Significant morbidity and expense result from frequent recurrences of non-muscle invasive bladder cancer (NMIBC) after standard treatment, and carcinoma in situ (Tis) is a poor prognostic factor. Predicated on observational and preclinical data strongly supporting cyclooxygenase-2 (COX-2) in the pathogenesis, and the activity of COX-2 inhibitors, in bladder cancer, we conducted a randomized, double-blind, placebo-controlled trial to determine if celecoxib could reduce the time-to-recurrence (TTR) in NMIBC patients at high risk for recurrence. 146 patients were randomized to celecoxib (200 mg) or placebo orally twice daily for at least 12 months. The average treatment duration was 1.25 years. Primary intent-to-treat analysis revealed celecoxib did not statistically significantly prolong TTR compared with placebo (P = 0.17, log-rank). With a median follow-up of 2.49 years the relative risk of recurrence in the celecoxib vs placebo arms was 0.64 (95% CI, 0.38, 1.17). The recurrence-free rate at 12 months with celecoxib was 88% (95% CI, 0.81,0.96) versus 78% (95% CI, 0.69, 0.89) with placebo. After controlling for covariates with Cox regression analysis, recurrence rates did not differ between the two study arms (HR = 0.69; 95% CI, 0.37,1.29). Celecoxib had a marginally significant effect on reducing metachronous recurrences (vs. placebo) with hazard ratio of 0.56 (95% CI, 0.3,1.06; P=0.075). Celecoxib was well tolerated, with similar adverse events and quality-of-life in both arms. Our clinical trial results do not show a clinical benefit for celecoxib in preventing NMIBC recurrence but further investigation of COX-2 inhibitors in this setting is warranted.
transitional cell carcinoma; non-muscle invasive bladder cancer; chemoprevention; celecoxib; COX-2 inhibitor
A challenge for implementing performance based Bayesian sample size determination is selecting which of several methods to use. We compare three Bayesian sample size criteria: the average coverage criterion (ACC) which controls the coverage rate of fixed length credible intervals over the predictive distribution of the data, the average length criterion (ALC) which controls the length of credible intervals with a fixed coverage rate, and the worst outcome criterion (WOC) which ensures the desired coverage rate and interval length over all (or a subset of) possible datasets. For most models, the WOC produces the largest sample size among the three criteria, and sample sizes obtained by the ACC and the ALC are not the same. For Bayesian sample size determination for normal means and differences between normal means, we investigate, for the first time, the direction and magnitude of differences between the ACC and ALC sample sizes. For fixed hyperparameter values, we show that the difference of the ACC and ALC sample size depends on the nominal coverage, and not on the nominal interval length. There exists a threshold value of the nominal coverage level such that below the threshold the ALC sample size is larger than the ACC sample size, and above the threshold the ACC sample size is larger. Furthermore, the ACC sample size is more sensitive to changes in the nominal coverage. We also show that for fixed hyperparameter values, there exists an asymptotic constant ratio between the WOC sample size and the ALC (ACC) sample size. Simulation studies are conducted to show that similar relationships among the ACC, ALC, and WOC may hold for estimating binomial proportions. We provide a heuristic argument that the results can be generalized to a larger class of models.
average coverage criterion; average length criterion; credible interval; coverage rate; interval length; worst outcome criterion
Oral premalignant lesions (OPLs) are precursors of oral squamous cell carcinoma (OSCC). Short telomeres in peripheral blood leukocytes are associated with increased risks of several cancers. However, whether short leukocyte telomere length (LTL) predisposes to OPL and OSCC is unclear.
LTLs were measured in PBLs of 266 patients with OPL (N=174) or OSCC (N=92) at diagnosis and 394 age- and gender-matched control subjects. The association between LTL and OPL or OSCC risk, as well as the interaction of telomere length, cigarette smoking and alcohol drinking on OPL or OSCC risk were analyzed.
The age-adjusted relative LTL was the shortest in OSCC (1.64±0.29), intermediate in OPL (1.75±0.43), and longest in controls (1.82±0.36) (P for trend < 0.001). When dichotomized at the median value in controls, adjusting for age, gender, smoking and alcohol drinking status, the odds ratio (OR) for OPL and OSCC risks associated with short LTL was 2.03 (95% CI = 1.29–3.21) and 3.47 (95% CI = 1.84–6.53), respectively, with significant dose-response effects for both associations. Among 174 OPL patients, 23 progressed to OSCC and the mean LTL was shorter than in progressors than non-progressors (1.66±0.35 vs. 1.77±0.44), although the difference did not reach statistical significance (P=0.258) likely due to the small number of progressors. Interaction analysis shows that short LTL, smoking, and alcohol drinking are independent risk factors for OPL and OSCC.
Short LTL is associated with increased risks of developing OPL and OSCC and likely predisposes to the malignant progression of OPL patients.
Telomere length; peripheral blood leukocyte; oral premalignant lesion; oral squamous cell carcinoma; smoking; alcohol drinking
To report the clinical efficacy of sorafenib and to evaluate biomarkers associated with sorafenib clinical benefit in the BATTLE program.
Patients and Methods
Patients with previously treated non-small–cell lung cancer (NSCLC) received sorafenib until progression or unacceptable toxicity. Eight-week disease control rate (DCR), progression-free survival (PFS), and overall survival (OS) were assessed. Prespecified biomarkers included K-RAS, EGFR, and B-RAF mutations, and EGFR gene copy number. Gene expression profiles from NSCLC cell lines and patient tumor biopsies with wild-type EGFR were used to develop a sorafenib sensitivity signature (SSS).
105 patients were eligible and randomized to receive sorafenib. Among 98 patients evaluable for 8-week DCR, the observed DCR was 58.2%. The median PFS and OS were 2.83 (95% confidence interval [CI], 2.04-3.58) and 8.48 months (95% CI, 5.78-10.97), respectively. Eight-week DCR was higher in patients with wt-EGFR than patients with EGFR mutation (P=0.012), and in patients with EGFR gene copy number gain (FISH positive) versus patients FISH negative (P=0.048). In wt-EGFR tumors, the SSS was associated with improved PFS (median PFS 3.61 months in high SSS versus 1.84 months in low SSS, P=0.026) but not with 8-week DCR. Increased expression of fibroblast growth factor-1, NF-kB and hypoxia pathways were identified potential drivers of sorafenib resistance.
Sorafenib demonstrates clinical activity in NSCLC, especially with wt-EGFR. SSS was associated with improved PFS. These data identify subgroups that may derive clinical benefit from sorafenib and merit investigation in future trials. ClinicalTrials.gov: NCT00411671.
multikinase inhibitor; non–small cell lung cancer; sorafenib; biomarkers; targeted treatment
We propose a randomized phase II clinical trial design based on Bayesian adaptive randomization and predictive probability monitoring. Adaptive randomization assigns more patients to a more efficacious treatment arm by comparing the posterior probabilities of efficacy between different arms. We continuously monitor the trial by using the predictive probability. The trial is terminated early when it is shown that one treatment is overwhelmingly superior to others or that all the treatments are equivalent. We develop two methods to compute the predictive probability by considering the uncertainty of the sample size of the future data. We illustrate the proposed Bayesian adaptive randomization and predictive probability design by using a phase II lung cancer clinical trial, and we conduct extensive simulation studies to examine the operating characteristics of the design. By coupling adaptive randomization and predictive probability approaches, the trial can treat more patients with a more efficacious treatment and allow for early stopping whenever sufficient information is obtained to conclude treatment superiority or equivalence. The design proposed also controls both the type I and the type II errors and offers an alternative Bayesian approach to the frequentist group sequential design.
Adaptive randomization; Bayesian inference; Clinical trial ethics; Group sequential method; Posterior predictive distribution; Randomized trial; Type I error; Type II error
Although the frequentist paradigm has been the predominant approach to clinical trial design since the 1940s, it has several notable limitations. The alternative Bayesian paradigm has been greatly enhanced by advancements in computational algorithms and computer hardware. Compared to its frequentist counterpart, the Bayesian framework has several unique advantages, and its incorporation into clinical trial design is occurring more frequently. Using an extensive literature review to assess how Bayesian methods are used in clinical trials, we find them most commonly used for dose finding, efficacy monitoring, toxicity monitoring, diagnosis/decision making, and for studying pharmacokinetics/pharmacodynamics. The additional infrastructure required for implementing Bayesian methods in clinical trials may include specialized software programs to run the study design, simulation, and analysis, and Web-based applications, which are particularly useful for timely data entry and analysis. Trial success requires not only the development of proper tools but also timely and accurate execution of data entry, quality control, adaptive randomization, and Bayesian computation. The relative merit of the Bayesian and frequentist approaches continues to be the subject of debate in statistics. However, more evidence can be found showing the convergence of the two camps, at least at the practical level. Ultimately, better clinical trial methods lead to more efficient designs, lower sample sizes, more accurate conclusions, and better outcomes for patients enrolled in the trials. Bayesian methods offer attractive alternatives for better trials. More such trials should be designed and conducted to refine the approach and demonstrate its real benefit in action.
adaptive trial design; Bayesian paradigm; clinical trial conduct; frequentist paradigm; trial efficiency; trial ethics
Enhancer of zeste homolog 2 (EZH2) promotes carcinogenesis by epigenetically silencing tumor suppressor genes. We studied EZH2 expression by immunohistochemistry in a large series of non-small cell lung carcinomas (NSCLC) in association with tumor characteristics and patient outcomes.
EZH2 immunohistochemistry expression was analyzed in 265 normal and premalignant bronchial epithelia, 541 primary NSCLCs [221 squamous cell carcinomas (SCCs) and 320 adenocarcinomas] and 36 NSCLCs with paired brain metastases. An independent set of 91 adenocarcinomas was also examined. EZH2 expression was statistically correlated with clinico-pathological information, and EGFR/KRAS mutation status.
EZH2 expression was significantly (P<0.0001) higher in SCCs compared to adenocarcinomas and in brain metastasis relative to matched primary tumors (P=0.0013). EZH2 expression was significantly (P<0.0001) elevated in bronchial preneoplastic lesions with increasing severity. In adenocarcinomas, higher EZH2 expression significantly correlated with younger age, cigarette smoking and higher TNM stage (P=0.02 to P<0.0001). Higher EZH2 expression in adenocarcinoma was associated with worse recurrence-free survival (RFS; P=0.025; HR 1.54) and overall survival (OS; P=0.0002; HR 1.96). Furthermore, lung adenocarcinomas with low EZH2 levels and high expression of the lineage-specific transcription factor, TTF-1, exhibited significantly improved RFS (P=0.009; HR 0.51) and OS (P=0.0011; HR 0.45) which was confirmed in the independent set of 91 adenocarcinomas.
In lung, EZH2 expression is involved in early pathogenesis of SCC and correlates with a more aggressive tumor behavior of adenocarcinoma. When EZH2 and TTF-1 expressions are considered together, they serve as a prognostic marker in patients with surgically resected lung adenocarcinomas.
EZH2; NSCLC; lung adenocarcinoma; lung squamous cell carcinoma; bronchial preneoplasia; brain metastasis; KRAS mutations; EGFR mutations
Target-matched treatment with PI3K/AKT/mTOR pathway inhibitors in patients with diverse advanced cancers with PIK3CA mutations have shown promise. Tumors from patients with colorectal cancer (CRC) were analyzed for PIK3CA, KRAS and BRAF mutations. PIK3CA mutated tumors were treated, whenever feasible, with agents targeting the PI3K/AKT/mTOR pathway. Of 194 patients analyzed, 31 (16%) had PIK3CA mutations, and 189 (97%) were assessed for KRAS mutations. Patients with PIK3CA mutations had a higher prevalence of simultaneous KRAS mutations than patients with wild-type (wt) PIK3CA (71%, 22/31 vs. 43%, 68/158; p=0.006). Of 31 patients with PIK3CA mutations, 17 (55%) were treated with protocols containing PI3K/AKT/mTOR pathway inhibitors (median age, 57; median number of prior therapies, 4; mTORC1 inhibitors , PI3K inhibitors  or an AKT inhibitor ). None (0/17) had a partial or complete response (PR/CR) and only 1 (6%, 95% CI 0.01–0.27) had stable disease (SD)≥6 months, which was not significantly different from a SD≥6 month/PR/CR rate of 16% (11/67; 95% CI 0.09–0.27) in CRC patients without PIK3CA mutations treated with PI3K/AKT/mTOR pathway inhibitors (p=0.44). Median progression-free survival was 1.9 months (95% CI 1.5–2.3). In conclusion, our data provide preliminary evidence that in heavily pretreated patients with PIK3CA-mutant advanced CRC, protocols incorporating PI3K/AKT/mTOR inhibitors have minimal activity. PIK3CA mutations are associated with simultaneous KRAS mutations, possibly accounting for therapeutic resistance.
PIK3CA mutation; PI3K/AKT/mTOR; Colorectal cancer; Target-matched therapy; KRAS mutation; Phase I trial
Epidemiologic and preclinical data support the oral-cancer prevention potential of green tea extract (GTE). We randomly assigned patients with high-risk oral premalignant lesions (OPLs) to receive GTE at 500 mg/m2, 750 mg/m2, or 1000 mg/m2 or placebo TID for 12 weeks, evaluating biomarkers in baseline and 12-week biopsies. The OPL clinical response rate was higher in all GTE arms (n=28; 50%) versus placebo (n=11; 18.2%; p=0.09) but did not reach statistical significance. However, the 2 higher-dose GTE arms (58.8% [750, 1000 mg/m2], 36.4% [500 mg/m2], and 18.2%, [placebo], p=0.03) had higher responses, suggesting a dose-response effect. GTE treatment also improved histology (21.4% versus 9.1%, p=0.65), though not statistically significant. GTE was well-tolerated although higher doses increased insomnia/nervousness but produced no grade-4 toxicity. Higher mean baseline stromal VEGF correlated with a clinical (p=0.04) but not histologic response. Baseline scores of other biomarkers (epithelial VEGF, p53, Ki-67, cyclin D1, and p16 promoter methylation) were not associated with a response or survival. Baseline p16 promoter methylation (n=5) was associated with a shorter cancer-free survival. Stromal VEGF and cyclin D1 expression were downregulated in clinically responsive GTE patients and upregulated in non-responsive patients at 12 weeks (versus at baseline). An extended (median 27.5 months) follow-up showed a median time to oral cancer of 46.4 months. GTE may suppress OPLs, in part through reducing angiogenic stimulus (stromal VEGF). Higher doses of GTE may improve short-term (12 week) OPL outcome. The present results support longer-term clinical testing of GTE for oral cancer prevention.
green tea extract; oral premalignant lesions; chemoprevention
In a previous trial, we found that combined 13-cis retinoic acid (13-cRA), interferon-α and α-tocopherol more effectively reversed advanced premalignant lesions of the larynx than of the oral cavity and that cyclin D1 (CD1)G/A870 single nucleotide polymorphism correlated with cancer risk. We conducted the present trial primarily to confirm the clinical activity of the combination in advanced laryngeal premalignancy and to confirm and extend our findings on CD1, both genotype and protein expression, in association with cancer risk in this setting. Twenty-seven moderate-to-severe laryngeal dysplasia patients underwent induction with combined 13-cRA daily, α-interferon twice weekly, and α-tocopherol daily for one year; 14 non-progressing patients then were randomized to maintenance fenretinide or placebo for two years. During induction, 2 patients had pathological complete responses, 6 had partial responses (30% overall response rate), and 5 developed laryngeal cancer. There were no significant differences between maintenance fenretinide and placebo in response or cancer rates. Ten patients developed cancer overall. Twenty-four patients were evaluated for the CD1 G/A870 genotype, and 23 for pre- and post-treatment CD1 protein expression. Consistent with our earlier report, shorter cancer-free survival was associated with the CD1 AA/AG genotype (p = 0.05). Extending our earlier work, high CD1 expression was associated with worse cancer-free survival overall (p= 0.04) and within each CD1 genotype group. These findings support CD1 genotype and protein expression as important risk markers for laryngeal cancer and suggest future trials targeting upstream regulators of CD1 transcription.
Premalignant lesions; larynx; biochemoprevention; cyclin D1 genotype; cyclin D1 protein expression
Outcome-adaptive randomization (AR) allocates more patients to the better treatments as the information accumulates in the trial. Is it worth to apply outcome-AR in clinical trials? Different views permeate the medical and statistical communities. We provide additional insights to the question by conducting extensive simulation studies. Trials are designed to maintain the type I error rate, achieve a specified power, and provide better treatment to patients. Generally speaking, equal randomization (ER) requires a smaller sample size and yields a smaller number of non-responders than AR by controlling type I and type II errors. Conversely, AR produces a higher overall response rate than ER with or without expanding the trial to the same maximum sample size. When there exist substantial treatment differences, AR can yield a higher overall response rate as well as a lower average sample size and a smaller number of non-responders. Similar results are found for the survival endpoint. The differences between AR and ER quickly diminish with early stopping of a trial due to efficacy or futility. In summary, ER maintains balanced allocation throughout the trial and reaches the specified statistical power with a smaller number of patients in the trial. If the trial’s result is positive, ER may lead to early approval of the treatment. AR focuses on treating patients best in the trial. AR may be preferred when the difference in efficacy between treatments is large or when limited patients are available.
Adaptive and fixed randomization; Bayesian clinical trial design; Efficacy and futility early stopping; Type I error and statistical power; Patient population; Sample size
Clinical trial designs for targeted therapy development are progressing toward the goal of personalized medicine. Motivated by the need of ongoing efforts to develop targeted agents for lung cancer patients, we propose a Bayesian two-step Lasso procedure for biomarker selection under the proportional hazards model. We seek to identify the key markers that are either prognostic or predictive with respect to treatment from a large number of biomarkers. In the first step of our two-step strategy, we use the Bayesian group Lasso to identify the important marker groups, wherein each group contains the main effect of a single marker and its interactions with treatments. Applying a loose selection criterion in the first step, the goal of first step is to screen out unimportant biomarkers. In the second step, we zoom in to select the individual markers and interactions between markers and treatments in order to identify prognostic or predictive markers using the Bayesian adaptive Lasso. Our strategy takes a full Bayesian approach and is built upon rapid advancement of Lasso methodologies with variable selection. The proposed method is generally applicable to the development of targeted therapies in clinical trials. Our simulation study demonstrates the good performance of the two-step Lasso: Important biomarkers can typically be selected with high probabilities, and unimportant markers can be effectively eliminated from the model.
Adaptive Lasso; Bayesian Lasso; Clinical trials; Group Lasso; Proportional hazards model; Targeted therapy design; Variable selection
Lung cancer is the leading cancer cause of mortality worldwide; large-scale trials have failed to improve clinical outcomes of patients with chemorefractory non-small-cell lung cancer (NSCLC).
Following an initial equal randomization period, BATTLE adaptively randomized patients with chemorefractory NSCLC to erlotinib, vandetanib, erlotinib plus bexarotene, or sorafenib based on molecular biomarkers of NSCLC pathogenesis in fresh core needle biopsy specimens. The primary end point was disease control rate (DCR) at 8 weeks.
Of 255 patients randomly assigned to erlotinib (59 patients), vandetanib (54), erlotinib plus bexarotene (37), and sorafenib (105), 244 were eligible for the DCR analysis. Pneumothorax after lung biopsy occurred in 11.5% and treatment-related toxicities grade 3–4 in 6.5% of patients. Overall results were a 46% 8-week DCR, 1.9-month median progression-free survival, 9-month median overall survival, and 35% 1-year survival. Individual markers predicting a significantly superior DCR for a treatment included: epidermal growth factor receptor (EGFR) mutation (P=0.04) for erlotinib; cyclin D1 positivity (P=0.01) or EGFR amplification (P=0.006) for erlotinib plus bexarotene; vascular endothelial growth factor receptor 2 positivity (P=0.05) for vandetanib; and absence of EGFR mutation (P=0.01) or of EGFR high polysomy (P=0.05) for sorafenib. A better 8-week DCR occurred with sorafenib versus all other regimens (64% versus 33%; P<0.001) among EGFR wild-type patients and versus all other regimens (61% versus 32%; P=0.11) among mutant-KRAS patients. The prespecified biomarker groups were less predictive than the individual biomarkers analyzed in this study.
The first completed biopsy-mandated study in pretreated NSCLC, BATTLE confirmed our pre-specified hypotheses regarding biomarker and targeted treatment interactions, establishing a new paradigm for personalizing therapy for patients with NSCLC. (ClinicalTrials.gov numbers, NCT00409968, NCT00411671, NCT00411632, NCT00410059, NCT00410189.)
We review the semiparametric approach previously proposed by Kong and Lee and extend it to a case in which the dose-effect curves follow the Emax model instead of the median effect equation. When the maximum effects for the investigated drugs are different, we provide a procedure to obtain the additive effect based on the Loewe additivity model. Then, we apply a bivariate thin plate spline approach to estimate the effect beyond additivity along with its 95% point-wise confidence interval as well as its 95% simultaneous confidence interval for any combination dose. Thus, synergy, additivity, and antagonism can be identified. The advantages of the method are that it provides an overall assessment of the combination effect on the entire two-dimensional dose space spanned by the experimental doses, and it enables us to identify complex patterns of drug interaction in combination studies. In addition, this approach is robust to outliers. To illustrate this procedure, we analyzed data from two case studies.
Additivity; Antagonism; Bivariate splines; Combination therapy; Emax model; the Loewe additivity model; Synergy; Review
Delay in the outcome variable is challenging for outcome-adaptive randomization, as it creates a lag between the number of subjects accrued and the information known at the time of the analysis. Motivated by a real-life pediatric ulcerative colitis trial, we consider a case where a short-term predictor is available for the delayed outcome. When a short-term predictor is not considered, studies have shown that the asymptotic properties of many outcome-adaptive randomization designs are little affected unless the lag is unreasonably large relative to the accrual process. These theoretical results assumed independent identical delays, however, whereas delays in the presence of a short-term predictor may only be conditionally homogeneous. We consider delayed outcomes as missing and propose mitigating the delay effect by imputing them. We apply this approach to the doubly adaptive biased coin design (DBCD) for motivating pediatric ulcerative colitis trial. We provide theoretical results that if the delays, although non-homogeneous, are reasonably short relative to the accrual process similarly as in the iid delay case, the lag is also asymptotically ignorable in the sense that a standard DBCD that utilizes only observed outcomes attains target allocation ratios in the limit. Empirical studies, however, indicate that imputation-based DBCDs performed more reliably in finite samples with smaller root mean square errors. The empirical studies assumed a common clinical setting where a delayed outcome is positively correlated with a short-term predictor similarly between treatment arm groups. We varied the strength of the correlation and considered fast and slow accrual settings.
outcome adaptive randomization; delayed outcome; doubly adaptive biased coin design; imputation
Investigate the mechanisms of regulation and role associated with EZH2 expression in lung cancer cells.
We investigated the mechanisms of EZH2 expression associated with the vascular endothelial growth factor (VEGF)/VEGF receptor 2 (VEGFR-2) pathway. Furthermore, we sought to determine the role of EZH2 in response of lung adenocarcinoma to platinum-based chemotherapy, as well as the effect of EZH2 depletion on VEGFR-2–targeted therapy in lung adenocarcinoma cell lines. Additionally, we characterized EZH2 expression in lung adenocarcinoma specimens and correlated it with patients’ clinical characteristics.
In this study, we demonstrate that VEGF/VEGFR-2 activation induces expression of EZH2 through the upregulation of E2F3 and HIF-1α, and downregulated expression of miR-101. EZH2 depletion by treatment with 3-deazaneplanocin A and knockdown by siRNA decreased the expression of EZH2 and H3K27me3, increased PARP-C level, reduced cell proliferation and migration, and increased sensitivity of the cells to treatment with cisplatin and carboplatin. Additionally, high EZH2 expression was associated with poor overall survival in patients who received platinum-based adjuvant therapy, but not in patients who did not receive this therapy. Furthermore, we demonstrated for the first time that the inhibition of EZH2 greatly increased the sensitivity of lung adenocarcinoma cells to the anti-VEGFR-2 drug AZD2171.
Our results suggest that VEGF/VEGFR-2 pathway plays a role in regulation of EZH2 expression via E2F3, HIF-1α and miR-101. EZH2 depletion decreases the malignant potential of lung adenocarcinoma and sensitivity of the cells to both platinum-based and VEGFR-2–targeted therapy.
EZH2; NSCLC; VEGF/VEGFR-2 pathway; DZNep
Both the lungs and oral cavity are exposed to tobacco carcinogens in smokers. We hypothesized that the oral epithelium undergoes molecular alterations similar to those in lungs and therefore may be used as a surrogate tissue to assess tobacco-induced molecular alterations.
Promoter methylation of p16 and FHIT genes was analyzed with methylation-specific PCR in 1,774 oral and bronchial brush specimens (baseline and 3 months after intervention) from 127 smokers enrolled in a prospective randomized placebo-controlled chemoprevention trial. The association between methylation patterns in oral tissues and bronchial methylation indices (methylated sites/total sites per subject) was analyzed blindly.
At baseline, promoter methylation was observed in 23%, 17%, and 35% of the bronchial tissues for p16, FHIT, and either of the two genes, respectively, which were comparable to the 19%, 15%, and 31% observed in the oral tissues. Among the 125 individuals with available data from both oral and bronchial tissues, strong correlations were observed between tissues from the two sites (P<0.0001 for both p16 and FHIT). Among the 39 individuals with oral tissue methylation in either of the two genes, the mean bronchial methylation index was 0.53 (± 0.29) compared with only 0.27 (± 0.26) for the 86 subjects without oral tissue methylation (P<0.0001). Similar correlations were also observed in samples obtained at 3 months after chemopreventive intervention.
The oral epithelium may be used as a surrogate tissue to assess tobacco-induced molecular damage in lungs, which has an important implication in conducting biomarker-based lung cancer prevention trials.
Recent advances in biotechnology have led to discoveries resulting in major improvements in the therapy of refractory malignancies, although most advanced cancers remain incurable. Thus, there is global consensus around the need to streamline the drug approval process for effective agents. Accelerated approval and Breakthrough Therapy Designation hold the promise of making new treatments available sooner through the use of smaller studies employing intermediate endpoints. Here we consider the inherent limitations of smaller studies and discuss strategies for hastening oncology drug development while maintaining high efficacy standards.
In a previous phase II trial, we demonstrated that fenretinide 200 mg/day had limited activity in retinoid-refractory leukoplakia (34% response rate), possibly due to the lack of achievement of high serum levels which would be required to elicit retinoid-receptor independent apoptosis in pre-malignant cells. We therefore designed this single-arm, phase II trial to investigate whether fenretinide at a higher dose would improve the leukoplakia response rate from our previous study’s historical control.
Patients and Methods
Patients with high-risk leukoplakia were treated with 4 three-week cycles of fenretinide (900 mg/m2 orally twice daily, days 1–7). At week 12, objective clinical responses were determined and blood samples were collected for serum drug level assessment. The original sample size was 25 patients to detect a 55% response rate (90% power, one-sided 10% type I error rate). A futility analysis was planned after accrual of the first 16 patients to allow for early trial closure if ≤4 patients responded.
Fenretinide was well tolerated, with only one grade 3 toxicity (diarrhea) observed. However, only 3 of the initial 15 patients (20%) had a partial response, leading to early trial termination due to lack of efficacy. Serum levels of fenretinide rose from 0 (baseline) to 0.122 μM ± 0.093 (week 12), indicating that high serum levels of the drug were achieved during the initial days of the cycle.
Despite high serum levels, fenretinide for oral leukoplakia, at the dose and schedule studied herein, failed to improve historical response rates.
fenretinide; oral pre-malignant lesion; leukoplakia; chemoprevention
EGFR and Src are frequently activated in non-small cell lung cancer (NSCLC). In preclinical models, combining EGFR and Src inhibition has additive synergistic effects. We conducted a phase I/II trial of the combination of Src inhibitor dasatinib with EGFR inhibitor erlotinib to determine the maximum tolerated dose (MTD), pharmacokinetic drug interactions, biomarkers, and efficacy in NSCLC.
The phase I 3+3 dose-escalation study enrolled patients with solid tumors to determine the MTD. The phase II trial enrolled patients with advanced NSCLC who had undergone no previous treatments to determine progression-free survival (PFS) and response. Pharmacokinetic and tissue biomarker analyses were performed.
MTD was 150 mg of erlotinib and 70 mg of dasatinib daily based on 12 patients treated in the phase I portion. No responses were observed in phase I. The 35 NSCLC patients treated in phase II had an overall disease control rate of 59% at 6 weeks. Five patients (15%) had partial responses; all had activating EGFR mutations. Median PFS was 3.3 months. Epithelial-mesenchymal transition markers did not correlate with outcomes.
The combination of erlotinib and dasatinib is safe and feasible in NSCLC. The results of this study do not support use of this combination in molecularly unselected NSCLC.
Lung cancer is the leading cause of cancer death, developing over prolonged periods through genetic and epigenetic changes induced and exacerbated by tobacco exposure. Many epigenetic changes including DNA methylation and histone methylation and acetylation are reversible, and agents that can modulate these aberrations are a potentially effective approach to cancer chemoprevention. Combined epigenetic-targeting agents have gained interest for their potential to increase efficacy and lower toxicity. The present study applied recently developed statistical methods to validate the combined effects of the demethylating agent 5-aza-2-deoxycytidine (5-AZA-CdR, or AZA, or decitabine) and the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA, or vorinostat). This validation compared AZA alone with SAHA alone and with their combinations (at later or earlier time points and in varying doses) for inhibiting the growth of cell lines of an in vitro lung carcinogenesis system. This system comprises isogenic premalignant and malignant cells that are immortalized (earlier premalignant), transformed (later premalignant), and tumorigenic human bronchial epithelial (HBE) cells (immortalized BEAS-2B and its derivatives 1799 [immortalized], 1198 [transformed], and 1170-I [tumorigenic]). AZA alone and SAHA alone produced a limited (< 50%) inhibition of cell growth, whereas combined AZA and SAHA inhibited cell growth more than did either agent alone, reaching 90% inhibition under some conditions. Results of drug-interaction analyses in the Emax model and semiparametric model supported the conclusion that the drug combinations exert synergistic effects, i.e., beyond additivity in the Loewe model. The present results demonstrate the applicability of our novel statistical methodology for quantitatively assessing drug synergy across a wide range of doses of agents with complex dose-response profiles, a methodology with great potential for advancing the development of chemopreventive combinations.
lung cancer; epigenetics; vorinostat; decitabine; premalignant; epithelial cells
The ETS2 transcription factor is an evolutionarily conserved gene that is deregulated in cancer. We analyzed the transcriptome of lung adenocarcinomas and normal lung tissue by expression profiling and found that ETS2 was significantly down-regulated in adenocarcinomas. In this study, we probed the yet unknown functional role of ETS2 in lung cancer pathogenesis.
Lung adenocarcinomas (n=80) and normal lung tissues (n=30) were profiled using the Affymetrix Human Gene 1.0 ST platform. Immunohistochemical (IHC) analysis was performed to determine ETS2 protein expression in NSCLC histological tissue specimens (n=201). Patient clinical outcome, based on ETS2 IHC expression, was statistically assessed using the log-rank and Kaplan-Meier tests. RNA interference and over-expression strategies were employed to assess effects of ETS2 expression on the transcriptome and on various malignant phenotypes.
ETS2 expression was significantly reduced in lung adenocarcinomas compared to normal lung (p<0.001). Low ETS2 IHC expression was a significant predictor of shorter time to recurrence in NSCLC (p=0.009, HR=1.89) and adenocarcinoma (p=0.03, HR=1.86). Moreover, ETS2 was found to significantly inhibit lung cancer cell growth, migration and invasion (p<0.05), and microarray and pathways analysis revealed significant (p<0.001) activation of the HGF pathway following ETS2 knockdown. In addition, ETS2 was found to suppress MET phosphorylation and knockdown of MET expression significantly attenuated (p<0.05) cell invasion mediated by ETS2-specific siRNA. Furthermore, knockdown of ETS2 augmented HGF-induced MET phosphorylation, cell migration and invasion.
Our findings point to a tumor suppressor role for ETS2 in human NSCLC pathogenesis through inhibition of the MET proto-oncogene.
NSCLC; ETS2; tumor suppressor; MET; HGF
We propose a class of phase II clinical trial designs with sequential stopping and adaptive treatment allocation to evaluate treatment efficacy. Our work is based on two-arm (control and experimental treatment) designs with binary endpoints. Our overall goal is to construct more efficient and ethical randomized phase II trials by reducing the average sample sizes and increasing the percentage of patients assigned to the better treatment arms of the trials. The designs combine the Bayesian decision-theoretic sequential approach with adaptive randomization procedures in order to achieve simultaneous goals of improved efficiency and ethics. The design parameters represent the costs of different decisions, e.g., the decisions for stopping or continuing the trials. The parameters enable us to incorporate the actual costs of the decisions in practice. The proposed designs allow the clinical trials to stop early for either efficacy or futility. Furthermore, the designs assign more patients to better treatment arms by applying adaptive randomization procedures. We develop an algorithm based on the constrained backward induction and forward simulation to implement the designs. The algorithm overcomes the computational difficulty of the backward induction method, thereby making our approach practicable. The designs result in trials with desirable operating characteristics under the simulated settings. Moreover, the designs are robust with respect to the response rate of the control group.
sequential method; response adaptive randomization; Bayesian decision–theoretic approach; backward induction; forward simulation
Non-small-cell lung cancer (NSCLC) is the primary cause of cancer-related death in Western countries. One important approach taken to address this problem is the development of effective chemoprevention strategies. In this study, we examined whether the cyclooxygenase-2 (COX-2) inhibitor celecoxib, as evidenced by decreased cell proliferation, is biologically active in the bronchial epithelium of current and former smokers.
Patients and Methods
Current or former smokers with at least a 20 pack-year (pack-year = number of packs of cigarettes per day times number of years smoked) smoking history were randomized into one of four treatment arms (3-month intervals of celecoxib then placebo, celecoxib then celecoxib, placebo then celecoxib, or placebo then placebo) and underwent bronchoscopies with biopsies at baseline, 3 months, and 6 months. The 204 patients were primarily (79.4%) current smokers; 81 received either low-dose celecoxib or placebo and 123 received either high-dose celecoxib or placebo. Celecoxib was originally administered orally at 200 mg twice daily and the protocol subsequently increased the dose to 400 mg twice daily. The primary endpoint was change in Ki-67 labeling (from baseline to 3 months) in bronchial epithelium.
No cardiac toxicities were observed in the participants. Although the effect of low-dose treatment was not significant, high-dose celecoxib decreased Ki-67 labeling by 3.85% in former smokers and by 1.10% in current smokers—a significantly greater reduction (P = 0.02) than that seen with placebo after adjusting for metaplasia and smoking status.
A 3–6-month celecoxib regimen proved safe to administer. Celecoxib 400 mg bid was biologically active in the bronchial epithelium of current and former smokers; additional studies on the efficacy of celecoxib in NSCLC chemoprevention may be warranted.
Simon’s two-stage design is commonly used in phase II single-arm clinical trials because of its simplicity and smaller sample size under the null hypothesis compared to the one-stage design. Some studies extend this design to accommodate more interim analyses (i.e., three-stage or four-stage designs). However, most of these studies, together with the original Simon’s two-stage design, are based on the exhaustive search method, which is difficult to extend to high-dimensional, general multi-stage designs. In this study, we propose a simulated annealing (SA)-based design to optimize the early stopping boundaries and minimize the expected sample size for multi-stage or continuous monitoring single-arm trials. We compare the results of the SA method, the decision-theoretic method, the predictive probability method, and the posterior probability method. The SA method can reach the smallest expected sample sizes in all scenarios under the constraints of the same type I and type II errors. The expected sample sizes from the SA method are generally 10–20% smaller than those from the posterior probability method or the predictive probability method, and are slightly smaller than those from the decision-theoretic method in almost all scenarios. The SA method offers an excellent alternative in designing phase II trials with continuous monitoring.
Simulated annealing; Simon’s design; Early stopping; Adaptive design; Bayesian inference; Phase II trial; Type I error; Type II error; Optimal design