Personalized Therapy at AACR

- Clinical Applications of Genomics and Biomarker Cancer Research

Ari M. VanderWalde, MD, MPH

The paradigm in cancer care is shifting toward individualized therapy. This was the message reiterated by speaker after speaker at the 101st Annual Meeting of the American Association for Cancer Research (AACR), held in Washington DC, April 17-21, 2010.

As Levi Garraway of the Dana-Farber Cancer Institute in Boston, Massachusetts, explained at Monday's plenary session, the hope is that any patient with cancer would have their tumor biopsied and profiled. The profile would then be displayed as a unique genetic signature, which would in turn predict which therapy is most likely to work.

Personalized Therapy

An early example of this approach was described by Edward Kim[2] from the MD Anderson Cancer Center in Houston during Sunday's plenary session when he presented the first results from the BATTLE (Biomarker-integrated Approaches of Targeted Therapy for Lung Cancer Elimination) trial in lung cancer.

In this trial, real-time biopsies were performed and an adaptive approach toward enrollment was used that, for the first time, attempted to predict which molecular markers in individual cancers would respond to which agents.

The BATTLE study identified 4 molecular markers: epidermal growth factor receptor (EGFR), Ras-Raf, RXR/cyclin D1, and vascular endothelial growth factor (VEGF). Patients were then randomly assigned to receive to erlotinib (an EGFR inhibitor), sorafenib (a VEGF inhibitor), vandetanib, or a combination of erlotinib and bexarotene.

The investigators found that 8-week disease control as determined by biopsy before and after this period was a good predictor of overall survival. Specifically, patients who had 8-week disease control had overall survival of 11.3 months vs only 7.3 months in those who did not have 8-week disease control.

Total disease response across all groups was 46%. However, response to sorafenib was very good in all non-EGFR groups; response was especially robust in the KRAS mutation group (79% disease control). Erlotinib yielded the best outcomes in the EGFR mutant group, but response to this agent was poor in the other groups.

Developing and Implementing Predictive Therapeutic Markers for Cancer

During Sunday's plenary session, Martine Piccart-Gebhart, founder and chair of the Breast International Group (BIG), presented the structure of the Neo-BIG program, which is being developed by BIG to accelerate drug development and biomarker discovery in early breast cancer.

The plan is to develop several concurrent clinical trials in the neoadjuvant setting that will evaluate genomics to determine predictive models of efficacy that can later be validated in the adjuvant setting.

Dr. Piccart-Gebhart explained that attempts to bring various tumor markers into the clinical setting have been unsuccessful owing to subjectivity of interpretation of histology or gene expression, as well as lack of specificity of the marker in predicting outcome or response to treatment. However, she noted that "genomic grade" seems to be a promising tool.

She explained that "early readout" of proliferation or signaling in the genetic signature of the tumor may predict response of the tumor to various agents, particularly in luminal-B cancers, a more aggressive genetic group of breast cancer. These data will be incorporated into the next Neo-BIG trial.

Elucidating the Cancer Genome

Monday's plenary session focused on how to make current research elucidating the cancer genome into something meaningful for clinical practice. Most speakers commented that cancer research is in the midst of a paradigm shift. They stated that the field of oncology is moving away from an anatomy- and histology-based view of cancer toward a genomic view the disease.

Bert Vogelstein of The Johns Hopkins University in Baltimore, Maryland, described the astounding progress scientists have made in identifying cancer-specific mutations since the announcement of the completion of the human genome project in 2003.

Dr. Vogelstein's laboratory studied the genomes of 68 tumors and found that genomic alterations in tumor cells occur at different rates and amounts in different tumors, with mutations accumulating as the tumor ages.

Most mutations are "passengers," meaning that they do not fundamentally determine the behavior of the tumor. However, using genetic mapping techniques, the researchers found that about 15% of a tumor is composed of "drivers": 2 or 3 oncogenes and tumor suppressor genes in each tumor, or about 320 "driver" genes across all human cancers combined, that have activity in 12 key cellular or nuclear pathways.

Using the cancer genome, Dr. Vogelstein identified 2 approaches toward development of targeted agents, which he dubbed the "mutant-gene" and "mutant-pathway" approaches. The benefit of the mutant-pathway approach is that 90% of drivers are tumor suppressor mutations that turn cellular genes off.

Because a therapeutic agent cannot turn a gene back on but can disrupt an overexpressed gene, and because resistance patterns are more likely if only 1 gene is targeted, he advocated using the mutant-pathway approach and suggested that it is likely to be more useful than targeting an individual gene.

He concluded with the inspiring assertion that we may now, for the first time, truly understand cancer. The hope is that what we do know can be used to help people.

The Human Genome and Cancer Risk

Rather than describing genomics of tumors themselves, Stephen Chanock[5] of the National Cancer Institute in Bethesda, Maryland, described progress in the identification of causal and contributory mutations in cancer genetic germ-line syndromes. This is the search for such syndromes as BRCA and Li-Fraumeni syndromes, wherein a mutation that affects every cell in the body predisposes for various types of cancer.

Whereas focus had previously been on looking for candidate genes, it is now shifting toward genome-wide association studies, which identify "soft spots" in the genome that seem to confer susceptibility.

For example, multiple cancers have been mapped to 2 distinct regions: 8q24 and 5p15.33. 8q24 is a 600-megabase region toward the middle of the chromosome from the MYC gene that harbors a series of independent markers associated with chronic lymphocytic leukemia and with breast, colorectal, prostate, and bladder cancer.

5p15.33 harbors the TERT-CLPTM1L locus and contains several variants associated with lung, brain, skin, and pancreatic cancers. Dr. Chanock noted that the success of genome-wide association studies has opened new horizons for exploration, and he expressed hope that common variants will be identified and applied in clinical and public health venues.

Conclusions

At the AACR Annual Meeting, cancer researchers expressed hope that the time to use genomics to create personalized cancer care is coming. The insights from their research and from their future planned studies confirmed this hope and left the audience optimistic and inspired.