Comparative assessment of synthetic time series generation approaches in healthcare: leveraging patient metadata for accurate data synthesis.

BACKGROUND: Synthetic data is an emerging approach for addressing legal and regulatory concerns in biomedical research that deals with personal and clinical data, whether as a single tool or through its combination with other privacy enhancing technologies. Generating uncompromised synthetic data could significantly benefit external researchers performing secondary analyses by providing unlimited access to information while fulfilling pertinent regulations. However, the original data to be synthesized (e.g., data acquired in Living Labs) may consist of subjects' metadata (static) and a longitudinal component (set of time-dependent measurements), making it challenging to produce coherent synthetic counterparts. METHODS: Three synthetic time series generation approaches were defined and compared in this work: only generating the metadata and coupling it with the real time series from the original data (A1), generating both metadata and time series separately to join them afterwards (A2), and jointly generating both metadata and time series (A3). The comparative assessment of the three approaches was carried out using two different synthetic data generation models: the Wasserstein GAN with Gradient Penalty (WGAN-GP) and the DöppelGANger (DGAN). The experiments were performed with three different healthcare-related longitudinal datasets: Treadmill Maximal Effort Test (TMET) measurements from the University of Malaga (1), a hypotension subset derived from the MIMIC-III v1.4 database (2), and a lifelogging dataset named PMData (3). RESULTS: Three pivotal dimensions were assessed on the generated synthetic data: resemblance to the original data (1), utility (2), and privacy level (3). The optimal approach fluctuates based on the assessed dimension and metric. CONCLUSION: The initial characteristics of the datasets to be synthesized play a crucial role in determining the best approach. Coupling synthetic metadata with real time series (A1), as well as jointly generating synthetic time series and metadata (A3), are both competitive methods, while separately generating time series and metadata (A2) appears to perform more poorly overall.

Hemodynamic Impact of the C-Pulse Cardiac Support Device: A One-Dimensional Arterial Model Study.

The C-Pulse is a novel extra-aortic counter-pulsation device to unload the heart in patients with heart failure. Its impact on overall hemodynamics, however, is not fully understood. In this study, the function of the C-Pulse heart assist system is implemented in a one-dimensional (1-D) model of the arterial tree, and central and peripheral pressure and flow waveforms with the C-Pulse turned on and off were simulated. The results were studied using wave intensity analysis and compared with in vivo data measured non-invasively in three patients with heart failure and with invasive data measured in a large animal (pig). In all cases the activation of the C-Pulse was discernible by the presence of a diastolic augmentation in the pressure and flow waveforms. Activation of the device initiates a forward traveling compression wave, whereas a forward traveling expansion wave is associated to the device relaxation, with waves exerting an action in the coronary and the carotid vascular beds. We also found that the stiffness of the arterial tree is an important determinant of action of the device. In settings with reduced arterial compliance, the same level of aortic compression demands higher values of external pressure, leading to stronger hemodynamic effects and enhanced perfusion. We conclude that the 1-D model may be used as an efficient tool for predicting the hemodynamic impact of the C-Pulse system in the entire arterial tree, complementing in vivo observations.

MRI Assessment of Diastolic and Systolic Intraventricular Pressure Gradients in Heart Failure.

A deep phenotypic characterization of heart failure (HF) is important for a better understanding of its pathophysiology. In particular, novel noninvasive techniques for the characterization of functional abnormalities in HF with preserved ejection fraction are currently needed. While echocardiography is widely used to assess ventricular function, standard echocardiographic techniques provide a limited understanding of ventricular filling. The application of fluid dynamics theory, along with assessments of flow velocity fields in multiple dimensions in the ventricle, can be used to assess intraventricular pressure gradients (IVPGs), which in turn may provide valuable insights into ventricular diastolic and systolic function. Advances in imaging techniques now allow for accurate estimations of systolic and diastolic IVPGs, using noninvasive methods that are easily applicable in clinical research. In this review, we describe the basic concepts regarding intraventricular flow measurements and the derivation of IVPGs. We also review existing literature exploring the role of IVPGs in HF.

Performance comparison of ultrasound-based methods to assess aortic diameter and stiffness in normal and aneurysmal mice.

OBJECTIVE: Several ultrasound-based methods are currently used to assess aortic diameter, circumferential strain and stiffness in mice, but none of them is flawless and a gold standard is lacking. We aimed to assess the validity and sensitivity of these methods in control animals and animals developing dissecting abdominal aortic aneurysm. METHODS AND RESULTS: We first compared systolic and diastolic diameters as well as local circumferential strains obtained in 47 Angiotensin II-infused ApoE(-/-) mice with three different techniques (BMode, short axis MMode, long axis MMode), at two different abdominal aortic locations (supraceliac and paravisceral), and at three different time points of abdominal aneurysm formation (baseline, 14 days and 28 days). We found that short axis BMode was preferred to assess diameters, but should be avoided for strains. Short axis MMode gave good results for diameters but high standard deviations for strains. Long axis MMode should be avoided for diameters, and was comparable to short axis MMode for strains. We then compared pulse wave velocity measurements using global, ultrasound-based transit time or regional, pressure-based transit time in 10 control and 20 angiotensin II-infused, anti-TGF-Beta injected C57BL/6 mice. Both transit-time methods poorly correlated and were not able to detect a significant difference in PWV between controls and aneurysms. However, a combination of invasive pressure and MMode diameter, based on radio-frequency data, detected a highly significant difference in local aortic stiffness between controls and aneurysms, with low standard deviation. CONCLUSIONS: In small animal ultrasound the short axis view is preferred over the long axis view to measure aortic diameters, local methods are preferred over transit-time methods to measure aortic stiffness, invasive pressure-diameter data are preferred over non-invasive strains to measure local aortic stiffness, and the use of radiofrequency data improves the accuracy of diameter, strain as well as stiffness measurements.

Association between thigh muscle development and the metabolic syndrome in adults.

BACKGROUND: Previous research has demonstrated an association between the metabolic syndrome (MS) and muscle mass; however, no studies have shown any relationship with a particular segment of the body, which would be more useful in clinical settings. AIMS: To investigate the association between muscle development of different segments of the body and presence of the MS in adults. METHODS: We used fractionation of body mass to calculate the development of muscle mass and correlated this with presence of the MS in a cross-sectional study in adults. RESULTS: The mean age and body mass index were 42.7 ± 6.6 years and 25.3 ± 3.7 kg/m(2), respectively. 23.1% of adults suffered from the MS. After adjusting for multiple variables, the Z score of both thigh and chest muscle girths were significantly associated with the MS. There were significant differences between adults with or without the MS in the Z score of thigh [-0.686; 95% confidence interval (95% CI) -1.020 to -0.351], mid-thigh (-0.566; 95% CI -0.931 to -0.200) and chest (0.611; 95% CI 0.260-0.962) girths. CONCLUSIONS: There is an association between muscle development and the MS; moreover, muscle thigh perimeter was larger in those without the MS. The use of muscle development of the thigh as an indicator of cardiovascular health-related metabolic alterations is proposed.

Pituitary tumors in patients with MEN1 syndrome.

We briefly review the characteristics of pituitary tumors associated with multiple endocrine neoplasia type 1. Multiple endocrine neoplasia type 1 is an autosomal-dominant disorder most commonly characterized by tumors of the pituitary, parathyroid, endocrine-gastrointestinal tract, and pancreas. A MEDLINE search for all available publications regarding multiple endocrine neoplasia type 1 and pituitary adenomas was undertaken. The prevalence of pituitary tumors in multiple endocrine neoplasia type 1 may vary from 10% to 60% depending on the studied series, and such tumors may occur as the first clinical manifestation of multiple endocrine neoplasia type 1 in 25% of sporadic and 10% of familial cases. Patients were younger and the time between initial and subsequent multiple endocrine neoplasia type 1 endocrine lesions was significantly longer when pituitary disease was the initial manifestation of multiple endocrine neoplasia type 1. Tumors were larger and more invasive and clinical manifestations related to the size of the pituitary adenoma were significantly more frequent in patients with multiple endocrine neoplasia type 1 than in subjects with non-multiple endocrine neoplasia type 1. Normalization of pituitary hypersecretion was much less frequent in patients with multiple endocrine neoplasia type 1 than in subjects with non-multiple endocrine neoplasia type 1. Pituitary tumors in patients with multiple endocrine neoplasia type 1 syndrome tend to be larger, invasive and more symptomatic, and they tend to occur in younger patients when they are the initial presentation of multiple endocrine neoplasia type 1.

Neurocysticercosis, meningioma, and silent corticotroph pituitary adenoma in a 61-year-old woman.

We report here the case of a 61-year-old woman who presented with hydrocephalus and cystic and solid lesions in sella turcica, suprasellar areas, and third ventricle. After ventriculoperitoneal shunt she developed cognitive changes and the cystic lesions enlarged. Magnetic resonance imaging (MRI) demonstrated multiple cysts and a solid lesion in the sella and around the anterior clinoid process. With diagnosis of neurocysticercosis she underwent craniotomy. Pathologic examination documented two different lesions: viable and dead cysticerci with inflaming infiltration and a left anterior clinoidal meningioma. At the second surgery, six weeks later via transnasal transsphenoidal approach a silent corticotroph pituitary adenoma was removed which was studied by histology, immunohistochemistry, and electron microscopy. To our knowledge, the occurrence of these three different lesions in the sellar area was not described before.

Pituitary tumors in patients with MEN1 syndrome

We briefly review the characteristics of pituitary tumors associated with multiple endocrine neoplasia type 1. Multiple endocrine neoplasia type 1 is an autosomal-dominant disorder most commonly characterized by tumors of the pituitary, parathyroid, endocrine-gastrointestinal tract, and pancreas. A MEDLINE search for all available publications regarding multiple endocrine neoplasia type 1 and pituitary adenomas was undertaken. The prevalence of pituitary tumors in multiple endocrine neoplasia type 1 may vary from 10% to 60% depending on the studied series, and such tumors may occur as the first clinical manifestation of multiple endocrine neoplasia type 1 in 25% of sporadic and 10% of familial cases. Patients were younger and the time between initial and subsequent multiple endocrine neoplasia type 1 endocrine lesions was significantly longer when pituitary disease was the initial manifestation of multiple endocrine neoplasia type 1. Tumors were larger and more invasive and clinical manifestations related to the size of the pituitary adenoma were significantly more frequent in patients with multiple endocrine neoplasia type 1 than in subjects with non-multiple endocrine neoplasia type 1. Normalization of pituitary hypersecretion was much less frequent in patients with multiple endocrine neoplasia type 1 than in subjects with non-multiple endocrine neoplasia type 1. Pituitary tumors in patients with multiple endocrine neoplasia type 1 syndrome tend to be larger, invasive and more symptomatic, and they tend to occur in younger patients when they are the initial presentation of multiple endocrine neoplasia type 1.