Purpose: In superficial basal cell carcinomas treated with photodynamic therapy with topical δ-aminolevulinic acid, we examined effects of light irradiance on photodynamic efficiency and pain. The rate of singlet-oxygen production depends on the product of irradiance and photosensitizer and oxygen concentrations. High irradiance and/or photosensitizer levels cause inefficient treatment from oxygen depletion in preclinical models.

Experimental Design: Self-sensitized photobleaching of protoporphyrin IX (PpIX) fluorescence was used as a surrogate metric for photodynamic dose. We developed instrumentation measuring fluorescence and reflectance from lesions and margins during treatment at 633 nm with various irradiances. When PpIX was 90% bleached, irradiance was increased to 150 mW/cm2 until 200 J/cm2 were delivered. Pain was monitored.

Results: In 33 superficial basal cell carcinomas in 26 patients, photobleaching efficiency decreased with increasing irradiance above 20 mW/cm2, consistent with oxygen depletion. Fluences bleaching PpIX fluorescence 80% (D80) were 5.7 ± 1.6, 4.5 ± 0.3, 7.5 ± 0.8, 7.4 ± 0.3, 12.4 ± 0.3, and 28.7 ± 7.1 J/cm2, respectively, at 10, 20, 40, 50, 60 and 150 mW/cm2. At 20-150 mW/cm2, D80 doses required 2.5-3.5 min; times for the total 200 J/cm2 were 22.2-25.3 min. No significant pain occurred up to 50 mW/cm2; pain was not significant when irradiance then increased. Clinical responses were comparable to continuous 150 mW/cm2 treatment.

Conclusions: Photodynamic therapy with topical δ-aminolevulinic acid using ∼40 mW/cm2 at 633 nm is photodynamically efficient with minimum pain. Once PpIX is largely photobleached, higher irradiances allow efficient, rapid delivery of additional light. Optimal fluence at a single low irradiance is yet to be determined.

Photodynamic therapy (PDT) is a selective, three-part therapy using a combination of photosensitizer, light, and oxygen to generate singlet oxygen in tissues that leads to cell death, vascular shutdown, and host inflammatory and immune responses (1, 2). The singlet oxygen is produced from ground-state oxygen. The rate of tissue photodynamic oxygen consumption is proportional to the product of photosensitizer concentration and light irradiance. PDT is more efficient when administered at low irradiances in preclinical models (38). High irradiances and/or photosensitizer concentrations can cause depletion of available tissue oxygen, leading to inefficient and ineffective treatment (37, 9).

Topically administered δ-aminolevulinic acid (ALA)-PDT is a therapy in which the prodrug ALA is converted into the photosensitizer protoporphyrin IX (PpIX) in the skin via the heme cycle (10). Similarly, methyl-ALA-PDT employs a methyl ester that is converted into ALA by intracellular esterases. ALA-PDT increasingly is used to treat a variety of conditions, including nonmelanoma skin cancers such as superficial basal cell carcinoma (sBCC). It has advantages of tumor selectivity (11), favorable cosmetic outcomes compared with surgery and cryotherapy (12), and easy delivery (13). A drawback to topical ALA-PDT is the stinging, burning pain that can accompany treatment.

In many clinical situations using PDT, the light irradiance is not controlled and may vary significantly depending, for example, on the distance between the light source and the patient (1416). In addition, topical PDT can produce high local concentrations of PpIX, making irradiance-dependent photodynamic oxygen depletion a particular concern. It is difficult to define optimum irradiances in a clinical trial, because the high complete response rates necessitate large treatment groups for each treatment condition. In addition, the optimum light doses are uncertain, and the time to deliver a light dose can be unacceptably large at low irradiances.

Self-sensitized photobleaching of PpIX is a surrogate metric for PDT efficiency that allows monitoring effects of irradiance. Because the PpIX can be destroyed by singlet oxygen and other reactive species generated by PDT, the rate of disappearance of PpIX with delivered light is an indirect measure of the efficiency of the treatment: greater efficiency corresponds to more rapid photobleaching. PpIX levels within the skin can be monitored by fluorescence, particularly when using red light excitation to maximize the depth of the interrogated tissue volume. We devised a means to continuously measure changes in PpIX fluorescence excited by the 633 nm treatment light without interrupting treatment and report the first determination of photobleaching kinetics in patients as a function of irradiance. The light dose causing disappearance of a fixed fraction of the PpIX fluorescence provides a measure of the efficiency of the treatment at different irradiances, and we determined the irradiance dependence of the fluence required to reduce PpIX fluorescence by 80% (D80).

We also postulated that pain could serve as a measure of normal tissue injury and that decreased pain with increased therapeutic selectivity could be achieved at lower irradiances (17, 18). This trial also examined the dependence of treatment-induced pain on irradiance and established a threshold irradiance, below which patients did not experience significant pain or require anesthetics or other interventions. Taken together, the results presented here motivate further clinical evaluation of the use of low irradiance in ALA-PDT and provide preliminary guidance for the choice of irradiance and fluence in trials designed to optimize pain-free, efficacious, and time-efficient protocols.

Patients. Selection criteria included sBCC measuring at least 5 mm but not >20 mm in diameter. No patient contributed more than two lesions to the study at each irradiance. No effort was made to select patients for skin tone, lesion pigmentation, lesion location, or other criteria. After a baseline study at 150 mW/cm2, groups of five lesions were examined with escalating irradiances starting at 10 mW/cm2. A stopping rule was employed such that if 2 patients had visual analogue scale (VAS) pain scores of ≥4 at a given irradiance then therapy at a higher initial irradiance would not be done. Instead, an irradiance midway between that irradiance and the previous irradiance would be evaluated. In accordance with the U.S. Food and Drug Administration guidelines, patients were informed of the investigational nature of this study and signed a consent form. The Office of Research Subject Protection and Scientific Integrity at Roswell Park Cancer Institute approved this study. sBCC was confirmed for these lesions by histologic biopsy 2 weeks before treatment, allowing the site to heal before treatment, or immediately following treatment.

ALA administration. ALA (DUSA Pharmaceuticals) was freshly prepared at a 20% (w/w) concentration in an oil-in-water cream (Moisturel; Westwood Pharmaceuticals). The ALA was covered by an occlusive bandage and opaque material to prevent photoactivation before therapy. Following application, the ALA remained on the skin for 4 h ± 15%. The area was cleaned with normal saline before irradiation to remove any residual cream. After the 4-h incubation period, the presence of PpIX was confirmed visually with fluorescence excited from a 405 nm LED flashlight. Visual inspection revealed lesion-to-lesion variation in initial PpIX fluorescence, which was confirmed with fluorescence spectroscopy measurements made during the therapy.

Irradiation. Six irradiances were used in this study: 150, 10, 20, 40, 50, and 60 mW/cm2. At 150 mW/cm2, 200 J/cm2 was delivered as is the standard at our institution for BCC. At the lower irradiances, PDT treatment was delivered in two parts: the initial therapy was delivered at low irradiance until the fluorescence contribution from PpIX in the lesion bleached by 90% of its initial value, at which point the treatment was continued at 150 mW/cm2 until 200 J/cm2 was delivered. This protocol insured that all patients received the same total light dose and allowed treatment to be completed in a reasonable timeframe.

Pain assessment. Pain experienced as a result of the therapy was determined with a VAS score, in which a score of 0 represents no pain, 10 represents unbearable pain, and 4 represents moderate pain, requiring anesthetic, cooling fan, or other measure; VAS ≤ 3 did not require interventions. Each patient was questioned about pain every 5 min and encouraged to notify the staff about changes in pain between these intervals. A pain score of ≥4 was treated by injections with 1% xylocaine. After the irradiance was increased to 150 mW/cm2, pain was assessed immediately and then every 5 min as before. Pain was not assessed, however, during the continuous 150 mW/cm2 treatment, as it has been established previously that treatment at this irradiance generally required pain intervention and had VAS score > 5. Pain was mediated in this group by conscious sedation or 1% xylocaine injection before therapy, repeated as necessary, and by fan cooling.

Clinical follow-up. Lesions were evaluated clinically at 3 and 6 months and then about every 6 months. Complete clinical responses were determined at ≥6 months.

Instrumentation. PpIX fluorescence photobleaching was measured using instrumentation and data reduction described previously (19) and described briefly herein. Fluorescence and reflectance spectra were taken repeatedly throughout PDT at two, 4-mm-diameter, spatially resolved regions. One region was located within the sBCC lesion and the other within the perilesion margin. The system used two light sources: the treatment source was an argon-ion-pumped dye laser (Coherent), band-pass filtered to operate at 632.8 nm, and the broadband reflectance source consisted of a SMA-coupled tungsten-halogen source (Avantes). Both the treatment and broadband beams were delivered through a GRIN lens-terminated fiber mounted in an off-surface probe 80 mm off the skin's surface (Fig. 1). Fluorescence spectra were collected through a microlens-terminated detection fiber, long-pass filtered at 647.1 nm (Semrock), and directed to a spectrometer with a cooled diode array detector that measured fluorescence between 665 and 770 nm with 2 nm resolution (BWTek). Reflectance spectra, taken in the same delivery-detection geometry, were directed to a second spectrometer, measuring 475 to 800 nm with 2 nm resolution. Fluorescence spectra were corrected for dark counts and also corrected for the instrument spectral response using a NIST-traceable calibration lamp (Ocean Optics).

Fig. 1.

Treatment and measurement cycles during system operation include (A) PDT delivery at 632.8 nm with fluorescence monitoring in the lesion, (B) PDT delivery with fluorescence monitoring in the perilesion margin, (C) broadband interrogation with reflectance monitoring in the lesion, and (D) broadband interrogation with reflectance monitoring in the margin. E, normalized fluorescence emission curves for PpIX and photoproduct I illustrating the relative location of the 632.8 nm treatment beam to PpIX and photoproduct I fluorescence and the detection window in which fluorescence was monitored with this instrument.

Fig. 1.

Treatment and measurement cycles during system operation include (A) PDT delivery at 632.8 nm with fluorescence monitoring in the lesion, (B) PDT delivery with fluorescence monitoring in the perilesion margin, (C) broadband interrogation with reflectance monitoring in the lesion, and (D) broadband interrogation with reflectance monitoring in the margin. E, normalized fluorescence emission curves for PpIX and photoproduct I illustrating the relative location of the 632.8 nm treatment beam to PpIX and photoproduct I fluorescence and the detection window in which fluorescence was monitored with this instrument.

Close modal

A laptop computer (R51; IBM) running a LabVIEW (National Instruments) program controlled system operation. The LabVIEW interface allowed control of patient data and variables governing the PDT time course, treatment variables such as fluence and fluence rate, and the data acquisition schedule. Once the treatment was initiated, the instrument cycled through several steps as illustrated in Fig. 1A to D.

In the first cycle, the treatment beam was delivered to the entire 25 mm treatment region, including lesion and perilesion margin, and a fluorescence spectrum was collected from the lesion. The excitation wavelength relative to the fluorescence emission of PpIX and the detection window of the instrumentation is shown in Fig. 1E. In the next cycle, the treatment beam remained on and fluorescence was collected from the margin. For the next two cycles, broadband light was delivered to the same treatment region through the treatment fiber, and reflectance spectra were captured from the lesion and perilesion margin, sequentially, by the same detection fibers as used for fluorescence. Fluorescence data were acquired at intervals of between 0.3 and 3 J/cm2, depending on which irradiance was used, and reflectance data were acquired at intervals between 0.5 and 18 J/cm2. Spectra were saved to files and analyzed in real-time using a Matlab (The MathWorks) routine called by LabVIEW. After PpIX fluorescence was decreased by 90%, the instrument automatically switched to 150 mW/cm2 irradiance and delivered the remainder of the total 200 J/cm2 fluence, which was calculated by the LabVIEW program.

Data analysis. To obtain intrinsic tissue fluorescence, spectra were corrected for tissue absorption and scattering using an empirical correction technique explored by Wu et al. (20) in which the measured spectrum was divided by the broadband reflectance spectrum measured in the same geometry and in the same spectral window. In practice, we found that reflectance spectra between 665 and 800 nm were relatively flat and stable during the therapy. This correction was therefore omitted during the real-time fluorescence analysis. Real-time corrections were instead made with representative reflectance spectra averaged from several preclinical experiments. The corrected fluorescence spectra were analyzed using a singular value decomposition linear fitting algorithm in a LabVIEW-called Matlab routine, which incorporated NIST-traceable, calibrated fluorescence basis spectra for PpIX, the primary fluorescent photoproduct of PpIX, photoproduct I (photoprotoporphyrin), and tissue and fiber autofluorescence. A Fourier series was also included in the spectral basis library to identify unaccounted-for contributions in the measured spectra and any possible autofluorescence changes induced by the therapy (8). In the subsequent post-treatment analysis of the data, the fluence at which 80% of the initial PpIX had bleached (D80) was used to determine bleaching efficiency. This point was chosen because bleaching in the perilesion margin field was found experimentally to proceed more slowly than in the monitored lesion field, and D80 therefore represented a comparable bleaching metric available in each data set. D80 was determined by linear interpolation of the calculated PpIX fluorescence curve for each lesion and margin. Photobleaching and photoproduct data were normalized by the PpIX fluorescence at the start of treatment.

Statistics. Fisher's exact test was used to compare the clinical responses from the 10 to 60 mW/cm2 irradiances to results of the prior trial at 150 mW/cm2. Statistical comparisons of initial PpIX photobleaching rates were done using single-factor ANOVA. D80 fluences for lesion versus perilesion were compared using a two-tailed t test.

Patients and lesions. Twenty-six patients, 11 male and 15 female, with mean age 52.2 years (range 17-86), contributed 33 sBCC lesions located on the face, arms, back, and chest. Seven patients had nevoid basal cell carcinoma syndrome. Data were collected on 7 carcinomas at 150 mW/cm2, 5 each at 10, 20, 40, and 60 mW/cm2, and 6 at 50 mW/cm2.

PpIX photobleaching. Using treatment beam excitation, sequential fluorescence measurements were taken from both sBCC and clinically normal perilesional margin. Singular value decomposition analysis done on each measured spectrum during the course of a treatment reports the relative amplitudes of the fluorophores represented in the fit. A representative fluorescence spectrum obtained from an ALA-sensitized sBCC lesion treated at 50 mW/cm2 and analyzed with singular value decomposition is shown in Fig. 2A. Here, the excellent fit to the corrected fluorescence spectrum is dominated by a weighted sum of contributions from PpIX, photoproduct I, and autofluorescence with only a negligible contribution from the Fourier terms. Thus, there are no significant distortions introduced by either the tissue optical properties or the correction algorithms, and quantitative analysis of the PpIX bleaching and photoproduct accumulation and subsequent bleaching is therefore possible. PpIX and photoproduct I (photoprotoporphyrin) contributions from the same lesion are shown in Fig. 2B. As PpIX monotonically bleached away with increased fluence, the contribution to fluorescence from photoproduct I was seen to increase, peaking at ∼3 J/cm2, after which it underwent photobleaching, although with a bleaching rate that was slower than that of PpIX. PpIX and photoproduct amplitudes for a lesion treated at 150 mW/cm2 are shown in Fig. 2C. Compared with 50 mW/cm2, it is evident that both PpIX and photoprotoporphyrin bleaching was slower at the higher irradiance.

Fig. 2.

A, representative singular value decomposition fit-corrected fluorescence spectrum taken from sBCC lesion being treated with ALA-PDT at 50 mW/cm2 with 632.8 nm light. The linear decomposition shows the fluorescence contribution from PpIX, autofluorescence, photoproduct I, and Fourier terms to the corrected fluorescence signal. Several data points have been removed for clarity. B, PpIX and photoproduct I contributions to the corrected fluorescence from the lesion measurement field throughout ALA-PDT in the same lesion normalized to the initial PpIX contribution. For this sBCC lesion, initial rapid PpIX bleaching was accompanied by a corresponding increase in photoproduct I fluorescence, which turns over and was also bleached during the course of therapy. C, PpIX and photoproduct contributions to fluorescence from a lesion treated at 150 mW/cm2 for 200 J/cm2. In B and C, error bars (crosses) are smaller than plot symbols.

Fig. 2.

A, representative singular value decomposition fit-corrected fluorescence spectrum taken from sBCC lesion being treated with ALA-PDT at 50 mW/cm2 with 632.8 nm light. The linear decomposition shows the fluorescence contribution from PpIX, autofluorescence, photoproduct I, and Fourier terms to the corrected fluorescence signal. Several data points have been removed for clarity. B, PpIX and photoproduct I contributions to the corrected fluorescence from the lesion measurement field throughout ALA-PDT in the same lesion normalized to the initial PpIX contribution. For this sBCC lesion, initial rapid PpIX bleaching was accompanied by a corresponding increase in photoproduct I fluorescence, which turns over and was also bleached during the course of therapy. C, PpIX and photoproduct contributions to fluorescence from a lesion treated at 150 mW/cm2 for 200 J/cm2. In B and C, error bars (crosses) are smaller than plot symbols.

Close modal

The initial PpIX fluorescence varied as much as 8-fold between lesions, but there was no correlation between initial fluorescence and the subsequent photobleaching curves (data not shown). Accordingly, the fluorescence data for each lesion was normalized to the initial amplitude in that lesion. For each of the irradiance groups, average ± SE bleaching curves paired by lesion and perilesion margin are shown in Fig. 3A; they show similar trends for all low-irradiance groups but diverge at 150 mW/cm2, where the normal skin bleached more slowly than the lesion. At the high irradiance, the normal skin PpIX did not fall below 20% of its initial value, consistent with oxygen depletion limiting singlet-oxygen production. However, for each irradiance, the initial bleaching efficiencies (represented by the slope of the first two points in the bleaching data) were comparable for lesion and margin (P > 0.17), suggesting that differential oxygen depletion did not begin immediately.

Fig. 3.

A, comparison of the average ± SE normalized PpIX contribution to the corrected fluorescence as a function of total fluence for lesion and margin fields treated with 10, 20, 40, 50, 60, and 150 mW/cm2 irradiances. PpIX contribution shows fluorescence photobleaching of PpIX during irradiation and suggests comparable bleaching trends between margin and lesion in the lower irradiance groups as well as a clear separation between average bleaching curves at 150 mW/cm2. B, average ± SE normalized PpIX contribution to the corrected fluorescence as a function of delivered fluence from the lesion and margin fields during ALA-PDT of 33 sBCC lesions at 10, 20, 40, 50, 60, and 150 mW/cm2. C, D80 values for individual sBCC (Δ) and their normal skin margins (○) at each irradiance. Horizontal lines, mean. †, at 150 mW/cm2, three margins had not reached 80% photobleaching at the maximum light dose of 200 J/cm2 and are not included in the mean.

Fig. 3.

A, comparison of the average ± SE normalized PpIX contribution to the corrected fluorescence as a function of total fluence for lesion and margin fields treated with 10, 20, 40, 50, 60, and 150 mW/cm2 irradiances. PpIX contribution shows fluorescence photobleaching of PpIX during irradiation and suggests comparable bleaching trends between margin and lesion in the lower irradiance groups as well as a clear separation between average bleaching curves at 150 mW/cm2. B, average ± SE normalized PpIX contribution to the corrected fluorescence as a function of delivered fluence from the lesion and margin fields during ALA-PDT of 33 sBCC lesions at 10, 20, 40, 50, 60, and 150 mW/cm2. C, D80 values for individual sBCC (Δ) and their normal skin margins (○) at each irradiance. Horizontal lines, mean. †, at 150 mW/cm2, three margins had not reached 80% photobleaching at the maximum light dose of 200 J/cm2 and are not included in the mean.

Close modal

The variation in normalized lesional and perilesional PpIX fluorescence at the different irradiances is summarized in Fig. 3B. As seen in this figure, photobleaching typically occurred more efficiently at lower irradiance, consistent with the work of other investigators (5, 6, 8, 21). The bleaching kinetics were not single exponential, and the slopes changed most substantially at higher irradiances. Photosensitizer bleaching was rapid, with the lesions in the slowest group (150 mW/cm2) bleaching by 80% in 28.7 ± 7.1 J/cm2, representing <20% of the 200 J/cm2 prescribed fluence (Table 1). Single-factor ANOVA showed that initial bleaching rates in lesions and margins for treatments conducted at 10 and 20 mW/cm2 were statistically similar (P > 0.7). PDT done at irradiances of 40, 50, and 60 mW/cm2 also produced bleaching rates that were statistically indistinguishable (P > 0.38). When data from these two statistically similar treatment groups were pooled, statistical significance emerged between photobleaching rates observed at 10 and 20 mW/cm2 versus those lesions and margins treated at irradiances of 40 to 60 mW/cm2 (P < 0.01). Similarly, the bleaching rates for the lesions and margins subjected to PDT at 40 to 60 mW/cm2 were significantly greater than that measured for PDT at 150 mW/cm2 (P < 0.01).

Table 1.

D80 fluence and T80 time required for each irradiance group to bleach PpIX by 80% and delivery time required to treat at low irradiance to D80 and then deliver the remainder of 200 J/cm2 at 150 mW/cm2

Fluence rate (mW/cm2)Lesion D80 ± SE (J/cm2)Lesion T80 ± SE (min)Total (low + high irradiance) treatment time (min)Margin D80 ± SE (J/cm2)Margin T80 ± SE (min)
10 5.7 ± 1.6 9.5 ± 2.6 31.2 8.7 ± 1.6 14.5 ± 2.6 
20 4.5 ± 0.3 3.7 ± 0.2 25.4 4.4 ± 0.8 3.7 ± 0.7 
40 7.5 ± 0.8 3.1 ± 0.3 24.5 6.3 ± 0.7 2.6 ± 0.3 
50 7.4 ± 0.2 2.5 ± 0.1 23.9 11.8 ± 2.5 3.9 ± 0.8 
60 12.4 ± 3.0 3.4 ± 0.8 24.3 14.3 ± 3.4 4.0 ± 0.9 
150 28.7 ± 7.1 3.2 ± 0.8 22.2 59.5 ± 10 6.6 ± 1.1 
Fluence rate (mW/cm2)Lesion D80 ± SE (J/cm2)Lesion T80 ± SE (min)Total (low + high irradiance) treatment time (min)Margin D80 ± SE (J/cm2)Margin T80 ± SE (min)
10 5.7 ± 1.6 9.5 ± 2.6 31.2 8.7 ± 1.6 14.5 ± 2.6 
20 4.5 ± 0.3 3.7 ± 0.2 25.4 4.4 ± 0.8 3.7 ± 0.7 
40 7.5 ± 0.8 3.1 ± 0.3 24.5 6.3 ± 0.7 2.6 ± 0.3 
50 7.4 ± 0.2 2.5 ± 0.1 23.9 11.8 ± 2.5 3.9 ± 0.8 
60 12.4 ± 3.0 3.4 ± 0.8 24.3 14.3 ± 3.4 4.0 ± 0.9 
150 28.7 ± 7.1 3.2 ± 0.8 22.2 59.5 ± 10 6.6 ± 1.1 

Individual PpIX bleaching curves were used to determine the D80 for each carcinoma and margin using linear interpolation of the data points. For the sBCC, there were significant interlesional variations in bleaching rates, particularly at very low and high irradiances. As shown in Fig. 3C and summarized in Table 1, the mean D80 values were relatively constant between 10 and 20 mW/cm2 and then increased slowly with irradiance for irradiances greater than 20 mW/cm2. The D80 fluences between lesion and margin were similar for all the low-irradiance groups. A two-tailed t test showed that D80 values between lesion and margin were significantly different only at 150 mW/cm2 (P < 0.03), where three of the seven margins failed to reach 80% bleaching by 200 J/cm2 (Fig. 3).

Bleaching efficiency in terms of the time required to deliver the D80 fluence was also considered: the time required to bleach PpIX by 80% (T80) was calculated by dividing D80 by the irradiance (Table 1). The lesion T80 values were remarkably consistent for 20, 40, 50, 60, and 150 mW/cm2 ranging from 2.5 ± 0.1 min at 50 mW/cm2 to 3.7 ± 0.2 min at 20 mW/cm2. The 10 mW/cm2 irradiance exhibited a significantly larger T80 of 9.5 ± 2.6 min. The time required for our two-irradiance protocol, in which D90 is delivered at low irradiance and then the remainder of the fluence is delivered at 150 mW/cm2, also is presented in Table 1. Under this protocol, only modest increases to treatment time are required for initial irradiances ≥20 mW/cm2.

Clinical outcomes. Twenty-four of the 26 lesions receiving low irradiances were available for evaluations ≥6 months after PDT (range, 6-20 months); 2 BCC did not have sufficiently long follow-up. Clinically, 21 of 24 evaulable lesions treated with the two irradiance protocol were complete responses (88%). This is not significantly different from the single treatment 96% CCR rate for a larger study of 112 sBCC at 150 mW/cm2 (P = 0.148, Fisher exact test).4

4

Oseroff et al., in preparation.

There were no differences in initial PpIX fluorescence levels or shapes of photobleaching curves between the successes and failures (data not shown).

Reflectance spectra. Reflectance spectra between 475 and 800 nm were collected separately for both lesion and the perilesion margin and were used to correct fluorescence spectra. Figure 4 shows representative reflectance measured from one lesion and adjacent margin treated at 50 mW/cm2 before PDT and after 1.0 J/cm2 was delivered. Data between 625 and 640 nm are excluded because of bleed-through of the treatment laser into the reflectance channel. The spectra are remarkably similar before and after 1.0 J/cm2 PDT, with the characteristic features of oxyhemoglobin being prominent. Rigorous analysis of the reflectance spectra acquired over the full range of irradiances will be reported elsewhere.

Fig. 4.

Representative reflectance spectra taken from (A) a sBCC lesion and (B) adjacent margin taken before and after 1.0 J/cm2 PDT delivered at 50 mW/cm2. Data between 625 and 640 nm are excluded due to bleed-through of the treatment laser into the reflectance channel.

Fig. 4.

Representative reflectance spectra taken from (A) a sBCC lesion and (B) adjacent margin taken before and after 1.0 J/cm2 PDT delivered at 50 mW/cm2. Data between 625 and 640 nm are excluded due to bleed-through of the treatment laser into the reflectance channel.

Close modal

Pain. Pain was scored on a VAS scale for 26 lesions treated at 10, 20, 40, 50, and 60 mW/cm2 initial irradiance. Lesions treated at an initial irradiance of 150 mW/cm2 were excluded from pain measurements because of a previously established VAS score of >5. Pain experienced during the therapy ± SE is shown in Fig. 5A. Pain increased with escalating irradiance as anticipated by pilot evidence and supported by the literature (12, 17, 22). In this study, a pain threshold was established at an irradiance where ≥2 patients reported a VAS score of ≥4, requiring pain intervention. This threshold was found to be 60 mW/cm2, at which 2 of 5 patients reported VAS scores of 6 and 7.

Fig. 5.

A, VAS pain scores ± SE experienced with ALA-PDT of sBCC measured in this study along with results from Wang et al. (2), Algermissen et al. (11), and Holmes et al. (5), measured at higher irradiances. B, VAS pain scores ± SE during low-irradiance delivery and following changeover to 150 mW/cm2 during ALA-PDT of sBCC in our study, showing only moderate increases. No pain intervention was required for patients treated at 10 to 50 mW/cm2 before or after changeover. †, two individuals in group received 1% xylocaine injections after reporting a VAS > 4.

Fig. 5.

A, VAS pain scores ± SE experienced with ALA-PDT of sBCC measured in this study along with results from Wang et al. (2), Algermissen et al. (11), and Holmes et al. (5), measured at higher irradiances. B, VAS pain scores ± SE during low-irradiance delivery and following changeover to 150 mW/cm2 during ALA-PDT of sBCC in our study, showing only moderate increases. No pain intervention was required for patients treated at 10 to 50 mW/cm2 before or after changeover. †, two individuals in group received 1% xylocaine injections after reporting a VAS > 4.

Close modal

As mentioned previously, after the initial PpIX fluorescence contribution in the lesion had bleached by 90%, treatment irradiance was increased to 150 mW/cm2 and the treatment continued until 200 J/cm2 was delivered. Figure 5B shows the resulting change in average VAS score ± SE resulting from the increased irradiance. Only modest increases of >0.5 were noted in the 20, 40, and 50 mW/cm2 groups. The decrease in VAS score at 150 mW/cm2 for the 60 mW/cm2 group was primarily the result of the single 1% xylocaine injection given to the patients who reported VAS scores of 6 and 7 during the low-irradiance interval. Patients treated at 10 mW/cm2 reported an average increase of ∼1. There was no correlation between initial PpIX levels as determined by florescence and pain levels (data not shown).

During topical ALA-PDT, illumination consumes ground-state molecular oxygen supplied from the upper and mid-dermal capillaries and from diffusion from the atmosphere (23). High light dose rates together with significant photosensitizer levels can lead to oxygen depletion with loss of photodynamic efficiency. High irradiances also can increase pain during treatment. This work represents the first systematic clinical investigation of the effect of irradiance on PpIX photobleaching rate, PDT efficiency, and pain.

Beyond 20 mW/cm2, the initial slopes of the photobleaching curves increase with irradiance, indicating some oxygen limitations even at low PDT dose rates. The D80 also increases with irradiance, as shown in Fig. 3 and Table 1, illustrating less efficient bleaching at high-irradiance therapy and providing a rational way to determine a bleaching-equivalent fluence for a given treatment protocol.

PpIX photobleaching plots for each irradiance (Fig. 3) were generated by averaging the normalized bleaching curves for each lesion. The uncertainties shown in the irradiance averages are dominated by biological and patient-to-patient variances as shown in Fig. 3C. This was evident in the lesions treated at 10 mW/cm2, where one lesion had a D80 of 11.5 J/cm2 compared with an average of 4.3 J/cm2 for the other four lesions. However, we found no correlation between D80 values or initial PpIX fluorescence and clinical responses.

Our PpIX fluorescence measurements were done using the treatment beam as the fluorescence excitation source. This was done in part to maximize the overlap between the treatment volume and the volume from which fluorescence data were acquired. Nevertheless, there are undoubtedly heterogeneities in PpIX accumulation with depth into the skin and the tumor, and the fluorescence measurements preferentially sample the upper portion of the epidermis and the carcinoma. For these reasons, the measured fluorescence may not reflect the full distribution of PpIX and the bleaching may not report the heterogeneity in photodynamic dose deposition. Therefore, the total light dose required to fully use the PpIX within the lesion is unknown. Prior work at high irradiances in multiple lesions in a small number of patients suggested a sigmoidal dose-response, with the upper knee of the curve ∼100 to 150 J/cm2 (24). A single 633 nm light dose of 75 J/cm2 at 50 mW/cm2 produced a complete clinical response rate of 89% at 12 months (25), and 635 nm light at a dose of 12.6 J/cm2 at 7 mW/cm2 gave a 1-year complete clinical response rate of 84% (26). To deliver our “standard” light dose of 200 J/cm2 to all lesions, we increased our light delivery rate to 150 mW/cm2 after 90% photobleaching. There was no additional pain, and this approach would represent a modest time penalty of <14% for initial irradiances of 20 to 50 mW/cm2. Alternatively, because 200 J/cm2 delivered at 150 mW/cm2 is ∼7 times the D80 fluence, it is possible that a bleaching-equivalent, and similarly efficacious, fluence would be deposited by 7 times the D80 fluence at a lower irradiance. At the highest pain-free irradiance in this study, 50 mW/cm2, this corresponds to a fluence of 51.3 J/cm2 and a treatment time of 17.1 min. However, a fluence-response study may be necessary to determine the appropriate low-irradiance light dose.

The persistence of perilesional PpIX fluorescence at high irradiances with three of seven sites having >20% of initial fluorescence at 200 J/cm2 (Fig. 3) is consistent with lack of oxygenation presumably due to lack of perfusion from reversible vasospasm. The mechanisms leading to discrepancies in D80 between lesion and margin at 150 mW/cm2 are not fully understood, although there are several physiologic differences between these regions. sBCC blood vessels, for example, are disorganized, occupy more tissue volume, are larger in diameter, and do not have the ability to contract due to the absence of smooth muscle (2729). Although oxygenation changes have been studied in nodular BCC during PDT with systemically administered Photofrin (9), we are not aware of any experiments to date that have investigated blood perfusion during ALA-PDT in human subjects, although investigators have reported on perfusion in the lesion and perilesion margin before and following ALA-PDT of sBCC, showing dissimilar trends between these regions. It has been shown that, before ALA-PDT, perfusion in the sBCC lesion was several times that found in the margin, and blood velocity was also elevated (11, 27, 30, 31). Additionally, investigators have found an increase in the ratio of perfusion in the lesion compared with the margin immediately following PDT (30, 31). Tumor margins and small islands of carcinoma cells may be fed predominately by normal vessels and could be protected from photodynamic injury.

Broadband reflectance measurements provide estimates of blood oxygen saturation and can be used to correct fluorescence spectra. To make reflectance measurements, treatment irradiation was periodically paused and broadband light transmitted through the treatment fiber. The interruptions were brief and infrequent to prevent any treatment fractionation effects; acquisition times were ∼1 s and the duty cycle was <10%. The differences in initial efficiency of photobleaching shown in Fig. 3 suggest that tissue deoxygenation was rapid. The representative reflectance data presented in Fig. 4, however, do not reflect any apparent changes in hemoglobin oxygen saturation. The reasons for this are uncertain, but it is possible that these superficial lesions are supplied with oxygen primarily via diffusion from the air (23), and as a result, hemoglobin oxygen saturation may not be perturbed even if photochemically relevant oxygen depletion occurs. The measured hemoglobin oxygen saturation during PDT is influenced by photochemical oxygen consumption, changes in blood flow, and possible changes in metabolic oxygen consumption. The influences of these factors are being investigated using a newly developed comprehensive PDT model (32). This analysis will inform the design of future studies.

Acute pain associated with ALA-PDT is the primary side effect of therapy. It can last for several hours (22) and has been noted to depend on light source, irradiance (18), lesion location (33), and number and size of treatment locations (33, 34). Pain is a significant obstacle to the general clinical acceptability of ALA-PDT, particularly for patients who are among the best candidates for and benefit most from PDT, such as those with nevoid basal cell carcinoma syndrome, who often require multiple simultaneous treatments and large treatment fields. Therefore, an important goal is to find optimal treatment protocols that equal or exceed the efficacy found using current protocols while limiting patient discomfort, reducing the need for anesthesia, and minimizing time in the clinic. Some investigators have noticed less pain associated with ALA-PDT delivered at lower fluence rates (17, 18) or excitation wavelengths that do not coincide with PpIX absorption peaks (35), which is consistent with our evidence that pain is reduced with lower absorbed irradiance. The majority of ALA-PDT studies cite reactive pain-mitigating techniques including analgesics, water spray, airstreams, and conscious sedation (13, 17, 22, 34), yet limited work has been done on the relationship between PDT-induced pain and various treatment variables. In this study, which examined treatment fields 25 mm in diameter during escalating irradiance, irradiances below 60 mW/cm2 alleviated the majority of discomfort associated with ALA-PDT and allowed delivery without requiring pain intervention.

Based on the pain assessments in this study, irradiances using a ∼633 nm laser below 60 mW/cm2 are good candidates for a low-irradiance treatment protocol. In addition to laser or LED sources corresponding to the PpIX absorption peak, however, several investigators perform PDT using broadband lamps and off-peak absorption laser sources (36). As shown by Ericson et al. (21), for a given irradiance, these sources are less efficient at exciting PpIX and they therefore have a lower effective irradiance than a laser source operating near peak PpIX absorption as was used in this study. This suggests that alternative irradiation sources operating below 60 mW/cm2 would also be below the pain threshold for sBCC lesions of comparable size. Because pain is expected to increase with irradiated areas, even lower irradiances may be necessary in some clinical situations. Note that PDT dose rate can be decreased by lowering either irradiance or PpIX concentration (e.g., with short application times). Both will decrease pain and increase efficiency. However, because of loss of PpIX through photobleaching, the use of low PpIX risks destroying the sensitizer before oxidative damage sufficient to cure the lesion is caused. Note that although this study used ALA in a cream vehicle, we expect similar results from Levulan Kerasticks applied with at least two passes and then occluded. To the extent that use of the methyl ester of ALA produces comparable levels of PpIX, our results should be at least qualitatively transferable.

Based on photodynamic efficiency, determined by bleaching, and the establishment of a pain threshold at 60 mW/cm2 for single, 2.5-cm-diameter treatment fields, we conclude that for multiple and larger lesions ∼40 mW/cm2 at ∼633 nm may represent an optimal irradiance for ALA-PDT of sBCC, motivating a low-irradiance trial investigating efficacy at this irradiance. Assessment of treatment efficacy, durability, and cosmetic outcome is continuing for all irradiance groups. We expect comparable photobleaching rates for sBCC of other sizes, although, as noted, large or multiple treatment areas may increase pain and benefit from a reduction in irradiance. Currently, we are measuring irradiance-dependent photobleaching and pain in nodular BCC and in squamous cell carcinoma in situ. Although we have not found a significant change in efficacy, ALA-PDT responses depend on both direct cell kill and host inflammatory and immune responses; it is not evident how these vary with irradiance in patients. In addition to the low-irradiance clinical trial, we are working on models of PDT dose deposition in skin and also anticipate studies assessing blood perfusion in sBCC during ALA-PDT.

DUSA Pharmaceuticals provided A.R. Oseroff with ALA used in this research.

Grant support: NIH grant P01 CA55719 (A.R. Oseroff and T.H. Foster), Roswell Park Cancer Institute Cancer Center support grant CA16056, and T32 HL66988 (W.J. Cottrell).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Sari Fien and Anne Calvaruso for clinical expertise, Barbara Henderson for a critical review of the manuscript, and Leslie Blumenson for statistical assistance in development of the protocol.

1
Dougherty TJ, Gomer CJ, Henderson BW, et al. Photodynamic therapy.
J Natl Cancer Inst
1998
;
90
:
889
–905.
2
Zeitouni NC, Oseroff AR, Shieh S. Photodynamic therapy for nonmelanoma skin cancers. Current review and update.
Mol Immunol
2003
;
39
:
1133
–6.
3
Foster TH, Murant RS, Bryant RG, Knox RS, Gibson SL, Hilf R. Oxygen consumption and diffusion effects in photodynamic therapy.
Radiat Res
1991
;
126
:
296
–303.
4
Sitnik TM, Henderson BW. The effect of fluence rate on tumor and normal tissue responses to photodynamic therapy.
Photochem Photobiol
1998
;
67
:
462
–6.
5
Robinson DJ, de Bruijn HS, van der Veen N, Stringer MR, Brown SB, Star WM. Fluorescence photobleaching of ALA-induced protoporphyrin IX during photodynamic therapy of normal hairless mouse skin: the effect of light dose and irradiance and the resulting biological effect.
Photochem Photobiol
1998
;
67
:
140
–9.
6
Robinson DJ, de Bruijn HS, van der Veen N, Stringer MR, Brown SB, Star WM. Protoporphyrin IX fluorescence photobleaching during ALA-mediated photodynamic therapy of UVB-induced tumors in hairless mouse skin.
Photochem Photobiol
1999
;
69
:
61
–70.
7
Gibson SL, VanDerMeid KR, Murant RS, Raubertas RF, Hilf R. Effects of various photoradiation regimens on the antitumor efficacy of photodynamic therapy for R3230AC mammary carcinomas.
Cancer Res
1990
;
50
:
7236
–41.
8
Finlay JC, Conover DL, Hull EL, Foster TH. Porphyrin bleaching and PDT-induced spectral changes are irradiance dependent in ALA-sensitized normal rat skin in vivo.
Photochem Photobiol
2001
;
73
:
54
–63.
9
Henderson BW, Busch TM, Vaughan LA, et al. Photofrin photodynamic therapy can significantly deplete or preserve oxygenation in human basal cell carcinomas during treatment, depending on fluence rate.
Cancer Res
2000
;
60
:
525
–9.
10
Peng Q, Berg K, Moan J, Kongshaug M, Nesland JM. 5-Aminolevulinic acid-based photodynamic therapy: principles and experimental research.
Photochem Photobiol
1997
;
65
:
235
–51.
11
Wennberg AM, Larko O, Lonnroth P, Larson G, Krogstad AL. δ-Aminolevulinic acid in superficial basal cell carcinomas and normal skin—a microdialysis and perfusion study.
Clin Exp Dermatol
2000
;
25
:
317
–22.
12
Wang I, Bendsoe N, Klinteberg CA, et al. Photodynamic therapy vs. cryosurgery of basal cell carcinomas: results of a phase III clinical trial.
Br J Dermatol
2001
;
144
:
832
–40.
13
Morton CA, Brown SB, Collins S, et al. Guidelines for topical photodynamic therapy: report of a workshop of the British Photodermatology Group.
Br J Dermatol
2002
;
146
:
552
–67.
14
Angell-Petersen E, Christensen C, Muller CR, Warloe T. Phototoxic reaction and porphyrin fluorescence in skin after topical application of methyl aminolaevulinate.
Br J Dermatol
2007
;
156
:
301
–7.
15
Horn M, Wolf P, Wulf HC, et al. Topical methyl aminolaevulinate photodynamic therapy in patients with basal cell carcinoma prone to complications and poor cosmetic outcome with conventional treatment.
Br J Dermatol
2003
;
149
:
1242
–9.
16
Soler AM, Warloe T, Berner A, Giercksky KE. A follow-up study of recurrence and cosmesis in completely responding superficial and nodular basal cell carcinomas treated with methyl 5-aminolaevulinate-based photodynamic therapy alone and with prior curettage.
Br J Dermatol
2001
;
145
:
467
–71.
17
Algermissen B, Osterloh D, Philipp CM, Berlien HP. Management of ALA-PDT induced pain sensations.
Med Laser Appl
2003
;
18
:
57
–64.
18
van Oosten EJ, Kuijpers IM, Thissen MRTM. Different pain sensations in photodynamic therapy of nodular basal cell carcinoma: results from a prospective trial and a review of the literature.
Photodiag Photodyn
2006
;
3
:
61
–8.
19
Cottrell WJ, Oseroff AR, Foster TH. Portable instrument that integrates irradiation with fluorescence and reflectance spectroscopies during clinical photodynamic therapy of cutaneous disease. Rev Sci Instrum 2006;77:064302.
20
Wu J, Feld MS, Rava RP. Analytical model for extracting intrinsic fluorescence in turbid media.
Appl Opt
1993
;
32
:
3595
.
21
Ericson MB, Sandberg C, Stenquist B, et al. Photodynamic therapy of actinic keratosis at varying fluence rates: assessment of photobleaching, pain and primary clinical outcome.
Br J Dermatol
2004
;
151
:
1204
–12.
22
Holmes MV, Dawe RS, Ferguson J, Ibbotson SH. A randomized, double-blind, placebo-controlled study of the efficacy of tetracaine gel (Ametop) for pain relief during topical photodynamic therapy.
Br J Dermatol
2004
;
150
:
337
–40.
23
Stucker M, Struk A, Altmeyer P, Herde M, Baumgartl H, Lubbers DW. The cutaneous uptake of atmospheric oxygen contributes significantly to the oxygen supply of human dermis and epidermis.
J Physiol
2002
;
538
:
985
–94.
24
Oseroff AR, Shieh S, Frawley NP, et al. Treatment of diffuse basal cell carcinomas and basaloid follicular hamartomas in nevoid basal cell carcinoma syndrome by wide-area 5-aminolevulinic acid photodynamic therapy.
Arch Dermatol
2005
;
141
:
60
–7.
25
de Haas ER, Kruijt B, Sterenborg HJ, Martino Neumann HA, Robinson DJ. Fractionated illumination significantly improves the response of superficial basal cell carcinoma to aminolevulinic acid photodynamic therapy.
J Invest Dermatol
2006
;
126
:
2679
–86.
26
Langmack K, Mehta R, Twyman P, Norris P. Topical photodynamic therapy at low fluence rates—theory and practice.
J Photochem Photobiol B
2001
;
60
:
37
–43.
27
Stanton AW, Drysdale SB, Patel R, et al. Expansion of microvascular bed and increased solute flux in human Basal cell carcinoma in vivo, measured by fluorescein video angiography.
Cancer Res
2003
;
63
:
3969
–79.
28
Newell B, Bedlow AJ, Cliff S, Drysdale SB, Stanton AW, Mortimer PS. Comparison of the microvasculature of basal cell carcinoma and actinic keratosis using intravital microscopy and immunohistochemistry.
Br J Dermatol
2003
;
149
:
105
–10.
29
Sitnik TM, Hampton JA, Henderson BW. Reduction of tumour oxygenation during and after photodynamic therapy in vivo: effects of fluence rate.
Br J Cancer
1998
;
77
:
1386
–94.
30
Wang I, Andersson-Engels S, Nilsson GE, Wardell K, Svanberg K. Superficial blood flow following photodynamic therapy of malignant non-melanoma skin tumours measured by laser Doppler perfusion imaging.
Br J Dermatol
1997
;
136
:
184
–9.
31
Enejder AM, AF Klinteberg C, Wang I, et al. Blood perfusion studies on basal cell carcinomas in conjunction with photodynamic therapy and cryotherapy employing laser-Doppler perfusion imaging.
Acta Derm Venereol
2000
;
80
:
19
–23.
32
Wang KK, Mitra S, Foster TH. A comprehensive mathematical model of microscopic dose deposition in photodynamic therapy.
Med Phys
2007
;
34
:
282
–93.
33
Grapengiesser S, Ericson M, Gudmundsson F, Larko O, Rosen A, Wennberg AM. Pain caused by photodynamic therapy of skin cancer.
Clin Exp Dermatol
2002
;
27
:
493
–7.
34
Sandberg C, Stenquist B, Rosdahl I, et al. Important factors for pain during photodynamic therapy for actinic keratosis.
Acta Derm Venereol
2006
;
86
:
404
–8.
35
Fritsch C, Stege H, Saalmann G, Goerz G, Ruzicka T, Krutmann J. Green light is effective and less painful than red light in photodynamic therapy of facial solar keratoses.
Photodermatol Photoimmunol Photomed
1997
;
13
:
181
–5.
36
Brancaleon L, Moseley H. Laser and non-laser light sources for photodynamic therapy.
Lasers Med Sci
2002
;
17
:
173
–86.