Treatment planning for photodynamic therapy of abscess cavities using patient-specific optical properties measured prior to illumination.

Photodynamic therapy (PDT) is an effective antimicrobial therapy that we used to treat human abscess cavities in a Phase 1 clinical trial. This trial included pre-PDT measurements of abscess optical properties, which affect light dose (light fluence) at the abscess wall and PDT response. This study simulated PDT treatment planning for 13 subjects that received optical spectroscopy prior to clinical PDT, to determine the impact of measured optical properties on ability to achieve fluence rate targets in 95% of the abscess wall. Retrospective treatment plans were evaluated for 3 conditions: (1) clinically delivered laser power and assumed, homogeneous optical properties, (2) clinically delivered laser power and measured, homogeneous optical properties, and (3) with patient-specific treatment planning using measured, homogeneous optical properties. Treatment plans modified delivered laser power, intra-cavity Intralipid (scatterer) concentration, and laser fiber type. Using flat-cleaved laser fibers, the proportion of subjects achieving 95% abscess wall coverage decreased significantly relative to assumed optical properties when using measured values for 4 mW cm-2(92% versus 38%,p= 0.01) and 20 mW cm-2(62% versus 15%,p= 0.04) thresholds. When measured optical properties were incorporated into treatment planning, the 4 mW cm-2target was achieved for all cases. After treatment planning, optimal Intralipid concentration across subjects was 0.14 ± 0.09%, whereas 1% was used clinically. Required laser power to achieve the 4 mW cm-2target was significantly correlated with measured abscess wall absorption (ρ= 0.7,p= 0.008), but not abscess surface area (ρ= 0.2,p= 0.53). When using spherical diffuser fibers for illumination, both optimal Intralipid concentration (p= 0.0005) and required laser power (p= 0.0002) decreased compared to flat cleaved fibers. At 0% Intralipid concentration, the 4 mW cm-2target could only be achieved for 69% of subjects for flat-cleaved fibers, compared to 100% for spherical diffusers. Based on large inter-subject variations in optical properties, individualized treatment planning is essential for abscess photodynamic therapy. (Clinical Trial Registration: The parent clinical trial from which these data were acquired is registered on ClinicalTrials.gov as 'Safety and Feasibility Study of Methylene Blue Photodynamic Therapy to Sterilize Deep Tissue Abscess Cavities,' with ClinicalTrials.gov identifier NCT02240498).

First in human measurements of abscess cavity optical properties and methylene blue uptake prior to photodynamic therapy by in vivo diffuse reflectance spectroscopy.

Significance: Efficacious photodynamic therapy (PDT) of abscess cavities requires personalized treatment planning. This relies on knowledge of abscess wall optical properties, which we report for the first time in human subjects. Aim: The objective was to extract optical properties and photosensitizer concentration from spatially resolved diffuse reflectance measurements of abscess cavities prior to methylene blue (MB) PDT, as part of a phase 1 clinical trial. Approach: Diffuse reflectance spectra were collected at the abscess wall of 13 human subjects using a custom fiber-optic probe and optical spectroscopy system, before and after MB administration. A Monte Carlo lookup table was used to extract optical properties. Results: Pre-MB abscess wall absorption coefficients at 665 nm were 0.15±0.1 cm-1 (0.03 to 0.36 cm-1) and 10.74±15.81 cm-1 (0.08 to 49.3 cm-1) post-MB. Reduced scattering coefficients at 665 nm were 8.45±2.37 cm-1 (4.8 to 13.2 cm-1) and 5.6±2.26 cm-1 (1.6 to 9.9 cm-1) for pre-MB and post-MB, respectively. Oxygen saturations were found to be 58.83%±35.78% (5.6% to 100%) pre-MB and 36.29%±25.1% (0.0001% to 76.4%) post-MB. Determined MB concentrations were 71.83±108.22 µM (0 to 311 µM). Conclusions: We observed substantial inter-subject variation in both native wall optical properties and MB uptake. This underscores the importance of making these measurements for patient-specific treatment planning.

Application of transfer learning for rapid calibration of spatially resolved diffuse reflectance probes for extraction of tissue optical properties.

Significance: Treatment planning for light-based therapies including photodynamic therapy requires tissue optical property knowledge. This is recoverable with spatially resolved diffuse reflectance spectroscopy (DRS) but requires precise source-detector separation (SDS) determination and time-consuming simulations. Aim: An artificial neural network (ANN) to map from DRS at multiple SDS to optical properties was created. This trained ANN was adapted to fiber-optic probes with varying SDS using transfer learning (TL). Approach: An ANN mapping from measurements to Monte Carlo simulation to optical properties was created with one fiber-optic probe. A second probe with different SDS was used for TL algorithm creation. Data from a third were used to test this algorithm. Results: The initial ANN recovered absorber concentration with RMSE=0.29 µM (7.5% mean error) and µs' at 665 nm (µs,665') with RMSE=0.77 cm-1 (2.5% mean error). For probe 2, TL significantly improved absorber concentration (0.38 versus 1.67 µM RMSE, p=0.0005) and µ's,665 (0.71 versus 1.8 cm-1 RMSE, p=0.0005) recovery. A third probe also showed improved absorber (0.7 versus 4.1 µM RMSE, p<0.0001) and µs,665' (1.68 versus 2.08 cm-1 RMSE, p=0.2) recovery. Conclusions: TL-based probe-to-probe calibration can rapidly adapt an ANN created for one probe to similar target probes, enabling accurate optical property recovery with the target probe.

Application of Transfer Learning for Rapid Calibration of Spatially-resolved Diffuse Reflectance Probes for Extraction of Tissue Optical Properties.

Significance: Treatment planning for light-based therapies including photodynamic therapy requires tissue optical property knowledge. These are recoverable with spatially-resolved diffuse reflectance spectroscopy (DRS), but requires precise source-detector separation (SDS) determination and time-consuming simulations. Aim: An artificial neural network (ANN) to map from DRS at short SDS to optical properties was created. This trained ANN was adapted to fiber-optic probes with varying SDS using transfer learning. Approach: An ANN mapping from measurements to Monte Carlo simulation to optical properties was created with one fiber-optic probe. A second probe with different SDS was used for transfer learning algorithm creation. Data from a third were used to test this algorithm. Results: The initial ANN recovered absorber concentration with RMSE=0.29 µM (7.5% mean error) and µ s ' at 665 nm (µ s,665 ' ) with RMSE=0.77 cm -1 (2.5% mean error). For probe-2, transfer learning significantly improved absorber concentration (0.38 vs. 1.67 µM, p=0.0005) and µ s,665 ' (0.71 vs. 1.8 cm -1 , p=0.0005) recovery. A third probe also showed improved absorber (0.7 vs. 4.1 µM, p<0.0001) and µ s,665 ' (1.68 vs. 2.08 cm -1 , p=0.2) recovery. Conclusions: A data-driven approach to optical property extraction can be used to rapidly calibrate new fiber-optic probes with varying SDS, with as few as three calibration spectra.

Treatment planning for photodynamic therapy of abscess cavities using patient-specific optical properties measured prior to illumination.

Background: Photodynamic therapy (PDT) is an effective antimicrobial therapy that we used to treat human abscess cavities in a recently completed Phase 1 clinical trial. This trial included pre-PDT measurements of abscess optical properties, which affect the expected light dose to the abscess wall and eventual PDT response. Purpose: The objective of this study was to simulate PDT treatment planning for the 13 subjects that received optical spectroscopy prior to clinical abscess PDT. Our goal was to determine the impact of these measured optical properties on our ability to achieve fluence rate targets in 95% of the abscess wall. Methods: During a Phase 1 clinical trial, 13 subjects received diffuse reflectance spectroscopy prior to PDT in order to determine the optical properties of their abscess wall. Retrospective treatment plans seeking to achieve fluence rate targets in 95% of the abscess wall were evaluated for all subjects for 3 conditions: (1) at the laser power delivered clinically with assumed optical properties, (2) at the laser power delivered clinically with measured optical properties, and (3) with patient-specific treatment planning using these measured optical properties. Factors modified in treatment planning included delivered laser power and intra-cavity Intralipid (scatterer) concentration. The effects of laser fiber type were also simulated. Results: Using a flat-cleaved laser fiber, the proportion of subjects that achieved 95% abscess wall coverage decreased significantly when incorporating measured optical properties for both the 4 mW/cm 2 (92% vs. 38%, p=0.01) and 20 mW/cm 2 (62% vs. 15%, p=0.04) fluence rate thresholds. However, when measured optical properties were incorporated into treatment planning, a fluence rate of 4 mW/cm 2 was achieved in 95% of the abscess wall for all cases. In treatment planning, the optimal Intralipid concentration across subjects was found to be 0.14 ± 0.09% and the optimal laser power varied from that delivered clinically but with no clear trend (p=0.79). The required laser power to achieve 4 mW/cm 2 in 95% of the abscess wall was significantly correlated with measured µ a at the abscess wall (ρ=0.7, p=0.008), but not abscess surface area (ρ=0.2, p=0.53). When using spherical diffuser fibers as the illumination source, the optimal intralipid concentration decreased to 0.028 ± 0.026% (p=0.0005), and the required laser power decreased also (p=0.0002), compared to flat cleaved fibers. If the intra-cavity lipid emulsion (Intralipid) was replaced with a non-scattering fluid, all subjects could achieve the 4 mW/cm 2 fluence rate threshold in 95% of the abscess wall using a spherical diffuser, while only 69% of subjects could reach the same criterion using a flat cleaved fiber. Conclusions: The range of optical properties measured in human abscesses reduced coverage of the abscess wall at desirable fluence rates. Patient-specific treatment planning including these measured optical properties could bring the coverage back to desirable levels by altering the Intralipid concentration and delivered optical power. These results motivate a future Phase 2 clinical trial to directly compare the efficacy of patient-specific-treatment planning with fixed doses of Intralipid and light.Clinical Trial Registration: The parent clinical trial from which these data were acquired is registered on ClinicalTrials.gov as "Safety and Feasibility Study of Methylene Blue Photodynamic Therapy to Sterilize Deep Tissue Abscess Cavities," with ClinicalTrials.gov identifier NCT02240498 .

First in human measurements of abscess cavity optical properties and methylene blue uptake prior to photodynamic therapy by in vivo diffuse reflectance spectroscopy.

Significance: Efficacious photodynamic therapy (PDT) of abscess cavities requires personalized treatment planning. This relies on knowledge of abscess wall optical properties, which we report for the first time in human subjects. Aim: The objective was to extract optical properties and photosensitizer concentration from spatially-resolved diffuse reflectance measurements of abscess cavities prior to methylene blue (MB) PDT, as part of a Phase 1 clinical trial. Approach: Diffuse reflectance spectra were collected at the abscess wall of 13 human subjects using a custom fiber-optic probe and optical spectroscopy system, before and after MB administration. A Monte Carlo lookup table was used to extract optical properties. Results: Pre-MB abscess wall absorption coefficients at 665 nm were 0.15±0.1 cm -1 (0.03-0.36 cm -1 ) and 10.74±15.81 cm -1 (0.08-49.3 cm -1 ) post-MB. Reduced scattering coefficients at 665 nm were 8.45±2.37 cm -1 (4.8-13.2 cm -1 ) and 5.6±2.26 cm -1 (1.6-9.9 cm -1 ) for pre-MB and post-MB, respectively. Oxygen saturations were found to be 58.83±35.78% (5.6-100%) pre-MB and 36.29±25.1% (0.0001-76.4%) post-MB. Determined MB concentrations were 71.83±108.22 µM (0-311 µM). Conclusions: We observed substantial inter-subject variation in both native wall optical properties and methylene blue uptake. This underscores the importance of making these measurements for patient-specific treatment planning.

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

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.