By Martine Piccart
Personalized medicine is one of the objectives of the European Union’s Horizon 2020 funding program and a high priority on the ESMO board’s strategic agenda: indeed, in oncology, many therapies share a narrow therapeutic index and high cost, implying that prescribing the wrong treatment to the wrong patient at the wrong time can have very negative consequences for both patients and public health systems.
While personalized medicine is the dream of every oncologist and the legitimate expectation of every cancer patient, it can only materialize in the next decade if extensive collaborative research efforts are launched and accompanied by a revolution in the sociology of medical research. The latter implies moving away from small-scale biomarker research projects with prolonged sequestration of data and entering an era of “team science” using high-throughput technologies and broad data sharing.
There is indeed a huge gap between our rapidly growing knowledge about the complex molecular-biological landscape of cancer and the extremely slow development of clinically useful and validated biomarkers. As a result, oncologists face serious challenges in daily clinical practice, including both the over- and undertreatment of patients and the prescription of ineffective therapies.
This remains true despite our living in the era of “targeted therapies”, given that the presence of the target does not equal efficacy of the corresponding targeted drug. A relevant example here is the anti-HER2 monoclonal antibody trastuzumab, which is widely prescribed for women with HER2-positive breast cancer, both in the metastatic and the adjuvant setting. While this agent is considered to be the “star” of targeted drugs in view of its positive impact on breast cancer survival, oncologists know that roughly 50 percent of all patients with HER2-positive tumors do not actually benefit from the drug. The problem is that without validated biomarkers of drug resistance, we’re simply unable to identify those patients upfront.
Greater progress in treatment tailoring has been made in colorectal and lung cancer: the monoclonal antibody cetuximab, directed at the EGFR receptor, is known to be ineffective in the presence of RAF/RAS mutations and, therefore, its administration requires pre-testing of these “negative”-predictive biomarkers. Similarly, the targeted drugs gefitinib and erlotinib are prescribed in non-small cell lung cancer only when EGFR mutations have been identified. In such cases, these drugs can be administered with an almost 70-80 percent chance of antitumor activity.
The problems of potential overtreatment or undertreatment are particularly relevant to breast and colorectal cancers, for which adjuvant chemotherapy has been shown to improve survival “on average”. But there is no “average” patient in the clinic!
Each patient is unique, and therefore each treatment must be unique.
Our ability to identify the unique features of an individual patient’s cancer has been increasing dramatically. The last 15 years have witnessed the development and testing of multi-gene expression signatures in breast cancer, several of which have shown robust prognostic abilities. In other words, they identify a subgroup of women with hormone receptor-positive tumors who are very unlikely to experience a recurrence if treated with adjuvant endocrine therapy only. Two of these signatures − Mammaprint™ and OncotypeDX® – are undergoing prospective validation of their clinical utility in two large trials, MINDACT (EORTC10041/BIG3-04) and TailoRx, which are expected to report results in 2015-2016. Similar attempts to provide level-1 evidence for prognostic biomarkers in other solid tumors have not taken place, illustrating the huge commitment that is needed to change standard clinical practice.
The search for predictive multigene expression signatures − based on RNA expression – has been far less successful, and it is losing ground in favor of next generation DNA sequencing, or molecular screening. The dramatic decrease in the cost of the latter technology renders it accessible to researchers, treating physicians, and wealthy patients. The hope here is to use identify the actionable mutations within a tumor, in other words the “driver” mutations for which targeted drugs exist and that are expected to generate clinical benefit.
A word of caution is needed here: there is considerable danger of cancer micromanagement if this technology leaves the academic world for commercial laboratories too soon, since there is still little evidence for the majority of cancers that drug therapy directed by molecular screening actually improves patient survival.
Academic researchers need to be prepared to embrace complexity, since the likelihood that a single biomarker will explain a cancer’s behavior is extremely low: not only do tumors require in-depth characterization, but their microenvironment and the genetic background of the individual patient (host) need to be considered as well. This multidimensional approach can also benefit from modern molecular imaging techniques, implying the need for much closer collaboration with imaging experts.
Finally, a clear and straightforward regulatory path will need to be created for new diagnostic tests. There is currently very little incentive for companies to invest in molecular diagnostics in the European Union, and personalized oncology will remain an empty shell without a rigorous yet feasible methodology for obtaining regulatory approval of new “multiplex” molecular tests.
Martine J. Piccart, MD, PhD, is Professor of Oncology at the Université Libre de Bruxelles (ULB) and Director of Medicine at the Jules Bordet Institute, in Brussels, Belgium. Earning her medical degrees at the ULB and oncology qualifications in New York and London, she is also member of the Belgian Royal Academy of Medicine. Dr. Piccart is active in numerous professional organizations. Since January 2012, she has been President of the ESMO. She is immediate past-president of the EORTC, president-elect of ECCO and served on the ASCO Board. She is the author of the article ‘Personalised cancer management: closer, but not here yet’, which is published in Annals of Oncology.
Annals of Oncology is a multidisciplinary journal that publishes articles addressing medical oncology, surgery, radiotherapy, paediatric oncology, basic research, and the comprehensive management of patients with malignant diseases. It is published by OUP on behalf of the European Society of Medical Oncology (ESMO).
Subscribe to the OUPblog via email or RSS.
Subscribe to only health and medicine articles on the OUPblog via email or RSS.
Image credit: Medical doctor comforting senior patient. By michaeljung, via iStockphoto.