159 cases of rhabdomyosarcoma (RMS), 42 from the NCI and 117 from St. Jude Children's Research Hospital treated similarly and staged according to the Intergroup Rhabdomyosarcoma Study (IRS) criteria were reviewed by 4 pathologists and subclassified according to 3 histologic schemes, i.e. the classical scheme (embryonal, alveolar, pleomorphic), Palmer's scheme (monomorphous round cell, anaplastic, other) and the NCI scheme (""""""""bad prognosis RMS, including the alveolar and solid variant alveolar RMS and good prognosis RMS, including all subcategories of embryonal RMS). Survival curves by histology, age, site grouping and stage were constructed according to the Kaplan and Meier method and compared for homogeneity and a pre-specified trend with the simple and stratified Mantel-Haenszel test. Finally, a multivariate Cox regression analysis was employed for each one of the 3 histologic schemes to evaluate the independent influence of histology on survival and the possible superiority of one scheme over the other two.
