For decades, approaches to chemotherapy have sought to kill as many cancer cells in a tumor as possible, prompting doses as high as patients can tolerate. Arguing for a radically different strategy, skeptics have challenged that mindset. Instead of trying to eliminate cancer, they say, clinicians should sometimes try to stabilize tumors and keep them under long-term control.
A pioneer in this line of emergent line of thinking is Robert Gatenby, M.D., a radiologist who directs the Cancer Biology and Evolution Program at the Moffitt Cancer Center in Tampa, Fla. Last February, Gatenby and colleagues published results in Science Translational Medicine showing that mouse models of aggressive human breast cancer live longer with fewer side effects if they get successively lower doses of chemotherapy instead of standard high-dose treatment. The study served as a proof of concept for what Gatenby calls adaptive therapy, which strives to keep drug-sensitive and drug-resistant cancer cells in prolonged equilibrium. The approach has no curative intent, “so it’s really only for patients with metastatic cancer for which no curative therapies are available,” Gatenby said.
Modeling Dose Decisions
Gatenby grounds his adaptive strategy in mathematical modeling and evolutionary theory. A central tenant of Darwinian evolution is that the fittest life forms survive, and according to Gatenby, high-dose therapies boost the fitness of drug-resistant cancer cells by killing off their drug-sensitive competitors. James DeGregori, Ph.D., a professor at the University of Colorado School of Medicine in Denver, who was not involved in the research, agreed. He proposed that cancer cell fitness has less to do with a time-dependent accumulation of genetic mutations than with selective pressures imposed by the tumor’s microenvironment. During Gatenby’s new study, mouse tumors treated with successively lower doses were less necrotic and better vascularized than tumors given high dose therapy, “and that’s a state that should favor the growth of more normal cells,” DeGregori said.
After tumors in the mice had grown large enough, researchers divided the animals into three treatment groups and a control. One treated group got standard therapy (ST), paclitaxel at 20mg per kg of body weight given intraperitoneally twice a week for 2.5 weeks. The two other groups were treated with adaptive therapy using dose schedules calculated by a mathematical algorithm. One group (AT-1) received the same initial ST dose followed by progressively lower doses given at equal intervals, in amounts determined on the basis of tumor size reduction. The second group (AT-2) received consistent doses of 15mg/kg each, but doses could be skipped if the tumors didn’t grow or if they shrank after treatment.
At first, all three regimens slowed tumor growth similarly. But only the AT-1 regimen produced long-term stabilization: Tumors in most animals treated with that regimen never grew beyond 1,000mm3. And in four animals, tumors remained stable even after therapy was withdrawn. By contrast, tumors in the ST group grew quickly after the scheduled treatments were finished. Though the AT-2 group had longer tumor stabilization, they generally got worse without any benefit in progression-free survival.
Gatenby said he and his team investigated the AT-2 regimen because they anticipated that skipping doses might be more practical in the clinic than lowering doses over time. “But we found that if we skipped even a single dose, we could lose control of the tumor altogether,” he said. Asked why, Gatenby speculated that even when given intermittently, the 15-mg/kg dose was killing off too many drug-sensitive cells and leaving populations that weren’t large enough to contain their drug-resistant counterparts. “It was only by decreasing doses at consistent intervals that we could keep the sensitive cells at more stable populations,” he said.
Adaptive Therapy in the Clinic
Still, dose skipping appears to be working in a pilot study of adaptive therapy at the Moffitt Center with patients who have metastatic stage IV prostate cancer. Led by Gatenby’s colleague, Jingsong Zhang, M.D., a genitourinary oncologist, the study allows patients to skip doses of abiraterone, an androgen synthesis inhibitor, if their prostate-specific antigen (PSA) levels drop by 50% or more below pretreatment baselines. With nine patients enrolled so far, and a target accrual of 40, the “study is too small to draw statistically robust conclusions,” Zhang said. “The next step would be a randomized phase II clinical trial.” But patients have so far been able to go off treatment for an average of 3 months per treatment cycle. “We watch their labs closely and give surveillance scans every 12 weeks, and if PSA levels return to the pretreatment baseline then we go back to abiraterone therapy,” Zhang said. Reiterating that the objective is a sustained balance between treatment-sensitive and treatment-resistant clones, Zhang added, “The whole idea is to prolong responses and minimize side effects” while reducing the financial burden for patients since abiraterone is expensive.
Larry Norton, M.D., a medical oncologist at the Memorial Sloan–Kettering Cancer Center in New York and Deputy Physician-in-Chief for Breast Cancer Programs at Memorial Sloan Kettering and Medical Director of the Evelyn H. Lauder Breast Center, said the adaptive approach is consistent with his view that mathematical models should play a greater role in designing dose schedules. “You need models that consider [not only] the number of cells killed but also the recovery of cells between treatments,” he said. “We’ve done the math and found that the best bet is a balance between appropriate doses given at optimal intervals.” Referring to the Zhang’s pilot study, he added, “I’d like to see trials like this in breast cancer too—there’s lots of room for improvement based on this kind of thinking.”
Louis Weiner, M.D., director of the Georgetown Lombardi Comprehensive Cancer Center in Washington, D.C., pointed out that Gatenby is using the same evolutionary concepts used to control weed pests in agriculture. Instead of giving high-dose herbicides that select for resistant strains, farmers increasingly opt for “integrated pest management” schemes that protect the environment by using only enough of the chemicals to suppress weeds at levels below those that would cause unacceptable crop damage. “Gatenby’s doing the same thing,” he said. “He’s trying to head off resistant strains at the pass so they don’t cause trouble down the road.”
Louis Weiner, M.D.
But Weiner pointed out that adaptive therapy would probably also require more frequent office visits and imaging. The Zhang study has the advantage of relying on PSA, an unusually accessible and sensitive tumor growth marker. “For other cancers, we don’t know what the frequency of the evaluations might be,” Weiner said. “For now, we can say that the approach seems to have merit and warrants continued development in preclinical studies and clinical trials. Time will tell if this is a modality that extends to other diseases, and other types of therapy, including targeted and immunotherapies. But I think it can be done.”
