Detection of haematologic and nonhaematologic cancer by bone biopsy.

A retrospective study was carried out to test the efficacy of routine bone marrow biopsies for the diagnosis, classification, and prognosis of different forms of neoplastic involvement. Trephine and needle biopsies of the iliac crest of 3,626 patients with haematologic and 838 patients with nonhaematologic neoplasias were embedded without prior decalcification. 43 histologic variables were evaluated in 3-millimicrons sections of each biopsy, stained by five different techniques. The incidence of bone marrow involvement, in decreasing order of frequency, was as follows: plasmacytoma 55% and 95% of 428 cases, malignant lymphoma 37% and 79% of 1.112 cases, metastatic carcinoma 20% and 63% of 838 cases, and Hodgkin disease 3% and 28% of 772 cases each without and with manifest systemic dissemination. In the group of the metastatic carcinomas, there was a striking incidence of bone marrow involvement--82%--due to occult primary tumours. From a comparison of these figures with those reported in the literature, it is concluded that the large variations in positive and negative results are due to 1) differences in the size and the preparation of the specimens, 2) extent of the neoplastic dissemination at the time of the biopsy, and 3) the incidence of bone marrow involvement characteristic for a particular type of neoplasia. In addition, a subclassification of the chronic myeloproliferative disorders is proposed; it is based on histologic criteria whose prognostic relevance was tested and demonstrated by statistical analysis of the survival rates. The high incidence of detection reported in this study in patients without other evidence of systemic spread, or even in patients with occult neoplasias, provides a strong justification for the use of bone marrow biopsy as a primary diagnostic tool as well as a staging procedure, in both haematologic and nonhaematologic cancer.