Effect of training-sample size and classification difficulty on the accuracy of genomic predictors.

Introduction: As part of the MicroArray Quality Control (MAQC)-II project, this analysis examines how the choice of univariate feature-selection methods and classification algorithms may influence the performance of genomic predictors under varying degrees of prediction difficulty represented by three clinically relevant endpoints. Methods: We used gene-expression data from 230 breast cancers (grouped into training and independent validation sets), and we examined 40 predictors (five univariate feature-selection methods combined with eight different classifiers) for each of the three endpoints. Their classification performance was estimated on the training set by using two different resampling methods and compared with the accuracy observed in the independent validation set. Results: A ranking of the three classification problems was obtained, and the performance of 120 models was estimated and assessed on an independent validation set. The bootstrapping estimates were closer to the validation performance than were the cross-validation estimates. The required sample size for each endpoint was estimated, and both gene-level and pathway-level analyses were performed on the obtained models. Conclusions: We showed that genomic predictor accuracy is determined largely by an interplay between sample size and classification difficulty. Variations on univariate feature-selection methods and choice of classification algorithm have only a modest impact on predictor performance, and several statistically equally good predictors can be developed for any given classification problem.

Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients.

The purpose of this study was to determine the prognostic accuracy of perfusion computed tomography (CT), performed at the time of emergency room admission, in acute stroke patients. Accuracy was determined by comparison of perfusion CT with delayed magnetic resonance (MR) and by monitoring the evolution of each patient's clinical condition. Twenty-two acute stroke patients underwent perfusion CT covering four contiguous 10mm slices on admission, as well as delayed MR, performed after a median interval of 3 days after emergency room admission. Eight were treated with thrombolytic agents. Infarct size on the admission perfusion CT was compared with that on the delayed diffusion-weighted (DWI)-MR, chosen as the gold standard. Delayed magnetic resonance angiography and perfusion-weighted MR were used to detect recanalization. A potential recuperation ratio, defined as PRR = penumbra size/(penumbra size + infarct size) on the admission perfusion CT, was compared with the evolution in each patient's clinical condition, defined by the National Institutes of Health Stroke Scale (NIHSS). In the 8 cases with arterial recanalization, the size of the cerebral infarct on the delayed DWI-MR was larger than or equal to that of the infarct on the admission perfusion CT, but smaller than or equal to that of the ischemic lesion on the admission perfusion CT; and the observed improvement in the NIHSS correlated with the PRR (correlation coefficient = 0.833). In the 14 cases with persistent arterial occlusion, infarct size on the delayed DWI-MR correlated with ischemic lesion size on the admission perfusion CT (r = 0.958). In all 22 patients, the admission NIHSS correlated with the size of the ischemic area on the admission perfusion CT (r = 0.627). Based on these findings, we conclude that perfusion CT allows the accurate prediction of the final infarct size and the evaluation of clinical prognosis for acute stroke patients at the time of emergency evaluation. It may also provide information about the extent of the penumbra. Perfusion CT could therefore be a valuable tool in the early management of acute stroke patients.

Sensitivity of predictive species distribution models to change in grain size

Predictive species distribution modelling (SDM) has become an essential tool in biodiversity conservation and management. The choice of grain size (resolution) of environmental layers used in modelling is one important factor that may affect predictions. We applied 10 distinct modelling techniques to presence-only data for 50 species in five different regions, to test whether: (1) a 10-fold coarsening of resolution affects predictive performance of SDMs, and (2) any observed effects are dependent on the type of region, modelling technique, or species considered. Results show that a 10 times change in grain size does not severely affect predictions from species distribution models. The overall trend is towards degradation of model performance, but improvement can also be observed. Changing grain size does not equally affect models across regions, techniques, and species types. The strongest effect is on regions and species types, with tree species in the data sets (regions) with highest locational accuracy being most affected. Changing grain size had little influence on the ranking of techniques: boosted regression trees remain best at both resolutions. The number of occurrences used for model training had an important effect, with larger sample sizes resulting in better models, which tended to be more sensitive to grain. Effect of grain change was only noticeable for models reaching sufficient performance and/or with initial data that have an intrinsic error smaller than the coarser grain size.

Right coronary artery flow velocity and volume assessment with spiral K-space sampled breathhold velocity-encoded MRI at 3 tesla: accuracy and reproducibility.

PURPOSE: To evaluate accuracy and reproducibility of flow velocity and volume measurements in a phantom and in human coronary arteries using breathhold velocity-encoded (VE) MRI with spiral k-space sampling at 3 Tesla. MATERIALS AND METHODS: Flow velocity assessment was performed using VE MRI with spiral k-space sampling. Accuracy of VE MRI was tested in vitro at five constant flow rates. Reproducibility was investigated in 19 healthy subjects (mean age 25.4 +/- 1.2 years, 11 men) by repeated acquisition in the right coronary artery (RCA). RESULTS: MRI-measured flow rates correlated strongly with volumetric collection (Pearson correlation r = 0.99; P < 0.01). Due to limited sample resolution, VE MRI overestimated the flow rate by 47% on average when nonconstricted region-of-interest segmentation was used. Using constricted region-of-interest segmentation with lumen size equal to ground-truth luminal size, less than 13% error in flow rate was found. In vivo RCA flow velocity assessment was successful in 82% of the applied studies. High interscan, intra- and inter-observer agreement was found for almost all indices describing coronary flow velocity. Reproducibility for repeated acquisitions varied by less than 16% for peak velocity values and by less than 24% for flow volumes. CONCLUSION: 3T breathhold VE MRI with spiral k-space sampling enables accurate and reproducible assessment of RCA flow velocity.

Testing demographic models of effective population size.

Due to practical difficulties in obtaining direct genetic estimates of effective sizes, conservation biologists have to rely on so-called 'demographic models' which combine life-history and mating-system parameters with F-statistics in order to produce indirect estimates of effective sizes. However, for the same practical reasons that prevent direct genetic estimates, the accuracy of demographic models is difficult to evaluate. Here we use individual-based, genetically explicit computer simulations in order to investigate the accuracy of two such demographic models aimed at investigating the hierarchical structure of populations. We show that, by and large, these models provide good estimates under a wide range of mating systems and dispersal patterns. However, one of the models should be avoided whenever the focal species' breeding system approaches monogamy with no sex bias in dispersal or when a substructure within social groups is suspected because effective sizes may then be strongly overestimated. The timing during the life cycle at which F-statistics are evaluated is also of crucial importance and attention should be paid to it when designing field sampling since different demographic models assume different timings. Our study shows that individual-based, genetically explicit models provide a promising way of evaluating the accuracy of demographic models of effective size and delineate their field of applicability.

Assessment of early-stage optic nerve invasion in retinoblastoma using high-resolution 1.5 Tesla MRI with surface coils: a multicentre, prospective accuracy study with histopathological correlation.

OBJECTIVES: To assess the accuracy of high-resolution (HR) magnetic resonance imaging (MRI) in diagnosing early-stage optic nerve (ON) invasion in a retinoblastoma cohort. METHODS: This IRB-approved, prospective multicenter study included 95 patients (55 boys, 40 girls; mean age, 29 months). 1.5-T MRI was performed using surface coils before enucleation, including spin-echo unenhanced and contrast-enhanced (CE) T1-weighted sequences (slice thickness, 2 mm; pixel size <0.3 × 0.3 mm(2)). Images were read by five neuroradiologists blinded to histopathologic findings. ROC curves were constructed with AUC assessment using a bootstrap method. RESULTS: Histopathology identified 41 eyes without ON invasion and 25 with prelaminar, 18 with intralaminar and 12 with postlaminar invasion. All but one were postoperatively classified as stage I by the International Retinoblastoma Staging System. The accuracy of CE-T1 sequences in identifying ON invasion was limited (AUC = 0.64; 95 % CI, 0.55 - 0.72) and not confirmed for postlaminar invasion diagnosis (AUC = 0.64; 95 % CI, 0.47 - 0.82); high specificities (range, 0.64 - 1) and negative predictive values (range, 0.81 - 0.97) were confirmed. CONCLUSION: HR-MRI with surface coils is recommended to appropriately select retinoblastoma patients eligible for primary enucleation without the risk of IRSS stage II but cannot substitute for pathology in differentiating the first degrees of ON invasion. KEY POINTS: • HR-MRI excludes advanced optic nerve invasion with high negative predictive value. • HR-MRI accurately selects patients eligible for primary enucleation. • Diagnosis of early stages of optic nerve invasion still relies on pathology. • Several physiological MR patterns may mimic optic nerve invasion.

Test result-based sampling: an efficient design for estimating the accuracy of patient safety indicators.

OBJECTIVE: Accuracy studies of Patient Safety Indicators (PSIs) are critical but limited by the large samples required due to low occurrence of most events. We tested a sampling design based on test results (verification-biased sampling [VBS]) that minimizes the number of subjects to be verified. METHODS: We considered 3 real PSIs, whose rates were calculated using 3 years of discharge data from a university hospital and a hypothetical screen of very rare events. Sample size estimates, based on the expected sensitivity and precision, were compared across 4 study designs: random and VBS, with and without constraints on the size of the population to be screened. RESULTS: Over sensitivities ranging from 0.3 to 0.7 and PSI prevalence levels ranging from 0.02 to 0.2, the optimal VBS strategy makes it possible to reduce sample size by up to 60% in comparison with simple random sampling. For PSI prevalence levels below 1%, the minimal sample size required was still over 5000. CONCLUSIONS: Verification-biased sampling permits substantial savings in the required sample size for PSI validation studies. However, sample sizes still need to be very large for many of the rarer PSIs.

Accuracy and predictability in use of AO three-dimensionally preformed titanium mesh plates for posttraumatic orbital reconstruction: a pilot study.

The aim of this study was to prospectively evaluate the accuracy and predictability of new three-dimensionally preformed AO titanium mesh plates for posttraumatic orbital wall reconstruction.We analyzed the preoperative and postoperative clinical and radiologic data of 10 patients with isolated blow-out orbital fractures. Fracture locations were as follows: floor (N = 7; 70%), medial wall (N = 1; 1%), and floor/medial wall (N = 2; 2%). The floor fractures were exposed by a standard transconjunctival approach, whereas a combined transcaruncular transconjunctival approach was used in patients with medial wall fractures. A three-dimensional preformed AO titanium mesh plate (0.4 mm in thickness) was selected according to the size of the defect previously measured on the preoperative computed tomographic (CT) scan examination and fixed at the inferior orbital rim with 1 or 2 screws. The accuracy of plate positioning of the reconstructed orbit was assessed on the postoperative CT scan. Coronal CT scan slices were used to measure bony orbital volume using OsiriX Medical Image software. Reconstructed versus uninjured orbital volume were statistically correlated.Nine patients (90%) had a successful treatment outcome without complications. One patient (10%) developed a mechanical limitation of upward gaze with a resulting handicapping diplopia requiring hardware removal. Postoperative orbital CT scan showed an anatomic three-dimensional placement of the orbital mesh plates in all of the patients. Volume data of the reconstructed orbit fitted that of the contralateral uninjured orbit with accuracy to within 2.5 cm(3). There was no significant difference in volume between the reconstructed and uninjured orbits.This preliminary study has demonstrated that three-dimensionally preformed AO titanium mesh plates for posttraumatic orbital wall reconstruction results in (1) a high rate of success with an acceptable rate of major clinical complications (10%) and (2) an anatomic restoration of the bony orbital contour and volume that closely approximates that of the contralateral uninjured orbit.

Effects of sample size on the performance of species distribution models

A wide range of modelling algorithms is used by ecologists, conservation practitioners, and others to predict species ranges from point locality data. Unfortunately, the amount of data available is limited for many taxa and regions, making it essential to quantify the sensitivity of these algorithms to sample size. This is the first study to address this need by rigorously evaluating a broad suite of algorithms with independent presence-absence data from multiple species and regions. We evaluated predictions from 12 algorithms for 46 species (from six different regions of the world) at three sample sizes (100, 30, and 10 records). We used data from natural history collections to run the models, and evaluated the quality of model predictions with area under the receiver operating characteristic curve (AUC). With decreasing sample size, model accuracy decreased and variability increased across species and between models. Novel modelling methods that incorporate both interactions between predictor variables and complex response shapes (i.e. GBM, MARS-INT, BRUTO) performed better than most methods at large sample sizes but not at the smallest sample sizes. Other algorithms were much less sensitive to sample size, including an algorithm based on maximum entropy (MAXENT) that had among the best predictive power across all sample sizes. Relative to other algorithms, a distance metric algorithm (DOMAIN) and a genetic algorithm (OM-GARP) had intermediate performance at the largest sample size and among the best performance at the lowest sample size. No algorithm predicted consistently well with small sample size (n < 30) and this should encourage highly conservative use of predictions based on small sample size and restrict their use to exploratory modelling.

Validity of silhouette showcards as a measure of body size and obesity in a population in the African region : a practical research tool for general-purpose surveys.

BACKGROUND: The purpose of this study is to validate the Pulvers silhouette showcard as a measure of weight status in a population in the African region. This tool is particularly beneficial when scarce resources do not allow for direct anthropometric measurements due to limited survey time or lack of measurement technology in face-to-face general-purpose surveys or in mailed, online, or mobile device-based surveys. METHODS: A cross-sectional study was conducted in the Republic of Seychelles with a sample of 1240 adults. We compared self-reported body sizes measured by Pulvers' silhouette showcards to four measurements of body size and adiposity: body mass index (BMI), body fat percent measured, waist circumference, and waist to height ratio. The accuracy of silhouettes as an obesity indicator was examined using sex-specific receiver operator curve (ROC) analysis and the reliability of this tool to detect socioeconomic gradients in obesity was compared to BMI-based measurements. RESULTS: Our study supports silhouette body size showcards as a valid and reliable survey tool to measure self-reported body size and adiposity in an African population. The mean correlation coefficients of self-reported silhouettes with measured BMI were 0.80 in men and 0.81 in women (P < 0.001). The silhouette showcards also showed high accuracy for detecting obesity as per a BMI ≥ 30 (Area under curve, AUC: 0.91/0.89, SE: 0.01), which was comparable to other measured adiposity indicators: fat percent (AUC: 0.94/0.94, SE: 0.01), waist circumference (AUC: 0.95/0.94, SE: 0.01), and waist to height ratio (AUC: 0.95/0.94, SE: 0.01) amongst men and women, respectively. The use of silhouettes in detecting obesity differences among different socioeconomic groups resulted in similar magnitude, direction, and significance of association between obesity and socioeconomic status as when using measured BMI. CONCLUSIONS: This study highlights the validity and reliability of silhouettes as a survey tool for measuring obesity in a population in the African region. The ease of use and cost-effectiveness of this tool makes it an attractive alternative to measured BMI in the design of non-face-to-face online- or mobile device-based surveys as well as in-person general-purpose surveys of obesity in social sciences, where limited resources do not allow for direct anthropometric measurements.

Why are offspring born larger when it is colder? Phenotypic plasticity for offspring size in the cladoceran Simocephalus vetulus (Müller)

It has been predicted on theorerical grounds (Sibly & Calow, 1983; Taylor & Williams, 1984) that optimal offspring size should be highly sensitive to juvenile growth and survival rates. To test such models, genetically-identical individuals of Simicephalus vetulus were reared at different temperatures and monitored for offspring size and juvenile growth rate. As adult size correlates negatively with temperature, an analysis of covariance was performed to separate the effects of temperature and maternal size. The result is that offspring size indeed correlates negatively with juvenile growth rate. Comparisons are made with field observation of several authors on seasonal variation of offspring size and alternative explanations are discussed. It is concluded that present experiments support the prediction of the theoretical models.

Sexual dimorphism in resource acquisition and deployment: both size and timing matter.

Background and Aims The males and females of many dioecious plant species differ from one another in important life-history traits, such as their size. If male and female reproductive functions draw on different resources, for example, one should expect males and females to display different allocation strategies as they grow. Importantly, these strategies may differ not only between the two sexes, but also between plants of different age and therefore size. Results are presented from an experiment that asks whether males and females of Mercurialis annua, an annual plant with indeterminate growth, differ over time in their allocation of two potentially limiting resources (carbon and nitrogen) to vegetative (below-and above-ground) and reproductive tissues.Methods Comparisons were made of the temporal patterns of biomass allocation to shoots, roots and reproduction and the nitrogen content in the leaves between the sexes of M. annua by harvesting plants of each sex after growth over different periods of time.Key Results and Conclusions Males and females differed in their temporal patterns of allocation. Males allocated more to reproduction than females at early stages, but this trend was reversed at later stages. Importantly, males allocated proportionally more of their biomass towards roots at later stages, but the roots of females were larger in absolute terms. The study points to the important role played by both the timing of resource deployment and the relative versus absolute sizes of the sinks and sources in sexual dimorphism of an annual plant.

Non-assisted versus neuro-navigated and XperCT-guided external ventricular catheter placement: a comparative cadaver study.

BACKGROUND AND PURPOSE: Accurate placement of an external ventricular drain (EVD) for the treatment of hydrocephalus is of paramount importance for its functionality and in order to minimize morbidity and complications. The aim of this study was to compare two different drain insertion assistance tools with the traditional free-hand anatomical landmark method, and to measure efficacy, safety and precision. METHODS: Ten cadaver heads were prepared by opening large bone windows centered on Kocher's points on both sides. Nineteen physicians, divided in two groups (trainees and board certified neurosurgeons) performed EVD insertions. The target for the ventricular drain tip was the ipsilateral foramen of Monro. Each participant inserted the external ventricular catheter in three different ways: 1) free-hand by anatomical landmarks, 2) neuronavigation-assisted (NN), and 3) XperCT-guided (XCT). The number of ventricular hits and dangerous trajectories; time to proceed; radiation exposure of patients and physicians; distance of the catheter tip to target and size of deviations projected in the orthogonal plans were measured and compared. RESULTS: Insertion using XCT increased the probability of ventricular puncture from 69.2 to 90.2 % (p = 0.02). Non-assisted placements were significantly less precise (catheter tip to target distance 14.3 ± 7.4 mm versus 9.6 ± 7.2 mm, p = 0.0003). The insertion time to proceed increased from 3.04 ± 2.06 min. to 7.3 ± 3.6 min. (p < 0.001). The X-ray exposure for XCT was 32.23 mSv, but could be reduced to 13.9 mSv if patients were initially imaged in the hybrid-operating suite. No supplementary radiation exposure is needed for NN if patients are imaged according to a navigation protocol initially. CONCLUSION: This ex vivo study demonstrates a significantly improved accuracy and safety using either NN or XCT-assisted methods. Therefore, efforts should be undertaken to implement these new technologies into daily clinical practice. However, the accuracy versus urgency of an EVD placement has to be balanced, as the image-guided insertion technique will implicate a longer preparation time due to a specific image acquisition and trajectory planning.

Octopus Standard Automated Perimetry, Pulsar Perimetry, Moorfields Motion Displacement Test and Heidelberg Retinal Tomography in Glaucoma Detection - A Comparison of Diagnostic Accuracy

Purpose: To investigate the accuracy of 4 clinical instruments in the detection of glaucomatous damage. Methods: 102 eyes of 55 test subjects (Age mean = 66.5yrs, range = [39; 89]) underwent Heidelberg Retinal Tomography (HRTIII), (disc area<2.43); and standard automated perimetry (SAP) using Octopus (Dynamic); Pulsar (TOP); and Moorfields Motion Displacement Test (MDT) (ESTA strategy). Eyes were separated into three groups 1) Healthy (H): IOP<21mmHg and healthy discs (clinical examination), 39 subjects, 78 eyes; 2) Glaucoma suspect (GS): Suspicious discs (clinical examination), 12 subjects, 15 eyes; 3) Glaucoma (G): progressive structural or functional loss, 14 subjects, 20 eyes. Clinical diagnostic precision was examined using the cut-off associated with the p<5% normative limit of MD (Octopus/Pulsar), PTD (MDT) and MRA (HRT) analysis. The sensitivity, specificity and accuracy were calculated for each instrument. Results: See table Conclusions: Despite the advantage of defining glaucoma suspects using clinical optic disc examination, the HRT did not yield significantly higher accuracy than functional measures. HRT, MDT and Octopus SAP yielded higher accuracy than Pulsar perimetry, although results did not reach statistical significance. Further studies are required to investigate the structure-function correlations between these instruments.