«Novel Biophotonic Imaging Techniques for Assessing Women’s Reproductive Health by Tyler Kaine Drake Department of Biomedical Engineering Duke ...»
The cohort consisted of healthy women between the ages of 18 and 45 who had a regular menstrual cycle and were not pregnant.
There were no restrictions or ethnicity or race.
Gel distribution was measured 10 minutes after insertion of the product.
Also included in the study were two pilot measurements of long term dynamics, where measurements were made 60 minutes after insertion.
The test gel consisted of 3.5 mL of fluorescein- labeled Replens (0.1% w/w) vaginal moisturizer (Lil Drugstore).
This gel is a polycarbophil- based gel and is considered a quality biophysical 69 model of some microbicide gel products.18, 27, 28 The study protocol called for the participant to
(1) remain seated for 1 min, stand for 1 min, sit for 1 min, and then remain supine for the remaining time (referred to as sit/stand/sit);
or (2) remain supine for the entire time interval (referred to as supine).76 The following procedure was used for data collection for each of the fifteen study sessions.
First, the tube was sterilized and cleaned by submerging it in an enzymatic detergent (EmPower, Metrex Research) for at least 1 min.
Next, the tube was submerged in CIDEX OPA solution (Advanced Sterilization Products) for a minimum of 12 minutes.
It was then rinsed with water for three 1 min rises, wiped with alcohol and dried.
The women entered the examination room and assumed a supine position on the exam table.
Initially, a background scan was performed and the tube was then lubricated for insertion with a small amount of nonlabeled Replens gel.
The physician then inserted the tube and endoscope assembly (Figure 3.23(b)), positioned the endoscope to the vaginal fornix, and adjusted an opaque flange which serves to block ambient light from the exam room.
The endoscope is then zero positioned at the fornix, and the physician manually advances the endoscope inside the tube which remains stationary.
Data are collected every 1- mm axially for mLCI and every 5- mm axially for fluorimetry.
Once the endoscope is advanced to the introitus, it is rotated azimuthally inside the tube 45- deg and the scan continues axially back towards the fornix, as in Figure 3.16.
This process is repeated until a 360- deg sweep of the vaginal epithelium is completed.76 70 Following the background scan, 3.5 mL of fluorescein- labeled Replens gel was inserted into the participant’s vagina via a syringe- type applicator.
The gel was inserted towards the fornix.
The participant then underwent one of the posture protocols (sit/stand/sit or supine), the probe was re- inserted into the vagina to the same depth as the background scan, and the imaging procedure was completed.
The data set typically consisted of about 160 measurements sites for fluorimetry and 800 measurement points for mLCI.
These numbers varied with the changes in anatomy between participants, but 8 azimuthal angles were scanned and about 75 to 100 mm of axial tissue surface was examined.
The procedure lasted on average 49 min per participant for the 10 min protocol and 1 h 40 min for the 60 min protocol.76
4.3 Data analysis and summary measures of coating As described in Section 3.2.4, a few steps must be taken to register the data between the modalities and compensate for the field of view variations.
First, the mLCI data is rotated 180- deg and shifted
- 20 mm to reconcile the mLCI imaging data with fluorimetry77.
Five consecutive mLCI scans are then averaged to compensate for variations due to sampling and the larger field of view of the fluorimetry device.
The background scan was then subtracted from the mLCI test scan to remove any effects due to the non- labeled Replens gel, which was used for lubrication during insertion.
A set of summary measurements, adopted from Henderson et al., was used to evaluate features of gel coating thickness distributions.27 The summary measures were 71 designed in order to compare vaginal drug delivery based on coating distributions for different gels.
However, unlike Henderson et al., the current study is aimed at assessing the feasibility of using mLCI for measuring gel coating thicknesses and not comparing the performances of different gel products.27 In addition to fraction of vaginal tissue surface with coating, the primary measure used by Henderson et al., a new measure is
azimuthally averaged coating thickness as a function of position along the vaginal canal.
This measure graphically represents the mass distribution of gel after insertion, and provides an indication of how far the gel has spread down the canal from application near the fornix.
4.4 Results The results from the fifteen study sessions are provided below in Table 4.1.
A total of 22,344 mLCI scans were collected in the study, but 23.7% were rejected because of low signal intensity, resulting in a total of 17, 048 useable mLCI scans.
The fluorimetric data set consisted of 2232 scans, all of which had sufficient signal for data analysis.
Average fraction of tissue with coating as measured by mLCI was 0.667 ± 0.206 (mean ± std dev), and fluorimetry measured 0.645 ± 0.222.
These results were found to be not differ statistically (p = 0.9771) using a paired t- test after an arcsin transformation of the percentages.
The average absolute difference in fraction of surface with coating between the modalities was found to be 0.065 ± 0.053.
The results of the two modalities are compared in Figure 4.1, on a per experiment basis.
A linear regression was performed and a line- of- best fit was calculated.
The line had a slope of 0.9967 (p = 0.9771) which was determined to not be statistically different 73 from unity.
An intercept value of
- 0.0197 (p = 0.8055) was found which was not statistically different from zero.
Fraction of vaginal surface with coating, as seen in Drake et al.76 mLCI values are on the x- axis with fluorimetric on the y- axis.
The shaded region indicates the 95% confidence curve for the fit, and the dotted lines show the 95% confidence interval for the individual observations.
Example data of a typical coating thickness distribution acquired at the 10 min interval after insertion is shown below in Figure 4.2.
The protocol also called for a sit/stand/sit routine for this data set.
The data are presented as a topological map, with the colorbar indicating gel thickness.
Example human in vivo gel coating measurements from the dual- modality instrument.
This data is from a sit/stand/sit protocol with a 10 min interval.
The colorbar reveals coating thickness from the fluorimetric technique (left) and the mLCI device (right).
The red line demarcates the margin for areas with coating greater than 100 µ;m.
The contour lines are binned at 50 µ;m changes in coating.
Figure taken from Drake et al.76 The zero- degree angle of the x- axis in Figure 4.2, is aligned to the vertical position as the participant is supine on the exam table.
The axial location of 20 mm is located closest towards the fornix and gel insertion location.
Figure 4.3 shows the azimuthally averaged coating thickness as a function of distance from the fornix for six experiments (3 sit/stand/sit;
The mLCI data were truncated at 20 mm in Figure 4.2 and Figure 4.3 to match the fluorimetric dataset.
mLCI failed to measure full thickness at large coating layers, but the modalities show agreement in the length of axial coating.
Furthermore, the axial position at which azimuthally averaged coating fell below 100 µ;m is an important metric which is plotted for both modalities in Figure 4.4.
The mLCI extent is on the x- axis and fluorimetry is on the y- axis and a linear regression was performed for the scatterplot.
The line- of- best fit had a slope of 0.9155 (p = 0.7896) which does not differ statistically from unity, and a y- intercept of
- 5.5449 (p = 0.3345) which does not differ statistically from zero.
The regression shows no systematic offset between the data and the relationship between measurements for the two modalities is linear, as expected.
Axial extent at which azimuthally averaged coating falls below 100 µ;m thickness.
mLCI is again on the x- axis and fluorimetry is plotted on the y- axis.
The shaded region indicates the 95% confidence interval for the fit, and the dotted region shows the 95% confidence interval for the individual observations.
Figure is from Drake et al.76 Additional coating metrics were also assessed and these are provided in Table 4.2.
In this table, percentages of tissue surface with coating thicknesses less than 100 µ;m were found, as well as the average axial extent of azimuthally averaged coating with thicknesses greater than 100 µ;m.
From Table 4.2, it is concluded that mLCI and fluorimetry measure similar fractions of surface with low coatings.
The modalities also agreed in identifying the point at which azimuthally averaged coating falls below 100 µ;m.
77 Table 4.2:
Fraction of coating thickness distribution with less than 100 µ;m thickness and axial extent of azimuthally averaged coating with thickness greater than 100 µ;m, as shown in Drake et al.76 (Results presented as mean ± SEM)
4.5 Discussion Studying the in vivo physical distribution of microbicide gels is critical in planning their rational design.
Coating extent and uniformity in the vagina govern the release and transport of APIs to target fluids and tissues, and also forms a physical barrier to HIV migration into the epithelial tissue surface.
The dual- modality optical imaging instrument is capable of measuring details of microbicide gel coating in vivo with depth resolution on the order of 10 µ;m.
This data can be used in objective computations of candidate microbicide gel performance.
Measured fraction of vaginal surface with coating was similar between the modalities, but mLCI was limited in its 78 ability to measure relatively large local coatings (500 µm).
This was partly due to the optical aberrations caused by the probe’s polycarbonate endoscopic tube (addressed in Section 3.2.4), which has been improved for future studies.
Also, the mLCI field of view is much smaller than the fluorimetric, so mLCI detects local variations in coating more finely than fluorimetry.
At thick coating layers, where mLCI points are lost due to imaging depth limitations, fewer mLCI data points are averaged in the processing routine, and the effect of local coating variations has a greater influence.
mLCI is advantageous in measuring microbicide gel coating thicknesses, however, in that there is no need for exogenous contrast agents to be added to the gel.
These contrast agents can limit the accuracy of coating measurements since they may diffuse out of the gel and into adjacent tissues and fluids, falsely indicating that gel is present.
Photobleaching, or photochemical destruction of the contrast fluorophore, also may occur which leads to additional inaccuracies in measurement.
Label- free modalities, like mLCI, are preferable for extended time studies, which are biologically relevant in understanding gel mechanisms for delivering APIs to target tissues.
Chapter4 presented the results of an in vivo human trial which measured vaginal microbicide gel thickness distributions.
The study used the dual- modality optical imaging instrument in imaging a cohort of 9 participants that completed 15 study sessions.
Microbicide gel thickness measurements from fluorimetry and mLCI were 79 compared in order to assess the ability of mLCI in measuring such distributions as compared to the standard of fluorimetry.
It was concluded that fraction of vaginal tissue with coating values found by the separate modalities were not statistically different (p = 0.9771), and the mLCI device successfully measured in vivo gel thickness distributions.
mLCI imaging is also advantageous in that it covers broad tissue area without the need of exogenous contrast agents.
This could allow extended time studies to be performed in the future, where gel thickness measurements are made at long time intervals (6- 12 hours) after insertion of the product.
Such studies would be useful in providing thickness distribution information of gels as they would be used in real- world practice.
80 5 Measuring dilution of microbicide gels
5.1 Introduction An in vivo clinical trial using the dual- modality optical imaging instrument to measure microbicide gel thickness distribution in the vaginal lumen was discussed in Chapter 4.
Here, the instrument is used in a different application for a proof of concept study.
The dual- modality instrument is used to measure dilution of a microbicide gel.
The method compares gel thickness measurements from fluorimetry and mLCI to calculate dilution of the gel.
As a microbicide gel becomes more dilute, its fluorimetry measurement decreases in signal intensity, as its mLCI measurement remains constant.
The difference between the measurements can then be related to the extent of dilution.
A preliminary validation study is performed and it is described in this chapter.
The study design is presented as well as unique data processing needed for this application.
Placebo gel measurements were taken in a calibration socket and the results are presented.
Changes in the slope of correlation between the measurements can be related to dilution, and a calibration curve is generated in Section 5.3, by repeating measurements with serial dilutions with a vaginal fluid simulant (VFS).
An example dilution calculation is presented in Section 5.3, and the study is summarized.
This methodology can provide valuable dilution information on candidate microbicide products in order to further study their in vivo behavior.