Range Of Cutpoints Research Articles

Predicting risk of suicidal behavior from insurance claims data vs. linked data from insurance claims and electronic health records.

Observational studies assessing effects of medical products on suicidal behavior often rely on health record data to account for pre-existing risk. We assess whether high-dimensional models predicting suicide risk using data derived from insurance claims and electronic health records (EHRs) are superior to models using data from insurance claims alone. Data were from seven large health systems identified outpatient mental health visits by patients aged 11 or older between 1/1/2009 and 9/30/2017. Data for the 5 years prior to each visit identified potential predictors of suicidal behavior typically available from insurance claims (e.g., mental health diagnoses, procedure codes, medication dispensings) and additional potential predictors available from EHRs (self-reported race and ethnicity, responses to Patient Health Questionnaire or PHQ-9 depression questionnaires). Nonfatal self-harm events following each visit were identified from insurance claims data and fatal self-harm events were identified by linkage to state mortality records. Random forest models predicting nonfatal or fatal self-harm over 90 days following each visit were developed in a 70% random sample of visits and validated in a held-out sample of 30%. Performance of models using linked claims and EHR data was compared to models using claims data only. Among 15 845 047 encounters by 1 574 612 patients, 99 098 (0.6%) were followed by a self-harm event within 90 days. Overall classification performance did not differ between the best-fitting model using all data (area under the receiver operating curve or AUC = 0.846, 95% CI 0.839-0.854) and the best-fitting model limited to data available from insurance claims (AUC = 0.846, 95% CI 0.838-0.853). Competing models showed similar classification performance across a range of cut-points and similar calibration performance across a range of risk strata. Results were similar when the sample was limited to health systems and time periods where PHQ-9 depression questionnaires were recorded more frequently. Investigators using health record data to account for pre-existing risk in observational studies of suicidal behavior need not limit that research to databases including linked EHR data.

Predictive ability of the Desire to Avoid Pregnancy scale

BackgroundA longstanding gap in the reproductive health field has been the availability of a screening instrument that can reliably predict a person’s likelihood of becoming pregnant. The Desire to Avoid Pregnancy Scale is a new measure; understanding its sensitivity and specificity as a screening tool for pregnancy as well as its predictive ability and how this varies by socio-demographic factors is important to inform its implementation.MethodsThis analysis was conducted on a cohort of 994 non-pregnant participants recruited in October 2018 and followed up for one year. The cohort was recruited using social media as well as advertisements in a university, school, abortion clinic and outreach sexual health service. Almost 90% of eligible participants completed follow-up at 12 months; those lost to follow-up were not significantly different on key socio-demographic factors. We used baseline DAP score and a binary variable of whether participants experienced pregnancy during the study to assess the sensitivity, specificity, area under the ROC curve (AUROC) and positive and negative predictive values (PPV and NPV) of the DAP at a range of cut-points. We also examined how the predictive ability of the DAP varied according to socio-demographic factors and by the time frame considered (e.g., pregnancy within 3, 6, 9 and 12 months).ResultsAt a cut-point of 2 on the 0–4 range of the DAP scale, the DAP had a sensitivity of 0.78, a specificity of 0.81 and an excellent AUROC of 0.87. In this sample the cumulative incidence of pregnancy was 16% (95%CI 13%, 18%) making the PPV 43% and the NPV 95% at this cut-point. The DAP score was the factor most strongly associated with pregnancy, even after age and number of children were taken into account. The association between baseline DAP score and pregnancy did not differ across time frames.ConclusionsThis is the first study to assess the DAP scale as a screening tool and shows that its predictive ability is superior to the limited pre-existing pregnancy prediction tools. Based on our findings, the DAP could be used with a cut-point selected according to the purpose.

Improving population-level refractive error monitoring via mixture distributions.

Sampling and describing the distribution of refractive error in populations is critical to understanding eye care needs, refractive differences between groups and factors affecting refractive development. We investigated the ability of mixture models to describe refractive error distributions. We used key informants to identify raw refractive error datasets and a systematic search strategy to identify published binned datasets of community-representative refractive error. Mixture models combine various component distributions via weighting to describe an observed distribution. We modelled raw refractive error data with a single-Gaussian (normal) distribution, mixtures of two to six Gaussian distributions and an additive model of an exponential and Gaussian (ex-Gaussian) distribution. We tested the relative fitting accuracy of each method via Bayesian Information Criterion (BIC) and then compared the ability of selected models to predict the observed prevalence of refractive error across a range of cut-points for both the raw and binned refractive data. We obtained large raw refractive error datasets from the United States and Korea. The ability of our models to fit the data improved significantly from a single-Gaussian to a two-Gaussian-component additive model and then remained stable with ≥3-Gaussian-component mixture models. Means and standard deviations for BIC relative to 1 for the single-Gaussian model, where lower is better, were 0.89 ± 0.05, 0.88 ± 0.06, 0.89 ± 0.06, 0.89 ± 0.06 and 0.90 ± 0.06 for two-, three-, four-, five- and six-Gaussian-component models, respectively, tested across US and Korean raw data grouped by age decade. Means and standard deviations for the difference between observed and model-based estimates of refractive error prevalence across a range of cut-points for the raw data were -3.0% ± 6.3, 0.5% ± 1.9, 0.6% ± 1.5 and -1.8% ± 4.0 for one-, two- and three-Gaussian-component and ex-Gaussian models, respectively. Mixture models appear able to describe the population distribution of refractive error accurately, offering significant advantages over commonly quoted simple summary statistics such as mean, standard deviation and prevalence.

Prognostic value of plasma circulating tumor DNA fraction across four common cancer types: a real-world outcomes study

Genomic analysis of circulating tumor DNA (ctDNA) is increasingly incorporated into the clinical management of patients with advanced cancer. Beyond tumor profiling, ctDNA analysis also can enable calculation of circulating tumor fraction (TF), which has previously been found to be prognostic. While most prognostic models in metastatic cancer are tumor type specific and require significant patient-level data, quantification of TF in ctDNA has the potential to serve as a pragmatic, tumor-agnostic prognostic tool. This study utilized a cohort of patients in a nationwide de-identified clinico-genomic database with metastatic castration-resistant prostate cancer (mCRPC), metastatic breast cancer (mBC), advanced non-small-cell lung cancer (aNSCLC), or metastatic colorectal cancer (mCRC) undergoing liquid biopsy testing as part of routine care. TF was calculated based on single-nucleotide polymorphism aneuploidy across the genome. Clinical, disease, laboratory, and treatment data were captured from the electronic health record. Overall survival (OS) was evaluated by TF level while controlling for relevant covariables. A total of 1725 patients were included: 198 mCRPC, 402 mBC, 902 aNSCLC, and 223 mCRC. TF ≥10% was highly correlated with OS in univariable analyses for all cancer types: mCRPC [hazard ratio (HR) 3.3, 95% confidence interval (CI) 2.04-5.34, P < 0.001], mBC (HR 2.4, 95% CI 1.71-3.37, P < 0.001), aNSCLC (HR 1.68, 95% CI 1.34-2.1, P < 0.001), and mCRC (HR 2.11, 95% CI 1.39-3.2, P < 0.001). Multivariable assessments of TF had similar point estimates and CIs, suggesting a consistent and independent association with survival. Exploratory analysis showed that TF remained consistently prognostic across a wide range of cutpoints. Plasma ctDNA TF is a pragmatic, independent prognostic biomarker across four advanced cancers with potential to guide clinical conversations around expected treatment outcomes. With further prospective validation, ctDNA TF could be incorporated into care paradigms to enable precision escalation and de-escalation of cancer therapy based on patient-level tumor biology.

Traumatic Brain Injury: What Is a Favorable Outcome?

Traumatic brain injury (TBI) results in disparate outcomes ranging from persistent disorders of consciousness to symptom resolution. Despite the breadth and complexity of TBI recovery, most clinical trials dichotomize outcome by establishing an arbitrary cut-point, above and below which recovery is described as "favorable" and "unfavorable," respectively. For example, the widely used eight-level Glasgow Outcome Scale-Extended (GOSE) is typically collapsed into these two categories. Dichotomizing the GOSE into "favorable" and "unfavorable" outcome may limit detection of treatment effects in TBI clinical trials, contribute to imprecise prognostic counseling, and unduly influence decision-making with regard to withdrawal of life-sustaining therapy. We illustrate the lack of standardization in defining "unfavorable" and "favorable" TBI outcome on the GOSE by identifying the broad range of cut-points, from a score of 3 (part-time supervision in the home required) to 7 (presence of some residual of symptoms), that have been used to dichotomize the GOSE. We also highlight the ethical concerns related to characterizing TBI outcomes solely from the perspective of investigators and clinicians, rather than patients and caregivers. Finally, we suggest that a pragmatic, immediate solution to GOSE dichotomization is to report the likelihood of achieving each of the eight GOSE outcome levels and propose a study design for a new patient- and caregiver-centered TBI outcome metric.

Abstract PS5-02: Assessment of early ctDNA dynamics to predict efficacy of targeted therapies in metastatic breast cancer: Results from plasmaMATCH trial

Abstract Background: Early changes in circulating tumour DNA (ctDNA) levels may identify which patients respond to therapy earlier than imaging, with ctDNA levels falling rapidly in patients who respond to therapy. The plasmaMATCH trial assessed the utility of ctDNA testing with an error-corrected 73-gene targeted panel (Guardant360, Guardant Health) to allocate patients to four mutation matched therapy cohorts. ESR1-extended fulvestrant (A), HER2-neratinib +/- fulvestrant (B), AKT1-capivasertib + fulvestrant (C), AKT basket-capivasertib (D). Here, we report paired baseline and early on treatment ctDNA analysis from plasmaMATCH, to establish the optimal criteria for predicting progression free survival (PFS). Methods: In plasmaMATCH treatment cohorts, plasma samples were collected for ctDNA analysis pre-treatment at cycle 1-day 1 (C1D1) and cycle 2-day 1 (C2D1) timepoints, and sequenced with the Guardant 360 assay. Patients were included if they had a minimum of 14 days of treatment in the first cycle. Multiple different methods were investigated to integrate variant allele fractions (VAF) of mutations identified at each timepoint to estimate the level of ctDNA, including maximum VAF, mean VAF and weighted mean VAF, and weighted mean VAF of clonal mutations at C1D1. Variants with a VAF &amp;lt;0.3%, set as the limit of detection, in C1D1 were excluded. Genes frequently mutated in CHIP were excluded (GNAS, JAK2, IDH1, IDH2 and ATM) from the weighted mean VAF of clonal mutations method. The circulating DNA ratio (CDR) was calculated as the ratio of C2D1 level relative to C1D1 level. The optimal cut-point for predicting PFS was assessed by fitting a range of cutpoints for each VAF integration method, identifying the cut-point with the highest Harrell’s C-index. Results: A total of 142 patients were enrolled into plasmaMATCH cohorts A-D, 79 patients had samples sent for paired C1D1-C2D1 plasma ctDNA sequencing, 1 failed sequencing and 1 insufficient treatment, and 77 (54%) patients had assessable C1D1-C2D1 plasma ctDNA sequencing results (45 cohort A, 12 cohort B, 12 cohort C, 8 cohort D). A weighted mean of clonal mutations in C1D1 ctDNA sequencing was the optimal method for integrating VAF, with peak C-Index 0.67. At the optimal C-index cutoff of 0.132, median PFS with high ctDNA CDR was 2.4 months (95% CI 2.0-3.7) and with suppressed ctDNA CDR was 9.9 months (95% CI 7.0-13.7) (HR 4.3, 95% CI 2.4-7.6, p&amp;lt;0.0001). Early changes in ctDNA level were also predictive in cohorts A extended dose fulvestrant alone (HR 5.8, 95% CI 2.2-16, p=0.0001) and cohorts B-D of targeted therapy (HR 3.8, 95% CI 1.7-8.6, p=0.00063). In analysis that was not pre-planned, patients with undetectable ctDNA at C2D1 had a particularly good outcome (p&amp;lt;0.0001, table 1). Conclusions: We identify an optimal methodology for assessing early dynamic changes in ctDNA that predicts treatment efficacy in patients with metastatic breast cancer. This methodology will require validation in independent data-sets, and if validated would allow trials of adapting therapy on the basis of early ctDNA dynamics. Table 1ctDNA dynamics categoryMedian PFS months (95%CI)6-month PFSORRUndetectable (N=11) CDR=018.2 (10.2-NA)91%9/11 (82%)Suppressed (N=14) CDR &amp;lt;0.132 and &amp;gt;05.4 (4.6-NA)48%6/14 (43%)High (N=52) CDR &amp;gt;=0.1322.4 (2.0-3.7)8%4/52 (8%) Citation Format: Javier Pascual, Rosalind J Cutts, Belinda Kingston, Sarah Hrebien, Lucy S Kilburn, Sarah Kernaghan, Laura Moretti, Katie Wilkinson, Andrew M Wardley, Iain R Macpherson, Richard D Baird, Rebecca Roylance, Michael Hubank, Giselle Walsh, Iris Faull, Kimberly C Banks, Richard B Lanman, Isaac Garcia-Murillas, Judith M Bliss, Alistair Ring, Nicholas C Turner. Assessment of early ctDNA dynamics to predict efficacy of targeted therapies in metastatic breast cancer: Results from plasmaMATCH trial [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS5-02.

Calibration and validation of accelerometry to measure physical activity in adult clinical groups: A systematic review

A growing body of research calibrating and validating accelerometers to classify physical activity intensities has led to a range of cut-points. However, the applicability of current calibration protocols to clinical populations remains to be addressed. The aim of this review was to evaluate the accuracy of the methods for calibrating and validating of accelerometers to estimate physical activity intensity thresholds for clinical populations. Six databases were searched between March and July to 2017 using text words and subject headings. Studies developing moderate-to-vigorous intensity physical activity cut-points for adult clinical populations were included. The risk of bias was assessed using the health measurement instruments and a specific checklist for calibration studies. A total of 543,741 titles were found and 323 articles were selected for full-text assessment, with 11 meeting the inclusion criteria. Twenty-three different methods for calibration were identified using different models of ActiGraph and Actical accelerometers. Disease-specific cut-points ranged from 591 to 2717 counts·min-1 and were identified for two main groups of clinical conditions: neuromusculoskeletal disorders and metabolic diseases. The heterogeneity in the available clinical protocols hinders the applicability and comparison of the developed cut-points. As such, a mixed protocol containing a controlled laboratory exercise test and activities of daily-life is suggested. It is recommended that this be combined with a statistical approach that allows for adjustments according to disease severity or the use of machine learning models. Finally, this review highlights the generalisation of cut-points developed on healthy populations to clinical populations is inappropriate.

Methodology to Derive Objective Screen-State from Smartphones: A SMART Platform Study.

Time on screens (screen time) on multiple digital devices (computers, mobile phones, tablets, television screens, etc.) due to varied motivations (work, leisure, entertainment, gaming, etc.) has become an integral part of population behaviour. However, a significant evidence gap exists in screen time accumulated over ubiquitous mobile devices such as smartphones. This study aimed to develop an accurate, reliable and replicable methodology to derive objective screen time (i.e., screen-state) from all types of citizen-owned smartphones. A convenience sample of 538 adults (≥18 years) from two largest urban centres in Saskatchewan, Canada (Regina and Saskatoon) was recruited in 2017 and 2018. Participants used a custom-built smartphone application to provide objective and subjective data. A novel methodology was developed to derive objective screen-state, and these data were compared with subjective measures. The findings showed that objective screen-state from smartphones can be derived and assessed across a range of cut-points that take into consideration varied measurement errors. When objective measures were compared with subjective reporting, the results indicated that participants consistently underreported screen time. This study not only provides a methodology to derive objective screen-state from ubiquitous mobile devices such as smartphones but also emphasises the need to capture context via subjective measures.

Clinical and analytical validation of Ki-67 in 9069 patients from IBCSG VIII + IX, BIG1-98 and GeparTrio trial: systematic modulation of interobserver variance in a comprehensive in silico ring trial.

Ki-67 has been clinically validated for risk assessment in breast cancer, but the analytical validation and cutpoint-definition remain a challenge. Intraclass correlation coefficients (ICCs) are a statistical parameter for Ki-67 interobserver performance. However, the maximum degree of variance among pathologists allowed for meaningful biomarker results has not been defined. Different amounts of variance were added to central pathology Ki-67 data (n = 9069) from three cohorts (IBCSGVIII + IX, BIG1-98, GeparTrio) by simulation of 4500 evaluations for each cohort, which were grouped by ICCs, ranging from excellent (ICC = 0.9) to poor concordance (ICC = 0.1). Endpoints were disease-free survival (DFS) and pathological complete response (pCR, GeparTrio). Ki-67 was a significant continuous prognostic marker for DFS over a wide range of cutpoints between 8% and 30% in all three cohorts. In our modelling approach, Ki-67 was a stable prognostic marker despite increased interpathologist variance. Even for a poor ICC of 0.5, one or more significant Ki-67 cutoffs were detected in 86.8% (GeparTrio), 92.4% (IBCSGVIII + IX) and 100% of analyses (BIG1-98). Similarly, in GeparTrio, even with an extremely low ICC of 0.2, 99.6% of analyses were significant for pCR. Our study shows that Ki-67 is a continuous marker which is extremely robust to pathologist variation. Even if only 50% of variance is attributable to true Ki-67-based proliferation (ICC = 0.5), this information is sufficient to obtain statistically significant differences in clinical cohorts. This stable performance explains the observation that many Ki-67 studies achieve significant results despite relevant interobserver variance and points to a high clinical validity of this biomarker. For clinical decisions based on analysis of individual patient data, ongoing efforts to further reduce interobserver variability, including ring trials and standardized guidelines as well as image analysis approaches, should be continued.

Abstract TP267: Performance of Components of the Canadian TIA Score to Predict Acute InfarctionAmong Patients Managed in an Emergency Department Observation Unit

Introduction: TIA risk stratification can help inform of impending stroke, but some scores lack sensitivity while others are impractical in the emergency department (ED) setting. The Canadian TIA Score has shown good discriminative ability at predicting early stroke recurrence and is designed to be applied in the ED, but awaits further validation. Objective: We assessed the performance components of the Canadian TIA Score in predicting diffusion weighted MRI (DWI) abnormalities and 30-day adverse events in a cohort of clinically suspected TIAs placed in an ED observation unit (OU). Methods: Patients in a large, urban, academic ED with suspected TIA deemed appropriate for the OU from 4/2013-7/2018 were included. Rates of acute infarct on DWI and adverse 30-day events were assessed. Logistic regression was performed to determine the odds of DWI-confirmed stroke from 12 of the 13 items on the weighted Canadian TIA Score (we did not collect platelet count); and for poor versus good health outcomes at 30-days. The AUC with 95% CIs were also calculated for significant models, and the The Youden index (J statistic) was determined to assess maximum effectiveness of diagnostic test across a range of cut-points. Results: Of 1208 patients admitted over the time period, 1097 had DWI performed (90.1%). Clinical features are described in Table 1 . The logistic regression model for the Canadian TIA Score predicting acute stroke on DWI was significant (Wald χ2 (1) = 36.6, p &lt; 0.001), with an odds ratio of 1.27 (95%CI: 1.16, 1.33), and AUC = 0.65 (95%CI: 0.61, 0.69). In predicting acute stroke, the maximum J value = 0.20; Canadian rule score ≥ 4; sensitivity = 0.68, specificity = 0.52. Of 593 patients reached at 30-days, 510 (86%) reported no adverse events. Of the 83 who did not, 1 had a disabling stroke and 1 had died. Discussion: In this overall low-risk cohort of ED patients with suspected TIA managed in an OU, the Canadian TIA Score performed reasonably well at predicting acute DWI abnormalities.

Coffee consumption and calcified atherosclerotic plaques in the coronary arteries: The NHLBI Family Heart Study

While a recent meta-analysis of prospective studies reported that coffee consumption is associated with a lower risk of cardiovascular disease mortality, limited and inconsistent data are available on the relation of coffee intake with subclinical disease. Thus, the aim of the present study was to see the association of coffee consumption with the prevalence of atherosclerotic plaque in the coronary arteries in NHLBI Family Heart Study. In a cross-sectional design, we studied 1929 participants of the NHLBI Family Heart Study without known coronary heart disease. Coffee consumption was assessed by a semi-quantitative food frequency questionnaire and coronary-artery calcium (CAC) was measured by cardiac computed tomography. We defined prevalent CAC as an Agatston score of ≥100 and used generalized estimating equations to calculate prevalence ratios of CAC as well as a sensitivity analysis at a range of cutpoints for CAC. Mean age was 56.7 years and 59% of the study subjects were female. In adjusted analysis for age, sex, BMI, smoking, alcohol, physical activity, field center, and energy intake, prevalence ratio (95% CI) for CAC was 1.0 (reference), 0.92 (0.57-1.49), 1.34 (0.86-2.08), 1.30 (0.84-2.02), and 0.99 (0.60-1.64) for coffee consumption of almost never, <1/day, 1/day, 2-3/day, and ≥4 cups/day, respectively. In a sensitivity analysis, there was no evidence of association between coffee consumption and prevalent CAC when CAC cut points of 0, 50, 150, 200, and 300 were used. These data do not provide evidence for an association between coffee consumption and prevalent CAC in adult men and women.

Moving Forward with Backward Compatibility: Translating Wrist Accelerometer Data.

This study aimed to provide a means for calibrating raw acceleration data from wrist-worn accelerometers in relation to past estimates of children's moderate-to-vigorous physical activity (MVPA) from a range of cut points applied to hip-worn ActiGraph data. This is a secondary analysis of three studies with concurrent 7-d accelerometer wear at the wrist (GENEActiv) and hip (ActiGraph) in 238 children age 9-12 yr. The time spent above acceleration (ENMO) thresholds of 100, 150, 200, 250, 300, 350, and 400 mg from wrist acceleration data (≤5-s epoch) was calculated for comparison with MVPA estimated from widely used children's hip-worn ActiGraph MVPA cut points (Freedson/Trost, 1100 counts per minute; Pate, 1680 counts per minute; Evenson, 2296 counts per minute; Puyau, 3200 counts per minute) with epochs of ≤5, 15, and 60 s. The optimal ENMO thresholds for alignment with MVPA estimates from ActiGraph cut points determined from 70% of the sample and cross validated with the remaining 30% were as follows: Freedson/Trost = ENMO 150+ mg, irrespective of ActiGraph epoch (intraclass correlation [ICC] ≥ 0.65); Pate = ENMO 200+ mg, irrespective of ActiGraph epoch (ICC ≥ 0.67); Evenson = ENMO 250+ mg for ≤5- and 15-s epochs (ICC ≥ 0.69) and ENMO 300+ mg for 60-s epochs (ICC = 0.73); Puyau = ENMO 300+ mg for ≤5-s epochs (ICC = 0.73), ENMO 350+ mg for 15-s epochs (ICC = 0.73), and ENMO 400+ mg for 60-s epochs (ICC = 0.65). Agreement was robust with cross-validation ICC = 0.62-0.71 and means within ∣7.8∣% ± 4.9% of MVPA estimates from ActiGraph cut points, except Puyau 60-s epochs (ICC = 0.42). Incremental ENMO thresholds enable children's acceleration data measured at the wrist to be simply and directly compared, at a group level, with past estimates of MVPA from hip-worn ActiGraphs across a range of cut points.

Effects of Varying Case Definition on Carpal Tunnel Syndrome Prevalence Estimates in a Pooled Cohort

ObjectiveTo analyze differences in carpal tunnel syndrome (CTS) prevalence using a combination of electrodiagnostic studies (EDSs) and symptoms using EDS criteria varied across a range of cutpoints and compared with symptoms in both ≥1 and ≥2 median nerve–served digits. DesignPooled data from 5 prospective cohorts. SettingHand-intensive industrial settings, including manufacturing, assembly, production, service, construction, and health care. ParticipantsEmployed, working-age participants who are able to provide consent and undergo EDS testing (N=3130). InterventionsNone. Main Outcome MeasuresCTS prevalence was estimated while varying the thresholds for median sensory latency, median motor latency, and transcarpal delta latency difference. EDS criteria examined included the following: median sensory latency of 3.3 to 4.1 milliseconds, median motor latency of 4.1 to 4.9 milliseconds, and median-ulnar sensory difference of 0.4 to 1.2 milliseconds. EDS criteria were combined with symptoms in ≥1 or ≥2 median nerve–served digits. EDS criteria from other published studies were applied to allow for comparison. ResultsCTS prevalence ranged from 6.3% to 11.7%. CTS prevalence estimates changed most per millisecond of sensory latency compared with motor latency or transcarpal delta. CTS prevalence decreased by 0.9% to 2.0% if the criteria required symptoms in 2 digits instead of 1. ConclusionsThere are meaningful differences in CTS prevalence when different EDS criteria are applied. The digital sensory latency criteria result in the largest variance in prevalence.

Exploring medical diagnostic performance using interactive, multi-parameter sourced receiver operating characteristic scatter plots

Determining diagnostic criteria for specific disorders is often a tedious task that involves determining optimal diagnostic thresholds for symptoms and biomarkers using receiver-operating characteristic (ROC) statistics. To help this endeavor, we developed softROC, a user-friendly graphic-based tool that lets users visually explore possible ROC tradeoffs. The software requires MATLAB installation and an Excel file containing threshold symptoms/biological measures, with corresponding gold standard diagnoses for a set of patients. The software scans the input file for diagnostic and symptom/biomarkers columns, and populates the graphical-user-interface (GUI). Users select symptoms/biomarkers of interest using Boolean algebra as potential inputs to create diagnostic criteria outputs. The software evaluates subtests across the user-established range of cut-points and compares them to a gold standard in order to generate ROC and quality ROC scatter plots. These plots can be examined interactively to find optimal cut-points of interest for a given application (e.g. sensitivity versus specificity needs). Split-set validation can also be used to set up criteria and validate these in independent samples. Bootstrapping is used to produce confidence intervals. Additional statistics and measures are provided, such as the area under the ROC curve (AUC). As a testing set, softROC is used to investigate nocturnal polysomnogram measures as diagnostic features for narcolepsy. All measures can be outputted to a text file for offline analysis. The softROC toolbox, with clinical training data and tutorial instruction manual, is provided as supplementary material and can be obtained online at http://www.stanford.edu/~hyatt4/software/softroc or from the open source repository at http://www.github.com/informaton/softroc.

Ki67 levels as predictive and prognostic parameters in pretherapeutic breast cancer core biopsies: a translational investigation in the neoadjuvant GeparTrio trial

BackgroundThe proliferation marker Ki67 has been suggested as a promising cancer biomarker. As Ki67 needs an exact quantification, this marker is a prototype of a new generation of tissue-based biomarkers. In this study, we have systematically evaluated different cut points for Ki67 using three different clinical end points in a large neoadjuvant study cohort. Patients and methodsWe have evaluated pretherapeutic Ki67 levels by immunohistochemistry in 1166 breast cancer core biopsies from the neoadjuvant GeparTrio trial. We used the standardized cutoff-finder algorithm for three end points [response to neoadjuvant chemotherapy (pCR), disease-free (DFS) and overall-survival (OS)]. The analyses were stratified for hormone receptor (HR) and HER2 status by molecular subtype radar diagrams (MSRDs). ResultsA wide range of Ki67 cut points between 3%–94% (for pCR), 6%–46% (for DFS) and 4%–58% (for OS) were significant. The three groups of Ki67 ≤ 15% versus 15.1%–35% versus >35% had pCR-rates of 4.2%, 12.8%, and 29.0% (P < 0.0005), this effect was also present in six of eight molecular subtypes. In MSRD, Ki67 was significantly linked to prognosis in uni- and multivariate analysis in the complete cohort and in HR-positive, but not triple-negative tumors. ConclusionsKi67 is a significant predictive and prognostic marker over a wide range of cut points suggesting that data-derived cut point optimization might not be possible. Ki67 could be used as a continuous marker; in addition, the scientific community could define standardized cut points for Ki67. Our analysis explains the variability observed for Ki67 cut points in previous studies; however, this should not be seen as weakness, but as strength of this marker. MSRDs are an easy new approach for visualization of biomarker effects on outcome across molecular subtypes in breast cancer. The experience with Ki67 could provide important information regarding the development and implementation of other quantitative biomarkers.

Operationalizing hippocampal volume as an enrichment biomarker for amnestic mild cognitive impairment trials: effect of algorithm, test-retest variability, and cut point on trial cost, duration, and sample size

The objective of this study was to evaluate the effect of computational algorithm, measurement variability, and cut point on hippocampal volume (HCV)-based patient selection for clinical trials in mild cognitive impairment (MCI). We used normal control and amnestic MCI subjects from the Alzheimer's Disease Neuroimaging Initiative 1 (ADNI-1) as normative reference and screening cohorts. We evaluated the enrichment performance of 4 widely used hippocampal segmentation algorithms (FreeSurfer, Hippocampus Multi-Atlas Propagation and Segmentation (HMAPS), Learning Embeddings Atlas Propagation (LEAP), and NeuroQuant) in terms of 2-year changes in Mini-Mental State Examination (MMSE), Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog), and Clinical Dementia Rating Sum of Boxes (CDR-SB). We modeled the implications for sample size, screen fail rates, and trial cost and duration. HCV based patient selection yielded reduced sample sizes (by ∼40%–60%) and lower trial costs (by ∼30%–40%) across a wide range of cut points. These results provide a guide to the choice of HCV cut point for amnestic MCI clinical trials, allowing an informed tradeoff between statistical and practical considerations.

Dose-Response Relationship Between Tissue Concentrations of UC781 and Explant Infectibility with HIV Type 1 in the RMP-01 Rectal Safety Study

A retrospective correlational analysis of UC781 (0.1, 0.25%) gel pharmacokinetics (PK) and pharmacodynamics (PD) was undertaken using data generated in the RMP-01/MTN-006 Phase 1 rectal safety study of the UC781 microbicide gel, where strong UC781-related inhibition of ex vivo biopsy infectibility (PD) was seen. Precision analysis, linear and logistical correlational methods were applied to model the dose-response relationship. Four analyses of explant virus growth were compared to determine tissue concentrations of UC781 needed to maintain ex vivo virus growth below a range of cut-points. SOFT, a cross-sectional index from a growth curve, and cumulative p24 endpoints were the most precise measurement of ex vivo HIV infection and significantly (p<0.01) correlated with rectal tissue UC781 concentrations. Cut-points reflecting infectibility, ranging from 200 to 1300 p24 pg/ml, provided EC(50,90,95) tissue levels of UC781. A cut-point of 200 p24 pg/ml provided an EC(50) of 2148 UC781 ng/g tissue; a cut-point of 1100 p24 predicted a lower EC(50) of 101 UC781 ng/g. A 30- to 170-fold EC(90):EC(50) ratio was found. Higher p24 cut-points provided more predictive models. Tissue UC781 levels and ex vivo infectibility data were correlated to model dose-response drug efficacy in this small Phase 1 trial. Logistic regression analyses showed EC(50,90,95) values were inversely related to p24 cut-point levels, providing clinically relevant insights into tissue drug concentration necessary for ex vivo suppression of HIV tissue infectibility. This first PK-PD assessment of topical microbicides demonstrates feasibility in Phase 1 trials, enabling comparisons of microbicide efficacy (i.e., EC(50,90,95)) between formulations, compartments, and application methods. (ClinicalTrials.gov; #NCT00408538).

Risk Terrain Modeling: Brokering Criminological Theory and GIS Methods for Crime Forecasting

The research presented here has two key objectives. The first is to apply risk terrain modeling (RTM) to forecast the crime of shootings. The risk terrain maps that were produced from RTM use a range of contextual information relevant to the opportunity structure of shootings to estimate risks of future shootings as they are distributed throughout a geography. The second objective was to test the predictive power of the risk terrain maps over two six‐month time periods, and to compare them against the predictive ability of retrospective hot spot maps. Results suggest that risk terrains provide a statistically significant forecast of future shootings across a range of cut points and are substantially more accurate than retrospective hot spot mapping. In addition, risk terrain maps produce information that can be operationalized by police administrators easily and efficiently, such as for directing police patrols to coalesced high‐risk areas.

Sex Differences in Performance of Fecal Occult Blood Testing

Immunological and guaiac-based fecal occult blood tests (iFOBTs and gFOBTs) are widely used for early detection of colorectal cancer (CRC). We aimed to assess potential sex differences in performance of iFOBTs and gFOBT in the screening setting. The sensitivity, specificity, and positive and negative predictive values for detection of advanced colorectal neoplasms (CRC or advanced adenoma) were assessed by sex for a range of cutpoints of a quantitative iFOBT (RIDASCREEN Hemoglobin), six qualitative iFOBTs, and a gFOBT (HemOccult) among 1,157 male and 1,167 female participants of the German screening colonoscopy program (mean age: 63.0 and 62.0 years, respectively). The prevalence of advanced colorectal neoplasms was much higher among men (13.5%) than among women (7.5%). At any cutpoint of the quantitative iFOBT, and for all qualitative iFOBTs and the gFOBT, the sensitivity and positive predictive value were substantially higher, and specificity and negative predictive value were substantially lower among men than among women. At the cutpoint of 2 μg/g stool given by the manufacturer of the quantitative iFOBT, sensitivity, specificity, and the positive and negative predictive values were 47.6, 85.0, 33.0, and 91.3% among men, and 30.7, 89.5, 19.3, and 94.1% among women (P value for sex differences 0.001, 0.002, 0.004, and 0.019, respectively). Sex differences in predictive values were mostly explained by sex differences in the prevalence of advanced colorectal neoplasms. There are major sex differences in the performance of fecal occult blood testing, which might require careful attention in the interpretation of test results, and in the design, modeling, and evaluation of CRC screening strategies.

Prediction of Early Readmission in Medical Inpatients Using the Probability of Repeated Admission Instrument

In the absence of an instrument to predict risk of early readmission, examination of the well-validated probability of repeated admission (Pra) for this new purpose is indicated. The objective of this study was to examine the use of the Pra in accurately identifying and predicting adult medical inpatients at risk of early readmission. Over 20 months, 1,077 consecutively admitted medical patients were enrolled in this prospective cohort study at a Midwestern tertiary care medical center. Pra score values were calculated within 2 days of discharge. Databases at the index medical center and other institutions were queried to identify readmission within 41 days. Prevalence of readmission was 14% (confidence interval = 12.4%-15.6%). Pra score values ranged from .16 to .75. Indices to identify and predict readmission for a range of cut points were reported to minimize loss of information. The likelihood ratio for patients with a Pra score value > or = .53 was 1.67. Using a Pra cut point of > or = .45, readmission of patients with a high Pra was 2.3 times more likely than that of patients with a low Pra (p < .001, confidence interval = 1.63-3.27). Comparisons between cohorts indicated that differences existed with four of the eight variables used to calculate the Pra score: diabetes (p = .01), self-rated health status (p = .007), and number of doctor visits (p < .001) and hospitalizations (p < .001) in the past year. Within this heterogeneous sample, prediction of readmission using the Pra was better than chance. These findings may facilitate development of a better predictive model by combining select Pra variables with other variables associated with early readmission.