Biopsy of a biopsy: validation of immunoprofiling in gastric cancer biopsy tissue microarrays

Aims: Tissue microarrays offer an efficient way of examining a large number of tumour cases on a single glass slide. A major concern, however, is tumour heterogeneity. Also, the use of tissue microarrays in biopsy material is unexplored. The purpose of the present study was to assess the possibility and validity of arraying three 0.6-mm cores per case in endoscopic gastric cancer biopsies for immunophenotyping. Methods and results: Thirty-eight cases were studied with immunohistochemical staining for p53, CD44v6 and vascular endothelial growth factor. Full tissue sections were compared with triple core-tissue microarrays. Thirty-six cases contained three cores with tumour, one case contained two cores with tumour and one case contained only a single core with viable tumour and was excluded. Three further cores had been lost from three separate cases on the sections for immunohistochemistry. kappa values for whole-sections versus tissue microarrays ranged between 0.77 and 0.94. p53 immunohistochemical staining (interpretation as + or -) yielded the best result with only 1/37 mismatches, whereas CD44v6 (graded both for intensity and extent) showed 3/37 mismatches. The small depth of tissue in cores from biopsies necessitates all cores being arrayed flush with the face of the recipient wax block for maximizing the number of sections available. Compared with the first section over 30 additional 4-mum sections were available before the first case (with one core left) had to be excluded and 80 sections before half the tissue cores were lost. Conclusions: It is impracticable to array more than 120-150 cores per block. Tissue microarray with three cores per case is feasible and valid for studying biopsy material.