Diagnostic imaging capabilities of the Ocelot -Optical Coherence Tomography System, ex-vivo evaluation and clinical relevance

OCT devices description

The Ocelot system is designed to cross chronic total occlusions (CTOs) facilitated
by OCT in the peripheral vasculature. The Dragonfly System is intended for OCT guided
diagnostic imaging of the coronary arteries. Both imaging consoles use swept-source
optical coherence tomography and obtain a radial tomograph of the artery by acquiring
individual A-lines as the optical beam is rotated along the axis of the catheter inside
the artery. Table 1 outlines the comparison for the basic operational features of the Ocelot and the
Dragonfly devices. The two systems differ with regards to the A-line acquisition modes,
the positioning of the optical fiber on the catheter, the catheter rotation speed,
the optical fiber rotation speed, and the modes of OCT image display.

Table 1. Ocelot and Dragonfly OCT system technical specifications

Tissue preparation

In order to compare the diagnostic information obtained by the Ocelot and Dragonfly
OCT systems we utilized human cadaveric vessels: the arterial segments were excised
from the superficial femoral artery to the peroneal and sectioned into individually
labeled segments (Fig. 1).

Fig. 1. Human cadaveric vessel segments that were utilized to compare the diagnostic information
obtained by the Ocelot and Dragonfly OCT systems

In order to eliminate the need for blood dispersion and enable the ex-vivo arterial
imaging, the cadaveric arterial segments were held fixed in saline bath solutions.

The cadaveric arteries were procured from: Science Care, 21210 N19th Ave. Phoenix,
AZ 85027. All donors (or their next of kin) consented in writing, and all the written
consents are on file with Science Care.

The donor procurement was in compliment with the Declaration of Helsinki and the local
research ethics committee: Medical Affairs at Avinger, approved the donation protocol
as well as the study utilization.

Device tissue imaging

Both the Ocelot and Dragonfly catheters were inserted into identical vessel segments
for image acquisition and comparison. The image acquisition was performed according
to standard operational procedures respective for each device. The cadaveric vessel
segment subsets that were used for comparative image analysis were chosen based on
screening for the following features (further described in Table 2): intact vessel-wall architecture of intima, media and adventitia in a single segment
(defined as Layered Structure, Table 2) well preserved external elastic lamina (EEL), arterial segment with bifurcation,
dissection, stent, or demonstrating arterial disease, including calcium, fibrin, lipid,
thrombus, or a combination thereof. Table 2 also outlines the nomenclature that was used to standardize the OCT imaging evaluation.
Accordingly, a total of 52 image sets (from n?=?9 donors and 36 vessels) were selected and OCT images were acquired with the Ocelot
and Dragonfly catheters. Figs. 2, 3, 4, 5 and 6 depict a representative images obtained with the Ocelot OCT system (Left panel -A),
images obtained with the Dragonfly OCT system (Middle panel -B), and the matching
histology slides (Right panel -C). Figure 2 shows representative images of the following features: Layered Structures (LS), e.g.
the vessel-wall architecture of intima, media and adventitia (LS, Fig. 2), arterial Bifurcation or arterial Branch (b, Fig. 2) and the External Elastic Lamina (EEL, Fig. 2). Concordantly, Fig. 3 shows representative OCT image of an arterial dissection as captured by the Ocelot
OCT system (Fig. 3, left panel -a), by the Dragonfly OCT system (Fig. 3, middle panel -b) and the matching histology slide (Fig. 3, right panel -c). Figure 4 shows representative imaging of a stented artery segment as captured by Ocelot (Fig. 4, left panel -a), or by Dragonfly OCT (Fig. 4, middle panel -b) systems. The matching histological slide (Fig. 4, right panel -c) shows representative slide of a stented arterial segment, notice the histological
processing necessitated the removal of the metallic stent from the fixed artery. Figure 5 shows representative image of calcification in the arterial segment as captured by
the Ocelot OCT system (Fig. 5, left panel -a), by the Dragonfly OCT system (Fig. 5, middle panel -b) and the matching histology slide (Fig. 5, right panel -c). Representative atherosclerotic plaque in the arterial segment is shown in Fig. 6. The boundary of the fibrous cap encapsulating the necrotic core and the distortion
of the normal vessel-wall architecture are apparent from the representative histological
slide (Fig. 6, left panel –c).

Table 2. Nomenclature for OCT evaluated features

Fig. 2. Arterial segments, ex-vivo, depicting Layered Structure (LS), Bifurcation (B), and
External Elastic Lamina (EEL) as seen in Ocelot (left), Dragonfly (center) and histology
(right)

Fig. 3. Ex-vivo segment with Dissection (D) seen in Ocelot (left), Dragonfly (center) and
histology (right)

Fig. 4. Stented cadaveric arterial segment, Non-Layered Structure (NLS) and External Elastic
Lamina (EEL) as seen in Dragonfly (left), Ocelot (center) and histology (right) images

Fig. 5. Ex-vivo segment with Non Layered Structures (Calcification)

Fig. 6. Ex-vivo segment with Non Layered Structures (Necrotic Core with Fibrous Cap)

Representative images of these features are shown respectively, as captured by the
Ocelot OCT imaging system (Fig. 6, left panel-a) or the Dragonfly OCT system (Fig. 6, middle panel-b).

Figures 7 and 8 demonstrate how physician interpreters were presented the image and answer options
for a matched image set. These images appeared independently and were randomly ordered
within the full test cohort.

Fig. 7. Sample Test Question (Ocelot)

Fig. 8. Sample Test Question (Dragonfly)

Study design

The cadaveric study was designed to evaluate the concordance between the Ocelot and
the Dragonfly OCT systems.

Three independent physician interpreters with significant experience using both Dragonfly
and Ocelot OCT systems (Table 3) evaluated the images for the presence of each feature, based on their clinical understanding
of OCT and the published literature 15], 18].

Table 3. Independent Physician Interpreter OCT Experience

In order to prevent any bias and or comparison amongst image sets, all images were
presented in a random fashion within each system, meaning the image sets were not
shown together. Each physician interpreter was blinded to a unique sequence of the
images. During the physicians review process, an electronic survey (Survey Gizmo,
Boulder, CO, USA) captured all answer sets.

The clinical images were obtained during therapeutic interventions that included CTO
crossing using the Ocelot catheter. The diagnostic information was obtained after
an informed consent.

Histological analysis

Third party histologic evaluation (Pathology Research Laboratories, South San Francisco,
CA) was performed using core lab standard rating scales.

Statistical analysis

Statistical boundaries for image comparison were set in concordance with diagnostic
imaging standards. Tolerating a less than or equal to ?20.0 % difference in sensitivity
(positive features identification) or specificity (false positive features identification)
for imaging performance between Ocelot and Dragonfly was deemed acceptable for an
individual arterial feature. The composite across all arterial features was tightened
to less than or equal to ?15.0 % to match OCT vessel features with verified histology,
establishing a minimum of 85 % accuracy for both Ocelot and Dragonfly.

Sample size determination was based on the binomial distribution of one-sided 95 %
lower confidence boundary for the sensitivity difference between Ocelot and Dragonfly
with a minimum number of matched sets providing a high probability (80 %) of success
with a specified assessment compared to histology.

The calculated lower bound required a minimum of 15 matched image sets, one Ocelot
and one Dragonfly fly image, per arterial feature. A total sample size of 52-matched
image sets was used to meet the requisite individual and composite feature requirements
or analysis.

Following blinded physician analysis, kappa statistics were calculated to determine
the significance of inter-observer variability across arterial feature identification.