Joint inference about the AUC and Youden index for paired biomarkers

2021 ◽  
Author(s):  
Jingjing Yin ◽  
Hani Samawi ◽  
Lili Tian
Keyword(s):  
2020 ◽  
pp. bjophthalmol-2020-317825
Author(s):  
Yonghao Li ◽  
Weibo Feng ◽  
Xiujuan Zhao ◽  
Bingqian Liu ◽  
Yan Zhang ◽  
...  

Background/aimsTo apply deep learning technology to develop an artificial intelligence (AI) system that can identify vision-threatening conditions in high myopia patients based on optical coherence tomography (OCT) macular images.MethodsIn this cross-sectional, prospective study, a total of 5505 qualified OCT macular images obtained from 1048 high myopia patients admitted to Zhongshan Ophthalmic Centre (ZOC) from 2012 to 2017 were selected for the development of the AI system. The independent test dataset included 412 images obtained from 91 high myopia patients recruited at ZOC from January 2019 to May 2019. We adopted the InceptionResnetV2 architecture to train four independent convolutional neural network (CNN) models to identify the following four vision-threatening conditions in high myopia: retinoschisis, macular hole, retinal detachment and pathological myopic choroidal neovascularisation. Focal Loss was used to address class imbalance, and optimal operating thresholds were determined according to the Youden Index.ResultsIn the independent test dataset, the areas under the receiver operating characteristic curves were high for all conditions (0.961 to 0.999). Our AI system achieved sensitivities equal to or even better than those of retina specialists as well as high specificities (greater than 90%). Moreover, our AI system provided a transparent and interpretable diagnosis with heatmaps.ConclusionsWe used OCT macular images for the development of CNN models to identify vision-threatening conditions in high myopia patients. Our models achieved reliable sensitivities and high specificities, comparable to those of retina specialists and may be applied for large-scale high myopia screening and patient follow-up.


2021 ◽  
pp. jnnp-2021-326914
Author(s):  
Dario Saracino ◽  
Karim Dorgham ◽  
Agnès Camuzat ◽  
Daisy Rinaldi ◽  
Armelle Rametti-Lacroux ◽  
...  

ObjectiveNeurofilament light chain (NfL) is a promising biomarker in genetic frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). We evaluated plasma neurofilament light chain (pNfL) levels in controls, and their longitudinal trajectories in C9orf72 and GRN cohorts from presymptomatic to clinical stages.MethodsWe analysed pNfL using Single Molecule Array (SiMoA) in 668 samples (352 baseline and 316 follow-up) of C9orf72 and GRN patients, presymptomatic carriers (PS) and controls aged between 21 and 83. They were longitudinally evaluated over a period of >2 years, during which four PS became prodromal/symptomatic. Associations between pNfL and clinical–genetic variables, and longitudinal NfL changes, were investigated using generalised and linear mixed-effects models. Optimal cut-offs were determined using the Youden Index.ResultspNfL levels increased with age in controls, from ~5 to~18 pg/mL (p<0.0001), progressing over time (mean annualised rate of change (ARC): +3.9%/year, p<0.0001). Patients displayed higher levels and greater longitudinal progression (ARC: +26.7%, p<0.0001), with gene-specific trajectories. GRN patients had higher levels than C9orf72 (86.21 vs 39.49 pg/mL, p=0.014), and greater progression rates (ARC:+29.3% vs +24.7%; p=0.016). In C9orf72 patients, levels were associated with the phenotype (ALS: 71.76 pg/mL, FTD: 37.16, psychiatric: 15.3; p=0.003) and remarkably lower in slowly progressive patients (24.11, ARC: +2.5%; p=0.05). Mean ARC was +3.2% in PS and +7.3% in prodromal carriers. We proposed gene-specific cut-offs differentiating patients from controls by decades.ConclusionsThis study highlights the importance of gene-specific and age-specific references for clinical and therapeutic trials in genetic FTD/ALS. It supports the usefulness of repeating pNfL measurements and considering ARC as a prognostic marker of disease progression.Trial registration numbersNCT02590276 and NCT04014673.


Author(s):  
Ruka Nakasone ◽  
Kazumichi Fujioka ◽  
Yuki Kyono ◽  
Asumi Yoshida ◽  
Takumi Kido ◽  
...  

To date, the difference in neurodevelopmental outcomes between late preterm infants (LPI) born at 34 and 35 gestational weeks (LPI-34 and LPI-35, respectively) has not been elucidated. This retrospective study aimed to evaluate neurodevelopmental outcomes at 18 months of corrected age for LPI-34 and LPI-35, and to elucidate factors predicting neurodevelopmental impairment (NDI). Records of all LPI-34 (n = 93) and LPI-35 (n = 121) admitted to our facility from 2013 to 2017 were reviewed. Patients with congenital or chromosomal anomalies, severe neonatal asphyxia, and without developmental quotient (DQ) data were excluded. Psychomotor development was assessed as a DQ using the Kyoto Scale of Psychological Development at 18 months of corrected age. NDI was defined as DQ < 80 or when severe neurodevelopmental problems made neurodevelopmental assessment impossible. We compared the clinical characteristics and DQ values between LPI-34 (n = 62) and LPI-35 (n = 73). To elucidate the factors predicting NDI at 18 months of corrected age, we compared clinical factors between the NDI (n = 17) and non-NDI (n = 118) groups. No significant difference was observed in DQ values at 18 months of corrected age between the groups in each area and overall. Among clinical factors, male sex, intraventricular hemorrhage (IVH), hyperbilirubinemia, and severe hyperbilirubinemia had a higher prevalence in the NDI group than in the non-NDI group, and IVH and/or severe hyperbilirubinemia showed the highest Youden Index values for predicting NDI. Based on the results of this study, we can conclude that no significant difference in neurodevelopmental outcomes at 18 months of corrected age was observed between LPI-34 and LPI-35. Patients with severe hyperbilirubinemia and/or IVH should be considered to be at high risk for developing NDI.


Author(s):  
Qing Zhang ◽  
Hao-Yang Gao ◽  
Ding Li ◽  
Chang-Sen Bai ◽  
Zheng Li ◽  
...  

Abstract Background Few mortality-scoring models are available for solid tumor patients who are predisposed to develop Escherichia coli–caused bloodstream infection (ECBSI). We aimed to develop a mortality-scoring model by using information from blood culture time to positivity (TTP) and other clinical variables. Methods A cohort of solid tumor patients who were admitted to hospital with ECBSI and received empirical antimicrobial therapy was enrolled. Survivors and non-survivors were compared to identify the risk factors of in-hospital mortality. Univariable and multivariable regression analyses were adopted to identify the mortality-associated predictors. Risk scores were assigned by weighting the regression coefficients with corresponding natural logarithm of the odds ratio for each predictor. Results Solid tumor patients with ECBSI were distributed in the development and validation groups, respectively. Six mortality-associated predictors were identified and included in the scoring model: acute respiratory distress (ARDS), TTP ≤ 8 h, inappropriate antibiotic therapy, blood transfusion, fever ≥ 39 °C, and metastasis. Prognostic scores were categorized into three groups that predicted mortality: low risk (< 10% mortality, 0–1 points), medium risk (10–20% mortality, 2 points), and high risk (> 20% mortality, ≥ 3 points). The TTP-incorporated scoring model showed excellent discrimination and calibration for both groups, with AUC being 0.833 vs 0.844, respectively, and no significant difference in the Hosmer–Lemeshow test (6.709, P = 0.48) and the chi-square test (6.993, P = 0.46). Youden index showed the best cutoff value of ≥ 3 with 76.11% sensitivity and 79.29% specificity. TTP-incorporated scoring model had higher AUC than no TTP-incorporated model (0.837 vs 0.817, P < 0.01). Conclusions Our TTP-incorporated scoring model was associated with improving capability in predicting ECBSI-related mortality. It can be a practical tool for clinicians to identify and manage bacteremic solid tumor patients with high risk of mortality.


2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Hui Zhu ◽  
Xiaoxiao Jin ◽  
Yuqing Xu ◽  
Weihua Zhang ◽  
Xiaodan Liu ◽  
...  

Abstract Background Non-invasive prenatal screening (NIPS) is widely used as the alternative choice for pregnant women at high-risk of fetal aneuploidy. However, whether NIPS has a good detective efficiency for pregnant women at advanced maternal age (AMA) has not been fully studied especially in Chinese women. Methods Twenty-nine thousand three hundred forty-three pregnant women at AMA with singleton pregnancy who received NIPS and followed-up were recruited. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), receiver operating characteristic (ROC) curves and the Youden Index for detecting fetal chromosomal aneuploidies were analyzed. The relationship between maternal age and common fetal chromosomal aneuploidy was observed. Results The sensitivity, specificity, PPV, NPV of NIPS for detecting fetal trisomy 21 were 99.11, 99.96, 90.98, and 100%, respectively. These same parameters for detecting fetal trisomy 18 were 100, 99.94, 67.92, and 100%, respectively. Finally, these parameters for detecting trisomy 13 were 100, 99.96, 27.78, and 100%, respectively. The prevalence of fetal trisomy 21 increased exponentially with maternal age. The high-risk percentage incidence rate of fetal trisomy 21 was significantly higher in the pregnant women at 37 years old or above than that in pregnant women at 35 to 37 years old. (Youden index = 37). Conclusion It is indicated that NIPS is an effective prenatal screening method for pregnant women at AMA.


2020 ◽  
Vol 79 (Suppl 1) ◽  
pp. 404.1-404
Author(s):  
V. Hax ◽  
R. Santo ◽  
L. Santos ◽  
M. Farinon ◽  
M. De Oliveira ◽  
...  

Background:Because the method of diagnosing sarcopenia is complex and is considered to be difficult to introduce into routine practice, the European Working Group on Sarcopenia in Older People’s (EWGSOP) recommends use of the SARC-F questionnaire as a way to introduce assessment and treatment of sarcopenia into clinical practice1. Only recently, some studies have focused their attention on the presence of sarcopenia in systemic sclerosis (SSc) and there is no data about the performance of SARC-F in this population.Objectives:To test the diagnostic properties of the SARC-F questionnaire for sarcopenia screening in SSc patients.Methods:Cross-sectional study, including 94 SSc patients assessed by clinical evaluation, laboratory and pulmonary function tests. Sarcopenia was evaluated using the EWGSOP diagnostic criteria updated in 2019 (EWGSOP2): dual-energy X-ray absorptiometry, handgrip strength, and short physical performance battery (SPPB)1. Participants also completed the SARC-F questionnaire. The questionnaires’ performances were evaluated through receiver operating characteristic (ROC) curves and standard measures of diagnostic accuracy were computed using the EWGSOP2 criteria as the gold standard for diagnosis of sarcopenia.Results:Sarcopenia was identified in 15 (15,9%) SSc patients by the EWGSOP2 criteria. Area under the ROC curve of SARC-F screening for sarcopenia was 0.588 (95% confidence interval (CI) 0.482, 0.688) (figure 1). The results of sensitivity, specificity, positive likelihood ratio (PLR) and negative likelihood ratio (NLR) with the EWGSOP2 criteria as the reference standard were 35.71 [95% CI, 12.76-64.86], 81.01 (95% CI, 70.62-88.97), 1.88 (95% CI, 0.81-4.35) and 0.79 (95% CI, 0.53-1.19), respectively. The optimal cut-off point of SARC-F in our sample was ≥ 4 (Youden index: 0.21), the same cut-off point recommended in the literature2,3. Only 6 (40%) out of the 15 participants with sarcopenia were identified by the SARC-F questionnaire in our population. However, the SARC-F properly identified 4 out of 5 patients who had severe sarcopenia.Conclusion:This is the first study to evaluate the performance of SARC-F questionnaire for sarcopenia screening in patients with SSc. Although it appropriately identifies severe cases of sarcopenia, the SARC-F alone may not be an adequate screening tool in high-risk populations, such as SSc, that may benefit from early intervention and treatment.References:[1]Cruz-Jentoft AJ, Bahat G, Bauer J, et al. Sarcopenia: Revised European consensus on definition and diagnosis. Age Ageing. 2019;48(1):16-31.[2]Malmstrom TK, Morley JE. SARC-F: A simple questionnaire to rapidly diagnose sarcopenia. J Am Med Dir Assoc. 2013;14(8):531-532.[3]Ida S, Kaneko R, Murata K. SARC-F for Screening of Sarcopenia Among Older Adults: A Meta-analysis of Screening Test Accuracy. J Am Med Dir Assoc. 2018;19(8):685-689.Disclosure of Interests:Vanessa Hax: None declared, Rafaela Santo: None declared, Leonardo Santos: None declared, Mirian Farinon: None declared, Marianne de Oliveira: None declared, Guilherme Levi Três: None declared, Andrese Aline Gasparin: None declared, M Bredemeier: None declared, Ricardo Xavier Consultant of: AbbVie, Pfizer, Novartis, Janssen, Eli Lilly, Roche, Rafael Mendonça da Silva Chakr: None declared


2021 ◽  
Vol 10 (5) ◽  
pp. 949
Author(s):  
Bartosz Szmyd ◽  
Małgorzata Biedrzycka ◽  
Filip Franciszek Karuga ◽  
Magdalena Rogut ◽  
Iwona Strzelecka ◽  
...  

Serious complications in both mother and newborn arising as a result of fetal macrosomia indicate the need for early diagnosis and prevention. Unfortunately, current predictors, such as fetal biometry, fundal height, and amniotic fluid index, appear to be insufficient. Therefore, we decided to assess the predictive potential of interventricular septal thickness (IVST), as measured at ≥33 weeks of gestation. Two hundred and ninety-nine patients met the inclusion criteria: complete medical history including all necessary measurements—namely, IVST obtained by M-mode echocardiography, fetal biometry, and birth weight. The Statistica 13.1 PL software was used to generate the receiver operating curve. The optimal cut-off point (IVST of 4.7 mm) was selected using the Youden index method. The analysis of fetal biometry abnormalities resulted in 46.6% of macrosomia cases being correctly predicted; however, IVST analysis detected 71.4% of cases. IVST at ≥4.7 mm appears to have a higher sensitivity and negative predictive value (NPV) than routine ultrasound.


Author(s):  
Luma Cordeiro Rodrigues ◽  
Silvia Ferrite ◽  
Ana Paula Corona

Abstract Purpose This article investigates the validity of a smartphone-based audiometry for hearing screening to identify hearing loss in workers exposed to noise. Research Design This is a validation study comparing hearing screening with the hearTest to conventional audiometry. The study population included all workers who attended the Brazilian Social Service of Industry to undergo periodic examinations. Sensitivity, specificity, the Youden index, and positive (PPV) and negative predictive values (NPV) for hearing screening obtained by the hearTest were estimated according to three definitions of hearing loss: any threshold greater than 25 dB hearing level (HL), the mean auditory thresholds for 0.5, 1, 2, and 4 kHz greater than 25 dB HL, and the mean thresholds for 3, 4, and 6 kHz greater than 25 dB HL. Note that 95% confidence intervals were calculated for all measurements. Results A total of 232 workers participated in the study. Hearing screening with the hearTest presented good sensitivity (93.8%), specificity (83.9%), and Youden index (77.7%) values, a NPV (97.2%), and a low PPV (69.0%) for the identification of hearing loss defined as any auditory threshold greater than 25 dB HL. For the other definitions of hearing loss, we observed high specificity, PPV and NPV, as well as low sensitivity and Youden index. Conclusion The hearTest is an accurate hearing screening tool to identify hearing loss in workers exposed to noise, including those with noise-induced hearing loss, although it does not replace conventional audiometry.


Sign in / Sign up

Export Citation Format

Share Document