Long-term recurrence rates of actinic keratosis: A systematic review and pooled analysis of randomized controlled trials

Abstract

To the Editor: Actinic keratoses (AK) are common lesions in light-skinned individuals, which can progress to invasive cutaneous squamous cell carcinoma.1Criscione V.D. Weinstock M.A. Naylor M.F. et al.Actinic keratoses: natural history and risk of malignant transformation in the Veterans Affairs Topical tretinoin Chemoprevention Trial.Cancer. 2009; 115: 2523-2530Google Scholar Most treatment options are superior to placebo regarding lesion clearance.2Werner R.N. Stockfleth E. Connolly S.M. et al.Evidence- and consensus-based (S3) guidelines for the treatment of actinic keratosis - International League of Dermatological Societies in cooperation with the European dermatology forum - short version.J Eur Acad Dermatol Venereol. 2015; 29: 2069-2079Google Scholar The majority of studies focus on short-term clearance evaluated within 3 to 6 months after treatment, although AK are considered a chronic condition.2Werner R.N. Stockfleth E. Connolly S.M. et al.Evidence- and consensus-based (S3) guidelines for the treatment of actinic keratosis - International League of Dermatological Societies in cooperation with the European dermatology forum - short version.J Eur Acad Dermatol Venereol. 2015; 29: 2069-2079Google Scholar,3Heppt M.V. Leiter U. Steeb T. et al.S3 guideline for actinic keratosis and cutaneous squamous cell carcinoma - short version, part 1: diagnosis, interventions for actinic keratoses, care structures and quality-of-care indicators.J Dtsch Dermatol Ges. 2020; 18: 275-294Google Scholar Here, we systematically investigated the recurrence rates of AK reported from parallel-arm randomized controlled trials (RCT).The protocol for this study was registered online (PROSPERO CRD42018095903) and published previously.4Steeb T. Heppt M.V. Becker L. Kohl C. French L.E. Berking C. Long-term efficacy of interventions for actinic keratosis: protocol for a systematic review and network meta-analysis.Syst Rev. 2019; 8: 237Google Scholar Outcomes of interest were the clinically assessed participant- and lesion-specific recurrence rates reported ≥12 months after the end of treatment, i.e. the number of relapsing patients/lesions divided by the number of patients/lesions with complete initial clearance. Pooled proportions with 95% confidence intervals (CI) were calculated with the inverse variance method using the function “metaprop” of the R package “meta” (R Core Team, 2020). We used a random-effects model because clinical and methodologic heterogeneity between the studies was likely. Heterogeneity was quantified with the I2 statistic. The risk of bias was evaluated with the revised Cochrane risk-of-bias tool for RCTs (RoB version 2.0).5Sterne J.A.C. Savovic J. Page M.J. et al.RoB 2: a revised tool for assessing risk of bias in randomised trials.BMJ. 2019; 366: l4898Google ScholarThe systematic search in the databases (Medline, Embase, and PubMed Central) and trial registers (controlled-trials.com, clinicaltrials.gov, anzctr.org.au, who.int/trialsearch, and clinicaltrialsregister.eu) identified 2741 references, of which 37 underwent full-text review. Seven publications referring to 9 RCTs met the eligibility criteria. They showed different risks of bias (Table I). In the 9 RCTs, data for participant-specific recurrence rates were reported (Fig 1). Cryotherapy and photodynamic therapy with 5-aminolevulinate showed the lowest recurrence rates of 39%, respectively, with a 95% CI of 20% to 62% for cryotherapy and a 95% CI of 33% to 45% for photodynamic therapy with 5-aminolevulinate. They were followed by placebo (44%; 95% CI 25% to 65%), photodynamic therapy with methyl-aminolevulinate (45%; 95% CI 38% to 51%), and imiquimod 5% (45%; 95% CI 14% to 81%). The highest participant-specific recurrence rates were observed for diclofenac (85%; 95% CI 70% to 94%), followed by ablative laser treatment (54%; 95% CI 18% to 86%), and 5-fluorouracil 5% (52%; 95% CI 38% to 66%). Data on lesion-specific recurrence rates were available from 4 RCTs. They were lowest for placebo (15%; 95% CI 10% to 22%), followed by photodynamic therapy with 5-aminolevulinate (20%; 95% CI 16% to 24%), and photodynamic therapy with methyl-aminolevulinate (34%; 95% CI 23% to 47%).Table IBaseline characteristics of the included studies and participant-specific and lesion-specific recurrence rates for 9 RCTs from 7 publicationsDesignSample sizeIntervention/controlParticipant-specific recurrence rate (%, n)Lesion-specific recurrence rate (%, n)Risk of biasMulticenter, double-blinded study (ALA-AK-CT003)122ALA-PDT30.2% (16 of 53)16.7% (59 of 353)LowPlacebo60.0% (3 of 5)13.3% (6 of 45)Multicenter, investigator-blinded study (ALA-AK-CT002)571ALA-PDT41.7% (80 of 192)21.7% (249 of 1147)LowMAL-PDT44.6% (70 of 157)25.5% (244 of 958)Placebo23.1% (3 of 13)15.5% (13 of 84)Two multicenter, open-label phase IV studies (LEIDA 1/2 study)479Imiquimod 5%90.7% (215 of 237)NRLowDiclofenac82.6% (200 of 242)NRMulticenter, double-blinded study246Imiquimod 5%39.0% (23 of 59)NRHighPlacebo57.1% (8 of 14)NRSingle-center, open-label study75Imiquimod 5%13.6% (3 of 22)NRSome concerns5-Fluorouracil 5%43.5% (10 of 23)NRCryotherapy58.8% (10 of 17)NRSingle-center, open-label study555-Fluorouracil 5%60.0% (15 of 25)NRHighErbium:Yttrium Aluminium Garnet-mediated laser ablation25.9% (7 of 27)NRSingle-center, open-label study200Cryotherapy27.4% (20 of 73)33.2% (63 of 190)Some concernsCO2 laser ablation78.1% (50 of 64)63.0% (87 of 138)Single-center, open-label study200Diclofenac72.0% (20 of 73)79.1% (140 of 177)Some concernsMAL-PDT44.8% (30 of 67)43.5% (248 of 570)ALA, 5-aminolevulinate; MAL, methyl-aminolevulinate; n, number of patients; NR, not reported; PDT, photodynamic therapy. Open table in a new tab We conclude that long-term recurrence rates were surprisingly high, even after the initial clearance of lesions. In particular, the participant-specific recurrence rates ranged from 39% to 85%, indicating that the majority of patients will experience recurrent lesions in the long term. Although lesion-specific recurrence rates were lower, ranging from 15% to 34%, we infer that the majority of patients require regular surveillance and retreatments. The recurrence rates were similar for most active interventions and were not superior to placebo. However, the recurrence rates could be subject to a large increment if only a few lesions or patients show a relapse, because the number of the denominator was lower for the placebo arms than for the active interventions. Nevertheless, these data challenge the value of short-term clearance rates (which are the values that are usually reported), and they serve as primary endpoints in RCTs and highlight the nature of AK as a chronic health condition. To the Editor: Actinic keratoses (AK) are common lesions in light-skinned individuals, which can progress to invasive cutaneous squamous cell carcinoma.1Criscione V.D. Weinstock M.A. Naylor M.F. et al.Actinic keratoses: natural history and risk of malignant transformation in the Veterans Affairs Topical tretinoin Chemoprevention Trial.Cancer. 2009; 115: 2523-2530Google Scholar Most treatment options are superior to placebo regarding lesion clearance.2Werner R.N. Stockfleth E. Connolly S.M. et al.Evidence- and consensus-based (S3) guidelines for the treatment of actinic keratosis - International League of Dermatological Societies in cooperation with the European dermatology forum - short version.J Eur Acad Dermatol Venereol. 2015; 29: 2069-2079Google Scholar The majority of studies focus on short-term clearance evaluated within 3 to 6 months after treatment, although AK are considered a chronic condition.2Werner R.N. Stockfleth E. Connolly S.M. et al.Evidence- and consensus-based (S3) guidelines for the treatment of actinic keratosis - International League of Dermatological Societies in cooperation with the European dermatology forum - short version.J Eur Acad Dermatol Venereol. 2015; 29: 2069-2079Google Scholar,3Heppt M.V. Leiter U. Steeb T. et al.S3 guideline for actinic keratosis and cutaneous squamous cell carcinoma - short version, part 1: diagnosis, interventions for actinic keratoses, care structures and quality-of-care indicators.J Dtsch Dermatol Ges. 2020; 18: 275-294Google Scholar Here, we systematically investigated the recurrence rates of AK reported from parallel-arm randomized controlled trials (RCT). The protocol for this study was registered online (PROSPERO CRD42018095903) and published previously.4Steeb T. Heppt M.V. Becker L. Kohl C. French L.E. Berking C. Long-term efficacy of interventions for actinic keratosis: protocol for a systematic review and network meta-analysis.Syst Rev. 2019; 8: 237Google Scholar Outcomes of interest were the clinically assessed participant- and lesion-specific recurrence rates reported ≥12 months after the end of treatment, i.e. the number of relapsing patients/lesions divided by the number of patients/lesions with complete initial clearance. Pooled proportions with 95% confidence intervals (CI) were calculated with the inverse variance method using the function “metaprop” of the R package “meta” (R Core Team, 2020). We used a random-effects model because clinical and methodologic heterogeneity between the studies was likely. Heterogeneity was quantified with the I2 statistic. The risk of bias was evaluated with the revised Cochrane risk-of-bias tool for RCTs (RoB version 2.0).5Sterne J.A.C. Savovic J. Page M.J. et al.RoB 2: a revised tool for assessing risk of bias in randomised trials.BMJ. 2019; 366: l4898Google Scholar The systematic search in the databases (Medline, Embase, and PubMed Central) and trial registers (controlled-trials.com, clinicaltrials.gov, anzctr.org.au, who.int/trialsearch, and clinicaltrialsregister.eu) identified 2741 references, of which 37 underwent full-text review. Seven publications referring to 9 RCTs met the eligibility criteria. They showed different risks of bias (Table I). In the 9 RCTs, data for participant-specific recurrence rates were reported (Fig 1). Cryotherapy and photodynamic therapy with 5-aminolevulinate showed the lowest recurrence rates of 39%, respectively, with a 95% CI of 20% to 62% for cryotherapy and a 95% CI of 33% to 45% for photodynamic therapy with 5-aminolevulinate. They were followed by placebo (44%; 95% CI 25% to 65%), photodynamic therapy with methyl-aminolevulinate (45%; 95% CI 38% to 51%), and imiquimod 5% (45%; 95% CI 14% to 81%). The highest participant-specific recurrence rates were observed for diclofenac (85%; 95% CI 70% to 94%), followed by ablative laser treatment (54%; 95% CI 18% to 86%), and 5-fluorouracil 5% (52%; 95% CI 38% to 66%). Data on lesion-specific recurrence rates were available from 4 RCTs. They were lowest for placebo (15%; 95% CI 10% to 22%), followed by photodynamic therapy with 5-aminolevulinate (20%; 95% CI 16% to 24%), and photodynamic therapy with methyl-aminolevulinate (34%; 95% CI 23% to 47%). ALA, 5-aminolevulinate; MAL, methyl-aminolevulinate; n, number of patients; NR, not reported; PDT, photodynamic therapy. We conclude that long-term recurrence rates were surprisingly high, even after the initial clearance of lesions. In particular, the participant-specific recurrence rates ranged from 39% to 85%, indicating that the majority of patients will experience recurrent lesions in the long term. Although lesion-specific recurrence rates were lower, ranging from 15% to 34%, we infer that the majority of patients require regular surveillance and retreatments. The recurrence rates were similar for most active interventions and were not superior to placebo. However, the recurrence rates could be subject to a large increment if only a few lesions or patients show a relapse, because the number of the denominator was lower for the placebo arms than for the active interventions. Nevertheless, these data challenge the value of short-term clearance rates (which are the values that are usually reported), and they serve as primary endpoints in RCTs and highlight the nature of AK as a chronic health condition. Dr Berking has been a member of advisory boards for Almirall Hermal, Biofrontera, Galderma, Leo Pharma, and Sanofi Genzyme, has received speaker's honoraria from Almirall Hermal, Galderma, and Leo Pharma, and has received funding for clinical research from Leo Pharma . Dr Heppt has been member of advisory boards for Almirall Hermal and Sanofi-Aventis and has received speaker's honoraria from Galderma and Biofrontera. Dr Schmitz has been a member of advisory boards for Almirall Hermal and has received speaker's honoraria from Almirall Hermal, Biofrontera, Galderma, Mylan, and Sanofi-Aventis. Dr Brinker is the owner of Smart Health Heidelberg, GmbH, which is outside the scope of the submitted work. Drs Steeb, Wessely, and Petzold and author Schöffski have no conflicts of interest to disclose. The present work was performed in partial fulfillment of the requirements for obtaining the degree “Dr. rer. biol. hum.” for Dr Heppt at the Friedrich-Alexander University Erlangen-Nürnberg (FAU).