Prognostic value of immunocytologic detection of bone marrow metastases in neuroblastoma.

BACKGROUND Morphologic evaluation of bone marrow for neuroblastoma cells is a routine and important component of clinical staging. Specific immunostaining of malignant cells with monoclonal antibodies should be more sensitive, however, and may improve the detection of metastases and provide additional prognostic information. METHODS We looked for tumor cells in bone marrow from 197 patients with newly diagnosed neuroblastoma, using immunoperoxidase staining with monoclonal antibodies (immunocytologic analysis) and examination of smears and specimens obtained by trephine biopsy (conventional analysis). RESULTS Routine smears and trephine-biopsy specimens were positive for tumor cells in 46 percent of the patients, whereas 67 percent were positive on immunocytologic analysis (P less than 0.0001). Immunocytologic analysis detected bone marrow metastases in 34 percent of patients considered to have only localized or regional disease (Stage I, II, or III). It also identified tumor cells that were not detected by conventional analysis in patients with widespread disease (Stage IV or IVS). Tumor content, as determined by immunocytologic analysis, predicted clinical outcome in relation to the age of the patient at diagnosis. Patients with Stage II or III disease diagnosed after one year of age who did not have occult marrow metastases did well, whereas those with metastases did poorly (P = 0.006). Patients in whom Stage IV disease was diagnosed before they were one year of age did well if bone marrow metastases were few or absent, but had poor survival if the marrow contained more than 0.02 percent tumor cells (P = 0.03). CONCLUSIONS Immunocytologic analysis of bone marrow aspirates is more sensitive than conventional analysis in detecting tumor cells and provides prognostic information. The relations among marrow metastases, age at diagnosis, and clinical outcome illustrate the biologic heterogeneity of neuroblastoma.