Oncology drug lag in Japan: has it improved over the last decade?

Existing studies and statistics on the drug lag between Japan and the United States (US) for anti-cancer drugs indicate that it has decreased, whereas more drugs are left unapproved in Japan. This study aimed to quantify the impact of unapproved drugs on the drug lag. Information on 136 anti-cancer drugs approved in the US between 2011 and 2022 was collected. The approval lag, defined as the number of days from the date of approval in the US to the date of approval in Japan, was calculated for all selected drugs, and the median was calculated using the Kaplan–Meier method. The approval lag for drugs not approved in Japan was treated as censored data. Factors potentially associated with the approval lag were explored using Cox regression analysis. The median approval lags for the first half-period (2011–2016) and the last half-period (2017–2022) were 961 days (2.6 years) and 1547 days (4.2 years), respectively (Log-rank test: p = 0.0687). The participation of Japan in the global pivotal trial was associated with a shorter approval lag, and new drug applications by non-Japanese companies that did not rank in the global sales top 20 were associated with a longer approval lag. Drug lag has not decreased over the last decade. The percentage of pivotal trials for US approval that included Japan has increased but should be further increased in the future. Japan may require a scheme to encourage smaller non-Japanese companies to include Japan in their global clinical development plan.


Introduction
Accessibility to innovative medicines, particularly those for life-threatening diseases, such as cancer, is an important public health concern.In most cases, there is a time lag from when a drug is granted marketing authorisation in the first country to when it obtains marketing authorisation in Japan [1].This time lag causes an issue that Japanese patients cannot use many drugs that are already available outside Japan.This issue is called "drug lag," and has been attracting public attention.Many studies have been conducted to understand the status and trends of drug lag.In 2011, one study reported that the approval lag time between the United States (US) and Japan for anti-cancer drugs approved in Japan between 2000 and 2009 was 29.9 months, presumably because of delays in the initiation of drug development and a longer regulatory review period in Japan [2].
From the late 2000s to the early 2010s, the Ministry of Health, Labour and Welfare (MHLW) issued several guidelines and notifications to promote the use of multinational clinical trials to support the marketing authorisation of new drugs in Japan.One of the objectives of promoting multinational clinical trials is to alleviate drug lag.Since then, several reports have suggested that drug lags have diminished [3,4].These reports followed the median lag time between marketing authorisation in the US and Japan, and concluded that it decreased over time.The Pharmaceuticals and Medical Devices Agency (PMDA), the regulatory authority of Japan for new drug reviews, has been publishing a summary of annual median drug lags since 2006 [5].The median drug lag in 2006 was 2.4 years, which decreased to 0.4 years in 2021.

3
A recent report discussed drug lag from a different angle [6], focusing on drugs that were approved outside Japan (e.g. the US) but not in Japan.The percentage of unapproved drugs has been increasing since 2016, creating an access gap to new medicines for patients in Japan.
Many existing studies and statistics on drug lag, including the summary provided by the PMDA, have only focused on the approval lag for drugs that had been approved in Japan.With this approach, the effect of unapproved drugs on drug accessibility was not considered, and the drug lag might have been underestimated.
The present study aimed to quantify the approval lag of anti-cancer drugs approved in the US over a given time period using survival analyses.Analyses were performed to understand whether and how drug lags in Japan have changed over time.The same analyses were performed for drug lag between the US and the European Union (EU) to compare the status of drug lag in Japan with that in the EU.Factors that might have affected drug lag were also explored.

Criteria for drug selection
This study included new molecular entities (NMEs) that were approved for cancer treatment by the Center for Drug Evaluation and Research (CDER), part of the US Food and Drug Administration (FDA), between 2011 and 2022.Drugs reviewed and approved by the Center for Biologics Evaluation and Research were excluded, as they follow different laws/regulations for development and approval, and the trend in terms of drug lag might be different.Products for cancer care with no anti-tumour effects, such as drugs to alleviate the side effects of chemotherapy and drugs that are used for diagnostic purposes, were not included.

Data sources
Data were collected from the websites of the regulatory agencies.The list of NMEs approved by the CDER is available on the FDA website.
The FDA's review report, PMDA's review report (if the product was approved in Japan), and clinical trial information on ClinicalTrials.govwere the main sources of key information such as the start date of the first-in-human clinical trial, start date of the first-in-Japan clinical trial, start date of the pivotal trial, participating countries/regions in the trial, date of new drug application (NDA), date of approval, and applicant (company) of the NDA.Publications from clinical trials or company press releases were also used to supplement missing information.When the date information was incomplete and only year and month were available, the date was assumed to be "15th."

Outcome measures and statistical analyses
The US-Japan approval lag, defined as the number of days from the date of approval in the US to the date of approval in Japan, was calculated for products that met the selection criteria and was used as a primary outcome measure to represent the drug lag between the US and Japan.The median approval lag was calculated using the Kaplan-Meier method.The approval lag for products not approved in Japan was treated as censored data in the analysis.If the date of approval in Japan was earlier than that in the US, the approval lag was handled as 0 (considered no approval lag for such products).
To assess the change in approval lag during the study period, the median approval lags of the first half-period (2011-2016) and last half-period (2017-2022) were also calculated.The cut-off points for approval information for the first and last half-periods were set on 31 December 2016, and 31 December 2022, respectively, so that the observation periods for both groups were the same.
Similarly, the approval lag between the US and the EU was calculated by subtracting the date of approval in the US from the date of approval in the EU.The values of US-Japan approval lag and US-EU approval lags were compared for reference.The review period, defined as the number of days from the NDA to approval, was also calculated for the US, EU, and Japan.
In addition to the approval lag for all drugs approved in the US, the US-Japan approval lag for drugs approved in Japan and the US-EU approval lag for drugs approved in the EU were calculated to assess the impact of the number of unapproved drugs on the Kaplan-Meier-based approval lag calculation.
Factors potentially associated with the US-Japan approval lag were explored using Cox regression analysis.The following factors were analysed as explanatory variables: US NDA applicant company type (large non-Japanese company, smaller-non-Japanese company, or Japanese company), treatment category (molecular targeted therapy including antibody drug conjugate, chemotherapy, immunomodulatory therapy, or hormonal therapy), molecular size (small or large), US accelerated approval (no or yes), EU conditional approval (no or yes), primary endpoint in pivotal trial for US approval (overall survival or event free survival, or response rate), different pivotal trials for US and EU approvals (no or yes), different applicants for US and EU NDAs (no or yes), different primary endpoints for US and EU approvals (no or yes), inclusion of Japan in pivotal trial for US approval (no or yes), inclusion of Japan in first-in-human trial (no or yes), number of patients in pivotal trial for US approval (continuous variable), lag between start of first-in-human trial and start of first-in-Japan trial (continuous variable), and lag between start of pivotal trial for US approval and start of first-in-Japan trial (continuous variable).Potential factors were screened using univariable analysis.Variables with a p-value of less than 0.05 were considered for multivariable analysis.Before selecting variables for multivariable analysis, the correlations between explanatory variables were reviewed to avoid multicollinearity.Statistical analyses were performed using StatsDirect version 3.3.6(StatsDirect).

Drug products selected
From 01 January 2011 to 31 December 2022, 498 NMEs were approved by the CDER (Fig. 1), among which 145 were approved for cancer indications and 136 for cancer treatment.In Japan, 105 NMEs were approved for cancer treatment between 2011 and 2022; 18 of them were approved in the US before 2011, 12 were not approved in the US, and remaining 75 were approved in the US between 2011 and 2022, which were included in the analyses.In the EU, 124 NMEs were approved for cancer treatment between 2011 and 2022.Of which, 6 were approved in the US before 2011, 9 were not approved in the US, and remaining 109 were approved in the US between 2011 and 2022, which were included in the analyses.

Descriptive statistics of metrics surrounding drug lag
The Kaplan-Meier curves for time to approval in Japan after US approval for the analysed anti-cancer drugs are presented in Fig. 2. The median for all 136 products was 1278 days (95% CI 971-1384).The medians for the first and last half periods were 961 days (95% CI 755-1167) and 1547 days (95% CI 1194-1899), respectively.The results suggested that the FDA-PMDA approval lag in the last half of the period was longer than that in the first half, although the difference was not statistically significant (Log-rank test: p = 0.0687).
The Kaplan-Meier curves for the time to submission in Japan after US submission are presented in Fig. 3.The submission lag in the last half-period was longer than that in the first half-period.The observation was similar to the approval lag, but the difference in the submission lag was statistically significant (Log-rank test: p = 0.0257).
The Kaplan-Meier curves for the time to approval in the EU after US approval are presented in Fig. 4 and 5.The curve for all periods is presented together with the curve for the time to approval in Japan after the US approval.The median US-EU approval lag for all 136 products was 316 days (95% CI 260-372), which was significantly shorter than the US-Japan approval lag in the same period.The medians for the first and the last half periods were 237 days (95% CI 170-304) and 395 days (95% CI 255-535), respectively, and the approval lag in the last half was significantly longer (Log-rank test: p = 0.0007).
The median review periods (calculated by subtracting the date of the NDA from the date of approval) in the US, EU, and Japan are summarised in Table 1.There were no notable differences in the median review periods between the first half-period (2011-2016) and last half-period (2017-2022) for any of the three countries/regions.
The median of US-Japan approval lag for products approved in Japan showed a different trend (Table 2).The median approval lag was shorter in the latter half of the period, suggesting that the approval lag for approved drugs decreased.The proportion of approved drugs was lower during the latter half of the study period.The median US-EU approval lag for products approved in the EU was Fig. 1 Flowchart of selecting anti-cancer drugs investigated in this study larger in the latter half, which was the opposite of what was observed in Japan.However, there was an increasing trend in the percentage of unapproved drugs, which is similar to that observed in Japan.

Review of potential factors associated with the US-Japan approval lag
The results of the univariable Cox regression analyses to screen for potential factors are summarised in Table 3. Six variables with a p-value of less than 0.05 were considered for the multivariable analysis, but the lag between the start of the pivotal trial for FDA approval and the start of the first-in-Japan trial was excluded as it was highly correlated with other two variables and potentially caused multicollinearity.As a result, five variables were included in the multivariable analysis.The results from multivariable analysis indicated a significant association of "Japan's participation in pivotal trial for US FDA approval" and "US NDA applicant company type" with the approval lag (Table 4).For these variables, the number of products by category by year is summarised in bar charts (Figs. 6 and 7 for all drugs, Figs. 8 and 9 for drugs approved in Japan).

Discussion
The data from our study indicate that drug lag has not decreased over the last decade, but has been increasing.Meanwhile, the approval lag of drugs approved in Japan has been shorter for more recent periods, which is consistent with existing studies on drug lag [3,4].This suggests that the efforts made by the Japanese government/regulatory authorities, academia, and industry to tackle drug lag were at least partly successful.For example, the median review period by the PMDA was 18-22 months in the late 2000s but accelerated to 9-12 months in the early 2010s [7].The review period of the PMDA in the present study was approximately 9-10 months throughout the study period.Considering the median review periods of other regulatory authorities, the performance of PMDA was competitive.In addition, data from multinational clinical trials were used in the NDA packages of many products approved in Japan in 2011 or later, and the approval lag for such products was generally shorter [8].
The main reason for the increase in the drug lag in the present study was the increase in the proportion of unapproved drugs in Japan.Many of these unapproved drugs were developed by companies that did not have Japanese branches or subsidiaries, and Japan did not seem to be included in global clinical development [6,9].Although the present study did not directly investigate the impact of the absence of Japanese branches on the approval lag, Cox regression analyses revealed that products developed by non-Japanese companies that did not rank in the global top 20 by sales had a smaller chance of product approval in Japan than those developed by the global top 20 companies.As the proportion of products developed by emerging biopharma companies has been soaring [10], Japan may need a specific scheme to encourage such companies to include Japan in their clinical development plan, or to promote the licensing of promising compounds from them.
Interestingly, the proportion of unapproved drugs in Europe increased during the same period.The majority of these emerging companies are based in the US (in the present study, the US NDA applicant was "smaller non-Japanese companies" for 58 drugs.Among these, 44 drugs were developed by US-based companies and 14 drugs were developed by European companies); therefore, they might have been focusing more on the US market than the European market.
Another factor considered to significantly impact the approval lag was the participation of Japan in a pivotal trial for US approval.This factor is known to have a significant impact on the approval lag [11][12][13], and the results of the current study confirmed that this was also true when the approval lag of unapproved drugs was included in the analysis.Among the factors investigated in this study, this was considered the most significant in determining the approval lag based on the hazard ratio.
The path towards the marketing authorisation of pharmaceutical products in Japan is broadly classified into following three categories: multinational clinical trials, bridging, and domestic development strategies.The bridging strategy, which involves performing a Japanese trial to show similarities in efficacy and safety in the Japanese population in reference to the global pivotal trial, was the main strategy for the marketing authorisation in Japan in the 2000s.Since guidelines on multinational clinical trials were issued in the late 2000s, multinational clinical trials have become a mainstream clinical development strategy in Japan [3].
The data from the present study show that the percentage of US pivotal trials that included Japan has increased.However, this was still less than half in the last half-period (2017-2022) and only two out of 11 products approved in 2022.This is not sufficient to compensate for the decrease in the number of drugs approved through bridging and domestic development strategies.Participation in global pivotal trials or global clinical development has become even more important to prevent the expansion of the drug lag, as many companies have shifted the focus of development towards rare tumours and rare fractions [10,14].It is often difficult to conduct a Japan-only clinical trial for such a small population, and if Japan does not participate in a pivotal global trial, the drug is likely to be left unapproved.
Japan's participation in the first-in-human trial was identified as a factor potentially affecting the approval lag in univariable analysis; however, this was not confirmed in multivariable analysis.This might be because participation in the first-in-human trial, defined as participation in the initial dose escalation part and not including the expansion part, was confirmed only in a small fraction of products Table 2 Median United states (US)-Japan approval lag of drugs approved in Japan and median US-European Union (EU) approval of drugs approved in the EU The drugs included in the US approval year 2011-2016 category were those that were approved in the US between 2011 and 2016 AND approved in Japan between 2011 and 2016, i.e., the ones approved in the US between 2011 and 2016 but approved in Japan in 2017 or later were not included in this category, in order to apply the same follow-up period for two groups (2011-2016 vs 2017-2022).Therefore, the sum of the number of drugs in these two groups is not equal to the number of drugs in the entire period (seven out of 87 products included in the multivariable analysis), potentially resulting in insufficient statistical power.Regardless of whether it decreases the approval lag, participation in first-in-human trials is an important opportunity to generate data on Japanese patients from the early phase and potentially provide input to later-phase global clinical development.
In 2021, the MHLW issued a revised version of the guidelines for clinical evaluation of anti-cancer drugs [15].The revised version addresses the ever-increasing cases of nontraditional clinical development approaches, such as using the Phase I expansion part to explore the efficacy and skipping Phase II trials or planning Phases II and III as one seamless trial.It is expected that the revised guidelines will Table 3 Univariable Cox regression analyses on the association of potential explanatory variables and approval lag NDA new drug application a One drug was given both accelerated approval and regular approval at the same time.The drug was excluded from this analysis b Drugs that were not approved in the EU were excluded from the analysis c Drugs for which the primary endpoint was not OS/EFS or response were excluded from the analysis The present study has some limitations.The follow-up period for drugs approved by the FDA in recent years was generally short.Therefore, the results may differ if the same analysis is performed for the same drugs in a few years.The sample size for the multivariable Cox regression analysis was relatively small owing to the unavailability of data.
In conclusion, the approval lag for drugs approved in Japan has decreased over the last decade, but the overall drug lag has been increasing, owing to an increase in unapproved drugs in Japan.The approval lag for drugs developed by small non-Japanese companies was longer, and the increase in the number of such drugs may be one reason for the increasing number of unapproved drugs.Japan's participation in global pivotal trials has increased, but this needs to be further increased, as countries that do not participate in global pivotal trials are likely to be left out.

Fig. 4 Fig. 5 2022 Table 1
Fig. 4 Kaplan-Meier curves for time to approval in Japan after the United States (US) approval, and time to approval in the European Union (EU) after the US approval, for anti-cancer drugs approved in the US between 2011 and 2022

Table 4
Multivariable Cox regression analysis on the association of explanatory variables and approval lagNDA new drug applicationIn the multivariable analysis, drugs for which data are available for all five items were included (n = 87) Number of anti-cancer drugs approved in the United States (US) and Japan with or without the participation of Japan in the pivotal trial, by the year of approval in the US Number of anti-cancer drugs approved in the United States (US) and Japan by type of new drug application (NDA) applicant and by the year of approval in the US.*Criterion for large company: global sales ranking top 20 in the year 2020