Staff Writer

In a plenary talk about cancer genomics at last week's 100th Annual Meeting of the American Association for Cancer Research, Eric Lander, founding director of the Broad Institute, said he hopes to see complete cancer genome analyses become a routine part of oncology research within five years and a standard of care for patients within 10 years.

Genetic analysis of patients' tumors already is helping to drive oncology treatment decisions. Some drugs - like F. Hoffmann-La Roche Ltd.'s Herceptin (trastuzumab) for breast cancer and Novartis AG's Gleevec (imatinib) for chronic myeloid leukemia - are prescribed according to the results of biomarker-based companion diagnostic tests. Meanwhile, data on the KRAS biomarker are influencing prescriptions of Amgen Inc.'s Vectibix (panitumumab) and Eli Lilly and Co.'s Erbitux (cetuximab).

At AACR, the focus on biomarkers and personalized medicine was "astonishing," JMP Securities Inc. analyst Charles Duncan told BioWorld Financial Watch. The buzz at the conference was that "this is the wave of the future and the future's not that far off," he said.

Duncan pointed to several reasons why personalized medicine - long a biotech buzzword - took center stage at AACR. He credited scientific progress first and foremost. Two years ago there were scarcely any important biomarkers for the top three cancers; today, there are dozens. But increased investment in personalized medicine also has played a role, as has political and economic pressure to control rising healthcare costs, he said.

Molecular diagnostics have the potential to significantly impact healthcare costs. The price of a premium diagnostic - perhaps a few thousand dollars - pales in comparison to that of a cancer biologic, which can top $50,000 annually. That's why Duncan predicts diagnostics eventually will become the larger industry.

"In the future, we may not have companion diagnostics; we may have companion drugs," he said.

Diagnostics take less time and money to develop than drugs, which translates into cost-savings for the consumer. But they also provide less of a barrier to entry for the competition, which can make it challenging for diagnostics companies to achieve profitability.

Earlier this month, Fremont, Calif.-based Vermillion Inc. filed for bankruptcy despite encouraging data with its ovarian biomarker panel, and Huntingdon Valley, Pa.-based Immunicon Corp. filed for bankruptcy last year despite its partnership with Johnson & Johnson subsidiary Veridex LLC on the market-leading CellSearch cancer diagnostic.

Even Redwood City, Calif.-based Genomic Health Inc., poster child of the molecular diagnostics industry with its Oncotype DX assay for determining whether or not breast cancer patients will benefit from chemotherapy, has yet to reach profitability.

The RxDx Business Model Challenge

Duncan maintained that companies can make "a ton of money" in diagnostics. But the old business model of focusing on low-value in vitro tests isn't going to do the trick.

For example, it would be difficult to make money developing a companion diagnostic for KRAS, explained David Halbert, founder and chairman of Caris Diagnostics Inc. Despite the buzz around KRAS, "your market would be limited to colon cancer patients and you'd be one of many companies offering a $500 test," Halbert explained.

Duncan said companies will need to focus instead on high-value, patent-protected, clinically validated diagnostics. For example, Genomic Health can charge $3,800 for Oncotype DX because the product involves a 21-gene panel rather than just a single gene, he said.

But Caris' Halbert argued that as more and more biomarkers are discovered, another company can always come along with a bigger or better panel.

Duncan agreed that diagnostics companies have to provide value beyond their panels. In the future, he predicted, the industry will depend less on knowledge of biomarkers, which he likened to composition of matter patents, and more on how the markers are used, which he likened to methods patents.

For Genomic Health, the methods include knowing which biomarkers to include and which questions to ask. But the company has other barriers to competitive entry as well, like its clinical validation, relationships with doctors, acceptance in practice guidelines and reimbursement. All in all, the barriers are "much higher than the large number of wannabes hope," Duncan said.

Caris Diagnostics - which made quite a splash at AACR - takes a different approach to its method of using biomarkers. The Irving, Texas-based company analyzes tumor samples using standard assays - some available in the public domain and some purchased from other companies - and then cross-references the genetic outcomes against its database of biomarker information gathered from analyses of more than 100,000 scientific papers. The results of this meta-analysis provide doctors with a list of treatments - both approved and in clinical trials - that might work for the patient.

AACR Data Point to New Models

At AACR, Daniel Von Hoff of the Translational Genomics Research Institute presented data from a trial in which 66 relapsed and refractory cancer patients were analyzed and treated according to the biomarker analysis provided by Caris' Target Now product.

The 66 patients in the study hailed from nine trial sites and had failed an average of about four prior therapies, although some patients had failed as many as 13 treatments. Their tumors included breast, ovarian, colorectal, prostate, lung, melanoma and other types. Caris was able to identify biomarkers in all patients pointing to the potential utility of various available drugs.

Of the 66 patients treated according to the biomarker analysis, 18 patients (27 percent) had a significant improvement in progression-free survival (p=0.007). After four months, 14 patients (21 percent) had not progressed. Additionally, 47 patients (71 percent) experienced tumor shrinkage, and 6 patients (9 percent) experienced a partial or complete response.

Interestingly, for the 18 patients with a PFS benefit, there were no matches between the treatment recommended by the biomarker analysis and the treatment the physician would have chosen.

Many of the treatments were anything but intuitive. Caris' senior vice president of regulatory affairs and reimbursement, Alan Wright, explained that one patient had adenoid cystic carcinoma of the breast, a rare cancer with no established treatment guidelines. Target Now demonstrated that the vitamin D receptors in her tumor were amplified, so her physician gave her vitamin D, and her progression slowed.

Another patient presented with rare cystic adenocarcinoma of the parotid gland. She had failed multiple therapies including radiation, two metastectomies and treatment in clinical trials with kinase inhibitors. Her Target Now analysis indicated high expression of GART and linked the biomarker to positive outcomes with Alimta (premetrexed, Eli Lilly). Based on the report, her insurance agreed to cover Alimta, and her tumor burden decreased by half, allowing her to go into partial remission for about a year.

Wright said Caris has "maybe 100" similar stories. The privately held company serves about 2,500 patients each day and is "very profitable," according to Halbert, who founded AdvancePCS and sold it to CaremarkRx for $7.4 billion before starting Caris.

The company's next step will be to gather clinical evidence from prospectively designed trials. In the study presented at AACR, the patients served as their own controls, with a benefit in PFS counted as a ratio of 1.3 or greater for the PFS on the new treatment regimen compared to PFS on the patient's last prior treatment regimen.

Duncan said Caris is an example of the "innovation in terms of the business model" in molecular diagnostics.

He added that clinical validation of biomarkers and new business models for using those markers likely will be a prominent topic at the upcoming annual meeting of the American Society of Clinical Oncology (ASCO).