Standard of Care Redefined

Cancer chemotherapy has a long history of mixed success. It's prescribed according to established protocols based on the type of cancer, pathology findings, and, to a lesser degree, tumor staging and whether treatment is adjuvant or first-line.

Choice of chemotherapy agent or combination is based on results from thousands of patients treated in large clinical trials. With few exceptions, no attempt is made to differentiate patients. Thus, advanced resectable lung cancer patients often receive an adjuvant combination regimen consisting of a platinum-based drug and at least one other agent.

Similarly, patients with inoperable tumors or those who experience recurrence are treated sequentially with first-line agents or combinations, followed by second-, third-, etc. Patients are referred to experimental treatment after failing some number of established therapies.

Only a fraction of patients respond to first-line chemotherapy. Many do not respond and must endure side effects from the drug's toxicity. Patients also lose valuable time, during which their tumor may continue growing or spreading. By the time an effective treatment is found, the patient may be too sick to benefit.

The efficacy of adjuvant chemotherapy varies significantly as well. For example, most women with very early stage breast cancer will be cured, but all of them receive adjuvant chemotherapy even though just 10% will experience recurrence. Because adjuvant therapy improves survival rate and most women tolerate the treatment, this practice has become the standard.

This situation is clearly not optimal. If clinicians had a better grasp as to which patients were at very low risk for recurrence, the percentage of women receiving adjuvant therapy could fall dramatically, depending on the sensitivity and specificity of the biomarker test. Thus, a large percentage of women might be spared chemotherapy and the health care system would save billions of dollars per year.

Given the tools already available for characterizing cancer at the molecular level, and the growing number of therapeutic options, the medical community should be able to do better than the current standard of oncology care. For example, rather than treating every patient according to protocol with first, second, and third-line therapies, some patients may benefit from receiving chemotherapy agents in a different order, or nonstandard combinations.

Biomarkers and cell culture

Biomarkers could play a significant role in selecting the proper cancer treatment, but only a handful of validated drug–biomarker combinations exist today.

Genomic Health, a genetic testing company, offers OncotypeDX, a multi-gene assay that provides a quantitative assessment of the likelihood of a distant breast cancer recurrence. By identifying these women, the test claims to predict who will benefit from adjuvant chemotherapy and who might safely forego it.

Similarly, correlative biomarkers such as Her2-neu, ERCC1, and K-ras serve as the rationale for treating with Herceptin, platinum compounds, and Erbitux, respectively, but provide no insight into what other therapies might work for biomarker-negative patients.

An assay that defines sensitivity to platinum, for example, says nothing about what other drugs might work, or what might be the next step in treatment.

For example, Response Genetics, Labcorp, and others offer a test for over-expression of the ERCC1 gene, which predicts sensitivity to platinum-based chemotherapy agents. A positive test increases the odds of a tumor responding to platinum to about 75%. This test provides some guidance, but only for one agent. Only about half of lung cancer patients respond to platinum.

All current cancer biomarker tests are correlative to one specific therapy. They confirm or eliminate one treatment option but offer nothing about other available treatments.

In vitro tests

Cultured tumor cell sensitivity tests involve culturing fresh tumor cells and testing them against one or more chemo agents, either in vitro or after implantation into test animals.

These tests predict resistance to chemotherapy agents reasonably well, but not sensitivity. In other words, if the cultured tumor does not respond, the patient's tumor will likely not respond. However, when the tumor does respond, the test's predictability is only about 50%. Cell culture-based testing takes about four to six weeks to complete.

The use of cultured cells to predict a drug's efficacy, while scientifically dubious, is widely practiced today for its ease of use and convenience. Cultured cells rapidly lose the primary tumor's physical, biochemical, and even genetic characteristics, and they continue to diverge with each generation. Biologists have known for decades that cells self-select for the ability to grow under culture conditions.

Xenotransplantation

Biomerk Tumorgraft (tumor xenograft) technology, under development by Champions Biotechnology, Inc., represents a dramatic break from the paradigms of correlative biomarker testing and cell- or tissue-based chemotherapy assays.

Tumorgrafts involve harvesting tumor tissue from patients through biopsy or surgery, and implanting tumor fragments in immune-incompetent mice. The tumors are "grown" in these animals and then retransplanted as fragments to expand the test animal population to dozens or hundreds of mice. This is not a new idea.

Early embodiments of tumor xenografts were based on implantation of cells extracted from patients and grown in vitro, sometimes for many generations. But xenografts obtained in this manner are poor predictors of a clinical efficacy. Xenografts are distinct from purely in vitro tumor cell sensitivity tests, which are conducted on cells that have undergone hundreds or thousands of divisions. In both cases, the cells have self-selected for survival under cell culture conditions, and are almost always significantly different from the original tumor.

A Tumorgraft established in mice provides a living, unlimited test bed for investigating existing anti-neoplastic agents and combinations, including experimental agents.

Tumorgrafts growing in test animals retain the salient characteristics of the original human cancer. The animals may be used to screen anticancer drugs and combinations to personalize treatment for the patient donating the tumor tissue, or can serve as screening platforms for new drugs and combinations and for discovering biomarkers.

Unlike cultured tumor cells in vitro or after transplantation into test animals, Tumorgrafts have been shown to predict both resistance and susceptibility to chemotherapy agents. The technique allows investigators to pinpoint specific drugs or combinations, even among new agents, which would take years to uncover through human studies, generating data that might take years to uncover.

Personalizing therapy

Champions Biotechnology has applied its Biomerk Tumorgraft technology to provide physicians a Personalized Tumorgraft option for cancer patients.

At this stage of development, Personalized Tumorgrafts are not appropriate for every cancer patient. Where biomarker results come back in a week, and cell culture assays approximately six weeks, Personalized Tumorgrafts typically require four to six months to expand the population of mice bearing the patient's tumor to the point where sufficient animals are available for testing. A patient that is a viable candidate for Personalized Tumorgraft testing must therefore be expected to live at least six months.

Unresectable patients who have a high likelihood of benefiting from first-line therapy may undergo their first chemotherapy regimen while their tumor line is expanding in the test animals. Patients expected to receive adjuvant therapy will usually wait until tests are completed before committing to treatment. These Personalized Tumorgraft patients can take comfort in the fact that if their disease returns they will receive a treatment based on analysis of multiple drug options tested against their own unique tumor.

Personalized Tumorgraft development and testing is also an expensive process that is not currently covered by insurance.

Patient benefits

Testing of drugs and combinations. Approximately 40 drugs have been approved in the United States for treating cancer. That number is sure to rise substantially if only a small fraction of the drugs in clinical trials are licensed. Combinations increase the number of possibilities by at least a factor of 10. Personalized Tumorgrafts can allow testing of all drugs and combinations, including permutations that would not normally be considered or that are difficult to screen through correlative studies.

Higher level of genotype predictability. Tumors classified phenotypically (lung, breast) may differ genetically and epigenetically. Targeting these specific genotypes with individualized therapies may be possible at some future time. Until then we are limited to treating the phenotype with the added possibility of "personalizing" therapy based on various markers (kRas, Her2neu, etc.).

Given these constraints, the model with the highest level of predictability would be based on the patient's own tumor grown in a living animal. We know, from molecular studies, that these Personalized Tumorgrafts retain all salient characteristics of the patient's tumor.

Targeted treatments. Personalized Tumorgrafts offer the potential to discover treatments that work, while sparing patients from drug toxicities that will be ineffective in their case. In addition to the resulting treatment efficacy, the patient may also be afforded a significant reduction in the costs of care as well as improvement in quality of life.

Uncovering chemotherapy efficacy. In addition to cancer type, stage, origin, patient age/sex, and treatment history, Tumorgraft characterization could include gene expression, mutations, comparative genomic hybridizations, methylation, and analysis of the proteome and microRNA. This creates the possibility for long-term studies of the efficacy of chemotherapy in specific cancer genotypes.

Correlating a tumor's molecular characteristics with phenotypic responses will provide a resource on how and why chemotherapy works. A database of such information might eventually eliminate the need to grow tumors in mice.

Use for personalized medicine. Tumorgrafts, when coupled with advanced biomarker and genetic testing, will enable investigators to hone-in on cancers that are distinct within their phenotype and accurately match them with the therapy most likely to work. Even a 10% improvement in predicting drug failures before clinical trials could save millions of dollars in wasted development costs.

David Sidransky, MD, is chairman of the board of directors of Champions Biotechnology, Inc., Science & Technology Park at Johns Hopkins, 855 N. Wolfe Street, Baltimore, MD 21205; director of the Head and Neck Cancer Research Division at Johns Hopkins University School of Medicine; professor of Oncology, Otolaryngology, Cellular & Molecular Medicine, Urology, Genetics, and Pathology at John Hopkins University and Hospital, email: sidranskyd@championsbiotechnology.com