Performance of toxicity probability interval based designs in contrast to the continual reassessment method

Should the choice of BOIN design parameter p.tox only depend on the target DLT rate?

IMPORTANCE

On December 10, 2021, the FDA published a Determination Letter, along with a Statistical Review and Evaluation Report, and concluded that under the non-informative prior, the local Bayesian optimal interval design (BOIN) design, in its revised form, can be designated fit-for-purpose for identifying the maximum tolerated dose (MTD) of a new drug, assuming that dose-toxicity relationship is monotonically increasing. Although setting the BOIN design parameter p.tox = 1.4 * target.DLT.rate is recommended in almost all BOIN methodology articles and is the default value in the R package BOIN, it's unclear if the choice of p.tox should only depend on the target DLT rate and whether certain range of p.tox could produce the same BOIN boundary table.

DESIGN

In this simulation study, following parameters were varied one at a time, using R package BOIN, to explore each parameter's effect on the equivalence intervals of p.saf and p.tox: 1) target DLT rate, 2) n.earlystop, 3) cutoff.eli, 4) cohortsize, and 5) ncohort. And a simple 3+3 design was used as an example to explore equivalent sets of BOIN design parameters that can generate the same boundary table.

RESULTS

When the early stopping parameter n.earlystop is relatively small or the cohortsize value is not optimized via simulation, it might be better to use p.tox < 1.4 * target.DLT.rate, or try out different cohort sizes, or increase n.earlystop, whichever is both feasible and provides better operating characteristics. This is because if the cohortsize was not optimized via simulation, even when n.earlystop = 12 and cohortsize > 3, the BOIN escalation/de-escalation rules generated using p.tox = 1.4 * target.DLT.rate could be exactly the same as those calculated using p.tox > 3 * target.DLT.rate, which might not be acceptable for some pediatric trials targeting 10% DLT rate.The traditional 3+3 design stops the dose finding process when 3 patients have been treated at the current dose level, 0 DLT has been observed, and the next higher dose has already been eliminated. If additional 3 patients were required to be treated at the current dose in the situation described above, the decision rules of this commonly used 3+3 design could be generated using BOIN design with target DLT rates ranging from 18% to 29%, p.saf ranging from 8% to 26%, and different p.tox values ranging from 39% to 99%. To generate this commonly used 3+3 design table, BOIN parameters also need to satisfy a set of conditions.

Optimal dose escalation methods using deep reinforcement learning in phase I oncology trials

In phase I trials of a novel anticancer drug, one of the most important objectives is to identify the maximum tolerated dose (MTD). To this end, a number of methods have been proposed and evaluated under various scenarios. However, the percentages of correct selection (PCS) of MTDs using previous methods are insufficient to determine the dose for late-phase trials. The purpose of this study is to construct an action rule for escalating or de-escalating the dose and continuing or stopping the trial to increase the PCS as much as possible. We show that deep reinforcement learning with an appropriately defined state, action, and reward can be used to construct such an action selection rule. The simulation study shows that the proposed method can improve the PCS compared with the 3 + 3 design, CRM, BLRM, BOIN, mTPI, and i3 + 3 methods.

An overview of the BOIN design and its current extensions for novel early-phase oncology trials

Bayesian Optimal Interval (BOIN) designs are a class of model-assisted dose-finding designs that can be used in oncology trials to determine the maximum tolerated dose (MTD) of a study drug based on safety or the optimal biological dose (OBD) based on safety and efficacy. BOIN designs provide a complete suite for dose finding in early phase trials, as well as a consistent way to explore different scenarios such as toxicity, efficacy, continuous outcomes, delayed toxicity or efficacy and drug combinations in a unified manner with easy access to software to implement most of these designs. Although built upon Bayesian probability models, BOIN designs are operationally simple in general and have good statistical operating characteristics compared to other dose-finding designs. This review paper describes the original BOIN design and its many extensions, their advantages and limitations, the software used to implement them, and the most suitable situation for use of each of these designs. Published examples of the implementation of BOIN designs are provided in the Appendix.

Early completion of phase I cancer clinical trials with Bayesian optimal interval design

Phase I cancer clinical trials have been proposed novel designs such as algorithm-based, model-based, and model-assisted designs. Model-based and model-assisted designs have a higher identification rate of maximum tolerated dose (MTD) than algorithm-based designs, but are limited by the fact that the sample size is fixed. Hence, it would be very attractive to estimate the MTD with sufficient accuracy and complete the trial early. O'Quigley proposed the early completion of a trial with the continual reassessment method among model-based designs when the MTD is estimated with sufficient accuracy. However, the proposed early completion method based on the binary outcome trees has a problem that the calculation cost is high when the number of remaining patients is large. Among model-assisted designs, the Bayesian optimal interval (BOIN) design provides the simplest approach for dose adjustment. We propose the novel early completion method for the clinical trials with the BOIN design when the MTD is estimated with sufficient accuracy. This completion method can be easily calculated. In addition, the method does not require many more patients treated for the determination of early completion. We confirm that the BOIN design applying the early completion method has almost the same MTD identification rate compared to the BOIN design through simulations conducted based on over 30 000 scenarios.

SPA: Single patient acceleration in oncology dose-escalation trials

Efficient identification of the optimal dose and dosing scheme is one of the most critical and challenging tasks in early-phase oncology trials. The results are far-reaching because advancing a sub-optimal dose to late-stage development may not only jeopardize patients' safety or fail to deliver desired efficacy, but also be costly to sponsors as refined doses must be evaluated further before seeking regulatory approval. A good dose-escalation design is anticipated to yield high accuracy of selecting the correct dose while using fewer patients and keeping the trial duration short. Recently, treating a single patient at each lower dose level until certain events are triggered to switch to larger cohorts has gained much popularity. We name this approach "Single Patient Acceleration" (SPA), which is essentially a variant of the Accelerated Titration Design (ATD) by Simon et al. [25]. Although literature on novel dose-escalation methods is abundant in the past decade, there is a surprisingly lack of research on evaluating the ATD/SPA framework. In this article, we conduct comprehensive simulations to evaluate the performance of dose-escalation designs with or without SPA, and show that SPA improves design efficiency with similar or better accuracy to those without the "single patient" component under certain circumstances (e.g., slow initial enrollment, or the true maximum tolerated dose is at higher candidate dose levels). Potential safety concerns as a cost of efficiency improvement are also investigated in a quantitative manner to illustrate a comprehensive benefit-risk profile of SPA. Practical considerations and recommendations in using SPA are also discussed.

Bayesian optimization for estimating the maximum tolerated dose in Phase I clinical trials

We introduce a Bayesian optimization method for estimating the maximum tolerated dose in this article. A number of parametric model-based methods have been proposed to estimate the maximum tolerated dose; however, parametric model-based methods need an assumption that dose-toxicity relationships follow specific theoretical models. This assumption potentially leads to suboptimal dose selections if the dose-toxicity curve is misspecified. Our proposed method is based on a Bayesian optimization framework for finding a global optimizer of unknown functions that are expensive to evaluate while using very few function evaluations. It models dose-toxicity relationships with a nonparametric model; therefore, a more flexible estimation can be realized compared with existing parametric model-based methods. Also, most existing methods rely on point estimates of dose-toxicity curves in their dose selections. In contrast, our proposed method exploits a probabilistic model for an unknown function to determine the next dose candidate without ignoring the uncertainty of posterior while imposing some dose-escalation limitations. We investigate the operating characteristics of our proposed method by comparing them with those of the Bayesian-based continual reassessment method and two different nonparametric methods. Simulation results suggest that our proposed method works successfully in terms of selections of the maximum tolerated dose correctly and safe dose allocations.

Continual reassessment method with regularization in phase I clinical trials

Many Phase I trial designs have been developed to improve upon the standard 3+3 design. These designs can be classified as long-memory designs, for example, the continual reassessment method (CRM), and short-memory designs such as the modified toxicity probability interval (mTPI) design. Long-term memory designs use all data but their performance can be negatively affected by the model misspecification. Short-term memory designs only use data at the current dose and might lose efficiency as a result. To overcome these issues, we propose a regularized CRM (rCRM). The rCRM offers a trade-off between long-term memory and short-term memory methods. The rCRM gives more weight to data obtained at the doses with the estimated probability of toxicity closer to the target toxicity rate. The addition of a regularization term has an effect of shrinking the dimension of the model and leads to improved performance of the 2-parameter CRM. The rCRM is a good design choice to guide assignments in an expansion cohort phase of a dose-finding trial since dose assignments do not seem to change as often as in corresponding CRMs.

PA-CRM: A continuous reassessment method for pediatric phase I oncology trials with concurrent adult trials

Pediatric phase I trials are usually carried out after the adult trial testing the same agent has started, but not completed yet. As the pediatric trial progresses, in light of the accrued interim data from the concurrent adult trial, the pediatric protocol often is amended to modify the original pediatric dose escalation design. In practice, this is done frequently in an ad hoc way, interrupting patient accrual and slowing down the trial. We developed a pediatric-continuous reassessment method (PA-CRM) to streamline this process, providing a more efficient and rigorous method to find the maximum tolerated dose for pediatric phase I oncology trials. We use a discounted joint likelihood of the adult and pediatric data, with a discount parameter controlling information borrowing between pediatric and adult trials. According to the interim adult and pediatric data, the discount parameter is adaptively updated using the Bayesian model averaging method. Numerical study shows that the PA-CRM improves the efficiency and accuracy of the pediatric trial and is robust to various model assumptions.

Evaluation of irrational dose assignment definitions using the continual reassessment method

BACKGROUND

This article studies the notion of irrational dose assignment in Phase I clinical trials. This property was recently defined by Zhou and colleagues as a dose assignment that fails to de-escalate the dose when two out of three, three out of six, or four out of six patients have experienced a dose-limiting toxicity event at the current dose level. The authors claimed that a drawback of the well-known continual reassessment method is that it can result in irrational dose assignments. The aim of this article is to examine this definition of irrationality more closely within the conduct of the continual reassessment method.

METHODS

Over a broad range of assumed dose-limiting toxicity probability scenarios for six study dose levels and a variety of target dose-limiting toxicity rates, we simulated 2000 trials of n = 36 patients. For each scenario, we counted the number of irrational dose assignments that were made by the continual reassessment method, according to the definitions of Zhou and colleagues. For each of the irrational decisions made, we classified the dose assignment as an underdose assignment, a target dose assignment, or an overdose assignment based on the true dose-limiting toxicity probability at that dose.

RESULTS

Across eight dose-toxicity scenarios, there were a total of 181,581 dose assignments made in the simulation study. Of these assignments, 8165 (4.5%) decisions were made when two out of three, three out of six, or four out of six patients had experienced a dose-limiting toxicity at the current dose. Of these 8165 decisions, 1505 (18.4%) recommended staying at the current dose level and would therefore be classified as irrational by Zhou and colleagues. Among the irrational decisions, 41.2% were misclassified, meaning they were made either at the true target dose (17.9%) or at a true underdose (23.3%). The remaining 58.8% were made at a true overdose and therefore truly irrational. Overall, irrational dose assignments comprised <1% of the total dose assignments made during the simulation study. Similar findings are reported in simulations across 100 randomly generated dose-toxicity scenarios from a recently proposed family of curves.

CONCLUSION

Zhou and colleagues argue that the behavior of the continual reassessment method is disturbing due to its ability to make irrational dose assignments. These definitions are based on rules that mimic the popular 3 + 3 design, which should not be the benchmark used to construct guidelines for trial conduct of modern Phase I methods. Our study illustrates that these dose assignments occur very seldom in the continual reassessment method and that even when they do occur, they can often be considered sensible when accounting for all accumulated data in the study.

Coherence principles in interval-based dose finding

This paper studies the notion of coherence in interval-based dose-finding methods. An incoherent decision is either (a) a recommendation to escalate the dose following an observed dose-limiting toxicity or (b) a recommendation to deescalate the dose following a non-dose-limiting toxicity. In a simulated example, we illustrate that the Bayesian optimal interval method and the Keyboard method are not coherent. We generated dose-limiting toxicity outcomes under an assumed set of true probabilities for a trial of n=36 patients in cohorts of size 1, and we counted the number of incoherent dosing decisions that were made throughout this simulated trial. Each of the methods studied resulted in 13/36 (36%) incoherent decisions in the simulated trial. Additionally, for two different target dose-limiting toxicity rates, 20% and 30%, and a sample size of n=30 patients, we randomly generated 100 dose-toxicity curves and tabulated the number of incoherent decisions made by each method in 1000 simulated trials under each curve. For each method studied, the probability of incurring at least one incoherent decision during the conduct of a single trial is greater than 75%. Coherency is an important principle in the conduct of dose-finding trials. Interval-based methods violate this principle for cohorts of size 1 and require additional modifications to overcome this shortcoming. Researchers need to take a closer look at the dose assignment behavior of interval-based methods when using them to plan dose-finding studies.

Evaluating the effects of design parameters on the performances of phase I trial designs

Numerous designs have been proposed for phase I clinical trials. Although studies have compared their performances, few have considered the effects of changing design parameters. In this article, we review a few popular designs, including the 3 + 3, continuous reassessment method (CRM), Bayesian optimal interval (BOIN) design, and Keyboard design, and evaluate how varying design parameters (such as number of dose levels, target toxicity rate, maximum sample size, and cohort size) could impact the performances of each design through simulations. Excluded from our analysis is the mTPI-2 design, which operates in the same way as the Keyboard. Our results suggest that regardless of the choices of design parameters, the 3 + 3 design performs worse than the other ones, and BOIN and Keyboard have comparable performance to CRM. For any design, the performance varies with the choice of parameters. In particular, it improves as sample sizes increase, but the magnitude of benefit from increasing sample sizes varies substantially across scenarios. The impact of cohort size on design performances seems to have no clear direction. Therefore, BOIN and Keyboard designs are generally recommended due to their simplicity and good performance. With regard to choices of sample size and cohort size in designing a trial, it is recommend that simulations be performed for the particular clinical settings to aid decision making.

Consequences of Performing Parallel Dose Finding Trials in Heterogeneous Groups of Patients

Patient heterogeneity, in which patients can be grouped by risk of toxicity, is a design challenge in early phase dose finding trials. Carrying out independent trials for each group is a readily available approach for dose finding. However, this often leads to dose recommendations that violate the known order of toxicity risk by group, or reversals in dose recommendation. In this manuscript, trials for partially ordered groups are simulated using four approaches: independent parallel trials using the continual reassessment method (CRM), Bayesian optimal interval design, and 3 + 3 methods, as well as CRM for partially ordered groups. Multiple group order structures are considered, allowing for varying amounts of group frailty order information. These simulations find that parallel trials in the presence of partially ordered groups display a high frequency of trials resulting in reversals. Reversals occur when dose recommendations do not follow known order of toxicity risk by group, such as recommending a higher dose level in a group of patients known to have a higher risk of toxicity. CRM for partially ordered groups eliminates the issue of reversals, and simulation results indicate improved frequency of maximum tolerated dose selection as well as treating a greater proportion of trial patients at this dose compared with parallel trials. When information is available on differences in toxicity risk by patient subgroup, methods designed to account for known group ordering should be considered to avoid reversals in dose recommendations and improve operating characteristics.

Combination Lenalidomide and Azacitidine: A Novel Salvage Therapy in Patients Who Relapse After Allogeneic Stem-Cell Transplantation for Acute Myeloid Leukemia

PURPOSE

Salvage options for patients who relapse after allogeneic stem-cell transplantation (allo-SCT) for acute myeloid leukemia (AML) and myelodysplasia (MDS) remain limited, and novel treatment strategies are required. Both lenalidomide (LEN) and azacitidine (AZA) possess significant antitumor activity effect in AML. Administration of LEN post-transplantation is associated with excessive rates of graft-versus-host disease (GVHD), but AZA has been shown to ameliorate GVHD in murine transplantation models. We therefore examined the tolerability and activity of combined LEN/AZA administration in post-transplantation relapse.

PATIENTS AND METHODS

Twenty-nine patients who had relapsed after allo-SCT for AML (n = 24) or MDS (n = 5) were treated with sequential AZA (75 mg/m² for 7 days) followed by escalating doses of LEN on days 10 to 30. Dose allocation and maximum tolerated dose (MTD) estimation were guided by a modified Bayesian continuous reassessment method (CRM).

RESULTS

Sequential AZA and LEN therapy was well tolerated. The MTD of post-transplantation LEN, in combination with AZA, was determined as 25 mg daily. Three patients developed grade 2 to 4 GVHD. There was no GVHD-related mortality. Seven of 15 (47%) patients achieved a major clinical response after LEN/AZA therapy. CD8+ T cells demonstrated impaired interferon-γ/tumor necrosis factor-α production at relapse, which was not reversed during LEN/AZA administration.

CONCLUSION

We conclude LEN can be administered safely post-allograft in conjunction with AZA, and this combination demonstrates clinical activity in relapsed AML/MDS without reversing biologic features of T-cell exhaustion. The use of a CRM model delivered improved efficiency in MTD assessment and provided additional flexibility. Combined LEN/AZA therapy represents a novel and active salvage therapy in patients who had relapsed post-allograft.

The Impact of Early-Phase Trial Design in the Drug Development Process

PURPOSE

Many of the therapeutic agents that are being used currently were developed using the 3+3 decision rule for dose finding. Over the past 30 years, several dose-finding designs have been proposed and evaluated, including the "continual reassessment method" (CRM) and the "Bayesian optimal interval design" (BOIN). This research investigates the role of the choice of an early-phase design on the likelihood that drugs entering the drug development pipeline will have 2 successful phase III trials.Experimental Design: Using simulation, each agent in a population of hypothetical agents was tracked through the drug development process, from initial dose finding to 2 confirmatory phase III trials. Varying the designs of the phase I, II, and III trials allows for an assessment of the effect of the choice of designs on the proportion of agents with successful phase III trials.

RESULTS

The results indicate that using the CRM or BOIN, rather than the 3+3, substantially enhances the proportion of effective agents that have successful phase III trials, with the CRM having a greater effect than BOIN. A larger phase II trial magnifies the effect of the phase I design.

CONCLUSIONS

The results underscore the importance of the choice of the early-phase designs. Use of the 3+3 results in fewer agents with successful phase III trials compared with the CRM or BOIN. The difference is more pronounced among highly effective agents. In addition, the results show the importance of a sufficiently powered phase II trial.

Shift models for dose-finding in partially ordered groups

BACKGROUND

Limited options are available for dose-finding clinical trials requiring group-specific dose selection. While conducting parallel trials for groups is an accessible approach to group-specific dose selection, this approach allows for maximum tolerated dose selection that does not align with clinically meaningful group order information.

METHODS

The two-stage continual reassessment method is developed for dose-finding in studies involving three or more groups where group frailty order is known between some but not all groups, creating a partial order. This is an extension of the existing continual reassessment method shift model for two ordered groups. This method allows for dose selection by group, where maximum tolerated dose selection follows the known frailty order among groups. For example, if a group is known to be the most frail, the recommended maximum tolerated dose for this group should not exceed the maximum tolerated dose recommended for any other group.

RESULTS

With limited alternatives for dose-finding in partially ordered groups, this method is compared to two alternatives: (1) an existing method for dose-finding in partially ordered groups which is less computationally accessible and (2) independent trials for each group using the two-stage continual reassessment method. Simulation studies show that when ignoring information on group frailty, using independent continual reassessment method trials by group, 30% of simulations would result in maximum tolerated dose selection that is out of order between groups. In addition, the two-stage continual reassessment method for partially ordered groups selects the maximum tolerated dose more often and assigns more patients to the maximum tolerated dose compared to using independent continual reassessment method trials within each group. Simulation results for the proposed method and the less computationally accessible approach are similar.

CONCLUSION

The proposed continual reassessment method for partially ordered groups ensures appropriate maximum tolerated dose order and improves accuracy of maximum tolerated dose selection, while allowing for trial implementation that is computationally accessible.

Overall success rate of a safe and efficacious drug: Results using six phase 1 designs, each followed by standard phase 2 and 3 designs

To evaluate the overall success rate of a new drug, phase 1, 2, and 3 trials were simulated using eight toxicity and two non-decreasing efficacy profiles. Six phase 1 designs including the standard 3 + 3, CCD, BOIN, mTPI, mTPI-2, and CRM were considered with standard phase 2 and 3 designs.

Based on our results, phase 1 design recommendations are provided when data informing the general shape of the dose-toxicity curve exist. If a large jump in toxicity between dose levels is expected, the standard 3 + 3 design is recommended; it more often recognized when the MTD was exceeded and had the highest overall success rates. If gradually increasing toxicity is expected, a nonstandard design other than the CRM is recommended. Nonstandard designs were more aggressive in dosing and MTD estimation than the standard 3 + 3 and had higher overall success rates, but the CRM was too aggressive and most frequently overestimated the true MTD. If fairly constant, safe toxicity is expected across dose levels, the BOIN or CRM designs are recommended; they escalated to the highest dose most frequently with superior overall success rates.

Without data informing the shape of the dose-toxicity curve, nonstandard phase 1 designs with a modified excessive toxicity rule more easily eliminating unsafe dose levels are recommended. With this modification, MTD overestimation error decreased and overall success rates were similar or higher with nonstandard designs. Among nonstandard designs, the modified CCD and BOIN perform well and are as transparent and simple to implement as the standard 3 + 3 design.

Revisiting isotonic phase I design in the era of model-assisted dose-finding

Background/aims In the conduct of phase I trials, the limited use of innovative model-based designs in practice has led to an introduction of a class of "model-assisted" designs with the aim of effectively balancing the trade-off between design simplicity and performance. Prior to the recent surge of these designs, methods that allocated patients to doses based on isotonic toxicity probability estimates were proposed. Like model-assisted methods, isotonic designs allow investigators to avoid difficulties associated with pre-trial parametric specifications of model-based designs. The aim of this work is to take a fresh look at an isotonic design in light of the current landscape of model-assisted methods. Methods The isotonic phase I method of Conaway, Dunbar, and Peddada was proposed in 2004 and has been regarded primarily as a design for dose-finding in drug combinations. It has largely been overlooked in the single-agent setting. Given its strong simulation performance in application to more complex dose-finding problems, such as drug combinations and patient heterogeneity, as well as the recent development of user-friendly software to accompany the method, we take a fresh look at this design and compare it to a current model-assisted method. We generated operating characteristics of the Conaway-Dunbar-Peddada method using a new web application developed for simulating and implementing the design and compared it to the recently proposed Keyboard design that is based on toxicity probability intervals. Results The Conaway-Dunbar-Peddada method has better performance in terms of accuracy of dose recommendation and safety in patient allocation in 17 of 20 scenarios considered. The Conaway-Dunbar-Peddada method also allocated fewer patients to doses above the maximum tolerated dose than the Keyboard method in many of scenarios studied. Overall, the performance of the Conaway-Dunbar-Peddada method is strong when compared to the Keyboard method, making it a viable simple alternative to the model-assisted methods developed in recent years. Conclusion The Conaway-Dunbar-Peddada method does not rely on the specification and fitting of a parametric model for the entire dose-toxicity curve to estimate toxicity probabilities as other model-based designs do. It relies on a similar set of pre-trial specifications to toxicity probability interval-based methods, yet unlike model-assisted methods, it is able to borrow information across all dose levels, increasing its efficiency. We hope this concise study of the Conaway-Dunbar-Peddada method, and the availability of user-friendly software, will augment its use in practice.

Report of the 14th International Conference on Malignant Lymphoma (ICML) Closed Workshop on Future Design of Clinical Trials in Lymphomas

The 14th ICML held in Lugano in June 2017 was preceded by a closed workshop (organized in collaboration with the American Association for Cancer Research and the European School of Oncology) where experts in preclinical and clinical research in lymphomas met to discuss the current drug development landscape focusing on critical open questions that need to be addressed in the future to permit a more efficient drug development paradigm in lymphoma. Topics discussed included both preclinical models that can be used to test new drugs and drug combinations, as well as the optimal design of clinical trials and the endpoints that should be used to facilitate accelerated progress. This report represents a summary of the workshop. Clin Cancer Res; 24(13); 2993-8. ©2018 AACR.

Dose Transition Pathways: The Missing Link Between Complex Dose-Finding Designs and Simple Decision-Making

The ever-increasing pace of development of novel therapies mandates efficient methodologies for assessment of their tolerability and activity. Evidence increasingly support the merits of model-based dose-finding designs in identifying the recommended phase II dose compared with conventional rule-based designs such as the 3 + 3 but despite this, their use remains limited. Here, we propose a useful tool, dose transition pathways (DTP), which helps overcome several commonly faced practical and methodologic challenges in the implementation of model-based designs. DTP projects in advance the doses recommended by a model-based design for subsequent patients (stay, escalate, de-escalate, or stop early), using all the accumulated information. After specifying a model with favorable statistical properties, we utilize the DTP to fine-tune the model to tailor it to the trial's specific requirements that reflect important clinical judgments. In particular, it can help to determine how stringent the stopping rules should be if the investigated therapy is too toxic. Its use to design and implement a modified continual reassessment method is illustrated in an acute myeloid leukemia trial. DTP removes the fears of model-based designs as unknown, complex systems and can serve as a handbook, guiding decision-making for each dose update. In the illustrated trial, the seamless, clear transition for each dose recommendation aided the investigators' understanding of the design and facilitated decision-making to enable finer calibration of a tailored model. We advocate the use of the DTP as an integral procedure in the co-development and successful implementation of practical model-based designs by statisticians and investigators. Clin Cancer Res; 23(24); 7440-7. ©2017 AACR.