PDT dose dosimetry for pleural photodynamic therapy

Abstract

PDT dose is the product of the photosensitizer concentration and the light fluence in target tissue. Although existing systems are capable of measuring the light fluence in vivo, the concurrent measurement of photosensitizer in the treated tissue so far has been lacking. We have developed and tested a new method to simultaneously acquire light dosimetry and photosensitizer fluorescence data via the same isotropic detector, employing treatment light as the excitation source. A dichroic beamsplitter is used to split light from the isotropic detector into two fibers, one for light dosimetry, the other, after the 665 nm treatment light is removed by a band-stop filter, to a spectrometer for fluorescence detection. The light fluence varies significantly during treatment because of the source movement. The fluorescence signal is normalized by the light fluence measured at treatment wavelength. We have shown that the absolute photosensitizer concentration can be obtained by an optical properties correction factor and linear spectral fitting. Tissue optical properties are determined using an absorption spectroscopy probe immediately before PDT at the same sites. This novel method allows accurate real-time determination of delivered PDT dose using existing isotropic detectors, and may lead to a considerable improvement of PDT treatment quality compared to the currently employed systems. Preliminary data in patient studies is presented.

1. INTRODUCTION

PDT treatment outcome depends on both the light dose and photosensitizer distribution¹. PDT dose is defined as a product of the photosensitizer concentration and the light fluence in target tissue. Although the existing systems are capable of measuring the light fluence in vivo, the concurrent measurement of photosensitizer in the treated tissue so far has been lacking.

This study demonstrates the results of ongoing patient study, where the PDT dose was measured using treatment light as a source. The same fiber detector that is employed for light dosimetry was also monitoring HPPH fluorescence in vivo, and photosensitizer concentration was determined by applying optical properties correction to the fluorescence measurements.

2. METHODS

2.1 Patient treatment

The patients in this study were enrolled in a clinical trial of HPPH-mediated PDT for the lung cancer treatment. They were administered 4 mg of HPPH per kg of body weight 48 hours before surgery, and treated with light therapy at 665 nm with a fluence of 25–30 J/cm². Light was delivered via a fiber optic embedded in a modified endotracheal tube filled with dilute intralipid. A dilute intralipid solution, used as a scattering medium, was placed inside the patients’ pleural cavity. The fluence rate and cumulative fluence were continuously monitored at seven sites during the treatment using isotropic detectors sewn to the pleural cavity wall², while the eighth isotropic detector was used to measure the light dose and HPPH fluorescence simultaneously. The treatment continued until the prescribed dose was reached at the monitored sites.

2.2 HPPH fluorescence detection during treatment

Treatment light from a diode 665 nm laser (B&W Tek, Newark, DE), source was sensed by an isotropic detector located inside the patient’s pleural cavity. A dichroic beamsplitter divided the light between two fibers, one used for light dosimetry, the other for fluorescence measurement. A band-stop filter centered at 658 nm (658 StopLine, Semrock, Rochester, NY) was used to remove most of the treatment light before fluorescence was recorded by a single-channel spectrometer (BTC-112E, B&W Tek, Newark, DE), (Fig. 1).

Figure 1.

Schematic diagram (a) and photograph (b) of the simultaneous light dose and photosensitizer fluorescence measurement setup.

2.3 Optical properties (OP) correction

The amount of HPPH fluorescence registered by a detector depends on the optical properties of the surrounding tissues. To determine the correction factor for HPPH concentration, absorption and scattering coefficients were measured for several intralipid, ink, and HPPH phantoms³. We have used the empirical OP correction function in the form^4–5:

$$\text{CF}=({\mathrm{C}}_{01}+{\mathrm{C}}_{02}{\mathrm{\mu }}_{{\mathrm{s}}^{\prime }})/{\mathrm{\mu }}_{{\mathrm{s}}^{\prime }}\ast exp[({\mathrm{b}}_1+{\mathrm{b}}_2{\mathrm{\mu }}_{{\mathrm{s}}^{\prime }}){\mathrm{\mu }}_{\text{eff}}]$$ (1)

where μ_eff = (3 μ_a μ_s′)^1/2, for semi-infinite geometry with detector located in water on top of a semi-infinite tissue or phantom medium.

Phantom optical properties are summarized in Table 1. Table 2 shows OP correction function parameters extracted from the phantom data fits. Fig. 2a demonstrates the results of applying correction function to the phantoms with constant HPPH concentration, while Fig. 2b shows the linear relationship between HPPH concentration and fluorescence in the variable HPPH phantom after the OP correction.

Table 1.

Phantom optical properties summary for the treatment wavelength (λ= 665 nm).

Optical properties	Intralipid	Ink	HPPH
Absorption coefficient μa, 1/cm	0.007±0.021	36.28*C(%)	0.061*C(mg/kg)
Scattering coefficient μs′, 1/cm	11.26*C(%)	0	0

Table 2.

Optical properties correction function parameters.

C01	C02	b1	b2
−0.7645	0.1919	0.5662	−0.0213

Figure 2.

(a) HPPH 0.5 mg/kg phantom: SVD with (dashed) and without (solid) optical properties correction; (b) HPPH fluorescence SVD signal as a function of HPPH concentration for a phantom with variable HPPH.

The OP correction factor was applied to the patient fluorescence data, based on pre-PDT OP measurements. To make sure that fluorescence signal was fully separated from the excitation light, a singular-value decomposition (SVD) was applied to the raw fluorescence data after background subtraction and normalization to the excitation signal.

3. RESULTS AND DISCUSSION

3.1 Patient data and raw fluorescence signal during treatment

So far, we have collected and analyzed data for four patients. Table 3 is the summary of patient data. It should be noted that the initial fluorescence value varies between patients, in part due to the difference between detector locations. The total treatment time depends on the size of the patient’s pleural cavity. The optical properties were measured before the PDT treatment.

Table 3.

A summary of patient data.

Patient Index	Treatment time, min	Total fluence, J/cm2	Initial raw fluorescence, arb. units	Absorption coeff, 1/cm	Scattering coeff, 1/cm
1	48	25	4.0	N/A	N/A
2	45	25	2.2	N/A	N/A
3	54	30	6.5	0.025	1.61
4	46	30	2.5	0.31	11.5

Fig. 3 shows typical raw fluorescence collected during treatment. The signal intensity at the detector position varied randomly throughout the entire treatment time, depending on the source location. This variation was the reason for further processing of the data before photosensitizer concentration could be obtained.

Figure 3.

(a) Typical raw HPPH fluorescence spectra collected during PDT treatment; (b) Fluorescence peak at 720 nm as a function of cumulative fluence.

3.2 Singular-value decomposition: basis vectors and components

Raw fluorescence data was background-subtracted and normalized to the signal around 640 nm, representing excitation light outside of the band-stop filter range. We have also applied thresholding to the data to skip time intervals with near-zero light conditions. Because there was a significant overlap between treatment light and HPPH fluorescence, a secondary fluorescence peak around 720 nm was chosen for data processing, and SVD was used to make sure that fluorescence signal is not contaminated by residual excitation light. Fig. 4 shows SVD basis vectors and decomposition results. HPPH fluorescence components for all four patients are shown in Fig. 5, both in terms of treatment time and cumulative fluence. Based on the smoothed results, we conclude that fluorescence did not change significantly during any of the four cases, decreasing at most by 15% by the end of treatment.

Figure 4.

(a) SVD basis vectors with band-stop filter: treatment laser light (blue) and HPPH fluorescence (green); (b) Results of SVD for typical patient data: treatment laser component (solid blue), HPPH fluorescence (solid green), background/Fourier component (solid red), and smoothed HPPH (dashed).

Figure 5.

Patient HPPH SVD components (plus smoothing) as a function of treatment time (a) and cumulative fluence (b); patients #1 (X), #2 (+), #3 (*), and #4 (Δ).

3.3 HPPH concentration during treatment

To obtain HPPH concentration during treatment, the OP correction function for HPPH, used with the patient optical properties data, was applied to the patient HPPH fluorescence SVD⁶. Fig. 6 shows HPPH concentration for patients #3 and #4. In both cases, photosensitizer concentration tends to decrease slightly (10 – 15%) over the course of the treatment. The absolute HPPH concentration, however, is different by a factor of seven between these patients.

Figure 6.

Patient HPPH concentration as a function of treatment time (a) and cumulative fluence (b) for patients #3 and #4.

4. CONCLUSIONS

We have demonstrated that photosensitizer fluorescence can be measured in clinical situations using existing isotropic detectors for light dosimetry. Fluorescence signal varied considerably during treatment, requiring extensive data processing. The HPPH concentration could be determined from fluorescence data using optical properties correction function.

In our clinical study, the amount of HPPH in the tissues did not change significantly during treatment time. However, we have observed HPPH concentration variation of a factor of seven among different patients.

Future plans include fluorescence collection at several detector locations per patient, real time processing of fluorescence data, and optimization of the OP correction function.