Computer-assisted image analysis of tumor sections for a new thrombospondin receptor

Background: A cell surface receptor (50 kd) has been recently identified in malignant cells that recognizes the tumor cell adhesive domain (ie, cysteine-serine-valine-threonine-cysteine-glycine [CSVTCG]) of thrombospondin (TSP). This CSVTCG-specific TSP receptor can be considered as a new tumor marker, and its concentration on the cell surface may correlate directly with the capacity of tumor cells to invade and metastasize. Materials and methods: Six patients with primary, stages III and IV squamous cell carcinomas of the head and neck were studied. Tumor sections were specifically stained for this receptor with immunohistochemical techniques. The stained specimens were then subjected to computer-assisted image analysis. The area of positive staining and the heterogeneity of the pattern of staining were compared to peritumoral angiogenesis and clinical outcome of the patients. Results: The results indicate that those patients with a high and homogenous positive stain score (mean ± standard error [SE] 78 ± 5%) for the CSVTCG-specific TSP receptor had high microvessel density and died from metastatic disease within 12 months of initial treatment (correlation coefficients = 0.95 and 1, respectively). Patients with a low and heterogenous positive stain score for receptor (mean ± SE 8 ± 2%; P <0.001) had low microvessel counts and remained disease-free for at least 2 years. There was no relationship between receptor density and histologic classification of the primary tumors. Conclusion: The CSVTCG-specific TSP receptor, quantified through image analysis of immunohistochemical stained tissue sections, is highly predictive of clinical outcome in patients with squamous cell carcinomas of the head and neck.