False-negative rates of tumor metastases in the histologic examination of bone marrow.

Recent reports have claimed that bone marrow histology is less sensitive for detecting metastatic tumor than are other methods of marrow examination. These reports, however, did not specify methodologic information, such as the number of slides or tissue sections reviewed. To demonstrate the importance of specifying such information, we reviewed archival, tumor-positive marrow biopsy specimens from our laboratory. By means of a statistical model, we calculated the false-negative rates that would have occurred had random histologic review entailed fewer than 10 slides. Our model demonstrates a progressive increase in the percentage of false-negative diagnoses with the review of fewer slides, e.g., from 1% with review of nine slides to as high as 24% with review of only two slides. We conclude that the number of histologic slides and tissue sections reviewed should be specified when comparing bone marrow histologic examination with other techniques of marrow assessment.