Su-jin Rhee

An efficient and effective ensemble of support vector machines for anti-diabetic drug failure prediction

We propose a method called E3-SVM, efficient and effective ensemble of SVMs.E3-SVM excludes superfluous data points when constructing an SVM ensemble.E3-SVM was applied to the drug failure prediction problem for type 2 diabetes.We confirmed the suitability of SVM with an accuracy of about 80%. The treatment of patients with type 2 diabetes is mostly based on drug therapies, aiming at managing glucose levels appropriately. As the number of patients with type 2 diabetes continually increases worldwide, predicting drug treatment failure becomes an important issue. Support vector machine (SVM) can be a good method for the anti-diabetic drug failure prediction problem; however, it is difficult to train SVM on large-scale medical datasets directly because of its high training time complexity O ( N 3 ) . To address the limitation, we propose an efficient and effective ensemble of SVMs, called E3-SVM. The proposed method excludes superfluous data points when constructing an SVM ensemble, thereby yielding a better classification performance. The proposed method consists of two phases. The first phase is to select the data points that are likely to be the support vectors by applying data selection methods. The second phase is to construct an SVM ensemble using the selected data points. We demonstrated the efficiency and effectiveness of the proposed method using the real-world dataset of the anti-diabetic drug failure prediction problem for type 2 diabetes. Experimental results show that the proposed method requires less training time to achieve comparable success, compared to the conventional SVM ensembles. Moreover, the proposed method obtains more reliable prediction results for each independent run of constructing an ensemble. In conclusion, firstly, the proposed method provides an efficient and effective way to use SVM for large-scale datasets. Secondly, we confirmed the suitability of SVM for the anti-diabetic drug failure prediction problem with an accuracy of about 80%.

Pre- and post-operative comorbidities in idiopathic carpal tunnel syndrome: cervical arthritis, basal joint arthritis of the thumb, and trigger digit

We retrospectively reviewed 633 hands in 362 patients who had idiopathic carpal tunnel syndrome and underwent carpal tunnel release between 1999 and 2009. Electrophysiological studies and simple radiographs of the wrist, cervical spine, and basal joint of the thumb were routinely checked, and patients were also assessed for the presence of trigger digit or de Quervain’s disease before and after surgery. Among 362 patients, cervical arthritis was found in 253 patients (70%), and C5-C6 arthritis was the most common site. Basal joint arthritis of the thumb was observed in 216 (34%) of the 633 hands. Trigger digit or de Quervain’s disease was observed in 85 of the 633 hands (13%) before surgery, and developed in 67 hands (11%) after surgery. Cervical arthritis, basal joint arthritis, and trigger digit commonly coexist with idiopathic carpal tunnel syndrome. Patient education about these disorders is very important when they coexist with idiopathic carpal tunnel syndrome.

Factors affecting late displacement following volar locking plate fixation for distal radial fractures in elderly female patients

The purpose of this study was to evaluate the risk of late displacement after the treatment of distal radial fractures with a locking volar plate, and to investigate the clinical and radiological factors that might correlate with re-displacement. From March 2007 to October 2009, 120 of an original cohort of 132 female patients with unstable fractures of the distal radius were treated with a volar locking plate, and were studied over a follow-up period of six months. In the immediate post-operative and final follow-up radiographs, late displacement was evaluated as judged by ulnar variance, radial inclination, and dorsal angulation. We also analysed the correlation of a variety of clinical and radiological factors with re-displacement. Ulnar variance was significantly overcorrected (p < 0.001) while radial inclination and dorsal angulation were undercorrected when compared statistically (p < 0.001) with the unaffected side in the immediate post-operative stage. During follow-up, radial shortening and dorsal angulation progressed statistically, but none had a value beyond the acceptable range. Bone mineral density measured at the proximal femur and the position of the screws in the subchondral region, correlated with slight progressive radial shortening, which was not clinically relevant. Volar locking plating of distal radial fractures is a reliable form of treatment without substantial late displacement.

Pharmacokinetic and pharmacodynamic interactions between metformin and a novel dipeptidyl peptidase-4 inhibitor, evogliptin, in healthy subjects

Evogliptin is a newly developed dipeptidyl peptidase-4 (DPP-4) inhibitor, which is expected to be combined with metformin for treating type 2 diabetes mellitus. We investigated the potential pharmacokinetic and pharmacodynamic interactions between evogliptin and metformin. A randomized, open-label, multiple-dose, six-sequence, three-period crossover study was conducted in 36 healthy male subjects. All subjects received three treatments, separated by 7-day washout intervals: evogliptin, 5 mg od for 7 days (EVO); metformin IR, 1,000 mg bid for 7 days (MET); and the combination of EVO and MET (EVO + MET). After the last dose in a period, serial blood samples were collected for 24 hours for pharmacokinetic assessments. During steady state, serial blood samples were collected for 2 hours after an oral glucose tolerance test, and DPP-4, active glucagon-like peptide-1, glucose, glucagon, insulin, and C-peptide were measured to assess pharmacodynamic properties. EVO + MET and EVO showed similar steady state maximum concentration and area under the concentration–time curve at steady state values for evogliptin; the geometric mean ratios (90% confidence interval) were 1.06 (1.01–1.12) and 1.02 (0.99–1.06), respectively. EVO + MET slightly reduced steady state maximum concentration and area under the concentration–time curve at steady state values for metformin compared to MET, with geometric mean ratios (90% confidence interval) of 0.84 (0.79–0.89) and 0.94 (0.89–0.98), respectively. EVO + MET and EVO had similar DPP-4 inhibition efficacy, but EVO + MET increased active glucagon-like peptide-1 and reduced glucose to larger extents than either EVO or MET alone. Our results suggested that EVO+MET could provide therapeutic benefits without clinically significant pharmacokinetic interactions. Thus, the EVO + MET combination is a promising option for treating type 2 diabetes mellitus.

Pharmacokinetics of the evogliptin/metformin extended-release (5/1,000 mg) fixed-dose combination formulation compared to the corresponding loose combination, and food effect in healthy subjects

A new fixed-dose combination formulation of evogliptin 5 mg and metformin extended-release (XR) 1,000 mg (FDC_EVO5/MET1000) was developed to improve medication adherence for type 2 diabetes mellitus. The pharmacokinetics of FDC_EVO5/MET1000 was compared to the corresponding loose combination in a randomized, open-label, crossover study in 36 healthy male subjects (Part 1), and the food effect on FDC_EVO5/MET1000 was assessed (under fasted or fed conditions) in a randomized, open-label, crossover study in 28 healthy male subjects (Part 2). Serial blood samples for pharmacokinetic analysis were collected up to 72 hours, and pharmacokinetic parameters of evogliptin and metformin were calculated using non-compartmental methods. The geometric mean ratios (fixed-dose combination to loose combination) and 90% confidence intervals of pharmacokinetic parameters for evogliptin and metformin were all within 0.800–1.250, suggesting bioequivalent pharmacokinetic. After a single oral dose of FDC_EVO5/MET1000, food did not significantly affect evogliptin pharmacokinetic while systemic exposure of metformin was increased about 47.5% under the fed condition, which is consistent with the already established food effect on metformin XR. FDC_EVO5/MET1000 was generally well tolerated without any drug-related serious adverse events. In conclusion, FDC_EVO5/MET1000 can be substituted for the loose combination of FDC_EVO5/MET1000, providing better compliance with convenient administration.

Characteristic radiographic features of the central ray in Apert syndrome

We reviewed retrospectively seven patients with Apert acrosyndactyly and measured the size of the capitate ossification centre relative to that of the hamate and determined the relative position of the middle finger metacarpal relative to the ring finger metacarpal. We then compared those parameters in 197 normal children. In all patients, the middle finger metacarpal bone had migrated proximally relative to the ring finger metacarpal and the size of the capitate ossification centre was smaller than that of the hamate. After surgical release of the middle finger, relative proximal migration of the middle finger metacarpal was partially relieved and catch-up growth of the capitate was observed within several months. As fusion of the distal phalanges creates a diamond-shaped configuration, bone growth is markedly impaired in the middle finger ray. Therefore, early separation of the middle finger may be as important as early separation of the border digits.

Reliable prediction of anti-diabetic drug failure using a reject option

The medical care for patients with type 2 diabetes generally involves ingestion of oral hypoglycemic agents in order to lower their glucose level. When predicting the result of the medication using a classification approach, high prediction accuracy of the classifier is essential because of high misclassification costs. The application of a reject option to this approach supports more accurate prediction, allowing for human experts to examine when the classifier is unreliable to predict. In this paper, we propose a reject option framework based on heterogeneous ensemble learning through a two-phase fusion. The first phase is to calculate confidence scores, which are used to determine whether to predict, and the second phase is to derive final prediction results by fusing the outputs from multiple heterogeneous classifiers. We confirm the effectiveness of the proposed method to the anti-diabetic drug failure prediction problem through experiments on actual electronic medical records data of type 2 diabetes. The proposed method yields a better trade-off between accuracy and rejection than other reject options with statistical significance. A lower prediction error is obtained for the same degree of rejection. We obtained desirable accuracy for the anti-diabetic drug failure problem by applying the proposed reject option, which allows using the classification approach in practice. The accurate prediction of drug failure at the moment of prescription can assist clinical decisions for patients. In addition, in-depth analysis can be considered for those prescriptions that are predicted as failure or rejected.

Population pharmacokinetics and dose-response relationship of levetiracetam in adult patients with epilepsy

Levetiracetam (LEV) is commonly used as a mono- or adjunctive therapy for treating patients with partial and generalized epilepsy. This study aimed to develop a population pharmacokinetic (PK) model of LEV, based on sparse data, and to explore LEV efficacy relative to its PK properties in patients with epilepsy. We included 483 LEV concentrations from 425 patients with epilepsy that received multiple oral LEV doses. We performed a population PK analysis, implemented in NONMEM (version 7.2). In addition, we explored the relationships between seizure control and PK variables (i.e., LEV dose, trough concentration, and the number of concomitant anti-epileptic drugs). LEV concentration-time profiles were adequately described with a one-compartment, open linear model, with first-order absorption, and additive residual error. The typical population estimates of the apparent clearance (CL/F) and the volume of distribution (V/F) were 3.9L/h and 65.3L, respectively. Body weight was a significant covariate for CL/F and V/F; the estimated glomerular filtration rate only significantly affected CL/F; and concomitant intake of other anti-epileptic drugs did not significantly affect either parameter. A cumulative percentage analysis revealed that over 95% of patients that remained seizure-free received LEV doses of 2000mg/day or lower. LEV trough concentrations were not significantly different between seizure-free and seizure groups, for each LEV dose. In conclusion, we successfully developed a population PK model of LEV, which enabled investigation of LEV efficacy, relative to its PK properties. The findings in this study can be utilized to optimize LEV dosing regimens in clinical practice.

Pediatric patients undergoing hematopoietic stem cell transplantation can greatly benefit from a novel once‐daily intravenous busulfan dosing nomogram

Busulfan, a bifunctional alkylating agent, has been used as a conditioning regimen prior to allogeneic hematopoietic stem cell transplantation (HSCT). The aim of this study was to derive a novel once‐daily intravenous (IV) busulfan dosing nomogram for pediatric patients undergoing HSCT using a population pharmacokinetic (PK) model. A population PK analysis was performed using 2183 busulfan concentrations in 137 pediatric patients (age: 0.6‐22.2 years), who received IV busulfan once‐daily for 4 days before undergoing HSCT. Based on the final population PK model, an optimal once‐daily IV busulfan dosing nomogram was derived. The percentage of simulated patients achieving the daily target area under the concentration‐time curve (AUC) by the new nomogram was compared with that by other busulfan dosing regimens including the FDA regimen, the EMA regimen, and the empirical once‐daily regimen without therapeutic drug monitoring (TDM). A one‐compartment open linear PK model incorporating patients body surface area, age, dosing day, and aspartate aminotransferase as a significant covariate adequately described the concentration–time profiles of busulfan. An optimal dosing nomogram based on the PK model performed significantly better than the other dosing regimens, resulting in >60% of patients achieving the target AUC while the percentage of patients exceeding the toxic AUC level was kept <25% during the entire treatment period. A novel once‐daily busulfan dosing nomogram for pediatric patients undergoing HSCT is useful for clinicians, particularly in a setting where TDM service is not readily available or to optimize the dose on day 1.

Physiologically Based Pharmacokinetic Modeling of Fimasartan, Amlodipine, and Hydrochlorothiazide for the Investigation of Drug–Drug Interaction Potentials

PurposeTo build a physiologically based pharmacokinetic (PBPK) model for fimasartan, amlodipine, and hydrochlorothiazide, and to investigate the drug–drug interaction (DDI) potentials.MethodsThe PBPK model of each drug was developed using Simcyp software (Version 15.0), based on the information obtained from literature sources and in vitro studies. The predictive performance of the model was assessed by comparing the predicted PK profiles and parameters with the observed data collected from healthy subjects after multiple oral doses of fimasartan, amlodipine, and hydrochlorothiazide. The DDI potentials after co-administration of three drugs were simulated using the final model.ResultsThe predicted-to-observed ratios of all the pharmacokinetic parameters met the acceptance criterion. The PBPK model predicted no significant DDI when fimasartan was co-administered with amlodipine or hydrochlorothiazide, which is consistent with the observed clinical data. In the simulation of DDI at steady-state after co-administration of three drugs, the model predicted that fimasartan exposure would be increased by ~24.5%, while no changes were expected for the exposures of amlodipine and hydrochlorothiazide.ConclusionsThe developed PBPK model adequately predicted the pharmacokinetics of fimasartan, amlodipine, and hydrochlorothiazide, suggesting that the model can be used to further investigate the DDI potential of each drug.