Dose Optimization in Oncology Drug Development: Risk Factors for Postmarketing Requirements and Commitments

Abstract

Optimal dosing of oncological drugs is historically determined based on the “higher is better” paradigm. However, a paradigm shift in optimal dose selection has occurred in the development of new modalities, including molecularly targeted drugs, antibody drugs, and immunotherapies. In 2021, Project Optimus was launched by the Food and Drug Administration Oncology Center of Excellence to reform the dose optimization and dose selection paradigm in oncology drug development. In August 2024, “Optimizing the Dosage of Human Prescription Drugs and Biological Products for the Treatment of Oncologic Diseases” was published, encouraging randomized evaluation of the benefit/risk profile of a range of doses before initiating a registration trial. Although Project Optimus offers general guidance on dose optimization, it does not specify which early clinical data requires a more cautious approach to dose optimization. This is the first comprehensive study to investigate newly approved oncology drugs by the FDA over a long period and to identify the risk factors for postmarketing requirement or commitment to dose optimization, using logistic regression analysis. Our findings show that when the labeled dose is the maximum tolerated dose, the percentage of adverse reactions leading to treatment discontinuation is increased, and an exposure‐safety relationship is established, the risk for postmarketing requirement or commitment to dose optimization is increased. Our study will provide actionable, data‐driven insights into dose optimization strategies by objectively and quantitatively evaluating risk factors. These findings will serve as valuable guidance for designing more effective early‐phase trials, complementing the FDA Project Optimus guidance.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

There has been a paradigm shift in the optimal dose selection for oncology drugs highlighted by Project Optimus, an initiative launched by the FDA to reform the approach to dose optimization and selection. This has led to increasing attention to dose optimization in oncology drug development.

• WHAT QUESTION DID THIS STUDY ADDRESS?

This study sought to identify the risk factors for PMR/PMC related to dose optimization in oncology drugs using logistic regression analysis and quantitatively evaluated the contribution of each risk factor to the PMR/PMC on dose optimization.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

By logistic regression analysis, this study objectively identified important risk factors, including “MTD being the labeled dose,” “an increased percentage of adverse reactions leading to treatment discontinuation,” and “the presence of an exposure‐safety relationship,” and quantitatively evaluated the contribution of each risk factor to the PMR/PMC on dose optimization. Our findings offer a more actionable and evidence‐based refinement of dose optimization strategies based on the emerging data obtained in early‐phase trials and observed risk factors.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

These findings underscore the critical importance of incorporating E‐R analyses during dose selection in early clinical trials. These results offer practical guidance for dose optimization strategies.

Optimal dosing of oncology drugs is historically determined based on the occurrence of toxicity under the assumption that the maximum tolerated dose (MTD) is the optimal dose. This approach stems from the “higher is better” paradigm, which was developed decades ago for cytotoxic agents. In fact, the effective concentration curve of most cytotoxic agents is steep, making it appropriate to select the MTD as the optimal dose based on the safety/tolerability‐driven dose selection strategy. ¹ However, with the development of new treatment modalities, such as molecularly targeted drugs, antibody drugs, and immunotherapies, the MTD paradigm has been challenged. This is due to the observation of non‐linear and even flat exposure‐response (E‐R) relationships during the clinical development of molecularly targeted therapies and immunotherapies. ² These drugs typically do not require dosing at an almost intolerable level but rather at a dose that achieves adequate target occupancy and inhabitation. As a result, the traditional system of defining the MTD based on toxicity limitations based on peak exposure is no longer appropriate. ³ Hence, doses below the MTD may have similar activity to the MTD with fewer toxicities, and a lower dose level than the MTD could be the optimal dose. ⁴ , ⁵ This highlights the need to set an appropriate dosage selection scheme according to drug profile.

Furthermore, 15.9% (24/151) of first‐cycle review failures among all drug applications first submitted to the Food and Drug Administration (FDA) between 2000 and 2012 for new molecular entities are due to the uncertainty related to dose selection. Dose optimization is, therefore, a key factor for the timely development of effective new drugs. ⁶

Moreover, if the optimal dose is not evaluated sufficiently in clinical trials and if the labeled dose is unnecessarily high, severe toxicity may occur without additional efficacy, leading to a high rate of dose reduction. Intolerable toxicities may lead to premature discontinuation and missed opportunities for continued benefit from the drug and a negative impact on overall survival. It may further decrease the patient's quality of life (QOL) and negatively impact the patient's financial burden. ⁷ Furthermore, modern oncology drugs, unlike cytotoxic agents, may be administered for many years, and over time, even low‐grade toxicities, which would not meet traditional definitions of dose‐limiting toxicities, may lead to a reduced QOL for patients, making the drugs intolerable. ⁸

In 2021, the FDA Oncology Center of Excellence launched Project Optimus, which provides the framework and guidance for dose optimization during stages of clinical development of anticancer drugs. ⁹ This initiative aims to reform the dose optimization and dose selection paradigm in oncology drug development. In August 2024, the guidance document titled “Optimizing the Dosage of Human Prescription Drugs and Biological Products for the Treatment of Oncologic Diseases” was published. ¹⁰ This guidance encourages the randomized evaluation of the benefit/risk profile across a range of doses before initiating the registration trial. Although Project Optimus offers general guidance on dose optimization, it does not specify the type of early clinical data for which a more cautious approach to dose optimization is necessary. In this study, new oncology drugs approved in the United States between 2010 and 2023 were evaluated to identify the risk factor for postmarketing requirement (PMR)/postmarketing commitment (PMC) on dose optimization using logistic regression. A key feature of this study is that risk factors associated with PMR/PMC related to dose optimization were objectively and quantitatively evaluated by statistical analysis using comprehensive, long‐term data. In the present study, using logistic regression analysis, we objectively identified the important risk factors and quantitatively evaluated the contribution of each risk factor to the PMR/PMC on dose optimization. Our study provides actionable, data‐driven insights into dose optimization strategies. These findings will serve as valuable guidance for designing more effective early‐phase trials, complementing the framework set by FDA Project Optimus.

MATERIALS AND METHODS

Data collection

Scope of the study

New oncology drugs approved in the United States between 2010 and 2023 were evaluated. Drugs with additional indications, combination drugs, and biosimilars were excluded from the analysis.

Data sourcing

Data were sourced from the publicly available Drugs@FDA database, ¹¹ including approval letters, FDA review reports (multiple‐discipline review report/medical review report/clinical pharmacology review report), and initial or current United States prescribing information. ¹¹

Data extraction

Two authors (HK and KM) manually reviewed and extracted excerpts for the identified information on each drug, including PMR and PMC on dose optimization, as well as explanatory variables. Inconsistencies were resolved through discussion. Except for cases of no applicable items, this study was prepared following the guidelines of Strengthening the Reporting of Observational Studies in Epidemiology (STROBE). ¹²

New oncology drugs with PMR/PMC on dose optimization

New oncology drugs with PMR/PMC on dose optimization were identified using an approval letter. “POSTMARKETING REQUIREMENTS UNDER 505(o) and POSTMARKETING COMMITMENTS SUBJECT TO REPORTING REQUIREMENTS UNDER SECTION 506B” in the approval letter are counted as PMR/PMC related to dose optimization. PMR/PMC for dose adjustment in special populations, food effect, or drug–drug interactions were not considered to be related to dose optimization. Furthermore, only cases of PMR/PMC that directly required a dose evaluation other than the approved dosage or regimen are considered PMR/PMC on dose optimization.

Explanatory variables for investigating risk factors for PMR/PMC on dose optimization

Explanatory variables for investigating PMR/PMC on dose optimization were companion diagnostics (CDx) (Yes/No [Y/N]), accelerated approval (Y/N), MTD, labeled dose, exposure‐efficacy relationship (Y/N), exposure‐safety relationship (Y/N), proportionality in pharmacokinetics, adverse reactions leading to treatment discontinuation (%), phase of pivotal study, type of pivotal study (placebo or active control study), presence of multiple dose evaluations, and randomized dose‐ranging trial.

In general, dose selection was justified based on clinical data, including pharmacokinetics, pharmacodynamics, safety, tolerability, dosage convenience, activity, and E‐R relationship, specified in the guidance of Project Optimus. ¹⁰ Therefore, we primarily selected variables related to these factors to investigate risk factors for PMR/PMC associated with dose optimization. Additionally, we included the study design as a variable to examine whether it impacted the PMR/PMC related to dose optimization.

CDx (Y/N), accelerated approval (Y/N), and labeled dose were extracted from the “indication and usage” and “dosage forms and strengths” sections of the United States prescribing information, respectively.

MTD was extracted from one of the following FDA review reports: multiple‐discipline, medical, or clinical pharmacology review reports. We excluded cases where the MTD was determined using a different dosing schedule or regimen rather than the labeled dose.

Adverse reactions leading to treatment discontinuation (%) were extracted from FDA review reports (multiple‐discipline and medical review reports) and from initial or current United States prescribing information.

The phase and type of pivotal study, and the presence of multiple dose evaluations and randomized dose‐ranging trials were extracted from the table of clinical studies from FDA review report.

Data regarding exposure‐efficacy relationship (Y/N) and exposure‐safety relationship (Y/N) were evaluated qualitatively based on the description in the clinical pharmacokinetics section, pharmacometrics review, summary of findings, key questions, and reviewer's analysis in FDA review reports.

To understand the landscape of PMR/PMC on dose optimization, the following were investigated:

1. Annual trends in newly approved oncology drugs for which the FDA requires PMR/PMC on dose optimization.

2. The percentage of newly approved oncology drugs with multiple dose evaluations and randomized dose‐ranging trials comparing among those with PMR/PMC and those without PMR/PMC.

Regarding the presence of multiple dose evaluations for newly approved oncology drugs, only cases in which multiple doses were evaluated during dose expansion in parallel or non‐parallel settings were considered. Cases where multiple dosages or regimens were evaluated in a dose escalation were excluded. Randomized dose‐ranging refers specifically to parallel randomized dose‐ranging trials.

Analysis

The data collected were subjected to descriptive statistics. A multivariate logistic regression model was used to examine the association between PMR/PMC and background factors. All analyses are conducted using SAS ver. 9.4.

Ethics statement

This study did not require institutional review board approval and patient informed consent because it was based on publicly available information and involved no patient records.

RESULTS

Trend in PMR/PMC on dose optimization

From 2010 to 2023, a total of 165 oncology drugs were newly approved by the FDA. Among them, 27 drugs, including those with additional indications, combination formulations, biosimilars, formulation modifications, and those with tentative approvals, were excluded from the analysis. As a result, 138 oncology drugs were included in the final analysis. Of these, 20 drugs (14.5%) had PMR/PMC, while the remaining 118 drugs (85.5%) did not have such requirements (Figure 1 ).

Figure 1.

Number of new oncology drugs approved with or without PMR/PMC on dose optimization. PMC, postmarketing commitment; PMR, postmarketing requirement.

No clear trend was observed in either the number or proportion of PMR/PMC on dose optimization issued for oncology drugs over the past decade. The annual number of approvals for oncology drugs ranged from 2 to 14, showing variability without a consistent increase or decrease (Figure 2 ).

Figure 2.

Annual trends in newly approved oncology drugs with PMR/PMC on dose optimization. PMC, postmarketing commitment; PMR, postmarketing requirement.

Among the 20 oncology drugs with PMR/PMC on dose optimization, three (15%) implemented randomized dose‐ranging and multiple dose evaluation studies, while the remaining 17 (85%) did not include dose evaluation studies in their initial approval package. For the 118 oncology drugs without PMR/PMC on dose optimization, 14 (11.9%) and 23 (19.5%) implemented randomized dose‐ranging and multiple dose evaluation studies, respectively, while the remaining 81 (68.6%) did not include dose evaluation studies in their initial approval package.

No significant difference was observed in the ratio of randomized dose‐ranging and multiple dose evaluation studies among drugs with and without dose optimization PMR/PMC (Figure 3 ).

Figure 3.

Percentage of newly approved oncology drugs with multiple dose evaluations and randomized dose‐ranging trials: comparison between drugs with and without PMR/PMC on dose optimization. PMC, postmarketing commitment; PMR, postmarketing requirement.

Analysis of risk factors for PMR/PMC on dose optimization

The background factors were summarized in Table 1 . As shown in Table 2 and Figure 4 , the relationships between PMR/PMC and background factors were examined using a multivariate logistic regression. By selecting the model with the smallest value for the Akaike information criterion among all possible models, the multivariate logistic model identified the following variables: “MTD as the labeled dose (Y/N),” “adverse reactions leading to treatment discontinuation (%),” “exposure‐efficacy relationship (Y/N),” and “exposure‐safety relationship (Y/N).” Other risk factors were not selected because of lower contributions, multicollinearity with the selected risk factors, and potential issues arising from highly correlated predictors.

Table 1.

Collected data on FDA newly approved oncology drugs from 2010 to 2023

Background factor	N = 138
CDx, n (%)
Yes	31 (22.5)
No	107 (77.5)
AA, n (%)
Yes	65 (47.5)
No	72 (52.6)
MTD, n (%)
Yes	54 (39.1)
No	84 (60.9)
MTD as labeled dose, n (%)
Yes	29 (21.0)
No	109 (79.0)
Phase of pivotal study, n (%)
1	7 (5.0)
1/2	24 (17.4)
2	44 (31.9)
3	62 (44.9)
2/3	1 (0.7)
Type of pivotal study ‐ randomization, n (%)
No	70 (50.7)
Placebo control	31 (22.5)
Active control	35 (25.4)
Dose ranging	2 (1.5)
Presence of dose‐evaluation study, n (%)
No	95 (68.8)
Randomized dose‐ranging	17 (12.3)
Multiple dose evaluation	26 (18.8)
PK proportion, n (%)
No data	6 (4.6)
Yes	97 (73.5)
More than proportional	20 (15.2)
Less than proportional	4 (3.0)
Non‐linear	5 (3.8)
Exposure‐efficacy relationship, n (%)
Yes	68 (49.3)
No	70 (50.7)
Exposure‐safety relationship, n (%)
Yes	89 (64.5)
No	49 (35.5)
Adverse reactions leading to treatment discontinuation (%)
Median	12.0
Min–max	1.2–36.0

AA, accelerate approval; CDx, companion diagnostics; Max, maximum; Min, minimum; MTD, maximum tolerated dose; PK, pharmacokinetics.

Table 2.

Multivariate logistic regression analysis based on a model with the smallest Akaike information criterion among all possible models relating to the PMR/PMC and background factors

Parameter	Estimate (SE)	P value
Intercept	−0.013 (0.072)	0.856
MTD as labeled dose (Yes/No)	0.150 (0.072)	0.038
Exposure‐efficacy relationship (Yes/No)	−0.132 (0.060)	0.028
Exposure‐safety relationship (Yes/No)	0.094 (0.065)	0.150
Adverse reactions leading to treatment discontinuation (%)	0.011 (0.004)	0.010

MTD, maximum tolerated dose; PMC, postmarketing commitment; PMR, postmarketing requirement.

Figure 4.

The relative impact of background factors on the PMR/PMC based on multivariate logistic regression analysis. PMC, postmarketing commitment; PMR, postmarketing requirement.

The analysis showed that the risk for PMR/PMC on dose optimization increased when MTD was the labeled dose, when the percentage of adverse reactions leading to treatment discontinuation was increased, and when an exposure‐safety relationship was identified in the E‐R analysis. Conversely, the risk of PMR/PMC on dose optimization decreased when the exposure‐efficacy relationship was identified in the E‐R analysis.

The distribution of the percentage of adverse reactions leading to treatment discontinuation in cases with and without a PMR/PMC is shown in Figure S1 .

DISCUSSION

Paradigm shift in dose optimization for oncology drugs

Recent advancements in oncology drug development have shifted away from the “higher is better” paradigm. Among the 138 oncology drugs analyzed in this study, MTD was determined for 54 (39.1%), and MTD was selected as the labeled dose for 29 (21%). This indicated that MTD was the labeled dose for approximately half of the oncology drugs for which MTD was determined. These findings are consistent with those reported by Mittapalli, R.K. et al. (2022), which reported that 63% of doses were set below the MTD or maximum studied dose (MSD), with approximately 30% being less than or equal to half of the MTD or MSD. ¹³

Project optimus and its impact

Project Optimus has played a pivotal role in shaping these recent strategies and encouraged the inclusion of randomized dose evaluations in early clinical trials. Historically, dose‐finding studies for oncology drugs have been limited to Phase 1 trials aimed at identifying the MTD, unlike other therapeutic areas, where randomized dose‐ranging trials are more common. Notably, according to our study investigation, only studies on 17 of 138 oncology drugs (12.3%) approved between 2010 and 2023 incorporated randomized dose‐ranging designs. Among these, seven (5%) utilized placebo‐controlled designs, five (3.6%) employed active‐controlled designs, and the remaining five (3.6%) used non‐controlled randomized dose evaluation designs. On the contrary, a recent study investigating the impact of Project Optimus on several pharmaceutical companies revealed that approximately half of the companies now prefer randomized dose‐expansion cohorts with two doses. Multiple doses/schedules, including lower active dose levels, are being evaluated using randomized studies during Phase 1/2 trials. ⁸ These shifts reflect growing alignment with the recommendations outlined by Project Optimus, and it is anticipated that the number of drugs undergoing randomized dose‐ranging studies during early clinical trials will continue to increase in the future.

Importance of pre‐approval dose optimization

Dose optimization in the postmarketing setting posed significant challenges, including delays in patient recruitment caused by challenges to access treatment for the target indication and differences in patient populations between PMR/PMC studies and the approved indication, which can complicate the interpretation of study findings. ¹⁴ These factors underscore the importance of completing dose optimization studies during clinical trials. By refining RP2D before approval, drugs can be delivered to patients with an optimal balance of efficacy and safety, reducing the need for costly and time‐consuming postmarketing adjustments.

Key findings and implications

In the present study, we identified risk factors for PMR/PMC related to dose optimization, providing critical insights for designing more effective clinical trials. We considered PMR/PMC on dose optimization an optimal endpoint for identifying risk factors, as it reflected whether the FDA deemed the data included in the NDA package sufficient to justify the validity of dose optimization. There are many reports published about dose optimization in oncology drugs since Project Optimus was launched in 2021. To the best of our knowledge, this is the first comprehensive study to identify the risk factor for PMR/PMC on dose optimization using logistic regression analysis by evaluating the newly approved oncology drugs in the United States. It objectively identified the important risk factors and quantitatively evaluates the contribution of each risk factor to the PMR/PMC on dose optimization.

In our study, no increasing trend was observed in the number of PMR/PMC related to dose optimization issued during the studied timeframe, even though the importance of dose optimization has been increasing recently. ¹⁵ , ¹⁶ , ¹⁷ This suggests that the timeframe selected for the study was appropriate for assessing risk factors for PMR/PMC related to dose optimization.

Importantly, our study results showed that multiple dose evaluations and randomized dose‐ranging trials alone do not consistently reduce the risk of PMR/PMC on dose optimization. However, a correlation was observed between PMR/PMC on dose optimization and safety‐related factors observed during clinical trials. Our analysis identified three key risk factors associated with PMR/PMC related to dose optimization.

MTD as the labeled dose, the percentage of adverse reactions leading to treatment discontinuation, and the presence of an exposure–safety relationship.

Our results emphasize the need to carefully monitor and address safety signals, particularly the percentage of adverse reactions leading to treatment discontinuation and when MTD is selected as the recommended phase 2 dose (PR2D), to mitigate the potential risks of receiving PMR/PMC related to dose optimization. Furthermore, these findings underscore the critical importance of incorporating E‐R analyses during dose selection in early clinical trials. In previous studies, the importance of E‐R analysis for dose optimization has been emphasized; this is consistent with our observations. ¹⁸ , ¹⁹

Challenges and practical recommendations

The guidance issued by Project Optimus generally encourages the randomized evaluation of the benefit/risk profile across a range of doses before initiating a registration trial.

However, implementing a randomized comparison of at least two doses increases the initial cost of drug development, as it requires a significantly larger study due to the expanded sample size. Additionally, this approach may extend the development timeline, leading to delays in delivering the drug to patients in need. Our results showed that multiple dose evaluations and randomized dose‐ranging trials alone did not consistently reduce the risk of PMR/PMC on dose optimization.

The randomized, parallel dose–response trial recommended by Project Optimus is merely a means to an end; what truly matters is the design to effectively justify the optimal dose. The trial design, including sample size, number of ranging doses, and selected dose levels, is more critical than simply conducting a randomized trial.

Our research suggests that in a case where MTD needs to be selected as RP2D, and there are signs of safety concerns, such as a high rate of adverse reactions leading to treatment discontinuation or a dose‐dependent increase in the frequency and severity of adverse events (AEs), it is essential to design the study carefully to enable effective dose optimization, in line with the Project Optimus guidance to mitigate the risk for PMR/PMC on dose optimization.

If the percentage of adverse reactions leading to treatment discontinuation is increased, even without reaching the MTD, it may be advisable to lower the high‐dose arm in a randomized trial to a level where early discontinuations are within an acceptable range. If maintaining a high dose is necessary to achieve sufficient efficacy, a three‐dose comparison may be a viable option. Conversely, if no risk factors are present and an E‐R efficacy relationship is confirmed, our results showed that the risk of PMR/PMC on dose optimization is relatively low. In such cases, a relatively small‐scale randomized trial with a reduced sample size could be a viable option compared to cases where risk factors are identified. Another option could be accumulating additional data through backfill cohorts in a dose‐escalation study instead of conducting a randomized trial, provided the pharmaceutical company considers that they can adequately justify the optimal dose selection. ²⁰ , ²¹ When implementing this approach, it is crucial to ensure that dose optimization is conducted with careful attention to selecting dose levels to expand. Specifically, the dose level should demonstrate early signs of anti‐tumor activity to avoid exposing patients to ineffective treatments or unreasonable risks. Furthermore, cohort number and size should be carefully determined to capture the general shape of the dose–response relationship and effectively guide dose selection. If the data only covers a narrow range of dosing exposure, it results in a lack of an E‐R relationship, which limits the value of justifying the optimized dose.

Therefore, it is essential to consider trial designs on a case‐by‐case basis and develop an efficient strategy while still acknowledging the value of randomized dose‐ranging evaluations. This approach is also emphasized in the research by Samineni, D. et al., ⁸ who highlight the importance of tailoring randomized dose‐finding trials to each unique scenario.

A robust dose optimization strategy is crucial for refining the RP2D and maximizing both safety and efficacy in subsequent trials. To achieve this, adaptive trial designs may need to be implemented during the Phase 1 study, providing the flexibility for adjustment based on emerging data and ensuring efficient and thorough evaluations. ²² , ²³

Regarding the dose level to be investigated in a dose‐ranging study, non‐regret doses are chosen in most cases. Additionally, it is important to compare the minimal biologically active dose, which is often estimated through pharmacokinetic‐pharmacodynamic modeling, with the highest tolerable dose. This comparison aims to evaluate whether there is a clear differential benefit with acceptable tolerability. To ensure distinct and meaningful comparisons, the selected doses should have non‐overlapping pharmacokinetic exposures, ideally separated by two to three‐fold. ²⁴

Conclusion

Overall, our findings suggest that randomized dose‐ranging and multiple dose evaluation studies alone do not inherently reduce the risk for PMR/PMC. Given the increased developmental costs and extension of the development timeline associated with such trials, careful consideration of their necessity and design is required. For drugs with the identified risk factors; “MTD as the labeled dose,” “an increased percentage of adverse reactions leading to treatment discontinuation,” and “the presence of an exposure‐safety relationship,” it is particularly important to follow the Project Optimus guidance and implement tailored dose optimization strategies.

By identifying the key risk factors for dose optimization PMR/PMC, our findings provide valuable insights into the cases where dose optimization should be performed with greater care.

Refining RP2D before approval enables drugs to be delivered to patients with an optimal balance of efficacy and safety, reducing the need for costly and time‐consuming postmarketing adjustments.

These results offer practical guidance for dose optimization strategies, contributing to a safer and more efficient drug development process. Ultimately, through timely approval of novel therapeutics at optimal doses, patient outcomes can be improved, and the overall efficiency of oncology drug development can be enhanced.

Limitations of the study

Some limitations of our study need further consideration. First, this study was limited to only approved drugs, as data on non‐approved drugs due to inadequate dose optimization are not publicly available. Consequently, the discussions in this study are limited to FDA‐approved drugs, and direct comparisons with non‐approved drugs could not be conducted.

Second, this study is limited by the availability of data on approved drugs and the relatively small sample size of PMR/PMC cases; it provides a foundational analysis of risk factors associated with dose optimization in oncology drug development. These findings offer a unique perspective that can inform future studies and regulatory practices.

Third, due to the limited sample size, we decided not to exclude new modalities (e.g., cell therapies, radiopharmaceuticals); dose‐finding strategies for these modalities may differ from those for small molecules or monoclonal antibodies.

In this study, we provided practical solutions based on risk factors observed during early‐phase trials. However, among the drugs approved from 2010 to 2023, those incorporating randomized dose‐ranging designs remain limited. To further validate the effectiveness of our proposed solutions, future evaluations should focus on newly approved drugs that incorporate randomized dose‐ranging studies.

Funding

No funding was received for this work.

Conflicts of interest

H.K. is an employee of Astellas Pharma Inc., K.T. is an employee of Astellas Pharma Global Development, Inc., All other authors declared no competing interests for this work.

Author contributions

H.K., K.T., and H.M. wrote the manuscript, H.K. and H.M. designed the research; H.K., K.M., and H.M. performed the research; H.K. and K.T. analyzed the data.

Supporting information

Figure S1

Data availability statement

The data presented in this study are available in the article/supplementary materials; further inquiries can be directed to the corresponding author.