Computer Assisted System for Detecting Pulmonary Embolism in Lungs

A pulmonary embolism (PE) occurs when a blood artery in the lungs becomes suddenly blocked, generally owing to a blood clot. PE is a frequent life-threatening illness that should be diagnosed as soon as possible. A novel approach for automatically detecting PE in contrast-enhanced CT images is suggested in this research. To identify PE, computerized tomography (CT) is the main test to capture images. It is quick test, incursive with good quality images, enhanced contrast and multi-sliced images can be obtained. Candidate identification, feature calculation, and classification are all part of the system. The major aims of candidate detection are to include PE with even entire occlusions and to eliminate erroneous diagnosis of tissue and parenchymal disorders. When calculating characteristics, the location and structure of the pulmonary vascular tree, as well as the severity, form, and size of an embolus, are all taken into consideration. The ability of the CAD tool to identify emboli in the sectional and sub sectional pulmonary Arterial Tree (PAT) was examined.

Pulse arrival time as a surrogate of blood pressure

Various models have been proposed for the estimation of blood pressure (BP) from pulse transit time (PTT). PTT is defined as the time delay of the pressure wave, produced by left ventricular contraction, measured between a proximal and a distal site along the arterial tree. Most researchers, when they measure the time difference between the peak of the R-wave in the electrocardiogram signal (corresponding to left ventricular depolarisation) and a fiducial point in the photoplethysmogram waveform (as measured by a pulse oximeter attached to the fingertip), describe this erroneously as the PTT. In fact, this is the pulse arrival time (PAT), which includes not only PTT, but also the time delay between the electrical depolarisation of the heart’s left ventricle and the opening of the aortic valve, known as pre-ejection period (PEP). PEP has been suggested to present a significant limitation to BP estimation using PAT. This work investigates the impact of PEP on PAT, leading to a discussion on the best models for BP estimation using PAT or PTT. We conducted a clinical study involving 30 healthy volunteers (53.3% female, 30.9 ± 9.35 years old, with a body mass index of 22.7 ± 3.2 kg/m$$^{2}$$ 2 ). Each session lasted on average 27.9 ± 0.6 min and BP was varied by an infusion of phenylephrine (a medication that causes venous and arterial vasoconstriction). We introduced new processing steps for the analysis of PAT and PEP signals. Various population-based models (Poon, Gesche and Fung) and a posteriori models (inverse linear, inverse squared and logarithm) for estimation of BP from PTT or PAT were evaluated. Across the cohort, PEP was found to increase by 5.5 ms ± 4.5 ms from its baseline value. Variations in PTT were significantly larger in amplitude, − 16.8 ms ± 7.5 ms. We suggest, therefore, that for infusions of phenylephrine, the contribution of PEP on PAT can be neglected. All population-based models produced large BP estimation errors, suggesting that they are insufficient for modelling the complex pathways relating changes in PTT or PAT to changes in BP. Although PAT is inversely correlated with systolic blood pressure (SBP), the gradient of this relationship varies significantly from individual to individual, from − 2946 to − 470.64 mmHg/s in our dataset. For the a posteriori inverse squared model, the root mean squared errors (RMSE) for systolic and diastolic blood pressure (DBP) estimation from PAT were 5.49 mmHg and 3.82 mmHg, respectively. The RMSEs for SBP and DBP estimation by PTT were 4.51 mmHg and 3.53 mmHg, respectively. These models take into account individual calibration curves required for accurate blood pressure estimation. The best performing population-based model (Poon) reported error values around double that of the a posteriori inverse squared model, and so the use of population-based models is not justified.

4 Day in dry immersion reproduces partially the aging effect on the arteries as observed during 6 month spaceflight or confinement

The objectives of this study were to determine whether 4 days of dry immersion (DI) induced similar arterial aging as spaceflight and to test the impact of thigh cuffs. Eighteen subjects underwent DI; nine wore thigh cuffs. Cardiac and arterial targets were assessed by ultrasound. No significant differences were found between the groups. The left ventricle volume, stroke volume (SV), and ejection fraction decreased with DI (p < 0.001). Carotid distensibility reduced (p < 0.05), carotid to femoral arterial tree became stiffer in 33% of the subjects, and femoral artery intima media thickness increased (p < 0.05). A reduction in plasma volume is likely to have caused the observed cardiac changes, whereas the arterial wall changes are probably best explained by hypokinesia and/or environmental stress. These changes are similar but lower in amplitude than those observed in spaceflight and mimic the natural aging effect on earth. The daytime-worn thigh cuffs had no acute or chronic impact on these arterial-focused measurements.

Abstract 1122‐000236: Intravascular Delivery and Crosslinking of Photosensitive Hydrogels for Embolizing Animal Models of AVMs and Tumors

Introduction : Embolization represents a minimally invasive treatment modality for arteriovenous malformations (AVMs), tumors, and other indications, but can be limited by currently available embolic agents, in terms of safety and efficacy. Discovery of new and improved agents could lead to better treatment outcomes. The goal of this project was to test a novel embolization methodology for the treatment of AVMs and tumors. Methods : We formulated low‐viscosity, shear‐thinning hydrogel formulations which were mixed with a photo‐initator agent and non‐ionic contrast medium. We then developed a method of intravascular hydrogel delivery with photo crosslinking at the tip of the catheter, using an integrated optical fibre. This allowed for rapid transition from a low viscosity liquid to a crosslinked solid‐state hydrogel to block blood flow to the vascular target. In addition, the UV intensity can be dynamically modulated, in real‐time, to modify the degree of crosslinking and thus the viscosity of the embolic agent. We utilized the swine rete mirabile as an animal model for AVMs, and the swine renal arterial tree (inferior segmental artery) as a model for hypervascular tumors. 5 animals were utilized without prior preparation. Embolization was graded based on degree of complete obliteration of the rete nidus or the renal arterial tree. Any non‐target embolization or other complications were recorded. Follow‐up angiography was performed at the 4‐week interval. Results : With a combination of shear‐thinning properties and dynamic modulation of photo crosslinking, we show that we are able to deliver an embolic agent with a viscosity range of up to 10^4 Pa*s through a single low viscosity precursor that is injectable through microcatheters (Figure 1). Using this methodology, hydrogel embolization was technically successful in all animals. Following embolization, 4/5 rete mirabile and 5/5 inferior renal arterial trees were completely obliterated. Representative angiographic images are shown in Figures 2 and 3. There were no instances of clinical or angiographic complications. Conclusions : We demonstrated a novel method of intravascular delivery of low viscosity photosensitive hydrogels, with photo crosslinking at the tip of the catheter, to successfully embolize animal models for AVMs and tumors. This promising technology will be investigated further with longer‐term comparative animal trials.

Development of an MRI-Guided Approach to Selective Internal Radiation Therapy Using Holmium-166 Microspheres

Selective internal radiation therapy (SIRT) is a treatment modality for liver tumours during which radioactive microspheres are injected into the hepatic arterial tree. Holmium-166 (166Ho) microspheres used for SIRT can be visualized and quantified with MRI, potentially allowing for MRI guidance during SIRT. The purpose of this study was to investigate the MRI compatibility of two angiography catheters and a microcatheter typically used for SIRT, and to explore the detectability of 166Ho microspheres in a flow phantom using near real-time MRI. MR safety tests were performed at a 3 T MRI system according to American Society for Testing of Materials standard test methods. To assess the near real-time detectability of 166Ho microspheres, a flow phantom was placed in the MRI bore and perfused using a peristaltic pump, simulating the flow in the hepatic artery. Dynamic MR imaging was performed using a 2D FLASH sequence during injection of different concentrations of 166Ho microspheres. In the safety assessment, no significant heating (ΔTmax 0.7 °C) was found in any catheter, and no magnetic interaction was found in two out of three of the used catheters. Near real-time MRI visualization of 166Ho microsphere administration was feasible and depended on holmium concentration and vascular flow speed. Finally, we demonstrate preliminary imaging examples on the in vivo catheter visibility and near real-time imaging during 166Ho microsphere administration in an initial patient case treated with SIRT in a clinical 3 T MRI. These results support additional research to establish the feasibility and safety of this procedure in vivo and enable the further development of a personalized MRI-guided approach to SIRT.

Structural Knowledge Transfer of Panoptic Kidney Segmentation to Other Stains, Organs, and Species

Panoptic segmentation networks are a newer class of image segmentation algorithms that are constrained to understand the difference between instance-type objects (objects that are discrete countable entities, such as renal tubules) and group-type objects (uncountable, amorphous regions of texture such as renal interstitium). This class of deep networks has unique advantages for biological datasets, particularly in computational pathology. We collected 126 periodic acid Schiff whole slide images of native diabetic nephropathy, lupus nephritis, and transplant surveillance kidney biopsies, and fully annotated them for the following micro-compartments: interstitium, glomeruli, globally sclerotic glomeruli, tubules, and arterial tree (arteries/arterioles). Using this data, we trained a panoptic feature pyramid network. We compared performance of the network against a renal pathologist's annotations, and the method's transferability to other computational pathology domain tasks was investigated. The panoptic feature pyramid networks showed high performance as compared to renal pathologist for all of the annotated classes in a testing set of transplant kidney biopsies. The network was not only able to generalize its object understanding across different stains and species of kidney data, but also across several organ types. We conclude panoptic networks have unique advantages for computational pathology; namely, these networks internally model structural morphology, which aids bootstrapping of annotations for new computational pathology tasks.

Abstract P113: IMPACT OF ARTERIAL SITE AND PRESSURE WAVE ACQUISITION METHOD ON ARTERIAL FORM FACTOR AND MEAN AND CENTRAL ARTERIAL PRESSURE

Accurate mean arterial pressure (MAP) measurement requires an integrated arterial pressure waveform but MAP is often estimated from systolic (S) and diastolic (D) BP using a fixed form factor (FF, usually 0.33): MAP = DBP + FF*pulse pressure (PP), also solvable for FF if MAP is known. MAP is constant in the arterial tree but central (c) PP is lower than peripheral (p) PP, so cFF should be higher than pFF, and MAP and FF values should be similar among devices. We tested these principles for 2 transfer function-dependent devices: SphygmoCor (SPG), which generates high fidelity radial arterial tonometry waveforms and integrates MAP, and Mobil-O-Graph (MOB), a 24-hour ambulatory oscillometric brachial pulse wave analysis system that reports a proprietary MAP. SPG data were obtained after 5-10 minutes of rest and MOB values were obtained at 20 min intervals. For SPG (n =371) [mean(SD)]: pBP 140(28)/80(15) and cBP 126(28)/81(15) mmHg, age 48(20) yr, BMI 28(6.6), 54% women, 17% black. SPG-pFF was lower than cFF [0.34(.056) vs 0.44(.049), p<0.000]; SPG-pFF and cFF were correlated (r 2 = 0.23, p<0.000); and on stepwise multiple regression, pFF was related to MAP > gender > age (multiple-r 2 = 0.37, p<0.000; BMI and race excluded). For MOB (n = 157): pBP 135(18)/81(12) and cBP 123(16)/83(12) mmHg, age 59(16) yr, BMI 30(6), 52% women, 24% black. MOB-pFF was a fixed value [0.46(0.00098)] that was lower than cFF [0.57(0.048), p<0.000] and MOB-cFF was not correlated with MAP or any other characteristic. To compare SPG and MOB directly, 101 pairs were matched precisely (SPSS software) for pSBP and pDBP (130.9/78.7 mmHg for each device group); SPG-MAP was lower than MOB-MAP [96.3 vs 102.6 mmHg (p<0.000)] and SPG-cSBP was lower than MOB-cSBP [117.5 (18.0) vs 119.7(14.6), p=.015], with respective FF’s similar to the full groups. We conclude that: 1) FF varies within-subjects by arterial location and (at constant MAP) cFF > pFF; 2) SPG-pFF (but not MOB-pFF) varies between-subjects and is affected by MAP, gender, and age; 3) the proprietary MOB-pFF is fixed and much higher than traditional FF or SPG-pFF; and 4) compared to SPG, MOB overestimates MAP by about 5% and underestimates cSBP by about 2%. These findings have implications for FF methods and proprietary devices such as Mobil-O-Graph.