Kaplan-Meier curves for disease-free rates for basal cell carcinoma and squamous cell carcinoma. Projected disease-free rate 36 months after therapy: 50% (95% confidence interval [CI], 43%-57%) for basal cell carcinoma and 8% (95% CI, 7%-9%) for squamous cell carcinoma (P<.001, log-rank test).
Recurrent basal cell carcinoma in a 66-year-old man 2 years after photodynamic therapy with δ-aminolevulinic acid (hematoxylin-eosin,×50).
Fink-Puches R, Soyer HP, Hofer A, Kerl H, Wolf P. Long-term Follow-up and Histological Changes of Superficial Nonmelanoma Skin Cancers Treated With Topical δ-Aminolevulinic Acid Photodynamic Therapy. Arch Dermatol. 1998;134(7):821-826. doi:10.1001/archderm.134.7.821
To investigate the immediate and long-term effects of photodynamic therapy with δ-aminolevulinic acid (ALA-PDT) on superficial basal cell carcinomas (BCC) and superficial squamous cell carcinomas (SCC).
Retrospective study with 60 months of maximal follow-up.
University-based hospital in Graz, Austria.
Forty-seven subjects with a total of 95 superficial BCC and 35 superficial SCC.
A compound of 20% δ-aminolevulinic acid was topically applied under an occlusive and light-shielding dressing before exposure to either UV-A or different wave bands of polychromatic visible light (full-spectrum visible light, >515, >570, or >610 nm).
Main Outcome Measures
Primary tumor responses and recurrence rates in the long-term follow-up, as well as histological changes associated with ALA-PDT, were studied.
The complete primary response rate for all wave bands of light was 86% (82/95) for superficial BCC and 54% (19/35) for superficial SCC. There was no statistically significant difference among the response rates to the different wave bands of light. After a median follow-up of 19 months (range, 3-60 months) for BCC and 8 months (range, 3-47 months) for SCC, the overall recurrence rate was 44% (36/81) and 69% (11/16), respectively. At 36 months after therapy, the projected disease-free rate was 50% (95% confidence interval, 43%-57%) for BCC vs 8% (95% confidence interval, 7%-9%) for SCC (P<.001, log-rank test). Histopathologic studies revealed a significant increase of fibrosis in the dermis after ALA-PDT and appearance of a sharp border between fibrotic and nonfibrotic tissue. In 15 of 16 BCC examined, the border between fibrotic and nonfibrotic tissue was deeper in the dermis than the maximum tumor thickness before therapy (P<.001, Wilcoxon signed rank test). Similar histopathologic observations were made in SCC.
Our study revealed poor long-term cure rates for superficial BCC and SCC treated with topical ALA-PDT and visible light. The histopathologic observations showing remarkable fibrosis in the dermis indicated that the effect of ALA-PDT reached deeper than the initial depth of invasiveness of the neoplastic tissue, suggesting in turn that the poor long-term results of ALA-PDT cannot be explained by insufficient penetration of the therapy effect.
THE INCREASING incidence of nonmelanoma skin cancers, ie, basal cell carcinomas (BCC) and squamous cell carcinomas (SCC), has led to the search for new therapeutic modalities. Photodynamic therapy with topically applied δ-aminolevulinic acid (ALA-PDT) is a novel form of therapy for neoplasms of the skin and other organs. The principle of ALA-PDT is that topically applied ALA is metabolized by the tumor cells into photosensitizing concentrations of endogenous porphyrins, particularly protoporphyrin IX. Irradiation with visible light then leads to the selective destruction of tumor tissue.1,2 Therapy with ALA-PDT is effective for superficial nonmelanoma skin tumors, and high primary clinical response rates with excellent cosmetic results have been reported.1,3- 5 Although ALA-PDT is used today experimentally in many centers around the world,6 most clinical studies of ALA-PDT of skin tumors have involved only a small number of cases and a clinical follow-up of only a few months. In the present study, we evaluated primary tumor responses and recurrence rates in a long-term follow-up (to 60 months) of 47 patients, with a total of 95 superficial BCC and 35 superficial SCC, who were treated with topical ALA and light of different wave bands.
Forty-seven patients (29 women and 18 men), between 47 and 90 years old (median age, 68 years), with a total of 95 superficial BCC (ie, defined as tumors with basaloid cells occurring no deeper than the papillary dermis) and 35 superficial SCC (ie, defined as SCC confined to the papillary dermis; none of the SCC were of the in situ type) were treated after giving their informed signed consent. Twenty-two patients accounted for 88 superficial BCC; 23 patients accounted for 32 superficial SCC. In addition, 1 man had 3 BCC and 1 SCC, and 1 woman had 4 BCC and 2 SCC. The primary clinical response of 32 BCC and 1 SCC in 4 of the 47 patients of the present study had been reported previously.2,5 Of the 95 BCC, 11 were located on the face, 4 on the scalp, 3 on the neck, 11 on the chest, 65 on the back, and 1 on a lower extremity. The diameter of the lesions ranged from 0.5 to 4.2 cm. Of the 35 SCC, 11 were located on the face, 8 on the scalp, 5 on the neck, 4 on the trunk, 4 on the upper extremities, and 3 on the lower extremities. The diameter of the lesions ranged from 1.5 to 6 cm.
Biopsy specimens from skin tumors were taken from each patient before therapy. If a patient had more than 1 tumor, a biopsy sample was taken only from 1 representative lesion. After therapy, biopsies were performed to confirm the response, recurrence, or both in cases in which the clinical evaluation was ambiguous. In addition, in some recurrent lesions the entire tumor tissue was examined after surgical excision. The specimens were embedded in paraffin, processed routinely, and stained with hematoxylin-eosin.
Lesions were treated by topical application of ALA (Fluka Chemie AG, Buchs, Switzerland) dissolved in a proprietary oil-in-water emulsion (Doritin) (Asta Medica Arzneimittel GmbH, Vienna, Austria). The ALA-containing emulsion was applied under occlusive (Opraflex, Lohmann GmbH & Co, Neuwied, Germany) and light-shielding dressings (Fablon, Firma Ludwig & Co, Graz, Austria) to the lesions and to approximately 0.5 cm of adjacent skin. The emulsion was left on for 4 hours to allow penetration of ALA into the tissue and synthesis of porphyrin.
After dressings were removed from ALA-treated areas, the presence of endogenous porphyrins was evaluated by rating the red fluorescence under a Wood light in a darkened room. The amount of surface fluorescence at a lesion's site was rated on a scale from 0 to 4, with 0 indicating negative; 1, weak; 2, moderate; 3, strong; and 4, very strong.
The skin tumors were irradiated with a slide projector (model P-2000, Leica, Leitz-Wetzlar, Germany) equipped with a 250-W lamp (flecta halogen Projektionslampe, 24 V/250 W, Firma reflecta, Schwabach, Germany) and a Vario-Elmaron-p-1:3.5, 110- to 200-mm lens (Firma Leica, Leitz-Wetzlar) at an irradiance ranging from 50 to 100 mW/cm2, as measured by a calibrated photodiode BPW 34 at 10 and 30 cm from the lens. Lesions were irradiated either with unfiltered full-spectrum visible light or with filtered visible light produced by the long-wave-pass color glass filters OG 515, OG 570, or RG 610 (Schott Glaswerke, Mainz, Germany) to eliminate wavelengths less than 515, 570, and 610 nm, respectively. These filters were used in certain cases because it had previously been found5 that they can reduce pain caused by ALA-PDT. The distance from the irradiation field to the lens ranged from 10 to 30 cm; the diameter of the irradiation field ranged from 7 to 11.5 cm. The irradiance of the light in the irradiation fields varied between±20% of the rates specified above. Certain lesions were treated with UV-A produced by UV-A equipment (Sellas GmbH, Gevelsberg, Germany) at an irradiance of 37 mW/cm2. The relative emission spectrum of the slide projector equipped with the different long-wave-pass color glass filters and the UV-A equipment is reported elsewhere.7
Of the 95 BCC, 70 were treated with full-spectrum visible light, 15 with filtered visible light of wavelengths greater than 570 nm (n=12) or greater than 610 nm (n=3), and 10 with UV-A (Table 1). Of the 35 SCC, 25 were treated with full-spectrum visible light, and 10 with filtered visible light of wavelengths greater than 515 nm (n=1), greater than 570 nm (n=1), or greater than 610 nm (n=8) (Table 1).
During ALA-PDT, the fluorescence of the treated lesions was determined under a Wood light at regular 5- to 10-minute intervals to monitor photobleaching. Lesions were exposed until the typical red porphyrin fluorescence disappeared, an edematous or vesiculous reaction occurred at the treated site, or both. The total exposure times for the different wave bands of visible light ranged from 3 to 32 minutes for BCC and from 5 to 30 minutes for SCC. These exposure times led to total light doses ranging from 18 to 131 J/cm2 (median total light dose, 60 J/cm2) for BCC and from 5.3 to 180 J/cm2 (median total light dose, 61 J/cm2) for SCC. The exposure time for each of the 10 BCC treated with UV-A was 30 seconds, resulting in a UV-A dose of 1.1 J/cm2.
The immediate phototoxic reaction at a lesion's site was rated within 5 to 10 minutes after light exposure by at least 1 of the investigators (R.F.-P. or P.W.) on a scale from 0 to 4, with 0 indicating no response; 1, faint erythema; 2, fiery red erythema; 3, erythema and edema; and 4, vesiculation.
Two to 4 weeks after ALA-PDT, clinical response to therapy was evaluated. Complete response was defined as the absence of a clinically evident lesion at the treatment site. Partial response was defined as a marked reduction (>50%) in tumor size as determined by clinical evaluation. No response was defined as insignificant reduction (<50%) in tumor size as determined by clinical evaluation.
The χ2 test or the Fisher exact test for small sample numbers was used to evaluate statistical significance of differences in clinical response and follow-up between different groups. The Student t test was used to compare quantitative variables. The Spearman rank correlation was used to evaluate the relation between lesion diameter, light dose, fluorescence, phototoxic reaction, and clinical response, as well as tumor thickness and time of recurrence of the lesions. Disease-free rates were calculated using the life-table method of Kaplan and Meier. Univariate disease-free interval analysis was done using Kaplan-Meier tables and the log-rank test. To evaluate differences in histopathologic signs, the Wilcoxon signed rank test and the paired sign test was used. A P value of .05 or less was considered to indicate statistical significance.8
For 66 BCC, the amount of surface fluorescence at the lesion site after ALA photosensitization was rated in 2 (3%) as weak; 33 (50%), moderate; 29 (44%), strong; and 2 (3%), very strong.
For 32 SCC, the amount of surface fluorescence was rated in 6 (19%) as weak; 14 (44%), moderate; 8 (25%), strong; and 4 (12%), very strong.
For 48 BCC, 4 (8%) showed faint erythema; 19 (40%), fiery red erythema; 22 (46%), erythema and edema; and 3 (6%), vesiculation.
For 28 SCC, 1 (3%) showed no response; 3 (11%), faint erythema; 10 (36%), fiery red erythema; 9 (32%), erythema and edema; and 5 (18%), vesiculation.
Taking all wave bands of light together, the complete primary response rate for BCC was 86% (82/95) (Table 1). The cosmetic results after ALA-PDT were excellent. There was a statistically significant correlation of clinical response of BCC with lesion diameter (P=.02), fluorescence (P<.001), light dose (P=.02), and phototoxic reaction (P<.001) (Table 2). These data suggest that high fluorescence and phototoxic reaction are the most important factors for successful ALA-PDT in BCC.
There was no statistically significant difference (P>.05, χ2 test) among the response rates to the different wave bands of light. The complete clinical response rate was 86% (60/70 BCC) for full-spectrum visible light, 87% (13/15 BCC) for filtered visible light (>570 or >610 nm), and 90% (9/10 BCC) for UV-A.
There was no site dependence of the clinical response of BCC. The complete clinical response rate was 89% (68/76) for lesions on the trunk vs 78% (14/18) for lesions on the face, scalp, or neck (P>.05, χ2 test). There was no difference between the clinical response rates of BCC in men and women (P>.05, Student t test).
Taking all wave bands of light together, the complete primary response rate for SCC was 54% (19/35). There was no statistically significant correlation (P>.05, Spearman rank correlation) of clinical response of SCC with lesion diameter, fluorescence, light dose, and phototoxic reaction of ALA-PDT. There was also no site dependence of the clinical response of SCC. The complete clinical response rate for lesions on the face, scalp, or neck was 63% (15/24 SCC) vs 36% (4/11 SCC) for lesions on the trunk or extremities (P>.05, Fisher exact test). There was no difference between the clinical response rates of SCC in men and women (P>.05, Student t test). There was no statistically significant difference (P>.05, Fisher exact test) between the primary response rate of SCC to full-spectrum visible light vs filtered visible light. The complete response rate was 60% (15/25) for SCC treated with full-spectrum visible light vs 40% (4/10) for SCC treated with filtered visible light (>515, >550, or >610 nm).
Of the 82 BCC that showed a complete primary response, 81 were followed up for 3 to 60 months (median follow-up, 19 months), while 1 was unavailable for follow-up. Of the 81 BCC, 36 (44%) developed recurrences, while 45 (56%) did not. The recurrent lesions were treated by excision (n=4), cryosurgery (n=17), or carbon dioxide laser (n=15). A punch biopsy before therapy of a recurrent lesion was only performed when the clinical diagnosis was ambiguous. The median time to recurrence was 27 months (range, 3-60 months).
Of the 19 SCC that showed a complete primary clinical response, 16 were followed up for 3 to 47 months (median follow-up, 8 months), while 3 SCC in 3 patients were unavailable for follow-up. Of these 16 SCC, 5 (31%) remained complete responders, while 11 (69%) developed recurrences.
The disease-free rates for BCC and SCC are shown in Figure 1. There was a statistically significant difference between them. For instance, 36 months after therapy, the projected disease-free rate was 50% (95% confidence interval, 43%-57%) for BCC vs 8% (95% confidence interval, 7%-9%) for SCC (P<.001, log-rank test).
In 16 of 36 recurrent BCC, we compared histological sections from punch biopsy specimens before ALA-PDT with sections from excised lesions (n=4) or punch biopsy specimens (n=12) after therapy.
Histopathologic studies revealed a significant homogeneous fibrosis in the dermis in all 16 cases after, but not before, ALA-PDT (Table 3). In all cases, there was a sharp border between fibrotic and nonfibrotic tissue in the dermis after ALA-PDT. In 15 of 16 lesions, the depth of fibrosis after therapy (mean depth, 1.04 mm; range, 0.5-1.5 mm) was greater than the tumor thickness before therapy (mean depth, 0.35 mm; range, 0.1-1.4 mm) (P<.001, Wilcoxon signed rank test). Figure 2 shows the histological changes in a BCC 2 years after ALA-PDT. Similar results were found in 9 SCC (comparison of sections of 9 punch biopsy specimens before ALA-PDT vs those of 5 punch biopsy specimens and 4 surgically excised lesions after therapy), but the fibrotic changes after ALA-PDT were not as prominent as in BCC (Table 4). There were no correlations of tumor thickness before therapy with time of recurrence for BCC as well as SCC (P>.05, Spearman rank correlation).
The treatment goals for BCC and SCC are complete tumor removal and minimization of cosmetic and functional defects, achieved mainly by excision surgery. However, a variety of other less invasive, similarly effective, but cosmetically more satisfactory treatment modalities are available to treat nonmelanoma skin cancer. These modalities include electrocurettage, carbon dioxide laser treatment,9 cryotherapy,10 topical chemotherapy,11 and treatment with biologic response modifiers such as the interferons.12 The cure rates for BCC and SCC treated with those modalities are usually very high. For instance, the cure rates for BCC treated by surgical excision vary from 90% to 98%.13- 15 Micrographic (Mohs) surgery was reported16 to result in a 98% to 99% cure rate for BCC and a 94% cure rate for SCC after 5 years of follow-up. Treatment of BCC with curettage combined with electrodesiccation, particularly for not too large and not too deeply growing tumors, produced long-term disease-free rates between 90% and 95%.17 The cure rate for BCC treated with x-rays falls between 88% and 90%.17 Litwin et al18 achieved an 86% cure rate for BCC and SCC using the topical chemotherapeutic agent fluorouracil (5-fluorouracil) (5%, 10%, and 20% concentrations) after follow-up periods ranging from 4 to 20 months (average, 9 months), while Stoll et al19 noted complete tumor resolution after treatment with 5% to 20% fluorouracil in 95% of superficial BCC without recurrence within 5 years. Kuflik and Gage10 reported a 5-year cure rate of 99% for BCC and 96% for SCC achieved by cryosurgery. Ikic et al20 demonstrated that interferon treatment can produce a persistent cure (5 years) in patients with BCC and SCC (recurrence rates of 2% for BCC and 4% for SCC).
In the present study, we examined the primary response rate and long-term recurrence rate of superficial BCC and SCC after ALA-PDT with polychromatic wave bands of light. The complete primary response was 86% for superficial BCC and 54% for superficial SCC. The amount of fluorescence after application of ALA and the immediate phototoxic reaction at the lesion site were statistically highly significantly correlated to the clinical response (BCC, P<.001). Thus, high fluorescence and strong phototoxic reaction at the site of a treated lesion seem to be the best predictive factors for clinical response. There was no site dependence of the clinical response of BCC or SCC. Also, the response rates did not statistically significantly differ with regard to the different wave bands used for light exposure in BCC or SCC. Interestingly, small doses of UV-A could be successfully used for the treatment of BCC. This observation may be because of the great overlap of the emission spectrum of the UV-A equipment used in the present study with the protoporphyrin-IX absorbance spectrum.7 The recurrence rate was 44% for BCC and 69% for SCC on follow-up. In the long-term follow-up, the projected disease-free rate was 50% for BCC vs 8% for SCC at 36 months (P<.001, log-rank test). Thus, our present long-term results for topical ALA-PDT with polychromatic light in the treatment of superficial skin cancers are poor, particularly for SCC.
There have been previous reports1,5,21- 23 on response rates for skin cancers treated with ALA-PDT. In an early report on topical ALA-PDT with polychromatic light, Kennedy et al1 noted a 90% complete remission of BCC at a follow-up of 2 to 3 months. In 1992, Kennedy and Pottier21 reported a 3-month complete remission rate of 79%, which somehow agrees with our primary response rate. Wolf et al5 previously reported a complete clinical response in 36 (97.3%) of 37 superficial BCC and 5 (83.3%) of 6 superficial SCC at 4 to 8 weeks after ALA-PDT. After a follow-up of 6 months, Wennberg et al22 reported a 92% cure rate for superficial BCC and a 61% cure rate for Bowen disease by using a 20% ALA solution and a filtered xenon lamp as the light source. Lui et al,23 who performed topical ALA-PDT with polychromatic visible light greater than 570 nm, achieved a clinical complete response in 7 (88%) of 8 superficial BCC and in 2 (67%) of 3 superficial SCC; in contrast, however, only 4 (50%) of 8 superficial BCC and 2 (67%) of 3 SCC achieved a complete histological response. Because of these data, Lui et al first suggested that there might be an unacceptably high tumor recurrence rate following ALA-PDT. However, higher long-term cure rates were reported in certain studies of ALA-PDT using laser light. For instance, Svanberg et al4 obtained a 100% cure rate for superficial BCC and 90% for Bowen disease after ALA-PDT using laser light emitted at 630 nm and a follow-up of 6 to 14 months. Calzavara-Pinton3 reported a cure rate of 87% for superficial BCC and 83% for SCC by repetitive PDT after topical application of ALA and irradiation with a 630-nm light, from data obtained with a dye laser at 24 to 36 months of follow-up. On the other hand, Cairnduff et al,24 who also used 630-nm light for ALA-PDT, showed that only 50% of patients with superficial BCC remained disease-free at a median follow-up of 17 months. Thus, the better results achieved in the ALA-PDT studies by Svanberg and Calzavara-Pinton cannot be solely attributed to the use of laser light. In any case, the reasons for the poor long-term results for BCC and particularly for SCC in our study remain unclear. However, the results demonstrate that long-term follow-up is necessary to evaluate the effect of ALA-PDT.
One factor that may have contributed to the failure of ALA-PDT is insufficient marking of tumors by protoporphyrin IX.25 Indeed, Martin et al25 previously demonstrated the incomplete presence of protoporphyrin-IX fluorescence in a significant number of BCC and suggested that topical ALA delivered with the present photodynamic protocols may not be a reliable regimen for PDT of BCC. Alternatively, they suggested that superficial tumors might contain tumor structures that lie too deep in the skin to be accessible to the topical ALA-PDT effect.
Our histopathologic studies indicated that ALA-PDT induces remarkable changes within the dermis of treated skin. In all 16 cases of BCC examined histologically before and after therapy, there was increased fibrosis after therapy and a sharp border had formed between fibrotic and nonfibrotic tissue in the dermis. The sharp border might reflect the result of the photodynamic threshold-dose, ie, the product of light and porphyrin accumulation that causes the death of cells and tissue.26 In 15 of 16 BCC examined, the depth of fibrosis after ALA-PDT was greater than tumor thickness before therapy (Table 3). Similar results were found in SCC, but the fibrosis was not as marked as in BCC (Table 4). These observations indicated that ALA-PDT induced changes within the skin that reached deeper than the tumor cells of superficial BCC and SCC. Thus, the poor long-term results of ALA-PDT in the present study cannot be easily explained by insufficient penetration of the therapy effect.
The mechanism of ALA-PDT–induced fibrosis remains unclear at present. However, sclerotic skin changes with thickening of collagen bundles can occur in certain porphyrias, and an increase of collagen biosynthesis was observed following incubation of fibroblasts with uroporphyrins.27 One theory holds that exposure to ionizing radiation induces a premature differentiation process in the fibroblast-fibrocyte cell system, resulting in an enhanced accumulation of postmitotic fibrocytes characterized by a severalfold increase in collagen synthesis.28 Transforming growth factor β is probably the major cytokine responsible for the fibrotic reaction in normal tissues following radiation therapy.29 Another theory suggests that transforming growth factor β is involved in the control of regression and cell death by apoptosis, which is a major determinant of growth in normal tissues and tumors.30 We speculate that PDT by the application of ALA and irradiation by visible light might induce similar processes on the cellular and cytokine levels within the tissue.
In conclusion, our study indicates that the long-term results of topical ALA-PDT with polychromatic light on superficial BCC and SCC are unsatisfactory at present. To improve ALA-PDT, other means of applying ALA, other vehicles for ALA, or both, supportive iontophoresis, and the use of chemical porphyrin inducers need to be studied. A recent study31 indeed suggests that the application of desferrioxamine mesylate, an agent for forming iron complexes, may optimize ALA-PDT of skin cancers. Another suggestion for achieving better penetration is the use of curettage before PDT.22 Ongoing studies, for instance, show that intralesional application of ALA in aqueous solution 4 hours before exposure to light can lead to complete eradication of nonmelanoma skin cancer and a long-term disease-free state.32 Furthermore, future studies will have to determine by direct comparison whether the use of monochromatic laser light is superior to that of polychromatic light in ALA-PDT.
Accepted for publication March 6, 1998.
Corresponding author: Peter Wolf, MD, Department of Dermatology, University of Graz, Auenbruggerplatz 8, A-8036 Graz, Austria (e-mail: peter.wolf @kfunigraz.ac.at).