Pathways Targeted in BATTLE-2
Targeted agents under study are shown in blue boxes.
Adapted from Herbst RS. Targeting the PI3K/AKT axis: what lies ahead in lung cancer.
Presented at: 13th International Lung Cancer Congress; Huntington Beach, CA; July 19-22, 2012.
As therapy based on cell-signaling pathways has become a priority in cancer research, so has the concept of designing clinical trials that can better target patient populations more likely to benefit from a particular regimen.
In that arena, the phase II BATTLE clinical trial stands out as a groundbreaking effort to advance therapies for patients with non-small cell lung cancer (NSCLC) through an innovative design that allowed researchers to simultaneously analyze multiple agents based on “real-time” biomarker analyses of the individual participants. The BATTLE program is continuing with two ongoing trials, as well as plans for an additional study.
The need for new methods for tackling NSCLC is particularly pressing. Despite treatment advances, the five-year survival rate for all stages of NSCLC is only 17%, and drops as low as 1% for those diagnosed as stage IV, according to the American Cancer Society. Moreover, only 15% of lung cancers are detected when the disease is still localized and outcomes are better.
Targeted agents used in combination with chemotherapy often have faltered in phase III trials in unselected patients with NSCLC, and systemic chemotherapy has remained the mainstay for metastatic disease, with oncologists making treatment decisions based on such clinical factors as age, gender, or performance status.
Against this backdrop, researchers at The University of Texas MD Anderson Cancer Center in Houston led the planning for the first BATTLE trial in 2004. At the time, the molecular landscape for NSCLC was far less characterized than it is today, but the trial’s full name is indicative of its ambitious goals: Biomarker-integrated Approaches of Targeted Therapy for Lung Cancer Elimination.
To test their hypotheses, researchers employed an adaptive trial design in which they used information collected early in the trial to assign patients to therapy based on the likelihood of success as the studies continued. The trial depended heavily on developing a biomarker profile for each patient.
“We felt the traditional way of testing drugs in combination with chemotherapy was not going to really help us move these drugs forward,” said principal investigator Edward S. Kim, MD, in an interview posted on MD Anderson’s website. Kim, now chair of the Department of Solid Tumor Oncology and Investigational Therapeutics at Carolinas HealthCare System’s Levine Cancer Institute, added, “We felt it was more important to really focus on the patientâŽ¯finding what was unique about that patient’s tumor.”
Choosing Therapies for First Trial
Four promising molecular targets were identified: endothelial growth factor receptor (EGFR), Ras/Raf, retinoid X receptor (RXR)/cyclin D1, and vascular endothelial growth factor (VEGF) or the angiogenesis cascade.
Investigators then selected four therapies to deploy against those signaling pathways: erlotinib (EGFR), vandetanib (VEGFR), erlotinib plus bexarotene (EGFR/ RXR), and sorafenib (KRAS/BRAF). Since no predominant biomarker had been identified yet in NSCLC, investigators used available research to match each patient’s molecular profile with the therapy most likely to benefit him or her.
In order to assess molecular tumor characteristics, the researchers recruited heavily pretreated patients with NSCLC and, during the course of the trial, performed two fresh core needle biopsiesâŽ¯with the challenging goal of completing biomarker analyses within 14 days.
In all, 255 of the 341 enrolled patients were randomized for treatment, with the first group of 97 patients randomized equally to the four treatment arms and the following 158 patients assigned through adaptive randomization to trial arms based on their biomarker status. The plan also allowed for patients who progressed to re-enter the trial and be randomized to another treatment, if appropriate.
The primary endpoint was the disease con trol rate (DCR) eight weeks after randomization, assessed through imaging studies, and many of the patients treated achieved that mark. Across all marker groups, 46% of those eligible for analysis (112 of 244 patients) exhibited disease control. In an analysis of 20 biomarker/treatment pairings (including a no-marker group), patients in eight groups exceeded the 0.8 historical probability of a DCR >30%.
For secondary endpoints, the overall median progression-free survival (PFS) was 1.9 months (95% CI, 1.8-2.4) and the median overall survival (OS) was 8.8 months (95% CI, 6.3-10.6). Median OS improved to 9.6 months for those who achieved a DCR at eight weeks, versus 7.5 months for those who did not.