Optical Coherence Tomography-Based Functional Stenosis Assessment: FUSION-A Prospective Multicenter Trial.

BACKGROUND: Intravascular imaging and intracoronary physiology may both be used to guide and optimize percutaneous coronary intervention; however, they are rarely used together. The virtual flow reserve (VFR) is an optical coherence tomography (OCT)-based model of fractional flow reserve (FFR) facilitating the assessment of the physiological significance of coronary lesions. We aimed to validate the VFR assessment of intermediate coronary artery stenoses. METHODS: FUSION (Validation of OCT-Based Functional Diagnosis of Coronary Stenosis) was a multicenter, prospective, observational study comparing OCT-derived VFR to invasive FFR. VFR was mathematically derived from a lumped parameter flow model based on 3-dimensional lumen morphology. Patients undergoing coronary angiography with intermediate angiographic stenosis (40%-90%) requiring physiological assessment were enrolled. Investigational sites were blinded to the VFR analysis, and all OCT and FFR data were reviewed by an independent core laboratory. The coprimary end points were the sensitivity and specificity of VFR against FFR as the reference standard, each of which was tested against prespecified performance goals. RESULTS: After core laboratory review, 266 vessels in 224 patients from 25 US centers were included in the analysis. The mean angiographic diameter stenosis was 65.5%±14.9%, and the mean FFR was 0.83±0.11. Overall accuracy, sensitivity, and specificity of VFR versus FFR using a binary cutoff point of 0.80 were 82.0%, 80.4%, and 82.9%, respectively. The 97.5% lower confidence bound met the prespecified performance goal for sensitivity (71.6% versus 70%; P=0.01) and specificity (76.6% versus 75%; P=0.01). The area under the curve was 0.88 (95% CI, 0.84-0.92; P<0.0001). CONCLUSIONS: OCT-derived VFR demonstrates high sensitivity and specificity for predicting invasive FFR. Integrating high-resolution intravascular imaging with imaging-derived physiology may provide synergistic benefits as an adjunct to percutaneous coronary intervention. REGISTRATION: URL: https://clinicaltrials.gov; Unique identifier: NCT04356027.

New Volumetric Analysis Method for Stent Expansion and its Correlation With Final Fractional Flow Reserve and Clinical Outcome: An ILUMIEN I Substudy.

OBJECTIVES: This study sought to compare conventional methodology (CM) with a newly described optical coherence tomography (OCT)-derived volumetric stent expansion analysis in terms of fractional flow reserve (FFR)-derived physiology and device-oriented composite endpoints (DoCE). BACKGROUND: The analysis of coronary stent expansion with intracoronary imaging has used CM that relies on the analysis of selected single cross-sections for several decades. The introduction of OCT with its ability to perform semiautomated volumetric analysis opens opportunities to redefine optimal stent expansion. METHODS: A total of 291 lesions treated with post-stent OCT and FFR were enrolled. The expansion index was calculated by using a novel volumetric algorithm and was defined as: ([actual lumen area / ideal lumen area] × 100) for each frame of the stented segment. The minimum expansion index (MEI) was defined as the minimum value of expansion index along the entire stented segment. MEI and conventional lumen expansion metrics were compared for the ability to predict post-stent low FFR (<0.90) and DoCE at 1 year. RESULTS: There was a stronger correlation between MEI and final FFR, compared with CM and final FFR (r = 0.690; p < 0.001) versus (r = 0.165; p = 0.044). MEI was significantly lower in patients with DoCE than those without DoCE (72.18 ± 8.23% vs. 81.48 ± 11.03%; p < 0.001), although stent expansion by CM was similar between patients with and without DoCE (85.05 ± 22.19% and 83.73 ± 17.52%; p = 0.858), respectively. CONCLUSIONS: OCT analysis of stent expansion with a newly described volumetric method, but not with CM, yielded data that were predictive of both an acute improvement in FFR-derived physiology and DoCE.

The physical origins of transit time measurements for rapid, single cell mechanotyping.

The mechanical phenotype or 'mechanotype' of cells is emerging as a potential biomarker for cell types ranging from pluripotent stem cells to cancer cells. Using a microfluidic device, cell mechanotype can be rapidly analyzed by measuring the time required for cells to deform as they flow through constricted channels. While cells typically exhibit deformation timescales, or transit times, on the order of milliseconds to tens of seconds, transit times can span several orders of magnitude and vary from day to day within a population of single cells; this makes it challenging to characterize different cell samples based on transit time data. Here we investigate how variability in transit time measurements depends on both experimental factors and heterogeneity in physical properties across a population of single cells. We find that simultaneous transit events that occur across neighboring constrictions can alter transit time, but only significantly when more than 65% of channels in the parallel array are occluded. Variability in transit time measurements is also affected by the age of the device following plasma treatment, which could be attributed to changes in channel surface properties. We additionally investigate the role of variability in cell physical properties. Transit time depends on cell size; by binning transit time data for cells of similar diameters, we reduce measurement variability by 20%. To gain further insight into the effects of cell-to-cell differences in physical properties, we fabricate a panel of gel particles and oil droplets with tunable mechanical properties. We demonstrate that particles with homogeneous composition exhibit a marked reduction in transit time variability, suggesting that the width of transit time distributions reflects the degree of heterogeneity in subcellular structure and mechanical properties within a cell population. Our results also provide fundamental insight into the physical underpinnings of transit measurements: transit time depends strongly on particle elastic modulus, and weakly on viscosity and surface tension. Based on our findings, we present a comprehensive methodology for designing, analyzing, and reducing variability in transit time measurements; this should facilitate broader implementation of transit experiments for rapid mechanical phenotyping in basic research and clinical settings.

Automatic feature extraction and statistical shape model of the AIDS virus spike.

We introduce a method to automatically extract spike features of the AIDS virus imaged through an electron microscope. The AIDS virus spike is the primary target of drug design as it is directly involved in infecting host cells. Our method detects the location of these spikes and extracts a subvolume enclosing the spike. We have achieved a sensitivity of 80% for our best operating range. The extracted spikes are further aligned and combined to build a 4-D statistical shape model, where each voxel in the shape model is assigned a probability density function. Our method is the first fully automated technique that can extract subvolumes of the AIDS virus spike and be used to build a statistical model without the need for any user supervision. We envision that this new tool will significantly enhance the overall process of shape analysis of the AIDS virus spike imaged through the electron microscope. Accurate models of the virus spike will help in the development of better drug design strategies.

Shape-based regularization of electron tomographic reconstruction.

We introduce a tomographic reconstruction method implemented using a shape-based regularization technique. Spatial models of known features in the structure being reconstructed are integrated into the reconstruction process as regularizers. Our regularization scheme is driven locally through shape information obtained from segmentation and compared with a known spatial model. We demonstrated our method on tomography data from digital phantoms, simulated data, and experimental electron tomography (ET) data of virus complexes. Our reconstruction showed reduced blurring and an improvement in the resolution of the reconstructed volume was also measured. This method also produced improved demarcation of spike boundaries in viral membranes when compared with popular techniques like weighted back projection and the algebraic reconstruction technique. Improved ET reconstructions will provide better structure elucidation and improved feature visualization, which can aid in solving key biological issues. Our method can also be generalized to other tomographic modalities.

Active segmentation of 3D axonal images.

We present an active contour framework for segmenting neuronal axons on 3D confocal microscopy data. Our work is motivated by the need to conduct high throughput experiments involving microfluidic devices and femtosecond lasers to study the genetic mechanisms behind nerve regeneration and repair. While most of the applications for active contours have focused on segmenting closed regions in 2D medical and natural images, there haven't been many applications that have focused on segmenting open-ended curvilinear structures in 2D or higher dimensions. The active contour framework we present here ties together a well known 2D active contour model [5] along with the physics of projection imaging geometry to yield a segmented axon in 3D. Qualitative results illustrate the promise of our approach for segmenting neruonal axons on 3D confocal microscopy data.

Computational Inversion of Electron Tomography Images Using L2-Gradient Flows.

In this paper, we present a stable, reliable and robust method for reconstructing a three dimensional density function from a set of two dimensional electric tomographic images. By minimizing an energy functional consisting of a fidelity term and a regularization term, an L2-gradient flow is derived. The flow is integrated by a finite element method in the spatial direction and an explicit Euler scheme in temporal direction. The experimental results show that the proposed method is efficient and effective.

Deformable density matching for 3D non-rigid registration of shapes.

There exists a large body of literature on shape matching and registration in medical image analysis. However, most of the previous work is focused on matching particular sets of features--point-sets, lines, curves and surfaces. In this work, we forsake specific geometric shape representations and instead seek probabilistic representations--specifically Gaussian mixture models--of shapes. We evaluate a closed-form distance between two probabilistic shape representations for the general case where the mixture models differ in variance and the number of components. We then cast non-rigid registration as a deformable density matching problem. In our approach, we take one mixture density onto another by deforming the component centroids via a thin-plate spline (TPS) and also minimizing the distance with respect to the variance parameters. We validate our approach on synthetic and 3D arterial tree data and evaluate it on 3D hippocampal shapes.