Medical College of Wisconsin
Research Publication Series:
November 2017 2 3 4 5
Soyoung Kim, PhD “Combining Biomarkers for Classification with Covariate Adjustment”
Matthew S. Karafin, MD, MS “Older Red Cell Units Are Associated with an Increased Incidence of Infection in Chronically Transfused Adults with Sickle Cell Disease”
Anna Palatnik, MD “Medical and Obstetric Complications Among Pregnant Women With Liver Cirrhosis”
Cassie Craun Ferguson, MD “Reducing Unnecessary Imaging for Patients With Constipation in the Pediatric Emergency Department”
6
Ha Son Nguyen, MD
6
Kai Yang, MD
7
7
“Intracranial Subependymoma: A SEER Analysis 2004-2013”
“Increasing Calcium Level Limits Schwann Cell Numbers In Vitro following Peripheral Nerve Injury”
Xiaojian Chen “Assessment of Treatment Response during Chemoradiation Therapy for Pancreatic Cancer Based on Quantitative Radiomic Analysis of Daily CTs: An Exploratory Study”
Selvi Kunnimalaiyaan, MS “Suberoylanilide Hydroxamic Acid, a Histone Deacetylase Inhibitor, Alters Multiple Signaling Pathways in Hepatocellular Carcinoma Cell Lines”
All rights reserved. Contents are the property of the authors and/or journals cited.
Soyoung Kim, PhD Assistant Professor Institute for Health & Equity Medical College of Wisconsin
I am an Assistant Professor in the Division of Biostatistics at Medical College of Wisconsin. Currently, I am serving as PhD statistician in Center for International Blood & Marrow Transplant Research. My current research interests include the development and application of statistical methods for research problems: survival analysis, casual inference, biomarker evaluation, HIV/AIDS vaccine trials, missing data, and case-cohort studies.
“Combining Biomarkers for Classification with Covariate Adjustment” Statistics in Medicine. 2017;36(15):2347-2362. Combining multiple markers can improve classification accuracy compared with using a single marker. In practice, covariates associated with markers or disease outcome can affect the performance of a biomarker or biomarker combination in the population. The covariate-adjusted receiver operating characteristic (ROC) curve has been proposed as a tool to tease out the covariate effect in the evaluation of a single marker; this curve characterizes the classification accuracy solely because of the marker of interest. In this research, we examine the effect of covariates on classification performance of linear marker combinations and propose to adjust for covariates in combining markers by maximizing the nonparametric estimate of the area under the covariate-adjusted ROC curve. We apply the proposed method to a biomarker study from an human immunodeficiency virus vaccine trial. Table 1. Data Analysis Results for Combining Two Biomarkers: X1=lgA Antibody Binding to Env. Reported Are Estimate Marker Combinations and Corresponding Gender-Specific AUC Values Estimated by Cross-Validation.
AAUC: Covariate-adjusted AUC; LGT: logistic regression; LZ: parametric estimator proposed by Liu and Zhou; S^E(^β): standard error of ^β estimated by 200 bootstrap resamples.
Matthew S. Karafin, MD, MS Associate Medical Director, Medical Sciences Institute BloodCenter of Wisconsin Assistant Professor, Department of Pathology & CTSI Medical College of Wisconsin
I am board certified physician in Anatomic Pathology, Clinical Pathology, and Transfusion Medicine, and I have a Master’s of Science in Clinical and Translational Research from the CTSI. I currently serve as Assistant Director of the Froedtert Hospital Transfusion Service. My research focuses on the variable efficacy and quality of red cell blood products when used for adults with sickle cell disease and the factors that influence this variability.
“Older Red Cell Units Are Associated with an Increased Incidence of Infection in Chronically Transfused Adults with Sickle Cell Disease” Transfusion & Apheresis Science. 2017;56(3):345-351. We performed a retrospective cohort study of adults with sickle cell disease (SCD) managed with prophylactic simple transfusion regimens. We analyzed our data using a mixed effects model, which accounts for a subject’s contribution to repeated transfusion encounters, to investigate the association between storage age and the incidence of hospital encounters for infection and pain crises prior to the next red cell transfusion. Twenty-eight steady-state adults with SCD received 627 units via simple transfusion over 281 outpatient encounters. Overall median unit storage age was 22 days. Receipt of older units was significantly associated with an increased incidence of admission for infection (p=0.04). Consequently, we provide evidence that receipt of older units is associated with a higher rate of admission for infection. Future prospective studies will need to validate these findings. Figure. 1. Median distribution of red cell units provided as simple outpatient transfusions during the three year study period.
(A) Box-plot of all red cell unit ages provided over the study period (N = 627 RBC units, units = 22 days). (B) Box-plot of average red cell unit ages by transfusion encounter and separated by those that resulted in a subsequent clinical infection (interval infection: N = 19 encounters, units = 29 days; no interval infection: N = 262 encounters, units = 23 days). (C) Box-plot of average red cell unit ages by transfused patient (N = 28 patients, units = 23 days).
I am an Assistant Professor in the Department of Obstetrics and Gynecology, Division of Maternal Fetal Medicine. I received my Medical degree from Sackler School of Medicine at Tel-Aviv University, Israel, and went on to complete my residency in Obstetrics and Gynecology here at MCW. I then completed my fellowship in Maternal Fetal Medicine at the Northwestern University before joining MCW faculty in 2016. My research interests include clinical obstetrics, preeclampsia and gestational diabetes.
Anna Palatnik, MD Assistant Professor Obstetrics and Gynecology Division of Maternal Fetal Medicine Medical College of Wisconsin
“Medical and Obstetric Complications Among Pregnant Women With Liver Cirrhosis” Obstetrics and Gynecology. 2017;129(6):1118-1123. In this study we found that women whose pregnancy is complicated by liver cirrhosis had higher rates of primary composite outcome that included fetal or neonatal death, placental abruption, preeclampsia, preterm delivery <37 weeks, and small-for -gestational age infant was significantly more frequent among women with cirrhosis, 19 (61%) vs. 15 (12%), p<0.001, with adjusted OR of 9.4. In addition, we found that gestational age at delivery was also significantly lower among women with cirrhosis, with adjusted mean difference of about 3 weeks. Birth weight was also lower among women with cirrhosis with adjusted mean difference of 600g. The most common cause of liver cirrhosis was alcoholic liver disease. Table 3. Pregnancy Outcomes of Women With Liver Cirrhosis and Women in a Control Group Outcomes Gestational age at delivery (wks) Birth weight (kg) Primary composite outcome* Fetal or neonatal demise Placental Abruption Preeclampsia Preterm delivery Spontaneous preterm delivery Small-for-gestational age infant Mode of delivery Cesarean delivery Vaginal delivery Postpartum hemorrhage Apgar score <7 at 5 minutes pH ≤ 7.0 Peripartum maternal ICU admission Transfusion
Women w/ Liver Cirrhosis (n=31) 35.2 ± 6.0 2.7 ± 0.8 19 (61.3) 3 (9.7) 3 (9.7) 6 (19.4) 14 (45.2) 7 (22.6) 5 (16.1)
39.1± 2.1 3.4 ± 0.6 15 (12.1) 0 (0.0) 0 (0.0) 5 (4.0) 7 (5.7) 4 (3.2) 6 (4.8)
8 (26.7) 22 (73.3) 2 (6.5) 1 (3.9) 1(3.9) 1 (3.2) 1 (3.2)
38 (30.7) 86 (69.4) 6 (4.8) 1 (0.8) 0 (0.0) 0 (0.0) 0 (0.0)
Controls (n=124)
Estimated mean diff between cohorts -3.9 -0.7 11.5 5.7 13.8 7.1 4.1
95% Confidence Interval -5.2 - -2.6 -0.9 - -0.4 4.7 – 28.4 1.6 – 20.2 4.9 – 38.9 1.9 – 27.2 1.2 – 14.5
<0.001 <0.001 <0.001 0.007 0.007 0.007 <0.001 0.007 0.029
Adjusted mean diff between cohortsa -3.1 -0.6 9.4 4.6 11.4 1.9 3.2
95% Confidence Interval -4.5 - -1.7 -0.8 - -0.3 3.4 – 26.2 1.1 – 19.8 3.4 – 38.1 0.3 – 13.4 0.7 – 14.0
<0.001 <0.001 <0.001 0.040 <0.001 0.537 0.116
0.8
0.3 – 2.0
0.67
0.7
0.3 – 2.1
0.580
1.4 4.9 -
0.3 – 7.1 0.3 – 81.3 -
0.718 0.318 0.020 0.020 0.020
1.1 10.6 -
0.2 – 7.8 0.6 – 186.5 -
0.926 0.106 -
P Value
P Value
All data presented as mean ± standard deviation or n (%). *Primary composite outcome includes fetal or neonatal demise, placental abruption, preeclampsia, preterm delivery, small-for-gestational-age. A ͣ djusted for race, chronic hypertension and alcohol use during pregnancy. ICU= intensive care unit
Cassie Craun Ferguson, MD Assistant Professor of Pediatrics Section of Emergency Medicine Medical College of Wisconsin
I completed my undergraduate degree in Psychology at UCLA, followed by medical school at MCW. I spent the next six years completing my Pediatric and Pediatric Emergency Medicine training at the University of Colorado before returning to MCW as faculty in 2010. Within the Department of Pediatrics, I serve as the Director of Quality Improvement for the Section of Emergency Medicine. Additionally, I am the Course Director for the Quality Improvement and Patient Safety Pathway within the medical school. My own quality improvement work is focused on improving the effectiveness of pediatric emergency care by decreasing unnecessary laboratory and imaging studies in the emergency department (ED). I also have a special interest in improving the ED care of pediatric patients with chronic illnesses.
â&#x20AC;&#x153;Reducing Unnecessary Imaging for Patients With Constipation in the Pediatric Emergency Departmentâ&#x20AC;? Pediatrics. 2017;140(1):e20162290. Children diagnosed with constipation in the emergency department may undergo an abdominal radiograph (AXR) as part of their diagnostic work-up despite studies that suggest an AXR in a patient suspected of being constipated is unnecessary and potentially misleading. We aimed to decrease the percentage of patients between 6 months and 18 years diagnosed with constipation that undergo an AXR in our ED from 60% to 20% over 12 months. Process analysis was done using a cause-and-effect diagram. 4 plan-do-study-act cycles were completed over 9 months. Interventions included education, sharing best practices, metric reporting, and academic detailing. Rational subgrouping and stratification on SPC charts were used to target the interventions. We observed a significant and sustained decrease from mean rate of 62% to 24% in the utilization of AXRs in the ED for patients with constipation. Figure 1. Annotated P chart displaying AXR rate by month during pre-intervention, intervention, and postintervention periods (Nov. 2012 â&#x20AC;&#x201C; Oct. 2016).
“Intracranial Subependymoma: A SEER Analysis 2004-2013” World Neurosurgery. 2017;101:599-605. Subependymomas are rare, slow-growing, benign tumors. We explored the SEER database to evaluate prognostic and treatment factors associated with intracranial subependymoma. Four hundred sixty-six cases were identified in the database. The overall incidence of intracranial subependymoma is 0.055 per 100,000 person-years (95% confidence interval, 0.05-0.06). Through multivariate analysis, age <40 years (hazard ratio [HR], 0.21; P = 0.03), female sex (HR, 0.34; P = 0.03), location within ventricles or near brainstem (HR, 0.49; P = 0.04), and occurrence of surgery (HR, 0.50; P = 0.02) were significant independent positive prognostic factors. Receipt of radiation did not show a significant relationship.
Ha Son Nguyen, MD Resident Physician Department of Neurosurgery
Kai Yang, MD Resident (PGY 4) Department of Plastic Surgery
“Increasing Calcium Level Limits Schwann Cell Numbers In Vitro following Peripheral Nerve Injury” Journal of Reconstructive Microsurgery. 2017;33(6):435-440. After peripheral nerve injury, there is an increase in calcium concentration in the injured nerves. The increase in calcium concentration correlated well with degree of nerve injury and that local infusion of calcitonin has a beneficial effect on nerve recovery. Meanwhile, Schwann cells play a pivotal role in regeneration and recovery. Thus, we aimed to examine cultured Schwann cell survivals in various concentrations of calcium-containing growth media and the effect of adding calcitonin. We found that in mediums with increasing calcium concentration, there was a significant decrease in the number of Schwann cells. However, the addition of calcitonin has significant cytoprotective effects.
“Assessment of Treatment Response during Chemoradiation Therapy for Pancreatic Cancer Based on Quantitative Radiomic Analysis of Daily CTs: An Exploratory Study” PLOS One. 2017;12(6):e0178961. CT radiomic features were quantitatively analyzed for pancreatic cancer using longitudinal diagnostic-quality CT, acquired daily with a CT-on-rails during routine CT-guided radiation therapy. Significant changes in various CT radiomic features of the pancreas were observed during the course of radiation delivery. In cases of good response, patients tend to have large reductions in mean CT number and skewness, along with large increases in standard deviation and kurtosis. These changes may be potentially used for early assessment of treatment response and stratification for therapeutic intensification.
Xiaojian Chen Medical Physicist Department of Radiation Oncology
Selvi Kunnimalaiyaan, MS Research Technologist Department of Surgery Division of Surgical Oncology
“Suberoylanilide Hydroxamic Acid, a Histone Deacetylase Inhibitor, Alters Multiple Signaling Pathways in Hepatocellular Carcinoma Cell Lines” American Journal of Surgery. 2017;213(4):645-651. Among the histone deacetylase inhibitors, suberoylanilide hydroxamic acid (SAHAVorinostat) has demonstrated clinical efficacy at tolerated doses but trials have stalled due to limited effectiveness as a single agent therapy. In this study, we examined why SAHA showed limited effectiveness. Our study in hepatocellular carcinoma (HCC), demonstrated that SAHA inhibits Notch, AKT, and Raf-1 oncogenic signaling pathways but activates STAT3 pathway. We believe that activation of STAT3 pathway may lead to cancer cell progression and thus reducing SAHA’s efficacy. These findings allow future translational work focusing on combination therapy, importantly of SAHA and STAT3 inhibitor in cancers including HCC.
8701 W Watertown Plank Road Milwaukee, WI 53226-0509 mcw.edu/office-of-research.htm