Category-free Net Reclassification Index Research Articles

Predicted Heart Mass in Obese Heart Transplant Donors and Recipients: An Analysis of the ISHLT Registry

<h3>Purpose</h3> Predicted heart mass (PHM) was neither derived nor validated in an obese population. Our objectives were to determine survival according to size mismatch defined by actual PHM and recalculated PHM using ideal body weight (IBW) in obese recipients or recipients of obese donors, and to evaluate whether size mismatch using IBW-adjusted PHM improves risk assessment. <h3>Methods</h3> We included HT recipients (≥18 y) with BMI ≥30 kg/m² or recipients of donors with BMI ≥30 kg/m² from the ISHLT registry between 2008-2017. IBW was calculated using Devine's formula. Donor-recipient PHM was evaluated as a continuous variable and according to established groups: <-30%, -30 to -20%, -20 to 20%, 20 to 30% and >30%. We used Kaplan-Meier survival and multivariable Cox PH analyses to evaluate post-HT mortality by actual/IBW-adjusted PHM groups. We compared the 2 models using net reclassification index (NRI). <h3>Results</h3> We included 10,817 HT recipients, age 55 (IQR 46-62) yrs, 23% female, BMI 31 kg/m² (28-33) with follow up of 3.1 (1.2-6.0) yrs. Included donors were 34 (24-44) yrs, 31% female, and BMI 31 kg/m² (26-34). 692 (6%) and 335 (3%) recipient-donor pairs were reclassified as undersized by 20-30% and >30% respectively using IBW-adjusted PHM. 462 (4%) and 431 (4%) recipient-donor pairs were reclassified as oversized by 20-30% and >30% respectively when using IBW-adjusted PHM. Actual PHM was associated with lower 1-year survival (<i>p</i><0.001); however, differences in 1-year survival between IBW-adjusted PHM groups were not discernible (<i>p</i>=0.25). Undersizing using actual PHM was associated with higher 1-year mortality (<i>p</i><0.001), not seen with IBW-adjusted PHM (p=0.15 Fig). Models based on IBW-adjusted PHM had worse risk quantification for post-HT mortality than using actual PHM (category-free NRI -0.078, 95% CI -0.158, 0.006). <h3>Conclusion</h3> Calculation of PHM using IBW was not associated with post-HT survival. Actual PHM can be used for size matching when assessing mortality risk in recipients of obese donors or obese recipients.

Plasma soluble suppression of tumorigenicity 2 and depression after acute ischemic stroke.

Soluble suppression of tumorigenicity 2 (sST2) might be related to stroke and depression, but the association of sST2 with poststroke depression (PSD) is unclear. The study aimed to prospectively assess the association between plasma sST2 levels and PSD. A total of 635 acute ischemic stroke patients with sST2 measurements from the China Antihypertensive Trial in Acute Ischemic Stroke were included in this analysis. We used the 24-item Hamilton Rating Scale for Depression to assess depression at 3months, and PSD was defined as a score of ≥8. Logistic regression analysis was performed to estimate the risk of PSD associated with sST2, and net reclassification index (NRI) and integrated discrimination improvement (IDI) were calculated to assess the predictive value of sST2. Two hundred fifty (39.4%) patients developed depression at 3months after ischemic stroke. Patients with PSD had higher sST2 levels than patients without PSD (172.7 vs. 153.8pg/ml; p=0.003). After adjustment for age, sex, education, National Institutes of Health Stroke Scale score, and other covariates, the odds ratio for the highest quartile of sST2 compared with the lowest quartile was 1.84 (95% confidence interval, 1.10-3.08) for PSD. Adding sST2 to a conventional model notably improved risk prediction for PSD (category-free NRI=19.34%, 95% confidence interval=4.39%-34.28%, p=0.017; IDI=1.20%, 95% confidence interval=0.25%-2.15%, p=0.014). Increased plasma sST2 levels in the acute phase of ischemic stroke were significantly associated with the increased risk of PSD, independently of conventional risk factors.

A Panel of 6 Biomarkers Significantly Improves the Prediction of Type 2 Diabetes in the MONICA/KORA Study Population.

ContextImproved strategies to identify persons at high risk of type 2 diabetes are important to target costly preventive efforts to those who will benefit most.ObjectiveThis work aimed to assess whether novel biomarkers improve the prediction of type 2 diabetes beyond noninvasive standard clinical risk factors alone or in combination with glycated hemoglobin A1c (HbA1c).MethodsWe used a population-based case-cohort study for discovery (689 incident cases and 1850 noncases) and an independent cohort study (262 incident cases, 2549 noncases) for validation. An L1-penalized (lasso) Cox model was used to select the most predictive set among 47 serum biomarkers from multiple etiological pathways. All variables available from the noninvasive German Diabetes Risk Score (GDRSadapted) were forced into the models. The C index and the category-free net reclassification index (cfNRI) were used to evaluate the predictive performance of the selected biomarkers beyond the GDRSadapted model (plus HbA1c).ResultsInterleukin-1 receptor antagonist, insulin-like growth factor binding protein 2, soluble E-selectin, decorin, adiponectin, and high-density lipoprotein cholesterol were selected as the most relevant biomarkers. The simultaneous addition of these 6 biomarkers significantly improved the predictive performance both in the discovery (C index [95% CI], 0.053 [0.039-0.066]; cfNRI [95% CI], 67.4% [57.3%-79.5%]) and the validation study (0.034 [0.019-0.053]; 48.4% [35.6%-60.8%]). Significant improvements by these biomarkers were also seen on top of the GDRSadapted model plus HbA1c in both studies.ConclusionThe addition of 6 biomarkers significantly improved the prediction of type 2 diabetes when added to a noninvasive clinical model or to a clinical model plus HbA1c.

Usefulness of atherogenic index of plasma for estimating reduced eGFR risk: insights from the national health and nutrition examination survey

ABSTRACT Aims Previous studies have identified Atherogenic index of plasma (AIP) as a simple measure of atherosclerosis. Because atherosclerosis plays a role in the development of renal damage, our study aims to evaluate the effect of AIP on the risk of reduced eGFR and assess its usefulness to refine the risk stratification of reduced estimated glomerular filtration rate (eGFR). Methods Our study included 15,836 participants from the National Health and Nutritional Survey (NHANES) 2009–2016. Association was investigated by logistic regression. AIP was calculated as log (triglycerides/high-density lipoprotein cholesterol). Reduced eGFR was determined as eGFR < 60 ml/min per 1.73 m*2. Results The prevalence of reduced eGFR was 8.01%. In the full model, each SD increase of AIP leaded to 27.4% additional risk for reduced eGFR. After dividing AIP into quartiles, the fourth quartile had a 1.649 times risk than the first quartile. Moreover, smooth curve fitting suggested that the risk of reduced eGFR elevated linearly with the increase of AIP. Subgroup analysis demonstrated that the association between AIP and reduced eGFR was robust in sex, body mass index, hypertension, and diabetes subpopulation, but the association was significantly stronger in black race and people aged less than 50 years old. Additionally, AUC displayed an advancement when introducing AIP into established risk factors (0.875 cs. 0.897, P < 0.001), category-free net reclassification index (0.249, 95% CI: 0.192–0.306, P < 0.001) and integrated discrimination index (0.007, 95% CI: 0.004–0.009, P < 0.001) also suggested the improvement from AIP. Conclusion The present work suggested a linear association between AIP and reduced eGFR. Furthermore, the results showed that the association was stronger in black race and people aged less than 50 years old. Most importantly, our work implicated the usefulness of AIP to refine the risk stratification of reduced eGFR.

Evaluation of the usefulness of red blood cell distribution width in critically ill pediatric patients.

Red blood cell distribution width (RDW) is a component of routine complete blood count, which reflects variability in the size of circulating erythrocytes. Recently, there have been many reports about RDW as a strong prognostic marker in various disease conditions in the adult population. However, only a few studies have been performed in children. This study aimed to investigate the association between RDW and pediatric intensive care unit (PICU) mortality in critically ill children. This study includes 960 patients admitted to the PICU from November 2012 to May 2018. We evaluated the associations between RDW and clinical parameters including PICU mortality outcomes. The median age of the study population was 15.5 (interquartile range, 4.8–54.5) months. The mean RDW was 15.6% ± 3.3%. The overall PICU mortality was 8.8%. As we categorized patients into 3 groups with respect to RDW values (Group 1: ≤14.5%; Group 2: 14.5%–16.5%; and Group 3: >16.5%) and compared clinical parameters, the higher RDW groups (Groups 2 and 3) showed more use of vasoactive-inotropic drugs, mechanical ventilator support, higher severity scores, including pediatric risk of mortality III, pediatric sequential organ failure assessment, pediatric logistic organ dysfunction-2 (PELOD-2), and pediatric multiple organ dysfunction syndrome scores, and higher PICU mortality than the lower RDW group (Group 1) (P < .05). Based on multivariate logistic regression analysis adjusted for age and sex, higher RDW value (≥14.5%) was an independent risk factor of PICU mortality. Moreover, adding RDW improved the performance of the PELOD-2 score in predicting PICU mortality (category-free net reclassification index 0.357, 95% confidence interval 0.153–0.562, P = .001). In conclusion, higher RDW value was significantly associated with worse clinical parameters including PICU mortality. RDW was an independent risk factor of PICU mortality and the addition of RDW significantly improved the performance of PELOD-2 score in predicting PICU mortality. Thus, RDW could be a promising prognostic factor with advantages of simple and easy measurement in critically ill pediatric patients.

Endostatin as a novel prognostic biomarker in acute ischemic stroke

Background and aimsEndostatin is implicated in the atherosclerosis process and serves as a promising cardiovascular biomarker, while its clinical significance in ischemic stroke patients remains unclear. We aimed to examine the association between endostatin and mortality and disability after ischemic stroke. MethodsA total of 3463 acute ischemic stroke patients with measured plasma endostatin from the China Antihypertensive Trial in Acute Ischemic Stroke were included in this study. The primary outcome was death or severe disability (modified Rankin scale score of 4–6), and secondary outcomes included death and vascular events. ResultsAfter 3-month follow-up, 402 (11.61%) participants experienced severe disability or died. Compared with the lowest quartile of endostatin, odds ratios or hazard ratios (95% confidence intervals) for the highest quartile were 1.47 (1.04–2.09) for the primary outcome, and 2.36 (1.23–4.54) for death after adjustment for multiple covariates, including age, sex, admission NIH Stroke Scale score and systolic blood pressure. Each 1-SD higher log-transformed endostatin was associated with a 20% (6%–36%) increased risk for primary outcome. Adding plasma endostatin to the basic model constructed with conventional factors significantly improved risk stratification of primary outcome, as observed by the category-free net reclassification index of 20.5% (95% CI 10.1%–30.8%; p < 0.001) and integrated discrimination improvement of 0.3% (95% CI 0.01%–0.6%; p = 0.04). ConclusionsIncreased baseline plasma endostatin levels in acute ischemic stroke were associated with increased risk of mortality and severe disability at 3 months. Plasma endostatin may serve as an important prognostic marker for risk stratification in patients with ischemic stroke.

6097Aminoterminal proB-type natriuretic peptide: a key parameter to optimise therapeutic management of low-flow, low-gradient aortic stenosis

Abstract Background B-type natriuretic peptide (BNP) and aminoterminal-proBNP (NT-proBNP) are well established surrogates of LV function impairment. However, data are scarce regarding their prognostic value to risk-stratify patients with classical low-flow, low-gradient aortic stenosis (LFLG-AS, with low left ventricular [LV] ejection fraction). Methods The TOPAS study is a prospective observational cohort of 240 patients with aortic valve area &lt;0.6 cm2/m2, mean gradient&lt;40 mmHg and LVEF&lt;50%. True severe AS was adjudicated using flow independent grading schemes. Results BNP significantly predicted one-year (area under the receiver operating-characteristic curve [AUC]) 0.62±0.04, p=0.026) but not three-year mortality. After adjustment for the severity of AS, initial treatment (aortic valve replacement [AVR] vs. conservative management [ConsRx]), age, sex and the EuroSCORE (Model#1), BNP-ratio&gt;550 pg/ml had a trend to predict time to death (HR=2.14 [1.00–4.58], p=0.05). In contrast, NT-proBNP ratio significantly predicted both one and three-year mortality (AUC=0.67±0.04 and 0.66±0.05, both p=0.001), and independently predicted time to death (HR=1.39 per 1 unit of Log transformed NT-proBNP [1.11–1.74], p=0.004). In a head-to-head comparison (108 patients with both biomarkers), the AUCs to predict one and thre-year mortality were significantly higher with NT-proBNP versus BNP (p&lt;0.009). NT-proBNP but not BNP independently predicted mortality and significantly improved Model#1 (Likelihood ratio test Chi2=15.95, p&lt;0.001). The category-free net reclassification index of NT-proBNP was 0.71 (p=0.008) versus 0.38 (p=0.15) for BNP. Furthermore, there was a marked survival benefit associated with AVR in patients with NT-proBNP ≥1700 pg/ml (adjusted hazard ratio (aHR) associated to AVR vs conservative management=0.52 [0.31–0.85], p=0.009), while those&lt;1700 pg/ml had excellent one-year survival under ConsRx (only one death [4.5±4.4%] at one year as compared to 23 [37±6.2%] for ConsRx-NTproBNP&gt;1700, aHR=0.11 [0.01–0.83], p=0.033). The survival benefit associated with AVR interacted with NT-proBNP (p&lt;0.001) but not with true or pseudosevere AS (p=0.53 for interaction), suggesting that NT-proBNP might identify moderate AS patients but sufficiently severe valvulo-ventricular disease to justify AVR. Survival according to NT-proBNP and AVR Conclusion NT-proBNP appears to be an excellent biomarker for the clinical purpose of risk-stratifying classical LFLG-AS. A threshold of 1700 pg/ml i.e. close to the diagnostic threshold for heart failure in acute dyspnea, was a strong independent determinant of the survival benefit associated with aortic valve replacement. Our findings suggest that NT-proBNP should be preferred over BNP. Acknowledgement/Funding Canadian Institute of Health Research

Value of triglyceride-glucose index for the estimation of ischemic stroke risk: Insights from a general population

Background and aimsRecent studies have recognized triglyceride-glucose index (TyG) as a practical surrogate of insulin resistance. Previous studies have demonstrated that insulin resistance contributes to ischemic stroke via multiple mechanisms. Our study aimed to investigate the association between TyG and prevalent ischemic stroke, exploring the value of TyG to optimize the risk stratification of ischemic stroke. Methods and resultsThis cross-sectional study included 10,900 subjects (mean age: 59.95 years, 59.8% females) from rural areas of northeast China between September 2017 to May 2018. TyG was calculated as ln[fasting triglyceride (mg/dl) × fasting plasma glucose (mg/dl)/2]. The prevalence of ischemic stroke was 5.49%. After adjusting for all covariates, each SD increment of TyG caused 22.8% additional risk for ischemic stroke. When dividing TyG into quartiles, the top quartile had a 1.776 times risk for ischemic stroke against the bottom category. Furthermore, smoothing curve fitting demonstrated this association was linear in the whole range of TyG. Finally, AUC revealed an improvement when introducing TyG into clinical risk factors (0.746 vs 0.751, p = 0.029). Consistently, category-free net reclassification index (0.195, 95% CI: 0.112–0.277, P < 0.001) and integrated discrimination index (0.003, 95% CI: 0.001–0.004, P < 0.001) confirmed the improvement by TyG to stratify ischemic stroke risk. ConclusionThe prevent ischemic stroke correlated proportionally with the increment of TyG, implicating the linearity of TyG as an indicator of ischemic stroke. Our findings suggest the potential value of TyG to optimize the risk stratification of ischemic stroke in a general population.

Usefulness of Triglyceride-glucose Index for estimating Hyperuricemia risk: Insights from a general Population

ABSTRACTObjectives: Hyperuricemia is a metabolic abnormality that has cast an enormous burden on global healthcare. Previous studies have revealed the close association between insulin resistance and hyperuricemia. Therefore, monitoring insulin sensitivity may be a possible way to prevent hyperuricemia. Recent studies have demonstrated the usefulness of triglyceride-glucose index (TyG) as a simple surrogate of insulin resistance. Hence, our study aimed to explore the impact of TyG on hyperuricemia and its value to improve the risk stratification and prevention of hyperuricemia.Methods: This cross-sectional study included 6466 subjects (mean age: 59.57 years, 60.19% females) from northeast China between September 2017 to May 2018. TyG was determined as ln[fasting TG(mg/dL)×FPG(mg/dL)/2]. Hyperuricemia was defined as serum uric acid ≥ 357 μmol/L for females and ≥ 417 μmol/L for males.Results: The prevalence of hyperuricemia was 5.24%. In the full model, each SD increment of TyG caused a 12.528 μmol/L elevation of serum urate concentration and a 54.1% additional risk for hyperuricemia. When dividing TyG into quartiles, the top quartile had a 2.730 times risk for hyperuricemia than the bottom one. Moreover, smooth curve fitting demonstrated this association was linear. Additionally, subgroup analysis revealed the association was robust to several risk factors of hyperuricemia. Finally, AUC displayed an improvement when introducing TyG into clinical risk factors (0.751 vs 0.772, p < 0.001), category-free net reclassification index (0.304, 95% CI: 0.195–0.413, P < 0.001) and integrated discrimination index (0.009, 95% CI: 0.004–0.013, P < 0.001) also showed the improvement from TyG.Conclusion: Our work revealed the linear and robust association between TyG and hyperuricemia. Furthermore, our results suggest the importance of simultaneous glycemic and lipids control in the prevention of hyperuricemia. Most importantly, our findings implicate the value of TyG to optimize the risk stratification and prevention of hyperuricemia.

SAT-349 Expression of Cyclin B1, Transferrin Receptor and Fibronectin Correlates with the Prognosis of Patients with Adrenal Cortical Carcinoma

Adrenal cortical carcinoma (ACC) is a rare cancer with poor prognosis. The main treatment of ACC is surgery, and the role of systemic chemotherapy is limited. However, up to 74% of patients are known to experience recurrence after surgery, even following complete resection (1). Therefore, it is important to know the biomarkers which can predict the prognosis beyond clinical stages. Previous studies have reported several prognostic factors of ACC including Ki67 and markers revealed by multi-omics approaches (2-4). However, the proteomic approach for predicting prognosis has not been investigated. Therefore, we aimed to investigate the prognostic biomarkers for ACCs using proteomic approach. We used the reverse phase protein array data from The Cancer Proteome Atlas. Differentially expressed proteins in metastatic and non-metastatic ACCs were identified using Benjamini-Hochberg method. We validated the identified proteins using the multivariate Cox’s proportional hazard regression analysis for metastasis including age and clinical staging. Additive prognostic values were analyzed using C-index and category free Net reclassification index (cfNRI). Among 46 ACC patients, cyclin B1, transferrin receptor (TFRC) and fibronectin were significantly overexpressed in patients with distant metastasis. In the Cox regression model, during median follow-up of 3.1 years, high expression status of cyclin B1 and TFRC were significantly associated with high mortality (hazard ratios [HR] (95% confidence interval [CI]), 6.13 (1.02-36.7), and 6.59 (1.14-38.2), respectively). However, high expression status of fibronectin was not associated with mortality (HR [95% CI], 3.92 [0.75-20.4]). When we combined protein expression signatures, high cyclin B1/high TFRC, high cyclin B1/high fibronectin and high TFRC/high fibronectin were highly associated with mortality: HR (95% CI) 13.72 (1.89-99.66), HR 9.22 (1.34-63.55) and HR 18.59 (2.54-135.88), respectively. Compared to age and staging model, addition of cyclin B1, TFRC, fibronectin, or combination of these proteins models all improved the risk prediction by C-index (0.780 to 0.818~0.864) and cfNRI (0.741~1.027, all P<0.05). Among ACC patients, overexpression of various combinations of cyclin B1, TFRC and fibronectin proteins were associated with poor prognosis. Also, these combinatory biomarkers showed additive prognostic values beyond age and clinical staging, and combination of cyclin B1 and TFRC showed poorest prognosis and highest predictive value among them. Thus, the combination of these three proteins can be useful prognostic markers for ACC. Reference: (1) Freire DS et al., Clin Endocrinol (Oxf). 2013 Oct;79(4):468-75. (2) Beuschlein F et al., J Clin Endocrinol Metab. 2015 Mar;100(3):841-9. (3) Assié G et al., Nat Genet. 2014 Jun;46(6):607-12. (4) Zheng S et al., Cancer Med. 2016 Aug;5(8):2117-25.

Estimate of reduced glomerular filtration rate by triglyceride-glucose index: insights from a general Chinese population

ABSTRACTObjectives: Recent studies have identified triglyceride-glucose index (TyG) as a surrogate of insulin resistance. Since insulin resistance correlates with renal damage, our study aims to investigate the impact of TyG on the risk of reduced eGFR and explore its value to improve the risk stratification and prevention of reduced eGFR.Methods: This cross-sectional study included 6466 participants (mean age:59.57 years, 60.2% females) from rural areas of northeast China between September 2017 to May 2018. TyG was calculated as ln[fasting triglyceride (mg/dL)×fasting plasma glucose (mg/dL)/2]. Reduced eGFR was defined as eGFR<60ml/min per 1.73m2.Results: The prevalence of reduced eGFR was 2.94%. After full adjustment, each SD increase of TyG caused 42.6% additional risk for reduced eGFR. When dividing TyG into quartiles, the top quartile had a 1.934 times risk than the bottom quartile. Furthermore, the risk of reduced eGFR increased linearly with the increment of TyG. Subgroup analysis also revealed the association was robust to several risk factors of renal damage. Finally, category-free net reclassification index (0.204, 95% CI: 0.060–0.349, p = 0.005) and integrated discrimination index (0.010, 95% CI: 0.005–0.016, p < 0.001) demonstrated the value of TyG to refine the risk stratification of reduced eGFR.Conclusion: Our work reveals the robust association between TyG and reduced eGFR. Moreover, the present study implicates the potential role of TyG as a risk indicator to optimize the prevention of reduced eGFR. Lastly, the findings also suggest the importance of simultaneous glycemic and lipid control to avoid the development and progression of reduced eGFR.

Risk Stratifying Low-Flow, Low-Gradient Aortic Stenosis using N-Terminal Pro B-Type Natriuretic Peptide: Results from TOPAS study

In low-flow, low gradient aortic stenosis (LFLG-AS i.e. with low left ventricular [LV] ejection fraction), there is no data on the value of the clinical activation ratios of B-type natriuretic peptide (BNP-ratio) versus aminoterminal-proBNP (NT-proBNP-ratio) as surrogates of LV impairment to risk-stratify the patients. BNP and NT-proBNP-ratios were calculated by dividing the actual serum level by the upper normal reference value for age and sex in 238 prospectively recruited LFLG-AS patients. After adjustment for the severity of AS, initial treatment (aortic valve replacement [AVR] vs. conservative management), age, sex and the euroSCORE (Model#1), BNP-ratio > 7.4 had a trend to predict time to death (HR = 2.14[1.00–4.58], P = 0.05). NT-proBNP ratio significantly predicted one and three-year mortality (area under the curve [AUC] = 0.67 ± 0.04 and 0.66 ± 0.05, both P = 0.001), and independently predicted mortality (HR = 1.39 [1.11–1.74], per unit LogNT-proBNP-ratio, P = 0.004). In a head-to-head comparison, the AUCs for one and three-year mortality were higher with NT-proBNP-ratio versus BNP-ratio ( P < 0.009). NT-proBNP-ratio but not BNP-ratio independently predicted mortality and significantly improved Model#1 (Likelihood ratio test Chi 2 = 15.953, P = 0.0003). The category-free net reclassification index of NT-proBNP-ratio when added to the multivariable model was 0.71 ( P = 0.008) versus 0.38 ( P = 0.15) for BNP-ratio. Furthermore, there was a marked survival benefit associated with AVR in patients with NT-proBNP-ratio > 11 (62% with severe AS), while those < 11 (46% with severe AS) had excellent short-term survival under conservative management ( Fig. 1 ). NT-proBNP-ratio strongly predicts the risk of mortality and is superior to BNP-ratio to risk stratify LFLG-AS patients. The assessment of LV function impairment using NT-proBNP-ratio has important clinical implications and should be complementary to the determination of true AS severity.

CT-based peritumoral radiomics signatures to predict early recurrence in hepatocellular carcinoma after curative tumor resection or ablation

ObjectiveTo construct a prediction model based on peritumoral radiomics signatures from CT images and investigate its efficiency in predicting early recurrence (ER) of hepatocellular carcinoma (HCC) after curative treatment.Materials and methodsIn total, 156 patients with primary HCC were randomly divided into the training cohort (109 patients) and the validation cohort (47 patients). From the pretreatment CT images, we extracted 3-phase two-dimensional images from the largest cross-sectional area of the tumor. A region of interest (ROI) was manually delineated around the lesion for tumoral radiomics (T-RO) feature extraction, and another ROI was outlined with an additional 2 cm peritumoral area for peritumoral radiomics (PT-RO) feature extraction. The least absolute shrinkage and selection operator (LASSO) logistic regression model was applied for feature selection and model construction. The T-RO and PT-RO models were constructed. In the validation cohort, the prediction efficiencies of the two models and peritumoral enhancement (PT-E) were evaluated qualitatively by receiver operating characteristic (ROC) curves, calibration curves and decision curves and quantitatively by area under the curve (AUC), the category-free net reclassification index (cfNRI) and integrated discrimination improvement values (IDI).ResultsBy comparing AUC values, the prediction accuracy in the validation cohort was good for the PT-RO model (0.80 vs. 0.79, P = 0.47) but poor for the T-RO model (0.82 vs. 0.62, P < 0.01), which was significantly overfitted. In the validation cohort, the ROC curves, calibration curves and decision curves indicated that the PT-RO model had better calibration efficiency and provided greater clinical benefits. CfNRI indicated that the PT-RO model correctly reclassified 47% of ER patients and 32% of non-ER patients compared to the T-RO model (P < 0.01); additionally, the PT-RO model correctly reclassified 24% of ER patients and 41% of non-ER patients compared to PT-E (P = 0.02). IDI indicated that the PT-RO model could improve prediction accuracy by 0.22 (P < 0.01) compared to the T-RO model and by 0.20 (P = 0.01) compared to PT-E.ConclusionThe CT-based PT-RO model can effectively predict the ER of HCC and is more efficient than the T-RO model and the conventional imaging feature PT-E.

Comparison of the Kansas City Cardiomyopathy Questionnaire and Minnesota Living With Heart Failure Questionnaire in Predicting Heart Failure Outcomes

Patient-reported outcome measures (PROMs) are relevant independent outcomes in heart failure (HF) care and are predictive of subsequent hospitalization and death in HF. The Kansas City Cardiomyopathy Questionnaire (KCCQ) and the Minnesota Living with Heart Failure Questionnaire (MLHFQ) are the 2 most widely adopted PROMs specific to HF. We compared their prognostic abilities in a prospective cohort of HF patients. A prospective cohort of subjects from a single-center registry was analyzed with regard to baseline KCCQ and MLHFQ scores and the outcomes of death, transplant, or left ventricular assist device implantation and hospitalization. A total of 516 subjects with reduced left ventricular ejection fraction (HFrEF) and 151 subjects with preserved left ventricular ejection fraction (HFpEF) were included. Discrimination was assessed using c-statistics based on time-to-event analyses and receiver-operator curves. The additive contribution of MLHFQ was assessed through the change in c-statistic, incremental discrimination index, and category-free net reclassification index. Overall, KCCQ was superior to MLHFQ for predicting death/transplant/ventricular assist device (c-statistic 0.702 [0.666 to 0.738] and 0.658 [0.621 to 0.695] respectively, p value for difference <0.001) and hospitalization (c-statistic 0.640 [0.613 to 0.666] and 0.624 [0.597 to 0.651], respectively, p value for difference 0.022). However, this difference was statistically nonsignificant in the HFpEF group alone. When analyzing the additional prognostic information afforded by adding MLHFQ to KCCQ in the overall, HFrEF, and HFpEF groups there was no significant improvement, although adding KCCQ to MLHFQ did significantly improve risk stratification. Scoring based upon the abbreviated KCCQ-12 did not reduce the prognostic accuracy of KCCQ. In conclusion, KCCQ is more prognostic of death/transplant/left ventricular assist device and hospitalization than MLHFQ in a combined cohort of patients with HFrEF and HFpEF, although the effect in HFpEF was less pronounced. KCCQ should be the preferred PROM for patients with HF if prognostication is a desired goal of using the PROMs.

Prognostic Implications of Relative Increase and Final Fractional Flow Reserve in Patients With Stent Implantation

ObjectivesThe study sought to investigate the prognostic implications of relative increase of fractional flow reserve (FFR) with PCI in combination with post–percutaneous coronary intervention (PCI) FFR. BackgroundFFR, measured after PCI has been shown to possess prognostic implications. The relative increase of FFR with PCI can be determined by the interaction of baseline disease pattern, adequacy of PCI, and residual disease burden in a target vessel. However, the role of relative increase of FFR with PCI has not yet been evaluated. MethodsA total of 621 patients who underwent PCI using second-generation drug-eluting stents based on low pre-PCI FFR (≤0.80) and available post-PCI FFR were analyzed. The relative increase of FFR was calculated by %FFR increase with PCI ([post-PCI FFR – pre-PCI FFR]/pre-PCI FFR × 100). Patients were divided according to the optimal cutoff values of post-PCI FFR (<0.84) and %FFR increase (≤15%). The primary outcome was target vessel failure (TVF) (a composite of cardiac death, target vessel–related myocardial infarction, and clinically driven target vessel revascularization) at 2 years. ResultsAmong the total population, 66.0% showed high post-PCI FFR (≥0.84) and 69.2% showed high %FFR increase (>15%). Patients with low post-PCI FFR showed a higher risk of 2-year TVF than did those with high post-PCI FFR (9.1% vs. 2.6%; hazard ratio [HR]: 3.367; 95% confidence interval [CI]: 1.412 to 8.025; p = 0.006). Patients with low %FFR increase also showed a higher risk of 2-year TVF compared with those with high %FFR increase (9.2% vs. 3.0%; HR: 3.613; 95% CI: 1.543 to 8.458; p = 0.003). Among the high post-PCI FFR group, there were no significant differences in clinical outcomes according to %FFR increase. Conversely, among the low post-PCI FFR group, those with low %FFR increase showed a significantly higher risk of TVF than did those with high %FFR increase (14.3% vs. 4.1%; HR: 4.334; 95% CI: 1.205 to 15.594; p = 0.025). Percent FFR increase significantly increased discriminant and reclassification ability for the occurrence of TVF when added to a model with clinical risk factors and post-PCI FFR (C-index 0.783 vs. 0.734; relative integrated discrimination improvement 0.702; p = 0.009; category-free net reclassification index 0.479; p = 0.031). ConclusionsPercent FFR increase with PCI showed similar prognostic implications with post-PCI FFR. Adding the relative increase of FFR to post-PCI FFR would enable better discrimination of high-risk patients after stent implantation. (Influence of FFR on the Clinical Outcome After Percutaneous Coronary Intervention [PERSPECTIVE]; NCT01873560)

Prognostic implication of thermodilution coronary flow reserve in patients with indeterminate pressure-bounded coronary flow reserve

BackgroundRecently, the concept of pressure-bounded coronary flow reserve (PB-CFR) has been introduced. However, using pressure-only data, a substantial proportion of patients could not be classified into high or low PB-CFR and remained as indeterminate PB-CFR. The current study evaluated the prognostic implication of thermodilution-based CFR (thermo-CFR) in patients with indeterminate PB-CFR. MethodsAmong 199 patients (211 lesions) with indeterminate PB-CFR, 170 patients (179 lesions) with deferral of revascularization were analyzed for the current study. The rates of patient-oriented composite outcomes (POCO, a composite of all-cause mortality, any myocardial infarction, and any ischemia-driven revascularization) were compared according to thermo-CFR. All patients underwent fractional flow reserve (FFR) and thermo-CFR measurements. Thermo-CFR ≤ 2.0 was classified as low thermo-CFR. The median follow-up duration was 1350.0 (Q1-Q3 1252.0–1468.0) days. ResultsMean angiographic percent diameter stenosis, FFR, and thermo-CFR were 42.3 ± 13.9, 0.84 ± 0.06, and 3.10 ± 1.15, respectively. Among 170 patients, 36 patients (21.2%) showed low thermo-CFR. Patients with low thermo-CFR showed significantly higher rate of POCO compared to those with high thermo-CFR (30.6% vs. 3.0%, HR 12.117, 95% CI 3.854–38.091, p < 0.001). Adding thermo-CFR to a prediction model with FFR significantly increased discrimination and reclassification index for the risk of POCO (c-index 0.545 vs. 0.766, p = 0.002, category-free net reclassification index 1.169, p < 0.001, relative integrated discrimination index 31.828, p < 0.001). ConclusionsPatients with low thermo-CFR showed a significantly higher risk of POCO compared to those with high thermo-CFR among patients with indeterminate PB-CFR. Thermo-CFR showed additional prognostic implication, in addition to FFR, in patients with indeterminate PB-CFR.

Identification of High-Risk Plaques Destined to Cause Acute Coronary Syndrome Using Coronary Computed Tomographic Angiography and Computational Fluid Dynamics

ObjectivesThe authors investigated the utility of noninvasive hemodynamic assessment in the identification of high-risk plaques that caused subsequent acute coronary syndrome (ACS). BackgroundACS is a critical event that impacts the prognosis of patients with coronary artery disease. However, the role of hemodynamic factors in the development of ACS is not well-known. MethodsSeventy-two patients with clearly documented ACS and available coronary computed tomographic angiography (CTA) acquired between 1 month and 2 years before the development of ACS were included. In 66 culprit and 150 nonculprit lesions as a case-control design, the presence of adverse plaque characteristics (APC) was assessed and hemodynamic parameters (fractional flow reserve derived by coronary computed tomographic angiography [FFRCT], change in FFRCT across the lesion [△FFRCT], wall shear stress [WSS], and axial plaque stress) were analyzed using computational fluid dynamics. The best cut-off values for FFRCT, △FFRCT, WSS, and axial plaque stress were used to define the presence of adverse hemodynamic characteristics (AHC). The incremental discriminant and reclassification abilities for ACS prediction were compared among 3 models (model 1: percent diameter stenosis [%DS] and lesion length, model 2: model 1 + APC, and model 3: model 2 + AHC). ResultsThe culprit lesions showed higher %DS (55.5 ± 15.4% vs. 43.1 ± 15.0%; p < 0.001) and higher prevalence of APC (80.3% vs. 42.0%; p < 0.001) than nonculprit lesions. Regarding hemodynamic parameters, culprit lesions showed lower FFRCT and higher △FFRCT, WSS, and axial plaque stress than nonculprit lesions (all p values <0.01). Among the 3 models, model 3, which included hemodynamic parameters, showed the highest c-index, and better discrimination (concordance statistic [c-index] 0.789 vs. 0.747; p = 0.014) and reclassification abilities (category-free net reclassification index 0.287; p = 0.047; relative integrated discrimination improvement 0.368; p < 0.001) than model 2. Lesions with both APC and AHC showed significantly higher risk of the culprit for subsequent ACS than those with no APC/AHC (hazard ratio: 11.75; 95% confidence interval: 2.85 to 48.51; p = 0.001) and with either APC or AHC (hazard ratio: 3.22; 95% confidence interval: 1.86 to 5.55; p < 0.001). ConclusionsNoninvasive hemodynamic assessment enhanced the identification of high-risk plaques that subsequently caused ACS. The integration of noninvasive hemodynamic assessments may improve the identification of culprit lesions for future ACS. (Exploring the Mechanism of Plaque Rupture in Acute Coronary Syndrome Using Coronary CT Angiography and Computational Fluid Dynamic [EMERALD]; NCT02374775)

Posttreatment National Institutes of Health Stroke Scale Is Superior to the Initial Score or Thrombolysis in Cerebral Ischemia for 3-Month Outcome.

The majority of ischemic stroke patients receiving endovascular recanalization therapy (EVT) experience variable changes of neurological severities during the hyperacute period. We hypothesized that the National Institutes of Health Stroke Scale (NIHSS) score after EVT is a better prognostic factor compared with the initial NIHSS score or revascularization status. We identified 566 stroke patients who received EVT at Seoul National University Bundang Hospital between April 2008 and December 2015. We prospectively collected post-EVT NIHSS score, which was measured in the angiography suite by on-duty physicians after completion of EVT. Model 1 included baseline predictors including an initial NIHSS score. In model 2, 3, and 4, revascularization status, post-EVT NIHSS score, or both were additionally included. The discrimination powers for modified Rankin Scale score of 0 to 2 at 3 months were assessed using C statistic, integrated discrimination index, and category-free net reclassification index. The median of initial and post-EVT NIHSS score were 14 (9-19) and 11 (5-17) points, respectively (an improvement, 58.8%; no change, 20.7%; deterioration, 20.5%). A modified Rankin Scale score 0 to 2 at 3 months was achieved in 47%. Based on the results of differences among the C statistics, both model 3 and model 4 (C statistics: 0.896 and 0.906) showed significantly increased discrimination power for modified Rankin Scale score 0 to 2 at 3 months than the model 1 or 2 (C statistics: 0.802 and 0.834, P values<0.001 for all comparisons). Model 4 showed significant improvement of both integrated discrimination index and net reclassification index as compared with all other models, but the magnitude of improvement from model 3 to model 4 (integrated discrimination index, 0.021; net reclassification index, 0.322) was modest. Incorporation of post-EVT NIHSS score conferred better discrimination power to the statistical models for functional recovery. Post-EVT NIHSS score may be an appropriate baseline factor when evaluating an intervention after hyperacute period.

Abstract No. 615 Performance of MELD score versus MELD-Na in predicting 90-day survival after TIPS placement

The Model for End-Stage Liver Disease (MELD) score was originally developed to assess prognosis after TIPS placement. The addition of sodium to the MELD score (MELD-Na) has been proposed as a more accurate prognostic score for overall survival. We assessed the performance of MELD vs. MELD-Na score in predicting 90-day survival after TIPS placement. Retrospective review of TIPS placements for ascites and hemorrhage at a major academic healthcare network between 1994 and 2014. MELD/MELD-Na scores were assessed for their ability to predict 90-day survival using area under the receiver operative curves (AuROC) and compared using change in AuROC, category free net reclassification index (NRI), and integrated discrimination improvement (IDI). Stratified analysis by TIPS indication was performed. Three hundred seventy seven TIPS placements were analyzed (225 for hemorrhage, 152 for ascites/hydrothorax). 90-day survival was 75% (282/377) for all subjects, 76% (170/225) for the hemorrhage subgroup, and 74% (112/150) for the ascites subgroup. For all subjects, median MELD score was 16 [IQR 11, 21] with AuROC of 0.77. Median MELD-Na was 19 [IQR 14, 23] with AuROC of 0.77. MELD and MELD-Na were highly correlated (r = 0.94; p < 0.001). MELD AuROC was 0.82 for the bleeding subgroup and 0.70 for the ascites subgroup. Addition of Na to the MELD score did not improve its performance by change in AuROC, NRI, or IDI for all patients or when stratified by TIPS indication, except for NRI in the bleeding subgroup (see Table). MELD and MELD-Na performed similarly when predicting 90-day survival after TIPS. MELD-Na score performs better in TIPS placed for hemorrhage compared to TIPS placed for ascites. Further study is required to improve predictive scores in this population.Tabled 1All TIPSn = 377Hemorrhagen = 225Ascitesn = 152MELDMELD-NaP ValueMELDMELD-NaP-valueMELDMELD-NaP ValueAuROC0.770.770.620.820.830.510.700.700.26Category-free NRI0.050.680.330.050.210.26IDI0.00040.530.010.20-0.000020.98 Open table in a new tab

Abstract TP41: Impact of Post-Endovascular Treatment Stroke Severity to Discriminate Long-Term Functional Outcome After Ischemic Stroke

Background and Aims: About 40% of patients have poor functional outcome even after modern endovascular treatment (EVT), which is strongly requiring post-EVT therapies targeting damages from ischemia-reperfusion. Futile or effective reperfusion can be a significant noise in assessing efficacy of post-EVT therapies. Thus, we investigated whether post-EVT NIHSS score better discriminates long-term outcomes after EVT. Methods: We identified 566 stroke patients who received EVT at Seoul National University Bundang Hospital between Apr 2008 and Dec 2015. We prospectively collected post-EVT NIHSS score, which was measured by on-duty physician right after completion of EVT. The prognostic values of post-EVT NIHSS score, recanalization status and other baseline profiles in discriminating favorable 3-month outcome (modified Rankin Scale 0-2) were assessed using 4 different logistic regression models (Figure). The c-statistic was used as the primary measure of model discrimination. The integrated discrimination index (IDI) and the category-free net reclassification index (NRI) were also compared. Results: The post-EVT NIHSS score was 11.8 ± 7.8 pts, which showed overall improvement of 2.3 ± 6.3 pts compared to the baseline NIHSS score (14.0 ± 7.1 pts). Any change of NIHSS score was present in 79%. Favorable 3-month outcome was achieved in 47% of patients. Based on the results of differences among the c-statistics, both model 3 and model 4 were significantly more discriminative of favorable outcome than model 1 or model 2 (Figure). Model 4 showed significant improvement of both IDI and NRI as compared to all other models, but the magnitude from model 3 to model 4 was modest (IDI, 0.021; NRI, 0.322). Conclusions: Post-EVT NIHSS score was the strongest factor in discriminating long-term outcomes after EVT, which was superior to baseline NIHSS or recanalization status. Therefore, post-EVT NIHSS score should to be considered when assessing the efficacy of post-EVT therapies.