33.04 Resection for Anal Melanoma: Is There an Optimal Approach?

Posted on January 18, 2018

A. T. Hawkins1, T. Geiger1, R. Muldoon1, B. Hopkins1, M. Ford1  1Vanderbilt University Medical Center,Colon & Rectal Surgery,Nashville, TN, USA

Introduction:
Anal melanoma is a lethal disease but its rarity makes understanding the behavior and effects of intervention challenging.  Local resection (LR) and abdominal perineal resection (APR) are the proposed treatments for non-metastatic disease and have each gone in and out of favor over the years. We hypothesize that there will be no difference in overall survival between the two types of resection. 

Methods:
The National Cancer Database (NCDB 2004-2014) was queried for adults with a diagnosis of anal melanoma who underwent curative resection. Patients with metastatic disease were excluded.  Patients were divided into two groups – those who underwent local resection (LR) and those who underwent abdominal perineal resection (APR).  Bivariate and multivariable analyses were used to examine the association between resection type and R0 resection rate, short term survival and overall survival.  

Results:
570 patients with anal melanoma who underwent resection were identified over the study period.  The median age was 68 and 59% of patients were female.  383 (67%) underwent LR.  Rate of LR did not change significantly by year. Factors associated with the use of LR included older age, government insurance, and treatment at a high volume center. LR was associated with lower rates of R0 resection rates (LR 73% vs. APR 91%; p<0.001). Overall five year survival for the entire cohort was 20%. There was no significant difference in five-year overall survival (LR 17% vs. APR 21%; p=0.31). (SEE FIGURE)  Even when adjusting for confounding variables including age, gender, comorbidity, and R0 resection in a Cox proportional hazard multivariable model there was no significant survival difference between resection methods (OR 0.84; 95%CI 0.66-1.06; p=0.15).  In addition, there was no improvement in overall survival for patients who underwent R0 resection (OR 1.18; 95%CI 0.90-1.56; p=0.22). 

Conclusion:
Anal melanoma has an abysmal prognosis, with only 1 out of 5 patients alive at five years.  Older age, government insurance, and treatment at a high volume center were associated with local resection. Although local resection was associated with lower rates of R0 resection, there was no discernable difference in overall survival in both unadjusted and adjusted analysis. Given the known morbidity of APR resection, local resection should be considered in cases of anal melanoma.  

Figure- Kaplan-Meier Curve for Overall Survival Comparing Method of Resection

 

33.01 Use and Outcomes of Abdominal Evacuation in Ruptured Endovascular Aortic Aneurysm Repairs

Posted on January 18, 2018

M. Pherson2, J. Richman2, A. Beck1, E. Spangler1  1University Of Alabama At Birmingham,Department Of Surgery, Division Of Vascular Surgery And Endovascular Therapy,Birmingham, AL, USA 2Univiersity Of Alabama At Birmingham,Department Of Surgery,Birmingham, AL, USA

Introduction: Endovascular aneurysm repair (EVAR) for ruptured abdominal aortic aneurysm (AAA) has evolved over the past 10 years to become a feasible treatment mechanism with potentially decreased morbidity and mortality for patients with appropriate anatomy.  On an institutional basis, reports of similar rates of abdominal compartment syndrome were found following creation of EVAR protocols for treatment of ruptured AAAs, but with overall decreased mortality.  We seek to examine EVAR use over time and outcomes of abdominal evacuation in EVAR for rupture in clinical practice as assessed by a national vascular quality database.

Methods: Registry data on open AAA and EVAR repairs from 2003-2016 in the Vascular Quality Initiative (VQI) were obtained (a total of 40,450 procedures).  Our cohort was then restricted to the 3,424 cases where rupture was the indication for repair.  This cohort was analyzed for change in use of modality of repair over time, variation in repair use by region (clustered into North, South, East and West), and survival outcomes by modality of repair.  Comparisons of demographics were performed via ANOVA and chi squared analyses as indicated, and time to event analyses included Kaplan Meier curves and log rank testing.

Results: In total from 2003-2016, 3424 rupture repairs were performed within the VQI: 1605 open repairs and 1819 EVAR repairs.  Of the EVAR repairs, 1597 were performed without abdominal evacuation, while 222 required abdominal evacuation.  Trends in modality of repair over time showed a distinct rise in utilization of any form of EVAR repair from none in 2003 to above 60% of repairs by 2016.  No significant variation in use of EVAR by geographic region (north, south, east or west) was seen.  As seen in Figure 1, EVAR repairs not requiring abdominal evacuation had the greatest survival, while EVAR repairs requiring abdominal evacuation had a lower survival than open AAA repair. 
Factors which were significantly different in the group requiring abdominal evacuation included age (p=.01), and intraoperative packed red blood cell transfusion (p=.02).  However, the percentage of patients considered unfit for open repair did not differ significantly between patients receiving or not receiving abdominal exploration (28.5% vs 26.2%, p=.48).

Conclusion: EVAR use has increased over time, however the proportion undergoing abdominal evacuation has remained relatively stable.  Patients requiring abdominal evacuation after EVAR fared worse than those undergoing EVAR repair without abdominal evacuation or open AAA repair, likely as a surrogate for occurrence of abdominal compartment syndrome.

 

33.02 Safety and Efficacy of Brain Injury Guidelines at a Level III Trauma Center

Posted on January 18, 2018

G. E. Martin1, C. Carroll2, Z. Plummer2, D. Millar1, T. Pritts1, A. T. Makley1, B. Joseph3, L. B. Ngwenya2, M. D. Goodman1  1University Of Cincinnati,Surgery,Cincinnati, OH, USA 2University Of Cincinnati,Neurosurgery,Cincinnati, OH, USA 3University Of Arizona,Surgery,Tucson, AZ, USA

Introduction: Patients with mild-to-moderate traumatic brain injury (TBI) are increasingly managed primarily by trauma/acute care surgeons. The Brain Injury Guidelines (BIG) were developed at an ACS-accredited level 1 trauma center to triage mild-to-moderate TBI patients and facilitate identification of patients warranting neurosurgical consultation. The BIG have not been validated at a level III trauma center. We hypothesized that BIG criteria can be safely adapted to an ACS-accredited level III trauma center to guide transfers to a higher echelon of care.

Methods:  We reviewed the trauma registry at a level III trauma center to identify TBI patients who presented with an Abbreviated Injury Severity-Head score >0. Demographic data, injury details, and clinical outcomes were abstracted with primary outcome measures of worsening on repeat head CT, neurosurgical intervention, transfer to a level I trauma center, and in-hospital mortality.  Patients were classified using the BIG criteria. After validating the BIG in our cohort, we reclassified patients using updated BIG criteria, including: mechanism of injury, anticoagulation or antiplatelet use into BIG-2 or BIG-3, and replacing the “neurologic exam” component with stratification by admission Glasgow Coma Scale (GCS) score. 

Results: From July 2013 to June 2016, 332 TBI patients were identified: 114 BIG-1, 26 BIG-2, and 192 BIG-3. Patients requiring neurosurgical intervention (n=30) or who died (n=29) were BIG-3 with one exception. Patients with GCS <12 had worse outcomes than those with GCS ≥12, regardless of BIG classification. Anticoagulant or antiplatelet use was not associated with worsened outcomes in patients not meeting other BIG-3 criteria. The updated BIG resulted in more patients in BIG-1 (n=109) and BIG-2 (n=100) without negatively impacting outcomes.

Conclusion: The BIG can be applied in the level III trauma center setting. Updated BIG criteria can aid triage of mild-to-moderate TBI patients to a level I trauma center and may reduce secondary overtriage.

 

32.11 Risk Factors of Residual Disease After Partial Mastectomy: A Single Institution Experience

Posted on January 18, 2018

L. M. DeStefano1, L. Coffua2, E. Wilson3, J. Tchou4, L. N. Shulman5, M. Feldman6, A. Brooks7, D. Sataloff7, C. S. Fisher8  1Mercy Catholic Medical Center,Department Of Surgery,Darby, PA, USA 2Philadelphia College Of Osteopathic Medicine,Philadelphia, PA, USA 3Perelman School Of Medicine,Philadelphia, PA, USA 4Hospital Of The University Of Pennsylvania,Department Of Surgery, Division Of Endocrine And Oncologic Surgery,Philadelphia, PA, USA 5Hospital Of The University Of Pennsylvania,Department Of Medicine, Division Of Hematology And Oncology,Philadelphia, PA, USA 6Hospital Of The University Of Pennsylvania,Department Of Pathology And Laboratory Medicine, Division Of Surgical Pathology,Philadelphia, PA, USA 7Pennsylvania Hospital,Department Of Surgery, Division Of Endocrine And Oncologic Surgery,Philadelphia, PA, USA 8Indiana University School Of Medicine,Department of Surgery, Division Of Endocrine And Oncologic Surgery,Indianapolis, IN, USA

Introduction:
For women with invasive breast cancer (IBC), the incidence of a close or positive margin after partial mastectomy (PM) ranges widely in the literature from 20-70%. The additional surgery required for margins leads to a delay in adjuvant treatment and an increased emotional, financial and cosmetic burden for the patient.  Criteria for re-excision are traditionally based on the proximity of the margin.  We hypothesize that based on a more comprehensive review of the initial pathology, there are additional factors that can better predict the likelihood of finding residual disease. 

Methods:
After IRB approval, we retrospectively identified patients diagnosed with Stage I-III IBC who underwent PM and re-excision at our institution from July 2010 – June 2015. We excluded patients if they had undergone neoadjuvant chemotherapy, had multicentric disease, concurrent contralateral cancer, if current cancer was a recurrence, or if the initial surgery was an excisional biopsy. Bivariate analyses were conducted using two-sample t-tests for continuous variables and Fisher’s Exact tests for categorical variables. A multivariate logistic regression was then performed on significant bivariate analyses variables. A statistical significance was accepted for p <0.05.

Results:
We identified 425 patients who underwent PM and re-excision.  Of these patients, 241 (56.7%) were excluded. The remaining 184 patients were included in our analysis and divided into two groups; those with residual disease on re-excision (87 or 47.3%) and those without residual disease (97 or 52.7%). Patients with residual disease were more likely to have higher T and N stages (p=0.02 and p=0.03, respectively), have undergone PM with shave margins (p=0.002) and have only DCIS at their margins (p = 0.02).  Of the patients who had residual disease, pure DCIS was found in 14 (16%), invasive disease in 4 (4.6%), and both DCIS and invasive disease in the remaining 69 (79%) patients. The number of positive margins at initial surgery varied significantly between the two groups, with fewer positive margins in patients with no residual disease (p<0.001).  In a multivariate logistic regression, surgery with or without shave margins) (p=0.004), number of positive margins (p<0.011), and type of disease present at margin (p=0.026) remained predictive of residual disease at re-excision. 

Conclusion:
Our study adds to a growing body of literature on the evaluation of margins after PM.  Our data can assist in making decisions regarding the absolute need for additional surgery.  While the numbers are unable to predict patients without residual disease with necessary accuracy, they can certainly assist in highlighting which patients will likely have residual disease. Future research should focus on the clinical significance of residual burden of disease. 
 

32.09 HCAHPS Scores are Influenced by Social Determinants of Health

Posted on January 18, 2018

S. F. Markowiak1, S. M. Pannell1, M. J. Adair1, C. Das1, W. Qu1, F. C. Brunicardi1, M. M. Nazzal1  1University Of Toledo,Department Of Surgery, College Of Medicine And Life Sciences,Toledo, OHIO, USA

Introduction:

The 15-year-old Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey was created to compare hospitals by measuring and publicly reporting patients’ perspectives of their care. The survey instrument has been found to be valid and reliable for this purpose. Since the passage of the Affordable Care Act, however, the HCAHPS survey has tied hospital reimbursement to patient satisfaction – a purpose for which the survey was not designed. The purpose of this study was to determine whether HCAHPS scores were influenced by social determinants of health (SDOH).

Methods:

Data were gathered from Centers for Medicare and Medicaid Services (CMS) and US Census Bureau archives.  We created a database pairing individual hospital HCAHPS data to corresponding census measures at the county level.  FY2013 was excluded because its HCAHPS performance period did not match with a single census period.  Multivariate analysis and Pearson’s Correlation Coefficient (Pearson r) were used to test 54 SDOH against HCAHPS score.

Results:

1,150 hospitals in 136 counties were analyzed.  Of the 54 SDOH analyzed, 27 had a statistically significant negative correlation with HCAHPS score.  19 had a statistically significant positive correlation with HCAHPS score.  Hospitals in communities with higher proportions of government insured patients had statistically lower HCAHPS scores (Pearson r -0.158, p < 0.001) compared to communities with higher rates of private coverage. Hospitals in communities with larger percentages of unemployment had statistically lower HCAHPS scores compared to those with a stronger workforce (Pearson r -0.153, p < 0.001). Hospitals in communities with predominantly Caucasian ethnicity had statistically higher HCAHPS scores than those in ethnically diverse ones (Pearson r 0.153, p < 0.001). For economic SDOH, hospitals located in communities where rent costs exceed 30% of income had lower scores (Pearson r -0.207, p < 0.001). Hospitals located in communities with higher rates of poverty had statistically lower scores (Pearson r -0.045, p 0.003). Hospitals in communities with higher rates of homeownership had higher HCAHPS scores (Pearson r 0.091, p < 0.001).

Conclusion:

Of all SDOH analyzed, 85% had a statistically significant correlation with HCAHPS scores.  59% of SDOH were negative correlations, indicating lower HCAHPS scores for hospitals located in communities with higher rates of poverty, less educational achievement, and more ethnic diversity. 41% of SDOH were positive correlations, indicating that hospitals in wealthier communities with high rates of homeownership, private insurance, and Caucasian ethnicity had better HCAHPS scores. HCAHPS scores are directly tied to a hospital’s CMS reimbursement by federal law. Because of disparities in SDOH, the HCAHPS survey will shift CMS reimbursement from hospitals in poor and diverse communities to hospitals in wealthier ones.
 

32.10 Routine pre-thyroidectomy laryngoscopy is not necessary in the era of intraoperative neuromonitoring

Posted on January 18, 2018

S. Goare1, E. Forrest1, J. Serpell1,2, S. Grodski1,2, J. C. Lee1,2  1The Alfred Hospital,Monash University Endocrine Surgery Unit,Melbourne, VICTORIA, Australia 2Monash University,Department Of Surgery,Melbourne, VICTORIA, Australia

Introduction:  Routine pre-operative vocal cord (VC) assessment with laryngoscopy in patients undergoing thyroid surgery allows clear documentation of baseline VC function, aides in surgical planning in patients with preoperative palsy, and facilitates the interpretation of intraoperative neuromonitoring (IONM) findings. This has been the practice in our institution for the last 20 years. In this study, we aimed to determine the rate of pre-operative vocal cord palsy (VCP) in our patient cohort;  to evaluate the associated risk factors for preoperative VCP; and therefore, build a case for a selective approach to pre-operative laryngoscopic VC assessment.

Methods:  :  This retrospective review study recruited patients from the Monash University Endocrine Surgery Unit database from 2000 to 2016. Patients who had a VC assessment by fiberoptic laryngoscopy prior to undergoing thyroid surgery were included. Case files were reviewed for potential indicators of VCP, including hoarseness and other symptoms, previous neck surgery, largest nodule dimension, and history of head and neck irradiation. 

Results: Of the 5 279 patients who had pre-operative laryngoscopy, 36 (0.68%) patients were found to have a VCP. Of these, 16 had a nodule > 3.5 cm, 15 had a hoarse voice, 12 had previous neck surgery, and 5 had a malignant cytology. More than one risk factor was present in 11 of these patients. Furthermore, the first 3 of these features would account for all 36 patients with pre-operative VCP. Pre-operative knowledge of malignancy was associated with palsy in 5 patients. However, all of these 5 patients also presented with either a hoarse voice or a nodule > 3 cm. Therefore, malignancy by itself was not an indicator of potential palsy. Approximately two-thirds of the 5 279 included patients had none of these 3 features and also did not have a VCP. Therefore, using these 3 pre-operative factors (hoarseness, previous surgery, nodule > 3.5 cm) as selection criteria, up to two-thirds of our patients could do without a pre-operative laryngoscopy and no palsy would have been missed. As this is a retrospective study, these data need to be interpreted with that in mind.

Conclusion: Using this large dataset, we have established that a VCP is extremely unlikely in the absence of previous neck surgery, hoarseness, or a large nodule. Therefore, in the era of intraoperative neuromonitoring, where the recurrent laryngeal nerve can be directly assessed, we support a selective approach to pre-operative laryngoscopy using the aforementioned criteria. 

 

32.08 Very Early vs. Early Readmissions in General and Vascular Surgery Patients

Posted on January 18, 2018

L. N. Clark1, M. C. Helm1, S. Singh1, J. C. Gould1  1Medical College Of Wisconsin,Milwaukee, WI, USA

Introduction:  Readmission rates are an important surgical quality metric.  Readmissions up to 30 days after discharge following a procedure are the most commonly examined metric.  We hypothesize that ‘very early’ readmissions (0-3 days after discharge) have a significantly different root cause than ‘early’ readmissions (4-30 days after discharge).

Methods:  The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) datasets from 2014-2015 were used to identify patients undergoing a general or vascular surgery procedure. Patients were excluded if they died during the index admission, were discharged greater than 30 days from the operation, or did not have readmission data entered. Patient demographics, medical comorbidities present at the time of surgery, and data regarding postoperative morbidity were analyzed. Complications were graded according to the Clavien-Dindo classification.  Binary logistic regression was used to compare age, functional status, comorbidities, discharge destination and complications to determine their relationship to any 30-day readmission as well as readmission within 3 days compared to 4-30 days following discharge.

Results: A total of 850,043 patients met inclusion criteria: 55.5% female, average age 55 years (range 18-89). Of these patients, 55,212 (6.5%) were readmitted within 30 days and 13,570 (1.6%) were readmitted within three days of discharge. These very early readmissions comprised 24.6% of all readmissions (Figure 1).  When evaluating all readmissions from 0-30 days regardless of timing, age ≥ 65 (Odds Ratio [OR] 1.5; 95% Confidence Interval [CI]; 1.5-1.6, p<0.0001), ≥3 comorbidities (OR 2.7; 95% CI; 2.7-2.8, p<0.0001), preoperative functional dependent status (OR 3.1; 95% CI; 2.8-3.3, p<0.0001), discharge to facility other than home (OR 2.8; 95% CI; 2.7-2.9, p<0.0001), any grade three or four complication prior to discharge (OR 2.4; 95% CI; 2.4-2.5, p<0.0001), and any grade three or four complication after discharge (OR 84.7; 95% CI; 81.3-88.1, p<0.0001) were all identified as risk factors.  The only factor found to be significantly associated with very early readmission compared to early readmission was any grade three or four complication prior to discharge (OR 1.3; 95% CI; 1.2-1.4, p<0.0001).

Conclusion: Readmissions within 3 days of surgery constitute a large portion of all 30-day readmissions. Grade 3 and 4 complications prior to initial discharge are significantly associated with an increased risk of readmission, especially within the first 3 days. Further research is needed determine if effective and targeted strategies can be developed to prevent very early readmission. 
 

32.06 Quantifying the Degree of Postoperative Opioid Over-Prescription

Posted on January 18, 2018

D. S. Swords1,2, S. Vijayakumar1, S. Brimhall1, B. Ostlund1, P. Narayanan1, J. Prochazka1, D. E. Skarda1,2  1Intermountain Healthcare,Surgical Services,Salt Lake City, UT, USA 2University Of Utah,Surgery,Salt Lake City, UT, USA

Introduction:  Previous studies have found that surgeons commonly over-prescribe opioid pain medications to surgical patients, representing an opportunity to decrease the flow of unnecessary opioids into the community. However, few resources are available for surgeons regarding the quantity of opioid medications that patients actually require after various surgeries. The goals of this study were to assess prescribing practices of surgeons and patient opioid utilization after a wide variety of surgical procedures.

Methods:  Between January 15 and August 21, 2017, patients who underwent surgical procedures at 1 of 27 Intermountain Healthcare (IHC) facilities were sent an email-based survey 2-3 weeks postoperatively. Surveys were re-sent to initial non-responders 2 additional times at 3 day intervals. The survey included questions about preoperative and postoperative opioid utilization. The IHC Enterprise Data Warehouse was queried for information about postoperative opioid prescriptions.

Results: During the study period, 6673/21434 (31.1%) patients responded to the survey. Sixty-nine percent of patients were opioid naïve, and 31.0% had taken opioids in the month prior to surgery. The cohort was comprised of 38.2% orthopedic surgery patients, 17.6% general surgery, 7.8% ENT, 7.2% gynecology,  7.3% urology , 4.9% neurosurgery, 4.9% plastics/maxillofacial and 12.1% other specialties. Narcotic naïve patients were prescribed a median of 30 pills (interquartile range [IQR] 24, 50), but used a median of only 4 (IQR 0, 15). Patients who had taken narcotics in the month prior to surgery were prescribed a median of 30 pills (IQR 24, 60) and also took a median of only 4 (IQR 0, 20). When examined on a procedure-specific basis, there was also substantial over-prescription of opioids for most examined procedures. Results for 5 representative procedures are shown in the Table.

Conclusion: The majority of patients undergoing surgery are substantially over-prescribed opioids postoperatively, representing a significant source of unnecessary opioids into the community. Surgeons have an opportunity to increase the appropriateness of postoperative opioid prescribing by prescribing patients fewer opioids. Our next step will be to provide our surgeons with recommendations regarding the number of doses that will satisfy the pain needs of the majority of patients.
 

32.07 Cost Effectiveness of Immediate Biopsy vs. Surveillance of Intermediate Suspicion Thyroid Nodules

Posted on January 18, 2018

E. J. Kuo1, J. X. Wu1, K. A. Zanocco1  1David Geffen School Of Medicine,Section Of Endocrine Surgery,Los Angeles, CA, USA

Introduction:
In an effort to reduce the overdiagnosis and overtreatment of low-risk thyroid cancer, recent American Thyroid Association guidelines increased the size-based biopsy thresholds for some sonographic categories of thyroid nodules. However, fine-needle aspiration (FNA) biopsy continues to be recommended for intermediate-suspicion nodules greater than 1cm in diameter. We hypothesize that the quality-adjusted life expectancy of patients with sonographically intermediate suspicion thyroid nodules would be improved and costs would be decreased by raising the size threshold for biopsy from 1.0 cm to 1.5 cm.

Methods:
A Markov transition-state model was constructed to compare the cost-effectiveness of immediate FNA versus ultrasound surveillance of an incidentally detected 1.5 cm thyroid nodule with intermediate-suspicion sonographic features (hypoechoic, smooth-margined solid nodule without microcalcifications, extrathyroidal extension, or taller than wide shape). Treatment outcome probabilities and their corresponding utilities were estimated based on literature review. Nonlinear growth modeling techniques were used to predict changes in the observed nodule size over time. Effectiveness was measured in quality-adjusted life years (QALYs). Costs were estimated using Medicare reimbursement data. A 3% annual discount rate was applied to all future costs and QALYs. The threshold for cost-effectiveness was defined as an incremental cost-effectiveness ratio of less than $100,000/QALY. Univariate and multivariate sensitivity analyses were used to examine the uncertainty of cost, probability, and utility estimates in the model.

Results:
The expected cost of routine ultrasound surveillance was $3,024 with an effectiveness of 23.8 QALYs. Ultrasound surveillance was $1,053 less costly and 0.01 QALY more effective than immediate FNA, making ultrasound surveillance the dominant strategy. Ultrasound surveillance decreased the lifetime rate of surgery from 26.5% to 23.7%. Immediate FNA became cost-effective during one-way sensitivity analysis when the pretest probability of malignancy increased from 15% to 71% or the cost of ultrasound examination increased from $130 to $570. Two-way sensitivity analysis demonstrated that routine FNA was cost effective if the quality adjustment factor for observation following a benign biopsy result exceeded the quality adjustment factor for observation without a biopsy. The model was not sensitive to the cost or complication rates of surgical therapy.

Conclusion:
Ultrasound surveillance is more cost-effective than immediate FNA for small thyroid nodules with intermediate-suspicion sonographic imaging characteristics, unless the probability of malignancy exceeds 71%. This model is highly sensitive to the utility differences between patients undergoing sonographic surveillance and patients with benign biopsy results. Therefore, additional primary investigation of health-related quality of life in these groups is necessary.

32.04 Impact of Frailty on Failure to Rescue After Low Risk and High Risk Inpatient Surgery

Posted on January 18, 2018

R. Shah1, K. Attwood6, S. Arya2, D. E. Hall3, J. M. Johanning5, N. N. Massarweh4  1Henry Ford Health System,General Surgery,Detroid, MI, USA 2Emory University School Of Medicine,Division Of Vascular And Endovascular Therapy/ Department Of Surgery,Atlanta, GA, USA 3University Of Pittsburg,Center For Health Equity Research And Promotion, Veterans Affairs Pittsburgh Healthcare System,Pittsburgh, PA, USA 4Baylor College Of Medicine,VA HSR&D Center For Innovations In Quality, Effectiveness And Safety, Michael E DeBakey VA Medical Center,Houston, TX, USA 5University Of Nebraska College Of Medicine,Department Of Surgery,Omaha, NE, USA 6Roswell Park Cancer Institute,Surgical Oncology,Buffalo, NY, USA

Introduction:  Failure to rescue (FTR), or death after a potentially preventable complication, is a nationally endorsed, publically reported quality measure. However, little is known about the impact of frailty on FTR—in particular, after lower risk surgical procedures.

Methods:  Retrospective cohort study of 984,550 patients from the National Surgical Quality Improvement Program (2005-2012) who underwent inpatient general, vascular, thoracic, cardiac and orthopedic operations. Frailty was assessed using the clinically applicable Risk Analysis Index (RAI) and patients were stratified into five groups based on RAI score (<=10, 11-20, 21-30, 31-40 and >40). Procedures were categorized as low (≤1%) or high mortality risk (>1%). The association between RAI, the number of post-operative complications (0, 1, 2, 3+), and FTR was evaluated using hierarchical modeling. 

Results: Among the most frail (RAI >30) patients in the cohort, ~20% were aged 55 years or younger. Regardless of procedural risk, increasing RAI score was associated with both an increased occurrence of post-operative complications and the number of complications. For those who underwent low risk surgery, major complication rates were 3.2%, 8.6%, 13.5%, 23.8% and 36.4% for RAI scores of <=10, 11-20, 21-30, 31-40 and > 40, respectively and for patients undergoing high risk surgery, the corresponding rates of major complications were 13.5%, 23.7%, 31.1%, 42.5% and 54.4%, respectively. Stratifying by the number of complications, significant increases in FTR rates were observed across RAI categories after both low and high risk procedures (Figure 1; trend test, p<0.001 for all). Increasing RAI was associated with an increased risk of FTR that was most pronounced after low risk procedures. For instance, the odds ratios (ORs) for FTR after 1 major complication for patients undergoing a low risk procedure were 4.8 (3.7, 6.2), 8.1 (5.9, 11.2), 19.3(12.6, 29.6) and 48.8 (22.7, 104.9) for RAI scores of 11-20, 21-30, 31-40 and > 40, respectively and for patients undergoing a high risk procedure, the corresponding ORs were 2.6 (2.4, 2.8), 5.2 (4.8, 5.6), 9.3 (8.5, 10.3) and 19.5 (16.8, 22.6) respectively. 

Conclusion: Frailty has a dose-response relationship with complications and FTR that is similarly apparent after low and high risk inpatient surgical procedures.  Tools facilitating rapid assessment of frailty during preoperative assessment, may help provide patients with more accurate estimates of surgical risk and could improve patient engagement in peri-operative interventions that enhance physiologic reserve and can potentially mitigate aspects of procedural risk.

 

32.05 Bariatric Surgery Reduces the Incidence of Estrogen Receptor Positive Breast Cancer

Posted on January 18, 2018

T. Hassinger1, J. H. Mehaffey1, R. B. Hawkins1, B. D. Schirmer1, P. T. Hallowell1, A. T. Schroen1, S. L. Showalter1  1University Of Virginia,Department Of Surgery,Charlottesville, VA, USA

Introduction:  Bariatric surgery is an effective treatment for morbid obesity with long-lasting weight loss. Additionally, elevated body mass index (BMI) is known to be an important risk factor for the development of breast cancer, one of the most common cancer diagnoses among women in the United States. Therefore, we hypothesized that patients undergoing bariatric surgery would have a decreased incidence of estrogen receptor (ER) positive breast cancer when compared to a propensity-matched non-surgical cohort.

Methods:  The bariatric population for this study included all female patients that underwent bariatric surgery at a single institution between 1985 and 2015. Patients from all routine outpatient visits were identified from the clinical data repository (CDR) and matched 1:1 with bariatric patients using body mass index (BMI), relevant comorbidities, demographics, and insurance status. The primary outcome of interest was ER positive breast cancer. Chart review was performed on all patients with a breast cancer diagnosis. Univariate analyses were performed to compare the two groups.

Results: A total of 4,860 patients were included in this study, with 2,430 in both the bariatric surgery and non-surgery groups. Median follow-up time from date of surgery or date of initial morbid obesity diagnosis (non-surgery group) was 5.6 years. There was no difference in median age (42.0 [35.0-51.0] vs. 42.0 [31.0-53.0]; p=0.29) or medical comorbidities aside from gastroesophageal reflux disease (713 [29.3%] vs. 149 [6.1%]; p<0.0001). Seventeen (0.7%) patients in the bariatric surgery group were diagnosed with any breast cancer after surgery compared to 32 (1.3%) patients in the non-surgery group (p=0.03). The non-surgery group had more ER positive tumors (4 [36.4%] vs. 22 [71.0%]; p=0.04) as well as larger median tumor size (p=0.02). 

Conclusion: Morbidly obese female patients who underwent bariatric surgery were found to have fewer subsequent diagnoses of any breast cancer and ER positive breast cancer when compared to a propensity-matched cohort. These results suggest the possibility of an oncologic benefit to weight-loss surgery.

32.03 Operating Room Teams: Does Familiarity Make a Difference?

Posted on January 18, 2018

S. Fitzgibbons1,2, S. Kaplan3, X. Lei3, S. Safford5, S. Parker4  1MedStar Georgetown University Hospital,Surgery,Washington, DC, USA 2Georgetown Univeristy Medical Center,Washington, DC, USA 3George Mason University,Psychology,Fairfax, VA, USA 4Virginia Tech Carilion School Of Medicine And Research Institute,Human Factors,Roanoke, VA, USA 5Carillion Clinic,Pediatric Surgery,Lynchburg, VA, USA

Introduction: The composition of any given operating room team may vary procedure-to-procedure. Studies in healthcare have documented that greater familiarity between certain team member pairs or dyads (ex. surgeon and scrub) corresponds to improved effectiveness, with outcomes ranging from shorter cross clamp times during cardiopulmonary bypass to shorter operative times during mammoplasty. We sought to further our understanding of this effect beyond simple dyads by developing an OR team familiarity score reflective of the larger and more complex group, and determining the impact of the larger group familiarity on surgical processes and clinical outcomes.  

Methods: Data from a diverse, primarily urban healthcare system including 6 acute-care hospitals was extracted from a system-wide electronic medical record.  All knee arthroplasty cases performed between 2013 and 2016 were included in the data set.  Information regarding individual OR team participants and their roles in the surgery were collected, in addition to patient demographics (ASA class, age, gender, race, ethnicity), case information (surgical procedure, date and time of the operation)  and outcome variables (length of procedure, length of hospital stay).  Team familiarity was calculated using a previously published formula from Huckman, Staats, and Upton (2009).  A multilevel regression (i.e., random coefficient modeling) framework was applied to examine the impact of a team’s familiarity score on case length and post-op length of stay. In addition, specific familiarity scores for each possible dyad on the team was calculated and analyzed. Dyads were defined as pairs of core team members: surgeon, scrub, circulator, anesthesiologist.

Results:A total of 4546 knee arthroplasty cases were included in the data set with an average case length of 92.68 minutes and an average length of hospital stay of 3.22 days.  When controlling for patient age, gender, hospital, and ASA class, a team’s familiarity score during a case was significantly associated with a shorter case length, with 10 previous team member interactions predicting a decreased case length of approximately 1.1 minutes (p=.012).  Similarly, an increased team familiarity score predicted a decreased length of stay, with 10 previous team member interactions predicting a decrease in hospital length of stay of 0.1 days.  With respect to the impact of specific dyad familiarity, all dyads involving the circulator predicted a shorter length of hospital stay, while all three dyads between the surgeon, scrub and circulator predicted a shorter case length.

Conclusion:Overall team member familiarity in the operating room is associated with a small but significant decrease in the case length and hospital length of stay for patients undergoing total knee arthroplasty.

 

31.09 Adjuvant Chemotherapy after Neoadjuvant Chemoradiation in Esophageal Cancer

Posted on January 18, 2018

C. Takahashi1, R. Shridhar2, A. Patel3, J. Huston4, K. Meredith3  1Midwestern University,Glendale, AZ, USA 2University Of Central Florida,Orlando, FL, USA 3Florida State University,Tallahassee, FL, USA 4Sarasota Memorial Hospital,Sarasota, FL, USA

Introduction:  Patients with locally advanced esophageal cancer (EC) have poor long-term survival despite improvements in multi-modality care. Neoadjuvant chemoradiation (NCR) followed by surgical resection remains standard of care. However, the utilization of adjuvant therapy continues to be debated. Our study reviews the effectiveness of adjuvant therapy after neoadjuvant therapy in resected EC.

Methods:  Utilizing the National Cancer Database (NCDB) we identified patients with esophageal cancer who underwent NCR followed by esophagectomy and received adjuvant therapy compared to those who did not. Propensity score matched (PSM) analysis was performed. Baseline univariate comparisons of patient characteristics were made for continuous variables using both the Mann-Whitney U and Kruskal Wallis tests as appropriate. Pearson’s Chi-square test was used to compare categorical variables. Unadjusted survival analyses were performed using the Kaplan-Meier method comparing survival curves with the log-rank test. All statistical tests were two-sided and α (type I) error <0.05 was considered statistically significant.

Results: We identified 1,816 patients from the NCDB with EC, 1,664 (91.6%) with adenocarcinoma and 134 (7.4%) with squamous cell carcinoma. Both the adjuvant therapy group and the no adjuvant group had 908 patients with a median age of 60 years (26-83). There were 1,596 (87.9%) males and 220 (12.1%) females. Location of the tumor was 121 (6.7%) middle, 1,267 (7.0%) lower, and 371 (20.4%) at the gastroesophageal junction. Univariate analysis revealed age, R0 resection, T-stage, N-stage, grade, <10 lymph nodes removed and adjuvant therapy were predictors of survival. All patients who received adjuvant therapy revealed greater median and overall survival, 36.4 months and 34.5% versus 30.9 months and 33.2%, p=0.02. Node negative patients did not show a significance in survival with adjuvant therapy 57.2 and 55.4 months respectively, p=0.4. However node positive patients demonstrated improved median and overall survival with adjuvant therapy 31.1 months and 27% respectively compared to the no adjuvant therapy group 25.7 months and 24.3%, p=0.03. Multivariate analysis revealed node positive patients T-stage (p=0.002), R0 resection (p<0.001), and number of lymph nodes removed (p<0.001) were predictors of survival. Adjuvant therapy failed to be a predictor of survival (p=0.2). However, PSM revealed that patients who received adjuvant therapy exhibited an improved median survival over those who did not 36.4 months and 30.9 months, p=0.02.

Conclusion: Adjuvant therapy in all EC patients after neoadjuvant therapy does show improved median and overall survival. Similar to other studies, R0 resection and T-stage continue to influence survival. However, node negative EC patients were found to have no survival benefit with the addition of adjuvant therapy.

 

31.10 Patterns of Disease Recurrence Following Neoadjuvant Therapy for Localized Pancreatic Cancer

Posted on January 18, 2018

C. Barnes1, M. Aldakkak1, K. Christians1, C. Clarke1, P. Ritch2, B. George2, M. Aburajab5, M. Griffin4, B. Erickson3, W. Hall3, D. Evans1, S. Tsai1  1Medical College Of Wisconsin,Department Of Surgery,Milwaukee, WI, USA 2Medical College Of Wisconsin,Department Of Hematology And Oncology,Milwaukee, WI, USA 3Medical College Of Wisconsin,Department Of Radiation Oncology,Milwaukee, WI, USA 4Medical College Of Wisconsin,Department Of Radiology,Milwaukee, WI, USA 5Medical College Of Wisconsin,Department Of Gastroenterology And Hepatology,Milwaukee, WI, USA

Introduction: Among patients with pancreatic cancer (PC) who are treated with a surgery-first approach, median survival is approximately two years and over 20% have local disease as the first site of recurrence; likely related to the high rates of node positive (~60%) and margin positive (~40%) disease. In contrast, neoadjuvant therapy and surgery have been associated with median survival durations of greater than three years and decreased rates of lymph node and margin positivity. The improved survival implies a greater level of systemic disease control; the importance of local disease control is controversial largely due to a lack of available data.

Methods: Consecutive patients with localized PC who received neoadjuvant therapy and surgery underwent post-treatment radiographic surveillance at 3-4 month intervals for the first 2 years and at 6 month intervals thereafter. The first site(s) of failure was classified as local recurrence (LR) for peripancreatic or perivascular recurrences, regional recurrence (RR) for peritoneal or abdominal wall recurrences, and distant recurrence (DR) for all other recurrence sites.

Results: Neoadjuvant therapy and surgery was completed in 231 consecutive patients; 115 (50%) with resectable and 116 (50%) with borderline resectable PC. Neoadjuvant therapy consisted of chemoradiation (n=75, 32%), chemotherapy alone (37, 16%), or both (119, 52%). Of the 231 patients, 137 (60%) had node negative disease and 207 (90%) had margin negative resections. Postoperative adjuvant therapy was completed in 138 (60%) of the 231 patients, including 27 (12%) patients who received adjuvant chemoradiation. At a median follow-up of 24.3 months, disease recurrence was present in 128 (55%) of the 231 patients (the first site(s) of recurrence are summarized in Figure 1). Of the 231 patients, 221 (96%) received radiation and 10 (4%) did not. Isolated LR occurred in 3 (30%) of the 10 patients with no radiation and 16 (7%) of the 221 patients who received radiation (p=0.04). Median overall survival (OS) was 43 months; not reached, 31.5, 19.4, and 24.8 months for patients with no recurrence, isolated LR, any RR without DR, and any DR, respectively.

Conclusion: Patients who successfully complete all intended neoadjuvant therapy and surgery have a median OS greater than three years, and a greater than 50% reduction in isolated LR. Despite the increased length of survival observed with neoadjuvant therapy, LR rates have not increased. As survival duration increases, neoadjuvant chemoradiation may be an important treatment component in minimizing isolated LR, which may be a preventable cause of patient death. 

 

31.08 Online Information on Surgery for Pancreatic Cancer is Often Inadequate for Shared Decision-Making

Posted on January 18, 2018

C. Zhang1, A. Yang1, A. Halverson1  1Northwestern University,Chicago, IL, USA

Introduction:
Decision making regarding surgery for pancreatic cancer may be difficult for patients as surgery can improve survival but can also negatively impact quality of life. In order to more actively participate in decision making, patients often seek information on the Internet. The aim of this project was to assess the quality of publicly available online information regarding surgery for pancreatic cancer.

Methods:
This study was a cross-sectional survey of patient-centered websites that address surgery for pancreatic cancer. Two search engines (Google, Bing) were queried with the terms “pancreatic cancer treatment”, “pancreatic cancer surgery”, “Whipple procedure”, and “pancreaticoduodenectomy” to identify websites of interest. Each website was evaluated using the DISCERN instrument (www.discern.org.uk), a validated questionnaire developed to analyze written consumer health information on treatment choices. An additional questionnaire was used to evaluate website content specifically for pancreatic cancer surgical treatment. Two healthcare providers (surgeon, medical student) reviewed each website independently and inter-rater reliability (IRR) was calculated. In addition, one pancreatic cancer patient and one family member analyzed a randomly selected subgroup of study websites using the DISCERN instrument.

 

Results:

We identified 93 distinct websites; 45 met inclusion criteria. Website affiliations included: health care organizations (44%), non-profit organizations (22%), open-access general information (22%), and government/professional websites (11%). Using DISCERN, the two healthcare providers identified that only 24% of the websites had clear aims, 31% had identifiable references, and 36% noted the publication date (Figure). Overall, 4 websites (9%) were identified as excellent, and nine (20%) were of poor quality. In regard to pancreatic cancer surgery, 62% of websites discussed postoperative complications, 56% addressed quality-of-life (QOL) issues, and 53% acknowledged the surgery volume-outcome relationship. IRR was 0.75 for the 2 professional assessors on the overall rating. DISCERN assessment by patient/family evaluators demonstrated 83% agreement with the results by medical professionals. Major areas of disagreement included QOL and website bias.

Conclusion:

The quality of patient-centered online information on pancreatic cancer treatment is highly variable. Websites frequently lack updated information and references, and often do not provide adequate information for patients to make well-informed treatment decisions. However, patients and family members demonstrated the ability to learn strategies to critically evaluate online health information.

31.06 Neutrophil to Lymphocyte Ratio is a Preoperative Risk Factor in Intrahepatic Cholangiocarcinoma

Posted on January 18, 2018

S. Buettner2, B. Groot Koerkamp2, M. Weiss3, S. Alexandrescu4, H. P. Marques5, J. Lamelas5, L. Aldrighetti6, T. Gamblin7, S. K. Maithel8, C. Pulitano9, T. W. Bauer10, F. Shen11, G. A. Poultsides12, J. Marsh13, J. N. IJzermans2, T. M. Pawlik1  1Ohio State University,Columbus, OH, USA 2Erasmus MC University Medical Center,Dept. Of Surgery,Rotterdam, ZUID-HOLLAND, Netherlands 3Johns Hopkins University School Of Medicine,Baltimore, MD, USA 4Fundeni Clinical Institute,Bucharest, ROMANIA, Romania 5Curry Cabral Hospital,Lisbon, PORTUGAL, Portugal 6Ospedale San Raffaele,Milan, ITALY, Italy 7Medical College Of Wisconsin,Milwaukee, WI, USA 8Emory University School Of Medicine,Atlanta, GA, USA 9University Of Sydney,Sydney, NSW, Australia 10University Of Virginia,Charlottesville, VA, USA 11Eastern Hepatobiliary Surgery Hospital,Shanghai, CHINA, China 12Stanford University,Palo Alto, CA, USA 13University Of Pittsburg,Pittsburgh, PA, USA

Introduction: Alterations in the Neutrophil-to-Lymphocyte ratio (NLR) may be indicative of host immune response to cancer. We sought to determine whether preoperative NLR was associated with long-term outcomes among patients undergoing surgery for intrahepatic cholangiocarcinoma (ICC).

Methods: Patients who underwent resection for ICC between 1990-2015 were identified from 12 major HPB centers. NLR was calculated by dividing the absolute number of neutrophils by the absolute number of lymphocytes. The Kaplan-Meier method and Cox regression models were used to evaluate factors, including NLR, that potentially were associated with overall survival. Factors associated with survival based on the multivariable Cox model were utilized to create a point-based preoperative score to predict prognosis. 

Results: Among 1,000 patients who underwent resection for ICC, the majority of patients had a solitary tumor (n=810, 82.6%) and median tumor size was 6.0 cm (interquartile range [IQR]: 4.3-9.0); 168 (19.5%) patients had lymph node metastases on preoperative imaging. Preoperative jaundice was present in 98 (9.8%) patients, while 27 (2.7%) patients had extrahepatic disease. Median NLR was 2.68 (2.05-4.00).  Overall survival at 1-, 3-, and 5- years was 78.7%, 61.5%, and 39.1%, respectively.  NLR was associated with prognosis as patients with a NLR ≥3 had a median survival of 31.3 months (95%CI: 23.9-38.8) compared with 53.1 months (95%CI: 37.1-69.2) among patients with a lower NLR (p<0.001).  In addition, 1-, 3-, and 5-year survival was 73.5%, 46.4%, and 38.8% versus 84.6%, 61.1%, and 47.8% among patients with NLR ≥3 versus NLR < 3, respectively (p<0.001).  Using backwards selection based on the Akaike Information Criterion, a preoperative risk score was derived. Specifically, on multivariable analysis, number of lesions (Hazard Ratio [HR]: 1.17; p<0.001), tumor size (HR: 1.04; p<0.001), preoperative lymph node metastases (HR: 1.32; p=0.022), preoperative jaundice (HR: 1.65; p<0.001) and extrahepatic disease (HR: 2.39, p<0.001) were each predictive of survival. After controlling for these competing risk factors, NLR remained independently associated with long-term survival (HR: 1.04, 95%CI: 1.02-1.07; p=0.001).  The combination of these preoperative factors into a prognostic model had fair discrimination to predict post-operative survival (area under the curve, 0.65).

Conclusion: Elevated NLR was an independent predictor of worse long-term outcomes among patients with ICC undergoing resection. NLR may be an important factor associated with survival and can help estimate postoperative survival when used in a preoperative predictive model.

 

31.07 Obstruction as a risk factor in the staging of colon cancer: a secondary analysis of the N0147 trial

Posted on January 18, 2018

F. S. Dahdaleh1, S. Sherman1, A. Benjamin1, E. Poli1, K. K. Turaga1  1The University Of Chicago Medicine,Section Of General Surgery/Surgical Oncology,Chicago, IL, USA

Introduction: Presentation with obstruction is frequent in patients with colon (non-rectal) cancer (CC). Large series have reported obstruction among “high-risk” features, yet data from prospective cohorts on its specific prognostic influence are lacking. We hypothesized that obstruction is an independent risk factor for poor prognosis in patients with stage III CC.

Methods: N0147 was a randomized trial conducted between 2004-09 including patients with Stage III CC randomized to adjuvant FOLFOX/FOLFIRI regimens with or without cetuximab. We obtained patient-level data for those in the control chemotherapy-only arms. Patient, tumor and treatment characteristics were abstracted. Disease-free and overall survival (DFS and OS) were estimated by the Kaplan-Meier method. Proportions were compared by Chi-square and Fisher-exact tests. Uni- and Multivariate survival analyses were performed using Cox-proportional hazards models.

Results: Of 1,543 patients with stage-III CC, 250 (16.2%) presented with obstruction. Obstructive tumors were more likely to be K-ras mutant (35% vs. 30%, p=0.07) and poorly differentiated (28.8% vs. 24.6% vs. p=0.17) but these did not reach statistical significance. Obstructed patients were no less likely to complete 12 cycles of adjuvant chemotherapy (75.6% vs. 77.0% p=0.62). With median follow-up time of 30.1 months among survivors, 5-year overall and disease-free survival was significantly worse among obstructed patients (OS 67.7% vs. 78.0%, p<0.001; DFS 53.9% vs. 67.0%, p<0.0001, Figure). After adjusting for conventional AJCC staging variables including T and N-stage, high-grade histology, and host characteristics, obstruction remained significantly associated with worse survival (OS HR 1.61, 95% CI 1.16-2.24, p=0.005; DFS 1.54, 95% CI 1.21-1.98, p<0.001). This difference was more pronounced in the cohort receiving FOLFOX (HR 2.13, 95% CI 1.46-3.11, p<0.001).

Conclusion: In this prospectively-followed cohort of Stage-III CC patients treated with standard-of-care adjuvant chemotherapy, obstruction was significantly associated with worse survival. Moreover, this effect was independent of T- and N-stage, and histology. Conclusions from this secondary analysis of a randomized trial are less likely to be biased by diagnostic and therapeutic factors, and by disparities in access to care than retrospective series. Inclusion of obstruction in the AJCC staging system might help better stratify these patients at high risk of recurrence and death.

31.04 Clinical Fate of T0N1 Esophageal Cancer: Results from the National Cancer Database

Posted on January 18, 2018

C. Takahashi1, R. Shridhar2, A. Patel3, J. Huston4, K. Meredith3  1Midwestern University,Glendale, AZ, USA 2University Of Central Florida,Orlando, FL, USA 3Florida State University,Tallahassee, FL, USA 4Sarasota Memorial Hospital,Sarasota, FL, USA

Introduction:  The long-term survival for patients with locally advanced esophageal cancer (EC) remains poor despite improvements in multi-modality care. Neoadjuvant chemoradiation (NCR) followed by surgical resection remains pivotal in the management of patients with ECC.  However the outcomes of patients whose primary tumor exhibits a complete response with residual regional nodal disease (T0N1) remains unclear as well as the role for adjuvant therapy.

Methods:  Utilizing the National Cancer Database we identified patients with esophageal cancer who underwent NCR followed by esophagectomy who had subsequent pathology of T0N1.  Baseline univariate comparisons of patient characteristics were made for continuous variables using both the Mann-Whitney U and Kruskal Wallis tests as appropriate. Pearson’s Chi-square test was used to compare categorical variables. Unadjusted survival analyses were performed using the Kaplan-Meier method comparing survival curves with the log-rank test. All statistical tests were two-sided and α (type I) error <0.05 was considered statistically significant.

Results: We identified 7,116 patients diagnosed with EC (6,235 (87.6%) adenocarcinoma (AC), 881 (12.4%) squamous cell carcinoma (SCC) with a median age of 62 (21 – 88) years.  There were 6,031 (84.8%) males and 1,085 (15.2%) females. R0 resections were achieved in 6,668 (93.7%) patients and this correlated to improved survival, median survival 55.4 (RO) and 24.4 (R1) months respectively, p<0.001. The median nodes harvested were 13 (0-83) with a median positive LN’s of 1.4 (2.9%). Complete response (pCR) was achieved in 1,334 (18.7%), partial response (pPR) 2,812 (39.5 %) and non-response (pNR) 2,970 (41.7%).  There were 230 (3.2%) patients deemed as pathologic T0N1.  The median survival of patients with pCR was 61.7 months compared to 32.1 months in the T0N1 patients p<0.001. T0N1 patients did not demonstrate an improved survival over T1/2 patients who had a median survival of 30.5 months, p=0.77. However, T0N1 did reveal an improved survival over T3/4 patients who had a median survival of 24.6 months, p=0.02. Adjuvant chemotherapy in T0N1 did not provide a benefit in survival, median survival adjuvant versus no adjuvant 30.8 vs 32.1 months respectively, p=0.08. Multivariate analysis in T0N1 patients demonstrated only number of LN’s positive, histology SCC vs ACC, and margin as predictive of survival, HR 1.23 (1.10-1.36) p<0.001, HR 0.38 (0.22-0.67), p=0.001, HR 1.97 (1.7-2.27) p<0.001,respectively.

Conclusion: Patients with esophageal cancer who exhibit a pathologic T0N1 after NCR have oncologic fates similar to node positive patients.  Patients with pCR of the primary tumor and regional lymph nodes continue to demonstrate significant survival benefits over all remaining pathologic cohorts.

31.05 Total Neoadjuvant Therapy in Pancreatic Cancer

Posted on January 18, 2018

N. Goel1, A. Nadler1, W. H. Ward1, K. J. Ruth1, A. Karachristos1, J. P. Hoffman1, S. S. Reddy1  1Fox Chase Cancer Center,Surgical Oncology,Philadelphia, PA, USA

Introduction:

There is increasing interest in total neoadjuvant therapy (TNT) for pancreatic cancer. This entails systemic chemotherapy followed by chemoradiotherapy, and then definitive surgery as long as the disease remains localized. The perceived benefits are that all patients receive multimodality therapy, eradication of occult systemic disease, and the potential to downstage borderline resectable (BR) tumors for curative resection. This study evaluates whether TNT has an overall survival (OS) benefit compared to neoadjuvant chemoradiation therapy and upfront surgery. 

Methods:

This is a retrospective study of 182 pancreatic adenocarcinoma patients treated at our institution from 2000-2015 who underwent a pancreaticoduodenectomy (PD) or a total pancreatectomy. Patients undergoing distal pancreatectomy, those with macroscopic disease discovered at the time of surgery, and those with stage IV disease were excluded. 

Results:

The analytic cohort consisted of 66 patients in the TNT group, 29 in the neoadjuvant chemoradiation group, and 87 in the surgery first group. Median age at diagnosis was 67 in the TNT cohort, 72 in the neoadjuvant cohort, and 70 in the surgery first cohort. 46(69.7%) of the TNT patients were initially borderline resectable (BR) or unresectable and were clinically stage 2A or higher. 67(77%) of the upfront surgery patients were clearly resectable and were clinically stage 2A or lower. Ten(15.2%) of the TNT patients had a complete pathologic response. 55(83.3%) of the TNT patients had an R0 resection. On univariate analysis, treatment(p=0.0016), age(p=0.037), nodal status(p=0.01), and margin status(p=0.001) were statistically significant. On multivariate analysis, treatment(p=0.024), age(p=0.0014), and margin status(p<0.0001) remained significant. Surgery first had a statistically significant greater hazard ratio compared to the TNT group at 1.81. Median OS from date of diagnosis was 18 months in the upfront surgery group, 25 months in the neoadjuvant group, and 36.7 months in the TNT group (p=0.0019). The TNT group also had the greatest CA19-9 drop with 77% having at least a 50% response.

Conclusion:

Although limited data exist on the utility of TNT, our review, which is one of the largest to date, shows a statistically significant improvement in OS for the TNT group, despite more advanced disease. We achieved a median OS of 36.7 months and 5 year OS 31.5% with TNT in a patient population that was more advanced, including both BR and unresectable patients. TNT therefore offers favorable short- and long-term outcomes, as well as the benefit of optimally selecting patients for surgery based on fitness for TNT and tumor biology.

 

31.02 A Comparison of Colon versus Rectal Adenocarcinoma using Molecular Profiling

Posted on January 18, 2018

J. Purchla1, W. H. Ward1, F. Lambreton1, N. Nweze1, T. Li2, N. Goel1, S. Reddy1, E. Sigurdson1, J. M. Farma1  1Fox Chase Cancer Center,Department Of Surgical Oncology,Philadelphia, PA, USA 2Fox Chase Cancer Center,Philadelphia, PA, USA

Introduction: The role of molecular profiling (MP) in the evaluation and management of colorectal adenocarcinoma continues to grow. In conjunction with standard genotyping for disease-related mutations, MP provides a genetic blueprint of the analyzed colorectal tumors. Although similar histologically, the clinical behavior and treatment of colon and rectal cancers can be quite dissimilar. The purpose of this investigation is to compare the colon and rectal MPs.

Methods: A retrospective study was performed using MP data of colorectal patients of any stage who were treated at our tertiary cancer center between 2006 and 2017, and underwent a Targeted Cancer Panel and/or specific single gene tests. Those who did not undergo molecular profiling were excluded. Demographic, clinical, and pathological data were collected and analyzed. The Wilcoxon test and Chi-square test were used for statistical analysis.

Results: 355 colon cancers underwent MP, and 66.5% had a mutation affecting 42 different genes. 126 rectal cancers underwent MP and 79.4% has a mutation affecting 45 different genes (P=0.007). In the colon group, 53.2% were male, and in the rectal group, 63.5% were male (P=0.04). Additionally, in the colon group, 80.6% were white, and in the rectal group, 89.7% were white (P=0.02). Stage IV colon cancer patients comprised 45.0% of the cohort, and stage IV rectal cancer patients comprised 48.8% (P=0.6). KRAS mutation was seen in 43.4% of colon and 46.1% of rectal (P=0.65), BRAF mutation in 11.6% of colon and 5.0% of rectal (P=0.06), P53 mutation in 57.4% of colon and 62.6% of rectal (P=0.4), APC mutation in 47.4% of colon and 58.2% of rectal (P=0.08), SMAD4 mutation in 10.0% of colon and 13.7% of rectal (P=0.3), PIK3CA mutation in 17.0% of colon and 7.3% of rectal (P=0.02), and defective mismatch repair/microsatellite instability (dMMR/MSI) in 12.6% of colon and 1.6% of rectal (P=0.0004). No mutations were seen in 33.5% of colon and 20.6% of rectal tumors, 1 mutation was seen in 20.9% of colon and 23.8% in rectal, 2 mutations were seen in 20.9% of colon and 13.5% of rectal, and 24.8% of colon and 42.1% of rectal had more than 3 mutations (P=0.0005). In colon cancer patients with mutations: 41.5% were located in the right colon, 10.9% in the transverse colon, 10.0% in the left colon, 37.6% in the sigmoid/recto-sigmoid colon, and 7 patients were unreported.

Conclusions: In colon and rectal patients who underwent MP had mutations affecting more than 40 unique genes. Colon tumors had higher rates of BRAF and PIK3CA mutation, and dMMR/MSI in comparison to rectal cancers. Based on this descriptive analysis, further investigation with a larger sample size is needed to affirm these patterns and may affect future treatment decision making. 

« Previous Page
Next Page »