Science Editor

George Sledge, outgoing president of the American Society of Clinical Oncology (ASCO), has a rather catchy way of summarizing where cancer treatment needs to be going.

In a nutshell, he told his audience last week at the ASCO annual meeting, oncologists and drug developers are now smarter than "stupid cancers." And they need to become smarter than "smart cancers."

More technically speaking, the cancer field needs to move from the era of targeted therapies to the genomic era.

Targeted therapies, Sledge said, work on tumors with a single dominant mutation, where monotherapy is effective and resistance emerges relatively rarely, relatively late, and can as likely as not be overcome with the next monotherapy.

But the most challenging cancers, Sledge said, are "smart cancers" whose defining feature is their overall high mutational load, rather than one single driving mutation. Lung cancer is the poster child for such cancers; Sledge said that by some estimates, the tumors of smokers have one mutation for every three cigarettes they have smoked.

How to deal with these cancers is the forefront of oncology – and for now, there are more questions than answers.

One thing is obvious: Because they have no single driving mutation, "there will be no magic bullet for these tumors," Sledge said.

Instead, such tumors will need combination treatments – and clinical trials will need to catch up.

Maximums, Minimums, Optimums

MD Anderson Cancer Center Director John Mendelsohn said in an ASCO session that while the current clinical trial standard involves testing one investigational drug at a time, he expects the future will involve testing combinations of investigational drugs – yielding "better results, maybe with more toxicity."

The field is already moving in that direction. At ASCO, researchers were abuzz with the news that the two super-star melanoma agents of this year's meeting, recently approved Yervoy (ipilimumab, Bristol-Myers Squibb Co.) and hopefully soon-to-be-approved Vemurafenib (PLX4032/RG7204, Plexxikon Inc./Roche AG), are entering into combination trials.

Yervoy lead investigator Jedd Wolchok told reporters at a press conference that the two agents – Vemurafenib with its spectacular, but seldom durable results, and Yervoy with its lower response rate, but relatively high rate of long-term success, "complement each other nicely."

But Yervoy and Vemurafenib first went through trials by themselves. The obstacles to simultaneously beginning to explore two clinical unknowns are manifold.

The most obvious concern is toxicity.

That toxicity should be synergistic in combination trials is not inevitable. A presentation of combination data on the oral Mek inhibitor GSK212 and the oral BRAF inhibitor GSK436 showed that when the two agents were combined at their full monotherapy doses, they had fewer side effects than as single agents. But that is likely to be the exception rather than the rule.

One possible approach to avoiding toxicity is searching for a so-called "optimum biological dose," or even a minimum inhibitory dose rather than the standard maximum tolerated dose (MTD). Such an optimum biological dose would, in theory, be lower than the MTD and allow combinations with less toxicity risk. The concept is analogous to that used in HIV: HIV inhibitors are given as cocktails, and Phase I trials more often focus on inhibition of viral load rather than maximum tolerated dose.

But in practice, the obstacles are formidable – in part, again, because smart cancers have such a high mutational load, Julie Cherrington told BioWorld Insight.

Cherrington was trained as a virologist and is now the CEO of San Francisco-based Pathway Therapeutics Inc., which has a PI3 kinase inhibitor in early trials.

"Viral replication is pretty straightforward," Cherrington said. Tumors obviously have extremely complex genetics. And a cocktail of even three to four drugs – as has been spectacularly successful in converting HIV infection from a sure death sentence to a chronic disease – might not be enough for complex tumors.

Additionally, in practice, even targeted therapeutics "just tend to have a very narrow window," Cherrington said. "Often, the optimum biological dose and the maximum tolerated dose are the same" in practice, or very nearly so.

From Phase 0 to 60

Phase 0 trials are another way in which researchers are hoping to get a handle on testing drug combinations. In such trials, a single "microdose" of the drug candidate is given before the initiation of Phase I trials. The dose is as low as 1/50th of the dose that is planned for the Phase I trial, and the goal is to see whether the drug activates the same pathways in humans as it did in the preclinical animal models – and if it activates any unexpected ones.

Here too though, the devil is in the details, though in this case, the risk is the opposite. Not the possibility of a dose that is so high it produces unexpected toxicity, but in using one that is so low that it fails to activate the pathways that a clinically relevant dose would.

That risk was driven home in an ongoing study from the University of Washington, where scientists attempted to validate the Phase 0 concept by testing it with a clinically approved drug, the mTOR inhibitor Torisel (temsirolimus, Pfizer Inc.).

The study is ongoing, with final conclusions yet to be forthcoming. But at doses of 1/50th the clinical dose, the scientists failed to see the biological markers that temsirolimus activates at clinical dose levels, suggesting that microdoses may need to be increased.

In his plenary talk, Sledge had other suggestions for how to tackle the genomics era. Among them, to develop "selectively promiscuous" drugs rather than highly targeted agents; to target more general cellular processes with approaches such as immune-based treatments, stem cells, and DNA damage repair; and to minimize exclusion criteria.

But overall, it's clear that this will be a long road. Smart cancers are characterized by genomic chaos – and today's clinical trials system, he said, "is not designed for chaos."