Department of Surgery

Research Day Poster Session

Tuesday, May 23 7:00-9:00 am

Welcome, Sayeed Ikramuddin, Interim Department Chair 7:00am, PWB 11-157

Award Presentation, Greg Beilman, Deputy Chair 8:40am, PWB 11-157

Poster Locations: #1-32: PWB Room 11-157 and PWB 11th Floor

#33-49: Moos 11th Floor Lobby

Poster # Presenting Author Mentor Title

1 Andrea Wolf Beilman Safety and Efficacy of Novel Formulations of a Resuscitation Fluid for Hemorrhagic Shock

2 Beth Lusczek Beilman Circadian Rhythms are Blunted in Critically Ill Patients

3 Megan Schmidt Iaizzo Time variate comparison of in situ and in vitro monophasic action potential recordings

4 Arpan Patel Iaizzo Physiological Response of Skeletal Muscle to Irreversible Electroporation

5 Dahlia Maxon Davydova Application of Oncolytic Adenovirus as a Novel Therapy for Melanoma

6 Mizuho Sato-Dahlman

Yamamoto A fiber and hexon-modified infectivity-selective oncolytic adenovirus escapes liver sequestration after systemic administration in pancreatic cancer model

7 Malavika Chandrashekar

Harmon Mini-laparotomy Mesenteric Vein Infusion vs. Percutaneous Transhepatic Portal Vein Infusion: Safety and Complication Rates during Clinical Islet Allograft Transplantation

8 Malavika Chandrashekar

Harmon

Safety and Efficacy of 4-Factor Prothrombin Complex Concentrate in Patients Receiving Vitamin K Antagonists Compared to Patients not Receiving Vitamin K Antagonists: A Single Center Retrospective Review.

9 Lucas Mutch Graham Training nonhuman primates for cooperation with medical device management to enhance human factor prediction and usability evaluations

10 Jody Janacek Graham Development of a novel ultrafiltrate perfusion device bioengineered for high density islet cell transplantation without immunosuppression

11 Alex Mattson Iaizzo 3-Dimensional Reconstruction and Analysis of Epicardial Adipose the Human Heart

12 Lars Mattison Iaizzo Physiological Assessment of Cardiac Muscle Post-Irreversible Electroporation Therapy

13 Carolina Fernandez Branson

Chipman How to Improve Resident Communication when Disclosing Bad News

14 Nora Christensen Chipman The Role of Ultrasound Guided Inguinal Lymph Node Biopsy in Studying HIV Viral Reservoirs

Poster # Presenting Author Mentor Title

15 Xianda Zhao Subramanian Tumor location impacts immune response in mouse models of colon cancer

16 Ce Yuan Subramanian Integrative analysis of colorectal cancer tissue microRNA and microbiome reveals microRNAs as potential mediator of host-microbiome interaction

17 Heena Shah Harmon The Undergraduate Surgery Interest Group (USIG): A New Venue for Premedical Student Exposure to the Field of Surgery

18 Andrew Sundin Harmon Introducing Bedside Ultrasound in Human Anatomy Laboratory: A Quantitative Feasibility Assessment

19 Elizabeth Lindemann

Melton-Meaux

Representation of Occupation Information in Clinical Texts: An Analysis of Free-Text Clinical Documentation in Multiple Sources

20 Steven Skube Melton-Meaux

Accelerating American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) Surgical Site Infection Abstraction with a Semi-Automated Approach

21 Mariya Skube Beilman Characteristics of 171 Combat-Associated Small Bowel Injuries

22 Mariya Skube Beilman Islet Graft Function is Preserved after Pregnancy in Patients with Previous Total Pancreatectomy with Islet Autotransplant

23 Ashish Singal Faizer Aortic Center Directed Risk Factor Verification Optimizes An EMR Based AAA Screening Program

24 Ashish Singal Faizer Pre-operative Ankle-Brachial Index Stratification for Patients with Peripheral Arterial Disease

25 John Spratt Loor Lung Transplantation from Donation after Circulatory Death Donors after Prolonged Normothermic Ex Vivo Lung Perfusion

26 Sam Miotke Mohr The State of Firework Laws and Injury Tracking

27 Victor Vakayil Harmon Urinary Tract Infection Sensitivities in an Intensive Care Setting: A Potential for Novel Antibiotic

29 Mikayle Holm Iaizzo Reconstruction of Full Body Skeleton from Computed Tomography Images of a Fresh Human Cadaver

30 Jorge Zhingre Sanchez

Iaizzo 3D modeling and visualization of human cadaver organs for clinical and educational applications

31 Kristin Colling Beilman Mortality from Necrotizing Pancreatitis in the Era of Endoscopic Drainage; An Analysis of 337 Cases

Poster # Presenting Author Mentor Title

32 Elliot Arsoniadis Kwaan Increased Variability in Treatment Pathways for African Americans with Squamous Cell Carcinoma of the Anal Canal

33 Lihua Li Subramanian Colon cancer derived extracellular vesicles regulate host immune response via microRNA mediated immunosuppression

34 Lisa Koodie Davydova Deletion of the Adenovirus Death Protein from Adenoviral E3 region Improves Imaging Potential of Oncolytic Adenovirus Therapy

35 Lisa Koodie Davydova Human Oncolytic Adenovirus Bio-distribution and Toxicity differ between Pigs and Mice

36 Josh Wilhelm Hering Glucose-responsive oxygen consumption rate in islets from chronic pancreatitis patients is size dependent: novel islet quality assessment through bioenergetic phenotyping

37 Jacob Ricks Hering/ Wilhelm

Histomorphometry of insulin-stained pancreas sections and correlation with islet isolation outcomes for patients undergoing total pancreatectomy and islet autotransplantation

38 Amanda Lord Harmon Therapeutic strategies for severe and severe-complicated Clostridium difficile infection

39 Cyrus Jahansouz Ikramuddin Disruption of Intestinal Microbiota Contributes to Failure of Vertical Sleeve Gastrectomy

40 Erik Gaasedelen Iaizzo Using Smartphone-Based Virtual Reality to Explore Internal Anatomy of 3D Heart Models

41 Charles W. LeNeave

Iaizzo Pharmacologic Post-Conditioning of DCD Swine Lungs on EVLP with Whole Blood: N-Acetyl Cysteine to Minimize Ischemic-Reperfusion Injury

42 Brent Bauman Harmon Multiple Simulation Modalities Enhanced a Preparation for Surgical Internship Course

43 Jing Li Huang Yamamoto CD133 infectivity-selective oncolytic adenovirus effectively targets radiation-resistant colorectal cancer cells

44 Victor Vakayil Harmon Indicators of Mortality for Patients on Veno-Arterial Extracorporeal Membrane Oxygenation (ECMO): A Single center experience

45 Victor Vakayil Harmon Perforated Peptic Ulcer Disease - Safety and Postoperative Complications during Laparoscopic vs. Open Repair: an ACS-NSQIP Analysis

46 Amanda Salzwedel Yamamoto Use of Oncolytic Adenovirus Expressing IFN- to develop improved IFN therapy against pancreatic cancer

Poster # Presenting Author Mentor Title

47 Shannen Kizilski Faizer The effect of rigid stent-grafts on the propagation of pressures in aortic dissection: A lumped parameter mathematical model of flow through the descending thoracic aorta

48 Laura Hocum Stone

Kelly Stem cell cardiac patch as an adjunctive therapy during revascularization of Chronically Ischemic Myocardium

49 David Heller Hering/ Wilhelm

Quantitative Analysis of Microbial Contamination in Total Pancreatectomy and Islet Auto Transplantation for Development of Bioburden Reduction Strategies

Thank you to our judges! Anne Bantle, MD Medicine

David Bernlohr, PhD CBS Scott Dehm, PhD LMP

Peter Gordon, MD Pediatrics Pamala Jacobson, PharmD - Pharmacy

Mahmoud Khalifa, MD, PhD LMP Emil Lou, MD - Medicine

Javed Mohammed, DVM, PhD Dermatology Ann Parr, MD, PhD Neurosurgery Jill Siegfried, PhD Pharmacology

Tim Starr, PhD OB/GYN Boris Winterhoff, MD OB/GYN

#1 Safety and Efficacy of Novel Formulations of a Resuscitation Fluid for Hemorrhagic Shock Wolf A; Thakral S; Mulier KE; Suryanarayanan R; Beilman GJ Introduction: There is significant need for treatments preventing mortality after blood loss and injury. A combination of melatonin (M) and D-beta-hydroxybutyrate (BHB) increases survival in preclinical models of hemorrhagic shock and trauma. The standard formulation (4M BHB/43mM M/20% DMSO) contains the solvent dimethylsulfoxide (DMSO), which is suboptimal for clinical use. We have developed two novel BHB/M solutions in which DMSO is omitted. The aim of this study was to test the safety and efficacy of these prototypes in a rat hemorrhagic shock model. Methods: Rats were exposed to gradual withdrawal of 40% total blood volume, followed by a four-hour shock interval. Treatment solutions were administered intravenously halfway throughout blood withdrawal. Treatment groups were 1) no hemorrhage, 2) control (lactated Ringers solution, LR), 3) 1x BHB/M/DMSO, 4) 0.5x BHB/M/DMSO, 5) BHB/M/Solvent 1 or 6) BHB/M/Solvent 2. All BHB/M-treated animals received the same dose of BHB and M. Survival was analyzed, as well as blood pressure, heart rate, free plasma hemoglobin, blood gases and markers of organ injury. Results: There was a significant difference in survival, which was highest for BHB/M/Solvent 1 (8/10), followed by BHB/M/Solvent 2 (6/10), 0.5x and 1x BHB/M/DMSO (5/10), LR (3/10) and no treatment (0/11). Upon completion of blood withdrawal, free plasma hemoglobin, a measure of in vivo hemolysis (an adverse effect of DMSO), was significantly lower in the prototype groups than in 1x BHB/M/DMSO rats, and lower in the BHB/M/Solvent 2 group than in 0.5x BHB/M/DMSO animals. Blood pressure, heart rate, blood gases and markers of organ injury did not differ between BHB/M-treated groups throughout the experiment. Translation: BHB/M significantly improves survival in preclinical hemorrhagic shock and injury models, but DMSO

constitutes a suboptimal diluent An optimized BHB/M formulation may improve safety and efficacy during clinical application Conclusion: In this preclinical small animal study, BHB/M solutions without DMSO exhibited improved efficacy and safety when compared to the standard solution.

#2 CIRCADIAN RHYTHMS ARE BLUNTED IN CRITICALLY ILL PATIENTS Elizabeth Lusczek, Germaine Cornelissen, Greg Beilman Introduction: Synchronized circadian rhythms plays a key role in coordinating physiologic health while desynchronized circadian rhythms may predispose individuals to disease. The ICU environment may desynchronize the circadian rhythm with unregulated light/dark cycles, continuous parenteral feeding, high noise levels, high levels of sedation, and fragmented sleep. Its effects on health outcomes of critically ill patients are expected to be numerous, and directly supporting circadian rhythms may improve patient outcomes. This pilot study was undertaken to determine if circadian rhythms are altered in ICU patients relative to healthy controls. Methods: Circadian rhythms were monitored in 5 healthy controls and 5 ICU patients for 24 hours. Healthy controls, age 45-72, stayed at the University of Minnesotas Masonic Clinical Research Institute and were monitored by trained staff. Patients, age 43-66, were recruited from the University of Minnesota Medical Center Intensive Care Unit. Beginning at 9:00 AM, heart rate and blood pressure were collected every 30 minutes; temperature was taken every hour in all patients. Blood cortisol levels were measured every 4 hours. Bedside light levels were measured every 20 seconds and sound levels were measured every second.

Results: Healthy controls had a significant (p

Time variate comparison of in situ and in vitro monophasic action potential recordings

Megan M. Schmidt1, 2

Paul A. Iaizzo 2

Department of Biomedical Engineering, University of Minnesota1 Department of Surgery, University of Minnesota 2

1 Background

Monophasic action potentials (MAPs) have long been used as a means to study the focal electrical activity of the myocardium. [1, 2] Upon the application of adequate contact force, the signals provide important insights into focal depolarization and repolarization, activation timing, and focal arrhythmic behaviors. [3-6]

Within our laboratory we have developed an isolated physiologic, four-chamber working, large mammalian heart model (the Visible Heart methodology) to study cardiac devices and their interactions with the myocardium. [7] Through the use of a modified Krebs-Henseleit buffer, we can uniquely visualize the device-tissue interface: in this study, the placement of catheters.

The purpose of this study was two-fold. First, we demonstrated the long term stability of MAP recordings in an in situ swine model. Second, we showed the relationship between MAPs recorded from in vitro and in situ preparations of each specimen.

2 Methods

Swine (n=12) weighing between 70 and 80 kg were used in these studies. Animals were initially anesthetized with 500 mg of Telazol and 500 mg of Methohexital. The animals were then intubated and maintained in a plane of deep anesthesia (MAC > 1.2). A median sternotomy was performed and a pericardial cradle was created to enable access to the epicardial surface of the heart. [8] Only specimens eliciting normal cardiac function (e.g., native sinus rhythm) were used.

For the recording of MAPs, four 7Fr MAP4 catheters, (Medtronic, PLC, Dublin, IE) were used simultaneously. First one catheter was placed endocardially in the right ventricle (RV) through an introducer in the jugular vein. A second catheter was then placed endocardially in the right atrium (RA). Next two catheters were placed on the epicardial surface of the RV and the left atrium (LA). Catheter locations were verified by direct visualization for epicardial catheters and fluoroscopy (Ziehm Vision R) for endocardial locations, as can be seen in Figure 1.

MAP signals were recorded using a CardioLab Recording System (GE Healthcare, Waukesha, Wisconsin). Baseline signals were collected for 5 minutes with no disturbances of the catheters. This served as the baseline for each specimen to

Figure 1. Location of MAPs catheter during in situ recordings. The image on the left shows the epicardial surface of the heart with 2 MAP4 catheters recording signals from the LA and the RV. The image on the right shows a fluoroscopic view of the MAP4 catheters (arrows) recording from the RA and RV endocardium.

which all signals would be compared. Signals were then continuously collected and monitored for two hours, with only slight adjustments in catheter position to maintain constant tip pressure.

Upon completion of in situ recordings, the heart was excised using previously described Visible Heart methodologies. The heart was then reanimated and MAPs were collected from the same 4 locations as seen in Figure 2.

All results were analyzed in a custom MATLAB script. Significance was determined through the use of ANOVA and Tukey HSD with a p-value

time points. Figure 3 shows the normalized APD90s for each of the four catheters in situ over the 2 hour recording period.

Figure 3. APD90 from in situ recordings normalized to their baseline recording for catheters on the RV epicardium and endocardium, RA endocardium and LA epicardium. There were no significant differences in normalized APD90 across any in situ time points.

In situ in vitro Comparisons The in situ baseline recording from each catheter was

compared to in vitro signals recorded at 0, 30, 60, 90 and 120 minute time points. The 0 minute time point represents the minute immediately following re-animation, or the restoration of full cardiac function. APD90 for all signals was normalized to the baseline in situ recording. Only endocardial RA and epicardial LA catheters showed significant differences in APD90; and only for the 0 minute time point. Both the RV endocardial and epicardial catheters showed no significant differences when compared to their relative baselines. The normalized APD90s for each catheter can be seen in Figure 4.

Figure 4. APD90 for in vitro recordings normalized to an in situ baseline recording for catheters on the RV epicardium and endocardium, RA endocardium and LA epicardium. The stars signify a significant difference from the in situ baseline for that location and time point.

4 Interpretation The results mentioned in this paper have provided

valuable insight into the recording of MAPs, both in situ and in vitro.

Over a period of 2 hours MAPs were successfully recorded in situ with only minor adjustments in catheter position/pressure. There were no significant differences in APD90 for each location when compared to their normalized baseline values. Further analysis of the relative amplitudes, depolarization rates and repolarization at other time points are being studied.

Similarly, when looking at the in vitro recordings there were no significant differences between the endocardial and epicardial RV catheters when compared to their normalized in situ baselines. Future studies will examine these results at time points greater than 120 minutes, to determine if there is a point at which the duration of the signals differ from in situ.

The atrial MAP recordings showed a significant difference from their in situ baselines immediately after re- animation. It is hypothesized this response is due to ischemic injury that occurs during the isolation and re-perfusion processes. Interestingly, by 30 minutes into the in vitro prep there were no significant differences in APD90 for either the RA or LA MAP recordings. Further analyses will be conducted to determine how long it takes for these locations to reach non-significant values.

From these results, we have gained two important insights into the recording of MAPs. First, the long-term monitoring of MAPs is possible with only slight adjustments in the catheter tip pressure and/or location. Second, the Visible Heart methodologies are not only a viable tool for the imaging of the device tissue interface, but also for the recordings of monophasic action potentials and other electrical signals. With this new information, the development of catheters designed specifically to record MAPs can be continued using the Visible Heart without compromised electrical activity, or diminished MAP signals.

References

1. Olsson SB. (1971) Monophasic action potentials from right atrial muscle recorded during heart catheterization. Acta Medica Scandinavica, 190: 369- 379.

2. Levine JH, Moore EN, Kadish AH, Guarnieri T, & Spear JF. (1986). The monophasic action potential upstroke: a means of characterizing local conduction. Circulation, 74(5): 1147-1155.

3. Franz MR. (1999). Current status of monophasic action potential recording: theories, measurements and interpretations. Cardiovascular Research, 41(1): 25-40.

4. Yang S, & Kittnar O. (2010) New insights into application of cardiac monophasic action potential. Physiol Res, 59(5):645-660.

5. Narayan SM, Wright M, Derval N, Jadidi A, Forclaz A, Nault I, Miyazaki S, Sacher F, Bordachar P, Clmenty J, Jas P, Hassaguerre M, & Hocini M. (2011) Classifying fractionated electrograms in human atrial fibrillation using monophasic action potentials and activation mapping: evidence for localized drivers, rate acceleration, and nonlocal signal etiologies. Heart Rhythm, 8(2):244-253.

6. Hassaguerre M, Sanders P, Hocini M, Takahashi Y, Rotter M, Sacher F, Rostock T, Hsu L, Bordachar P, Reuter S, Roubaut R, Clmenty J, & Jais P. (2005) Catheter ablation of longlasting persistent atrial fibrillation: critical structures for termination, Journal of cardiovascular electrophysiology, 16(11):1125-1137.

7. Chinchoy E, Soule CL, Houlton AJ, Gallagher WJ, Hjelle MA, Laske TG, Morissett J, & Iaizzo PA. (2000). Isolated four-chamber working swine heart model. Annals of Thoracic Surgery, 70: 1607-1614.

8. Iles, T.L., Howard, B., Howard, S., Quallich, S., Rolfes, C., Richardson, E., Iaizzo, H.R., Iaizzo, P.A. Testing the Efficacy of Pharmacological Agents in a Pericardial Target Delivery Model in the Swine. J. Vis. Exp. (113), e52600, doi:10.3791/52600 (2016).

00.20.40.60.8

1

0 12 24 36 48 60 72 84 96 108120132144156168180N

orm

alize

d Tw

itch

Forc

eTime Post-Ablation (minutes)

Normalized Twitch Force of Swine Diaphragm after Electroporation

0 Pulses 10 Pulses 20 Pulses 30 Pulses

40 Pulses 50 Pulses 60 Pulses 70 Pulses

Physiological Response of Skeletal Muscle to Irreversible Electroporation Arpan Patel MS, Lars M Mattison, Paul A Iaizzo PhD, FHRS

Background: Ablation technologies are emerging as treatment options for patients suffering from cardiac arrhythmias such as atrial fibrillation. One of the most cited complications from cardiac ablation is collateral damage to surrounding smooth and skeletal tissues causing adverse symptoms such as increased difficulty breathing and dizziness. Irreversible electroporation (IRE) is a technology that continues to be assessed as a treatment method for atrial fibrillation. Previous studies have shown that low doses of IRE result in physiological and biomechanical damage to skeletal muscle in swine models. There remains a gap in understanding additional underlying tissue responses to various IRE energies. This study aims to investigate isolated skeletal muscles physiologic responses to IRE. Methods: Yorkshire-cross swine (male, n=16) were intubated and anesthetized for approved protocols by the institutions animal care and use committee. From each, a muscle biopsy from the diaphragm was obtained. Each biopsy was carefully dissected into approximately uniform bundles (2-4 mm in diameter and 3-5 cm long). Suture loops were tied onto both ends of isolated bundles. These were subsequently suspended between a stationary rod and a force transducer. The setup was submerged in a carbogen gassed chamber containing Krebs-Henseleit buffer maintained at 37 degrees Celsius. Laterally placed platinum electrodes were used to supramaximally stimulate the bundles to contract every 10s. Force data was recorded using a custom LabView program. Optimal peak forces were identified (length-tension relationships). Next, the electrodes were removed from both sides of the muscle bundle and two NanoKnife (Angiodynamics, Latham, NY) needles were placed on both sides of

the bundle with a 1cm probe exposure length. When aligned, an FDA approved treatment of IRE varying from 10 pulses to 70 pulses at 500 V/cm with a pulse rate of 240 PPM and 90 s pulse width were applied to the bundles. A control of no IRE ablation was used in this experiment. The bundles were monitored for three hours after the ablation therapy. Results:

Figure 1. The graph shows normalized tissue twitch force post therapy. A noticeable difference is observed between the control group and treated groups.

Normalized twitch force was analyzed among the treatment and control groups. Average normalized forces were compared between the groups at 60 min, 120 min, and 180 minutes after treatment. T-test analysis identified a significant difference between the average normalized twitch forces between the control group (0 pulses) and each of the treatment groups at 60 min, 120 min, and 180 min post ablation (p

#5 Application of Oncolytic Adenovirus as a Novel Therapy for Melanoma Dahlia Maxon, Lisa Koodie and Julia Davydova Department of Surgery, University of Minnesota, MN, 55441 The use of the adenovirus has produced promising results in the field of oncology and gene therapy. Previous studies have shown that a modified chimeric vector with the Ad5 and Ad3 serotypes has enhanced killing ability of malignant cells. The modified chimeric vector, Ad5/Ad3, contains a fiber from the Ad5 serotype and a knob from the Ad3 serotype. This project investigated the use of the wildtype Ad5 virus and the modified Ad5/Ad3 virus in human melanoma cell lines (A375 and WM115), and a murine melanoma cell line (B16FO). Three experiments were performed in order to compare the effectiveness between the two vectors. A binding assay was completed to assess the ability of each virus to bind, but not internalize, to the melanoma cells. Next, a crystal violet assay was done to analyze the cytotoxicity of the two vectors in each of the cell lines. Finally, a luciferase expression assay was performed to evaluate the efficacy of gene delivery, the luciferase gene, for each of the viruses. Results and data analysis concluded that the modified Ad5/Ad3 vector had enhanced binding and killing ability for its application in the field of melanoma treatment.. In addition, results showed that the mouse model is not an efficient animal model for these studies and work with adenoviral treatment.

#6

A fiber and hexon-modified infectivity-selective oncolytic adenovirus escapes liver sequestration after systemic administration in pancreatic cancer model.

Mizuho Sato-Dahlman1, Yoshiaki Miura1, Jing Li Huang1, Praveensingh Hajeri1, Hideki Yoshida1,

Kari Jacobsen1, Julia Davydova1, and Masato Yamamoto1 1. Department of Surgery, University of Minnesota, Minneapolis, MN

Oncolytic adenovirus has high potential for systemic cancer therapy. However, its efficacy upon

systemic application has been quite limited to date. Unlike loco-regional therapy, systemic application of cancer therapy mandates better tumor distribution and transduction. When adenovirus vectors are injected intravenously into mice, most of the virus goes to the liver and can in high dosage lead to liver toxicity. One of the reason for liver tropism is that hepatocytes express high levels of the primary adenovirus receptor (CAR), and non-parenchymal liver cells, such as Kupffer cell and epithelial cell, also capture the viral particle. As a consequence of large sequestration of adenovirus by liver, the tumor transduction rate is low and the in vivo efficacy of systemic therapy is limited. Therefore, the improvement of cancer selective transduction and vector distribution to avoid liver sequestration would overcome the obstacles for systemic delivery and enable efficient systemic treatment of cancer with oncolytic Ad (OAd).

To improve the tumor transduction, we have generated the pancreatic cancer targeted-OAd in previous studies. AdML-VTIN, an oncolytic adenovirus targets the mesothelin protein, which is overexpressed in pancreatic cancer. AdML-VTIN showed selective and powerful anti-tumor effect against Panc-1 tumors in both intratumoral injection (i.t.) and intravenous (i.v.) injection. Interestingly, when we assessed viral distribution after i.v. injection to the nude mice, the liver sequestration of AdML-VTIN was lower than AdML-5WT (Ad5 fiber) at 48 hrs after injection. By day seven, the viral copy number of AdML-VTIN in the tumor was more than three orders of magnitude higher than WT. When we compared the therapeutic effect of i.v. and i.t. injections, the tumor volume significant decreased in both groups. Although the viral particle in the tumor with i.v. injection was about half of that with i.t. injection at 24 hrs after injection, the anti-tumor effect of systemic injection group was similar to that of local injection group. These results suggest that systemic injection of the tumor targeted-OAd showed significantly lower liver sequestration and better tumor accumulation.

Next, to further reduce liver sequestration and increases antitumor efficacy, we developed an improved vector which is a double-modified vector with both hexon- and fiber-modification. The mesothelin-targeted OAd was additionally modified by substitution of Ad5 hexon-hypervariable region 7 (HVR7) with Ad26 hexon-HVR7. We assessed liver distribution of the double-modified vector after i.v. injection. When the virus copy number was compared after 48 hrs, double-modification resulted in significantly decreased liver sequestration. These data suggest that hexon-modification allowed to escape from the liver. The detailed mechanism of this finding is currently being investigated and focuses on the viral uptake by hepatocytes and Kupffer cells.

We believe that the fiber and hexon double-modified vector might be a promising vector to reduce hepatotoxicity and enhance systemic antitumor effect for systemic treatment of pancreatic cancer.

#7 Mini-laparotomy Mesenteric Vein Infusion vs. Percutaneous Transhepatic Portal Vein Infusion: Safety and Complication Rates during Clinical Islet Allograft Transplantation Victor Vakayil1 MBBS, Malavika Chandrashekar1 B.S., Casey Yang2 MD, Joshua J. Wilhelm3 MS, Melena D. Bellin3 MD, Raja Kandaswamy1 MD, David E. Sutherland3 MD, PhD, David Hunter4 MD, Bernhard J. Hering3 MD, James V. Harmon Jr.1 MD, PhD 1. Department of Surgery, University of Minnesota, Minneapolis, MN, United States; 2. Department of Surgery, Hennepin County Medical Center, Minneapolis, MN, United States; 3. Schulze Diabetes Institute, University of Minnesota, Minneapolis, MN, United States; 4. Interventional Radiology, University of Minnesota, Minneapolis, MN, United States. Aims of the Study Assessment of the safety and complication rates of different surgical techniques helps to reduce the surgical risks involved in clinical islet allotransplantation. Percutaneous transhepatic (PT) portal vein infusion and mini-laparotomy (MLap) mesenteric vein infusion are two techniques used at our institution during islet transplantation. We reviewed the safety, complications, and technical aspects of the two techniques, with particular attention to the PT technique for tract closure that involves a sandwich technique of alternating coils and gelfoam. Methods We performed a retrospective chart review of 49 adult patients who underwent pancreatic islet allotransplantation at our center. We analyzed and compared the demographics, clinical variables, perioperative measures, and serious adverse event (SAE) associated with the PT and MLap groups (Table 1). Results A total of 70 islet allotransplants (19 PT and 51 MLap) were performed in 49 recipients (35 female, 14 male). Demographic, clinical, and perioperative variables such as age, BMI, coagulation profiles, LFTs, RFTs, ASA scores, and estimated blood loss did not differ between the two groups (all p values>0.1). There were no significant differences between the median durations of anesthesia or surgery for both procedures. The monitored anesthesia care (MAC) to general anesthesia (GA) conversion rates for the PT and MLap groups were 5.6% and 15.2%, respectively. In the PT group, the portal vein was successfully cannulated on the first attempt 78.9% of the time, and required >3 attempts in 10.5%. Three SAEs occurred in the PT group: one case of symptomatic bleeding and two cases of cholecystitis. One SAE occurred in the MLap group: a non-incarcerated incisional hernia. A partial left portal vein thrombus was noted in the MLap; this was not considered a SAE. All SAE were successfully managed and resolved. The symptomatic bleeding in the PT group resulted from multiple punctures to the liver due to difficulties in tracking the needle tip. No transhepatic tract bleeding occurred in this group. Conclusion Both PT and MLap approaches are safe methods for clinical allogeneic islet cell transplantation. Although complications for both procedures were minimal, the MLap group had a significantly lower SAE rate (p=0.02) compared to the PT group at our institution. Though PT is considered less invasive, the MLap technique may reduce injury to the liver and permit direct surgical control of bleeding.

Variable PT (n=19) MLap (n=51) p-value

Age 43.64.7 45.210.3 0.522

ASA Score 3[2,3] 3[2,3] 0.831

Duration of Anesthesia(min) 155[125,215] 180[143, 253] 0.101

Duration of Surgery(min) 110[72,167] 114[87,180] 0.342

Estimated Blood Loss(mL) 10[5,20] 10[10,20] 0.164

MAC to GA Conversion Rate

5.6% 15.2% 0.293

Vein Cannulation Attempts 1st: 78.9%

>3: 10.5%

1st: 92.2%

>3: 5.9%

1st: 0.123

>3: 0.123

Post-op Length of Stay(days)

2[2,2] 2[2,3] 0.621

Serious Adverse Events Rate 3/19(5.8%) 1/51(1.9%) 0.02

#8 Safety and Efficacy of 4-Factor Prothrombin Complex Concentrate in Patients Receiving Vitamin

K Antagonists Compared to Patients not Receiving Vitamin K Antagonists: A Single Center Retrospective Review

Victor Vakayil MBBS1, Emily Coler PharmD2, Jared Larson PharmD2, Malavika Chandrashekar BS1,

Amanda Lord BS1, Julie Welbig MLS(ASCP)SBB3, Nicole D. Zantek MD, PhD4, James V. Harmon MD, PhD, FACS1

1. Department of Surgery, University of Minnesota, Minneapolis, MN 2. Department of Pharmacy, University of Minnesota Medical Center, Minneapolis, MN 3. Laboratory Administration, Fairview Health Services, St. Paul, MN 4. Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN Background A 4-factor prothrombin complex concentrate (4F-PCC) has been approved by the FDA for the reversal of vitamin K antagonists (VKA) in patients with major bleeding or need for an emergent surgery/invasive procedure. In patients with other causes of coagulopathy, 4F-PCC may improve hemostasis, however safety and efficacy of 4F-PCC in these patients has not been established. Methods A single center retrospective review of patients who received 4F-PCC from 2014-2016 was performed. Demographic, clinical, laboratory, safety, and efficacy variables were analyzed. Analysis was performed using SPSS (IBM, Armonk, NY). Data is presented as percent of evaluable patients or median [interquartile range] unless otherwise specified. Results 4F-PCC was administered to a total of 91 patients (46.2% female, 53.8% male) with 79.1% of patients receiving VKA therapy prior to administration. The indications for 4F-PCC were bleeding (48.4%), emergent surgery/procedure (17.6%), both bleeding and emergent surgery/procedure (17.6%), and other (16.5%). Liver failure was a frequent comorbidity in patients not receiving VKA [VKA (n=6) 8.3% versus no VKA (n=7) 36.8%, p=0.002] (ref. Table). International normalized ratio (INR) prior to 4F-PCC (pre-INR) was available for 97.8% of patients and was 2.4 [1.95, 3.28], with 78.7% of patients having a pre-INR >2.0. For patients with pre-INR 1.5, the first INR value following 4F-PCC infusion (post-INR) was analyzed. INR values drawn within 1 hour post-infusion (n=25) were 1.40[1.22, 1.58] and INR values drawn within 8 hours post-infusion (n=73) were 1.38[1.23, 1.55]. Among patients with a pre-INR 1.5 and a post-INR obtained within 8 hours, the post-INR correction (INR

Variables Non-Vitamin K Antagonist Reversal [NVKA] (N=19)

Vitamin K Antagonist Reversal [VKA] (N= 72)

P-value

Dem

ogra

phic

s

Age 53.68 +/- 13.52 67.04 +/- 15.53 0.001*

Sex (F) 10 (52.6%) 32 (44.4%) 0.524

BMI 29.63 +/- 5.19 28.80 +/- 6.44 0.293

4F-PCC Dosage (Units/Kg)

34.83 +/- 13.56 28.67 +/- 8.59 0.201

Co-

mor

bidi

ties

Hypertension 9 (21.1%) 50 (69.4%) 0.073

Atrial Fibrillation 5 (26.3%) 47 (65.3%) 0.002*

Coronary Artery Disease 2 (10.5%) 24 (33.3%) 0.060

Heart Failure 6 (31.6%) 44 (61.1%) 0.021*

Liver Failure 3 (15.8%) 29 (40.3%) 0.002*

H/o Deep Vein Thrombosis 5 (26.3%) 15 (20.8%) 0.638

H/o Pulmonary Embolism 5 (26.3%) 15/72 (20.8%) 0.608

Clin

ical

In

dica

tion

for

Use

Bleeding 8 (42.1%) 52 (72.2%) 0.024*

Intracranial Hemorrhage 3 (15.8%) 29 (40.3%) 0.086

Emergent Procedure 7 (36.8%) 25 (34.7%) 0.863

Lab

orat

ory/

Clin

ica

l Var

iabl

es

Pre- 4F-PCC INR (NVKA= 18, VKA= 71)

1.7 [1.15, 2.34] 2.58 [2.17, 3.53] 0.001*

Post 4F-PCC INR (NVKA= 15, VKA= 59)

1.79 [1.37, 2.22] 1.31 [1.2, 1.45] 0.001*

Vitamin K Given 10 (52.6%) 64 (88.9%) 0.001*

Length of Hospital Stay 9 [5, 23] 8 [4, 15] 0.440

Safe

ty &

E

ffic

acy

INR corrected to

#9 Training nonhuman primates for cooperation with medical device management to enhance human factor prediction and usability evaluations Lucas A Mutch1; Jody L Janecek1; Rachael MZ Lee1; Melanie N Niewinski1; Lisa Yang1; Travis Navarro1, Melanie L Graham1,2 Department of Surgery, University of Minnesota1 Veterinary Population Medicine Department, University of Minnesota2 Implantable devices designed to encapsulate and support islet grafts are employing more sophisticated and dynamic designs, with some requiring continuous user management. In certain circumstances it is necessary to test devices in nonhuman primates (NHPs) to properly evaluate inflammatory and immunological responses. The biological complexity of the NHP and similar anatomical, physiological and behavioral characteristics with humans can improve prediction in studies that cannot be performed in vitro or in less sentient animal models. We have previously demonstrated that NHPs be trained to cooperate with clinical care necessary in disease modeling, and aim to expand training to manage medical devices with the perspective of most accurately modeling interactions with the user for clinical relevance. Optimization of device manipulations and user interactions, termed human factors, is a critical factor in maximizing the likelihood that the device will be safe and effective for use. Five adolescent male cynomolgus macaques (CM) and six adolescent male rhesus macaques (RM), all socially housed, were trained for cooperation with diabetic clinical care and management of a totally implantable bioartificial pancreas device (BPD) utilizing positive reinforcement. The training program included physical examination, presentation for blood collection and drug administration, hold for BPD charge events, and shifting. Specific training for BPD management included desensitization to a charging wand, positioning for wand placement, allowing the wand to be placed on the body, and holding for at least a five minute charge. The success of behavior acquisition and ongoing cooperation was evaluated. Since animals were trained to perform multiple behaviors, cooperation with device management was assessed with the perspective of usability. All animals completed the training program successfully, demonstrating competence with each of the behaviors. Testing inductive charging in cooperating animals allowed us to observe animals would prematurely end the planned charge session to withdraw from thermal stimulus generated by energy transfer while still remaining willing to interact with the trainer. Thermal withdrawal was not consistent with expected tolerance device temperatures. No behavioral regression was observed during followup. Devices moving to the clinic must have a strong safety and efficacy profile, but success also hinges on patient adoption, which is largely based on usability and willingness to interact with the device. NHPs can be trained to closely approximate clinical usability and identify factors that may affect device acceptance prior to clinical testing and enhance acceptance. Alongside the scientific benefit, cooperative handling improves animal wellbeing.

#10 Development of a novel ultrafiltrate perfusion device bioengineered for high density islet cell transplantation without immunosuppression Jody L Janecek1; Michael J Dalton4; Lucas A Mutch1; John Brekke5; Timothy D OBrien2,3; Melanie L Graham1,2 Department of Surgery, University of Minnesota, Minneapolis, MN1 Veterinary Population Medicine Department, University of Minnesota, Saint Paul, MN2 Stem Cell Institute, University of Minnesota, Minneapolis, MN3 Norfolk Medical Products, Inc, Skokie, IL4 BRTI Life Sciences, Two Harbors, MN5 Widespread clinical application of islet cell transplant is hampered by the limited availability of islets, immunosuppressive risk, and moderate numbers of patients experiencing graft functional decline over time. We are developing a subcutaneous component-based macroencapsulation device designed to address these barriers, eliminate the need for immunosuppression, advantage a xenogeneic cell source, plus permit non-invasive access for loading, reloading, biopsy and graft recovery. The Cell-SafeTM (C-S) device is designed to generate autologous ultrafiltrate for perfusion to an immunoisolation chamber containing a three-dimensional hydrogel, Islet-MateTM (I-M), high-density islet construct. Proof-of-concept studies were performed to evaluate the suitability of ultrafitrate as a perfusion medium to sustain immunoisolated islet viability and function using Lewis rats. Base characteristics including hematologic, chemistry, electrolyte, and oxygen content of ultrafiltrate generated by the accumulation disc were evaluated at 1, 3, and 6 weeks post-implant in 2 healthy rats. In a second group, porcine islets were loaded at a high density (>15,000IEQ) I-M embedded in the C-S in 3 non-immunosuppressed diabetic rats to evaluate immune protection and graft survival. Needle biopsies of the ultrafiltrate and I-M islet cell construct were performed at 0, 4, 10, and 20 days post-transplant (window 3 days). Interstitial fluid accumulated within the disc and had properties similar to serum with a neutral pH, normal electrolyte composition, but slightly lower protein profile. The average PO2 level was 467mmHg at 1 week, increasing to an average PO2 of 9719mmHg at follow-up timepoints. The tissue surrounding the fluid accumulation disc was highly vascularized with only sparse inflammatory cells observed at 8 weeks postimplant. In transplanted animals, porcine C-peptide levels in ultrafiltrate were >50ng/ml at 4, 10, and 20 days post-transplant. Histologic analysis of the I-M islet cell construct at biopsy showed numerous islets throughout the hydrogel that were both viable and functional with no detectable inflammatory cell infiltrates. The C-S device can be used to derive autologous blood ultrafiltrate for use as a perfusion medium to support I-M embedded islets in the cell immunoisolation chamber. The complete absence of infiltration and strong insulin staining of porcine islets recovered from the C-S device provides evidence of immunoprotective capability. Ultrafiltrate and graft samples were successfully obtained at biopsy timepoints, demonstrating the feasibility of non-invasive access to the graft. The C-S is a fundamentally new device platform engineered to support islet cell replacement in the absence of immunosuppression with potential to dramatically increase the longevity of therapeutic benefit and accessibility to a larger population of diabetic patients.

#11

3-Dimensional Reconstruction and Analysis of Epicardial Adipose the Human Heart

Alexander R. Mattson BS2,3, Traci Jones BS2,3, Susan Sun BA2, Paul A. Iaizzo PhD1

1Department of Surgery, University of Minnesota, Minneapolis, MN USA

2Department of Biomedical Engineering, University of Minnesota, Minneapolis, MN USA

3Medtronic, Mounds View, MN USA

Introduction

Epicardial pacing leads, and epicardial ventricular ablations hinge on reliable contact with ventricular myocardium. However, on the human heart, often a significant adipose layer lies superficial to the epicardial surface of the ventricles: i.e. leaving limited myocardium accessible. The objective of this study was to quantitatively assess the epicardial fat pad superficially laying on various human hearts, so to better inform cardiac device design and feasibility and aid clinicians in epicardial procedures. Here, we present a full MRI-based 3-Dimensional reconstruction of the epicardial adipose tissue on 24 human hearts.

Methods

Human hearts (n=24) deemed non-viable for transplant were obtained from LifeSource (an organ procurement agency; Minneapolis, MN). Hearts were MRI scanned using an mprage T1-weighted protocol. Mimics software (Materialise, Belgium) was used to segment epicardial adipose from myocardial tissue in each scan. A 3-matic (Materialise, Belgium) algorithm quantified a continuous adipose thickness distribution across each human heart. Adipose thickness was measured at 51 anatomical reference points evenly-spaced across the right and left ventricles Results

In this selection of hearts, epicardial adipose was found to cover an average of 80% of right ventricular myocardium and 70% of the left ventricular myocardium. Average adipose thickness ranged from Adipose thickness was typically the greatest along the interventricular sulcus, ventricular base, and right ventricular margin. Average adipose thickness was decreased in two elliptical patches on the diaphragmatic surface of the hearts, the left ventricular margin, and an L-shaped region on the right ventricular outflow tract.

Conclusions

Here, we provide a detailed quantitative mapping of the epicardial adipose that lays superficial to the human heart. This information may be utilized to inform surgical interventions in the pericardial space of the human heart.

#12 Physiological Assessment of Cardiac Muscle Post-Irreversible Electroporation Therapy Lars M. Mattison, Paul A. Iaizzo PhD, FHRS Ablations have become the gold clinical standard of drug resistant atrial fibrillation (AF). AF is projected to affect 50 million people by the year 2050. Today, two primary methods of ablation are used clinically: radio frequency and cryoablation. These ablation technologies are equally effective but still cause complications due to their nature as a thermal ablation. Irreversible Electroporation serves as a possible non-thermal alternative. This therapy is a train of high voltage (>500V/cm) short DC pulses that cause pores to form in the cell membrane. Very little study of this potential treatment relating to the heart has been done. The appeal of a potentially more predictable lesion would be highly desired in this clinical realm. Here we present initial investigations as to the functional response of cardiac tissue to electroporative energy via the NanoKnife.

Yorkshire-cross swine were intubated and anesthetized using approved protocols by the institutions animal care and use committee. Upon completion of another unrelated protocol, the heart was excised and trabeculae from both the left and right ventricle were carefully dissected. Following procurement and dissection of tissue, loops of suture were tied onto either end of isolated bundles: these were then suspended between a stationary support rod and a force transducer. This set up was then submerged in a temperature controlled, carbogen gassed chamber, containing a modified Krebs-Henseleit buffer maintained at 37 C. Laterally placed platinum electrodes were used to supramaximally stimulate the bundles to contract every 10 seconds. The force data was then recorded using a custom LabView program. Optimal muscle function was obtained i.e. length-tension relationships and stimulus amplitude. Following this, the stimulating electrodes were removed from either side of the muscle bundle and replaced with 1 cm long NanoKnife (Angiodynamics, Latham, NY) needles. Once in place, an FDA approved treatment of irreversible electroporation varying from 500V/cm to 1100 V/cm at a pulse rate of 90 (n=34) or 240 (n=70) PPM with a 90 s pulse width and 70 total pulses were applied to the bundles. The viability of the bundles was then monitored for three hours following the applied therapy.

These data show the normalized to baseline function of the bundles as they recover over the 3 hours following the ablation. Baseline force was defined as average twitch force for 15 minutes prior to application of the therapy. The control group received no therapy and serves as a reference to functional decay of the tissue over time. The following graphs show the dose responses of the tissue to 90 PPM and 240 PPM. For 90 PPM, voltages above 1000 V/cm caused a larger decrease in function compared to the other energies. 700V/cm and 800 V/cm also elicited a slight decrease in twitch force. The lower voltages appear to follow the function of the control group. The three different groups could potentially represent irreversible electroporation, reversible electroporation that is still recovering, and reversible electroporation that recovered quickly. While using 240 ppm, voltages above 800V/cm show a larger functional decrease while the lower voltages showing an initial decrease but long term trends that follow the control group. This suggest that the faster pulse rate causes more initial stunning of the tissue when compared to the slower pulse rate, but that the voltages deliver similar results across the different pulse rates.

These data suggest that irreversible electroporation can effectively cause a decrease in function of cardiac muscle. The exact minimum dose needed for this functional decrease in force is between 800V/cm and 900V/cm. Seeing this decrease in function suggests that lesions are created in cardiac tissue allowing this technology to be used for cardiac ablations. As this is a relatively new technology, there is still a need to develop a better understanding of how different tissue types respond to the therapy. Future work should include a larger sample sizes and looking at other factors of irreversible electroporation such as pulse length and number of pulses.

#13 How to Improve Resident Communication when Disclosing Bad News Branson CF, Chipman JG. Department of Surgery, University of Minnesota Background: In this study, we explore surgical residents communication with surrogate patients (SPs), in a portion of the Objective Structured Clinical Examination (OSCE) administered by the Department of Surgery at the University of Minnesota. Methods: We use discourse analysis (DA), a qualitative approach to understanding language, to evaluate our residents interactions with SPs. After identifying problematic communication patterns, we apply communication theory to discuss our findings and provide suggestions for improvement. Results: We found that residents consistently use bluntness and evasiveness to deliver difficult information; they tend to use neutral language when empathetic language is warranted; and many of them try to direct the response of SPs, who then become defensive. Residents use evasiveness most frequently, followed by bluntness. These delivery methods often result in poor communication. Conclusion: We recommend a communication-training program for residents that teaches forecasting and interpersonal skills, so they can convey support and empathy to patients, thereby enhancing care. ACGME core competencies: Patient Care Interpersonal and Communication Skills

#14 The Role of Ultrasound Guided Inguinal Lymph Node Biopsy in Studying HIV Viral Reservoirs

Nora F. Christensen BA, Timothy W. Schacker MD, Torfi Hoskuldsson MD, Greg J Beilman MD,

Jeffery G. Chipman MD. Department of Surgery, Department of Infectious Disease and International Medicine, University of Minnesota, Minneapolis, MN, 55455

Introduction: HIV has been shown to actively replicate in lymphatic tissue, even in the setting of viral suppression by antiretroviral agents. Additionally, it is these viral reservoirs that have shown to be the source of recrudescing virus upon cessation of HAART.1 Therefore, obtaining lymph nodes from HIV-positive individuals is critical in order to study this phenomenon. Lymph node biopsy has shown to be safe; however, palpating non-enlarged lymph nodes can be difficult and limit successful yield. We hypothesize that ultrasound guidance increases the yield of lymph nodes when utilized by surgeons experienced in POC ultrasound. Methods: We retrospectively reviewed prospectively collected data on 91 patients from both the United States and Uganda who underwent one or more inguinal lymph node biopsies. We compared the success rate (i.e. lymph tissue is obtained) of ultrasound guided biopsies versus biopsies guided on palpation alone. A Fischer exact test was used for this comparison. Amongst a smaller subgroup of patients (US patients only), we also examined the success rate of ultrasound guided biopsy in patients with a BMI >25 as well as compared lymph node excision times between ultrasound guided biopsies versus palpation guided biopsies. A Fischer exact test was used for the later two, as well as a Wilcoxon rank sum test for comparison of excision times. Results: Amongst both Uganda and US patients, ultrasound guided biopsy showed a significantly higher success rate compared to palpation alone (P=0.029293872, Table 1). Amongst a smaller subset of patients (US only), excision time and BMI were also considered. In those with BMI > 25, the odds ratio suggests that the success rate was higher with the use of ultrasound vs. palpation alone; however, this difference was not statistically significant (OR=2.23; P=0.53, Table 2). The mean excision time was not significantly different between ultrasound guided biopsy vs palpation guided biopsy, however the ultrasound guided biopsies seem to have a smaller variance (P=0.9621, Figure 1).

Ultrasound used No Yes

Lymph node obtained

No 7 7 Yes 28 92

Table 1: count table of lymph node obtained versus ultrasound used

Ultrasound used No Yes

Lymph node obtained

No 1 1 Yes 13 30

Table 2: count table of lymph node obtained versus ultrasound used for patients with BMI >25

Conclusion: Lymph node biopsy is extremely important in the study of HIV replication, as viral replication within lymph nodes has been shown to occur in the absence of a detectable viral load and subsequently remerges into systemic circulation upon cessation of HAART. We successfully rejected our first null hypothesis and found the following: ultrasound guided lymph node biopsy has a statistically significantly higher success rate in lymph node retrieval vs. palpation guided biopsy. We were unable to successfully reject the second and third null hypotheses, however this was likely due to a limited sample size. BMI data and excision times were only available for lymph node biopsies done in the United States and thus, the data from Uganda could not be analyzed. The odds ratio suggests, however, that ultrasound is associated with higher biopsy success rate in those with BMI >25. In conclusion, ultrasound is a safe and effective adjunct method to increase the effectiveness of lymph node biopsy. References 1. Fletcher, C. V. et al. Persistent HIV-1 is associated with lower antiretroviral drug concentrations in

lymphatic tissues. Proc. Natl Acad. Sci. USA 111, 23072312 (2014)

#15

Tumor location impacts immune response in mouse models of colon cancer

Xianda Zhao1, Lihua Li1, Timothy Starr2, and Subbaya Subramanian1,3*

1Department of Surgery, University of Minnesota Medical School, Minneapolis, MN 2Department of Obstetrics and Gynecology and Womens Health, University of Minnesota Medical School, Minneapolis, MN 3Masonic Cancer Center, University of Minnesota, Minneapolis, MN

Existing preclinical models of human colorectal cancer (CRC) that rely on syngeneic subcutaneous

grafts are problematic, because of increasing evidence that the immune microenvironment in

subcutaneous tissue is significantly different from the gastrointestinal tract. Similarly, existing

orthotopic models that use a laparotomy for establishing grafts are also problematic, because the

surgical procedure results in extensive inflammation, thereby creating a nonphysiologic tumor

microenvironment. To facilitate the bench-to-bedside translation of CRC immunotherapy

strategies, we developed a novel orthotopic model in mice that uses endoscopy-guided

microinjection of syngeneic cancer cells. When we compared immune system infiltration, we

found that tumors in the subcutaneous model had fewer T cells, B cells, and natural killer (NK)

cells, but more immunosuppressive myeloid cells; in contrast, tumors in our model had a higher

number of tumor-infiltrating T cells, B cells, and NK cells, with fewer immunosuppressive

myeloid cells. The number of immune-stimulating cytokines, such as interleukin (IL)-2, IL-6,

interferon (IFN)-gamma, and granzyme B, was also higher in tumors in our model, as compared

with the subcutaneous model. Those differences resulted in heightened sensitivity to immune

checkpoint blockade therapy in our endoscopy-guided orthotopic CRC model. Our study indicates

that tumor location affects immune response in CRC mouse models; choosing the appropriate

preclinical model is important when testing immunotherapy in CRC.

#16

Integrative analysis of colorectal cancer tissue microRNA and microbiome reveals microRNAs as potential mediator of host-microbiome interaction

Ce Yuan1,2, Michael Burns3, Subbaya Subramanian1,2*, Ran Blekhman1,4,5*

1. Bioinformatics and Computational Biology Program, University of Minnesota, Rochester, MN, 55904.

2. Department of Surgery, University of Minnesota, Minneapolis, MN, 55455. 3. Department of Biology, Loyola University Chicago, Chicago, IL, 60626 4. Department of Genetics, Cell Biology, and Development, University of Minnesota,

Minneapolis, MN, 55455. 5. Department of Ecology, Evolution, and Behavior, University of Minnesota, Saint Paul, MN,

55108.

Background Variation in the composition of the gut microbiome has been linked with colorectal cancer (CRC). However, it is unclear what factors may mediate the interactions between CRC and the microbiome. MicroRNAs (miRNAs) are known to regulate CRC progression and patient survival. Recent studies have found suggested that host miRNAs can also regulate bacterial growth and are important in shapingalter the composition of the gut microbiome. Here, we investigated the correlations between the gut microbiome compositions and miRNAs expression in human CRC tissues. Method We sequenced the small RNAs from patient-matched tumor and normal samples collected from 44 human CRC patients and jointly analyzed with the microbiome taxonomic composition data from the same samples. We then interrogated the function of bacteria correlated with differentially expressed miRNAs and the function of miRNAs correlated with two bacterial taxa, Fusobacterium and Providencia, both associated with CRC. Results We identified a total of 112 differentially expressed miRNAs in tumor tissues, including miR-182, miR-503, and miR-17~92 cluster, all are known oncogenic miRNAs in CRC. The predicted functional profiles of bacteria correlated with differentially expressed miRNAs show enrichment in nutrient metabolism and biomolecule biosynthesis pathways. The predicted targets of miRNAs correlated with Fusobacterium and Providencia show enrichment in CRC related pathways and genes. Conclusions This work investigated the correlations between the gut microbiome compositions and miRNAs expression in human CRC tissues. The results indicate miRNAs play a role in mediating a bi-directional host-microbiome interaction in CRC. Our research provides new evidence and new research directions for host-microbiome interactions in CRC with the potential to improve the efficiency of current therapies. Our data suggests a new paradigm that miRNAs mediate host-microbiome interactions.

#17 The Undergraduate Surgery Interest Group (USIG): A New Venue for Premedical

Student Exposure to the Field of Surgery Hedberg J1, Vakayil V, MBBS2; Kernahan, P, M.D., Ph.D., FACS2; Hendrickson, Lois, Vadhul R1, Prokhoda P1, Peyer M1, Malik S1, Harmon JV, M.D., Ph.D., FACS2

1University of Minnesota College of Biological Sciences, 2University of Minnesota Department of

Surgery Introduction Early exposure to the field of surgery through workshops, lectures, and mentors may bolster and prolong interest in surgical careers for undergraduate students. An Undergraduate Surgery Interest Group (USIG) was started through the initiative of students, seeking to create more opportunities through which younger students could develop and explore their motivations for advancing toward surgical practices. This USIG has received mentorship, resources, and support from the University of Minnesota Medical School Department of Surgery (DOS). Methods We conducted a comprehensive one-year assessment of our institutions USIG. We reviewed its organizational structure, funding resources, media promotion, and educational activities. We analyzed the evaluations completed by USIG event participants and compared them to similar questionnaires completed by medical students in the Surgery Interest Group (MSIG), a medical student organization at the University of Minnesota. Our comparative analysis sought to understand the perceptions of a surgical career among premedical and medical students in hopes of identifying the differences between the two groups. Results USIG was met with enthusiasm and participation by premedical students and faculty from the DOS. Currently there are over 300 active participants. Weekly board meetings are held to coordinate biweekly events, which include surgical skills workshops, tours of medical laboratories, and faculty-led seminars. Social media has helped in promoting USIG recognition and participation. Post-event evaluations further attested the positive impact USIG had on its participants; students stated increased confidence in their surgical knowledge and skill. A comparison between the premedical and medical student survey results helped identify potential areas for educational refinement. Conclusion Our USIG provides unique opportunities for premedical students to gain exposure to the field of surgery and venues to interact with medical students, residents, and surgical faculty, while promoting long-term commitment to a surgical career. Other academic institutions may find merit in investing resources to promote undergraduate surgical education by supporting similar groups.

#18

Introducing Bedside Ultrasound in Human Anatomy Laboratory: A Quantitative Feasibility Assessment

SUNDIN, Andrew N,1,2 Philip ROBAN,2 Mary A. MCNEIL,3 Victor R. VAKAYIL,1 Brent BAUMAN,1 Malavika CHANDRASHEKAR,1 Peter J. KERNAHAN,1, 2 Anthony J. WEINHAUS,2 James V. HARMON,1, 2 1Department of Surgery, University of Minnesota 2Department of Integrative Biology and Physiology, University of Minnesota 3Department of Emergency Medicine, University of Minnesota

INTRODUCTION: Rectus femoris muscle thickness (RFMT) measurement using point-of-care ultrasound (POC-U) is now being used to evaluate patient nutritional status. Introducing hands-on POC-U instruction in the medical school anatomy laboratory may be beneficial and clinically relevant to patient care. METHODS: Medical students were provided faculty-guided demonstrations in the use of POC-U to measure the RFMT in the parasagittal (PS) and transverse (T) planes. Measurements were made at a point one-third the distance along the line between the superior border of the patella and the anterior superior iliac spine. Each measurement was repeated three times by each student. Subsequently, students manually measured the RFMT using calipers immediately following the anatomic dissection of the lower extremity. In order to assess the reproducibility and accuracy of the students measurements, intra- and inter- observer reliability and inter-method reliability were analyzed using a repeated measure ANOVA and repeated measures factorial ANOVA. RESULTS: Two medical students measured the RFMT in 44 cadavers using POC-U and calipers. The average measure of RFMT using the POC-U in the T-plane was 7.8mm 2.7mm and 7.8mm 2.8 in the PS-plane. Average RFMT using calipers was 9.5mm 2.5mm. Comparative analysis of the students POC-U and caliper measurements of the T- and PS- planes revealed no significant differences in intra- and inter- observer measurements (P>0.05). RFMT measurements obtained from POC-U significantly differed from those gathered manually with calipers (P=0.001). SUMMARY: POC-U may be used to accurately to measure the RFMT to assess the clinical nutritional status of the patient. Observed differences between the two techniques may be due to (i) intrinsic differences in the two modalities, (ii) residual soft tissue, or (iii) increased muscle diameter from relaxation after dissection. Subsequent studies will directly compare POC-U and caliper in the dissected specimen.

Representation of Occupation Information in Clinical Texts: An Analysis of Free-Text Clinical Documentation in Multiple Sources

Elizabeth A. Lindemann, BS1, Elizabeth S. Chen, PhD2, Sripriya Rajamani, MBBS, PhD, MPH3, 4, Nivedha Manohar, MHI5, Yan Wang, PhD3, Genevieve B. Melton, MD, PhD1, 3

1Dept. of Surgery, 3Institute for Health Informatics, 4Public Health Informatics Program Univ. of Minnesota, Minneapolis, MN; 2Ctr. for Biomedical Informatics, Brown Univ.,

Providence, RI; 5West Side Community Health Services, St. Paul, MN Introduction

As electronic health record (EHR) system use and efforts to leverage associated information sources increases, the need to accurately capture occupation and other social history at the point of care becomes central for effective primary use in clinical care and secondary uses such as research. The National Institute for Occupational Safety and Health (NIOSH) has worked to promote documentation of occupation information and created the Occupational Data for Health (ODH) data model1. Previous work2, 3 has analyzed standards and EHR occupation data entry to inform representation and standardization of occupation information. Standardization in entry could lead to more refined standards, and ultimately, better understanding of how social determinants affect health outcomes. The objective of this study was to build on prior efforts by describing occupation within multiple clinical text sources and further inform standardization of this information in the EHR to provide a more complete picture of occupation.

Methods

This study utilized six clinical note sources: 491 Consult History and Physical MTSamples (MTS) notes and 200 de-identified H&P (History and Physical Examination) University of Pittsburgh Medical Center (UPMC) notes from the public domain, and four note types from Fairview Health Services Epic EHR. Sentences were anonymized using the Safe Harbor Method4. Annotation schema and guidelines were derived from the NIOSH ODH Model, providing the parent entities Occupational History, Usual Occupation, Employment Status, Occupational Injury, Occupational Exposure, and associated child entities. Three annotators (EL, SR, NM) annotated 25 sentences as a group, and 50 overlapping sentences individually to evaluate the annotation schema and guidelines (Figure 1). Entities were added for Subject, Negation, Temporal, Occupational Conditions, and Occupation Status. Annotations were made at the most specific level of detail possible (e.g., nurse is annotated as Occupation Description, and rather than the parent entity, Occupational History). The annotators then coded a set of 10% sentences to calculate inter-rater reliability (Cohens kappa of 0.76 and proportion agreement of 0.95).

Results

A total of 868 sentences were mapped to 41 entities (Table 1) for a total of 2,005 annotations. The most frequent entities were: Occupation Description (17.7%), Employment Status Not Specified (12.5%), Employer Name (11.0%), Subject (9.8%), and Industry Description (6.2%). There were no annotations for the Volunteer and Usual Occupation entities. The Fairview Social History Documentation sentences contained 39 (95%) of 41 possible entities and had the greatest variety of occupation information. Fairview Physical Therapy Notes contained 13 (32%) of the 41 possible entities, making it the note source with the least diversity in occupation information.

Discussion

The findings of this study demonstrate the broad nature of occupation information in clinical text. The ODH model was adequate in representing most of the occupation information, but several additions to the schema were needed for comprehensive representation in our clinical text corpus. The volume of Subject related information points to a need for structured entry of occupation information for spouses and parents of patients. Ambiguity persisted in several Occupation Status related entities, such as Self-Employed and Homemaker/Housewife. For example, the term stay at home mom implies both a caretaker role as well as a homemaker role. Self-employment could imply Full-time or Part-time Employment, in addition to Self-employment. In the future, additional work focusing on the distribution of values across entities will help in extending the ODH model and with standardized entry of occupation information to maintain data quality within the EHR ultimately improving patient care and secondary use by providing a basis for natural language processing efforts.

Figure 1: Overview of Evaluation of Occupation Information in Clinical Text

Sentences containing Occupation Information from six note

sources

Annotation Schema

Creation, Expert Agreement, Annotation

Process

Distribution of Annotations describing

Occupation in Clinical Texts

erinbSticky NoteUnmarked set by erinb

erinbTypewritten Text#19

Table 1: Distribution of occupation entities across sources.

Entities MTS

[171] (n=385)* UPMC

[44] (n=101)*

Fairview SH Documentation [553] (n=1295)*

Fairview SW Notes

[58] (n=113)*

Fairview OT Notes

[16] (n=65)*

Fairview PT Notes

[26] (n=46)*

All Sources [868]

(n=2005)* Occupational History - - 2 (0.2%) - 1 (1.5%) - 3 (0.1%)

Industry Description 22 (5.7%) 10 (9.9%) 80 (6.2%) 6 (5.3%) 5 (7.7%) 2 (4.3%) 125 (6.2%) Occupation Description 77 (20.0%) 24 (23.8%) 227 (17.5%) 15 (13.3%) 3 (4.6%) 8 (17.4%) 354 (17.7%) Job Duties 13 (3.4%) - 55 (4.2%) 3 (2.7%) 3 (4.6%) 2 (4.3%) 76 (3.8%) Employer Name 36 (9.4%) 7 (6.9%) 151 (11.7%) 18 (16.0%) 1 (1.5%) 8 (17.4%) 221 (11.0%) Employer Location 7 (1.8%) 4 (4.0%) 48 (3.7%) 3 (2.7%) - 3 (2.2%) 65 (3.2%)

Usual Occupation - - - - - - - Usual Industry Description 1 (0.3%) - 3 (0.2%) - - - 4 (0.2%) Usual Occupation Description 5 (1.3%) 1 (1.0%) 4 (0.3%) - - - 10 (0.5%)

Occupation Status 1 (0.3%) - 3 (0.2%) - - - 4 (0.2%) Employment Status 4 (1.0%) - 1 (0.1%) - - - 5 (0.2%)

Full-Time 2 (0.5%) 1 (1.0%) 24 (1.9%) 4 (3.5%) - 6 (13.0%) 37 (1.8%) Part-Time 3 (0.8%) - 9 (0.7%) 4 (3.5%) - 1 (2.2%) 17 (0.8%) Not Specified 46 (11.9%) 13 (12.9%) 141 (10.9%) 4 (3.5%) 41 (63.1%) 6 (13.0%) 251 (12.5%)

Self-Employed 4 (1.0%) 1 (1.0%) 2 (0.2%) 2 (1.8%) 1 (1.5%) 1 (2.2%) 11 (0.5%) Employed but temporarily not working 1 (0.3%) 1 (1.0%) 1 (0.1%) 4 (3.5%) 1 (1.5%) - 8 (0.4%) Not Employed 12 (3.1%) 10 (9.9%) 30 (2.3%) 3 (2.7%) - - 55 (2.7%) Retired 29 (7.5%) - 56 (4.3%) 1 (0.9%) - - 86 (4.3%) Disability (Not employed) 9 (2.3%) - 4 (0.3%) 1 (0.9%) - - 14 (0.7%) Military Service 2 (0.5%) 1 (1.0%) 1 (0.1%) - - - 4 (0.2%) Student - - 22 (1.7%) - - - 22 (1.1%)

Student Type 1 (0.3%) - 19 (1.5%) - - - 20 (1.0%) Student Status 5 (1.3%) - 9 (0.7%) 1 (0.9%) - - 15 (0.7%) Student Other 3 (0.8%) - 55 4.2%) 1 (0.9%) 1 (1.5%) 2 (4.3%) 62 (3.1%)

Volunteer - - - - - - - Not in Paid Workforce (Child) 3 (0.8%) - 25 (1.9%) - - - 28 (1.4%) Homemaker/Housewife 8 (2.1%) - 30 (2.3%) 1 (0.9%) 1 (1.5%) - 40 (2.0%) Caretaker 2 (0.5%) - 6 (0.5%) - - - 8 (0.4%)

Occupational Injury 2 (0.5%) 1 (1.0%) 3 (0.2%) - - - 6 (0.3%) Occupational Exposure 1 (0.3%) 2 (2.0%) 2 (0.2%) - - - 5 (0.2%) Occupational Conditions 8 (2.1%) - 7 (0.5%) 3 (0.2%) 1 (1.5%) 4 (8.7%) 23 (1.1%) Temporal 1 (0.3%) - 3 (0.2%) - - - 4 (0.2%)

Work Schedule 3 (0.8%) - 5 (0.4%) 2 (1.8%) 1 (1.5%) 2 (4.3%) 13 (0.6%) Start Date 4 (1.0%) - 18 (1.4%) 2 (1.8%) - - 24 (1.2%) End Date 2 (0.5%) - 3 (0.2%) - - - 5 (0.2%) Days Per Week - 1 (1.0%) 5 (0.4%) - 1 (1.5%) - 7 (0.3%) Hours Per Week - - 3 (0.2%) - 1 (1.5%) 1 (2.2%) 5(0.2%) Duration Years 1 (0.3%) 3 (3.0%) 10 (0.8%) 5 (4.4%) 1 (1.5%) - 33 (1.6%) Time Frame 38 (9.9%) 20 (19.8%) 55 (4.2%) 6 (5.3%) 1 (1.5%) - 120 (6.0%)

Subject 10 (2.6%) 1 (1.0%) 162 (12.5%) 23 (20.4%) 1 (1.5%) - 197 (9.8%) Negation 6 (1.6%) - 11 (0.8%) 1 (0.9%) - - 18 (0.9%) # of Entities 35 16 39 23 17 13 39 [# of sentences per source]; *number of annotations; SH=Social History; SW=Social Work; OT=Occupational Therapy; PT=Physical Therapy

Acknowledgements

This work was supported in part by National Library of Medicine grant R01LM011364.References

1. National Institute for Occupational Safety and Health (NIOSH). Structuring Patient Work Information in EHRs to Improve Patient Care and Public Health; Occupational Data for Health (ODH) Model. Public Health Informatics Conference; Atlanta, GA. 2014.

2. Rajamani S, Chen ES, Aldekhyyel R, Wang Y, Melton GB. Validating the Occupational Data for Health Model: An Analysis of Occupational Information in Reports, Standards, Surveys, and Measures. AMIA Annu Symp Proc. 2016: 115-116.

3. Aldekhyyel R, Chen ES, Rajamani S, Wang Y, Melton GB. Content and Quality of Free-Text Occupation Documentation in the Electronic Health Record. AMIA Annu Symp Proc. 2016: 1708-1716.

4. U.S. Department of Health and Human Services. Guidance on Satisfying the Safe Harbor Method. [cited 2016 September]. Available from: https://www.hhs.gov/hipaa/for-professionals/privacy/special-topics/de-identification/index.html#safeharborguidance.

#20 Accelerating American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) Surgical Site Infection Abstraction with a Semi-Automated Approach Skube SJ, Hu Z, Simon GJ, Wick EC, Arsoniadis EG, Jensen EH, Kwaan MR, Rothenberger DA, Ko CY, Melton GB Objective: Despite the value of ACS-NSQIP, 30-day outcome abstraction remains laborious. We applied supervised machine learning algorithms, previously developed for surgical site infection (SSI) detection from electronic health record structured clinical data and unstructured documents through natural language processing (NLP) at a single-institution for semi-automated SSI abstraction. Methods: A Lasso-penalized logistic regression model was trained on 2011-3 data. Training dataset performance was measured with 10-fold cross-validation (Table 1). A cutoff probability score was established to divide each dataset into negative and possible SSI using four unique algorithms (overall, superficial, deep, and organ space). Manual data abstraction was only performed on the possible group, semi-automating the data abstraction process. Algorithms were separately evaluated on 2014 data, 2015 data, and overall. Results: With 2011-3 data as training (N=6,188) and 2014-5 as evaluation (N=5,132), algorithms had good performance (Table 2). Negative groups had

2

Characteristics of 171 Combat-Associated Small Bowel Injuries Mariya E. Skube MD, Quinn Mallery BS, Elizabeth Lusczek PhD, Greg J. Beilman MD

Introduction: Although there are multiple studies regarding the management and outcomes of colonic injuries incurred in combat, the literature is limited in regards to small bowel injuries. This study seeks to address that void by providing the largest reported review of the characteristics of combat-associated small bowel injuries.

Methods: The Department of Defense Trauma Registry was queried for members of the United States Armed Forces who sustained hollow viscus injuries in the years of 2007 to 2012 during Operations Enduring Freedom, Iraqi Freedom, and New Dawn. Patients with other backgrounds (i.e. North Atlantic Treaty Organization troops, local nationals) were excluded. Service members with injuries of the small bowel were identified by diagnosis codes, and concomitant injuries, procedures, and complications were delineated. After summarizing the data, Fishers exact test was used to analyze the relationship of bowel injury pattern to rates of repeat laparotomy, fecal diversion, and complications.

Results: One hundred seventy-one service members had small bowel injuries. The mean age was 25.8 6.6 years with a mean injury severity score of 27.9 12.4. The majority of injuries were penetrating (94.2%, n=161) as a result of explosive devices (61.4%, n=105). The median blood transfusion requirement in the first 24 hours was 6.0 units (IQR, 1.0-17.3); 48 patients received at least 10 units. The most frequent concomitant injuries were large bowel (64.3%, n=110), pelvic fracture (35.7%, n=61), perineal (26.3%, n=45), liver (20.5%, n=35), and pelvic organ (19.9%, n=34). Fifty patients (29.2%) had a colostomy, and 9 patients (5.3%) had an ileostomy. 62.6% (n=107) of soldiers underwent more than one laparotomy. The mortality rate was 1.8% (n=3). Median length of stay was 13 days (IQR, 5-38). The most common complications were pneumonia (15.2%, n=26), deep vein thrombosis (14.6%, n=25), wound infection (14.6%, n=25), and pulmonary embolus (12.9%, n=22). The need for repeat laparotomy and fecal diversion (ileostomy and/or colostomy) were found to be significantly associated with injury pattern (p= 0.00052 and p=

3

Table 1. Repeat laparotomy, fecal diversion, and complications with respect to injury pattern. Outcome n (%) p value Repeat Laparotomy 0.00052 SB 21 (39.6) SB+LB 65 (71.4) SB+Rect 6 (75.0) SB+LB+Rect 15 (78.9) Fecal Diversion

2

Islet Graft Function is Preserved after Pregnancy in Patients with Previous Total Pancreatectomy with Islet Autotransplant

Mariya Skube MD, Pamela Mills MD MPH, James Hodges PhD, Gregory Beilman MD, Melena Bellin MD

Introduction: Pregnancy increases insulin demand, and this effect requires careful management in women with pregestational diabetes. In addition, diabetes increases the risk of maternal and fetal complications such as preeclampsia, preterm delivery, and infant macrosomia. Women who have underwent total pancreatectomy with islet autotransplantation (TPIAT) and subsequently become pregnant have not been well studied to examine effects on islet graft function as well as to identify potential pregnancy complications. We hypothesized that similar to women with pregestational diabetes, pregnancy in TPIAT patients temporarily stresses graft function with a subsequent return to baseline function after pregnancy. Methods: Women who underwent TPIAT at our institution and subsequently had successfully completed a pregnancy were eligible for enrollment. Data on graft function (glycosylated hemoglobin, insulin requirements, and functional classification) as well as pregnancy course were collected from the medical records and the institutional TPIAT database. A control group was selected from our female TPIAT patients without pregnancies who were of childbearing age at the time of transplant; matching was completed for year of transplant, age and body mass index at transplant, and islet yield (Table 1). Case-control analysis was performed using Pearsons chi-squared and Mantel-Haenszel tests. Results: Five patients (cases) representing 7 pregnancies were enrolled. The pregnancies spanned the years from 2006 to 2016. The median time from TPIAT to conception was 21 months (IQR, 8-163). 71.4% (n=5) of the pregnancies were complicated by preterm labor, and the caesarean section rate was 57.1% (n=4). One infant had a congenital anomaly (sacrococcygeal teratoma). Median gestational age at birth was 36 weeks (IQR, 35.5-38.0) with a median birth weight of 2800 grams (IQR, 2325-3254). There were no significant differences identified between the cases (post-partum) and controls in the primary outcomes of glycosylated hemoglobin (mean [SE], 6.8 [0.87] vs. 6.4 [0.62], p=0.68), insulin use (16.6 units [9.6] vs. 12.6 units [7.6], p=0.64), and graft function categorization of failed, partial, or full function (p=0.82) (Table 2). Of note, the mean follow-up time for cases was longer than control follow-up for each primary outcome: glycosylated hemoglobin (77 months [16.7] vs 41 months [12.3], p=0.045); insulin use (78 months [18.2] vs 39 months [13.4], p=0.051); graft function (75 months [18.1] vs 41 [12.6], p=0.10). Conclusions: Long-term graft function was comparable between the cases and their matched controls. Pregnancy did not have a negative impact on long-term graft function in women with a history of TPIAT.

erinbTypewritten Text#22

3

TABLE 1. Comparison of baseline characteristics of cases and controls.

Case (n=5) Control (n=14) p value

Age at transplant (years), mean (SE) 24.8 (2.8) 28.4 (2.0) 0.21

BMI (kg/m2), mean (SE) 22.0 (1.5) 22.6 (1.0) 0.75

Islet yield (IEQ/kg), mean (SE) 4474 (1065) 4336 (1054) 0.55

SE, standard error. BMI, body mass index. IEQ, islet equivalent.

TABLE 2. Analysis of primary endpoints for cases and controls. Case (n=5) Control (n=14) p value HgA1C, mean (SE) 6.8 (0.87) 6.4 (0.62) 0.68 Time to follow-up (mo) 77 (16.7) 41 (12.3) 0.045 Insulin (units/day), mean (SE) 17.2 (9.5) 12.6 (7.6) 0.59 Current insulin use, no (%) 2 (40) 5 (36) 0.23 Time to follow-up (mo) 78 (18.2) 39 (13.4) 0.051 Graft function 0.82 Failed, no (%) 1 (20) 3 (21) Partial, no (%) 1 (20) 8 (57) Full, no (%) 3 (60) 3 (21) Time to follow-up (mo) 75 (18.1) 41 (12.6) 0.10 HgA1C, glycosylated hemoglobin. Mo, month. SE, standard error.

#23 Aortic Center Directed Risk Factor Verification Optimizes An EMR Based AAA Screening Program Stacy Carda, Yifei Sun, Clarence Ojo, Michael Rosenberg, Alan Hirsch, Susan Shafii, Rumi Faizer Objectives The U.S. preventive services task force recommends one-time screening for abdominal aortic aneurysm (AAA) with ultrasonography in men aged 65-75 years who have ever smoked. The purpose of this study was to evaluate the implementation of the SAAAVE Act in identifying patients at risk of AAA between May 2016 and January 2017 in clinical practice at our institution. Methods Data was extracted from the EMR database to identify men aged 65-75 years who smoked greater than 100 cigarettes during their lifetime. Medical records were reviewed electronically to exclude patients with abdominal imaging studies within last 5 years that would have diagnosed an AAA. AAA was defined as aortic diameter of 3.0 cm or greater. Identified AAAs were triaged following consultation with vascular surgery or with their primary care physician. The number of eligible patients, unscreened patients, comorbidities, and AAAs identified were quantified. Results A total of 1268 men aged 65 to 75 years (69 3.2 years) were screened for an AAA between May 2016 to January 2017. Among the cohort, 41.7% patients were identified to be eligible for ultrasound scan, of which 13.0% were current smokers, 84.3% were former smokers, and 2.5% had never smoked. Of those who had never smoked, 23.1% had a positive family history for AAA (2.1% of total). In addition, 3.2% had an abdominal imaging study performed within the last 5 years that would have identified AAA. 1.1% had major co-morbidities such as heart and/or neurological disease, cancer, and end-stage renal failure which would have made the screening inappropriate. 0.6% of patients refused screening. Among the patient cohort eligible for ultrasonography scan, 26.7% of patients were found to be diabetic who may represent a subpopulation with a decreased yield of screening; and 4.7% had PAD which may represent a subpopulation with an increased yield of screening. A total of 17 aneurysms (8.3%) were found, that included 14 AAAs (6.8%) and diagnosis of 3 incidental iliac aneurysms (1.5%). Among this subgroup of patients, 35.3% had diabetes and 5.9% had PAD. Conclusions In summary, systematic AAA screening augmented by patient calls effectively eliminates the need to screen and reduce unnecessary burden on the imaging departments. 77% of patients never smoked history is the most common reason not to screen for AAA in our patient cohort.

#24 Pre-operative Ankle-Brachial Index Stratification for Patients with Peripheral Arterial Disease Alex Sun, Judy Madson, Eric Weinhandl, Clarence Ojo, Alan Hirsch, Susan Shafii, CJ Lee, Rumi Faizer Introduction: The Ankle-Brachial Index (ABI) is a well-accepted tool to assess severity of peripheral arterial disease (PAD). Categorization of ABI values to match clinical PAD severity (claudication, rest pain and tissue loss) is based on limited data of a few hundred patients from 1970 and 1996. The American Heart Association guidelines recommended in 2011 to change reporting of ABIs for cardiovascular risk stratification to normal, abnormal, borderline and non-compressible. As such, reporting categories for ABI in PAD need reevaluation. Methods: We analyzed data from the Vascular Quality Initiative (VQI) of 130,930 patients, who have undergone either peripheral vascular intervention (PVI) or infra-inguinal bypass (INFRA) between Jan 2003 to Feb 2017. The patients pre-operative ABI and toe-brachial index (TBI) were correlated with their disease severity in 3 major categories. The data is also stratified into diabetic and non-diabetic patients. The patients with ABI value between 0 and 1.3, and TBI value between 0 and 1.0 were included. Results: Overall, patients with symptoms of claudication, rest pain and tissue loss had mean ABIs of 0.66, 0.47 and 0.54, respectively. 80% of patients with claudication had ABIs ranging from 0.95 to 0.39. Similarly, 80% of patients with rest pain and tissue loss had ABIs ranging from 0.83 to 0, and 0.94 to 0.16, respectively. Subgroup analyses of patients with and without diabetes did not show a statistically significant difference except in patients with tissue loss. Pre-operative ABI and TBI in patients treated with bypass were lower than those in the PVI group. Conclusions: Analysis of a large, multicenter dataset of peripheral vascular procedures suggests that while mean ABI and TBI values are different for patients with claudication than for those with rest-pain or tissue loss, significant overlap exists between categories. As with ABI and cardiovascular risk, it may be more appropriate to move to report ABIs in PAD to categories of normal, abnormal, and non-compressible.

. . . . . .

INFRA

PVI

#25 Lung Transplantation from Donation after Circulatory Death Donors after Prolonged Normothermic Ex Vivo Lung Perfusion John R. Spratt MD, MA, Lars M. Mattison, BS, Paul A. Iaizzo, PhD, Carolyn Meyer, BS, Roland Z. Brown, BA, Tinen Iles, MS, Angela Panoskaltsis-Mortari, PhD, and Gabriel Loor, MD Purpose Portable normothermic ex vivo lung perfusion (EVLP) has been evaluated in human clinical trials of standard and extended-criteria donor lungs. We describe a swine model of lung transplant from donation after circulatory death (DCD) donors using prolonged normothermic EVLP to clarify the effects of pre-preservation warm ischemia on acute lung transplant outcomes. Methods Adult swine were anesthetized and heparinized. In the control group (n=4), lungs were procured, flushed, and transplanted immediately. In the treatment groups, swine underwent either beating-heart procurement (n=3) or hypoxic cardiac arrest followed by either 1 (n=4) or 2 hours (n=4) of ventilated warm ischemia. Lungs were then preserved for 24 hours using normothermic blood-based EVLP, followed by single lung transplant. Recipient pulmonary artery (PA) pressure, dynamic compliance (Cdyn), and oxygenation were then monitored for 4 hours. Results Prior to transplant, all EVLP groups demonstrated equivalent PA pressures, dynamic compliance, and oxygenation after 24 hours of preservation. After transplant, all EVLP groups showed equivalent systemic and allograft oxygenation, but all were inferior to control. Post-transplant PA pressure and Cdyn corresponded to donor ischemic injury severity. Post-transplant allograft oxygenation was best predicted by persistent PA pressure and Cdyn abnormalities and the severity of pulmonary edema on EVLP. Conclusion Donor warm ischemia does not independently predict acute post-transplant function in swine lungs after prolonged normothermic EVLP. Favorable post-transplant function is predicted by low PA pressures, good compliance, and minimal pulmonary edema during EVLP.

#26