FDA Approval Summary: Sonidegib for Locally Advanced Basal Cell Carcinoma

Basal cell carcinoma (BCC) is the most common human cancer, accounting for approximately 80% of nonmelanoma skin cancers (1). Abnormal activation of the hedgehog (Hh) signaling pathway is a major driver in BCC pathogenesis. The majority of mutations occur in the patched 1 (PTCH1) gene, and the aberrant gene product induces activation of smoothened (SMO), triggering constitutive activation of the Hh pathway (2). Risk factors for the development of BCC include fair skin pigmentation; radiation; exposure to arsenic or aromatic hydrocarbons; immunosuppression; and underlying genetic syndromes, such as nevoid BCC syndrome (NBCCS). NBCCS, also known as Gorlin syndrome, is a hereditary condition caused by germline mutations in PTCH1 and characterized by developmental abnormalities and a predisposition to developing cancers, predominantly medulloblastoma and BCC (3). The majority of patients with BCC are treated curatively with local therapies, including surgical resection, and in select cases, radiation. The recurrence rate is estimated to be 5% after local treatment (4).

Advanced BCC, including locally advanced and metastatic disease, accounts for approximately 1% to 10% of BCC and confers a poor prognosis (5). No formal criteria exist to define locally advanced BCC (laBCC), and treatment decisions with regard to choosing localized versus systemic therapies are at times controversial. This population generally includes patients with multiply recurrent tumors in which repeat resections would not be curative, could be disfiguring, or cause a major functional deficit (e.g., facial, periorbital, cranial nerve tumors). For patients with laBCC no longer amenable to local treatment, progressive disease can result in substantial morbidity due to tissue invasion that can cause infection, bleeding, severe disfigurement, and risk for development of metastatic disease (6). For patients with mBCC that has spread to the lungs, bone, or liver, average survival is 8 to 14 months (5).

Treatment with a SMO inhibitor is recommended for patients with advanced BCC when surgery would not be curative or would cause unacceptable morbidity, and for patients with disease refractory to radiation or when radiation is contraindicated (4). Prior to the approval of sonidegib, the first-in-class Hh pathway inhibitor vismodegib (ERIVEDGE; Genentech) was the only available systemic treatment for patients with advanced BCC (7, 8).

Sonidegib is a synthetically derived new molecular entity (chemical name: N-{6-[(2R,6S)-2,6-dimethylmorpholin-4-yl]pyridin-3-yl}-2-methyl-4′-(trifluoromethoxy)[1,1′-biphenyl]-3- carboxamide diphosphate; molecular formula: C26H26 F3N3O3• 2H3PO4) that disrupts the Hh signaling pathway by binding and inhibiting SMO receptors. Sonidegib is primarily metabolized by CYP3A in the liver and exhibits an elimination half-life of 28 days with 19-fold accumulation at steady state after daily dosing. The increase in systemic exposure to sonidegib is less than dose proportional at doses greater than 400 mg.

The FDA approval of this New Drug Application was based on results of one clinical study of sonidegib for the treatment of patients with laBCC not amenable to curative surgery or radiotherapy. Study LDE225A2201 (A2201) was an international, randomized, double-blind, noncomparative trial in 230 patients with laBCC or mBCC who were randomized 2:1 to receive sonidegib 800 mg (n = 151) or 200 mg (n = 79) once daily continuously. Randomization was stratified by stage of disease (locally advanced or metastatic), aggressive or nonaggressive histology (for patients with laBCC), and geographic region (Australia, Europe, and North America). Eligibility criteria included age of at least 18 years; histologically confirmed diagnosis of BCC; measurable disease not amenable to radiotherapy, curative surgery, or other local therapies; WHO performance status ≤2; adequate organ function; no prior use of Hh pathway inhibitors; and no known neuromuscular disorders or use of concomitant medications recognized to cause rhabdomyolysis. Women of child-bearing potential and sexually active males were required to use highly effective contraception throughout the study and for 6 months after the last administration of study treatment. Patients continued study treatment until disease progression, death, intolerable toxicity, or withdrawal of consent.

The primary efficacy endpoint was centrally reviewed, confirmed objective response rate (ORR) according to RECIST version 1.1 in patients with mBCC and a protocol-specific, modified RECIST (mRECIST) for patients with laBCC. The mRECIST guidelines involved a composite response assessment based on integration of radiologic (MRI), photographic, and histopathologic results (see Table 1). MRI response measurements were made according to RECIST. Digital photographs of target and nontarget lesions were taken at scheduled assessments, including close-up views with a ruler and lesion contouring to include both visible and palpable components for reliable size determination. Histopathology was assessed on tumor punch biopsies from accessible laBCC lesions taken at baseline, weeks 9 and 17, disease progression, and end of treatment. All modalities must have demonstrated absence of tumor to achieve a composite assessment of complete response (CR). Additional biopsies were performed to confirm CRs and to determine the presence or absence of tumor cells any time a response assessment was confounded by ulceration, cyst, or scarring. Radiology, photography, and pathology were centrally reviewed. An independent review committee consisting of two dermato-oncologists and one radiologist determined the composite overall response for each patient.

The primary efficacy analysis was conducted when all patients had been treated or followed for at least 24 weeks. Treatment with sonidegib was considered to be of clinical benefit if the observed ORR by central review on any treatment arm at the end of the study was 30% or higher, with the lower bound of the 95% confidence interval (CI) exceeding 20%. There were no planned statistical comparisons.

The trial enrolled 230 patients. The proportion of patients with mBCC in each treatment arm was similar [16% (n = 13) of the 200 mg group and 15% (n = 23) of the 800 mg group]. Sixteen patients (7%) had a diagnosis of nevoid BCC syndrome (i.e., Gorlin syndrome). More than 75% of patients on both arms had prior surgery and/or radiotherapy. The most common reasons for enrollment of patients with laBCC included tumor location not amenable to surgery or radiation, risk for severe disfigurement, and multiple recurrences after prior surgery or radiation.

The efficacy of sonidegib was evaluated separately for patients with laBCC and mBCC, because these subtypes differ in clinical presentation, management principles, and overall prognoses. The primary analysis was conducted 6 months after initiation of treatment for all patients. Response data from an updated analysis after 12 months of follow-up were also considered for the benefit–risk assessment and product labeling. Efficacy results are summarized in Table 2. For patients with laBCC, the ORR for both treatment arms exceeded the prespecified assumptions—that is, >30% with the lower bound of the 95% CI excluding an ORR of 20% or less. The 12-month data demonstrated that seven additional patients with laBCC in the 200 mg treatment arm achieved a response [CR = 1, partial response (PR) = 6] compared with 6-month data after longer exposure to sonidegib. For the subset of patients with mBCC, neither treatment arm met the assumptions.

Duration of response (DoR) and CR rate (CRR) were key secondary endpoints. In the 12-month analysis, the median DoR in patients with laBCC was nonestimable (NE) for the 200 mg arm and 15.7 months (95% CI, NE) for the 800 mg arm. Among the 38 patients with an objective response treated at the recommended 200-mg dose, 31 (82%) had ongoing responses ranging from 1.9+ to 18.6+ months. Four of the seven patients who eventually developed progressive disease maintained a response of 6 months or longer. The CRR for patients with laBCC was 5% for the 200 mg arm and 2% for the 800 mg arm. A prespecified sensitivity analysis using an alternative definition for CRR, defined as at least a PR according to MRI and/or photography and no evidence of tumor on biopsy of the residual lesion, yielded a CRR of 20%. There were no CRs observed at either dose for patients with mBCC.

The safety population included 229 patients with BCC who received at least one dose of sonidegib in Study A2201. Safety data from the 6-month primary analysis, as well as 12-month and 18-month analyses, were reviewed. The 18-month data were comprehensive and used to inform the product label given the anticipation for chronic dosing in some patients with laBCC. The median duration of treatment in the 18-month analysis was 11 months (range, 1.3–33.5) for patients in the 200 mg arm and 6.5 months (range, 0.3–31.5) for patients in the 800 mg arm. The shorter exposure in the 800 mg group was attributed to earlier discontinuation due to adverse events (AE).

Table 3 lists the most common adverse reactions occurring in ≥10% of patients treated with sonidegib at either dose. The frequencies of all common adverse reactions except diarrhea, arthralgia, asthenia, abdominal pain, and headache were greater in patients treated at the higher dose. Eight patients died while on treatment or within 30 days of receiving a dose of sonidegib. Two deaths were attributed to disease progression. None of the other six deaths were attributed to study treatment based on underlying comorbidities and risk factors. Nonfatal serious AEs (SAE) occurred in 18% of patients in the 200 mg group and 37% in the 800 mg group. In the 200 mg group, no more than one type of any SAE occurred. In the 800 mg group, SAEs that occurred in at least 2% (n = 3 or more) of patients included rhabdomyolysis (n = 5), blood creatine kinase (CK) elevation (n = 5), anemia (n = 5), vomiting (n = 4), pneumonia (n = 3), and nausea (n = 3).

Rhabdomyolysis is a serious risk of sonidegib. SMO inhibitors induce muscle contraction and muscle fiber twitching in primary human muscle cells by inducing changes in normal calcium influx (9). An Independent Safety Review and Adjudication Committee for Muscular Events (IAC) reviewed pooled safety data for musculoskeletal AEs occurring across the sonidegib development program. Rhabdomyolysis was defined as a greater than 10-fold serum CK elevation above baseline plus a 1.5-fold increase in serum creatinine from baseline or greater than 10 times upper limit of normal if no baseline level was reported. Across 12 clinical studies involving 571 patients treated with sonidegib at doses ranging from 100 to 3,000 mg, rhabdomyolysis occurred in one patient (0.2%). None of the five investigator-assessed events of rhabdomyolysis reported during A2201 met the IAC definition for rhabdomyolysis. These were adjudicated as myositis, myalgia, and/or muscle spasms.

The FDA performed additional analyses using the A2201 data to characterize the musculoskeletal toxicity of sonidegib. Musculoskeletal adverse reactions requiring medical interventions, including administration of analgesics, muscle relaxants, magnesium supplementation, intravenous hydration, or hospitalization, were experienced by 29% of patients in the 200 mg arm and 43% in the 800 mg arm. Serum CK elevation occurred in 61% and 67% of patients in the 200 and 800 mg treatment groups, respectively, and grade 3 or 4 CK elevations occurred in 8% (200 mg cohort) and 17% (800 mg cohort) of patients. Among patients in the 200 mg group with grade 2 or higher CK elevations, the median time to onset was 12.9 weeks (range, 2–39 weeks) and the median time to resolution (to ≤grade 1) was 12 days (95% CI, 8–14 days). The median time to onset in the 800 mg group was 6.7 weeks (range, 2–40 weeks), and the median time to resolution was 15 days (95% CI, 9–15 days). One patient (1%) in the 200 mg group required dose reduction for musculoskeletal adverse reactions. A dose lower than 200 mg daily was not studied, because patients in the 200 mg group were dose reduced to placebo per protocol. In the 800 mg group, 22 patients (15%) experienced musculoskeletal adverse reactions requiring dose reduction; 12 of 18 patients (67%) with documented follow-up experienced a recurrence of symptoms or serum CK elevation at the reduced dose.

The sonidegib label includes a black box warning for embryofetal toxicity based on the results of animal studies in which abortion, complete resorption of fetuses, and severe malformations, including vertebral, distal limb and digit, and craniofacial and other midline defects, were observed. Skeletal variations were observed when maternal exposure to sonidegib was below the limit of detection. Females and males (due to the potential risk of exposure through semen) of reproductive potential are advised to use effective contraception during treatment and for 20 months (for females) or 8 months (for males) after the last dose.

Sonidegib was granted regular approval on July 24, 2015, for the treatment of patients with laBCC that has recurred following surgery or radiotherapy, or those who are not candidates for surgery or radiotherapy. The ORR in the 66 patients with laBCC receiving 200 mg of sonidegib daily of 58% (95% CI, 44.8–69.7), with the median DoR not reached at the 12-month analysis, was of sufficient magnitude and durability to be deemed direct evidence of clinical benefit. Photographic evidence demonstrating substantial responses in disfiguring and morbid BCC lesions was also considered of direct clinical benefit. Finally, knowledge that sonidegib works through interruption of a well-described signaling pathway responsible for driving BCC pathogenesis provided increased confidence that the demonstrated antitumor effects were treatment related.

Key issues that arose during the FDA's review of the sonidegib application pertained to restriction of the proposed indication to patients with laBCC, dose selection, and provision of adequate risk description and reasonable dosing guidelines in product labeling.

The efficacy in the cohort of patients with mBCC was evaluated separately from the cohort of patients with laBCC. Neither treatment arm met the targeted ORR effect size and included response rates of less than 10% within the lower bound of the 95% CI. Novartis proposed that the disease control rate observed in patients with mBCC demonstrated evidence of clinical benefit. However, the FDA concluded that maintenance of stable disease for this population in an uncontrolled trial is confounded by the natural history of disease and, therefore, does not reliably represent clinical benefit.

In Study A2201, Novartis evaluated doses supported by data from preclinical models and from a dose-finding study in which doses ≥200 mg/day resulted in systemic exposures associated with antitumor activity. The maximum tolerated dose (MTD) was 800 mg daily. On the basis of an anticipated exposure–response relationship, a 2:1 unbalanced allocation to the 800 mg versus 200 mg treatment arms was an important design element of the A2201 trial. Although no exposure–response relationship for efficacy was actually observed, an exposure-dependent relationship was observed for safety, with an increase in the frequency and severity of AEs and a higher proportion of patients requiring discontinuation for adverse reactions resulting in a shorter median duration of treatment in the 800 mg arm. Therefore, the 200-mg dose was chosen as the recommended dose. These data suggest an argument for dose selection based on a biologic rationale that may produce similar and more durable antitumor effects with less risk for toxicity than using an MTD.

Sonidegib administration carries a risk for serious musculoskeletal toxicity, including rhabdomyolysis. However, as some patients who discontinued sonidegib for musculoskeletal toxicity had durable responses of more than 3 months following the last dose of sonidegib, the risk–benefit was considered favorable. Optimal management of serious musculoskeletal toxicity was not evaluated in the clinical development program; thus, the FDA relied on the dosing guidelines employed in Study A2201 and the observed clinical experience. It is recommended that clinicians obtain baseline serum CK and creatinine levels prior to initiating sonidegib, periodically during treatment, and at least weekly in patients with muscle symptoms and concurrent serum CK elevation until resolution of clinical signs and symptoms. Treatment interruption at the onset of muscle symptoms is recommended in product labeling based on data demonstrating that muscle symptoms usually present prior to experiencing laboratory evidence of CK elevation. Permanent discontinuation is recommended for patients who experience a grade 4 CK elevation based on the absence of data supporting that dose reduction is effective in managing this toxicity and on sonidegib's long terminal half-life.

The approval of sonidegib for the treatment of patients with laBCC not amenable to curative surgery or radiotherapy offers another treatment option for patients with a serious disease who are subject to significant morbidity and disfigurement. The observed response rate of 58%, with approximately half of the responses being durable for at least 6 months, represents direct evidence of clinical benefit. The risks of sonidegib can be mitigated by routine laboratory assessments, dose interruptions, or discontinuation in the setting of specific AEs or laboratory abnormalities, and use of effective contraception. These risks were deemed acceptable to the medical and patient community, as confirmed by advice provided by Special Government Employees with expertise in the clinical management of patients with advanced BCC consulted during the review.

No potential conflicts of interest were disclosed.

The Deputy Editor handling the peer review and decision-making process for this article has no relevant employment associations to disclose.

Conception and design: R. Pazdur

Development of methodology: R. Pazdur

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Chen, R. Pazdur

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D. Casey, S. Shord, H. Chen, K. He, A. Putman, P. Keegan, R. Pazdur

Writing, review, and/or revision of the manuscript: D. Casey, S. Demko, S. Shord, H. Zhao, H. Chen, K. He, A. Putman, W. Helms, P. Keegan, R. Pazdur

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Putman, R. Pazdur

The authors thank Anuja Patel, Senior Regulatory Health Project Manager.