Preclinical Strategies to Optimize Optical Imaging in Head and Neck Cancer

Abstract

Objective: There is a need for cancer imaging in the clinic and operating room. We have shown that fluorescently labeled antibodies can detect microscopic islands of head and neck cancer cells using modified intraoperative microscopes in preclinical animal studies. Here we evaluated the optimal targeting molecule for clinical translation. Method: We assessed in vitro binding of fluorescently labeled antibodies targeting EGFR (cetuximab/Erbitux and panitumumab/Vectibix) and an antibody mimetic (anti-EGFR affibody) in multiple cell lines (SCC5, OSC19, FADU). To compare in vivo imaging characteristics, HNSCC xenografts (SCC5, SCC1, FADU) were imaged after systemic injection of fluorescently labeled antibodies over 5 days. Results: The binding affinity was lower for the control antibody IgG (KD = 9.55 nM) compared with either panitumumab (KD = 6.82 nM) or cetuximab (KD = 2.86 nM). In vitro, panitumumab had a two fold increase in relative fluorescence compared with cetuximab and a four fold increase in fluorescence compared with control antibody (nonspecific IgG). Xenograft imaging in vivo demonstrated a variation in fluorescence intensity that correlated with circulating half-lives: peaks in fluorescence intensity were seen at 1.5 h (affibody) and 48 to 72 h (cetuximab, panitunumab). The pantitumumab and affibody images demonstrated greater tumor fluorescence and lower background signal compared to cetuximab. Conclusion: Panitumumab/Vectibix demonstrated optimal imaging characteristics which appeared to be related to its binding affinity and increased circulation time. These characteristics may allow for increased localization to the tumor site. Panitumumab may be the optimal antibody for clinical translation since it will decrease false positive rates associated with non-specific uptake.