Core Imaging Library - Part I: a versatile Python framework for tomographic imaging

We present the Core Imaging Library (CIL), an open-source Python framework for tomographic imaging with particular emphasis on reconstruction of challenging datasets. Conventional filtered back-projection reconstruction tends to be insufficient for highly noisy, incomplete, non-standard or multi-channel data arising for example in dynamic, spectral and in situ tomography. CIL provides an extensive modular optimization framework for prototyping reconstruction methods including sparsity and total variation regularization, as well as tools for loading, preprocessing and visualizing tomographic data. The capabilities of CIL are demonstrated on a synchrotron example dataset and three challenging cases spanning golden-ratio neutron tomography, cone-beam X-ray laminography and positron emission tomography. This article is part of the theme issue 'Synergistic tomographic image reconstruction: part 2'.

Core Imaging Library - Part I: a versatile Python framework for tomographic imaging

We present the Core Imaging Library (CIL), an open-source Python framework for tomographic imaging with particular emphasis on reconstruction of challenging datasets. Conventional filtered back-projection reconstruction tends to be insufficient for highly noisy, incomplete, non-standard or multi-channel data arising for example in dynamic, spectral and in situ tomography. CIL provides an extensive modular optimization framework for prototyping reconstruction methods including sparsity and total variation regularization, as well as tools for loading, preprocessing and visualizing tomographic data. The capabilities of CIL are demonstrated on a synchrotron example dataset and three challenging cases spanning golden-ratio neutron tomography, cone-beam X-ray laminography and positron emission tomography. This article is part of the theme issue 'Synergistic tomographic image reconstruction: part 2'.

Commentary on a combined approach to the problem of developing biomarkers for the prediction of spontaneous preterm labor that leads to preterm birth

INTRODUCTION

Globally, preterm birth has replaced congenital malformation as the major cause of perinatal mortality and morbidity. The reduced rate of congenital malformation was not achieved through a single biophysical or biochemical marker at a specific gestational age, but rather through a combination of clinical, biophysical and biochemical markers at different gestational ages. Since the aetiology of spontaneous preterm birth is also multifactorial, it is unlikely that a single biomarker test, at a specific gestational age will emerge as the definitive predictive test.

METHODS

The Biomarkers Group of PREBIC, comprising clinicians, basic scientists and other experts in the field, with a particular interest in preterm birth have produced this commentary with short, medium and long-term aims: i) to alert clinicians to the advances that are being made in the prediction of spontaneous preterm birth; ii) to encourage clinicians and scientists to continue their efforts in this field, and not to be disheartened or nihilistic because of a perceived lack of progress and iii) to enable development of novel interventions that can reduce the mortality and morbidity associated with preterm birth.

RESULTS

Using language that we hope is clear to practising clinicians, we have identified 11 Sections in which there exists the potential, feasibility and capability of technologies for candidate biomarkers in the prediction of spontaneous preterm birth and how current limitations to this research might be circumvented.

DISCUSSION

The combination of biophysical, biochemical, immunological, microbiological, fetal cell, exosomal, or cell free RNA at different gestational ages, integrated as part of a multivariable predictor model may be necessary to advance our attempts to predict sPTL and PTB. This will require systems biological data using "omics" data and artificial intelligence/machine learning to manage the data appropriately. The ultimate goal is to reduce the mortality and morbidity associated with preterm birth.

Automated angular and translational tomographic alignment and application to phase-contrast imaging

X-ray computerized tomography (CT) is a 3D imaging technique that makes use of x-ray illumination and image reconstruction techniques to reproduce the internal cross-sections of a sample. Tomographic projection data usually require an initial relative alignment or knowledge of the exact object position and orientation with respect to the detector. As tomographic imaging reaches increasingly better resolution, thermal drifts, mechanical instabilities, and equipment limitations are becoming the main dominant factors contributing to sample positioning uncertainties that will further introduce reconstruction artifacts and limit the attained resolution in the final tomographic reconstruction. Alignment algorithms that require manual interaction impede data analysis with ever-increasing data acquisition rates, supplied by more brilliant sources. We present in this paper an iterative reconstruction algorithm for wrapped phase projection data and an alignment algorithm that automatically takes 5 degrees of freedom, including the possible linear and angular motion errors, into consideration. The presented concepts are applied to simulated and real measured phase-contrast data, exhibiting a possible improvement in the reconstruction resolution. A MATLAB implementation is made publicly available and will allow robust analysis of large volumes of phase-contrast tomography data.

The association between angiogenic markers and fetal sex: Implications for preeclampsia research

OBJECTIVE

Current research suggests sexual dimorphism between the male and female fetoplacental units, but with unknown relevance for preeclampsia. We investigated the association between fetal sex and concentrations of the angiogenic markers soluble Fms-like kinase 1 (sFlt-1), placental growth factor (PlGF), and sFlt-1/PlGF ratio in first and second-third trimester in women with/without preeclampsia, and the impact of fetal sex on the prognostic value of angiogenic markers for preeclampsia.

STUDY DESIGN

Observational study in a prospective, population-based cohort of 2110 singleton pregnancies with 150 preeclampsia cases.

RESULTS

Higher sFlt-1 concentrations were observed for women carrying female fetuses in first trimester (all, 1107.65 vs. 992.27pg/ml; preeclampsia cases, 1118.79 vs. 934.49pg/ml, p<0.05) and in second-third trimester (all, 1130.03 vs. 1043.15pg/ml; preeclampsia, 1480.30 vs. 1152.86pg/ml, p<0.05), with similar findings for the sFlt-1/PlGF ratio concentrations in first (29.67 vs. 27.39 p<0.05) and second-third trimester (3.56 vs. 3.22, p<0.05). In first trimester, log transformed concentrations of PlGF, sFlt-1 and sFlt-1/PlGF (all participants) and sFlt-1 (preeclampsia cases) associated with fetal sex in adjusted analyses (p<0.05). In second-third trimester, only log(sFlt-1) associated with fetal sex (all, p=0.028; preeclampsia, p=0.067) In receiver operating curve analysis, prediction of early-onset preeclampsia by sFlt-1/PlGF tended to be superior in pregnancies with female vs. male fetuses (p=0.06).

CONCLUSION

Sexual dimorphism was observed for concentrations of angiogenic markers. Female fetal sex was associated to higher sFlt-1 and sFlt-1/PlGF ratio concentrations in both healthy pregnancies and women developing preeclampsia. Fetal sex should be considered in research and clinical use of angiogenic markers.

How little data is enough? Phase-diagram analysis of sparsity-regularized X-ray computed tomography

We introduce phase-diagram analysis, a standard tool in compressed sensing (CS), to the X-ray computed tomography (CT) community as a systematic method for determining how few projections suffice for accurate sparsity-regularized reconstruction. In CS, a phase diagram is a convenient way to study and express certain theoretical relations between sparsity and sufficient sampling. We adapt phase-diagram analysis for empirical use in X-ray CT for which the same theoretical results do not hold. We demonstrate in three case studies the potential of phase-diagram analysis for providing quantitative answers to questions of undersampling. First, we demonstrate that there are cases where X-ray CT empirically performs comparably with a near-optimal CS strategy, namely taking measurements with Gaussian sensing matrices. Second, we show that, in contrast to what might have been anticipated, taking randomized CT measurements does not lead to improved performance compared with standard structured sampling patterns. Finally, we show preliminary results of how well phase-diagram analysis can predict the sufficient number of projections for accurately reconstructing a large-scale image of a given sparsity by means of total-variation regularization.

Metabolic effects of lifestyle intervention in obese pregnant women. Results from the randomized controlled trial 'Lifestyle in Pregnancy' (LiP)

AIMS

The Lifestyle in Pregnancy intervention in obese pregnant women resulted in significantly lower gestational weight gain compared with the control group, but without improvement in rates of clinical pregnancy complications. The impact of the lifestyle intervention on metabolic measurements in the study participants is now reported.

METHODS

The Lifestyle in Pregnancy study was a randomized controlled trial among 360 obese women (BMI 30-45 kg/m²) who were allocated in early pregnancy to lifestyle interventions with diet counselling and physical activities or to the control group. Fasting blood samples, including plasma glucose, insulin, lipid profile and capillary blood glucose during a 2-h oral glucose tolerance test were carried out three times throughout pregnancy. Insulin resistance was estimated with the homeostasis model assessment of insulin resistance.

RESULTS

Three hundred and four women (84%) were followed until delivery. Women in the intervention group had a significantly lower change in insulin resistance (HOMA-IR) from randomization to 28-30 weeks' gestation compared with control subjects (mean ± SD: 0.7 ± 1.3 vs. 1.0 ± 1.3, P = 0.02). Despite a significantly lower gestational weight gain in the intervention group, there was no difference between the groups with respect to total cholesterol, HDL, LDL or triglycerides.

CONCLUSIONS

Lifestyle intervention in obese pregnant women resulted in attenuation of the physiologic pregnancy-induced insulin resistance. Despite restricted gestational weight gain, there were no changes in glucose or lipid metabolism between the groups.

Increasing use of antibiotics in pregnancy during the period 2000-2010: prevalence, timing, category, and demographics

OBJECTIVE

The aim of this study was to describe the use of antibiotics in a national population-based cohort of pregnant Danish women between 2000 and 2010.

DESIGN

Register-based, population-wide, cohort study.

SETTING

Denmark, from 2000 to 2010.

POPULATION

All pregnancies among Danish residents during the period 2000-2010 were included for analysis.

METHODS

Data were obtained from the Danish Medical Birth Registry, the Danish National Patient Registry, and the Registry of Medicinal Product Statistics. The filled prescriptions for systemic antibacterial, antimycotic, and antiviral drugs, as well as intravaginally applied antibiotics, were analysed. Associations with demographic variables were assessed using multivariate analysis.

MAIN OUTCOME MEASURES

Filled prescriptions for antibiotic drugs during pregnancy.

RESULTS

We included 987 973 pregnancies in Denmark from 2000 to 2010; 38.9% of women with a delivery and 14.8% of women with a miscarriage or termination of pregnancy had one or more antibiotic treatments during pregnancy. Systemic antibacterial drugs were the most frequently used drug group, with filled prescriptions for 33.4% of all deliveries and 12.6% of all abortions. This proportion increased from 28.4% in 2000 to 37.0% in 2010 among deliveries. The biggest change was seen for pivmecillinam, which increased among deliveries from 6.3% in 2000 to 19.5% in 2010. Obese (odds ratio 1.51; 95% CI 1.47-1.56), young (odds ratio 1.35; 95% CI 1.30-1.39), and low-educated women (odds ratio 1.37; 95% CI 1.35-1.1.39) tended to fill more prescriptions of antibiotics during pregnancy.

CONCLUSIONS

Overall, the number of women who filled prescriptions of antibiotics increased during the 11-year study period. In 2010, at least 41.5% of all deliveries were exposed to antibiotic therapy during pregnancy.

Parity and tanned white skin as novel predictors of vitamin D status in early pregnancy: a population-based cohort study

CONTEXT

In pregnancy, vitamin D insufficiency and deficiency, defined as serum 25-hydroxyvitamin D (25(OH)D) <50 nM, and <25 nM, respectively, may have adverse effects for both mother and child. Prevalence estimates, and identification of subgroups at special risk, may be useful for the planning of preventive strategies.

OBJECTIVE

To study the prevalence and risk factors of hypovitaminosis D in early pregnancy.

DESIGN AND METHODS

In a cross-sectional study of 1348 women in early pregnancy from the Odense Child Cohort, Denmark, 25(OH)D was determined and correlated to demographic and lifestyle variables (age, nationality, skin tone, parity, prepregnancy body mass index (BMI), smoking and sun exposure), using multiple linear and logistic regression analyses for all year, or stratified for summer and winter. The risk of vitamin D insufficiency was expressed as odds ratios (OR) with 95% confidence intervals in brackets.

RESULTS

The prevalence of vitamin D insufficiency and deficiency was estimated to 27·8% and 3·5% respectively. In adjusted analyses, vitamin D insufficiency was directly associated with winter season, OR = 1·89 (1·35-2·63); increasing prepregnancy BMI, OR = 1·06 (1·03-1·10); and smoking, OR = 2·7 (1·34-5·41); but was less frequent in nulliparous, OR = 0·47 (0·33-0·68) and tanned Caucasians, OR = 0·63 (0·41-0·97). Season-specific associations having parental origin from outside Europe in summer, OR = 4·13 (1·41-12·13); in winter smoking, OR = 3·15 (1·19-8·36); and prepregnancy BMI, OR = 1·12 (1·06-1·18).

CONCLUSIONS

Vitamin D insufficiency was widespread in early pregnancy. Associations to smoking, prepregnancy BMI and origin outside Europe varied with season. Multiparity and not being tanned in Caucasians represent new risk factors of vitamin D insufficiency.

Mechanisms of renal NaCl retention in proteinuric disease

In diseases with proteinuria, for example nephrotic syndrome and pre-eclampsia, there often are suppression of plasma renin-angiotensin-aldosterone system components, expansion of extracellular volume and avid renal sodium retention. Mechanisms of sodium retention in proteinuria are reviewed. In animal models of nephrotic syndrome, the amiloride-sensitive epithelial sodium channel ENaC is activated while more proximal renal Na(+) transporters are down-regulated. With suppressed plasma aldosterone concentration and little change in ENaC abundance in nephrotic syndrome, the alternative modality of proteolytic activation of ENaC has been explored. Proteolysis leads to putative release of an inhibitory peptide from the extracellular domain of the γ ENaC subunit. This leads to full activation of the channel. Plasminogen has been demonstrated in urine from patients with nephrotic syndrome and pre-eclampsia. Urine plasminogen correlates with urine albumin and is activated to plasmin within the urinary space by urokinase-type plasminogen activator. This agrees with aberrant filtration across an injured glomerular barrier independent of the primary disease. Pure plasmin and urine samples containing plasmin activate inward current in single murine collecting duct cells. In this study, it is shown that human lymphocytes may be used to uncover the effect of urine plasmin on amiloride- and aprotinin-sensitive inward currents. Data from hypertensive rat models show that protease inhibitors may attenuate blood pressure. Aberrant filtration of plasminogen and conversion within the urinary space to plasmin may activate γ ENaC proteolytically and contribute to inappropriate NaCl retention and oedema in acute proteinuric conditions and to hypertension in diseases with chronic microalbuminuria/proteinuria.

Calibration of a reflectance pulse oximeter in fetal lambs for arterial oxygen saturations below 70%

OBJECTIVE

To calibrate a reflectance pulse oximeter developed at the Division of Perinatal Physiology, University Hospital of Zurich, for monitoring of oxygen saturations below 70%.

METHODS

Oxygen saturation measured noninvasively with the reflectance pulse oximeter in fetal sheep was compared with oxygen saturation of arterial blood samples measured by a conventional two-wavelength oximeter. Oxygen saturation was varied by stepwise reduction of the inspired oxygen concentration of the ewe.

RESULTS

A total of 58 data pairs was obtained from five fetuses, with a range of 10.9-73.5% for the reference method and 6.8-69.2% for the pulse oximetry oxygen saturation values. There was good correlation between the results obtained by the two methods (r2 = .89, P < .0001), with a 95% confidence interval of +/- 12.5%. The pulse oximeter showed a precision of 6.2% oxygen saturation.

CONCLUSION

The Zurich reflectance pulse oximeter offers valid measurements of oxygen saturation levels down to 10% oxygen saturation, and is therefore suitable for monitoring the fetus during delivery.

Use of neural networks to diagnose acute myocardial infarction. I. Methodology

We investigated several aspects of using neural networks as a diagnostic tool: the design of an optimal network, the amount of patients' data needed to train the network, the question of training the network optimally while avoiding overfitting, and the influence of redundant variables. The specific clinical problem chosen for illustration was the diagnosis of acute myocardial infarction, given only the electrocardiogram and the concentration of potassium in serum at the time of admission. We found that, in contrast to usual practice, the termination of the training process should be based on the generalization performance and not on the training performance. We also found that a principal component analysis can be used to eliminate redundant variables, thereby reducing the data space. The diagnostic performance of the neural network we used was 78%--superior to that of linear discriminant function analysis but similar to that of quadratic discriminant function analysis.

Use of neural networks to diagnose acute myocardial infarction. II. A clinical application

We investigated the ability of neural networks to diagnose acute myocardial infarction (AMI) from laboratory data only. Several networks were trained with different combinations of data obtained at admission and within the first 12 h and 24 h after admission. The data used included the electrocardiogram (ECG) and the concentrations in serum of potassium, creatine kinase B-subunit (CKB), and lactate dehydrogenase isoenzyme 1 for 250 patients with suspected AMI. Based on admission data, the correct diagnosis was predicted for 76% of the patients in the test group from the ECG data only, and the best combination of ECG results with other variables yielded correct diagnoses for 85% of the test group. Using all of the data available within 24 h, the network predicted the correct diagnosis for 99% of the test data. Almost the same high predictability was obtained by using only two CKB values-recorded at admission and within 12 h after admission-or by using just the latter one. Neural networks and quadratic discriminant analysis performed similarly, but the neural networks were more robust for combinations with many laboratory data.

Use of neural networks to diagnose acute myocardial infarction. II. A clinical application

We investigated the ability of neural networks to diagnose acute myocardial infarction (AMI) from laboratory data only. Several networks were trained with different combinations of data obtained at admission and within the first 12 h and 24 h after admission. The data used included the electrocardiogram (ECG) and the concentrations in serum of potassium, creatine kinase B-subunit (CKB), and lactate dehydrogenase isoenzyme 1 for 250 patients with suspected AMI. Based on admission data, the correct diagnosis was predicted for 76% of the patients in the test group from the ECG data only, and the best combination of ECG results with other variables yielded correct diagnoses for 85% of the test group. Using all of the data available within 24 h, the network predicted the correct diagnosis for 99% of the test data. Almost the same high predictability was obtained by using only two CKB values-recorded at admission and within 12 h after admission-or by using just the latter one. Neural networks and quadratic discriminant analysis performed similarly, but the neural networks were more robust for combinations with many laboratory data.

Use of neural networks to diagnose acute myocardial infarction. I. Methodology

We investigated several aspects of using neural networks as a diagnostic tool: the design of an optimal network, the amount of patients' data needed to train the network, the question of training the network optimally while avoiding overfitting, and the influence of redundant variables. The specific clinical problem chosen for illustration was the diagnosis of acute myocardial infarction, given only the electrocardiogram and the concentration of potassium in serum at the time of admission. We found that, in contrast to usual practice, the termination of the training process should be based on the generalization performance and not on the training performance. We also found that a principal component analysis can be used to eliminate redundant variables, thereby reducing the data space. The diagnostic performance of the neural network we used was 78%--superior to that of linear discriminant function analysis but similar to that of quadratic discriminant function analysis.

Reflectance pulse oximetry in neonates

OBJECTIVE

To test the feasibility and reliability in neonates of an in-house reflectance pulse oximetry (RPOX) system as an alternative method of non-invasive oxygen saturation monitoring when transmission pulse oximetry (TPOX) cannot be used, as in fetal scalp oxygen saturation monitoring during delivery.

STUDY DESIGN

The study population consisted of 31 intensive care neonates. The RPOX sensor was attached by suction to the forehead, cheek, occiput and back; recordings were under simultaneous TPOX control.

RESULTS

There was close agreement between RPOX and TPOX oxygen saturation and heart rate values, even in periodic breathing. RPOX signals from the back were unreliable because of breathing artifacts. Differences in mean absolute RPOX and TPOX oxygen saturation values were due to different calibrations. Both systems were equally sensitive to motion artifacts.

CONCLUSION

As a feasible and reliable method of non-invasive oxygen saturation monitoring in neonates, RPOX has potential applications in fetal scalp monitoring.

Limitations of forehead pulse oximetry

During initial clinical tests to calibrate our reflectance pulse oximetry system, we observed serious physiologic limitations to the use of pulse oximetry in the forehead region. We present a case of simultaneous reflectance and transmission mode pulse oximetry monitoring in a child undergoing cardiac surgery for congenital cyanotic heart disease with a large intracardiac shunt. During general anesthesia, when the patient was endotracheally intubated and mechanically ventilated, the transmission mode saturation agreed well with arterial oxygen saturation measurements; but, our reflectance pulse oximeter, with the sensor applied to the forehead, displayed spuriously lower (-18%) oxygen saturations. Before and after anesthesia and surgery, there was fine agreement between reflectance and transmission mode saturation values. We suggest that the difference was caused by vasodilatation and pooling of venous blood due to compromised venous return to the heart, and a combination of arterial and venous pulsations in the forehead region. This means that the reflectance pulse oximeter measured a mixed arterial-venous oxygen saturation.

[Increased episcleral venous pressure in uveal effusion]

This paper reports on two patients with bilateral, spontaneous choroidal and retinal detachment. One patient had scleritis, while in the other no underlying eye disease was found. No signs of orbital venous congestion were detected. Although the pathogenesis of this disease is not fully known, the findings of (1) dilated episcleral veins and (2) elevated episcleral venous pressure suggest an obstruction of venous outflow from the uveal tract. The authors believe that these two findings may have been caused by shunting of venous blood anteriorly from the uveal tract into the episcleral veins due to partial obstruction of the vortex veins.

[Necrotizing sclerokeratitis following cataract extraction]

A number of cases of necrotic sclerokeratitis following eye surgery have been reported in recently published literature. The condition was presumably triggered by surgical inflammation and caused by localized occlusive vasculitis: in one case deposits of immune complexes in vessel walls were demonstrated. The authors report on three cases of necrotic sclerokeratitis which developed five to 16 months after uncomplicated cataract extraction. Clinical examination showed disappearance of vessels in affected sclera, together with tissue necrosis. None of the patients had any systemic immune disease and no immune complexes were found in biopsy material obtained from two eyes (of which one was enucleated and the other two healed under steroid treatment). In all three affected eyes cataract extraction was performed via a corneoscleral incision. In two of the three cases communicated here cataract extraction was performed in the fellow eye without any postoperative complications. Since postsurgical sclerokeratitis is presumably triggered by vascula inflammation, clear corneal incisions may be preferable to corneoscleral incisions in cases where immune reactions are anticipated.

[The role of episcleral venous pressure in the development of secondary glaucomas]

Between 1983 and 1987 the authors studied 64 patients with dilated episcleral vessels and open-angle glaucoma. Elevated episcleral venous pressure was found to be the cause of the increased intraocular pressure. Clinically, the cause of elevation of venous pressure included spontaneous carotid cavernous fistulas, Sturge-Weber syndrome, orbital tumors, endocrine exophthalmos, anterior scleritis and idiopathic cases. A variety of pathophysiological mechanisms causing elevated episcleral venous pressure were found. The clinical course and complications associated with elevated episcleral venous pressure are discussed.

Ophthalmoscopic findings in spontaneous carotid cavernous fistula: an analysis of 20 patients

The pathophysiological influence of spontaneous carotid cavernous fistula (CCF) on retinal/orbital circulation is discussed. Of 20 patients, 1 case of papilloedema, 4 of choroidal detachment (CD), 2 of exudative retinal detachment (ERD) and 3 cases of central venous thrombosis (CVT) were seen, associations that on rarely reported in the literature. An elevation in the episcleral venous pressure was measured that was more than twice the normal, which explains the ophthalmoscopic findings.

[Sturge-Weber syndrome: glaucoma with elevated episcleral venous pressure]

The present report describes clinical findings in 13 patients with Sturge-Weber syndrome, 11 of whom had glaucoma (22-52 mm Hg). The episcleral venous pressure (EVP) was high in all glaucomatous eyes (14-23 mm Hg). Tonography showed a marked decrease in outflow facility. The observations suggest that glaucoma in Sturge-Weber syndrome is caused by a) elevated episcleral venous pressure and b) secondary damage of the trabecular meshwork with reduction of outflow facility. The cause and pathogenesis of these lesions are discussed. In nine patients a diffuse angiomatosis involving most of the choroid (1.5-4.0 mm thick), as well as the episcleral perilimbal plexus, was demonstrated by ultrasonography.

[An unusual case: episcleral and orbital varicose veins]

A rare case of episcleral and orbital varix in a 16-year-old female is reported. The diagnosis was confirmed by CT scan and ultrasonography, and showed during Valsalva's maneuver a typical expansion of the orbital tumor. CT scan showed the tumor to be composed of a plexus of dilated vessels. During Valsalva's maneuver a 10 mm Hg rise in the venous pressure in the episcleral varix was noted, which suggests a direct communication between the episcleral and the orbital varices.

[Pathogenesis of unilateral dilated episcleral vessels and increase in intraocular pressure]

Between 1983 and 1986 the authors studied 5 patients with idiopathically dilated episcleral vessels. All of them had unilateral open-angle glaucoma (25-38 mm Hg) and elevated episcleral venous pressure (22-38 mm Hg). Orbital venography showed normal venous outflow. It is presumed that the etiological mechanism was a localized venous obstruction in the region of the extraocular muscles.