Let’s get personal about cancer

Daily Editorial by Matthias Preusser, Associate Editor of the ESMO Congress Daily - Sunday 30 September 2012

It has been a delight to listen to such excellent research presented so far at ESMO. One story particularly stands out: the incredible advances we have made in just a few years in the molecular profiling and genetic subtyping of cancers. Today, the detection and correlation of clinical responses to genetic variation has become a necessary element of almost every clinical trial.

The ongoing search for key mutations and hence molecular targets for therapies fuels the research community. Every tumorigenic mutation offers a potential target for treatment. At the same time, the increase in validated predictive biomarkers will help clinicians to select the most suitable drugs and treatment regimens for their patients. Many patients today are experiencing good response rates and longer survival thanks to detailed genetic subtyping of their tumours.

The ‘personalisation’ of medicine – still not the production of individually tailored therapeutics but at least an informed choice of agents and regimens for every individual – is becoming a reality.

In my area of brain tumours, we heard in the special symposium on molecular neuro-oncology about several genetic mutations and characteristics, which are prognostic, but also predictive for improved treatment outcomes. Studies have shown that testing for isocitrate dehydrogenase (IDH) mutations is not just helpful for the diagnosis of grade II and grade III anaplastic glioma, but also has significant prognostic implications. For glioblastoma, MGMT has been identified as important prognostic and for some patient populations, a predictive marker.

But here is the big question: do molecular biomarkers really help with decision making in the clinical setting? Of course, we know that genetic profiling is standard for some cancers such as HER2 testing for breast cancer, or KRAS mutations testing for colorectal cancer. But for most cancers, molecular profiling is still not clinically validated, although there is plenty of encouraging data emerging, including presentations here at ESMO, to suggest that this could change in the near future.

Can we look forward to a day when tumours are fully profiled for all known biomarkers as standard practice? And if they are, will clinicians know the full extent and implications of the results they receive and the nuances that particular combinations of markers signify in terms of treatment? Or will we continue to stick rigidly to standard therapies – perhaps afraid of the risks and repercussions of following non-standard treatments?

Fortunately, for glioblastoma, a deadly cancer, which we have not understood biologically at all for a long time, promising biomarkers of response to therapy are emerging. Most importantly, MGMT promoter methylation status has been singled out as a predictive marker for response (or non-response) to chemotherapy with temozolomide in elderly glioblastoma patients. Another potential ‘druggable’ molecular alteration is a specific type of epidermal growth factor receptor mutation, EGFRvIII and this is being explored in randomised clinical trials investigating vaccination strategies.
The number of validated biomarkers for virtually every cancer is set to explode and there are important advances not only in adult, but also in paediatric neurooncology. Yesterday’s symposium heard about the molecular heterogeneitiy of medulloblastoma. Beside known alterations (CTNNB1, PTCH1, MLL2, SMARCA4), an integrative deep-sequencing analysis has revealed that several genes not previously implicated in the disease (DDX3X, CTDNEP1, KDM6A) are recurrently mutated, often in subgroup-specific patterns; many of these genes are involved in chromatin remodeling. Hopefully, these advances will soon lead to the development of targeted agents for paediatric brain cancer subtypes, for example sonic hedgehog inhibitors in medulloblastoma patients.

But even as the pool of biomarkers expands, we observe that most treatment decisions for brain tumour patients are still based on age and performance status today, even though biomarkers with validated clinical performance have been identified. Indeed, this pattern is replicated across many therapy areas.

In my opinion, one major obstacle to bringing new biomarkers into use in everyday clinical work and for the benefit of patients is the lack of studies validating our laboratory assays. It is important to understand that assays that can separate patient populations in large studies often turn out not to be sensitive and specific enough to be used for treatment decisions in the individual patient sitting in front of you today. We need more high quality studies on the analytical performance of test methods to identify the best assays for a given biomarker. At present we are missing a selection process for biomarkers, which is similar to the process novel drugs have to go through, i.e. phase 1, 2, 3 trials. Clinical and tissue-based researchers need to come together to solve this problem in an interdisciplinary effort.

As a young oncologist I am thrilled about the speed at which oncology is moving forward at the moment and that becomes clear once again here at the ESMO 2012 meeting. Attending this meeting incredibly motivates me to continue being part and research in oncology to help develop this fascinating field so that we can ultimately defeat cancer some day. I hope many of my colleagues share this feeling and enjoy the congress here in the beautiful city of Vienna!