An assessment of volume-weighted mean nuclear volume estimates as a prognostic index for neuroblastoma

This study evaluates an objective, unbiased, stereologic parameter (volume-weighted mean nuclear volume vV) as a prognostic indicator for survival of neuroblastoma in comparison with three histopathologic grading systems. In this retrospective study 24 consecutive, nonselected patients from the registry of Royal Liverpool Children's Hospital, Alder Hey, England were analyzed. Only primary lesions obtained before chemotherapy were used. Follow-up time in surviving patients (n = 10) was 4.5 to 20 years. All lesions were regraded blind and twice by two pathologists. vV was estimated on routinely processed, hematoxylin and eosin-stained, 5-microns sections, requiring on average less than 20 min per patient. Results show an absolute cut-off point for survival at vV = 270 microns 3. No patient with vV less than this value is alive at present (n = 7). In addition, actuarial survival curves for nonsurvivors show a bimodal pattern of survival time, separating patients with vV greater than (n = 7) and less than (n = 7) 270 microns 3. In comparison with the same analysis for the results of regrading by means of the Hughes or Beckwith system, results of vV estimation were superior. In comparison with the Shimada system the results confirmed the strong dichotomy for survivors and nonsurvivors, although with more overlap. vV has the advantage of predicting length of survival in nonsurvivors. The combination of Shimada grading and vV measurement, on the basis of the material studied, seems to offer useful prediction of biologic behavior. vV was always estimated on the small cell population, reducing the problems of biopsy representability.