Comparative evaluation of bone marrow aspirate particle smears, biopsy imprints, and biopsy sections.

Bone marrow aspirate particle smears, biopsy imprints, and biopsy sections were compared to determine the accuracy of the three samples in assessing for overall cellularity, differential cell count, megakaryocyte density, iron stores, and tumor infiltration. Aspirate particle smears and biopsy imprints were stained by Wright-Giemsa method. Aspirate particle smears were also stained with Prussian-blue. Biopsy sections were 1 1/2-2 micron thick and were prepared from non-decalcified plastic embedded samples and stained with combined Prussian-blue-hematoxylin-eosin, and Giemsa. One hundred-eight sets of specimens from 99 patients were examined. In 20 cases, chi-square analysis showed a comparable degree of cellularity (p less than 0.001) and megakaryocyte density (p less than 0.001) among the three preparations. Differential count comparison by regression analysis indicated that mean percentages of neutrophilic cells in the proliferation compartment were comparable in the three groups (p less than 0.01). A better correlation was obtained among the three groups in the percent neutrophilic cells in the maturation-storage compartment, normoblasts, eosinophils, and plasma cells (p less than 0.001). Lymphocytes in the aspirate smears correlated with the biopsy imprints (p less than 0.01) but not with the biopsy sections (p greater than 0.05). Monocytes did not correlate in any of the groups (p greater than 0.05). In 47 cases, chi-square analysis of iron stores in the aspirate particle smears correlated well with those in the biopsy sections (p less than 0.001). Fifty-two marrows that were done for staging nonhematological malignancies revealed malignant cells in 21 cases, biopsy sections were positive in all, biopsy imprints were positive in 19 (90%), and aspirate particle smears were positive in 7 (33%). Thirty-six marrows done for staging non-Hodgkin's lymphoma showed malignant cells in 13 cases. Twelve (92%) biopsy sections, three (23%) biopsy imprints, and nine (69%) aspirate particle smears contained lymphoma cells. In conclusion, a satisfactory evaluation of marrow samples for diagnostic studies can be achieved by examination of biopsy sections along with aspirate particle smears or biopsy imprints. Any of the three marrow preparations alone is not sufficient for accurate diagnosis in all cases. The biopsy imprint is an accurate modality for identifying nonhematological tumor metastasis in the bone marrow.