Personalizing an End to Cancer

April 26, 2013

By Matt Luchette 
April 26, 2013 | BOSTONIn a presentation marked by equal parts melancholy and excitement, Dr. Kevin Hrusovsky spoke at the 2013 Bio-IT World Expo on some of the recent advancements and setbacks in cancer therapeutics. 
“This is the first time in 200 years that children’s life expectancies will be shorter than their parents’,” Hrusovsky remarked at the beginning of his talk, addressing the recent plateau in life expectancy gains. Cancer, the number one cause of death worldwide, is believed to be a major cause of these declining gains. While the global rise of cancer is apparent (in Japan, cancer’s mortality rate has tripled since the 1940s), and the pathophysiology of the disease has been characterized with, in many cases, molecular-level precision, the therapeutics that have sprung from these findings are only effective for 25% of patients.
Hrusovsky, the president of the Life Sciences and Technology division at PerkinElmer, argued that the fault of these first forays in cancer drug development fell from the “one size fits all” paradigm that researchers applied to their patients. As Stanford bioinformatics researcher Dr. Atul Butte once described the problem, “Why did we ever even think that cancers would look like each other, just because the same doctor takes care of them?”
Hrusovsky says the “Next Generation Sequencing” revolution holds promise for a more tailored approach to cancer therapeutics. Customizing a patient’s cancer treatment based on the results of genomic tests can increase the therapy’s efficacy to unprecedented levels. In one study, 81% of metastatic melanoma patients with a mutation in the gene BRAF were found to respond to the drug Vermurafenib. Currently, biomarker tests are required by the FDA for healthcare providers to administer 26 different cancer drugs, and it is recommended for 90 others.
While these promising personalized therapeutics may be the breakthroughs oncologists and patients have been hoping for, according to Hrusovsky, there are two main bottlenecks that prevent these technologies from being widely adopted: sample preparation and informatics.
The preparation of tumor samples has become increasingly cumbersome as the sample volumes decrease. Microfluidic devices that isolate single circulating tumor cells from the blood may one day replace tumor tissue biopsies as the staple of cancer diagnosis. Yet while researchers have found out how to capture these circulating cells, isolating the biomarkers and DNA within these single cells remains a challenge. Hrusovsky hopes that scientists will one day have the “ability to isolate a cell, look at ten biomarkers on that cell, and then open up the cell and do NGS,” maximizing the information content of these microscopic sample sizes.
Furthermore, once these samples are isolated, scientists and healthcare providers lack the sophisticated software required to quickly analyze the results. Informatics has a “user-friendliness need,” Hrusovsky remarked. Quoting Gail Vance, the director of Diagnostic Genomics at Indiana University, in order to analyze genomic results currently, “You have a 5-inch-thick set of papers on your desk for the bioinformatics,” Hrusovsky explained. “That’s where the cost is,” 
Hrusovsky touched on how PerkinElmer hopes to address these bottlenecks. Their program Spotfire, for example, provides an interface for “Big Data visualization of next-generation genome sequencing,” allowing researchers to compare their own data against these large data sets. The software is meant to help researchers uncover unexpected insights and reach actionable conclusions from their data faster through more vivid data visualization.
In addition, Hrusovsky hopes improved imaging technologies could help pathologists make more accurate diagnoses from tissue samples. Vectra, a high-throughput slide analysis system, uses pattern-recognition to analyze biomarker expression or signaling pathways in tissue samples. This software could be a significant step up from the antiquated process of analyzing minute malformations or distinguishing between subtle shades in a tissue stain, rendering more accurate diagnoses “better than a pathologist,” according to Hrusovsky.
Hrusovsky hopes these technologies will create better cancer diagnoses and treatments now, but also improve the healthcare system more generally in the future. “Our healthcare system has very poor productivity,” said Hrusovsky, adding that focusing on prevention instead of treatment could increase life expectancy while decreasing healthcare costs.
However genomic factors are only a part of the puzzle. “Only 20% of the disease can be linked to germline mutations,” Hrusovsky said. The rise in cancer mortality in the past century is believed to be linked to environmental factors, such as stress or diet. 
With these factors in mind, the goal, according to Hrusovsky, is to use genome sequencing technologies to give each new parent a summary of their child’s genomic and environmental health risks. Crowd-sourced initiatives, such as the American Society of Clinical Oncology’s CancerLinQ project, which collects treatment and outcome data from millions of cancer patients worldwide, could help researchers examine the effects of various genomic and environmental factors on cancer development throughout a patient’s lifetime. By personalizing a child’s healthcare from birth, Hrusovsky hopes to see a similar gain in life expectancy to that of cancer patients who received early personalized therapies.
While the uptake of new technologies is typically slow, as Hrusovsky pointed out “the revolution in NGS and imaging is unprecedented in terms of speed.” Eventually, he continued, “new technologies will likely replace many current pathology practices.”