Celldex designed the phase II study to test the theory that CDX-011 would work better in breast cancer patients whose tumors over-express the GPNMB protein. The 122 patients enrolled in the study had breast cancer in which at least 5% of the cells screened positive for GPNMB. All the patients had advanced breast cancer that was no longer responding to a median of five to six prior therapies. Two-thirds of patients were treated with injections of CDX-011, while the remaining third of patients were treated with chemotherapy of the doctor's choice.
Treatment with CDX-011 partially shrank tumors in 19% of all patients enrolled in the phase II study compared to 14% of patients treated with "investigator's choice" chemotherapy.
In patients with breast cancer tumors that contained high levels of the protein GPNMB, the CDX-011 response rate was 32% compared to 13% for patients treated with single-agent chemotherapy.
Back to the bear story for a moment: They raise questions about the legitimacy of GPNMB as an anti-cancer target for CDX-011. To them, the Celldex-GPNMB story is similar to Clovis Oncology's development of pancreatic cancer drug CO-101 in patients who were low expressers of hENT1 protein. Clovis' drug blew up in a phase III study because low hENT1 turned out to have no role at all in pancreatic cancer. Celldex bears believe GPNMB's role in breast cancer may also be a fallacy.-- Reported by Adam Feuerstein in Boston. Follow @AdamFeuerstein