Hedgehog Blockade for Basal Cell Carcinoma: Coming at a (Secondary Neoplastic) Price.

Abstract

Hedgehog inhibitors, more accurately known as smoothened inhibitors (SIs),were introduced for the treatmentofbasal cell carcinoma(BCC)withmuchsuccess.1Theclinicaldata supported the intriguing rationale of interfering with a tumor driver (overactive Sonic Hedgehog signaling in BCC) thatwas identified 20years ago. Beyond thewell-established surgical or conservative (radiotherapy) treatment options, treatmentwithSIsnowrepresentsanovelavenuethedermatooncologist can take when surgery or radiotherapy of BCC is contraindicated.2This extensionof the therapeutic armamentarium for advanced BCC, which is currently used in patients with highly advanced tumors or basal cell nevus syndrome, has also suggested using vismodegib or sonidegib treatment inaneoadjuvant setting.However,neoadjuvant therapywould be precluded by the development of secondary resistance to vismodegib, which is mediated in the vast majority of cases by secondary mutations in smoothened.3,4 Neoadjuvant treatment may induce regression of advancedBCC in patientswith comorbidities that do not permit surgery at reasonable risk orwhenanatomical limitationspreclude radiotherapy. Intermittent oral treatment for limitedperiods may permit either palliative regression of the tumor or convert a clinical situationnot suitable for surgery to one that would allow less radical surgery. However, such broader use of SIs would immediately have to take into account adverse effects. In particular, the occurrence of secondary independent invasivecutaneoussquamouscell carcinomas (SCCs)during treatment with SIs, which have thus far only been reported in an anecdotal manner, is of great relevance.5,6 In addition, low numbers of cases with putative transdifferentiationofvismodegib-treated tumorhavealsobeendescribed.7 Thesepotential adverse effects need tobe acknowledged, understood, andmanagedbeforewidespreaduse of SIs in aneoadjuvant setting. In this issueof JAMADermatology,Mohanet al8 retrospectively compared the incidenceof secondary skin cancers inpatients with advanced BCC treated with vismodegib vs that in controlpatientsnot treatedwithvismodegib.Although theauthors found an overall higher frequency of secondary cancers in the cohort of patients who did not receive vismodegib, the meanageof thecontrolgroupwassignificantlyhigher thanthat of the treatment group. As age is a strong risk factor for development of cancer, the authors consequently corrected the incidenceof secondarycancers for age.Whenthisdifferencewas accounted for, patients receivingvismodegib vs thosewhodid not receive SI treatment demonstrated ahighly enhancedhazard ratio for developing secondary epithelial skin cancers. The enhancedhazard ratio alsopersistedafter correcting for thediagnosis of basal cell nevus syndrome and for the histologic diagnosis of metatypical BCC. Unavoidably, owing to the nature of a “first-in-class”drug, the studydesignmaybebiasedby selectionand/or chronological biasbecauseof thenonparallel recruitment of the cohorts who received vismodegib and those who did not. In future studies, investigators will need to pay closeattentiontothesimilarityof theresponseanddropoutpatterns between the 2 groups. Thevalidity of the epidemiologic conclusionsdependson whether it was adequate to correct the incidence of secondary cancers for age or whether the adjustment for age within the treatment groups is incorrect. Incidence of cancer in general and of skin cancer specifically is strongly correlatedwith age. In the data set used by Mohan et al,8 the reduced frequency of cancer other than skin tumors in the younger patient cohort receiving vismodegib, compared with the older control group, at least supports this assumption. Still, onemay speculate that patientswhowere treatedwithvismodegibdeveloped advanced BCCs at a younger age owing to increased accumulation of UV light–inducedmutations in sun-exposed skin. Because themutational load is surprisingly high in clinicallynoncancerous skin thathasbeenexposed toUV lightover the long term,9anenhanced incidenceof secondarySCCwould then be independent of vismodegib treatment. Mohan et al8 tried to avoid such bias first by screening their patient database for those with advanced or metastatic BCC and by subsequentstratificationof thecohorts forexposure toknownmutagenic agents as well as for immunosuppression. Yet, the ultimate technique to exclude a bias would have consisted in performing a prospective study that incorporates assessment of mutational loads at multiple skin sites in patients who receivedvismodegib aswell as in the control group. Such elaborated studies may be performed in the future, given today’s increasingly broad accessibility of advanced unbiased ultradeep sequencing technologies that may change our understandingofUV-induced carcinogenesis.9Mohanet al8 laid the groundwork for such studies in future patients treated with vismodegib. Howcould the transition of BCC to SCCor the denovodevelopment of clinically apparent SCC, a tumor-promoting effect, be explained at themolecular level in patients receiving SI treatment? A possible scenario is that suppression of the Hedgehog pathway might select for tumor cells that have overcome Hedgehog pathway dependency by activation of Related article page 527 Opinion