Phase I-II clinical trial design: a state-of-the-art paradigm for dose finding.

Background: Conventional phase I algorithms for finding a phase-2 recommended dose (P2RD) based on toxicity alone is problematic because the maximum tolerated dose (MTD) is not necessarily the optimal dose with the most desirable risk-benefit trade-off. Moreover, the increasingly common practice of treating an expansion cohort at a chosen MTD has undesirable consequences that may not be obvious. Patients and methods: We review the phase I-II paradigm and the EffTox design, which utilizes both efficacy and toxicity to choose optimal doses for successive patient cohorts and find the optimal P2RD. We conduct a computer simulation study to compare the performance of the EffTox design with the traditional 3 + 3 design and the continuous reassessment method. Results: By accounting for the risk-benefit trade-off, the EffTox phase I-II design overcomes the limitations of conventional toxicity-based phase I designs. Numerical simulations show that the EffTox design has higher probabilities of identifying the optimal dose and treats more patients at the optimal dose. Conclusions: Phase I-II designs, such as the EffTox design, provide a coherent and efficient approach to finding the optimal P2RD by explicitly accounting for risk-benefit trade-offs underlying medical decisions.

Fludarabine with pharmacokinetically guided IV busulfan is superior to fixed-dose delivery in pretransplant conditioning of AML/MDS patients.

We hypothesized that IV busulfan (Bu) dosing could be safely intensified through pharmacokinetic (PK-) dose guidance to minimize the inter-patient variability in systemic exposure (SE) associated with body-sized dosing, and that this should improve outcome of AML/MDS patients undergoing allogeneic stem cell transplantation. To test this hypothesis, we treated 218 patients (median age 50.7 years, male/female 50/50%) with fludarabine 40 mg/m2 once daily x4, each dose followed by IV Bu, randomized to 130 mg/m2 (N=107) or PK-guided to average daily SE, AUC of 6000 µM min (N=111), stratified for remission status and allo-grafting from HLA-matched donors. Toxicity and GvHD rates in the groups were similar; the risk of relapse or treatment-related mortality remained higher in the fixed-dose group throughout the 80-month observation period. Further, PK-guidance yielded safer disease control, leading to improved overall and PFS, most prominently in MDS patients and in AML patients not in remission at allogeneic stem cell transplantation. We conclude that AML/MDS patients receiving pretransplant conditioning treatment with our 4-day regimen may benefit significantly from PK-guided Bu dosing. This could be considered an alternative to fixed-dose delivery since it provides the benefit of precise dose delivery to a predetermined SE without increasing risk(s) of serious toxicity and/or GvHD.

PR1 peptide vaccine induces specific immunity with clinical responses in myeloid malignancies.

PR1, an HLA-A2-restricted peptide derived from both proteinase 3 and neutrophil elastase, is recognized on myeloid leukemia cells by cytotoxic T lymphocytes (CTLs) that preferentially kill leukemia and contribute to cytogenetic remission. To evaluate safety, immunogenicity and clinical activity of PR1 vaccination, a phase I/II trial was conducted. Sixty-six HLA-A2+ patients with acute myeloid leukemia (AML: 42), chronic myeloid leukemia (CML: 13) or myelodysplastic syndrome (MDS: 11) received three to six PR1 peptide vaccinations, administered subcutaneously every 3 weeks at dose levels of 0.25, 0.5 or 1.0 mg. Patients were randomized to the three dose levels after establishing the safety of the highest dose level. Primary end points were safety and immune response, assessed by doubling of PR1/HLA-A2 tetramer-specific CTL, and the secondary end point was clinical response. Immune responses were noted in 35 of 66 (53%) patients. Of the 53 evaluable patients with active disease, 12 (24%) had objective clinical responses (complete: 8; partial: 1 and hematological improvement: 3). PR1-specific immune response was seen in 9 of 25 clinical responders versus 3 of 28 clinical non-responders (P=0.03). In conclusion, PR1 peptide vaccine induces specific immunity that correlates with clinical responses, including molecular remission, in AML, CML and MDS patients.

A phase II randomized trial of induction chemotherapy versus no induction chemotherapy followed by preoperative chemoradiation in patients with esophageal cancer.

BACKGROUND: The contribution of induction chemotherapy (IC) before preoperative chemoradiation for esophageal cancer (EC) is not known. We hypothesized that IC would increase the rate of pathologic complete response (pathCR). METHODS: Trimodality-eligibile patients were randomized to receive no IC (Arm A) or IC (oxaliplatin/FU; Arm B) before oxaliplatin/FU/radiation. Surgery was attempted ∼5-6 weeks after chemoradiation. The pathCR rate, post-surgery 30-day mortality, overall survival (OS), and toxic effects were assessed. Bayesian methods and Fisher's exact test were used. RESULTS: One hundred twenty-six patients were randomized dynamically to balance the two arms for histology, baseline stage, gender, race, and age. Fifty-five patients in Arm A and 54 in Arm B underwent surgery. The median actuarial OS for all patients (54 deaths) was 45.62 months [95% confidence interval (CI), 27.63-NA], with median OS 45.62 months (95% CI 25.56-NA) in Arm A and 43.68 months (95% CI 27.63-NA) in Arm B (P = 0.69). The pathCR rate in Arm A was 13% (7 of 55) and 26% (14 of 54) in Arm B (two-sided Fisher's exact test, P = 0.094). Safety was similar in both arms. CONCLUSIONS: These data suggest that IC produces non-significant increase in the pathCR rate and does not prolong OS. Further development of IC before chemoradiation may not be beneficial. Clinical trial no.: NCT 00525915 (www.clinicaltrials.gov).

Adenoviral infections in adult allogeneic hematopoietic SCT recipients: a single center experience.

Disseminated adenoviral infection (AI) is associated with profound immunosuppression and poor outcome after allogeneic hematopoietic SCT (allo-HSCT). A better understanding of AI in allo-HSCT recipients can serve as a basis to develop more effective management strategies. We evaluated all adult patients who received allo-HSCT at MD Anderson Cancer Center between 1999 and 2008. Among the 2879 allo-HSCT patients, 73 (2.5%) were diagnosed with AI. Enteritis (26%) and pneumonia (24%) were the most common clinical manifestations; pneumonia was the most common cause of adenovirus-associated death. A multivariable Bayesian logistic regression showed that when the joint effects of all covariates were accounted for, cord blood transplant, absolute lymphocyte count (ALC) ≤ 200/mm(3) and male gender were associated with a higher probability of disseminated AI. The OS was significantly worse for patients with AI that was disseminated rather than localized (median of 5 months vs median of 28 months, P<0.001) and for patients with ALC ≤ 200/mm(3) (P<0.001). Disseminated AI, in patients who received allo-HSCT, is a significant cause of morbidity and mortality. Strategies for early diagnosis and intervention are essential, especially for high-risk patients.

Intravenous BU plus Mel: an effective, chemotherapy-only transplant conditioning regimen in patients with ALL.

We investigated the administration of i.v. BU combined with melphalan (Mel) in patients with ALL undergoing allogeneic hematopoietic SCT. Forty-seven patients with a median age of 33 years (range 20-61) received a matched sibling (n=27) or matched unrelated donor transplant (n=20) for ALL in first CR (n=26), second CR (n=13), or with more advanced disease (n=8). BU was infused daily for 4 days, either at a fixed dose of 130 mg/m² (5 patients) or using pharmacokinetic (PK) dose adjustment (42 patients), to target an average daily area-under-the-curve (AUC) of 5000 µmol/min, determined by a test dose of i.v. BU at 32 mg/m². This was followed by a rest day, then two daily doses of Mel at 70 mg/m². Stem cells were infused on the following day. The 2-year OS, PFS and non-relapse mortality (NRM) rates were 35% (95% confidence interval (CI), 23-51%), 31% (95% CI, 21-48%) and 37% (95% CI, 23-50%), respectively. Acute NRM at 100 days was favorable at 12% (95% CI, 5-24%); however, the 2-year NRM was significantly higher for patients older than 40 years, 58% vs 20%, mainly due to GVHD.

Dynamic change in phosphorylated platelet-derived growth factor receptor in peripheral blood leukocytes following docetaxel therapy predicts progression-free and overall survival in prostate cancer.

In a placebo-controlled randomised study of the platelet-derived growth factor receptor (PDGFR) inhibitor imatinib mesylate and docetaxel in metastatic prostate cancer with bone metastases (n=116), no significant differences in progression-free and overall survival were observed. To evaluate pharmacodynamic correlates of outcomes, we assessed the association of plasma platelet-derived growth factor (PDGF) isoform kinetics and PDGFR inhibition with progression-free and overall survival by individual treatment arm. We found that in the docetaxel-placebo arm alone, the probability of decrease in PDGFR phosphorylation (Pr-Decr-pPDGFR) above 0.5 (vs 30 months (HR 3.1; P=0.04 in log-rank test). By contrast, in the docetaxel plus imatinib arm, the association of Pr-Decr-pPDGFR >0.5 with a rise in plasma PDGF isoform concentrations and inferior survival was not observed. The data suggest that dynamic changes in PDGFR phosphorylation in peripheral blood leukocytes predict docetaxel efficacy. Rising plasma PDGF concentrations may explain and/or mark docetaxel resistance. Validation and mechanistic studies addressing these unexpected findings should anticipate a confounding influence of concurrent PDGFR inhibitor therapy.

Phase I/II study of gemtuzumab ozogamicin added to fludarabine, melphalan and allogeneic hematopoietic stem cell transplantation for high-risk CD33 positive myeloid leukemias and myelodysplastic syndrome.

We investigated the hypothesis that gemtuzumab ozogamicin (GO), an anti-CD33 immunotoxin would improve the efficacy of fludarabine/melphalan as a preparative regimen for allogeneic hematopoietic stem cell transplantation (HSCT) in a phase I/II trial. Toxicity was defined as grades III-IV organ damage, engraftment failure or death within 30 days. 'Response' was engraftment and remission (CR) on day +30. We sought to determine the GO dose (2, 4 or 6 mg m(-2)) giving the best trade-off between toxicity and response. All patients were not candidates for myeloablative regimens. Treatment plan: GO (day -12), fludarabine 30 mg m(-2) (days -5 to -2), melphalan 140 mg m(-2) (day -2) and HSCT (day 0). GVHD prophylaxis was tacrolimus and mini-methotrexate. Diagnoses were AML (n=47), MDS (n=4) or CML (n=1). Median age was 53 years (range, 13-72). All but three patients were not in CR. Donors were related (n=33) or unrelated (n=19). Toxicity and response rates at 4 mg m(-2) were 50% (n=4) and 50% (n=4). GO dose was de-escalated to 2 mg m(-2): 18% had toxicity (n=8) and 82% responded (n=36). 100-day TRM was 15%; one patient had reversible hepatic VOD. Median follow-up was 37 months. Median event-free and overall survival was 6 and 11 months. GO 2 mg m(-2) can be safely added to fludarabine/melphalan, and this regimen merits further evaluation.

Comparison of 100-day mortality rates associated with i.v. busulfan and cyclophosphamide vs other preparative regimens in allogeneic bone marrow transplantation for chronic myelogenous leukemia: Bayesian sensitivity analyses of confounded treatment and center effects.

We evaluated the 100-day mortality rates associated with busulfan-based myeloablative conditioning regimens based on data from 1812 chronic myelogenous leukemia patients who underwent allogeneic blood or marrow transplantation (allotx). In all, 47 patients received intravenous (i.v.) busulfan and cyclophosphamide (i.v.BuCy2) with allotx at MD Anderson Cancer Center (MDACC) during 1995-1999. The remaining 1765 patients, whose data were supplied by the International Bone Marrow Transplant Registry (IBMTR), received alternative preparative regimens, primarily Cy-total body irradiation ( approximately 45%) or oral BuCy ( approximately 35%) during 1997-1998. As patients were not randomized between conditioning regimens, the i.v.BuCy2-versus-alternative treatment effect is confounded with a possible center effect due to nontreatment differences associated with factors differing between MDACC and the IBMTR centers. Additional complications are that the i.v.BuCy2-MDACC patients all survived 100 days, and three prognostic subgroups were included. Bayesian sensitivity analyses were performed to assess treatment effect on the probability of 100-day mortality, over a range of possible MDACC-versus-IBMTR center effects. For these patients, the posterior probability that i.v.BuCy2 was superior to alternative conditioning regimens ranges from 0.54 to 0.99, depending on prognosis and the magnitude of the assumed center effect.

Practical model-based dose-finding in phase I clinical trials: methods based on toxicity.

We describe two practical, outcome-adaptive statistical methods for dose-finding in phase I clinical trials. One is the continual reassessment method and the other is based on a logistic regression model. Both methods use Bayesian probability models as a basis for learning from the accruing data during the trial, choosing doses for successive patient cohorts, and selecting a maximum tolerable dose (MTD). These methods are illustrated and compared to the conventional 3+3 algorithm by application to a particular trial in renal cell carcinoma. We also compare their average behavior by computer simulation under each of several hypothetical dose-toxicity curves. The comparisons show that the Bayesian methods are much more reliable than the conventional algorithm for selecting an MTD, and that they have a low risk of treating patients at unacceptably toxic doses.

Phase II study of the antiangiogenesis agent thalidomide in recurrent or metastatic squamous cell carcinoma of the head and neck.

BACKGROUND: Thalidomide has been shown to have antiangiogenic effects in preclinical models as well as a significant antitumor effect in hematologic tumors such as multiple myeloma. The authors performed this Phase II study to determine the activity, toxicity profile, and antiangiogenic effect of thalidomide in patients with locoregionally recurrent or metastatic squamous cell carcinoma of the head and neck. METHODS: Twenty-one patients with recurrent or metastatic squamous cell carcinoma of the head and neck were treated with single-agent thalidomide. All patients had received radiation therapy, and most had undergone surgery (95%) and/or chemotherapy (90%). Thalidomide was initiated at 200 mg;3>daily and increased to a target dose of 1000 mg daily. Patients continued treatment until disease progression, unacceptable toxicity, or death occurred. RESULTS: All 21 patients eventually developed progressive disease. Median time to progression was 50 days (95% confidence interval, 28-70), with median overall survival time of 194 days (95% lower confidence boundary, 151), similar to the progression and survival times reported for this patient group with other agents. Thalidomide was generally well tolerated, with few patients experiencing Grades 3 to 4 toxicities. Serum vascular endothelial growth factor and basic fibroblast growth factor levels increased in six of seven patients, for whom paired serum samples were available and all of whom had progressive disease. CONCLUSIONS: In this heavily pretreated population of patients with advanced squamous cell carcinoma of the head and neck, thalidomide does not appear to have single-agent antitumor activity. Further evaluation of the mechanism of action of thalidomide is indicated. Potentially, future evaluations of thalidomide may be performed in combination with other antiangiogenic or cytotoxic agents in patients with earlier stage disease or in patients with minimal residual disease.

Comparison of idarubicin + ara-C-, fludarabine + ara-C-, and topotecan + ara-C-based regimens in treatment of newly diagnosed acute myeloid leukemia, refractory anemia with excess blasts in transformation, or refractory anemia with excess blasts.

It has been unclear whether regimens containing topotecan + ara-C (TA) or fludarabine + ara-C (FA) +/- idarubicin are superior to regimens containing idarubicin + ara-C (IA) without either fludarabine or topotecan for treatment of newly diagnosed acute myeloid leukemia (AML), refractory anemia with excess blasts in transformation (RAEB-t), or RAEB. Of 1279 patients treated here for these diagnoses between 1991 and 1999, 322 received IA regimens, 600 FA regimens, and 357 TA regimens. All regimens used ara-C doses of 1 to 2 gm/m(2)/d, given by continuous infusion in IA, and over 2 to 4 hours in FA and TA. Complete remission (CR) rates were lower with FA (55%) and TA (59%) than with IA (77%). Both event-free survival (EFS) in CR and survival were shorter: median EFS in CR (95% confidence interval) was 63 weeks (range, 55-76 weeks) for IA, 40 (range, 31-46 weeks) for FA, and 36 (range, 27-44 weeks) for TA; median survival was 77 weeks (range, 57-88 weeks) for IA, 30 (range, 27-35 weeks) for FA, and 41 (range, 35-50 weeks) for TA. These trials were not randomized, and patients with worse prognoses were disproportionately given the FA and TA regimens. Nonetheless, after accounting for prognosis the FA and TA regimens remained highly significantly associated with lower CR rates, shorter EFS in CR, and shorter survival. Accounting for possible effects of individual trials within each of the IA, FA, and TA groups did not alter these findings. It is unlikely that, as given here, either FA or TA is, in general, superior to IA, highlighting the need for new treatments.

Dose-finding based on feasibility and toxicity in T-cell infusion trials.

A new modality for treatment of cancer involves the ex vivo growth of cancer-specific T-cells for subsequent infusion into the patient. The therapeutic aim is selective destruction of cancer cells by the activated infused cells. An important problem in the early phase of developing such a treatment is to determine a maximal tolerated dose (MTD) for use in a subsequent phase II clinical trial. Dose may be quantified by the number of cells infused per unit body weight, and determination of an MTD may be based on the probability of infusional toxicity as a function of dose. As in a phase I trial of a new chemotherapeutic agent, this may be done by treating successive cohorts of patients at different dose levels, with each new level chosen adaptively based on the toxicity data of the patients previously treated. Such a dose-finding strategy is inadequate in T-cell infusion trials because the number of cells grown ex vivo for a given patient may be insufficient for infusing the patient at the current targeted dose. To address this problem, we propose an algorithm for trial conduct that determines a feasible MTD based on the probabilities of both infusibility and toxicity as functions of dose. The method is illustrated by application to a dendritic cell activated lymphocyte infusion trial in the treatment of acute leukemia. A simulation study indicates that the proposed methodology is both safe and reliable.

Optimal two-stage designs for clinical trials based on safety and efficacy.

In clinical trials designed to evaluate treatment efficacy, it is common practice to terminate a treatment arm in which the observed rate of an adverse event is unacceptably high. This practice may be formalized by a group-sequential test based on a multivariate outcome including both adverse and efficacy events. Recently, Thall and Cheng proposed a family of tests for randomized trials of an experimental treatment versus a standard where patient outcome is bivariate with entries characterizing efficacy and safety. The test is motivated by the idea that clinically meaningful improvements over the standard may be characterized by a two-dimensional parameter quantifying trade-offs between efficacy and safety. We provide optimal two-stage designs based on this test that minimize either the mean sample size under the null hypothesis of no treatment difference, or the maximum sample size if the trial continues to a second stage. A more general group-sequential version of the design also is described, an illustration is provided, and application to the special case of single-arm phase II trials is discussed.

Melphalan and purine analog-containing preparative regimens: reduced-intensity conditioning for patients with hematologic malignancies undergoing allogeneic progenitor cell transplantation.

A reduced-intensity preparative regimen consisting of melphalan and a purine analog was evaluated for allogeneic transplantation in 86 patients who had a variety of hematologic malignancies and were considered poor candidates for conventional myeloablative therapies because of age or comorbidity. Seventy-eight patients received fludarabine 25 mg/m(2) daily for 5 days in combination with melphalan 180 mg/m(2) (n = 66) or 140 mg/m(2) (n = 12). Eight patients received cladribine 12 mg/m(2) continuous infusion for 5 days with melphalan 180 mg/m(2). The median age was 52 years (range, 22-70 years). Disease status at transplantation was either first remission or first chronic phase in 7 patients, untreated first relapse or subsequent remission in 16 patients, and refractory leukemia or transformed chronic myelogenous leukemia in 63 patients. Nonrelapse mortality rates on day 100 were 37.4% for the fludarabine/melphalan combination and 87.5% for the cladribine/melphalan combination. The median percentage of donor cells at 1 month in 75 patients was 100% (range, 0%-100%). The probability of grade 2-4 and 3-4 acute graft-versus-host disease was 0.49 (95% CI, 0.38-0.60) and 0.29 (95% CI, 0.18-0.41), respectively. Disease-free survival at 1 year was 57% for patients in first remission or chronic phase and 49% for patients with untreated first relapse or in a second or later remission. On multivariate analysis the strongest predictor for disease-free survival was a good or intermediate risk category. In summary, fludarabine/melphalan combinations are feasible in older patients with associated comorbidities, and long-term disease control can be achieved with reduced-intensity conditioning in this population.

Decision-theoretic designs for pre-phase II screening trials in oncology.

A Bayesian decision-theoretic method is proposed for conducting small, randomized pre-phase II selection trials. The aim is to improve on the design of Thall and Estey (1993, Statistics in Medicine 12, 1197-1211). Designs are derived that optimize a gain function accounting for current and future patient gains, per-patient cost, and future treatment development cost. To reduce the computational burden associated with backward induction, myopic versions of the design that consider only one, two, or three future decisions at a time are also considered. The designs are compared in the context of a screening trial in acute myelogenous leukemia.

Evaluating multiple treatment courses in clinical trials.

In oncology, a patient's treatment often involves multiple courses of chemotherapy. The most common medical practice in choosing treatments for successive courses is to repeat a treatment that is successful in a given course and otherwise switch to a different treatment. Patient outcome thus consists of a sequence of dependent response variables and corresponding treatments. Despite the widespread use of such adaptive 'play-the-winner-and-drop-the-loser' algorithms in medical settings involving multiple treatment courses, most statistical methods for treatment evaluation characterize early patient outcome as a single response to a single treatment, resulting in a substantial loss of information. In this paper, we provide a statistical framework for multi-course clinical trials involving some variant of the play-the-winner-and-drop-the-loser strategy. The aim is to design and conduct the trial to more closely reflect actual clinical practice, and thus increase the amount of information per patient. The proposed design is similar to a multi-stage cross-over trial, with the essential difference that here all treatments after the first course are assigned adaptively. We illustrate the method by application to a randomized phase II trial for androgen independent prostate cancer. We consider the goals of selecting one best treatment, or selecting a best ordered pair of treatments with the second given if the first fails to achieve a patient success. A simulation study is reported, and extensions to trials involving toxicity or regimen-related death are discussed.

Approximate Bayesian evaluation of multiple treatment effects.

We propose an approximate Bayesian method for comparing an experimental treatment to a control based on a randomized clinical trial with multivariate patient outcomes. Overall treatment effect is characterized by a vector of parameters corresponding to effects on the individual patient outcomes. We partition the parameter space into four sets where, respectively, the experimental treatment is superior to the control, the control is superior to the experimental, the two treatments are equivalent, and the treatment effects are discordant. We compute posterior probabilities of the parameter sets by treating an estimator of the parameter vector like a random variable in the Bayesian paradigm. The approximation may be used in any setting where a consistent, asymptotically normal estimator of the parameter vector is available. The method is illustrated by application to a breast cancer data set consisting of multiple time-to-event outcomes with covariates and to count data arising from a cross-classification of response, infection, and treatment in an acute leukemia trial.

Maximum tolerated dose of nalmefene in patients receiving epidural fentanyl and dilute bupivacaine for postoperative analgesia.

BACKGROUND: This study investigated the ability of the modified continual reassessment method (MCRM) to determine the maximum tolerated dose of the opioid antagonist nalmefene, which does not reverse analgesia in an acceptable number of postoperative patients receiving epidural fentanyl in 0.075% bupivacaine. METHODS: In the postanesthetic care unit, patients received a single intravenous dose of 0.25, 0.50, 0.75, or 1.00 microg/kg nalmefene. Reversal of analgesia was defined as an increase in pain score of two or more integers above baseline on a visual analog scale from 0 through 10 after nalmefene administration. Patients were treated in cohorts of one, starting with the lowest dose. The maximum tolerated dose of nalmefene was defined as that dose, among the four studied, with a final mean probability of reversal of anesthesia (PROA) closest to 0.20 (ie., a 20% chance of causing reversal). The modified continual reassessment method is an iterative Bayesian statistical procedure that, in this study, selected the dose for each successive cohort as that having a mean PROA closest to the preselected target PROA of 0.20. RESULTS: The modified continual reassessment method repeatedly updated the PROA of each dose level as successive patients were observed for presence or absence of ROA. After 25 patients, the maximum tolerated dose of nalmefene was selected as 0.50 microg/kg (final mean PROA = 0.18). The 1.00-microg/kg dose was never tried because its projected PROA was far above 0.20. CONCLUSIONS: The modified continual reassessment method facilitated determination of the maximum tolerated dose ofnalmefene . Operating characteristics of the modified continual reassessment method suggest it may be an effective statistical tool for dose-finding in trials of selected analgesic or anesthetic agents.