MDCT Arthrography of the Hip: Value of the Adaptive Statistical Iterative Reconstruction Technique and Potential for Radiation Dose Reduction

Au cours des deux dernières décennies, la technique d'imagerie arthro-scanner a bénéficié de nombreux progrès technologiques et représente aujourd'hui une excellente alternative à l'imagerie par résonance magnétique (IRM) et / ou arthro-IRM dans l'évaluation des pathologies de la hanche. Cependant, elle reste limitée par l'exposition aux rayonnements ionisants importante. Les techniques de reconstruction itérative (IR) ont récemment été mis en oeuvre avec succès en imagerie ; la littérature montre que l'utilisation ces dernières contribue à réduire la dose d'environ 40 à 55%, comparativement aux protocoles courants utilisant la rétroprojection filtrée (FBP), en scanner de rachis. A notre connaissance, l'utilisation de techniques IR en arthro-scanner de hanche n'a pas été évaluée jusqu'à présent. Le but de notre étude était d'évaluer l'impact de la technique ASIR (GE Healthcare) sur la qualité de l'image objective et subjective en arthro-scanner de hanche, et d'évaluer son potentiel en terme de réduction de dose. Pour cela, trente sept patients examinés par arthro-scanner de hanche ont été randomisés en trois groupes : dose standard (CTDIvol = 38,4 mGy) et deux groupes de dose réduite (CTDIvol = 24,6 ou 15,4 mGy). Les images ont été reconstruites en rétroprojection filtrée (FBP) puis en appliquant différents pourcentages croissants d'ASIR (30, 50, 70 et 90%). Le bruit et le rapport contraste sur bruit (CNR) ont été mesurés. Deux radiologues spécialisés en imagerie musculo-squelettique ont évalué de manière indépendante la qualité de l'image au niveau de plusieurs structures anatomiques en utilisant une échelle de quatre grades. Ils ont également évalué les lésions labrales et du cartilage articulaire. Les résultats révèlent que le bruit augmente (p = 0,0009) et le CNR diminue (p = 0,001) de manière significative lorsque la dose diminue. A l'inverse, le bruit diminue (p = 0,0001) et le contraste sur bruit augmente (p < 0,003) de manière significative lorsque le pourcentage d'ASIR augmente ; on trouve également une augmentation significative des scores de la qualité de l'image pour le labrum, le cartilage, l'os sous-chondral, la qualité de l'image globale (au delà de ASIR 50%), ainsi que le bruit (p < 0,04), et une réduction significative pour l'os trabuculaire et les muscles (p < 0,03). Indépendamment du niveau de dose, il n'y a pas de différence significative pour la détection et la caractérisation des lésions labrales (n=24, p = 1) et des lésions cartilagineuses (n=40, p > 0,89) en fonction du pourcentage d'ASIR. Notre travail a permis de montrer que l'utilisation de plus de 50% d'ASIR permet de reduire de manière significative la dose d'irradiation reçue par le patient lors d'un arthro-scanner de hanche tout en maintenant une qualité d'image diagnostique comparable par rapport à un protocole de dose standard utilisant la rétroprojection filtrée.

Assessment of statistical iterative reconstructions to optimize CT examination protocols for children and young adults

La tomodensitométrie (CT) est une technique d'imagerie dont l'intérêt n'a cessé de croître depuis son apparition dans le début des années 70. Dans le domaine médical, son utilisation est incontournable à tel point que ce système d'imagerie pourrait être amené à devenir victime de son succès si son impact au niveau de l'exposition de la population ne fait pas l'objet d'une attention particulière. Bien évidemment, l'augmentation du nombre d'examens CT a permis d'améliorer la prise en charge des patients ou a rendu certaines procédures moins invasives. Toutefois, pour assurer que le compromis risque - bénéfice soit toujours en faveur du patient, il est nécessaire d'éviter de délivrer des doses non utiles au diagnostic.¦Si cette action est importante chez l'adulte elle doit être une priorité lorsque les examens se font chez l'enfant, en particulier lorsque l'on suit des pathologies qui nécessitent plusieurs examens CT au cours de la vie du patient. En effet, les enfants et jeunes adultes sont plus radiosensibles. De plus, leur espérance de vie étant supérieure à celle de l'adulte, ils présentent un risque accru de développer un cancer radio-induit dont la phase de latence peut être supérieure à vingt ans. Partant du principe que chaque examen radiologique est justifié, il devient dès lors nécessaire d'optimiser les protocoles d'acquisitions pour s'assurer que le patient ne soit pas irradié inutilement. L'avancée technologique au niveau du CT est très rapide et depuis 2009, de nouvelles techniques de reconstructions d'images, dites itératives, ont été introduites afin de réduire la dose et améliorer la qualité d'image.¦Le présent travail a pour objectif de déterminer le potentiel des reconstructions itératives statistiques pour réduire au minimum les doses délivrées lors d'examens CT chez l'enfant et le jeune adulte tout en conservant une qualité d'image permettant le diagnostic, ceci afin de proposer des protocoles optimisés.¦L'optimisation d'un protocole d'examen CT nécessite de pouvoir évaluer la dose délivrée et la qualité d'image utile au diagnostic. Alors que la dose est estimée au moyen d'indices CT (CTDIV0| et DLP), ce travail a la particularité d'utiliser deux approches radicalement différentes pour évaluer la qualité d'image. La première approche dite « physique », se base sur le calcul de métriques physiques (SD, MTF, NPS, etc.) mesurées dans des conditions bien définies, le plus souvent sur fantômes. Bien que cette démarche soit limitée car elle n'intègre pas la perception des radiologues, elle permet de caractériser de manière rapide et simple certaines propriétés d'une image. La seconde approche, dite « clinique », est basée sur l'évaluation de structures anatomiques (critères diagnostiques) présentes sur les images de patients. Des radiologues, impliqués dans l'étape d'évaluation, doivent qualifier la qualité des structures d'un point de vue diagnostique en utilisant une échelle de notation simple. Cette approche, lourde à mettre en place, a l'avantage d'être proche du travail du radiologue et peut être considérée comme méthode de référence.¦Parmi les principaux résultats de ce travail, il a été montré que les algorithmes itératifs statistiques étudiés en clinique (ASIR?, VEO?) ont un important potentiel pour réduire la dose au CT (jusqu'à-90%). Cependant, par leur fonctionnement, ils modifient l'apparence de l'image en entraînant un changement de texture qui pourrait affecter la qualité du diagnostic. En comparant les résultats fournis par les approches « clinique » et « physique », il a été montré que ce changement de texture se traduit par une modification du spectre fréquentiel du bruit dont l'analyse permet d'anticiper ou d'éviter une perte diagnostique. Ce travail montre également que l'intégration de ces nouvelles techniques de reconstruction en clinique ne peut se faire de manière simple sur la base de protocoles utilisant des reconstructions classiques. Les conclusions de ce travail ainsi que les outils développés pourront également guider de futures études dans le domaine de la qualité d'image, comme par exemple, l'analyse de textures ou la modélisation d'observateurs pour le CT.¦-¦Computed tomography (CT) is an imaging technique in which interest has been growing since it first began to be used in the early 1970s. In the clinical environment, this imaging system has emerged as the gold standard modality because of its high sensitivity in producing accurate diagnostic images. However, even if a direct benefit to patient healthcare is attributed to CT, the dramatic increase of the number of CT examinations performed has raised concerns about the potential negative effects of ionizing radiation on the population. To insure a benefit - risk that works in favor of a patient, it is important to balance image quality and dose in order to avoid unnecessary patient exposure.¦If this balance is important for adults, it should be an absolute priority for children undergoing CT examinations, especially for patients suffering from diseases requiring several follow-up examinations over the patient's lifetime. Indeed, children and young adults are more sensitive to ionizing radiation and have an extended life span in comparison to adults. For this population, the risk of developing cancer, whose latency period exceeds 20 years, is significantly higher than for adults. Assuming that each patient examination is justified, it then becomes a priority to optimize CT acquisition protocols in order to minimize the delivered dose to the patient. Over the past few years, CT advances have been developing at a rapid pace. Since 2009, new iterative image reconstruction techniques, called statistical iterative reconstructions, have been introduced in order to decrease patient exposure and improve image quality.¦The goal of the present work was to determine the potential of statistical iterative reconstructions to reduce dose as much as possible without compromising image quality and maintain diagnosis of children and young adult examinations.¦The optimization step requires the evaluation of the delivered dose and image quality useful to perform diagnosis. While the dose is estimated using CT indices (CTDIV0| and DLP), the particularity of this research was to use two radically different approaches to evaluate image quality. The first approach, called the "physical approach", computed physical metrics (SD, MTF, NPS, etc.) measured on phantoms in well-known conditions. Although this technique has some limitations because it does not take radiologist perspective into account, it enables the physical characterization of image properties in a simple and timely way. The second approach, called the "clinical approach", was based on the evaluation of anatomical structures (diagnostic criteria) present on patient images. Radiologists, involved in the assessment step, were asked to score image quality of structures for diagnostic purposes using a simple rating scale. This approach is relatively complicated to implement and also time-consuming. Nevertheless, it has the advantage of being very close to the practice of radiologists and is considered as a reference method.¦Primarily, this work revealed that the statistical iterative reconstructions studied in clinic (ASIR? and VECO have a strong potential to reduce CT dose (up to -90%). However, by their mechanisms, they lead to a modification of the image appearance with a change in image texture which may then effect the quality of the diagnosis. By comparing the results of the "clinical" and "physical" approach, it was showed that a change in texture is related to a modification of the noise spectrum bandwidth. The NPS analysis makes possible to anticipate or avoid a decrease in image quality. This project demonstrated that integrating these new statistical iterative reconstruction techniques can be complex and cannot be made on the basis of protocols using conventional reconstructions. The conclusions of this work and the image quality tools developed will be able to guide future studies in the field of image quality as texture analysis or model observers dedicated to CT.

Implementing statistical learning methods through Bayesian networks (Part 2): bayesian evaluations for results of black toner analyses in forensic document examination

This paper presents and discusses the use of Bayesian procedures - introduced through the use of Bayesian networks in Part I of this series of papers - for 'learning' probabilities from data. The discussion will relate to a set of real data on characteristics of black toners commonly used in printing and copying devices. Particular attention is drawn to the incorporation of the proposed procedures as an integral part in probabilistic inference schemes (notably in the form of Bayesian networks) that are intended to address uncertainties related to particular propositions of interest (e.g., whether or not a sample originates from a particular source). The conceptual tenets of the proposed methodologies are presented along with aspects of their practical implementation using currently available Bayesian network software.

Swain: Stata module to correct the SEM chi-square overidentification test in small sample sizes or complex models. Statistical Software Components S457617, Boston College Department of Economics.

Swain corrects the chi-square overidentification test (i.e., likelihood ratio test of fit) for structural equation models whethr with or without latent variables. The chi-square statistic is asymptotically correct; however, it does not behave as expected in small samples and/or when the model is complex (cf. Herzog, Boomsma, & Reinecke, 2007). Thus, particularly in situations where the ratio of sample size (n) to the number of parameters estimated (p) is relatively small (i.e., the p to n ratio is large), the chi-square test will tend to overreject correctly specified models. To obtain a closer approximation to the distribution of the chi-square statistic, Swain (1975) developed a correction; this scaling factor, which converges to 1 asymptotically, is multiplied with the chi-square statistic. The correction better approximates the chi-square distribution resulting in more appropriate Type 1 reject error rates (see Herzog & Boomsma, 2009; Herzog, et al., 2007).

Adaptive strategy for the statistical analysis of connectomes.

We study an adaptive statistical approach to analyze brain networks represented by brain connection matrices of interregional connectivity (connectomes). Our approach is at a middle level between a global analysis and single connections analysis by considering subnetworks of the global brain network. These subnetworks represent either the inter-connectivity between two brain anatomical regions or by the intra-connectivity within the same brain anatomical region. An appropriate summary statistic, that characterizes a meaningful feature of the subnetwork, is evaluated. Based on this summary statistic, a statistical test is performed to derive the corresponding p-value. The reformulation of the problem in this way reduces the number of statistical tests in an orderly fashion based on our understanding of the problem. Considering the global testing problem, the p-values are corrected to control the rate of false discoveries. Finally, the procedure is followed by a local investigation within the significant subnetworks. We contrast this strategy with the one based on the individual measures in terms of power. We show that this strategy has a great potential, in particular in cases where the subnetworks are well defined and the summary statistics are properly chosen. As an application example, we compare structural brain connection matrices of two groups of subjects with a 22q11.2 deletion syndrome, distinguished by their IQ scores.

Database construction and Peptide identification strategies for proteogenomic studies on sequenced genomes.

Since the advent of high-throughput DNA sequencing technologies, the ever-increasing rate at which genomes have been published has generated new challenges notably at the level of genome annotation. Even if gene predictors and annotation softwares are more and more efficient, the ultimate validation is still in the observation of predicted gene product( s). Mass-spectrometry based proteomics provides the necessary high throughput technology to show evidences of protein presence and, from the identified sequences, confirmation or invalidation of predicted annotations. We review here different strategies used to perform a MS-based proteogenomics experiment with a bottom-up approach. We start from the strengths and weaknesses of the different database construction strategies, based on different genomic information (whole genome, ORF, cDNA, EST or RNA-Seq data), which are then used for matching mass spectra to peptides and proteins. We also review the important points to be considered for a correct statistical assessment of the peptide identifications. Finally, we provide references for tools used to map and visualize the peptide identifications back to the original genomic information.

Statistical discrimination of black gel pen inks analysed by laser desorption/ionization mass spectrometry

Pearson correlation coefficients were applied for the objective comparison of 30 black gel pen inks analysed by laser desorption ionization mass spectrometry (LDI-MS). The mass spectra were obtained for ink lines directly on paper using positive and negative ion modes at several laser intensities. This methodology has the advantage of taking into account the reproducibility of the results as well as the variability between spectra of different pens. A differentiation threshold could thus be selected in order to avoid the risk of false differentiation. Combining results from positive and negative mode yielded a discriminating power up to 85%, which was better than the one obtained previously with other optical comparison methodologies. The technique also allowed discriminating between pens from the same brand.

Improved statistical evaluation of group differences in connectomes by screening-filtering strategy with application to study maturation of brain connections between childhood and adolescence.

Detecting local differences between groups of connectomes is a great challenge in neuroimaging, because the large number of tests that have to be performed and the impact on multiplicity correction. Any available information should be exploited to increase the power of detecting true between-group effects. We present an adaptive strategy that exploits the data structure and the prior information concerning positive dependence between nodes and connections, without relying on strong assumptions. As a first step, we decompose the brain network, i.e., the connectome, into subnetworks and we apply a screening at the subnetwork level. The subnetworks are defined either according to prior knowledge or by applying a data driven algorithm. Given the results of the screening step, a filtering is performed to seek real differences at the node/connection level. The proposed strategy could be used to strongly control either the family-wise error rate or the false discovery rate. We show by means of different simulations the benefit of the proposed strategy, and we present a real application of comparing connectomes of preschool children and adolescents.

HIERFSTAT, a package for R to compute and test hierarchical F-statistics

The package HIERFSTAT for the statistical software R, created by the R Development Core Team, allows the estimate of hierarchical F-statistics from a hierarchy with any numbers of levels. In addition, it allows testing the statistical significance of population differentiation for these different levels, using a generalized likelihood-ratio test. The package HIERFSTAT is available at http://www.unil.ch/popgen/softwares/hierfstat.htm.

Improved statistical analysis of low abundance phenomena in bimodal bacterial populations.

Accurate detection of subpopulation size determinations in bimodal populations remains problematic yet it represents a powerful way by which cellular heterogeneity under different environmental conditions can be compared. So far, most studies have relied on qualitative descriptions of population distribution patterns, on population-independent descriptors, or on arbitrary placement of thresholds distinguishing biological ON from OFF states. We found that all these methods fall short of accurately describing small population sizes in bimodal populations. Here we propose a simple, statistics-based method for the analysis of small subpopulation sizes for use in the free software environment R and test this method on real as well as simulated data. Four so-called population splitting methods were designed with different algorithms that can estimate subpopulation sizes from bimodal populations. All four methods proved more precise than previously used methods when analyzing subpopulation sizes of transfer competent cells arising in populations of the bacterium Pseudomonas knackmussii B13. The methods' resolving powers were further explored by bootstrapping and simulations. Two of the methods were not severely limited by the proportions of subpopulations they could estimate correctly, but the two others only allowed accurate subpopulation quantification when this amounted to less than 25% of the total population. In contrast, only one method was still sufficiently accurate with subpopulations smaller than 1% of the total population. This study proposes a number of rational approximations to quantifying small subpopulations and offers an easy-to-use protocol for their implementation in the open source statistical software environment R.

Statistical learning theory for geospatial data. Case study: Aral sea

In recent years there has been an explosive growth in the development of adaptive and data driven methods. One of the efficient and data-driven approaches is based on statistical learning theory (Vapnik 1998). The theory is based on Structural Risk Minimisation (SRM) principle and has a solid statistical background. When applying SRM we are trying not only to reduce training error ? to fit the available data with a model, but also to reduce the complexity of the model and to reduce generalisation error. Many nonlinear learning procedures recently developed in neural networks and statistics can be understood and interpreted in terms of the structural risk minimisation inductive principle. A recent methodology based on SRM is called Support Vector Machines (SVM). At present SLT is still under intensive development and SVM find new areas of application (www.kernel-machines.org). SVM develop robust and non linear data models with excellent generalisation abilities that is very important both for monitoring and forecasting. SVM are extremely good when input space is high dimensional and training data set i not big enough to develop corresponding nonlinear model. Moreover, SVM use only support vectors to derive decision boundaries. It opens a way to sampling optimization, estimation of noise in data, quantification of data redundancy etc. Presentation of SVM for spatially distributed data is given in (Kanevski and Maignan 2004).

Informing the judgments of fingerprint analysts using quality metric and statistical assessment tools

The aim of this research was to evaluate how fingerprint analysts would incorporate information from newly developed tools into their decision making processes. Specifically, we assessed effects using the following: (1) a quality tool to aid in the assessment of the clarity of the friction ridge details, (2) a statistical tool to provide likelihood ratios representing the strength of the corresponding features between compared fingerprints, and (3) consensus information from a group of trained fingerprint experts. The measured variables for the effect on examiner performance were the accuracy and reproducibility of the conclusions against the ground truth (including the impact on error rates) and the analyst accuracy and variation for feature selection and comparison.¦The results showed that participants using the consensus information from other fingerprint experts demonstrated more consistency and accuracy in minutiae selection. They also demonstrated higher accuracy, sensitivity, and specificity in the decisions reported. The quality tool also affected minutiae selection (which, in turn, had limited influence on the reported decisions); the statistical tool did not appear to influence the reported decisions.

Sample size and statistical power in the hierarchical analysis of variance: applications in morphometry of the nervous system.

Analysis of variance is commonly used in morphometry in order to ascertain differences in parameters between several populations. Failure to detect significant differences between populations (type II error) may be due to suboptimal sampling and lead to erroneous conclusions; the concept of statistical power allows one to avoid such failures by means of an adequate sampling. Several examples are given in the morphometry of the nervous system, showing the use of the power of a hierarchical analysis of variance test for the choice of appropriate sample and subsample sizes. In the first case chosen, neuronal densities in the human visual cortex, we find the number of observations to be of little effect. For dendritic spine densities in the visual cortex of mice and humans, the effect is somewhat larger. A substantial effect is shown in our last example, dendritic segmental lengths in monkey lateral geniculate nucleus. It is in the nature of the hierarchical model that sample size is always more important than subsample size. The relative weight to be attributed to subsample size thus depends on the relative magnitude of the between observations variance compared to the between individuals variance.