Photodynamic Therapy of Multiple Nonmelanoma Skin Cancers With Verteporfinand Red Light–Emitting Diodes: Two-Year Results Evaluating Tumor Response and Cosmetic Outcomes

Background  Efficient treatment of patients with multiple synchronous nonmelanomaskin cancers represents a therapeutic challenge.

Objective  To study the safety and efficacy of photodynamic therapy (PDT) withverteporfin and red light in the treatment of multiple nonmelanoma skin cancers.

Design  Open-label, randomized, multicenter, dose-ranging phase 2 study conductedat 4 North American university-based dermatology clinics.

Patients  Fifty-four patients with 421 multiple nonmelanoma skin cancers includingsuperficial and nodular basal cell carcinoma and squamous cell carcinoma insitu (Bowen disease).

Methods  A single intravenous infusion of 14 mg/m² of verteporfinfollowed 1 to 3 hours later by exposure of tumors to 60, 120, or 180 J/cm² of red light (688 ± 10 nm) from a light-emitting diode panel.

Main Outcome Measures  Pathologic response of treated sites was assessed at 6 months. Clinicaland cosmetic responses were assessed and graded at 6 weeks, 3 months, and6 months after verteporfin PDT, with optional follow-up visits at 12, 18,and 24 months.

Results  The histopathologic response, defined as absence of tumor on biopsyspecimens 6 months after verteporfin PDT, ranged from 69% at 60 J/cm² to 93% at 180 J/cm². At 24 months of follow-up (276 tumorsin 31 patients), the clinical complete response rate ranged from 51% at 60J/cm² to 95% at 180 J/cm². No significant systemic adverseevents were observed; most events occurred at the treated tumor sites andincluded events such as pain. Overall, 65% (95% confidence interval, 58%-71%)of tumors were judged to have good to excellent cosmesis at 24 months.

Conclusion  A single course of verteporfin PDT showed treatment benefit for patientswith multiple nonmelanoma skin cancers.

Nonmelanoma skin cancers (NMSCs) are the most common carcinomas affectingmen and women, with about 1.3 million cases reported annually in the UnitedStates.¹ Basal cell carcinoma (BCC) accountsfor approximately 75% of NMSC and 25% of all cancers diagnosed in the UnitedStates, whereas squamous cell carcinoma (SCC) accounts for 20% of NMSC.²

Photodynamic therapy (PDT) is a 2-step process consisting of administrationof a photosensitizer followed by light application.³ Verteporfin(benzoporphyrin derivative monoacid A ring), is a semisynthetic, second-generationporphyrin derivative that can be activated by red light (peak absorption,689 nm). This wavelength permits activation of the drug within the deep portionsof tumors. For the treatment of NMSC with verteporfin PDT, light-emittingdiodes (LEDs) can be used as the red light source.

The present study examines pathologic, clinical, and cosmetic outcomesfor patients with multiple NMSCs who underwent verteporfin PDT using LED light.The cosmetic assessments performed in this study are unique in that they wereprospectively defined, semiquantitative, and conducted periodically for upto 2 years after treatment.

Patients with at least 2 biopsy-proven nonpigmented NMSCs (either superficialor nodular BCC or SCC in situ [Bowen disease]) were recruited for this study(Table 1). All patients gave writteninformed consent, and approval from each participating institution's ethicscommittee or institutional review board was granted before the beginning ofthe study.

Patients were randomly assigned to receive 1 of 3 different light doses:60, 120, or 180 J/cm². If a second treatment was required, thepatient's light dose remained the same as that given in the first course.To avoid a marked difference in the number of tumors within each treatmentgroup, randomization was stratified by the number of tumors (2-5 tumors, 6-10tumors, or >10 tumors) and by center.

Patients received a 10-minute intravenous infusion of 14 mg/m² of verteporfin (QLT Inc, Vancouver, British Columbia) followed 1 to3 hours later by exposure of tumors to 60, 120, or 180 J/cm² ofred light (688 ± 10 nm) from a nonthermal LED panel. Up to 20 tumorswere exposed to light in 1 session. The exposed area included a peritumoralmargin of 3 to 4 mm. Tumors that could not be treated within the 1- to 3-hourpostinfusion time window could receive other standard therapy at the discretionof the investigator. These tumors were not included as part of this study.

Tumors could be retreated at 3 months after the initial treatment ifthe investigator judged that the tumor had not achieved a clinical completeresponse (CCR). For retreatment, the verteporfin dose was increased to 18mg/m², but the patient's light dose remained the same as that usedin the first course.

The LED panels consisted of a monolithic array of hybrid gallium aluminumarsenide LEDs that emitted diffused red light (Quantum Devices Inc, Barneveld,Wis). A photograph of the LED unit and treatment head is provided in Figure 1. The LED panel had a central wavelengthof 688 ± 10 nm, with a full-width half-maximum bandwidth of approximately25 nm. The emission spectrum of the LED unit compared with the absorptionspectrum of verteporfin is illustrated in Figure 2. The irradiance delivered to the skin by the LED unit was200 ± 40 mW/cm². The LED output was checked by a radiometer(model UDT268M; UDT Instruments, Baltimore, Md) before and after the tumorswere exposed to light.

The primary efficacy variable was the histopathologic response (HPR)rate, defined as the percentage of treated tumor sites with no residual tumoras assessed by a histopathologic review of a representative 2-mm punch biopsyspecimen taken from within the PDT treatment site 6 months after the firstverteporfin PDT course. Retreated tumors were included in the calculationof the HPR rate. Clinical response of the treated sites was assessed at 6weeks and 3, 6, 12, 18, and 24 months after verteporfin PDT. Follow-up beyond6 months was optional. The CCR rate was defined as the percentage of treatedsites that were judged by the investigator to be tumor free on clinical examination.

At each follow-up visit, the color, profile, and texture of each treatedsite (ie, tumor plus peritumoral margin) were assessed by the investigatorand assigned a numeric value depending on the answers to following questions:(1) Are you satisfied with how the color of the scar matches the color ofthe surrounding skin? (2) Are you satisfied that the scar surface is flushwith the surrounding skin? (That is, are you satisfied that the scar surfaceis not too elevated or depressed with respect to the surrounding tissue?)(3) Are you satisfied with how the surface texture of the scar blends withthe surrounding skin? A yes answer was scored 3 points; neutral, 2 points; and no, 1 point.For each treated site, the points of all 3 responses were tallied, and thecosmetic outcome was judged to be excellent (a score of 9), good (7 or 8),satisfactory (5 or 6), or poor (3 or 4).

Investigators also rated the cosmetic outcome of each site treated withverteporfin PDT as superior to, equivalent to, or worse than the cosmeticoutcome expected from a standard treatment such as surgical excision or electrodesiccationand curettage. This subjective global assessment was based on the investigators'prior general clinical experiences with cosmetic outcomes for standard treatmentof lesions of equivalent location and size. Patients who had received priorstandard therapy for skin tumors were also asked to rate the cosmetic outcomeof each of their verteporfin-treated sites as superior to, equivalent to,or worse than the cosmetic outcome obtained with a prior standard therapy.Only sites with a CCR were included in this analysis.

All statistical tests and confidence intervals (CIs) were 2-sided withan α of .05. The SAS statistical package (version 8.2) was used forthe analyses (SAS Institute Inc, Cary, NC). The 95% CIs were calculated basedon the binomial distribution. In addition to the overall response for alltumors combined, the HPR and CCR rates were also determined according to tumortype (nodular BCC, superficial BCC, or SCC), size (<1 cm, 1-2 cm, >2 cm),and location (head and neck, upper extremities, lower extremities, or trunk).

Only observed tumor data were included in the analyses. In addition,the last observation was carried forward for tumors that had been nonrespondersat the most recent visit, and values of nonresponders were imputed for tumorsthat were not assessed because they had received alternate therapy.

To adjust for correlation among tumor outcomes within a patient, weused the generalized estimating equation (GEE) method^4,5 witha robust covariance and an exchangeable working correlation structure to explorethe effect of light dose on tumor pathologic response at month 6, clinicalresponse at month 24, and investigator assessment of cosmetic response (excellentor good vs satisfactory or poor) at months 6 and 24. In addition, the effectsof age, sex, and tumor size, type, and location were evaluated in the GEEmodel. The Spearman rank correlation coefficient⁶ wasused to evaluate correlation between investigator and patient assessmentsof cosmetic outcome compared with previous standard therapy.

Fifty-four patients with 421 tumors received at least 1 course of verteporfinPDT (Table 1). The average numberof treated tumors per patient was 7.8 (range, 2-20). Patients ranged in agefrom 22 to 79 years, with a mean age of 55 years, and most had Fitzpatrickskin phototype II or III. Most tumors (92%) were BCC, and 15 patients hadnevoid BCC syndrome. Six patients (31 tumors), all of whom were treated witha light dose of 60 J/cm², received a second treatment with 18 mg/m² of verteporfin and the same light dose. Seven patients (51 tumors)withdrew from the study prior to month 6: 2 were lost to follow-up; 4 withdrewfor unspecified reasons; and 1 requested withdrawal owing to the inconvenienceof travel and treatment site pain requiring the use of codeine. Although thefollow-up visits after month 6 were optional, 31 patients (57%) with 276 tumors(66%) were available for a full 2 years of follow-up. The ages, sex distribution,and number and types of tumors were not different between those patients whocompleted 24 months of follow-up and those who were observed for only 6 months.

There was a dose-response relationship of HPR to verteporfin PDT, withHPR rates of 69% (95% CI, 61%-76%), 79% (95% CI, 70%-86%), and 93% (95% CI,86%-97%) noted for tumors receiving 60, 120, and 180 J/cm², respectively(n = 378; nonresponse carried forward for 8 tumors). The HPR rates accordingto tumor type, size, and location are listed in Table 2. The dose response relationship was confirmed by GEE modeling.Although the overall effect of light dose was not significant (P = .06), there was a trend indicating that the higher HPR was associatedwith a higher light dose. There was no significant difference in HPR ratebetween age, sex, tumor size, tumor location, and tumor type.

Treatment sites judged to be tumor free based on clinical examinationwere recorded as CCR. The CCR rate increased with the light dose. At 6 monthsafter verteporfin PDT, the CCR rates were 78% (95% CI, 71%-84%), 88% (95%CI, 81%-94%), and 98% (95% CI, 93%-100%) for tumors treated with 60, 120,and 180 J/cm², respectively (n = 378); 24 months after verteporfinPDT, the responses were 51% (95% CI, 42%-61%), 79% (95% CI, 67%-88%), and95% (95% CI, 89%-99%), respectively (n = 276) (Figure 3). At 24 months, the GEE model showed significant differencesin CCR between the light doses (180 vs 120 J/cm², P = .02; 180 vs 60 J/cm², P<.001;120 vs 60 J/cm², P = .05). Other factorssuch as age, sex, tumor size, tumor location, and tumor type were not predictiveof clinical response. In addition, tumors responded similarly in patientswith and without nevoid BCC syndrome.

The cosmetic outcomes of treated tumor sites as assessed by the investigatorsat month 24 (categorized as excellent, good, satisfactory, or poor) are presentedin Figure 4. The best cosmetic outcomeswere observed in tumors treated with 60 J/cm². The cosmetic outcomeof the tumors in the 180 J/cm² group improved markedly over time(data not shown). The GEE modeling showed significant differences in the cosmeticoutcome rate between certain light doses only at month 6 (180 vs 60 J/cm², P = .001; 120 vs 60 J/cm², P = .05) and not at month 24. Smaller tumors showed bettercosmesis at month 6 (P = .01) and month 24 (P = .02). At month 24, head and neck tumors showed significantlybetter cosmesis than tumors in other locations (P<.01for all). An example of an excellent cosmetic outcome is shown in Figure 5.

Investigators and patients were asked to compare the results of verteporfinPDT with prior standard treatments. At month 24, patients assessed the cosmeticoutcome of 58% of treated tumors as superior to prior therapy (Figure 6). The percentage of outcomes judged by investigators andpatients to be superior to prior therapy was highest at the 60 J/cm² light dose and lowest at the 180 J/cm² light dose. Despiteslight differences in the numbers of sites assessed, investigator assessmentsof cosmetic outcome were in agreement with patient assessments at month 6and month 24 (P<.001).⁶

Adverse events associated with treatment occurred in 48 (89%) of 54patients. In general, verteporfin PDT was well tolerated with no major systemicadverse events. The most common systemic associated adverse events were headache(5 patients, 9%) and photophobia (2 patients, 4%). Most of the associatedadverse events occurred at the treatment site.

Discomfort during light treatment was described variously as warmth,pain, burning/stinging, prickling, and pruritus, which in turn were collectivelydefined as a specific safety variable called the smartingresponse. To evaluate the smarting response, investigators interviewedthe patients during light application to assess the maximum degree of discomforton a scale of 1 (barely noticeable) to 5 (severe, requiring interruption oflight exposure). Overall, warmth and a burning/stinging sensation were themost commonly reported symptoms of discomfort during light exposure of thetumors, reported for 44% and 41% of the treated sites, respectively (n = 421).Severe treatment site smarting responses during light application were reportedby 9 patients (17%) overall with no clear-cut dose-response effect (Table 3). Smarting response symptoms werealleviated by interrupting light treatment briefly, applying ice packs locally,or injecting lidocaine into the site (4% of sites received lidocaine).

Treatment site adverse events occurring after verteporfin PDT were reportedby 48 (89%) of 54 patients (Table 3)and included pain (n = 39; 72%), pruritus (n = 13; 24%), burning (n = 12;22%), edema (n = 11; 20%), prickling (n = 9; 17%), erythema (n = 8; 15%),skin hypertrophy (n = 4; 7%), exudate (n = 3; 6%), hyperesthesia (n = 3; 6%),and stinging (n = 3; 6%). In most patients, these events were mild to moderatein intensity.

Twenty patients (37%) used oral analgesic preparations containing opiates(acetaminophen plus codeine) to control treatment site pain occurring afterverteporfin PDT (Table 3). Themedian duration of pain (Table 3)was similar for the 3 groups (9 days for 60 J/cm²; 12 days for120 J/cm²; and 14 days for 180 J/cm²). The number ofsevere pain events was higher at the 180 J/cm² light dose thanat the other 2 light doses.

Successful treatment of multiple NMSCs has been previously reportedwith a number of systemic photosensitizers, including porfimer sodium, hematoporphyrinderivative, tin ethyl etiopurpurin, and metatetrahydroxyphenylchlorin.^7-12 Oneof the problems associated with many of these photosensitizers is prolongedgeneralized cutaneous photosensitivity that can last for several months.¹³

Photodynamic therapy with topical photosensitizers such as aminolevulinicacid (ALA) has also been used for the treatment of NMSC with the advantageof inducing photosensitivity only at the application site. Good complete responserates have been reported for the treatment of superficial BCC and Bowen disease;however, the treatment of nodular BCC with ALA PDT has been associated withhigher recurrence rates despite the apparently favorable early responses basedon clinical assessment.^10,14-20 Thishigher recurrence rate might be explained by limited penetration of ALA PDTto deeper areas of the tumor.²¹ Topical methylesterALA has also been approved for use in Europe. Gentle curettage is performedbefore methylester ALA application to facilitate penetration of the photosensitizerprecursor.^20,22,23

Verteporfin, marketed with the brand name Visudyne (Novartis AG, Basel,Switzerland), is now approved in many countries for the treatment of a numberof ophthalmologic indications such as age-related macular degeneration. Oneof the advantages of verteporfin over other systemic photosensitizers is itsshort photosensitivity period of only a few days.²⁴ Inaddition, verteporfin can be activated with 688-nm red light, which penetratesdeeper into tumors than the 630- to 635-nm light used to activate porfimersodium and protoporphyrin IX, the endogenous photosensitizer generated byALA. The LED device used in this clinical study has a number of advantagesover other devices such as lasers and filtered broadband lamps for dermatologicPDT: (1) it is simple to use; (2) it does not have special electrical requirements;(3) it is solid-state and therefore reliable; and (4) it emits a narrow wavelengthrange.

The exact mechanism of action of verteporfin PDT in the treatment ofNMSC is not known. Light-activated verteporfin has been shown to induce cytochrome c release from mitochondria followed by caspase 3, 6, 7,and 8 activation culminating in apoptosis.²⁵ Theoccurrence of apoptosis following verteporfin PDT appears to be a predominantcytotoxic mechanism in vitro, but its relevance to clinical response followingtreatment of tumors in vivo has not been assessed. In rabbits, endothelialcell and pericyte damage was observed on biopsy specimens taken immediatelyafter verteporfin PDT when exposure to 690-nm red light took place 1 to 5hours after verteporfin administration.²⁶ At2 to 24 hours after verteporfin PDT, no epidermal phototoxic reaction wasobserved, but blood stasis, edema, and dermal-epidermal separation were present.These observations suggest that PDT at 1 to 5 hours after verteporfin administrationmight target the vascular compartment.

In the present study, the response rate of verteporfin PDT was highestat 180 J/cm² with a 93% histopathologic response. The analysisof CCR at month 6 and month 24 shows that at the 2 higher light doses, theclinical response was maintained over time. At the highest light dose, theCCR was 98% at month 6 and 95% at month 24, which compares favorably withstandard therapies at 24 months. At 60 J/cm², the CCR decreasedfrom 78% at 6 months to 51% at 24 months, suggesting that this lower lightdose did not result in the same sustained CCR achieved at the 2 higher lightdoses. The post-PDT biopsy specimens were limited to 2 mm to allow us to moreeasily observe the PDT cosmetic outcomes over 24 months, but this could potentiallyhave introduced a treatment site sampling error that overestimated the HPRrate.

Cosmetic outcomes were generally favorable. There was a significantinverse relationship between light dose and cosmesis at the 6-month assessment.Since the patients in this study had an average of 8 skin tumors, and somehad been diagnosed with nevoid BCC syndrome (which can lead to hundreds oftumors in a lifetime), the comparison of cosmesis between verteporfin PDTand prior therapy, although subjective, is relevant. The cosmetic outcomesof verteporfin PDT at 24 months were favorable, and more than 80% of treatedsites were considered by both patients and investigators to have a cosmeticoutcome equivalent or superior to previous standard therapies for NMSC. Notunexpectedly, small tumors were associated with more favorable cosmetic outcomes,as were head and neck tumors.

Other than headache in 5 patients, no significant pattern of systemicadverse events was observed in association with the study treatment. The smartingsensations of pain, burning/stinging, or warmth that patients reported atthe treatment sites during light application were easily alleviated with icepacks or local anesthetic. These symptoms promptly ceased on termination oflight exposure. Pain at the treatment sites after verteporfin PDT was reportedby most patients and was generally well controlled with oral analgesics. However,severe posttreatment pain was experienced by 35% of patients, and its frequencywas highest at the highest light dose. Patients with a higher number of tumorsdid not necessarily report more intense pain.

In conclusion, verteporfin PDT as a single treatment modality can inducegood response rates with favorable cosmesis. Further study in patients withmultiple NMSC is warranted.

Corresponding author and reprints: Harvey Lui, MD, Division of Dermatology,University of British Columbia, 835 W 10th Ave, Vancouver, British Columbia,Canada V5Z 4E8 (e-mail: hlui@interchange.ubc.ca).

Accepted for publication September 24, 2003.

This study was sponsored by QLT Inc, Vancouver, British Columbia.

This study was presented in part at the 11th Annual Meeting of the PhotomedicineSociety; February 21, 2002; New Orleans, La; and the International PhotodynamicAssociation Eighth World Congress of Photodynamic Medicine; June 8, 2001;Vancouver, British Columbia.

Writing and editing assistance was provided by Christy V. Costello,ELS, and Anne Fisher, MSc, of QLT Inc.