Clinical development of anticancer drugs can be enhanced using efficacy data of small population clinical trials

Although there are accelerated approval pathways based on data of small populations and surrogate endpoints, the concern that these pathways authorize the use of inefficacious drugs based on limited data from earlier phase clinical trials remains. We retrospectively investigated the efficacy of anticancer drugs, which were approved or whose development was terminated in small and large clinical trials, and verified whether small clinical trials could reflect the results for efficacy in large clinical trials.


| WHAT IS KNOWN AND OBJECTIVE
There are guidelines providing several pathways for the expedition of novel drug development and marketing authorization approval for unmet medical needs in the United States (US), Japan, and the European Union; moreover, a waiver may be granted to conduct confirmatory clinical trials under specific conditions. 1 Many anticancer drugs have been approved based on the data of small populations and surrogate endpoints without the data of confirmatory clinical trials by following an accelerated approval pathway; moreover, withdrawals of accelerated approvals based on the results of post-approval confirmatory trials are limited. [2][3][4] It was stated that the regulatory authority 2 | METHODS

| Clinical trial selection
In this investigation, the efficacy data of small and large clinical trials with comparable study design settings were extracted and compared.
We constructed two databases of compounds approved as anticancer drugs and those unapproved but expected to be useful as anticancer  Unapproved compounds were also within the scope of this investigation and were noted as 'development failure cases'. Unapproved compounds were identified from the repository of phase III clinical trials on ClinicalTrials.gov that enrolled cancer patients and was funded by the industry. The search terms used were 'cancer', status: 'terminated', completion date: 'from 2015 to 2019', study type: 'interventional', and study phase: 'phase III'. Paediatric phase III clinical trials were outside the scope of this study. After the development failure cases were identified, we constructed a dataset of small and large clinical trials that met the pairing criteria. We also confirmed that none of the compounds from the development failure cases had been approved in Japan.

| Data collection
We analysed the median overall survival (OS), median progressionfree survival (PFS), and overall response rate (ORR), which are key efficacy endpoints in almost all clinical trials of anticancer drugs.
These data of the approved anticancer drugs in Japan were extracted from the CTD registered on the PMDA's website or published articles.

| Data analysis
Simple linear regression analysis was performed to assess the relationship between the efficacy data of small and large clinical trials, as well as the predictability of efficacy data in large clinical trials based on that of small clinical trials. Spearman's rank correlation analysis was performed to assess the correlation of the data of each efficacy endpoint between large and small clinical trials. Furthermore, the paired sample t-test was performed to determine whether there were statistically significant differences in the efficacy data between small and large clinical trials. All statistical analyses were performed using EZR software version 1.54 with α = 0.05 as the threshold of statistical significance. 16 Table 1, and similar efficacy data were observed in small and large clinical trials.

| RESULTS AND DISCUSSION
Scatterplots of the median OS, median PFS and ORR in small and large clinical trials are shown in Figure 1.
The results of the simple linear regression analysis, Spearman's rank correlation analysis, and the paired sample t-test are shown in Table 2. A significant linear trend and correlation in the median OS, median PFS, and ORR between small and large clinical trials were observed. The correlation analysis showed that the correlation coefficients between both clinical trials were high for all endpoints, and the correlation coefficients of the median OS and ORR were >0.85, which was the threshold used to assess the validation of surrogate endpoints. 17 The paired sample t-test showed there was a statistically significant difference of median OS between small and large clinical trials (p = 0.02). In contrast, there were no statistically significant differences in the median PFS and ORR between small and large clinical trials, and the mean differences of the median PFS and ORR were À0.102 months and À1.531%, respectively. Additionally, it is worth noting that even when the sample size of the small clinical trial is limited (approximately 30-50), there were no significant differences in the median PFS and ORR. The endpoint used in clinical trials on the basis for granting accelerated approvals was mainly response rate. 3 Therefore, this finding supports the use of the accelerated approval pathway, which is based on promising response rate data obtained from small populations in the early phase. These data suggest that pharmaceutical companies can confidently rely on the efficacy data of small clinical trials and should not expect more efficacious data in future large clinical trials.
It has been reported that phase III trials of several specific compounds failed despite predecessor trials that showed promising results. 18,19 However, significant differences in the median OS, PFS, and ORR between comparable phase II and III clinical trials of anticancer drugs listed by the US Food and Drug Administration (FDA) were not observed, and their phase III trials failed to show the significant benefit of the investigated products, compared with comparators. In the report from Lara and Redman, 19 there was a discrepancy in the concomitant medication setting, which potentially affected the effi-    to regular approval or withdrawn must be conducted on a timely basis once the post-approval confirmatory trial is completed.
Finally, we extracted the data of compounds whose clinical trials in ClinicalTrials.gov were terminated as development failure cases.
However, there were also compounds in phase III clinical trials registered as "completed" whose development was stopped or paused by the pharmaceutical company due to unexpected efficacy and safety results in clinical trials. These cases were not evaluated in our datasets as development failure cases.

| WHAT IS NEW AND CONCLUSION
To summarize, our research suggested that there were no significant differences in median PFS and ORR between small and large clinical trials, and even when the sample size of the clinical trial was