Safety of bevacizumab therapy in subjects with brain metastases due to non-small cell lung cancer (NSCLC): PD3-3-3

Background:

Bevacizumab (Bv) in combination with carboplatin/paclitaxel received FDA approval in October 2006 for improving response rates and survival in first-line treatment of advanced NSCLC patients without brain metastases. Brain metastases were excluded based on a grade 3 intracerebral hemorrhage from a hepatocellular carcinoma brain metastasis in the original phase I study. Since NSCLC is the most common cause of brain metastases, a large fraction of patients who might benefit from Bv have never been evaluated formally. We report here on the safety of Bv therapy in 24 subjects with NSCLC following local therapy for brain metastases.

Methods:

Subjects were treated on protocols AVF3671g (ATLAS) or AVF3752g (PASSPORT):

The ATLAS phase III study includes platinum based chemotherapy with Bv, followed by Bv ± erlotinib (E) to disease progression, for subjects with advanced or metastatic non-squamous NSCLC. Brain metastases subjects had to have completed whole brain radiotherapy (WBRT).

The PASSPORT phase II study includes Bv combined with first-line or second-line systemic therapy in subjects with non-squamous NSCLC with treated brain metastases. Treatment for Brain metastases included radiosurgery, neurosurgery or WBRT.

Demographic characteristics, baseline disease characteristics, prior treatment, and on-study treatment are summarized by study. The overall rate of CNS hemorrhage and exact 95% confidence interval using the method of Blyth-Still-Casella are reported.

Results:

These results pertain to data collected on subjects enrolled from March 2006 to follow-up through January 2007.

The mean number of Bv doses (15 mg/kg/q3w) in the 24 treated subjects was 4 and the median was 3 (range 1-15). 2 subjects were never treated with Bv.

There were no CNS hemorrhages reported on either study (95 % upper confidence limit: 12.6%).

On ATLAS, no subjects experienced Grade 3/4 CNS events. One on-study death due to disease progression occurred while the subject was on the study treatment follow-up period.

On PASSPORT, 2 subjects experienced new onset Grade 3 CNS events of seizures. One of these 2 subjects was not treated with Bv; the other had residual resolving hemorrhage at baseline prior to study treatment.

Conclusions:

These preliminary data are encouraging with respect to safe administration of bevacizumab to subjects with treated brain metastases in these two ongoing studies. It is anticipated that data on 40 subjects will be available at presentation; additional follow-up and subjects are necessary to more precisely characterize the CNS hemorrhage rate and overall safety of Bv treatment in this setting.