Preliminary measurements of optical properties in human abscess cavities prior to methylene blue photodynamic therapy

Abstract

As part of our ongoing Phase 1 clinical trial to establish the safety and feasibility of methylene blue photodynamic therapy (MB-PDT) for human deep tissue abscess cavities, we have shown that determination of abscess wall optical properties is vital for the design of personalized treatment plans aiming to optimize light dose. To that end, we have developed and validated an optical spectroscopy system for the assessment of optical properties at the cavity wall, including a compact fiber-optic probe that can be inserted through the catheter used for the standard of care abscess drainage. Here we report preliminary findings from the first three human subjects to receive these optical spectroscopy measurements. We observed wide variability in concentrations of oxy- and deoxy-hemoglobin prior to MB administration, ranging from 7.3–213 μM and 0.1–47.2 μM, respectively. Reduced scattering coefficients also showed inter-patient variability, but recovered values were more similar between subjects (5.5–10.9 cm⁻¹ at 665 nm). Further, methylene blue uptake was found to vary between subjects, and was associated with a reduction in oxygen saturation. These measured optical properties, along with pre-procedure computed tomography (CT) images, will be used with our previously developed Monte Carlo simulation framework to generate personalized treatment plans for individual patients, which could significantly improve the efficacy of MB-PDT while ensuring safety.

1. INTRODUCTION

Deep tissue abscesses develop as a result of host inflammatory response to acute microbial infection, which leads to the formation of a fibrous pseudocapsule enclosing the pyogenic infection[1]. If left untreated, these can result in significant morbidity and mortality[2]. Despite widespread adoption of image-guided percutaneous drainage and perioperative antibiotics, response can vary widely due to differences in pathology and antibiotic resistance[3–5]. Based on this need for improved treatment strategies, we have commenced a phase 1 clinical trial aimed at establishing the safety and feasibility of photodynamic therapy (PDT) with methylene blue (MB) at the time of percutaneous abscess drainage (ClincalTrials.gov Identifier: NCT02240498).

Photodynamic therapy relies on the combination of a photosensitizer, visible illumination at a particular wavelength, and molecular oxygen to generate cytotoxic reactive oxygen species[6]. Assuming a sufficient oxygen supply, the antimicrobial efficacy of PDT is based upon two dose components: drug dose and light dose. Drug dose consists of the concentration of photosensitizer at the treatment site, and light dose consists of the fluence rate and fluence of illumination delivered to the target. For the case of light dose deposited in tissue, this is largely determined by the tissue optical properties: absorption and scattering. Multiple studies have shown the importance of patient-specific measures of tissue optical properties for PDT treatment planning purposes[7–9], as these can vary widely between individuals[10].

This is particularly true for abscesses, as morphology can be quite distinct between subjects and the integrating sphere effect can result in fluence rates much different than those predicted purely by geometry or diffusion of light[11]. In Monte Carlo simulation studies, we have shown that eligibility for abscess PDT and the necessary treatment parameters are highly dependent upon abscess wall optical properties[12,13]. However, the optical properties of human abscess cavities have never been measured. In order to remedy this, we have constructed and validated a spatially-resolved diffuse reflectance spectroscopy system that is capable of extracting tissue optical properties using a fiber-optic probe placed in contact with the abscess wall[14]. As part of our phase 1 clinical trial, we have recently begun collecting these diffuse reflectance data before and after administration of MB. The goal is to assess variability in native abscess wall optical properties, as well as MB uptake at the abscess wall. It is expected that these optical properties will diverge widely, as abscesses can form in a variety of tissue structures as the result of myriad bacterial species.

Here, we report preliminary diffuse reflectance spectroscopy results and recovered optical property spectra from three subjects enrolled in our phase 1 clinical trial. These results are presented for measurements made prior to and following administration of methylene blue, and represent the first optical property measurements performed in human abscesses.

2. MATERIALS AND METHODS

2.1. Clinical trial protocol

The goal of our phase 1 clinical trial (ClinicalTrials.gov Identifier: NCT02240498) is to establish the safety and feasibility of MB-PDT at the time of percutaneous abscess drainage. For each subject, routine image-guided percutaneous drainage of the abscess cavity is first performed. The fiber-optic probe described below is then inserted through the drainage catheter and advanced until it makes gentle contact with the cavity wall. Diffuse reflectance spectra are then captured sequentially at each of the detector fibers (Pre-MB spectroscopy). Subsequently, the fiber-optic probe is removed and 1 mg/mL MB (BPI Labs, LLC, Largo, FL) is instilled into the cavity and allowed to incubate for 10 minutes. MB is then aspirated, and the cavity is flushed twice with sterile saline. Reflectance measurements are again performed as described above (Post-MB spectroscopy). Finally, the cavity is filled with 1% Intralipid (Fresenius Kabi AG, Bad Homburg, Germany), and treatment light is delivered at 665 nm (ML7710-PDT, Modulight, Inc., Tampere, Finland) using a flat-cleaved optical fiber (VariLase Platinum Bright Tip, Vascular Solutions LLC, Minneapolis, MN) placed at the approximate center of the abscess cavity. Illumination time is escalated using a 3+3 study design[15], with times ranging from 5–30 minutes. For the subjects described here, the first received 10 minutes of illumination while the remaining two received 15 minutes of illumination. Finally, the treatment fiber is removed, Intralipid is aspirated, and the cavity is flushed twice with sterile saline.

As part of our ongoing trial, eight subjects have been treated with MB-PDT, with spatially-resolved diffuse reflectance spectroscopy performed in the most recent three of these. Subjects provided written informed consent, and all study procedures were approved by the University of Rochester Medical Center’s Research Subjects Review Board.

2.2. Spatially resolved diffuse reflectance measurements

Our optical spectroscopy system and fiber-optic probe were described in detail previously[14]. In short, the fiber-optic probe employed consists of one source fiber and eight detector fibers encapsulated in a flexible assembly with an outer diameter (OD) of 2mm. This allows for the probe to pass through the 12 French (F) drainage catheter typically employed for the standard of care procedure. The distal end of the fiber optic probe was designed in a “plus” pattern with spacings of 300 μm between adjacent fiber centers. The source-detector separations in the actual manufactured probe ranged from 317 μm to 1308 um (Figure 2), determined by imaging the distal end of the probe alongside a US air force resolution target with a stereomicroscope (SMZ1500, Nikon Instruments, Melville, NY). A fiber-coupled broadband tungsten halogen lamp (HL-2000-HP-FHSA, Ocean Optics, Inc., Largo, FL) was used as the source for the spectroscopy measurements, and spectra were captured sequentially for each detector fiber using a thermoelectrically-cooled spectrometer (QE Pro, Ocean Optics). Prior to spectroscopy measurements, a calibration measurement was made by placing the probe’s distal end flush with the input port of an integrating sphere (3P-GPS-020-SF, Labsphere, Inc., North Sutton, NH), and capturing spectra at each detector fiber with and without the system lamp enabled. Note that while the optical spectroscopy system has fluorescence capability[16], we only present diffuse reflectance results here.

Figure 2.

Fiber-optic probe face with source fiber (S) and detector fibers (D1-D8) circled in red. Source detector separations are indicated for each detector fiber.

2.3. Data analysis

Prior to optical property recovery, both clinical and calibration measurements were corrected for dark background, integration time, lamp power, and system throughput. Then, the experimental data from clinical measurements were divided by the corresponding calibration data for each detection fiber to account for possible changes in fiber throughput and lamp power.

A Monte Carlo (MC) based lookup table approach was used for optical property recovery. This recovery technique was thoroughly described previously[14]. Briefly, in order to recover optical property, corrected measured data were compared with a pre-built 3D MC library. As the measured data are of a different scale compared to that of the simulation library (counts/mW/ms vs. simulated photon weight), the measured data were scaled to the simulation library by dividing the measured data for each fiber by corresponding ratio values for each fiber. These ratio values were obtained by comparing previous phantom experiments to MC simulations at the same optical properties[14]. From the measured, corrected data (see Figure 3) it was apparent that the only absorbers present were hemoglobin and methylene blue. Thus, we employed a simultaneous multi-spectral fitting technique for optical property recovery. Absorption was assumed to be a superposition of known absorption spectral shapes and scattering was assumed to have a decaying power law form. The root-mean-squared deviation between simulated detected spectra and measured spectra was minimized simultaneously for all detectors using constrained, non-linear optimization to determine absorber concentrations and scattering parameters. Oxygen saturation (SO₂) was calculated by dividing the concentration of the oxy-hemoglobin concentration ([HbO₂]) by total hemoglobin concentration (THC = [Hb] + [HbO₂]).

Figure 3.

Representative corrected diffuse reflectance data (solid lines) for detector fibers 1–4 and resulting fits (open circles) prior to MB administration.

3. RESULTS

Representative corrected diffuse reflectance spectra and resulting fits prior to MB administration are shown in Figure 3 for subject two for detector fibers 1–4. Employing our optical property recovery technique on these corrected data, absorption and reduced scattering spectra were recovered, as shown in Figure 4 (pre-MB results corresponding to Figure 3 are displayed in blue). It is evident that prior to MB administration, oxy- (HbO₂) and deoxy-hemoglobin (Hb) are the primary absorbers, with a large contribution from MB after photosensitizer incubation and aspiration. As shown in figure 4b, there is a small increase in reduced scattering after MB administration.

Figure 4.

Recovered (a) absorption and (b) reduced scattering spectra for the subject shown in Figure 3 (Pre-MB and Post-MB).

The aforementioned optical property extraction technique was also used on the corrected diffuse reflectance spectra of the other subjects, with primary findings summarized in Table 1. For all subjects, prior to MB instillation (Pre-MB) cavity wall absorption was dominated by oxy- and deoxy-hemoglobin. The concentrations of HbO₂ and Hb varied widely between subjects. As our treatment light for MB-PDT is delivered at 665 nm, this wavelength is of particular importance. Pre-MB absorption coefficients at 665 nm were less than 1 cm⁻¹ for all subjects. Pre-MB scattering spectra exhibited similar wavelength dependence, although there was inter-patient variability in overall magnitude.

Table 1.

Summary of extracted parameters (Post-MB data for subject 1 was not usable due to poor signal-to-noise ratio).

Subject		1	2	3
Hb (μM)	Pre - MB	47.2	3.43	0.11
	Post - MB	-	24.8	21.4
HbO2 (μM)	Pre - MB	213.8	7.31	132.8
	Post - MB	-	0.32	55.8
SO2 (%)	Pre - MB	81.9	68.0	99.9
	Post - MB	-	1.3	72.3
μa (665 nm)	Pre - MB	0.53	0.08	0.14
	Post -MB	-	14.5	0.33
μs’ (665 nm)	Pre - MB	7.55	10.9	5.53
	Post - MB	-	12.2

In post-MB spectroscopy, we found differences between subjects in MB uptake. MB uptake for subject two was much higher compared to that of subject three. Thus, subject two’s post-MB absorption spectrum was dominated by MB absorption, as shown in the Figure 4a. For both subjects with post-MB spectra, we noticed a reduction in oxygen saturation (SO₂) for post-MB values compared to pre-MB values. For both subject two and three, reduced scattering at 665 nm was similar pre- and post-MB, with a slight increase after MB administration.

4. DISCUSSION

The spatially-resolved diffuse reflectance data collected for these first three subjects show a large degree of variability in the recovered optical property spectra. In particular, concentrations of oxy- and deoxy-hemoglobin ranged from 7.3–213 μM and 0.1–47.2 μM prior to administration of methylene blue, respectively. Reduced scattering spectra showed inter-patient variability, but recovered values were more similar between subjects. Further, methylene blue uptake was found to vary between subjects, and was associated with a reduction in tissue oxygen saturation. These preliminary findings highlight the importance of future patient-specific treatment planning.

This initial report represents the first measurements of the optical properties of human abscesses, making direct comparison to prior literature difficult. However, measured optical properties at locations with the human abdomen have been previously reported. Morales et al performed diffuse reflectance measurements at multiple sites within the pleural cavity before and after photodynamic therapy[17]. The absorption and reduced scattering coefficients at 665 nm determined here fall within the ranges reported for the pre-PDT conditions by these investigators. Pre-MB measured values here are also similar to those reported for the small bowel, large bowel, and peritoneum by Wang et al[18], although these investigators provide absorption and reduced scattering coefficients at 630 nm rather than 665 nm. Finally, we previously determined the optical properties of surgically extracted human kidneys[19]. At 665 nm, μ_a ranged from 0.3–1.7 cm⁻¹ and μ_s’ ranged from 12.5–39.4 cm⁻¹, which are again in a similar range to those shown here. Therefore, while we cannot comment on whether the current results are typical for abscesses, the measured optical properties are in reasonable ranges for nearby healthy tissues in the abdomen.

In the two subjects that had sufficient signal to noise ratio in their post-MB measurements, we observed a large decrease in oxygen saturation (SO₂) following MB administration. While previous reports have shown that MB administration can cause artificially low SO₂ readings when measured via standard pulse oximetry[20–22], these studies only utilized two wavelengths for analysis rather than the full spectra employed here. As we can distinguish between MB and hemoglobin absorption spectra, this reduction in SO₂ appears to be real and proportional to MB uptake. A potential explanation is localized conversion of hemoglobin to methemoglobin, which cannot bind oxygen[23], although we do not see evidence of methemoglobin absorption features[24] in the collected spectra. Crucially, we did not observe this same reduction in SO₂ remote from the abscess, indicating that this is not a systemic effect and is localized to the site of MB infusion. It is difficult to draw strong conclusions, due to the small sample size. We will explore this in greater detail upon completion of the ongoing clinical trial.

In addition to determination of typical optical properties within human abscesses, the overall goal of these measurements is to enable future patient-specific treatment planning for PDT of deep tissue abscesses. We have previously demonstrated that the ability to achieve a target light dose is highly dependent upon optical properties at the abscess wall, particularly absorption[12,13]. As our current phase 1 trial is examining the safety and feasibility of PDT and optical spectroscopy, we cannot currently modify the PDT treatment parameters based on measured optical properties. However, future phase 2 studies will utilize these optical properties to adjust the optical power delivered and Intralipid concentration utilized in order to optimize light delivery. We anticipate that this will increase the portion of the abscess wall that receives an efficacious target light dose, and will subsequently improve response to PDT.

We acknowledge some limitations in the current results. First, these represent preliminary findings from a limited number of research subjects. This clinical trial is ongoing, with additional data collection expected in the near future. Additionally, post-MB spectroscopy measurements for the first subject suffered from poor signal to noise ratio and could not be fully analyzed. This motivated software improvements that eliminated this issue in subsequent subjects, and will not influence future measurements. Finally, repeated spectroscopy measurements in the same subject to assess variability or the effects of probe surface contact were not possible for these subjects due to clinical time constraints. We will endeavor to perform these repeated measures in future subjects.

Figure 1.

Clinical trial protocol illustrating timing of standard of care abscess drainage, optical spectroscopy, and photodynamic therapy