01) and less likely to receive adjuvant therapy (24% vs 37% (NHB), 37% (HS), and 37% (AS), p<0.01). On multivariable analysis, of the patients without metastatic disease, AS were more likely to undergo resection (odds ratio [OR], 1.4; 95% confidence interval [CI], 1.27-1.53) while NHB were the least likely (OR, 0.80; 95% CI, 0.75-0.85). In addition, for patients with metastatic disease, NHB were the least likely to receive systemic chemotherapy (OR, 0.88; 95% CI, 0.82-0.95). AS, HS, NHW, and NHB had significantly different median overall survivals (24, 16, 13, and 12 months, respectively, log-rank p< 0.01). The survival advantage among AS and HS remained after adjusting for patient, tumor, and treatment factors (AS: hazard ratio [HR], 0.73; 95% CI, 0.70-0.76; HS: HR, 0.78;95% CI, 0.76-0.81).

Conclusion: Tumor presentation, treatment, and survival differ among racial/ethnic groups in gastric cancer. Notably, NHB are undertreated and HS have better survival despite more advanced disease. Racial/ethnic disparities in survival is not completely accounted for by tumor characteristics and treatment, and additional studies are need to determine the basis of these disparities.

 

45.08 Initial Experience of Rectal Cancer Staging with 7T Magnetic Resonance Imaging

Posted on January 26, 2017

J. J. Blank2, N. G. Berger2, P. M. Knechtges3, R. W. Prost3, C. Y. Peterson1, K. A. Ludwig1, T. J. Ridolfi1  1Medical College Of Wisconsin,General Surgery, Colorectal Division,Milwaukee, WI, USA 2Medical College Of Wisconsin,General Surgery,Milwaukee, WI, USA 3Medical College Of Wisconsin,Radiology,Milwaukee, WI, USA

Introduction:  The management of rectal cancer relies heavily on Magnetic Resonance Imaging (MRI) for proper staging. MRI is an invaluable but imperfect tool, with tumor depth being reported incorrectly in up to 1/3 of cases. If tumor depth is wrongly reported as too deep or lymph node status is wrongly reported as positive it will relegate the patient to additional unnecessary radiation and chemotherapy. Additionally, 10%–30% of individuals who undergo preoperative chemoradiation are found to have a complete pathologic response.  Some centers are now adopting a ‘‘watch and wait’’ approach, foregoing surgery for high operative risk individuals who have clinical evidence of a complete pathologic response. Resection is still considered standard of care in healthy individuals as no clinically available imaging technology is able to accurately predict complete response. Standard MRI technology currently operates at 3T, a unit of magnetic strength. The Medical College of Wisconsin Center of Imaging Research houses one of only twenty 7T MRIs worldwide. We hypothesize that 7.0 T MRI will accurately predict postoperative tumor depth and nodal status.

Methods:  Patients undergoing low anterior resection for rectal cancer were enrolled in the trial. Patients received neoadjuvant treatment based on current NCCN guidelines. Following excision, the surgical specimen was secured in a normal saline filled canister and subsequently imaged in the 7T MRI. A radiologist blinded to the pathologic data interpreted the specimen images for tumor depth and nodal status. Imaging data was then compared to the pathology reports to determine accuracy.

Results: Between July 2015 and July 2016, 5 patients met inclusion criteria. Radiologic and pathologic interpretation of specimen was identical regarding tumor depth in 3 of 5 patients (60%) Additionally, nodal status was correctly predicted by MRI in 3 of 5 patients (60%). One patient had perfect correlation between radiologic interpretation and pathologic specimen.  

Conclusion: Discrepancy between the pathologic and radiologic staging of the specimens was identified. This is not unexpected as the current project represents the first images of the human rectum obtained at 7T. We anticipate that 7T MRI radiologic staging will become more accurate as specific imaging characteristics of varying stages of rectal cancer are defined. 7T MRI holds promise in accurately staging rectal cancer and possibly predicting response to neoadjuvant therapy. 

 

45.07 Use of Exonal Mutation Analysis in Surgically-Treated Gastrointestinal Stromal Tumors

Posted on January 26, 2017

H. M. Dohnalek1,2,3, R. Acree1,2,3, D. Bihm1,2,3, K. Pulice1,2,3, H. Quadri1,2,3, N. G. Haddad1,2,3, L. B. Johnson1,2,3, J. Marshall1,2,3, W. Al-Refaie1,2,3  1Georgetown University Medical Center,Surgery,Washington, DC, USA 2MedStar Georgetown University Hospital,Surgery,Washington, DC, USA 3MedStar Georgetown Surgical Outcomes Research Center,Washington, DC, USA

Introduction:  While surgery remains the cornerstone treatment for Gastrointestinal stromal tumors (GIST), imatinib meselate has emerged as an effective targeted tyrosine kinase inhibitor in the adjuvant and advanced/ metastatic settings. Several clinically significant exonal mutations have been discovered in KIT and PDGFRA. Exonal mutational analysis (EMA) can identify such mutations and thus informs a patient’s prognosis and subsequent treatment strategies. Indeed, the National Comprehensive Cancer Network has recommended that patients undergo EMA following resection of high risk tumors (defined as per the Meittinen and Lasota criteria) or prior to initiation of TKI therapy in advanced or metastatic tumors. However, the use of EMA within a comprehensive cancer center remains unknown.  We hypothesize that use of exonal mutation analysis at our comprehensive cancer center is similar to that of published rates of 17.7% (Bischof et al. 2014). 

Methods:  An IRB-approved retrospective cohort study was performed on 59 patients who underwent surgical resection of pathologically-confirmed, biobanked and c-KIT positive GIST from 2006 through 2015. Patient’s electronic health records were accessed and information was then stored in a RedCap cloud-based database. To ensure data collection accuracy, two quality assurance evaluations of a 10% random selection of patients were performed. The NCCN guidelines based on the Meittinen and Lasota criteria for GIST risk stratification were used. 

Results: The most frequent location of GIST within our cohort was in the stomach (71.1%), followed by small bowel (20.3%), pancreas and biliary tract (6.8%) and large bowel (1.7%). Only 15% of our cohort were classified as high-risk tumors and 5% had metastatic or locally advanced tumors. Of our entire cohort of 59 patients, 15% underwent EMA. Only one third of high-risk and metastatic patients underwent EMA. The overall exonal mutational distributions were KIT exon 11 mutations in 78% and PDGFRA exon 18 mutations in 22%. Overall, the use of EMA has shown an increasing trend during the study period [Figure].

Conclusion: This hypothesis-generating exploratory study demonstrates uptake of EMA after resection of high-risk or advanced/metastatic GIST is relatively comparable to published rates. However, these rates are lower than guideline-recommended care. As personalized medicine becomes more common, further studies are needed to uncover reasons behind this low uptake while investigating its impact on outcomes and cost of care. 

 

45.06 IMRT is Superior to 3D-CRT for the Treatment of Anal Cancer

Posted on January 26, 2017

B. F. Gilmore1, Z. Sun1, M. A. Adam1, D. Spiegel2, B. Ezekian1, M. C. Turner1, U. Nag1, M. Palta2, J. Migaly1, C. Mantyh1  1Duke University Medical Center,Department Of Surgery,Durham, NC, USA 2Duke University Medical Center,Department Of Radiation Oncology,Durham, NC, USA

Introduction:
Current national guidelines recommend use of Intensity-Modulated Radiation Therapy (IMRT) over 3-Dimensional Conformal Radiation Therapy (3D-CRT) for anal squamous cell cancer given the benefits of reducing acute toxicity and minimizing treatment breaks. However, published data evaluating patient survival are limited. Our aim was to compare survival between patients treated with IMRT to 3D-CRT. 

Methods:
The 2004 – 2013 National Cancer Database Anal Cancer Participant Data Use File was queried for patients with anal squamous cell carcinoma who underwent chemotherapy and radiation therapy. Patients who received surgery or had metastatic disease were excluded. Patients were stratified by modality of radiation therapy (IMRT vs. 3D-CRT) and overall survival was compared between groups. 

Results:
Among 5,545 patients included, 4,758 (85.8%) underwent treatment with IMRT. Before adjustment, survival was similar between patients who received IMRT vs. those who received 3D-CRT (5-year overall survival 73% vs. 71%, p= 0.215). However, after adjustment for demographics, clinical, and tumor characteristics, patients receiving 3D-CRT had worse survival than those undergoing IMRT (adjusted hazard ratio 1.24, 95% CI: 1.046-1.468, p=0.013). Factors associated with use of IMRT include black race (odds ratio 1.60, p=0.011), academic hospital designation (odds ratio 1.496, p=0.007), and stage III disease (odds ratio 1.435, p=0.006).

Conclusion:
Use of IMRT is associated with superior survival compared to 3D-CRT for management of anal squamous cell carcinoma. In combination with the established data demonstrating improved acute toxicity profile and reduction in treatment related breaks associated with IMRT, these findings provide additional support for the current national guidelines.
 

45.05 Impact of Travel Distance on Survival from Early-Stage Breast and Pancreas Cancer

Posted on January 26, 2017

N. J. Lanzotti1, W. E. Zahnd1, J. F. Tseng2, J. D. Mellinger1, S. Ganai1  2Beth Israel Deaconess Medical Center, Harvard Medical School,Department Of Surgery,Boston, MA, USA 1Southern Illinois University School Of Medicine,Department Of Surgery,Springfield, IL, USA

Introduction:
While regionalization has been advocated for pancreas cancer due to volume-outcome relationships, it is uncertain whether centralization of care leads to access disparities for patients at greater distance from surgical services. We hypothesize that pancreas cancer patients will travel further than breast cancer patients, and that increased travel distance may negatively impact survival.

Methods:
A retrospective cohort study of the National Cancer Database from 1998-2012 was performed to assess the role of travel distance in patients with early-stage cT1-2N0 breast cancer (n=882,791) and cT1-2N0 pancreatic cancer (n=75,233). Travel distance by region was calculated for each group. Multivariable Cox regression was performed evaluating the impact of travel distance on survival controlling for age, gender, race, pathologic stage, facility type, income, insurance, region, and rural-urban designation. Travel distance was analyzed as a categorical variable (0-25 miles, 25-50 miles, 50–100 miles, 100–500 miles, and >500 miles).

Results:
Mean distance traveled was 20.6±86.3 miles for breast cancer and 46.6±151.0 miles for pancreas cancer (p<0.001). Median travel distance for pancreas cancer [12.4 miles (interquartile range, IQR 4.8-36.9)] was significantly greater than for breast cancer [8.0 miles (IQR 3.8-16.8)] across all regions (Figure, medians with IQR error bars). Travel distance differences between cancers varied by region, with New England having the smallest breast-pancreas differential (Δmedian, 1.8 miles; Δ75th%ile, 8.7 miles) and West North Central having the largest (Δmedian, 13.4 miles; Δ75th%ile, 52.7 miles). Outcome was associated with travel distance, with greatest survival benefit noted for patients with travel distance exceeding 500 miles for both breast (HR 0.80; 95% CI, 0.73-0.88; p<0.001) and pancreas cancer (HR 0.88; 95% CI, 0.81-0.95; p<0.001) compared to travel distances under 25 miles. Accounting for geography, heterogeneity in survival was noted between regions for both cancers. Rural location negatively impacts pancreas cancer (HR 1.05; 95% CI, 1.02-1.08; p<0.001), but not breast cancer survival (HR 0.98; 95% CI. 0.96-0.99; p=0.04).

Conclusion:
Contrary to our hypothesis, travel at extremes beyond the median is associated with better long-term survival for both early-stage breast and pancreatic cancer. The impact of travel distance may be influenced by region, rurality, insurance, and SES, and may also be an indicator of likelihood of receiving highly specialized care. Data are limited to those who sought treatment at an NCDB facility and likely reflect selection bias. Further exploration of access disparities in cancer care is warranted.
 

45.04 Willingness to Travel is Associated with Increased Survival for Patients Undergoing Pancreatectomy

Posted on January 26, 2017

M. Raoof1, P. H. Ituarte1, S. G. Warner1, Y. Woo1, G. Singh1, Y. Fong1, L. Melstrom1  1City Of Hope Comprehensive Cancer Center,Surgery,Duarte, CA, USA

Introduction:  Centralization of complex cancer surgery is associated with improved outcomes and recent data demonstrate extensive centralization over the past two decades. We hypothesize that certain patient factors and travel distance impose significant barriers to seeking pancreatic cancer surgery at high-volume centers.

Methods:  Using the National Cancer Database (NCDB), we analyzed patients undergoing pancreatic resections over a ten-year period (2003-2014). Socioeconomic and clinico-pathologic variables were compared between patients treated at high-volume centers (HVC, >25 cases/yr) or low-volume centers (LVC, <25 cases/yr). Multiple logistic regression was used to identify predictors of care at HVCs.

Results: A total of 35,804 eligible patients underwent pancreatic resections at 941 different hospitals. Independent predictors of pancreatic resection at HVC are summarized in Table1. Patients treated at HVCs traveled farther (Median 38.3 miles, IQR: 15.2-94.9) compared to those treated at LVCs (Median 13 miles, IQR: 5.4-35.1, p=0.0001). Ninety-day mortality was associated with distance traveled (9% for 0-10 miles vs. 6.9% for >10 miles, p<0.001). Of the patients treated at HVCs, those traveling >10miles had similar 90-day mortality to those traveling 0-10 miles (5.7% vs. 5.3%, p=0.5). These patients who travel >10 miles, are less likely to be old i.e.>60 years (OR 0.86, p=0.03), Black (OR 0.24,p<0.001), Hispanic (OR 0.25,p<0.001) and more likely to be from high income (OR 1.44,p<0.001) or high education county (OR 1.5,p<0.001).

Conclusion: Travel distance is a major barrier in access to centralized pancreatic cancer surgery, underlying significant socioeconomic and racial disparities. Patients that are able to travel farther have increased access to HVCs and lower mortality.

 

45.03 OPRT as a Predictor of Benefit from S-1 Adjuvant Chemotherapy for Cholangiocarcinoma Patients

Posted on January 26, 2017

K. Akahoshi1, D. Ban1, R. Kuboki1, S. Matsumura1, Y. Mitsunori1, T. Ochiai1, A. Kudo1, M. Tanabe1  1Tokyo Medical And Dental University,Hepato-Biliary-Pancreatic Surgery,Bunkyo-ku, Tokyo, Japan

Introduction:  To improve the prognosis of cholangiocarcinoma, we investigated biomarkers that can contribute to select the patients who can take benefit from adjuvant chemotherapy after surgical resection.

Methods:  Of 170 consecutive patients who underwent surgical resection for cholangiocarcinoma between 2004 and 2015, 26 patients who received gemcitabine adjuvant chemotherapy (the GEM group), 36 patients who received S-1 adjuvant chemotherapy (the S-1 group) and 103 patients who didn’t receive adjuvant chemotherapy (the NC group) were enrolled in this study. Propensity score matching was performed to adjust for differences in patient characteristics, then 36 patients were selected from the NC group. Immunohistochemical analysis of orotate phosphoribosyltransferase (OPRT) and human equilibrative nucleoside transporter 1 (hENT1) was performed, and the correlation between the expression and cancer recurrence was analyzed.

Results: Before the propensity matching selection, the NC group was associated with higher prevalence of distal cholangiocarcinoma (p=0.03) and higher ASA score (p=0.03) than the GEM and the S-1 groups. After the matching, there were no differences among the GEM(n=26), S-1(n=36) and the NC(n=36) groups. Immunohistochemistry of resected tumor tissues was performed. Of the 98 patients, 46 and 52 showed high OPRT and hENT1 expression, respectively. Distal cholangiocarcinoma was more likely to exhibit high OPRT and hENT1 expression (p<0.01). Among high OPRT patients, the disease free survival(DFS) rates of the GEM, S-1, and NC groups at 5 years were 28.8%, 50.8% and 22.9%, respectively. The DFS of the S-1 group was significantly better than the NC group (p=0.037) and the DFS of the GEM group was similar to the NC group (p=0.85). On the other hand, among low OPRT patients, the DFS rates of the GEM, S-1, and NC groups at 5 years were 23.1%, 18.0% and 27.6%, respectively. There were no significant differences of the DFS among them. The same analysis was performed about hENT1. No significant improve of the DFS was observed by adjuvant chemotherapy in both high and low hENT1 expression patients. Multivariate analysis of all patients (n=98) determined that residual tumor (HR=2.533; 95% CI:1.548-4.144; p=0.001) and lymph node metastases (HR=1.708; 95% CI:1.031-2.830; p=0.038) were independent prognostic factors for cancer recurrence, whereas S-1 adjuvant chemotherapy wasn’t a prognostic factor. However, multivariate analysis of the high OPRT patients (n=46) determined that S-1 adjuvant chemotherapy (HR=0.309; 95% CI:0.120-0.796; p=0.015) and lymph node metastases (HR=2.594; 95% CI:1.099-6.121; p=0.030) were prognostic factors for recurrence.

Conclusion:

S-1 adjuvant chemotherapy for cholangiocarcinoma patients was effective to improve DFS only in the patients who exhibited high tumoral OPRT expression. OPRT is expected to contribute to the optimization and personalization of adjuvant chemotherapy for cholangiocarcinoma.

 

« Previous Page
Next Page »