aPatient had a complete response (CR) on his last study visit prior to his withdrawal because of unresolved grade II skin pain; although maintenance of CR for 4 weeks could not be established, this patient was assigned a response of CR because these were the only available data. AST indicates aspartate aminotransferase; DLCO, diffusing capacity of the lungs for carbon monoxide; PD, progressive disease; PR, partial response; and SD, stable disease.
The dashed line represents 50% SWAT score reduction. Patients with a partial response (PR) had 79% to 95% clearance of mycosis fungoides. CR indicates complete response; PD, progressive disease; and SD, stable disease.
aThe patient had significant skin inflammation after 3 cycles and withdrew. On clinic follow-up 1 month after withdrawal, a skin biopsy was performed to determine whether the skin changes represented inflammation or persistent cutaneous T-cell lymphoma (CTCL); pathology testing showed no evidence of CTCL and the patient was assigned a response of SD.
bThe patient had a CR on his last study visit before his withdrawal because of unresolved grade II skin pain; although maintenance of CR for 4 weeks could not be established, this patient was assigned a response of CR because these were the only available data.
Red scaly mycosis fungoides plaques prior to treatment and after 12 cycles of O6-benzylguanine plus carmustine on the right medial thigh (A) with complete response (CR) (B), pubic skin (C) with CR (D), and left buttock (E) with partial response (F).
eTable 1. Predictors of Complete Response: Univariate Logistic Regression Analyses
eTable 2. Predictors of Time to Response: Univariate Regression Analyses
eTable 3. Frequency of Grade I and Grade II Adverse Events Experienced By Patients Based on the National Cancer Institute Common Terminology Criteria for Adverse Events Version 3.0 Grading System
eFigure 1. Representative Photos of Dermatitis
eFigure 2. Confocal Images
Tacastacas JD, Chan DV, Carlson S, Gerson SL, Dowlati A, Fu P, Lu K, Groft S, Rosenjack J, Honda K, McCormick T, Cooper KD. Evaluation of O6-Benzylguanine−Potentiated Topical Carmustine for Mycosis FungoidesA Phase 1-2 Clinical Trial. JAMA Dermatol. Published online February 15, 2017. doi:10.1001/jamadermatol.2016.5793
What is the maximum tolerated dose of topical carmustine in combination with dual dosing of intravenous O6-benzylguanine for the treatment of stage IA through stage IIA cutaneous T-cell lymphoma, mycosis fungoides type?
In this phase 1-2 clinical trial including 17 patients, the MTD of carmustine was 20 mg applied once, in combination with 2 daily doses of 120 mg/m2 of O6-benzylguanine. The overall clinical response rate was 88% and adverse events were mild.
Carmustine in combination with dual-dose O6-benzylguanine is an effective, well-tolerated, skin-directed therapy for early stage cutaneous T-cell lymphoma, mycosis fungoides type.
In a phase 1 trial, single-dose O6-benzylguanine with topical carmustine for patients with early stage (stage IA through stage IIA) cutaneous T-cell lymphoma, mycosis fungoides (MF) type, resulted in clinical responses proportional to inhibition of O6-alkylguanine DNA alkyltransferase activity, but a maximum tolerated dose (MTD) was not reached.
To determine whether dose escalation of carmustine in combination with dual-dose O6-benzylguanine to prolong alkyltransferase inhibition could reach an MTD.
Design, Setting, and Participants
A single-arm phase 1-2 clinical trial conducted at a university teaching hospital enrolled 17 adults with stage IA or stage IIA cutaneous T-cell lymphoma, MF type, to evaluate treatment using topical carmustine plus 2 subsequent daily doses of intravenous O6-benzylguanine, administered every 2 weeks for up to 24 weeks (12 cycles). All patients who received treatment were included in an intent-to-treat analysis of the response rate. The study was conducted from February 17, 2010, to April 8, 2014. Data analysis was performed from May 1, 2014, to December 1, 2015.
Topical carmustine and O6-benzylguanine.
Main Outcomes and Measures
Clinical disease response was assessed by the Severity-Weighted Assessment Tool (score range, 0-400; higher score indicates worse disease). Safety data were acquired by review of adverse events at study visits.
Of the 17 patients enrolled, 12 (71%) were men; mean (SD) age was 45.2 (14.6) years. There were 7 complete responses and 8 partial responses to combination O6-benzylguanine treatment. The overall clinical response rate was 88%, with a mean (SD) duration of complete response of 14.43 (6.6) months. The MTD was 20 mg of carmustine applied once in combination with 2 daily doses of 120 mg/m2 of O6-benzylguanine. Most adverse events (118 [67%]) were grade I. Of 15 patients with dermatitis, 5 individuals (33%) demonstrated grade II dermatitis that was unresponsive to topical corticosteroid therapy. The dermatitis was characterized by high levels of macrophage activation and clearance was associated with vitamin D3 administration.
Conclusions and Relevance
Compared with single-dose O6-benzylguanine and carmustine, dual-dose O6-benzylguanine resulted in higher overall response rates and reduced total carmustine doses but was associated with more cutaneous adverse events. The MTD for dual-dose O6-benzylguanine plus carmustine was also ascertained.Trial Registration clinicaltrials.gov Identifier: NCT00961220
The efficacy and safety of combination O6-benzylguanine and carmustine therapy for patients with stage I through stage IIA cutaneous T-cell lymphoma (CTCL), mycosis fungoides (MF) type, was first demonstrated in a phase 1 trial.2 However, the maximum tolerated dose (MTD) was not achieved. In that study, 21 patients were treated once every 2 weeks with a single dose of intravenous O6-benzylguanine followed by a single application of topical carmustine. Compared with conventional topical carmustine therapy alone, single-dose O6-benzylguanine with carmustine resulted in less total drug exposure per treatment course and allowed a longer duration of treatment owing to a lower incidence of cutaneous reactions.2,3 Clinical responses were proportional to inhibition of the DNA repair enzyme, O6-alkylguanine DNA alkyltransferase (AGT) activity,2 evidenced by a lower area under the curve of AGT activity that correlated with greater Severity-Weighted Assessment Tool (SWAT)4 reduction (score range, 0-400, with higher score indicating worse disease). The phase 1-2 clinical trial reported in this article was conducted to determine whether doubling the O6-benzylguanine dose to 2 subsequent daily doses would potentiate the efficacy and/or toxicity of a dose escalation of topical carmustine.
Ambulatory patients with CTCL, MF type, were recruited from a university teaching hospital from February 17, 2010, to November 12, 2013; the study was completed April 8, 2014. The full trial protocol is available in Supplement 1 (clinicaltrials.gov identifier NCT00961220). Inclusion criteria were age 18 years or older, histologically confirmed stage IA through stage IIA CTCL, MF type, recalcitrant to at least 1 prior therapy, Eastern Cooperative Oncology Group grade 0 to 2, adequate marrow function (white blood cell count at least 3500/µL [to convert to ×109/L, multiply by 0.001], absolute neutrophil count ≥1600/µL [to convert to ×109/L, multiply by 0.001], and platelets >100 × 103/µL [to convert to ×109/L, multiply by 1]), adequate hepatic function (bilirubin <1.5 mg/dL [to convert to micromoles per liter, multiply by 17.104] and aspartate aminotransferase within the reference range of 10-37 U/L [to convert to microkatals per liter, multiply by 0.0167]), adequate renal function (creatinine ≤1.5 mg/dL [to convert to micromoles per liter, multiply by 88.4]), normal serum electrolyte levels, and no CTCL therapy 4 weeks before the first dose of study medications, except topical corticosteroids, which may be used up to 2 weeks before enrollment. Exclusion criteria were prior treatment with carmustine or nitrosoureas, known central nervous system primary malignant tumors or involvement by lymphoma, pregnant or breast feeding, current reproductive potential without adequate contraception, active infection requiring hospitalization, and pulmonary disease with diffusing capacity of the lungs for carbon monoxide less than 70%.
The study was approved by the University Hospitals Cleveland Medical Center institutional review board, and the patients provided written informed consent. There was no financial compensation.
O6-benzylguanine and carmustine were given on 2 consecutive days every 2 weeks for up to 12 treatment cycles. At each cycle, O6-benzylguanine was administered intravenously at a fixed dose of 120 mg/m2 over 1 hour (day 1) and then 24 hours later (day 2). One hour after the first O6-benzylguanine infusion, carmustine solution was applied topically to the total skin surface, except for uninvolved areas of the scalp, face, lips, and eyelids. Preparation of the carmustine solution was described in the phase 1 study.2 The protocol was modified to remove the dose of carmustine on day 2 after initial observations of 3 patients with dose-limiting toxic (DLT) effects with respect to pruritus and/or pain in MF lesions treated with 20 mg of carmustine on days 1 and 2. In the modified protocol, 3 carmustine dose levels were set (20, 30, and 40 mg), given on day 1 only. Carmustine levels were escalated, starting at 20 mg, in patient cohorts of 1 each, unless there were DLT effects, defined as National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE), version 3.0, grade II or higher toxic effects, after 4 weeks of observation (2 cycles).5 If a DLT effect occurred at a certain dose level, 3 patients were required to be treated at that dose level. If 2 of 3 patients developed a DLT effect, no further dose escalation was performed and subsequently enrolled patients were given the lower carmustine dose level. If patients developed grade II dermatitis due to carmustine, a cycle of dosing was held only if the dermatitis covered more than 10% of the body surface area. If dermatitis covered less than 10% of the body surface area, carmustine was still applied, but affected areas were avoided until the dermatitis resolved. For unresolved grade II adverse events (AEs) related to carmustine, the next treatment cycle was held. Once the AEs resolved, carmustine was restarted at a dose 1 level lower. If the AE recurred and remained unresolved after dose reduction, resulting in an additional cycle in a row being held, the patient was withdrawn from the study. Other withdrawal criteria included the decision of patient, protocol-defined progressive disease, more than a 2-week delay for recovery, missing more than 1 cycle of therapy due to a National Cancer Institute CTCAE, version 3.0, grade II AE or worse, missing more than 2 consecutive treatment cycles for any reason, and the investigator’s decision to withdraw the patient.
Acetaminophen, diphenhydramine, prochlorperazine, promethazine hydrochloride, or ondansetron therapy was planned to manage potential pruritus, headache, and vomiting. Triamcinolone, 0.1%, cream or a topical corticosteroid of equivalent potency was used for symptomatic treatment of dermatitis AEs.
At screening, patients received a physical examination and functional status assessment (Eastern Cooperative Oncology Group grading) and the following tests were conducted: complete blood cell count and differential, blood urea nitrogen, creatinine, prothrombin time, calcium, phosphorus, uric acid, alkaline phosphatase, aspartate aminotransferase, total bilirubin, serum electrolytes, 12-lead electrocardiogram, chest radiograph, urinalysis, pulse oximetry, pregnancy test, and pulmonary function testing for predicted diffusing capacity of the lungs for carbon monoxide when needed for patients with significant pulmonary disease determined on history, physical examination, or chest radiograph. For the first 4 weeks of therapy, patients were monitored with weekly complete blood cell count and differential, blood urea nitrogen, creatinine, alkaline phosphatase, total bilirubin, and aspartate aminotransferase. Following the first 4 weeks, the above-mentioned laboratory studies were performed every 2 weeks during treatment visits plus pulse oximetry, serum electrolytes, urinalysis, calcium, phosphorus, uric acid, and prothrombin time. Physical examination was performed and vital signs were measured during each treatment visit. Assessment of target MF skin lesions and SWAT scoring were documented monthly, with the final assessment occurring 2 weeks after the last treatment.
One MF lesion amenable to biopsy was left untreated during the first cycle of therapy (carmustine untreated). For laboratory correlative studies, 6-mm punch samples for biopsies of carmustine-untreated and carmustine-treated MF lesions were obtained 24 or 48 hours after cycle 1 day 1 (patients alternatively assigned to either biopsy schedule as they entered the study). Four-millimeter punch samples for biopsies were obtained in carmustine-untreated MF skin at baseline, carmustine-treated MF skin 24 hours or 48 hours after cycle 1 day 1, and on cycle 1 day 8. Specimens were frozen with liquid nitrogen after embedding in optimal cutting temperature compound (Tissue-Tek O.C.T. compound; Sakura Finetek) and stored in a −80°C freezer. Patients with grade II cutaneous dermatitis AEs also underwent biopsy.
Clinical response was assessed by the modified SWAT.1,4 Photographs of 5 target MF lesions were obtained for each patient and used to assist with response confirmation. Clinical response criteria were complete response (CR), defined as no evidence of disease with 100% improvement for at least 4 weeks; partial response (PR), defined as a 50% or more decrease in the SWAT score compared with baseline and improvement maintained for at least 4 weeks; stable disease (SD), defined as less than 50% decrease in the SWAT score compared with baseline; and progressive disease, defined as an increase of 25% or more of the SWAT score compared with baseline while the patient was actively receiving the study drug. Time to response was months to first achieve protocol-defined CR or PR. Duration of CR was determined as the time interval (months) from the last treatment date to the date of recurrence documented by skin biopsy or the date the patient was last seen in the clinic. The descriptions and grading scales found in the revised National Cancer Institute CTCAE, version 3.0, were used for AE reporting.
The overall response rate (ORR) for patients with CTCL, MF type, was 55% to 80% when treated with conventional topical carmustine therapy as reported in a previous review.6 It was hypothesized that an expected ORR of 89% for O6-benzylguanine with carmustine would be sufficiently promising to justify further testing of the regimen and schedule. In contrast, an ORR of 55% or less would discourage further testing. The optimal 2-stage phase 2 design by Simon7 with a type I error of 0.05 (ie, a significance level P = .05) and a type II error of 0.1 (ie, 90% power) was used on clinical responses. Ten patients were enrolled in stage 1 of the trial. If there were more than 7 patients with CR or PR observed of these 10 patients, 5 additional patients would be enrolled in stage 2. If there were 12 or more responses observed in 15 patients, treatment would be considered to be more than 55% effective based on Simon’s7 design. Furthermore, assuming that 10% of the participants failed to complete the study, a total of 17 patients (to achieve 15 evaluable patients) were enrolled.
The true response (CR or CR plus PR) rate was estimated based on the number of responses using a binomial distribution theory. The rate of disease stabilization was estimated based on the number of CRs, PRs, or SDs. The 95% CIs for response rate and the rate of disease stabilization were estimated using the Wilson method.8 Pretreatment and posttreatment SWAT scores were compared using a 2-tailed, paired t test. Logistic regression was used to identify indicators of CR, and a linear regression model was used to identify the indicators of time to response. All tests were 2-sided and P values ≤.05 were considered statistically significant. All patients who met eligibility criteria and received any treatment during the trial were included in an intent-to-treat analysis of the response rate. Data analysis was performed from May 1, 2014, to December 1, 2015. Statistical testing was conducted using SAS, version 9.4 (SAS Institute Inc).
Twenty-five patients were screened, and 17 were enrolled (Figure 1). The enrollment was discontinued after 17 patients were included since this number was sufficient to achieve the study goals. Mean (SD) age of the patients was 45.2 (14.6) years (Table 1). Patients had experienced 2 to 7 prior treatments. Twelve (71%) participants were male, and 5 (29%) were black. All patients completed at least 3 cycles of treatment and could be assessed for response.
The Case Comprehensive Cancer Center Data Safety and Monitoring Committee reviewed and confirmed responses for all patients. A waterfall plot of tumor reduction after treatment from baseline, based on the SWAT score (Figure 2), demonstrates that 15 of 17 patients achieved CR or PR. This 88% ORR (95% CI, 0.66-0.97) exceeded the predefined primary end point of 55% for study continuance. The CR rate was 41% (7 patients) (95% CI, 0.22-0.64); rate of disease stabilization (CR+PR+SD) was 94% (95% CI, 0.73-0.99). Partial responses were substantial, with 79% to 95% SWAT score reduction (Figure 2). The mean (SD) time to response to attain CR was 3.54 (1.47) months; this result was not significantly different from a slightly shorter time to PR of 2.75 (1.17) months (Table 1). The CTCL burden was significantly reduced from a mean baseline SWAT score of 15.59 (21.83) to a mean posttreatment SWAT score of 2.36 (5.71) (P = .02). The mean reduction in the SWAT score was 86.2% (30.3%). Two patients had unconfirmed responses (Table 1). The first patient had a CR at his last study visit before withdrawal because of unresolved grade II skin pain; although maintenance of CR for 4 weeks could not be established, the patient was assigned a response of CR because these were the only available data. The second patient had significant skin inflammation after 3 cycles and withdrew. On clinic follow-up 1 month after withdrawal, a biopsy was performed to determine whether skin changes represented inflammation or persistent CTCL, and pathologic testing showed no evidence of CTCL. However, this patient was assigned a response of SD. Mean duration of CR was 14.43 (6.6) months after mean of 9 treatment cycles.
Age was somewhat indicative of CR. The odds of having CR (eTable 1 in Supplement 2) were reduced 11% per year increase of age (OR, 0.89; P = .06), and age was significantly associated with time to response (coefficient, −0.06; P = .001). The number of prior treatments was also significantly indicative of time to response (coefficient, −0.36; P = .04). With 1 additional prior treatment, time to response (eTable 2 in Supplement 2) was reduced by 0.36 months (10.8 days). Representative photographs of clinical responses are provided in Figure 3.
Of 166 AEs, 11 (67%) were grade I (eTable 3 in Supplement 2). Grade I AEs with at least 2% frequency were fatigue (18 [11%]), nausea (15 [9%]), hyperpigmentation (13 [8%]), aspartate aminotransferase level elevation (10 [6%]), headache (9 [5%]), pruritus (8 [5%]), skin pain (7 [4%]), uric acid level elevation (5 [3%]), and anemia (4 [2%]). Grade II AEs occurring with a frequency of 2% or more were erythema and desquamation (16 [10%]), skin pain (9 [5%]), headache (7 [4%]), pruritus (6 [4%]), and nausea (4 [2%]). There were no AEs of grade III or worse severity. The MTD was determined to be topical carmustine, 20 mg, applied once in combination with 2 daily doses of O6-benzylguanine, 120 mg/m2.
All 7 patients with CR and 5 of 8 patients (63%) with PR experienced grade II dermatitis (erythema and desquamation, pain, and pruritus). The time of onset of grade II dermatitis was variable, occurring during treatment cycles 1 to 9. Treatment of grade II dermatitis included topical corticosteroids alone or in combination with oral cephalexin or vitamin D, or withholding carmustine per protocol. Additional details regarding patients who obtained CR are reported in Table 2. All 3 patients who received back-to-back carmustine treatment during days 1 and 2 achieved CR but also experienced almost half of the grade II AEs.
Five patients (63%) with dermatitis underwent biopsies with the results showing interface dermatitis with dyskeratotic keratinocytes, macrophages, and some lymphocytes. A representative patient who developed red, painful patches related to carmustine application is shown in eFigure 1 in Supplement 2. Immunostaining revealed prominent macrophage infiltration and activation in both treated nonclinically inflamed MF lesions (eFigure 2B in Supplement 2) and in posttreatment dermatitis (eFigure 2C and D in Supplement 2). Neutrophils were not observed on hematoxylin-eosin staining, so CD11b immunostaining was mainly identifying macrophages and monocytes. The prominent expression of inducible nitric oxide synthase colocalization indicates significant activation by carmustine treatment amplified in the dermatitis AE tissue. Five patients (33%) with grade II dermatitis were unresponsive to topical corticosteroid therapy and all had insufficient or deficient 25 hydroxyvitamin D levels. Dermatitis resolved after administration of oral vitamin D3 (2000-4000 IU/d); 1 patient received a single vitamin D3 dose of 50 000 IU.
In this phase 1-2 trial with dual-dose O6-benzylguanine and carmustine, the 88% ORR was higher compared with the ORR achieved with single-dose O6-benzylguanine with carmustine in the phase 1 study (76%).2 In addition, the dual dose allowed a shorter mean duration of treatment of 18 weeks (28 weeks for single dose), and lower mean total carmustine dose (213.3 mg for dual dose vs 309.5 mg for single dose). Mean time to response was 3.54 months (CR) and 2.75 months (PR), and mean duration of response was 14.43 months. Responses to dual-dose O6-benzylguanine with carmustine therapy compare favorably with 2 US Food and Drug Administration–approved CTCL topical therapies: mechlorethamine hydrochloride and bexarotene gel. In a randomized, controlled, observer-blinded, phase 2 trial, 260 patients with early-stage CTCL, MF type, were randomized to receive mechlorethamine, 0.02%, either gel or ointment.9 Overall response for mechlorethamine gel ranged from 47% to 77% depending on the clinical end point used—either Composite Assessment of Index Lesion Disease Severity or modified SWAT score—as well as the patient population (intent-to-treat vs protocol-defined efficacy evaluable). Time to response, defined as estimated time to a 50% response rate, was 6.5 (gel) and 10.5 months (ointment). At 12 months (end of the trial), 86% of patients receiving the gel and 82% of patients receiving the ointment maintained their response.
Two open-label trials of bexarotene gel examined ORR, time of onset to response, and duration of response. In the phase 1-2 dose-escalation trial, 42 of 67 (63%) adult patients with stage IA or stage IIA CTCL, MF type, had a CR and PR based on the Physician’s Global Assessment.10 Kaplan-Meier estimated time to onset of response with 50% or more improvement (PR) was a median of 5 months (range, 1-21.5 months) and median duration of response from the beginning of therapy to relapse time was 25 months (range, 5-25 months). In the phase 3 trial of bexarotene, 1%, gel, 27 of 50 (54%) patients with stage IA or stage IIB CTCL, MF type, had a CR or PR based on 2 primary efficacy end points: the Physician’s Global Assessment and Composite Assessment of Index Lesion Disease Severity.11 Time to response with 50% or more improvement (PR) for these 27 patients was a median of 5 months (range, 1-17 months). Duration of response (time to relapse) ranged from 3 to 10 months.
In the present study, older patients had reduced odds of achieving a CR but earlier time to response. Reasons for these seemingly opposite findings are unclear. Patients who had more prior treatments had a decreased time to response instead of an increased time to response that is more characteristic of refractory disease.
Dual-dose O6-benzylguanine with carmustine was well tolerated. Dermatitis AEs occurred more frequently among patients with grade II AEs. Macrophage infiltration and hyperactivation may contribute to carmustine toxic effects given the colocalization of CD11b and iNOS on confocal microscopy.
Five patients with grade II dermatitis that was unresponsive to topical corticosteroid therapy were vitamin D insufficient or deficient. Dermatitis resolved with vitamin D3 supplementation. The link between low vitamin D levels and carmustine-related dermatitis is unknown. Vitamin D therapy may reduce inflammatory responses according to prior evidence and reports that vitamin D can modulate the innate and adaptive responses, playing a role in chronic illnesses.12 Among 9 patients with psoriasis and 16 individuals with vitiligo with low baseline levels of 25 hydroxyvitamin D, 6 months of vitamin D3 therapy (35 000 IU/d) resulted in significant improvement of Psoriasis Area and Severity Index scores and 25% to 75% repigmentation of vitiligo lesions without toxic effects.13 In addition, vitamin D administration was shown to suppress hyperactive macrophage-mediated skin inflammation induced by a similar alkylating agent: nitrogen mustard.14 It is not possible to know for certain that vitamin D therapy vs healing with time alone resulted in resolution of grade II dermatitis experienced by 5 patients.
There are concerns for risk of development of squamous and basal cell carcinomas in both sun-exposed and sun-protected skin of patients with MF treated with topical nitrogen mustard.15- 17 However, among 143 patients with CTCL treated with topical carmustine, no skin cancers related to therapy were identified (follow-up, 6 months to 12.9 years).3
This study is limited by a small sample size as well as the unavailability of AGT assays, so the effects of dual-dose O6-benzylguanine administration on the extent and duration of AGT depletion are still unknown. The practicality of this therapy may be enhanced by developing alternatives to facility-based intravenous O6BG administration.
Dual-dose O6-benzylguanine with carmustine for early-stage CTCL, MF type, resulted in a high ORR (88%) with a fairly durable mean CR duration (14.43 months). Compared with single-dose O6-benzylguanine with topical carmustine, dual-dose O6-benzylguanine with carmustine resulted in a shorter duration of treatment, lower mean total carmustine dose, and more cutaneous AEs. Alternative formulations of O6-benzylguanine (oral vs patch) for ease of administration should be examined. Furthermore, a clinical trial can be designed to determine the efficacy of high-dose oral vitamin D3 treatment on suppressing inflammatory skin responses in patients with CTCL exposed to nitrosoureas. Carmustine may exhibit a unique mechanism of toxicity, which may be macrophage hyperactivation.
Accepted for Publication: December 3, 2016.
Corresponding Author: Joselin D. Tacastacas, MD, Department of Dermatology, University Hospitals Cleveland Medical Center, Lakeside 3500, 11100 Euclid Ave, Cleveland, OH 44106 (email@example.com).
Published Online: February 15, 2017. doi:10.1001/jamadermatol.2016.5793
Author Contributions: Drs Tacastacas and Cooper 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: Gerson, Fu, Lu, McCormick, Cooper.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Tacastacas, Dowlati, Lu.
Critical revision of the manuscript for important intellectual content: Chan, Carlson, Gerson, Dowlati, Fu, Lu, Groft, Rosenjack, Honda, McCormick, Cooper.
Statistical analysis: Fu, Lu.
Obtained funding: Gerson, Lu, Cooper.
Administrative, technical, or material support: Tacastacas, Gerson, Dowlati, Lu, Groft, Rosenjack, Cooper.
Study supervision: Gerson, Dowlati, Lu, Honda, McCormick, Cooper.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by the National Cancer Institute/Investigational Drug Branch at the National Institutes of Health who provided O6-benzylguanine; Bristol-Myers Squibb and Heritage Pharmaceuticals, who provided carmustine (BiCNU) without charge; National Institutes of Health/National Institute of Arthritis and Musculoskeletal and Skin Diseases grant 5U01AR064144-02 (Dr Lu); and the National Cancer Institute at the National Institutes of Health grant 5R21CA115057-2 (Dr Cooper).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: S. Percy Ivy, MD (National Cancer Institute/Investigational Drug Branch, National Institutes of Health), provided O6-benzylguanine without charge.