Companion Diagnostics – meeting scientific and regulatory needs
The majority of drugs are currently prescribed empirically, but advances in research focusing on the molecular mechanisms of specific diseases have opened new pathways to matching patients with drugs that are more likely to be effective and safe. The term ‘personalised medicine’ is currently used to summarise different levels of this process. The route to approval for a drug/diagnostic combination, however, continues to be rocky.
In extreme cases, individualised medicines vary from patient to patient. Such therapeutic concepts include, for example, cell-based cancer vaccines. A less extreme level of personalisation is the concept of stratified medicine. Here, a patient can be classified according to a specific cohort that shows an efficient differential response to a given drug. This stratification is achieved through the use of a clinical biomarker that has been correlated to a differential response.
Improving on empirical medicine
To be implemented in stratified medicine as a companion diagnostic, a clinical biomarker should be predictive for the efficacy and safety of the treatment. The most frequently-used predictive clinical biomarker concepts at the moment include techniques such as gene-expression pattern, individual protein expression (on histology) and even imaging. A stratified medicine regime adds a clinical biomarker assessment to empirical medicine practice, and associates a patient with the best therapy available with respect to efficacy and safety. That helps to avoid treatments with toxic side effects and low or even no benefit for the patient. For the application of stratified medicine, the following three prerequisites are required: I) underlying disease with variability reflecting multifactorial aetiology, II) multiple treatment options and III) a predictive clinical biomarker.
Cetuximab – a pioneer in the field
Successful anti-cancer treatment requires not only a potent drug, but also finding the right drug for a specific patient in a timely manner. The concept of stratified medicine has so far therefore focused on cancer treatment where both quality and time are critical. In 2004, the approval of Cetuximab by the US Food & Drug Administration (FDA) added another option for the treatment of colorectal cancer through chemotherapy. However, subsequent clinical experience indicated benefit for patients in only 10 -20% of the cases. In the years that followed, an intensive search for a predictive biomarker led to stratification of the treatment with Cetuximab. Retrospective analysis revealed a lack of benefit associated with Cetuximab treatment in colorectal cancer patients harboring mutations in the KRAS gene. These studies resulted in a label restriction from both the FDA and the EMA (in 2009) for patients without a KRAS mutation. In Europe, a diagnostic test for detection of the KRAS mutation recently (2011) received a CE label.
Making the case-by-case decision
Despite the fact that retrospective clinical biomarker analysis with subsequent development of companion diagnostic tests remains common practice, both the EMA and FDA favor co-development of drug and diagnostic. The concept of co-development envisages the approval of a diagnostic test at the time of drug approval. One of the most important consequences of this concept is that screening for suitable predictive biomarkers should occur at a very early stage of drug development, with implementation of the biomarker hypothesis in clinical trial design. In recent years, both the EMA and FDA have noted the need for regulations and clear concepts to support the development of stratified/personalised medicines along with appropriate diagnostic tests. However, in Europe the situation remains uncertain. Most importantly, current regulations and concept papers do not allow for a coordinated approval of a drug and its associated companion diagnostic. And in contrast to the situation in the US, cross-labeling of drug and companion diagnostics in Europe is not required. Additionally, self-certification by IVD manufacturers and a lack of harmonised quality regulations in clinical laboratories providing companion diagnostic testing may serve to widen the quality window of test results. Overall, the current regulatory landscape in Europe and in the US implements a case-by-case approach for the development of companion diagnostics that is in line with the advice for a very early consultation with the regulatory agency.