Discussion

This study describes the clinical and regulatory scenarios of new FDA-approved drugs and indications for the four most incident cancers.

We demonstrate that drug approvals for lung, breast, colorectal, and prostate cancer increased from 2016 to 2018. The number of FDA-approved oncology drugs has changed over time. From the 1950s to the 1970s, the mean rate of FDA-approved new molecules was 1.1/year, reaching 4.9/year in the 2000s. The approval rate continued to increase until 2011–13, reaching an average value of 10.3/year (Reference Kinch7). According to the annual report on FDA drug approval, issued by the Center for Drug Evaluation and Research (CDER), 2018 had the highest number of new approvals in the preceding 10 years, with fifty-nine new molecules approved (5). Of these, 73 percent were in one or more expedited categories: fast track, breakthrough, priority review, and/or accelerated approval. These results are very similar to our study, with 78 percent of drugs in one or more expedited categories. In 2018, CDER also announced a pilot program, to further accelerate approval, based on real-time clinical trial information, before the company's official submission. This is likely to decrease the time for drug approval, impacting directly on healthcare systems. A report from the WHO estimated the global expenditure on anticancer drugs and related supportive care to be USD 133 billion in 2017. This report also discusses “value-based pricing” and the strengths and uncertainties about this new payment model (8;Reference Sun, Wei, Zhou and Wang9).

Our study showed that 86 percent of the studies for new drug/indication approvals by the FDA used a surrogate end point, and only 14 percent used overall survival as the primary outcome. Surrogate end point measures have been widely used in oncology. From 2009 to 2014, 66 percent of oncology drug approvals by the FDA were based on surrogate end points (Reference Kim and Prasad10). Surrogates seem to have low to moderate correlation with patient-important outcomes, such as overall survival (Reference Haslam, Hey, Gill and Prasad11;Reference Prasad, Kim, Burotto and Vandross12) and improvements in quality of life (Reference Davis, Naci, Gurpinar, Poplavska, Pinto and Aggarwal13;Reference Salas-Vega, Iliopoulos and Mossialos14).

Accelerated approvals have allowed submissions based on intermediate outcomes since 1992 for serious or life-threatening diseases (4). Our results from 2016 to 2018, regarding the most incident cancers, demonstrate that 25 percent of the total approvals used the accelerated pathway. Pharmaceutical companies that submitted a drug though accelerated approval are required to publish confirmatory post-marketing studies with patient-important outcomes. However, some drug approvals have taken up to 7.4 years, and in the worst case 12.9 years, to be converted into regular approvals, after publication of the delayed confirmatory data (Reference Johnson, Ning and Farrell15). A study published in 2019 showed that most confirmatory data also uses surrogate end points. Furthermore, only 21 percent showed confirmed benefit, and only marginal gains (20 percent) were seen in overall survival (Reference Gyawali, Hey and Kesselheim16).

The FDA label designations, breakthrough, and priority review, allow for more flexible requirements in drug approvals, and phase 2 and single-arm trials can be used as a single reference for drug approval. A meta-epidemiologic study published in 2019 evaluated all single-arm studies used for FDA approvals and demonstrated that the number of single-arm studies leading to approval increased substantially. Furthermore, the majority was based on studies with small sample sizes, and did not provide any other verifying studies (Reference Ladanie, Speich and Briel17). Recently, the EMA and FDA have been increasing the number of approvals based on non-randomized trials. From 2005 to 2017, the FDA and EMA approved 98 percent and 79 percent of these submissions, respectively. However, most FDA approvals required post-approval confirmatory data, and a quarter of EMA approvals were conditional on post-marketing RCTs or additional non-randomized trials (Reference Goring, Taylor and Müller18). In order to improve data for drug approval, the FDA recently launched a guideline for real-world evidence submission to generate information on effectiveness (19). However, some authors are concerned about the lack of standardized data collection and the allowance of retrospective studies (Reference Nabhan, Klink and Prasad20).

In the current study, of the thirty-six new drugs and indications approved in the 3-year period, 53 percent were for treatment of advanced lung cancer. Another study published in 2017 analyzed FDA approval rates from 1964 to 2014, showing that breast cancer drugs were responsible for the majority of drug approvals for solid tumors (twenty-seven new approvals) in the last 50 years, followed by lung cancer drugs with seventeen approvals (8;Reference Sun, Wei, Zhou and Wang9). The authors also showed an increase in targeted therapy approvals, mainly PKIs and monotherapies, in the 1990s (8;9). These results are consistent with our findings that PKIs were the most frequently approved drug class, being primarily indicated for lung cancer, and as oral monotherapies. Oral anticancer drugs are more convenient for patients; however, they can result in significantly increased healthcare costs, depending on insurance reimbursement (Reference Shen, Chien and Geynisman21).

Monoclonal antibodies, including immunotherapies, have been increasingly approved over the years, becoming a therapeutic option for most cancers. Currently, there are many approvals based on studies combining chemotherapy and immunotherapy in the intervention arm, thus increasing the number of indications for previously approved drugs (Reference Blumenthal and Pazdur22). The assessment of oncologic drugs, many of them approved for different indications, has been a challenge for HTA agencies, such as NICE (UK) (Reference Dillon and Landells23).

Among the approved drugs assessed in this study 73 percent had a biomarker in the indication. These were primarily genetic alterations. Jardim et al. (Reference Jardim, Schwaederle, Hong and Kurzrock24) evaluated biomarker-based selection strategies in drug development and demonstrated that targeted drugs were approved faster than cytotoxic drugs (Reference Blumenthal and Pazdur22). A French study with 17,664 NSCLC patients showed that genetic alterations changed 50 percent of the first-line treatment choices (Reference Dillon and Landells23). In our study, targeted drugs for lung cancer were mostly single oral agents. Biochemical tests used to identify or measure biomarkers are expensive, and their reimbursement policy remains unclear in public and private health care (Reference Barlesi, Mazieres and Merlio25).

Four tumor-agnostic therapies were approved since 2017. These therapies are defined by the NCI Dictionary as “therapy that uses drugs or other substances to treat cancer based on the cancer's genetic and molecular features without regard to the cancer type or where the cancer started in the body. Tumor-agnostic therapy uses the same drug to treat all cancer types that have the genetic mutation (change) or biomarker that is targeted by the drug” (Reference Lemery, Keegan and Pazdur26). HTA agencies have to be prepared to assess the limitations and quality of studies in which the subject populations are not defined by a specific disease, but by the mutations or expression of a particular feature. This represents a challenge for evidence-based medicine, especially because the majority of these therapies could be classified as rare disease treatment (Reference Lemery, Keegan and Pazdur26). There were no cytotoxic drugs approved during the period studied. Can we foresee the end of cytotoxic drug research and expect a new era of anticancer drugs?

With respect to biosimilars in oncology, a filgrastim biosimilar was previously approved as a therapy for hematologic complications after cytotoxic drugs (Reference Cornes27). For cancer treatment, bevacizumab was the first biosimilar, as shown in our study, followed by the trastuzumab biosimilar products. There is a global discussion about market access and cost-saving in oncology, and cancer biosimilar drugs may reduce the cost of and increase patient access to treatment (Reference Cornes27;Reference Schleicher and Seidman28).