Development of a Nurse Turnover Prediction Model in Korea Using Machine Learning.

Nurse turnover is a critical issue in Korea, as it affects the quality of patient care and increases the financial burden on healthcare systems. To address this problem, this study aimed to develop and evaluate a machine learning-based prediction model for nurse turnover in Korea and analyze factors influencing nurse turnover. The study was conducted in two phases: building the prediction model and evaluating its performance. Three models, namely, decision tree, logistic regression, and random forest were evaluated and compared to build the nurse turnover prediction model. The importance of turnover decision factors was also analyzed. The random forest model showed the highest accuracy of 0.97. The accuracy of turnover prediction within one year was improved to 98.9% with the optimized random forest. Salary was the most important decision factor for nurse turnover. The nurse turnover prediction model developed in this study can efficiently predict nurse turnover in Korea with minimal personnel and cost through machine learning. The model can effectively manage nurse turnover in a cost-effective manner if utilized in hospitals or nursing units.

Hepatocellular carcinoma pathologic grade prediction using radiomics and machine learning models of gadoxetic acid-enhanced MRI: a two-center study.

PURPOSE: To develop a radiomics-based hepatocellular carcinoma (HCC) grade classifier model based on data from gadoxetic acid-enhanced MRI. METHODS: This retrospective study included 137 patients who underwent hepatectomy for a single HCC and gadoxetic acid-enhanced MRI within 60 days before surgery. HCC grade was categorized as low or high (modified Edmondson-Steiner grade I-II vs. III-IV). We used the hepatobiliary phase (HBP), portal venous phase, T2-weighted image(T2WI), and T1-weighted image(T1WI). From the volume of interest in HCC, 833 radiomic features were extracted. Radiomic and clinical features were selected using a random forest regressor, and the classification model was trained and validated using a random forest classifier and tenfold stratified cross-validation. Eight models were developed using the radiomic features alone or by combining the radiomic and clinical features. Models were validated with internal enrolled data (internal validation) and a dataset (28 patients) at a separate institution (external validation). The area under the curve (AUC) of the validation results was compared using the DeLong test. RESULTS: In internal and external validation, the HBP radiomics-only model showed the highest AUC (internal 0.80 ± 0.09, external 0.70 ± 0.09). In external validation, all models showed lower AUC than those for internal validation, while the T2WI and T1WI models failed to predict the HCC grade (AUC 0.30-0.58) in contrast to the internal validation results (AUC 0.67-0.78). CONCLUSION: The radiomics-based machine learning model from gadoxetic acid-enhanced liver MRI could distinguish between low- and high-grade HCCs. The radiomics-only HBP model showed the best AUC among the eight models, good performance in internal validation, and fair performance in external validation.

Ectopic lesions in the abdomen and pelvis: a multimodality pictorial review.

Various ectopic lesions occur in the abdomen and pelvis and affect multiple organs including liver, gallbladder, pancreas, spleen, and organs of the genitourinary system. Ectopic organs may be present outside their normal positions, or ectopic tissues may develop while the original organ exists in its normal position. Both benign and malignant lesions can occur in ectopic organs and tissues. Owing to their unusual location, they can often be misdiagnosed as other lesions or even malignant lesions, such as metastasis or seeding. This multimodality pictorial review provides various cases of ectopic lesions in the abdomen and pelvis, which will help narrow the differential diagnosis and guide clinical decision-making.

Ultrasonographic index for the diagnosis of non-alcoholic steatohepatitis in patients with non-alcoholic fatty liver disease.

Background: Liver biopsy is a gold standard for the diagnosis of non-alcoholic steatohepatitis (NASH), but has several disadvantages including invasiveness, high cost, and sampling error. Ultrasonography (US) is a noninvasive imaging modality widely used in non-alcoholic fatty liver disease (NAFLD) patients. This study aimed: (I) to assess the feasibility of US in the prediction of NASH and (II) to develop various US indices combining US parameters and laboratory data for the detection of NASH in NAFLD patients and to compare the diagnostic performance of them. Methods: Sixty patients who underwent liver biopsy, gray-scale US [hepatorenal index (HRI) and shear-wave elastography (SWE)], and Fibroscan [controlled attenuation parameter (CAP) and transient elastography (TE)] for the evaluation of NASH were included. Patients were classified according to the NAFLD Activity Score (NAS) into the NASH (NAS ≥5) and non-NASH (NAS <5) groups. The diagnostic performance of HRI, CAP, SWE, TE, and laboratory data for grading steatosis, lobular inflammation, ballooning degeneration, and fibrosis was evaluated. After the identification of laboratory data that were independently associated with NASH through univariable and multivariable logistic regression analyses, various US indices were developed by combining US parameters with or without these laboratory data. The diagnostic performance of the US indices was assessed with obtaining area under the curve (AUC) and compared using DeLong test. Results: Twenty-five NASH and 35 non-NASH patients were included. The mean AUCs for grading steatosis were 0.871 using HRI and 0.583 using CAP. The mean AUCs for grading fibrosis and ballooning degeneration were 0.777 and 0.729 using SWE and 0.830 and 0.708 using TE, respectively. Aspartate aminotransferase (AST) was the only significant laboratory data associated with NASH (OR, 1.019; P=0.032). Using AST, the mean AUCs for grading lobular inflammation and ballooning degeneration were 0.712 and 0.775, respectively. Among various US indices, the index consisting of gray-scale US parameters (SWE and HRI) and AST showed the best diagnostic performance for the detection of NASH in NAFLD patients (AUC =0.806). Conclusions: The index combining gray-scale US parameters and AST is useful for the detection of NASH and may be used to exclude the need for liver biopsy in NAFLD patients.

MRI Diagnosis for Post-Traumatic Osteomyelitis of Extremities Using Conventional Metal-Artifact Reducing Protocols: Revisited.

RATIONALE AND OBJECTIVES: To assess the diagnostic power of MR imaging features for post-traumatic osteomyelitis (PTOM) of the extremities with metal implants using a 3.0-T machine with conventional metal-artifact reducing (MAR) protocols. MATERIALS AND METHODS: Between December 2012 and September 2016, 261 consecutive patients underwent MRI for clinical suspicion of PTOM in the appendicular skeleton at our hospital. A total of 72 enrolled patients with metal implants were divided into two groups, one with surgical, histologic, or microbiologic proven to be PTOM, and the other proven not to be PTOM. Their MR images were reviewed by two musculoskeletal radiologists who qualitatively analyzed various findings regarding PTOM, and concluded diagnosis of PTOM or without PTOM for each patient. All MR images were obtained using conventional MAR protocols. RESULTS: The sensitivity, specificity, positive and negative predictive values, and accuracy of MR diagnosis by two observers were 55%/38%, 81%/93%, 67%/40%, 73%/78%, and 71%/71%, respectively. Among the findings, T1 low signal intensity, medullary location, confluent pattern, typical features (concurrence of T1 low signal intensity, medullary, and confluent patterns), heterogeneous or no enhancement, and cortical destruction were statistically significant by univariate analysis (p<0.05). Among these features, only medullary involvement was significant by multivariate analysis (p = 0.007). Medullary involvement and no enhancement pattern were significant by step-wise discrimination analysis. Interobserver correlation was moderate with a weighted kappa value of 0.512 (confidence interval: 0.3-0.723). CONCLUSION: The overall accuracy for diagnosis of PTOM of the extremities using 3.0-T MRI with conventional metal-artifact reduction parameters was 71%. The strongest diagnostic MR imaging feature was medullary involvement of T1 low signal intensity. PTOM of the extremities can be accurately diagnosed with a 3.0-T MR machine using conventional MAR protocols.

Washout appearance in Gd-EOB-DTPA-enhanced MR imaging: A differentiating feature between hepatocellular carcinoma with paradoxical uptake on the hepatobiliary phase and focal nodular hyperplasia-like nodules.

PURPOSE: To identify the most reliable imaging features for differentiating hepatocellular carcinoma with paradoxical uptake on the hepatobiliary phase (HCCpara ) from focal nodular hyperplasia (FNH)-like nodules using Gd-EOB-DTPA-enhanced MRI. MATERIALS AND METHODS: This was a retrospective study. Twenty patients with HCCpara and 21 patients with FNH-like nodules were included. The following MRI features were evaluated using 3.0 Tesla unit by two radiologists: signal intensity (SI) on T1-, T2-, and diffusion-weighted imaging (DWI), arterial enhancement pattern, washout appearance on the portal venous phase (PVP) and/or transitional phase (TP), uptake pattern on the hepatobiliary phase (HBP), "T2 scar," "EOB scar," and chemical shift on in- and out-of-phase images. Multivariate logistic regression analysis was performed to assess MRI features for prediction of HCCpara . RESULTS: Compared with FNH-like nodules, HCCpara had significantly more frequent heterogeneous T1 SI (P < 0.0001), T2 hyperintensity (P = 0.032), heterogeneous arterial enhancement (P < 0.0001), washout appearance on the PVP and/or TP (P < 0.0001), heterogeneous uptake on the HBP (P < 0.0001), absence of "EOB scar" (P < 0.0001), and hyperintensity on DWI (P = 0.004). Multivariate logistic regression analysis revealed washout appearance as the only independent imaging feature associated with HCCpara (odds ratio, 7.019; P = 0.042). Washout appearance also showed the best diagnostic performance with a sensitivity of 90% and a specificity of 100%. CONCLUSION: Washout appearance on the PVP and/or TP is the most reliable imaging feature for differentiating HCCpara from FNH-like nodules. LEVEL OF EVIDENCE: 3 J. MAGN. RESON. IMAGING 2017;45:1599-1608.

Epigenetic regulation in the brain after spinal cord injury : a comparative study.

OBJECTIVE: After spinal cord injury (SCI), functional and structural reorganization occurs at multiple levels of brain including motor cortex. However, the underlying mechanism still remains unclear. The current study was performed to investigate the alterations in the expression of the main regulators of neuronal development, survival and death, in the brain following thoracic contusive SCI in a mouse model. METHODS: Eight-week-old female imprinting control region mice (n=60; 30-35 g) were used in this study. We analyzed the expression levels of regulators such as brain-derived neurotrophic factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), nerve growth factor (NGF) and histone deacetylase (HDAC) 1 in the brain following thoracic contusive SCI. RESULTS: The expression of BDNF levels were elevated significantly compared with control group at 2 weeks after injury (p<0.05). The expression of NGF levels were elevated at 2, 4 weeks compared with control group, but these difference were not significant (p>0.05). The GDNF levels were elevated at 2 week compared with control group, but these differences were not significant (p>0.05). The difference of HDAC1 levels were not significant at 2, 4 and 8 weeks compared with control group (p>0.05). CONCLUSION: These results demonstrate that the upregulation of BDNF may play on important role in brain reorganization after SCI.

Electrophysiological and behavioral changes by phosphodiesterase 4 inhibitor in a rat model of alcoholic neuropathy.

OBJECTIVE: Alcoholic neuropathy is characterized by allodynia (a discomfort evoked by normally innocuous stimuli), hyperalgesia (an exaggerated pain in response to painful stimuli) and spontaneous burning pain. The aim of the present study is to investigate the effect of rolipram, a phosphodiesterase 4 inhibitor, against alcohol-induced neuropathy in rats. METHODS: Allodynia was induced by administering 35% v/v ethanol (10 g/kg; oral gavage) to Spraue-Dawley rats for 8 weeks. Rolipram and saline (vehicle) were administered intraperitoneally. Mechanical allodynia was measured by using von Frey filaments. Somatosensory evoked potential (SEP) was proposed as complementary measure to assess the integrity of nerve pathway. RESULTS: The ethanol-induced mechanical allodynia began to manifest from 3 week, and then peaked within 1 week. Beginning from 3 week, latency significantly started to increased in control group. In rolipram treated rats, the shorter latency was sustained until 8 weeks (p<0.05). The mechanical allodynia, which began to manifest on the 3 weeks, intraperitoneal injections of rolipram sustained statistical difference until 8 weeks, the final week of the study (p<0.05). CONCLUSION: This study suggests that rolipram might alleviate mechanical allodynia induced by alcohol in rats, which clearly has clinical implication.

The effect of phosphodiesterase-4-specific inhibitor in the rat model of spinal nerve ligation.

OBJECTIVE: Peripheral neuropathy is characterized by hyperalgesia, spontaneous burning pain, and allodynia. The purpose of this study was to investigate the effect of rolipram, a phosphodiesterase-4-specific inhibitor, in a segmental spinal nerve ligation model in rats. METHODS: Both the L5 and L6 spinal nerves of the left side of the rats were ligated. Phosphodiesterase-4 inhibitor (rolipram) and saline (vehicle) were administered intraperitoneally. We measured mechanical allodynia using von Frey filaments and a nerve conduction study. RESULTS: The mechanical allodynia, which began to manifest on the first day, peaked within 2 days. Multiple intraperitoneal injections of rolipram ameliorated the mechanical allodynia. Furthermore, an intraperitoneal administration of rolipram improved the development of pain behavior and nerve conduction velocity. CONCLUSION: This study suggests that the phosphodiesterase-4 inhibitor, rolipram, alleviates mechanical allodynia induced by segmental spinal nerve ligation in rats. This finding may have clinical implications.

The effects of heat and massage application on autonomic nervous system.

PURPOSE: The objective of this study is to evaluate the effects of heat and massage application on autonomic nervous system. MATERIALS AND METHODS: One hundred thirty-nine subjects volunteered and completed this study. Heat and massage was daily applied for 40 minutes, 5 days a week for 2 weeks. Primary-dependent measures included heart rate variability, sympathetic skin response, and serum cortisol and norepinephrine levels. RESULTS: Serum cortisol levels were significantly decreased at 2 weeks compared to baseline (p=0.003). Plasma norepinephrine levels at 4 weeks were significantly decreased compared to baseline (p=0.010). Heart rate, using the power spectra, increased significantly after 2 weeks compared to baseline. Of autonomic nerve conduction measures, latency was significantly increased at 2 and 4 weeks compared to baseline (p=0.023, 0.012), and amplitude was significantly decreased at 4 weeks compared to baseline (p=0.008). There were no serious adverse events such as burns or other major complications. CONCLUSION: The results of this study suggest that heat and massage applications provide relaxation to the autonomic nervous system without serious adverse events.