The Avastin Saga Continues

We previously wrote about bevacizumab (Avastin) and its approval for breast cancer. The early clinical trials revealed evidence of improved time to disease progression. This surrogate measure for survival benefit had, over recent years, gained popularity, as time to disease progression is a measure of the impact of a given treatment upon the patient’s response durability. It was hoped and believed that time to progression would be an early measure of survival.

Unfortunately, the survival advantage for the Avastin-based therapies in breast cancer has not met statistical significance. As such, careful review by the oncology drug committee of the FDA lead to a unanimous decision to remove Avastin’s indication in breast cancer. Avastin has not been removed from the market, but instead, cannot be promoted or advertised, nor do insurers necessarily reimburse it. This decision, however, will have a very big impact on Medicare patients and many others who are in managed care programs (HMOs).

There are no villains here. Instead, dedicated physicians empowered to scrutinize the best data could not prove beyond any doubt that the drug improved survival. The time to progression data was favorable and the survival data also trended in a favorable direction. But, the final arbiter of clinical approval — statistically significant survival — was not met.

The physicians who want to provide this for the patients, the company that produces the drug and the patients who believe it offers benefit all have legitimate positions. As Jerome Groopman, MD, once said, in a similar situation with regard to the FDA approval of interleukin 2 (a biological agent with profound activity in a small minority of melanoma and renal cell cancer patients), “I am confronted with a dilemma of biblical proportions, how to help the few at the expense of the many.”

The Avastin saga is but one example of what will occur repeatedly. The one-size-fits-all paradigm is crumbling as individual patients with unique biological features confront the results of the blunt instrument of randomized clinical trials. Our laboratory has been deeply involved in these stories for 20 years. When we first observed synergy for purine analogs (2CDA and fludarabine) with cytoxan, and then recommended and used this doublet in advanced hematologic malignancies (highly successfully, we might add) we were a lone voice in the woods. Eventually, clinical trials conducted at M.D. Anderson and other centers confirmed the activity establishing these treatments as the standards of care for CLL and low-grade lymphoma.

The exact same experience occurred in our solid tumor work when we combined cisplatin plus gemcitabine in pancreatic, ovarian, breast, bladder, lung and other cancers. While our first patient (presumably the first patient in the world) received cisplatin plus gemcitabine for drug-resistant recurrent ovarian cancer in 1995 — providing her an additional five years of life — it wasn’t until 2006 that the FDA approved the closely related carboplatin plus gemcitabine for this indication.

We now confront an even greater hurdle. With our discoveries, using novel combinations of targeted agents, we are years (perhaps decades) ahead of the clinical trial process. We know that patients evaluated in our laboratory with favorable profiles can respond to some of the newest drugs, many of which have already completed Phase I of clinical trials. It is our fervent belief that we could accelerate the drug development process if we could join with the pharmaceutical companies and the FDA to put these hypotheses to a formal test.

Again, there are no villains here. Patients want, and should, receive active drugs. Doctors should be allowed to give them. The drug companies want to sell their agents and the FDA wants to see good therapies go forward.

The rancor that surrounds these emotionally charged issues will best be resolved when we introduce techniques that match patients to active therapies. We believe that the primary culture platform used in our laboratory, and a small number of dedicated investigators like us, may be the answer to this dilemma.

We will redouble our efforts to apply these methods for our patients and encourage our patients to lobby their health care insurers and representatives to sponsor these approaches. To date, we have been unsuccessful in convincing any cooperative group to test the predictive ability of these selection methodologies. In response, I reiterate that I will gladly participate and, to the best of my ability, support at least the laboratory component of any fair test of our primary culture methodologies.

We stand at the ready for the challenge.

Chronic Lymphocytic Leukemia (CLL) as a Platform for Functional Profiling

Among the most common forms of leukemia in adults is chronic lymphocytic leukemia. This neoplasm usually arises in a subset of lymphocytes known as B-cells. However, T-cell variants also occur. The disease presents clinically as an elevation of the circulating lymphocytes. This may be associated with enlarged lymph nodes, splenomegaly or liver enlargement.

The decision to treat patients is largely based upon clinical staging systems know as the Rai or Binet classifications. Low risk patients can often be observed without treatment, while more aggressive presentations (such as those associated with anemia and low platelet counts) require intervention. More recently, molecular determinants of aggressiveness have been applied in the prognosis of this disease. These include: CD38, VH gene mutation and Zap 70. Additional findings include ATM mutations, principally in the T-cell and pro-lymphocytic variants.

For more than 40 years, the treatment of choice for this disease was oral chlorambucil. Although effective, chlorambucil resulted in the development of resistance and was associated with rather significant myelosuppression over time. The introduction of fludarabine (FAMP) and 2-CDA revolutionized the management of this disease —providing high response rates with relatively tolerable toxicities.

The introduction of 2-CDA and fludarabine in the 1980s offered an opportunity for our laboratory to examine drug interactions in CLL patients. Combining the alkylating agents (of which, chlorambucil is a member) with 2-CDA revealed synergy (supra-additvity) in 100 percent of the CLL samples we studied (Nagourney, R; et al. British Journal of Cancer, 1993). Based on this observation, we began treating patients with CLL and related lymphoid malignancies with a combination of Cytoxan and 2-CDA, resulting in dramatic and durable remissions.

O’Brian, Keating and other investigators at the MD Anderson then undertook this work (using fludarabine), providing for the most effective therapy for CLL in today’s literature. Unfortunately, a percentage of patients who receive this combination develop deep myelo-suppression. Therefore, the administration of this combination requires careful monitoring by the physician.

One of the most interesting aspects of the high activity observed for fludarabine was the capacity of this “anti-metabolite” to induce cell death in short term cultures of CLL cells. It was well known that CLL cells were not highly proliferative, yet the anti-metabolite class of drugs was specifically designed to stop cell proliferation at the level of DNA synthesis. We realized that 2-CDA and Fludarabine had to be killing cells, not preventing their growth. This conundrum provided an opportunity for us to test a related anti-metabolite in this disease. We chose cytarabine (Ara-C), a drug not considered effective for CLL (e.g. low proliferative rate, no likelihood of DNA synthesis inhibition, no likelihood of cytotoxicity). To our delight, low doses or Ara-C proved highly effective in controlling even the most advanced cases of CLL as we then reported.

CLL became one of our favored models for the study of human tumor biology, enabling us to study drug responses at the molecular level. Many of the observations that we made in this hematologic malignancy granted us insights that we continue to apply in solid tumors today.