Recurrence rate of superficial basal cell carcinoma following successful treatment with imiquimod 5% cream: interim 2-year results from an ongoing 5-year follow-up study in Europe

ARTICLE

Auteur(s) : Harald Gollnick¹, Carlos Guillén Barona², Ronald GJ Frank³, Thomas Ruzicka⁴, Mosaad Megahed⁴, Veronica Tebbs⁵, Mary Owens⁶, Patti Stampone⁶

¹Universitätsklinik für Dermatologie und Venerologie, Otto-von-Guericke-Universität Magdeburg,Leipziger Str. 44, 39120 Magdeburg, Germany
²Instituto Valenciano de Oncología, Profesor Beltrán Báguena, 19, 46009-Valencia, Spain
³Medisch Spectrum Twente, Ariensplein 1, 7511 JX, Enschede, Netherlands
⁴Universitätshautklinik Düsseldorf, Moorenstr.5, 40225 Düsseldorf, Germany
⁵3M Healthcare Limited, 3M House, Morley Street, Loughborough, LE11 1EP, United Kingdom
⁶3M Pharmaceuticals, 3M Center Building 275-02W-14, St. Paul, Minnesota, 55144-1000 USA

accepté le 29 Juin 2005

Methods

Subjects ≥ 18 years old were eligible for study participation if they had 1 primary, noninfected, histologically confirmed sBCC located on the limbs, trunk (excluding the anogenital area), neck, or head (excluding areas within 1 cm of the hairline, eyes, ears, nose, or mouth). At screening, prebiopsy lesions were to have a minimum area of 0.5 cm² and a maximum diameter of 2.0 cm; be clinically and histologically consistent with sBCC without histological evidence of aggressive growth patterns (e.g., severe squamous metaplasia, morpheaform, basosquamous features, infiltrative/desmoplastic features); and have a maximum tumor depth ≤ 1 mm. Subjects were excluded if they had evidence of Gorlin syndrome, a metastatic tumor or a tumor with a high probability of metastatic spread, or had had a malignant tumor of the skin within the target tumor site and within 5 cm of the target site margins in all directions within the last 5 years. Subjects were also excluded if they had received immunomodulatory or immunosuppressive therapies, including systemic and topical steroids, 10 weeks prior to study treatment initiation.

Study intervals consisted of a screening visit; an initiation visit; a 6-week treatment period with interval visits at weeks 2, 4, and 6; and a 12-week post treatment period. Subjects with clinical clearance at the 12-week post treatment period continued in a 5-year follow-up period with interval visits at months 3, 6, 12 and annually thereafter up to year 5.

The goal for the screening period for this study was to enrol 160 subjects from 20 to 25 European study centers; each center was to enrol 8 subjects. At the screening visit, 3 mm punch biopsies were taken from up to 4 clinically identified sBCC tumors. Tissue samples were processed and reviewed independently by 2 expert dermatopathologists at separate dermatopathology laboratories according to predetermined histological criteria for sBCC, as previously defined above. After notification of eligibility based on histology, subjects who still met the inclusion criteria were enrolled and treatment was initiated for 1 selected target tumor. The tumor location, relative to anatomical and surface landmarks, was mapped using a template and photos to ensure that the target site could be accurately relocated for each assessment for up to 5 years. Dermatologists listed as the investigator or co-investigator who had a minimum of 3 years experience in diagnosing and treating non-melanoma skin cancer, evaluated the target tumor site at each assessment visit.

During the 6-week treatment period, subjects or a caregiver applied imiquimod 5% cream 5 ×/week (consecutive days), at approximately the same time of day by gently rubbing the cream on the tumor area and about 1 cm around it until the cream disappeared. Safety evaluations consisted of physical examinations, vital sign measurements, clinical laboratory tests (hematology, blood chemistry, and urinalysis), pregnancy tests (where indicated for women of reproductive potential), and use of concomitant medications. Local skin reactions (LSRs) of erythema, edema, induration, vesicles, erosion, ulceration, scaling/flaking, and scabbing/crusting were assessed by the investigator at the target tumor site and in the surrounding area at each clinic visit beginning with the initiation visit, and using a rating scale where 0 = none, 1 = mild, 2 = moderate, and 3 = severe. In addition, skin quality assessments were made at the target treatment site and surrounding area in order to assess cosmetic outcome.

Efficacy was evaluated by determining the non-recurrence rate, defined as the proportion of treated subjects who were clinically clear of BCC at the target tumor site at the 12- week post treatment visit (i.e., initial clearance rate) and remained clear at each time point during the 5-year follow-up period. The protocol specified that the Kaplan-Meier (KM) product-limit method would be used to determine the nonrecurrence rates and associated standard errors. A post-hoc analysis was also performed using the life table method. These life table estimates of non-recurrence rates were used to calculate the estimated proportion of subjects who will remain clear after a single course of imiquimod treatment (ie, the life table estimated rate of non-recurrence multiplied by the initial clearance rate).

Safety data were evaluated using the intent-to-treat (ITT) data set. The incidences for adverse events (percent of subjects reporting the adverse event at least once) were tabulated. In addition, application site reactions were summarized separately. The intensity of each of the 8 LSRs was summarized by study visits (including the follow-up visits) separately for assessments made by the investigator and the subject. In addition, for each LSR category, the most intense reaction over the course of the study was determined for each subject, and the frequency distributions of these scores were tabulated.

An analysis of the change in intensity for each of the skin quality assessments was performed to determine if there was a significant change in intensity observed during the post treatment and follow-up periods as compared to the baseline assessments. Changes in intensity ratings for each skin quality assessment as compared to baseline were analyzed using Wilcoxon signed rank tests.

Laboratory data were collected at the screening visit (baseline) and at the end of the treatment visit. Summary statistics of absolute data and change from baseline (end of treatment minus baseline) and shift tables of baseline vs. end of treatment values relative to the reference range were constructed for the laboratory data.

Results

Study population

All 182 enrolled subjects were white, of whom 48% had Fitzpatrick skin type II and 42% had skin type III (all Fitzpatrick skin types were eligible for this study) [13]. Subjects were between 21 and 89 years old, and the mean age was 65 years. At treatment initiation, the postbiopsy target tumor areas ranged from a minimum of 0.2 cm² to a maximum of 9.0 cm². The overall median tumor area was 1.0 cm². The majority of target tumors were on the trunk (n = 114, 62.6%) and were most commonly located on the upper posterior trunk (n = 55, 30.2%).

Efficacy outcomes

A total of 26 subjects discontinued from the study for reasons other than sBCC recurrence during long-term follow up (( figure 1 )). Fourteen subjects who had been deemed clinically clear at 12-weeks post treatment had clinical evidence of sBCC recurrence at the 24-month follow-up interval. Of these 14 subjects, histological results were available for 13 subjects, of whom 2 subjects had no histological evidence of sBCC (histopathological diagnosis for both: cicatricial fibrosis). Overall, the estimated proportion of imiquimod-treated subjects who will be clinically clear of their treated tumor after applying a single 6 week course of treatment at 5x/week shows a gradual decline from 87.3% at month 3 to 79.4% at month 24 (table 1). These non-recurrence estimates do not include the 2 subjects who were clinically positive for sBCC recurrence, but were subsequently found to be histologically negative.

Additional subgroup analyses were performed to explore any correlations between treatment response (ie, clinical clearance of sBCC) and each of following variables: rest periods, tumor location, and the intensity of the investigator-assessed LSRs. Of the 148 subjects who did not take a rest period, 130 (88%) were responders. With the exception of 1 subject, who rested for 15 days, all remaining subjects with a rest period had a clinical response at 12-weeks post treatment. There was no trend for differences in response based on tumor location. However, there was a trend for differences in response based on severity of LSRs. In general, clinical response rates increased as LSR severity increased – 73% to 86% response with no LSRs; 81% to 100% response with mild LSRs; and 90% to 100% response with moderate or severe LSRs.
Table 1 Initial clearance rate and nonrecurrence

^bEstimate of the probability that a subject is clinically clear at 12 weeks posttreatment and remains clear for the corresponding timepoint was calculated as: Clearance rate at 12 weeks posttreatment (0.896) multiplied by the life table nonrecurrence rate (not shown) for that timepoint.

Safety outcomes

At the time of this report, 29 subjects have reported 48 serious adverse events (SAEs), none of which were judged by the investigator to be related to the study drug. Of these, 4 resulted in death: 1 during the 12-week post-treatment period (aortic aneurysm), 1 during the 6-month follow-up period (myocardial infarction), and 2 during the 24-month follow-up period (1 due to myocardial infarction/ventricular fibrillation and 1 due to ovarian carcinoma, with a history of cystitis).

Thirty-four (19%) subjects took a rest period during treatment to allow LSRs to subside. The median number of doses missed due to a prescribed rest period was 5 (ranging from 1 dose to 15 doses). The first rest period occurred during week 1 of treatment and the latest rest period occurred during the final week (week 6) of treatment (with a median time to first rest period of 3 weeks). The mean number of doses applied during the 6-week treatment period was 28 of the 30 possible doses.

Most of the LSRs assessed by the investigator were mild to moderate in intensity. During the study, LSR intensity levels were the highest during the treatment period, but subsided to levels at or below baseline during the 12-week posttreatment and follow-up periods (figures 2 and 3).

At the 24-month follow-up visit, skin quality assessments indicated an increase from baseline in the intensity ratings of hypopigmentation, and a decrease from baseline in the intensity ratings of hyperpigmentation, degree of scarring, skin surface (rough/dry/scaly), mottled or irregular pigmentation, and atrophy at the target tumor site.

Laboratory results showed no abnormal findings that were drug related and no new safety concerns were noted. Findings from physical examinations and vital signs measurements were also consistent with the demographics of the study population.

Discussion

In this study, the safety and tolerability profile observed was consistent with that reported for other studies evaluating imiquimod 5% cream for the treatment of sBCC. The most frequently recorded treatment-emergent effects were the LSRs rated by the investigators and application site reactions reported by the subjects. Other adverse events reported during the treatment period were consistent for a patient population with a mean age of 65 (a range from 21 to 89 years of age) and active medical conditions.

The frequency and severity of LSRs observed in this study were greatest during the treatment period, and decreased during the post treatment period. No subject withdrew from this study due to LSRs. Rest periods proved useful in ensuring that treatment was tolerated, thereby reducing the intensity of local signs and symptoms, where necessary. Of the adverse events reported by the subject, application site reactions including itching and pain occurred most often; however, not at sufficient severity to cause the subject to withdraw from the study.

Imiquimod 5% cream is a topical treatment for sBCC that induces an immune response, including type 1 interferon locally at the lesion site, as well as cell-mediated immunity through T cells via Toll-like receptor 7 [15] and tumor cell apoptosis. Recent results detailed in the literature from several in vitro studies suggest that imiquimod enhances a pro-apoptotic state in squamous cell carcinoma and HaCaT (spontaneously immortalized human keratinocyte) cell lines [16], perforin in cytotoxic T lymphocytes [17], and activation of innate immunity via natural killer cells and adaptive immunity through stimulation of T lymphocytes [18]. Moreover, induction of apoptosis in tumor cells via the intracytoplasmic cytochrome C pathway is now reported [19]. An in vivo study in patients with BCC suggests that imiquimod enhances a pro-apoptotic state with a decrease in Bcl-2 and recruitment of CD4 and CD8 infiltrate into the treated BCCs [20]. The results of these studies imply that some of the effects of treatment, such as erythema, erosion, and ulceration, may be due to a directed immune response to the sBCC lesion, the mechanism of which is still being studied. Therefore, the local skin reactions commonly seen during treatment may prove to be useful in predicting which subjects will respond to imiquimod and have complete clearance of their sBCC. Absence of LSRs during imiquimod treatment may be indicative of either subject non-compliance with the therapy or a lack of antitumor response mechanisms.

With respect to skin quality assessments, the relative decrease in rough/dry/scaly skin surface and atrophy may be explained by the healing of the tumor. Any assessment of scarring is confounded by the requirement for all subjects to receive a pretreatment biopsy, which per se could lead to a scar or the presence of tumor, which in itself can be the product of a destructive process. As for changes in pigmentation, the overall visual impression of how dark skin appears is a complex phenomenon. Thus, the overall increase of hypopigmentation rating from the baseline assessment may be secondary to regenerated skin at the treated site that has not sustained UV damage, particularly when compared to the surrounding chronically sun-damaged skin.

At the 12-week post treatment visit, the initial clearance rate (the proportion of subjects without clinical evidence of sBCC at the target tumor site) was 90%, which is appreciably higher than the 81% histological clearance rates reported in the phase II studies or the 75% composite (clinical and histological) clearance rate reported by Geisse et al. [21] in two US phase III studies for the same dosing regimen. This apparent difference at the 12-week posttreatment visit in efficacy between these earlier phase II and phase III studies may be due to the fact that in this long-term follow-up study the clinician’s assessment of efficacy at that time-point may have been unconsciously biased towards success. This may have been a result of the clinician’s knowledge of further opportunity to follow up the patient at subsequent time-points so that even patients who were judged to be clinically clear at the 12-week post treatment would be re-examined to ensure proper treatment for the patient. These initial clearance assessments were performed at 12-weeks post treatment to minimize LSR interference with the clinical judgment of sBCC presence or absence; only those subjects who initially cleared entered the follow-up period.

Both the Kaplan-Meier method and life table method were used to estimate non-recurrence rates for these interim data. Non-recurrence rate estimates from the Kaplan-Meier method were virtually identical to the life table method (data not shown) for all follow-up intervals except the current 24-month follow-up interval (Kaplan-Meier estimate of 86.7% vs. life table estimate of 88.7%). This difference at 24 months is due to late events (i.e., censoring due to the timing of subject visits) near the end of the 24-month interval [14]. Because the life table method provides estimates when event times (i.e., time of sBCC recurrence) may not be precisely measured, but are known to have occurred within an interval (i.e., it uses the interval these events occur and not the date of the actual event), it is less sensitive to censoring artifacts. Thus, the life table method provides a more robust estimate of nonrecurrence rates for these 24-month interim data. The life table estimates were then used to calculate the overall proportion of treated subjects who remained clear after a single course of imiquimod treatment for each follow-up visit. Clinicians considering imiquimod treatment can use these overall proportion estimates to assess the recurrence risk over time.

Rowe et al. have performed a meta-analysis using the available literature on the short-term (follow-up of < 5 years) and long-term (5 years) recurrence rates for BCC with the following potential treatment choices, including Mohs micrographic surgery, surgical excision, curettage and electrodesiccation, radiation therapy, and cryotherapy [22]. As with any meta-analysis, the findings should be considered with caution due to the known issues with this method of determining results. Some of these include: selection bias, publication bias (negative results tend not to be published), reliance on data that may come from unreliable sources (often extracted from case histories, etc), the literature may report results for mixed populations without stating so, and drawing conclusions based on summary data that may not contain sufficient detail to feel confident in the outcomes. That said, Rowe’s meta-analysis is of high quality and the authors address many of these weaknesses and the impact on their conclusions. Based on Rowe’s meta-analysis, the short-term (< 5 years) and long-term (> 5 years) recurrence rates for non-Mohs modalities were 4.2% and 8.7%, respectively. Of these non-Mohs modalities, Rowe et al. reported short-term recurrence rates of 5.3% and 3.7% for radiotherapy and cryotherapy, respectively. These short-term rates reported by Rowe et al. are lower than the current short-term recurrence rate of 7.7% (14/182) seen in this study (data through 2 years of follow-up). Thissen et al. [23] and more recently Bath-Hextall et al. [24] have also reviewed the literature to assess various treatment modalities for BCC, and both agree that surgery appears to be the most effective treatment for BCC with low recurrence rates. For nonsurgical modalities, Bath-Hextall’s review of randomized controlled trials report short-term (1 year) recurrence rates of 6% to 39% for cryotherapy and 5% to 18% for photodynamic therapy. Thissen et al. note that few published studies contain sufficient data to calculate cumulative 5-year recurrence rates using the Kaplan-Meier or life table method, and Bath-Hextall et al. note that few studies used an ITT analysis. The study presented here uses both the survival curve and ITT approach. As discussed earlier, direct comparisons between meta-analyses, with their recognized limitations and clinical studies, prospectively conducted in a highly controlled manner and analyzed using an ITT approach, may potentially be misleading.

Subjects enrolled in this study will be required to return to the study center for reassessment of the target site for clinical evidence of sBCC for 7 visits during the 5-year follow-up period. If clinical evidence of sBCC is found at the target site, the subject will be treated according to local standard medical practice. This reliance on clinical assessment is appropriate as physicians prescribing imiquimod 5% cream for sBCC in clinical practice would not excise the target tumor site following treatment unless there was clinical evidence of residual tumor. This is consistent with standard practice following other therapies such as cryotherapy, curettage and electrodesiccation, interferon, radiotherapy, photodynamic therapy, topical retinoids, or systemic retinoids. Clinical evaluation is the standard follow up for these methods as well as for excision techniques.

Imiquimod is a non-invasive immunobiological approach to treating sBCC and may have an advantage over other current procedures mentioned above. Initial efficacy rates support the clinical utility of imiquimod and the rate of recurrence seen to date supports ongoing follow up of any patient treated with imiquimod 5% cream.

Acknowledgements

H. Gollnick, Universitätsklinik für Dermatologie und Venerologie/Med. Fakultät, Otto-von-Guericke-Universität Magdeburg, Magdeburg, Germany; M Hagedorn, Städtische Klinik Darmstadt, Abteilung Eberstadt, Hautklinik, Darmstadt, Germany; L Kowalzick, Vogtland-Klinikum Plauen GmbH, Hautklinik, Plauen, Germany; T Ruzicka, Medizinische Einrichtungen der Heinrich-Heine-Universität Düsseldorf, Hautklinik des Operativen Zentrums III, Düsseldorf, Germany; HJ Schulze, Fachklinik Hornheide, Münster, Germany; U Haustein, Universitätsklinikum Leipzig, Klinik und Poliklinik für Hautkrankheiten, Leipzig, Germany; RGJ Frank, Medisch Spectrum Twente, Department of Dermatology, Enschede, Netherlands; WR Faber, Algemeen Christelijk Ziekenhuis Eemland, Amersfoort, Netherlands; M Koedam, Algemeen Christelijk Ziekenhuis Eemland, Amersfoort, Netherlands; O Christensen, Ullevål Hospital, Department of Dermatology, Oslo, Norway; P Helsing, Hudavdelingen Rikshospitalet, Oslo, Norway; P Lidbrink, Huddinge Universitetssjukhus, Hudkliniken, Stockholm, Sweden; A Wennberg, Sahlgrenska Universitetssjukhus, Hudkliniken, Göteborg, Sweden; C Guillén Barona, Department of Dermatology, Instituto Valenciano de Oncología, Profesor Beltrán Báguena, Valencia, Spain; I Febrer, Department of Dermatology Hospital, General UniversitarioAvda. Tres Cruces, Valencia, Spain; ML Carrelero, Department of Dermatology, Hospital Clinic i Provincial, Barcelona, Spain; JF Iniesta, Department of Dermatology, Hospital Virgen de la Arixaca, El Palmar, Spain; MC Jiménez, Department of Dermatology, Hospital La Paz, Paseo de la Castellana, Madrid, Spain; JL Lopez Estebaranz,, Servicio Dermatología consultas externas, Fundación Hospital Alcorcón, Alcorcón-Madrid, Spain; D Gould, The Dermatology Unit, Treliske Hospital, Cornwall, UK, England; C Griffiths, Hope Hospital, Salford, UK, England; B Cribier, Clinique de Dermatologie, Strasbourg, France; JJ Grob, Service de Dermatologie, Hôpital Ste Marguerite, Marseille, France; JP Ortonne, CHU de Nice/Hôpital de l’Archet 2, Consultation de dermatologie, Nice, France; D Lambert, Service de Dermatologie, Hopital Universitaire Bocage, Dijon, France; P Celerier, Service de Dermatologie, Centre Hospitaler du Mans, Le Mans, France.

Expert dermatopathology review was conducted by:

M Megahed, Cytopathology Laboratory, Universitätshautklinik Düsseldorf, Düsseldorf, Germany; W Sterry, Universitätsklinikum Charité, Medizinische Fakultät der Humboldt-Universität, Berlin, Germany

We thank Ian Haynes for invaluable medical advice and study oversight; David Alton, Marion Carey-Yard, Jo Gilbert, Anne Grethe Tandsether, Jeanne Lacallion, Charo Lozano, Frederic Marmion, and Sandra van der Poel for study conduct; Jani Birnstengel, Deanne Lacerenza, and Carol Pedersen for data management; Ron Hawkinson and Shelley-Ann Walters for statistical analysis support; Cathryn Lloyd for clinical project management; and Anne Roush for manuscript preparation.

References

2 Preston DS, Stern RS. Nonmelanoma cancers of the skin. N Engl J Med 1992; 327: 1649-62.

3 MacKie RM. Basal cell Carcinoma Section of Epidermal skin tumours. In: Champion RH, Burton JL, Burns DA, Breathnach SM, eds. Textbook of Dermatology, Vol 2, 6th ed. Oxford UK: Blackwell Science Ltd, 1998: 1679-84.

4 Slade HB, Owens ML, Tomai MA, Miller RL. Imiquimod 5% Cream (Aldara™). Exp Opin Invest Drugs 1998; 7: 437-49.

5 Testerman TL, Gerster JF, Imbertson LM, et al. Cytokine induction by the immunomodulators imiquimod and S-27609. J Leuk Biol 1995; 58: 365-72.

6 Cornell RC, Greenway HT, Tucker SB, et al. Intralesional interferon therapy for basal cell carcinoma. J Am Acad Dermatol 1990; 23: 694-700.

7 Toma S, Vincenti M, Palumbo R, et al. Results of the association of intralesional recombinant alpha interferon-2a (α-IFN) plus 13-cisretinoic acid (13cRA) in the treatment of basal cell carcinomas (BCC) of the skin. Int J Oncol 1993; 3: 1149-54.

8 Greenway HT, Cornell RC, Tanner DJ, Peets E, Bordin GM, Nagi C. Treatment of basal cell carcinoma with interlesional interferon. J Am Acad Dermatol 1986; 15: 437-43.

9 Marks R, Gebauer K, Shumack S, et al. Imiquimod 5ream in the treatment of superficial basal cell carcinoma: Results of a multicenter 6-week dose-response trial. J Am Acad Dermatol 2001; 44: 807-13.

10 Geisse JK, Rich P, Pandya A, et al. Imiquimod 5ream for the treatment of superficial basal cell carcinoma: A double-blind, randomized, vehicle-controlled study. J Am Acad Dermatol 2002; 47: 390-8.

11 Sterry W, Ruzicka T, Herrera E, et al. Imiquimod 5ream for the treatment of superficial and nodular basal cell carcinoma: randomized studies comparing lowfrequency dosing with and without occlusion. Br J Derm 2002; 147: 1227-36.

12 Lang PG, Maize JC. Basal Cell Carcinoma. Chapter 4. In: Friedman RJ, Rigel D, Harris MN, Baler D, eds. Cancer of the Skin. Philadelphia, PA: WB Saunders Company, 1991: 35-73.

13 Fitzpatrick TB. The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol 1988; 124: 869-71.

14 Cantor AB, Shuster JJ. Parametric vs. non-parametric methods for estimating cure rates based on censored survival data. Stat Med 1992; 11: 931-7.

15 Hemmi H, Kaisho T, Takeuchi O, Sato S, et al. Small anti-viral compounds activate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat Immunol 2002; 3: 196-200.

16 Schön M, Bong AB, Drewniok C, et al. Tumorselective induction of apoptosis and the small-molecule immune response modifier imiquimod. J Nat Cancer Inst 2003; 95: 1138-49.

17 Ambach A, Bonnekoh B, Nguyen M, Schon MP, Gollnick H. Imiquimod, a Toll-like receptor-7 agonist, induces perforin in cytotoxic T lymphocytes in vitro. Mol Immunol 2004; 40: 1307-14.

18 Gaspari AA, Sauder DN. Immunotherapy of basal cell carcinoma: evolving approaches. Dermatol Surg 2003; 29: 1027-34.

19 Schön MP, Schön M. Immune modulation and apoptosis induction: two sides of the antitumoral activity of imiquimod. Apoptosis 2004; 9: 291-8.

20 Urosevic M, Maier T, Benninghoff B, Slade H, Burg G, Dummer R. Mechanisms underlying imiquimod-induced regression of basal cell carcinoma in vivo. Arch Dermatol 2003; 139: 1325-32.

21 Geisse J, Caro I, Lindholm J, et al. Imiquimod 5ream for the treatment of superficial basal cell carcinoma: results from two phase III, randomized, vehicle controlled studies. J Am Acad Dermatol 2004; 50: 722-33.

22 Rowe DE, Carroll RJ, Day CL. Long-term recurrence rates in previously untreated (primary) basal cell carcinoma: implications for patient follow-up. J Dermatol Surg Oncol 1989; 15: 315-28.

23 Thissen M, Neumann M, Schouten LJ. A systematic review of treatment modalities for primary basal cell carcinomas. Arch Dermatol 1999; 135: 1177-83.

24 Bath-Hextall F, Bong J, Perkins W, et al. Interventions for basal cell carcinoma of the skin: systemic review. BMJ 2004; 329(7468): 705.