Forbes Article Here

In a major speech to cancer doctors Saturday, the outgoing president of the American Society of Clinical Oncology told his colleagues to prepare for a new era in which rapidly advancing genetic technology will change the way cancer is treated for the better –  but  also force doctors to change the way they invent and test drugs and care for patients.

This bold declaration came as the first results of this way of thinking bear fruit. A drug made by pharmaceutical giants Roche and Daiichi Sankyo reduces the risk of death from melanoma by 63% – if the skin cancer tumor has a particular mutation. A Pfizer lung cancer treatment awaiting regulatory approval keeps lung cancer patients alive longer – if they are in the small minority that have a very specific genetic defect. Two clinical trials that have been presented at ASCO ‘s annual meeting in Chicago, which I’m attending, showed it is possible to pick drugs for patients using a panel of genetic tests.

George Sledge, the ASCO president and one of the country’s top breast cancer researchers dramatically and eloquently spelled out what these big changes will mean. This story is a condensation of his prepared remarks, which ASCO shared with me. Mistakes are mine. Credit for marshaling these facts and constructing this argument belongs to Sledge.

In This Story: A new era | Smart cancer, dumb cancer, and magic shotguns|Clinical trials and computer networks

Cancer doctors, Sledge said, are entering an era of “genomic chaos,” a phrase that describes both the genetic madness that makes healthy cells turn into deadly cancers and the instability that incorporating rapidly advancing genetic technology into cancer care will bring. The way clincial trials are run will need to change dramatically. New kinds of electronic health records will need to be created to collect data, inform doctors instantly of new results, and track how good a job physicians are doing. Sledge, whose one-year term as ASCO president is ending, argued doctors need to face up to this blast of new technology.

This is not just geeky talk about cool genetics — it is really a matter of life and death. The speech was dedicated to Sledge’s administrative assistant, who was recently diagnosed with triple-negative breast cancer, one of the hardest varieties of the disease to treat.

A million and a half Americans will be diagnosed with cancer this year, Sledge said. Cancer death rates are falling as a result of less tobacco smoking and better science, but malignancy kills half a million Americans every year. The cost of health care outpaces inflation. According to an article in the Journal of National Cancer Institute, costs for cancer care will increase by 27% or more. The workforce of doctors willing to do clinical trials is dropping; it is hard work, comparatively poorly compensated, and is a “labor of love.” And new cancer drugs fail most of the time. Cancer drugs succeed in late-stage trials 34% of the time, compared to the 60% success rate in respiratory disease, endocrinology, and immune system disorders.

Doctors have been trying to use localized treatments to attack tumors since the 19th century. In the 1940s and early 1950s, the first chemotherapies were used. Cancer drugs like Gleevec and Erbitux, targeted to attack specific mutations in cancer cells, have been the breakthroughs of the past decade. The next step, Sledge says, is DNA-sequencing technology, which is getting cheaper and more powerful at a rate faster than Moore’s Law, which predicts the regular reduction in the cost of microprocessors that has driven the computer revolution. Already, he notes, companies like Knome in Cambridge, Mass., will deliver a genome on a thumb drive.

“So what happens when, a few years from now, a patient walks into a doctor’s office and hands a physician a memory stick loaded with gigabytes of personal genomic data?” Sledge asks. His answer: the flood of data will help doctors and patients, but things will get “very, very complicated.”

DNA sequencing in cancer is very, very new, and is changing faster than any technology doctors have dealt with in the past. The first two human genomes were decoded in 2001 at a cost of over $3 billion, but the first complete sequences of human cancer genomes are only three years old. Now government and international efforts are sequencing several thousands of tumors. A recent paper in the Journal of the American Medical Association showed that sequencing an individual patient’s leukemia revealed a brand new mutation that was used to select a drug that would help that patient.

“We can look forward to a future in which the unraveling of the secrets of the genetic code is commonplace, expected, and routinely drives care,” Sledge said. “But this case, as wonderful as it is as a harbinger of our collective future, is not the whole story. Not every story will end this happily.”

The problem, Sledge says, is that some cancers are smarter than others. Cancers are caused by mutations, changes in the DNA code that somehow turn our own cells into something malignant and deadly. Some cancers have more mutations than others. Sledge showed data from Gaddy Getz, a researcher at the Broad Institute in Cambridge, Mass., that showed that while some childhood cancers and blood tumors have less than one mutation for every thousand bases, colon or lung cancer has close to 10 mutations per thousand DNA bases, and can have 100. Some cancers are 1,000-times more complex than others. In Sledge’s terminology, there are “stupid cancers” and “smart cancers.”

One danger of “stupid” cancer is it makes human beings feel smarter than we are. Novartis’ Gleevec is probably the best drug of the past ten years. It has turned a deadly cancer, chronic myelogenous leukemia, into a chronic disease. When it does fail, other drugs that work in the same way will. The reason is that the cancer is caused by a single rearrangement of DNA in blood cells. It turned out to be an easy target.

“I do not mean to denigrate either the groundbreaking research that led to [Gleevec] or the use of these drugs: this is a true victory for targeted therapy and demonstrates its very real promise for cancer patients,” Sledge said. “But this is a stupid cancer.”

What’s a smart cancer? The kind of lung cancer caused by smoking tobacco, which slowly causes mutations to pile up in the lungs over the course of decades. Researchers in one study were able to determine that patients had one mutation for every three cigarettes they had smoked in their lives. Drugs like Pfizer’s crizotinib or Roche’s Tarceva work best in those cases where the cancer happens to be driven by a single mutation. “This is smart cancer,” Sledge said. Another example is melanoma. Roche and Daiichi Sankyo’s new drug targets a specific gene that drives some forms of the disease. Sledge showed a chilling image of a man whose melanoma, visible as lumps throughout his body, entirely vanished on the new drug – and then came back. This drug hasn’t even hit the market yet and doctors already know of six ways that tumors outwit it.

Worse, the driving mutations for cancer are different in different people. One study of breast cancer patients found most mutations were present in fewer than 5% of patients. So in this new era, every new treatment will be a treatment for a rare disease. Even worse, many new drugs will need to be developed in combination with one another because they won’t work alone.

Sledge says:

Genomic chaos forms the basis for the “smart tumors” that cause so much harm…. These tumors aren’t hard targets because we haven’t found a single “magic bullet.” There will be no “magic bullet” for these tumors because they don’t have a single driving mutation: we need to think in terms of a “magic shotgun,” loaded with pellets aimed at multiple targets in multiple pathways.

So, let’s assume—because it is probably true more often than we would wish—that cancers have multiple drivers, and that to cure a cancer—and let us use the word cure, for our patients deserve no less—that targeting them simultaneously increases benefit. So now imagine cancers with two drivers, requiring two different kinase inhibitors. What is the number of patients we need to study the combination of two new [drugs]?

The answer: for every patient who would actually be entered into a clinical trial to study such a combination of drugs, Sledge calculates that doctors would have to screen 154 patients with the disease. One hundred and fifty three of them wouldn’t have the right combination of genetic mutations in their tumors to respond to the drugs, or wouldn’t participate for other reasons. Right now, he guesses doctors screen 14 patients for every one that gets entered into a clinical study.

Clinical trials, already a tough business, would get ten times harder. Forget testing a combination of three drugs – it would be simply impossible. The challenge, Sledge says, is “daunting.” And right now researchers are not even collecting the biomarkers — what researchers call experimental diagnostic tests — that are needed.

What happens when the next ten patients you see require eight different combinations based on their tumor genomes? Our current system is not designed to handle genomic chaos. It emphasizes single agent trials. It virtually never employs multiple biomarker-driven studies—and biomarkers will be required to validate the genomics. In most studies, biomarker development and analysis are of secondary importance at best. Finally, we have a regulatory apparatus that is ill-suited to the emerging biologic reality.

How will we meet the challenges of the genomic era as a profession? Will we be passive recipients of, or active participants in, this scientific revolution? I would suggest that we must work to meet the challenges of this new genomic era. We need a trained and motivated workforce. We need a vibrant clinical trials system. And we need to pioneer a rapid learning system for oncology.

Every oncologist will need to be “a clinical cancer biologist.” Sledge says. Clinical trials in cancer should soon involve genetic testing for all patients, moving to using DNA-sequencing technology like that made by Illumina as quickly as possible. The clinical trials groups are going to need to organize themselves to recruit patients for multiple studies at once, so that they won’t have to run through a hundred people or more just to get one who has the mutation they are testing. That will change the informed consent process, and the regulations that guide drug trials. “None of this will be easy,” Sledge said, “but all of it is necessary.”

In this new world, he said, ASCO itself would need to move into the field of information technology, becoming what is called a “rapid learning system” in which new science, data on individual patients, powerful real-time computing would change the way that doctors practice. “If the health system for oncology is to succeed,” Sledge said, “all its parts must be healthy and connected.”

Doctors will need real-time access to clinical data from all practice settings. This in turn will require interoperable databases using common terminology. Health information technology should offer on-the-spot decision support to oncologists and patients facing the increasingly complex tapestry revealed by modern genomics. It should provide individualized, ready access to a clinical trials systems. It should support appropriate coverage and reimbursement for services. And it should aggregate data so that we can learn from every patient’s experience.

There are real challenges facing us here, challenges involving cost, patient privacy, data ownership, and the dysfunctional silo mentality of health care systems across the globe. ASCO is not an electronic health records company, but we do believe we have an important organizing role to play in creating the [health information technology] systems of the genomic era.

Part of this system, Sledge argued, will need to include quality measures attached to electronic health records that would be used to track whether individual doctors are doing a good job. “Creating a unified set of measures and standards for our profession is far superior to having a legion of measures imposed on us by a multitude of dueling sources, something that is increasingly – and alarmingly – the case.”

This was a radical dose of future shock. Sledge ended his talk with a paean to medicine, showing the audience a classic painting.

This painting by Goya is entitled “Self-Portrait with Dr. Arrieta.” Though nearly two centuries old, and with nary a PET scan in sight, we have no trouble identifying this as a physician caring for a suffering patient. The words at the bottom of this portrait read in translation: “Goya gives thanks to his friend Arrieta: for the skill and care with which he saved his life during his acute and dangerous illness.” Skill, and care.

As we go forward in the genomic era, we must be willing to look back. Back to the humane standards that have forever guided our profession. Back to our belief that patients always come first. Back to the realization that the pathways forward all flow from that which is best in the human spirit: our thirst for useful knowledge, our compassion for our fellow beings, and our belief in their essential dignity. From our skill and care.

He ended with thank you to his assistant, fighting breast cancer, and to “all of the other cancer patients in the audience today: you give our lives purpose and meaning.” I’d like to thank Sledge for delivering such a well-thought out argument about what  new genetic technologies will mean for doctors and patients.