scholarly journals Defining an Optimal Cut-Point Value in ROC Analysis: An Alternative Approach

2017 ◽  
Vol 2017 ◽  
pp. 1-14 ◽  
Author(s):  
Ilker Unal
Keyword(s):  
Diagnostic Accuracy ◽  
Sensitivity And Specificity ◽  
Roc Curve ◽  
Simulated Data ◽  
Real Data ◽  
Roc Curve Analysis ◽  
New Approach ◽  
Cut Point ◽  
Alternative Approach ◽  
The Difference

ROC curve analysis is often applied to measure the diagnostic accuracy of a biomarker. The analysis results in two gains: diagnostic accuracy of the biomarker and the optimal cut-point value. There are many methods proposed in the literature to obtain the optimal cut-point value. In this study, a new approach, alternative to these methods, is proposed. The proposed approach is based on the value of the area under the ROC curve. This method defines the optimal cut-point value as the value whose sensitivity and specificity are the closest to the value of the area under the ROC curve and the absolute value of the difference between the sensitivity and specificity values is minimum. This approach is very practical. In this study, the results of the proposed method are compared with those of the standard approaches, by using simulated data with different distribution and homogeneity conditions as well as a real data. According to the simulation results, the use of the proposed method is advised for finding the true cut-point.

An Alternative Approach for Evaluating the Binormal ROC Curve

2013 ◽  
Vol 4 (1) ◽  
pp. 1-17 ◽  
Author(s):  
R. Amala ◽  
R. Vishnu Vardhan
Keyword(s):  
Roc Curve ◽  
Error Function ◽  
False Positive Rate ◽  
True Positive Rate ◽  
Roc Curve Analysis ◽  
New Approach ◽  
Alternative Approach ◽  
Positive Rate ◽  
Almost All ◽  
Two Populations

In recent years the ROC curve analysis has got its attention in almost all diversified fields. Basing on the data pattern and its distribution various forms of ROC models have been derived. In this paper, the authors have assumed that the data of two populations (healthy and diseased) follows normal distribution, it is one of the most commonly used forms under parametric approach. The present paper focuses on providing an alternative approach for the tradeoff plot of ROC curve and the computation of AUC using a special function of sigmoid shape called Error function. It is assumed that the test scores of particular biomarker are normally distributed. The entire work has been carried out for providing a new approach for the construction of Binormal ROC curve, which makes use of Error function which can be called as ErROC curve. The summary measure AUC of the resulting ErROC curve has been estimated and defined as ErAUC. The authors have also focused on deriving the expression for obtaining the optimal cut-off point. The new ErROC curve model will provide the true positive rate value at each and every point of false positive rate unlike conventional Binormal ROC model.

Baseline apparent diffusion coefficients: Validation study of new predictor of survival in patients with unresectable hepatocellular carcinoma following chemoembolization

2021 ◽  
pp. 1-10
Author(s):  
Lichao Xu ◽  
Shiqin Wang ◽  
Shengping Wang ◽  
Ying Wang ◽  
Wentao Li ◽  
...  
Keyword(s):  
Hepatocellular Carcinoma ◽  
Roc Curve ◽  
Cox Regression ◽  
Response To Treatment ◽  
Roc Curve Analysis ◽  
Adc Value ◽  
Apparent Diffusion Coefficients ◽  
Apparent Diffusion ◽  
Response Biomarker ◽  
The Difference

OBJECTIVES: To investigate whether the baseline apparent diffusion coefficient (ADC) can predict survival in the hepatocellular carcinoma (HCC) patients receiving chemoembolization. MATERIALS AND METHODS: Diffusion-weighted MR imaging of HCC patients is performed within 2 weeks before chemoembolization. The ADC of the largest index lesion is recorded. Responses are assessed by mRECIST after the start of the second course of chemoembolization. Receiver operating characteristic (ROC) curve analysis is performed to evaluate the diagnostic performance and determine optimal cut-off values. Cox regression and Kaplan–Meier survival analyses are used to explore the differences in overall survival (OS) between the responders and non-responders. RESULTS: The difference is statistically significant in the baseline ADC between the responders and non-responders (P <  0.001). ROC analyses indicate that the baseline ADC value is a good predictor of response to treatment with an area under the ROC curve (AUC) of 0.744 and the optimal cut-off value of 1.22×10–3 mm2/s. The Cox regression model shows that the baseline ADC is an independent predictor of OS, with a 57.2% reduction in risk. CONCLUSION: An optimal baseline ADC value is a functional imaging response biomarker that has higher discriminatory power to predict tumor response and prolonged survival following chemoembolization in HCC patients.

The Rowland Universal Dementia Assessment Scale (RUDAS): a multicultural cognitive assessment scale

2004 ◽  
Vol 16 (1) ◽  
pp. 13-31 ◽  
Author(s):  
Joella E. Storey ◽  
Jeffrey T. J. Rowland ◽  
David A. Conforti ◽  
Hugh G. Dickson
Keyword(s):  
Roc Curve ◽  
Predictive Accuracy ◽  
Geriatric Medicine ◽  
Community Dwelling ◽  
Assessment Scale ◽  
Elderly Persons ◽  
Roc Curve Analysis ◽  
Simple Method ◽  
Cognitive Domains ◽  
Cut Point

Objective: To develop and validate a simple method for detecting dementia that is valid across cultures, portable and easily administered by primary health care clinicians.Design: Culture and Health Advisory Groups were used in Stage 1 to develop culturally fair cognitive items. In Stage 2, clinical testing of 42 items was conducted in a multicultural sample of consecutive new referrals to the geriatric medicine outpatient clinic at Liverpool Hospital, Sydney, Australia (n=166). In Stage 3, the predictive accuracy of items was assessed in a random sample of community-dwelling elderly persons stratified by language background and cognitive diagnosis and matched for sex and age (n=90).Measurements: A research psychologist administered all cognitive items, using interpreters when needed. Each patient was comprehensively assessed by one of three geriatricians, who ordered relevant investigations, and implemented a standardized assessment of cognitive domains. The geriatricians also collected demographic information, and administered other functional and cognitive measures. DSM-IV criteria were used to assign cognitive diagnoses. Item validity and weights were assessed using frequency and logistic regression analyses. Receiver-operating characteristic (ROC) curve analysis was used to determine overall predictive accuracy of the RUDAS and the best cut-point for detecting cognitive impairment.Results: The 6-item RUDAS assesses multiple cognitive domains including memory, praxis, language, judgement, drawing and body orientation. It appears not to be affected by gender, years of education, differential performance factors and preferred language. The area under the ROC curve for the RUDAS was 0.94 (95% CI 0.87–0.98). At a cut-point of 23 (maximum score of 30), sensitivity and specificity were 89% and 98%, respectively. Inter-rater (0.99) and test-retest (0.98) reliabilities were very high.Conclusions: The 6-item RUDAS is portable and tests multiple cognitive domains. It is easily interpreted to other languages, and appears to be culturally fair. However, further validation is needed in other settings, and in longitudinal studies to determine its sensitivity to change in cognitive function over time.

scholarly journals Relation of Airway Responsiveness to Methacholine to Parent and Child Reporting of Symptoms Suggecting Asthma

1996 ◽  
Vol 3 (2) ◽  
pp. 115-123 ◽  
Author(s):  
JM FitzGerald ◽  
DE Fester ◽  
MM Morris ◽  
M Schulzer ◽  
FE Hargreave ◽  
...  
Keyword(s):  
Sensitivity And Specificity ◽  
Roc Curve ◽  
Airway Hyperresponsiveness ◽  
Respiratory Symptoms ◽  
Respiratory Illness ◽  
Airway Responsiveness ◽  
Methacholine Challenge ◽  
Cut Point ◽  
The Past ◽  
History Of

BACKGROUND:The lack of a relationship between airway responsiveness and respiratory symptoms in epidemiological studies of children may, in part, reflect inaccuracies in symptom reporting or inadequate knowledge by the parent of the child's symptoms.OBJECTIVE:To relate airway responsiveness to methacholine in children with symptoms of respiratory illness in the child as reported by the parent and as reported by the child.POPULATION:Eight- to 10-year-old (n=290) randomly sampled schoolchildren.SETTING:Seven randomly selected schools in Ontario.METHODS:Parents completed a mailed questionnaire regarding the child's respiratory health. Children completed a similar interview-administered questionnaire at school and underwent methacholine challenge testing by the tidal breathing method.RESULTS:The cumulative prevalence of a history of physician-diagnosed asthma was 9.0%, and of any wheezing it was 25.5%. A further 9% of children reported wheezing not documented by their parent. Of 229 children consenting to methacholine challenge, 78 (34.1%) showed airway responsiveness in the range generally associated with asthma in adults (provocation concentration of methacholine causing a 20% fall [PC20] in forced expired volume in 1 s [FEV1] 8 mg/mL or less); half of these children had no history of respiratory symptoms reported by the parent. The sensitivity of airway hyperresponsiveness defined by a cut-point for PC208 mg/mL or less in relation to any history of recurrent wheezing reported by the parent was 48% and did not improve if only symptoms within the past year were considered (sensitivity 44%); the specificity of the test for parent-reported symptoms ever was 71%, and 68% in those with symptoms in the past year. None of these sensitivities or specificities was increased by using symptoms reported by the child or by combining parent and child reported symptoms. Receiver operating characteristic (ROC) curves for sensitivity and specificity of the methacholine test were constructed for parent and child reports of symptoms. For all symptom strata, the cut-point of PC20producing optimal balance of sensitivity and specificity was between 4 and 8 mg/mL. A parental questionnaire positive for physician-diagnosed asthma was strongly related to methacholine response, producing an ROC curve with an area significantly different from 0.5 (P=0.006), as did all parent-reported wheezing (P=0.009). If the child reported asthma, there was an equally strong relationship, with a positive ROC curve (P=0.001), as there was for all child-reported wheezing (P=0.048).CONCLUSIONS:Airway hyperresponsiveness to methacholine in children relates closely with asthma and wheezing reported by either the parent or the child. In addition, the results confirm that respiratory symptoms and airway hyperresponsiveness are common in Canadian children, and that airway hyperresponsiveness may be found in children with no history of respiratory illness either at present or in the past.

scholarly journals Is There a Role for Prostate Tumour Overexpressed-1 in the Diagnosis of HGPIN and of Prostatic Adenocarcinoma? A Comparison with α-Methylacyl CoA Racemase

2012 ◽  
Vol 25 (1) ◽  
pp. 67-74 ◽  
Author(s):  
M. Scarpelli ◽  
R. Mazzucchelli ◽  
F. Barbisan ◽  
A. Santinelli ◽  
A. Lopez-Beltran ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Diagnostic Accuracy ◽  
Differential Display ◽  
Roc Curve ◽  
Intraepithelial Neoplasia ◽  
Prostatic Intraepithelial Neoplasia ◽  
Curve Analysis ◽  
High Grade ◽  
Roc Curve Analysis ◽  
Prostate Tumour

Prostate Tumour Overexpressed-1 (PTOV1) was recently identified as a novel gene and protein during a differential display screening for genes overexpressed in prostate cancer (PCa). α-Methyl-CoA racemose (AMACR) mRNA was identified as being overexpressed in PCa. PTOV1 and racemase were immunohistochemically evaluated in PCa, high-grade prostatic intraepithelial neoplasia (HGPIN), atrophy and normal-looking epithelium (NEp) in 20 radical prostatectomies (RPs) with pT2a Gleason score 6 prostate cancer with the aim of analyzing the differences in marker expression between PTOV1 and AMACR. The level of expression of PTOV1 and AMACR increased from NEp and atrophy through HGPIN, away from and adjacent to prostate cancer, to PCa. With the ROC curve analysis the overall accuracy in distinguishing PCa vs HGPIN away from and adjacent to cancer was higher for AMACR than for PTOV1. In conclusion, AMACR can be considered a more accurate marker than PTOV1 in the identification of HGPIN and of PCa. However, PTOV1 may aid in the diagnosis of PCa, at least to supplement AMACR as another positive marker of carcinoma and to potentially increase diagnostic accuracy.

scholarly journals The Diagnostic Value of Eosinophil Count And Platelet-To-Lymphocyte Ratio In Eosinophilic Gastroenteritis

2021 ◽  
Author(s):  
Qiong Lin ◽  
Renmin Zhou ◽  
Hao Wujuan ◽  
Zhumeng Ni ◽  
Xiaozhong Li
Keyword(s):  
Blood Cell ◽  
Roc Curve ◽  
Diagnostic Value ◽  
Eosinophilic Gastroenteritis ◽  
Control Group ◽  
Platelet To Lymphocyte Ratio ◽  
Roc Curve Analysis ◽  
Record System ◽  
Lymphocyte Ratio ◽  
The Difference

Abstract Objective: To evaluate the diagnostic value of eosinophil (EO) count and platelet-to-lymphocyte ratio (PLR) in eosinophilic gastroenteritis (EGE). Methods: In total, 91 patients with EGE and 83 age–sex matched patients without EGE were selected as study subjects during January 2018 to December 2020. Data on blood cell count, and serum, C-reactive protein (CRP), and albumin levels were obtained from the Wuxi children's hospital electronic medical record system; the neutrophil-to-lymphocyte ratio (NLR), PLR, and CRP-to-albumin ratio (CAR) in the peripheral blood were recorded. Independent sample t-test, non-parametric test, or χ2 test was used according the data type to compare the difference between two groups, and receiver operating characteristic (ROC) curve analysis was performed to evaluate the diagnostic value for EGE. Results: The EO counts and PLR were significantly higher in the EGE group than those in the control group, whereas differences in the white blood cell, lymphocyte, neutrophil, and platelet counts, and the CRP level, NLR, and CAR were not significant. After treatment(Corticosteroids, 1mg/kg.d, lasting for 2 weeks）, the EO counts and PLR in the EGE group decreased gradually and the difference was significant. The diagnostic value of EO counts and PLR was determined with an area under the ROC curve as 0.756 and 0.616, sensitivity was 75.00% and 34.29%, and specificity was 74.29% and 92.31%, respectively. Conclusions EO and PLR represent potential predictive markers for diagnosing EGE.

Diagnostic accuracy of pancreatic enzymes evaluated by use of multivariate data analysis

1993 ◽  
Vol 39 (9) ◽  
pp. 1960-1965 ◽  
Author(s):  
S C Kazmierczak ◽  
P G Catrou ◽  
F Van Lente
Keyword(s):  
Neural Network ◽  
Network Analysis ◽  
Diagnostic Accuracy ◽  
Roc Curve ◽  
Enzyme Activities ◽  
Laboratory Tests ◽  
Curve Analysis ◽  
Multivariate Techniques ◽  
Neural Network Analysis ◽  
Roc Curve Analysis

Abstract We analyzed pancreatic enzyme data from 508 patients with suspected pancreatitis by neural network analysis, by an Expert multirule generation protocol, and by receiver-operator characteristic (ROC) curve analysis of a single test result. Neural network analysis showed that use of lipase provided the best means for diagnosing pancreatitis. Diagnostic accuracies achieved by using amylase only, lipase only, and amylase and lipase in combination were 76%, 82%, and 84%, respectively. Use of the Expert rule generation protocol provided a diagnostic accuracy of 92% when rules for single and multiple samplings were combined. ROC curve analysis for initial enzyme activities showed the maximal diagnostic accuracy to be 82% and 85% for amylase and lipase, respectively; use of peak enzyme activities yielded accuracies of 81% and 88%, respectively. The evaluation of laboratory test data should include analysis of the diagnostic accuracy of laboratory tests by multivariate techniques such as neural network analysis or an Expert systems approach. Multivariate analysis should allow for a more realistic assessment of the diagnosis accuracy of laboratory tests because all the available data are included in the evaluation.

scholarly journals Serum Zonulin in HBV-Associated Chronic Hepatitis, Liver Cirrhosis, and Hepatocellular Carcinoma

2019 ◽  
Vol 2019 ◽  
pp. 1-6 ◽  
Author(s):  
Xin Wang ◽  
Ming-ming Li ◽  
Ye Niu ◽  
Xin Zhang ◽  
Ji-bin Yin ◽  
...  
Keyword(s):  
Hepatocellular Carcinoma ◽  
Liver Cirrhosis ◽  
Chronic Hepatitis ◽  
Diagnostic Accuracy ◽  
Roc Curve ◽  
Healthy Subjects ◽  
Serum Levels ◽  
Curve Analysis ◽  
Advanced Stage ◽  
Roc Curve Analysis

Background. The gut microbiota is involved in the occurrence and development of chronic liver diseases. Zonulin is considered a marker of intestinal permeability. The purpose of this study was to assess zonulin levels in patients with chronic hepatitis B (CHB), HBV-associated liver cirrhosis (LC), and HBV-associated hepatocellular carcinoma (HCC). Materials and Methods. The study population consisted of 90 HBV-associated HCC patients, 90 HBV-associated LC patients, 90 CHB patients, and 90 healthy subjects. Serum levels of zonulin and AFP were determined. The diagnostic accuracy of each marker was evaluated using receiver operating characteristic (ROC) curve analysis (AUC). Results. Serum zonulin levels were significantly higher in patients with HCC than in patients with LC or CHB or healthy subjects (p<0.001). Moreover, the zonulin levels were increased in the advanced stage of LC and HCC. ROC curve analysis revealed that serum zonulin could be used to differentiate CHB from cirrhosis. In addition, the combination of zonulin and AFP exhibited a significantly larger AUC compared with zonulin or AFP alone. Conclusions. Serum zonulin levels were significantly increased both in LC and in HCC and correlated with the advanced stage of LC and HCC. Moreover, the combination of zonulin and AFP confers significant benefit to diagnostic accuracy in differentiating LC from HCC.

Use of percent positive biopsy cores to predict prostate cancer–specific death in patients treated with dose-escalated radiotherapy.

2011 ◽  
Vol 29 (7_suppl) ◽  
pp. 35-35
Author(s):  
Y. Qian ◽  
F. Y. Feng ◽  
S. Halverson ◽  
K. Blas ◽  
H. M. Sandler ◽  
...  
Keyword(s):  
Prostate Cancer ◽  
Roc Curve ◽  
Risk Group ◽  
Curve Analysis ◽  
External Beam Radiation ◽  
Rank Test ◽  
Roc Curve Analysis ◽  
Cut Point ◽  
Log Rank Test ◽  
Positive Biopsy

35 Background: The percent of positive biopsy cores (PPC)-considered a surrogate of local disease burden-has been shown to predict biochemical failure (BF) after external beam radiation therapy (EBRT), but most series have used conventional dose RT. Dose-escalated RT has been demonstrated to improve prostate cancer outcomes, but the value of PPC is unclear in the setting of RT doses high enough to decrease local failure. Methods: A retrospective evaluation was performed of 651 patients treated to ≥75 Gy with biopsy core information available. Patients were stratified for PPC by quartile, and differences by quartile in BF, freedom from metastasis (FFM), cause specific survival (CSS), and overall survival (OS) were assessed using the log-rank test. Receiver operated characteristic (ROC) curve analysis was utilized to determine an optimal cut-point for PPC. Cox proportional hazards multivariate regression was utilized to assess the impact of PPC on clinical outcome when adjusting for risk group. Results: With median follow-up of 62 months the median number of cores sampled was 7 (IQR: 6–12) with median PPC in 38% (IQR: 17%-67%). On log-rank test, BF, FFM, and CSS were all associated with PPC (p < 0.005 for all), with worse outcomes only for the highest PPC quartile (>67%). There was no observed difference in OS based upon PPC. ROC curve analysis confirmed a cut-point of 67% as most closely associated with CSS (p<0.001, AUC=0.71). On multivariate analysis after adjusting for NCCN risk group and ADT use, PPC>67% increased the risk for BF (p<0.0001, HR:2.1 [1.4–3.0]), FFM (p<0.05, HR:1.7 [1.1 to 2.9]), and CSS (p<0.06 (HR:2.1 [1.0–4.6]). When analyzed as a continuous variable controlling for risk group and ADT use, increasing PPC increased the risk for BF (p < 0.002), metastasis (p < 0.05), and CSS (p < 0.02), with a 1–2% increase in relative risk of recurrence for each 1% increase in the PPC. Conclusions: For patients treated with dose-escalated RT, the PPC adds prognostic value but at a higher cut-point then previously utilized. Patients with PPC >67% remain at increased risk for failure even with dose-escalated EBRT and may receive benefit from further intensification of therapy. No significant financial relationships to disclose.

scholarly journals Comparison of Serum Creatinine and Cystatin C for Early Diagnosis of Contrast-Induced Nephropathy after Coronary Angiography and Interventions

2012 ◽  
Vol 58 (2) ◽  
pp. 458-464 ◽  
Author(s):  
Flavio Ribichini ◽  
Giovanni Gambaro ◽  
Maria Stella Graziani ◽  
Michele Pighi ◽  
Gabriele Pesarini ◽  
...  
Keyword(s):  
Serum Creatinine ◽  
Diagnostic Accuracy ◽  
Cystatin C ◽  
Roc Curve ◽  
Diagnostic Value ◽  
Roc Curve Analysis ◽  
Contrast Induced Nephropathy ◽  
Early Predictors ◽  
Patients At Risk

Abstract BACKGROUND The diagnostic accuracy of serum creatinine and cystatin C (Cys) as early predictors of contrast-induced nephropathy (CIN) has been debated. We investigated the diagnostic sensitivities, diagnostic specificities, and variations from baseline for serum creatinine and Cys in CIN. METHODS We prospectively evaluated 166 patients at risk for CIN at baseline, and at 12, 24, and 48 h after exposure to contrast media. CIN occurred in 30 patients (18%). Changes (Δ) compared to baseline in serum creatinine and Cys were evaluated at the predefined time points. ROC curve analysis was performed for the Δ 12-h basal serum creatinine and Cys. RESULTS The Δ serum creatinine at 12 h from baseline was the earliest predictor of CIN [area under the ROC curve (AUC) = 0.80; P &lt; 0.001]. The Δ serum creatinine 15% variation [0.15 mg/dL (13.2 μmol/L)] yielded 43% diagnostic sensitivity and 93% diagnostic specificity. The ΔCys at 12 h from baseline performed significantly worse than serum creatinine (AUC = 0.48; P = 0.74). CONCLUSIONS Variations from the serum creatinine baseline offer better diagnostic accuracy for predicting CIN at an earlier stage than similar variations in Cys. An additional diagnostic value of Cys over the determination of serum creatinine in the setting of CIN was not observed.

Sign in / Sign up

Export Citation Format

Share Document