Keratocystic odontogenic tumors (KCOTs) of the jaw affect more than 65% of patients with basal cell nevus syndrome (BCNS). Surgery frequently causes facial disfigurement and is not always curative. Most BCNS-related and some sporadic KCOTs have malignant activation of the Hedgehog signaling pathway.
We examined the effect of vismodegib (an oral Hedgehog pathway inhibitor) on KCOT size in patients with BCNS enrolled in a clinical trial testing vismodegib for basal cell carcinoma prevention (NCT00957229), using pretreatment and posttreatment magnetic resonance imaging. Four men and 2 women had pretreatment KCOTs (mean longest diameter, 2.0 cm; range, 0.7-3.3 cm), occurring primarily in the mandible. Patients were treated with vismodegib, 150 mg/d, for a mean (SD) of 18.0 (4.8) months (range, 11-24 months). Four patients experienced a size reduction and 2 had no change. Vismodegib reduced the mean longest diameter of KCOTs in all patients by 1.0 cm (95% CI, 0.03-1.94; P = .02) or 50% from baseline. We observed no enlargement of existing KCOTs or new KCOT development.
Conclusions and Relevance
Vismodegib shrinks some KCOTs in patients with BCNS and may offer an alternative to surgical therapy. These effects were maintained for at least 9 months after drug cessation in 1 patient. Further studies assessing long-term efficacy and optimal maintenance regimens should be performed.
Basal cell nevus (Gorlin) syndrome (BCNS) (OMIM 109400) is a rare autosomal dominant disease in which affected individuals can develop a panoply of phenotypic abnormalities, the most prominent of which are basal cell carcinomas (BCCs) of the skin and keratocystic odontogenic tumors (KCOTs) of the jaw.1 The incidence of KCOTs varies from 4% to 16.5% of oral pathology specimens2; however, they affect 65% to 100% of patients with BCNS.1 Multiple KCOTs, which typically grow rapidly, are often the earliest presenting feature of BCNS. Surgery is the standard of care, but it frequently causes facial disfigurement and often is not curative.2- 4 To our knowledge, there is no effective nonsurgical therapy for KCOTs.
Patients with BCNS inherit 1 defective copy of the tumor suppressor gene PATCHED 1 (PTCH1) (OMIM 601309), which encodes a primary inhibitor of the Hedgehog (HH) signaling pathway. Functional loss of the second allele leads to activation of HH pathway signaling.5 Keratocystic odontogenic tumors in patients with BCNS and in 30% of sporadic lesions harbor somatically acquired PTCH1 mutations4 and increased HH target gene expression (GLI1).6 They also have been seen in heterozygous Ptch1 knockout mice and in K5-Gli2 transgenic mice.6,7 This discovery highlights the possibility that the newly approved targeted HH pathway inhibitor, vismodegib (Erivedge), might be effective for KCOT treatment. We conducted a substudy in patients enrolled in a recent clinical trial testing vismodegib for BCC prevention8 to determine its effect on KCOT size in patients with BCNS.
This study received institutional review board approval by Children’s Hospital Oakland Research Institute, and all participants provided written consent.
Of 41 patients with BCNS enrolled in a clinical trial (NCT00957229) testing vismodegib (150 mg/d) for BCC prevention,8 18 participated in the KCOT substudy and had pretreatment (baseline) magnetic resonance imaging (MRI). We performed posttreatment MRI in 8 patients with KCOTs at baseline who had received at least 11 months of vismodegib after the pretreatment MRI (Figure 1).
A CONSORT flow diagram of patient selection highlighting reasons for exclusion from the case series, as well as baseline and posttreatment magnetic resonance imaging (MRI) results. KCOT indicates keratocystic odontogenic tumor.
We recorded reasons for not having an MRI, time interval between MRIs, treatment duration, and KCOT characteristics. We calculated the percentage reduction in KCOT size (longest diameter in centimeters) from baseline and used the Wilcoxon signed rank test to compare the change in KCOT diameters before and after vismodegib treatment.
We excluded patients without paired MRIs from the analysis. We examined face and neck MRI (1.5 or 3.0 T) with coronal, sagittal, and axial T1 fast spin-echo and multiplanar T2 to demonstrate the intrinsic high T2 signal of KCOTs. A single study radiologist (T.J.) compared all paired MRIs and/or reports.
Of the 41 patients enrolled in the original clinical trial,8 we obtained baseline MRIs in 18, of whom 9 (50%) had KCOTs. Twenty-three patients refused to participate in the MRI-KCOT study because they had no history of tumors or had recent negative dental imaging, were unable to fit into the MRI scanner, or perceived having an MRI scan as an inconvenience (Figure 1).
We excluded from our analysis 3 of the 9 patients with KCOTs diagnosed on baseline MRI because 2 had tumors removed and 1 had a movement artifact on posttreatment MRI, precluding exact measurement. We performed posttreatment MRIs in 2 patients with no KCOTs at baseline due to new jaw pain to assess for recent growth of KCOTs while receiving vismodegib. Thus, a total of 6 patients with paired KCOT MRIs are included in the final analysis (Table and Figure 1).
Study participants included 4 men and 2 women, with a mean (SD) age of 51 (8) years (range, 37-59 years) (Table). All patients had previous removal of KCOTs, with most having had at least 2 procedures.
In 6 patients with a total of 9 documented KCOTs, we obtained posttreatment MRIs within a mean (SD) of 19.5 (4.0) months (range, 13-23 months) after baseline imaging. During this interval, we treated these patients with vismodegib for a mean (SD) of 18.0 (4.8) months (range, 11-24 months); 3 patients took drug breaks (ie, a brief period off vismodegib treatment) (Table). The baseline MRI of 1 of the 6 patients was performed 5 months after starting vismodegib, with 2 KCOTs seen. The mean longest diameter of baseline KCOTs was 2.0 cm (range, 0.7-3.3 cm). In 6 of 9 patients, KCOTs occurred primarily in the mandible. Of the 6 patients with tumors, 4 experienced a reduction in KCOT size and 2 had no change. Vismodegib reduced the longest diameter of all monitored KCOTs by a mean of 1.0 cm (95% CI, 0.03-1.94; P = .02), that is, by 50% from baseline. No enlargement or new KCOTs occurred in any case, including the 2 patients without KCOTs at baseline. In one patient with 100% resolution of 1 KCOT with vismodegib, there was no recurrence 9 months after drug cessation (Figure 2). All patients experienced mild adverse effects (grade 1) of taste loss, muscle cramps, and hair loss, as previously reported.8 Three experienced grade 2 muscle cramps and gastrointestinal disturbance, necessitating brief drug breaks.
Serial sagittal T2-weighted magnetic resonance images demonstrating the size change in the left maxillary keratocystic odontogenic tumor (arrows) (A) from baseline, (B) after 11 months of vismodegib, and (C) 9 months after drug discontinuation.
We found that vismodegib can shrink some KCOTs in patients with BCNS, with a mean size reduction of 50% in our study. Two patients had complete resolution of 1 or more KCOTs, with no enlargement of existing tumors or development of new lesions in any case. We found no KCOT relapse after a 9-month drug discontinuation in 1 patient, who also had minimal BCC recurrence at that time.
Vismodegib is an oral HH pathway inhibitor approved for the indefinite treatment of locally advanced or metastatic BCCs. In one case report, Goldberg et al3 incidentally noted resolution of KCOTs in 1 patient with BCNS treated with vismodegib (270 mg/d) for 2 years. This dose was used in phase I studies before the data demonstrated the 150-mg/d dosage to be the most efficacious.
It is unclear why vismodegib was effective in some but not all of our patients. The 3 patients with multiple KCOTs in our study appeared to have had the best response, regardless of size, months of treatment, or drug breaks. This may reflect the increased proliferative activity associated with multiple tumors, a finding that has been associated with protein truncating PTCH1 mutations in both sporadic and syndrome-related KCOTs.9 Another hypothesis is that higher doses of vismodegib are required to induce shrinkage. Indeed, the dose used in our study was almost half of that used in the case report by Goldberg et al,3 and cyclopamine (another HH pathway inhibitor) has significantly arrested the growth of KCOT cells in a dose-dependent manner in vitro.10 However, since all our patients received the same US Food and Drug Administration–approved dose of vismodegib, we were not able to assess any dose-dependent response.
In 1 patient, there was no KCOT recurrence 9 months after drug cessation. In fact, there was further size reduction, albeit minimal, of the mandibular tumor. Interestingly, this patient had recurrence of only 1 BCC in the same period, a much lower recurrence rate than previously reported.8 The incidence of KCOT recurrence in reported series varies from 0% to 62%, with a higher recurrence rate in BCNS-related tumors (>80%) compared with sporadic tumors.11 This variability could relate to the diverse nature of the cases reported, the different treatment protocols used, and variations in posttreatment monitoring. The management of KCOTs remains controversial. Conservative therapy, such as simple enucleation or marsupialization, has a recurrence rate of 50% compared with a lower rate following aggressive complete resection.12 Recurrence, which typically develops within 5 years of treatment, is thought to occur due to incomplete removal of the original cysts, the presence of microscopic satellite cysts, or new cyst development in adjacent areas.13 Relapse has been reported even after 41 years,14 necessitating long-term follow-up, particularly since these tumors have locally aggressive behavior and, rarely, malignant potential (primary intraosseous squamous cell carcinoma).4
Limitations of this study include possible selection bias since only a subset of patients elected to undergo imaging, selecting for patients with recent or symptomatic KCOTs. Most of the 41 clinical trial patients reported previous surgical treatment for KCOTs. We studied only patients with BCNS, thus limiting the generalizability of the results to syndrome-related KCOTs. However, since 30% of sporadic KCOTs harbor a PTCH1 mutation,4 vismodegib could be effective for a significant percentage of sporadic tumors. In addition, analysis of KCOT tissue to assess for PTCH1 mutations could explain some variability in response. Strengths include the availability of pretreatment and posttreatment MRIs for KCOT size analysis, as well as, to our knowledge, the largest number to date of patients with KCOTs treated with vismodegib. We chose to image with MRI instead of dental radiographs to improve KCOT measurement and avoid exposure to ionizing radiation, which can induce BCCs in patients with BCNS.15
Our findings confirm functional involvement of the HH pathway in KCOTs and illustrate the usefulness of treatment with molecularly targeted drugs such as vismodegib. This is particularly important for patients with multiple KCOTs, who can be left with facial disfigurement and a speech impediment following multiple surgical procedures.3 Vismodegib offers a nonsurgical treatment option, potentially revolutionizing the management of KCOTs in BCNS. It is difficult to predict which patients develop KCOTs due to variable expression in BCNS; however, MRI may be used for screening during the initial diagnostic workup and for those with jaw symptoms. Further studies into optimal maintenance regimens of vismodegib for KCOT treatment are required. An ongoing clinical trial (NCT01556009) assessing its use as maintenance therapy for BCCs may yield more information.
Accepted for Publication: August 13, 2013.
Corresponding Author: Ervin H. Epstein Jr, MD, Children's Hospital Oakland Research Institute, 5700 Martin Luther King Jr Way, Oakland, CA 94609 (firstname.lastname@example.org).
Published Online: March 12, 2014. doi:10.1001/jamadermatol.2013.7444.
Author Contributions: Drs Ally and Epstein had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ally, Tang, Joseph, Mackay-Wiggan, Epstein.
Acquisition of data: All authors.
Analysis and interpretation of data: Ally, Tang, Joseph, Thompson, Chanana, Epstein.
Drafting of the manuscript: Ally, Tang, Lindgren, Raphael, Ulerio.
Critical revision of the manuscript for important intellectual content: Ally, Tang, Joseph, Thompson, Chanana, Mackay-Wiggan, Bickers, Epstein.
Statistical analysis: Ally, Tang, Epstein.
Obtained funding: Tang, Bickers, Epstein.
Administrative, technical, or material support: Ally, Tang, Joseph, Thompson, Lindgren, Raphael, Ulerio, Mackay-Wiggan.
Study supervision: Tang, Mackay-Wiggan, Epstein.
Conflict of Interest Disclosures: Drs Tang and Epstein reported being consultants for Genentech. Dr Epstein also reported being a consultant for Novartis and owning stock options in Curis and Infinity. No other disclosures were reported.
Funding/Support: This study was supported by Genentech, a Clinical and Translational Science Award (UL1RR02413) from the National Institutes of Health, grants 1K23AR056736 (Dr Tang) and 5P30AR044535-11 (Dr Bickers) from the National Institute of Arthritis and Musculoskeletal and Skin Diseases, grant R01CA109584 from the National Cancer Institute (Dr Epstein), a Clinical Investigator Award (CI-54-11) from the Damon Runyon Cancer Research Foundation (Dr Tang), and funding from the Swim Across America Foundation and the Michael J. Rainen Family Foundation.
Role of the Sponsor: Genentech supplied the medication and read drafts of the manuscript for clarity.
Ally MS, Tang JY, Joseph T, Thompson B, Lindgren J, Raphael MA, Ulerio G, Chanana AM, Mackay-Wiggan JM, Bickers DR, Epstein EH. The Use of Vismodegib to Shrink Keratocystic Odontogenic Tumors in Patients With Basal Cell Nevus Syndrome. JAMA Dermatol. 2014;150(5):542-545. doi:10.1001/jamadermatol.2013.7444