Discovering Salient Anatomical Landmarks by Predicting Human Gaze.

Anatomical landmarks are a crucial prerequisite for many medical imaging tasks. Usually, the set of landmarks for a given task is predefined by experts. The landmark locations for a given image are then annotated manually or via machine learning methods trained on manual annotations. In this paper, in contrast, we present a method to automatically discover and localize anatomical landmarks in medical images. Specifically, we consider landmarks that attract the visual attention of humans, which we term visually salient landmarks. We illustrate the method for fetal neurosonographic images. First, full-length clinical fetal ultrasound scans are recorded with live sonographer gaze-tracking. Next, a convolutional neural network (CNN) is trained to predict the gaze point distribution (saliency map) of the sonographers on scan video frames. The CNN is then used to predict saliency maps of unseen fetal neurosonographic images, and the landmarks are extracted as the local maxima of these saliency maps. Finally, the landmarks are matched across images by clustering the landmark CNN features. We show that the discovered landmarks can be used within affine image registration, with average landmark alignment errors between 4.1% and 10.9% of the fetal head long axis length.

Expected-value bias in routine third-trimester growth scans.

OBJECTIVES: Operators performing fetal growth scans are usually aware of the gestational age of the pregnancy, which may lead to expected-value bias when performing biometric measurements. We aimed to evaluate the incidence of expected-value bias in routine fetal growth scans and assess its impact on standard biometric measurements. METHODS: We collected prospectively full-length video recordings of routine ultrasound growth scans coupled with operator eye tracking. Expected value was defined as the gestational age at the time of the scan, based on the estimated due date that was established at the dating scan. Expected-value bias was defined as occurring when the operator looked at the measurement box on the screen during the process of caliper adjustment before saving a measurement. We studied the three standard biometric planes on which measurements of head circumference (HC), abdominal circumference (AC) and femur length (FL) are obtained. We evaluated the incidence of expected-value bias and quantified the impact of biased measurements. RESULTS: We analyzed 272 third-trimester growth scans, performed by 16 operators, during which a total of 1409 measurements (354 HC, 703 AC and 352 FL; including repeat measurements) were obtained. Expected-value bias occurred in 91.4% of the saved standard biometric plane measurements (85.0% for HC, 92.9% for AC and 94.9% for FL). The operators were more likely to adjust the measurements towards the expected value than away from it (47.7% vs 19.7% of measurements; P < 0.001). On average, measurements were corrected by 2.3 ± 5.6, 2.4 ± 10.4 and 3.2 ± 10.4 days of gestation towards the expected gestational age for the HC, AC, and FL measurements, respectively. Additionally, we noted a statistically significant reduction in measurement variance once the operator was biased (P = 0.026). Comparing the lowest and highest possible estimated fetal weight (using the smallest and largest biased HC, AC and FL measurements), we noted that the discordance, in percentage terms, was 10.1% ± 6.5%, and that in 17% (95% CI, 12-21%) of the scans, the fetus could be considered as small-for-gestational age or appropriate-for-gestational age if using the smallest or largest possible measurements, respectively. Similarly, in 13% (95% CI, 9-16%) of scans, the fetus could be considered as large-for-gestational age or appropriate-for-gestational age if using the largest or smallest possible measurements, respectively. CONCLUSIONS: During routine third-trimester growth scans, expected-value bias frequently occurs and significantly changes standard biometric measurements obtained. © 2019 the Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of the International Society of Ultrasound in Obstetrics and Gynecology.

Spatio-Temporal Partitioning and Description of Full-Length Routine Fetal Anomaly Ultrasound Scans.

This paper considers automatic clinical workflow description of full-length routine fetal anomaly ultrasound scans using deep learning approaches for spatio-temporal video analysis. Multiple architectures consisting of 2D and 2D + t CNN, LSTM, and convolutional LSTM are investigated and compared. The contributions of short-term and long-term temporal changes are studied, and a multi-stream framework analysis is found to achieve the best top-1 accuracy=0.77 and top-3 accuracy=0.94. Automated partitioning and characterisation on unlabelled full-length video scans show high correlation (ρ=0.95, p=0.0004) with workflow statistics of manually labelled videos, suggesting practicality of proposed methods.

Detection of carbapenemase activity in Pseudomonas aeruginosa by Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS).

A new MALDI-TOF MS imipenem hydrolysis assay was developed to optimize detection of carbapenemase-producing strains of Pseudomonas aeruginosa. The method correctly discriminated carbapenemase producers (n=74) from non-producers (n=95) in 99.4% of cases, and successfully detected 100% GES-5 strains (n=5).

SonoEyeNet: Standardized Fetal Ultrasound Plane Detection Informed by Eye Tracking.

We present a novel automated approach for detection of standardized abdominal circumference (AC) planes in fetal ultrasound built in a convolutional neural network (CNN) framework, called SonoEyeNet, that utilizes eye movement data of a sonographer in automatic interpretation. Eye movement data was collected from experienced sonographers as they identified an AC plane in fetal ultrasound video clips. A visual heatmap was generated from the eye movements for each video frame. A CNN model was built using ultrasound frames and their corresponding visual heatmaps. Different methods of processing visual heatmaps and their fusion with image feature maps were investigated. We show that with the assistance of human visual fixation information, the precision, recall and F1-score of AC plane detection was increased to 96.5%, 99.0% and 97.8% respectively, compared to 73.6%, 74.1% and 73.8% without using eye fixation information.

Effect of summer daylight exposure and genetic background on growth in growth hormone-deficient children.

The response to growth hormone in humans is dependent on phenotypic, genetic and environmental factors. The present study in children with growth hormone deficiency (GHD) collected worldwide characterised gene-environment interactions on growth response to recombinant human growth hormone (r-hGH). Growth responses in children are linked to latitude, and we found that a correlate of latitude, summer daylight exposure (SDE), was a key environmental factor related to growth response to r-hGH. In turn growth response was determined by an interaction between both SDE and genes known to affect growth response to r-hGH. In addition, analysis of associated networks of gene expression implicated a role for circadian clock pathways and specifically the developmental transcription factor NANOG. This work provides the first observation of gene-environment interactions in children treated with r-hGH.

Network analysis: a new approach to study endocrine disorders.

Systems biology is the study of the interactions that occur between the components of individual cells - including genes, proteins, transcription factors, small molecules, and metabolites, and their relationships to complex physiological and pathological processes. The application of systems biology to medicine promises rapid advances in both our understanding of disease and the development of novel treatment options. Network biology has emerged as the primary tool for studying systems biology as it utilises the mathematical analysis of the relationships between connected objects in a biological system and allows the integration of varied 'omic' datasets (including genomics, metabolomics, proteomics, etc.). Analysis of network biology generates interactome models to infer and assess function; to understand mechanisms, and to prioritise candidates for further investigation. This review provides an overview of network methods used to support this research and an insight into current applications of network analysis applied to endocrinology. A wide spectrum of endocrine disorders are included ranging from congenital hyperinsulinism in infancy, through childhood developmental and growth disorders, to the development of metabolic diseases in early and late adulthood, such as obesity and obesity-related pathologies. In addition to providing a deeper understanding of diseases processes, network biology is also central to the development of personalised treatment strategies which will integrate pharmacogenomics with systems biology of the individual.

Pharmacogenomics of insulin-like growth factor-I generation during GH treatment in children with GH deficiency or Turner syndrome.

Individual responses to growth hormone (GH) treatment are variable. Short-term generation of insulin-like growth factor-I (IGF-I) is recognized as a potential marker of sensitivity to GH treatment. This prospective, phase IV study used an integrated genomic analysis to identify markers associated with 1-month change in IGF-I (ΔIGF-I) following initiation of recombinant human (r-h)GH therapy in treatment-naïve children with GH deficiency (GHD) (n=166) or Turner syndrome (TS) (n=147). In both GHD and TS, polymorphisms in the cell-cycle regulator CDK4 were associated with 1-month ΔIGF-I (P<0.05). Baseline gene expression was also correlated with 1-month ΔIGF-I in both GHD and TS (r=0.3; P<0.01). In patients with low IGF-I responses, carriage of specific CDK4 alleles was associated with MAPK and glucocorticoid receptor signaling in GHD, and with p53 and Wnt signaling pathways in TS. Understanding the relationship between genomic markers and early changes in IGF-I may allow development of strategies to rapidly individualize r-hGH dose.

Pharmacogenomics related to growth disorders.

Growth disorders resulting in short stature are caused by a wide range of underlying pathophysiological processes. To improve height many of these conditions are treated with recombinant human growth hormone (rhGH). However, substantial inter-individual variability in growth response both in the short and long-term is recognised. Over the last decade, disease-specific growth prediction models have been developed that the clinician can use to define a child's potential response to rhGH and to optimise starting and maintenance doses of rhGH. These models, however, are not able to predict all the variations in treatment response. There has, therefore, been recent interest in using genetic information to contribute to the evaluation of responses to rhGH, including high-throughput technologies for assessing DNA markers (genome) and mRNA transcripts (transcriptome) as pharmacogenomic tools. This review will focus on how these pharmacogenomic approaches are being applied to growth disorders.

A pharmacogenomic approach to the treatment of children with GH deficiency or Turner syndrome.

OBJECTIVE: Individual sensitivity to recombinant human GH (r-hGH) is variable. Identification of genetic factors contributing to this variability has potential use for individualization of treatment. The objective of this study was to identify genetic markers and gene expression profiles associated with growth response on r-hGH therapy in treatment-naïve, prepubertal children with GH deficiency (GHD) or Turner syndrome (TS). DESIGN: A prospective, multicenter, international, open-label pharmacogenomic study. METHODS: The associations of genotypes in 103 growth- and metabolism-related genes and baseline gene expression profiles with growth response to r-hGH (cm/year) over the first year were evaluated. Genotype associations were assessed with growth response as a continuous variable and as a categorical variable divided into quartiles. RESULTS: Eleven genes in GHD and ten in TS, with two overlapping between conditions, were significantly associated with growth response either as a continuous variable (seven in GHD, two in TS) or as a categorical variable (four more in GHD, eight more in TS). For example, in GHD, GRB10 was associated with high response (≥ Q3; P=0.0012), while SOS2 was associated with low response (≤ Q1; P=0.006), while in TS, LHX4 was associated with high response (P=0.0003) and PTPN1 with low response (P=0.0037). Differences in expression were identified for one of the growth response-associated genes in GHD (AKT1) and for two in TS (KRAS and MYOD1). CONCLUSIONS: Carriage of specific growth-related genetic markers is associated with growth response in GHD and TS. These findings indicate that pharmacogenomics could have a role in individualized management of childhood growth disorders.

Optimization of growth hormone dosing in children born small for gestational age: an open-label, randomized study of children during the fourth year of therapy.

BACKGROUND: Optimal dosage for growth hormone (GH) therapy in short, prepubertal children born small for gestational age (SGA) is controversial. METHODS: SGA OPTIMIS (NCT00249821) is a multicenter, open-label, parallel-group, pilot study of short children born SGA who had received recombinant human GH (r-hGH) (57 µg/kg/day) for 3 years. Children were randomized 1:1 to receive either 57 or 35 µg/kg/day r-hGH during year 4. The primary endpoint was height gain during year 4. RESULTS: 22 children were randomized (57 µg/kg/day, n = 10; 35 µg/kg/day, n = 12) and 21 completed the fourth year of therapy; 22 were included in efficacy analyses. During year 4, mean [standard deviation (SD)] height velocity was 6.4 (1.4) and 4.4 (1.2) cm/year (p = 0.001) and height velocity SD score (SDS) was 0.3 (0.3) and -0.1 (0.2) (p = 0.002) in the 57 and 35 µg/kg/day groups, respectively. The 57 µg/kg/day group continued with catch-up growth, had a significantly higher mean weight gain (p = 0.015) and significantly higher insulin-like growth factor-I levels at 12 months (p = 0.038). Five treatment-emergent adverse events were reported; none was serious or caused study withdrawal. CONCLUSIONS: Children who continued receiving 57 µg/kg/day r-hGH in year 4 had significantly greater height gain than those receiving 35 µg/kg/day r-hGH.

Changes in insulin sensitivity and glucose metabolism during therapy with recombinant human growth hormone in short children born small for gestational age show a negative correlation with baseline measurements.

Recombinant human growth hormone (rhGH) is an effective therapy for children with short stature born small for gestational age (SGA); however, insulin resistance can develop during treatment. This retrospective analysis assessed the effect of rhGH treatment (0.067 mg/kg/day) on glucose metabolism and insulin secretion in children with short stature born SGA, and measured whether baseline characteristics correlated with changes in insulin resistance or glucose sensitivity during treatment. Baseline glucose area under the concentration-time curve (AUC) was negatively correlated with the change in glucose AUC (p<0.001). Similar negative correlations were seen between baseline insulin AUC and the change in insulin AUC during treatment (p<0.001); and between baseline HOMA-IR (homeostatic model of insulin resistance) and the change in HOMA-IR during treatment (p<0.001). Small but significant changes, not thought to be clinically significant, were seen in indicators of insulin sensitivity during rhGH treatment. Glucose levels remained within the normal range during oral glucose tolerance testing.

Recombinant human growth hormone for children born small for gestational age: meta-analysis confirms the consistent dose-effect relationship on catch-up growth.

BACKGROUND: The optimal treatment regimen of recombinant human GH (r-hGH) for short children born small for gestational age (SGA) is still under discussion. METHODS: A meta-analysis was performed of existing clinical trials that investigated the treatment of r-hGH in short children diagnosed SGA or with intrauterine growth retardation to determine the relationship between the daily r-hGH dose (placebo/no treatment; 0.033 mg/kg/day; 0.067 mg/kg/day) and the effect on growth [change in height-SD score (SDS) for chronological age]. A mathematical model describing the dose-response relationship was produced, and growth response (gain in height-SDS) to 2 yr of r-hGH 0.033 mg/kg/day [somatropin (rDNA origin) for injection; Serono] was estimated and compared with the response to other r-hGH formulations. RESULTS: The relationship between r-hGH dose and 2-yr growth response was described by an equation. The equation yielded a mean difference in height- SDS gain of 0.48 (0.35) between r-hGH 0.033 and 0.067 mg/kg/day in favor of the higher dose. The height-SDS gain after 2 yr of Serono r-hGH formulation, 0.033 mg/kg/day was estimated as 1.2. Comparison of this estimate to the growth response to 2-yr treatment at 0.033 mg/kg/day of other r-hGH formulations (mean difference in height-SDS 0.05, lower limit of the 95% confidence interval=-0.15) confirmed that growth response to Serono r-hGH formulation 0.033 mg/kg/day is an inferred response estimated to be within the range of observed responses to a (non-Serono formulation) r-hGH dose of 0.033 mg/kg/day. CONCLUSION: There is a clear dose-response relationship for r-hGH in the treatment of short children born SGA and the analysis confirmed that treatment with Serono r-hGH formulation 0.033 mg/kg/day should provide a meaningful therapeutic response.

Growth hormone therapy for short children born small for gestational age.

Children born small for gestational age may demonstrate continued growth retardation, resulting in persistent short stature. In the majority of the cases, this is linked with abnormal growth hormone secretion and also abnormal insulin-like growth factor levels. This review discusses the treatment of such children with recombinant human growth hormone. It illustrates the importance of starting therapy early, the dose-dependent response, and the advantages of continuous therapy and describes safety considerations.

Peripheral and central H1 histamine receptor occupancy by levocetirizine, a non-sedating antihistamine; a time course study in the guinea pig.

BACKGROUND AND PURPOSE: The H(1) receptor occupancy (H1RO) in brain is an indicator of central side effects of antihistamines. Here, we determined the kinetics of central and peripheral H1RO by levocetirizine in relation to its brain and plasma concentration, and investigated the role of the blood-brain barrier in any delay in brain H1RO. EXPERIMENTAL APPROACH: Concentration-time profiles in plasma and brain were obtained after 0.1 and 1 mg kg(-1) oral doses of levocetirizine in guinea pigs. H1RO in brain was measured ex vivo using [3H]-mepyramine and, in the periphery, by measuring the degree of inhibition of histamine-induced contractions of isolated guinea pig ileum. KEY RESULTS: The concentration-time profile of levocetirizine indicated lower levels (partition coefficient, K(p)=0.06-0.08), higher t(max) (2-4 h vs 1-1.5 h) and longer terminal half-life (4-5.6 h vs 2.1-2.8 h) in brain than plasma. The H1RO at 0.1 and 1 mg kg(-1) were 75% and 97%, respectively, at 1 hr in the periphery and, in the brain, were <20% and 28-67% respectively, at all time points studied. Brain H1RO vs plasma concentrations profile showed a delay, but not when compared to brain concentrations. CONCLUSIONS AND IMPLICATIONS: This study demonstrates an effective peripheral antihistamine effect of levocetirizine without central adverse effects at the dose close to human therapeutic dose. The slow increase in H1RO in the brain with time was caused by slow blood-brain barrier transport of levocetirizine. This demonstrates the importance of measuring time course of brain H1RO in relation to brain concentrations of drugs.

Contractile properties of various histaprodifen derivatives in guinea pig isolated ileum and trachea.

We characterized the histamine H(1) receptor agonism of various histaprodifen derivatives in guinea pig isolated ileum and trachea in comparison with histamine. Based on their affinity (calculated pK(A) values for ileum and trachea, respectively), the compounds were ranked as follows: suprahistaprodifen (8.31/8.08) > N(alpha)-(4-phenylbutyl)histaprodifen (7.22/5.93) >or= histamine (5.79/5.19) approximately methylhistaprodifen (5.57/6.07). Based on their efficacy (calculated tau values for ileum and trachea, respectively), the compounds were ranked as follows: methylhistaprodifen (37.67/2.50) > histamine (5.64/1.80) > suprahistaprodifen (1.63/1.42) >or= N(alpha)-(4-phenylbutyl)histaprodifen (0.083/1.54). In the ileum, histamine and methylhistaprodifen showed a high histamine H(1) receptor reserve while suprahistaprodifen and N(alpha)-(4-phenylbutyl)histaprodifen are devoid of any histamine H(1 )receptor reserve. On the trachea, no histamine H(1 )receptor reserve was demonstrable with the four tested agonists. The kinetic of contraction/relaxation of the ileum was faster with histamine and methylhistaprodifen than with suprahistaprodifen and N(alpha)-(4-phenylbutyl)histaprodifen. Histamine contracted the trachea faster than histaprodifen derivatives. Levocetirizine antagonized contractions induced by histamine and histaprodifen derivatives in both tissues. The differences observed in the calculated pA(2) (7.60-8.29) and/or pD'(2) values (6.28-7.90) depending on the tissue and/or the agonist are discussed.

Solubilization and immunopurification of rat brain synaptic vesicle protein 2A with maintained binding properties.

This study reports the solubilization of the rat synaptic vesicle protein SV2A, the brain binding site for the antiepileptic drug levetiracetam (LEV), and its characterization. N-dodecyl-beta-D-maltoside (DDM) was the best detergent at achieving a high percentage of SV2A solubilization and at maintaining the binding characteristics of a tritiated form of a more potent analogue of LEV, [3H]ucb 30889 ((2S)-2-[4-(3-azidophenyl)-2-oxopyrrolidin-1-yl]butanamide). Scatchard analysis revealed that approximately 25% of SV2A proteins from brain membranes are solubilized by DDM under optimal conditions. Competition binding experiments with a variety of LEV analogues indicated that [3H]ucb 30889 labels the same binding site in both crude homogenates and soluble extracts, with still high stereoselectivity. After immunoprecipitation of SV2A from solubilized rat brain membranes, binding properties of [3H]ucb 30889 to SV2A and association with synaptotagmin I were maintained. The two other isoforms SV2B and SV2C were found to be co-immunoprecipitated with SV2A. The solubilization and immunopurification of SV2A with unmodified ligand affinities and synaptotagmin I interaction provides the starting point for future protein-protein interactions and structural studies.