The regression coefficients for four variables were

 significant at the .05 level. These included age, stage,

 initial treatment, and subsequent treatment. The partial

 regression coefficient of the variable histology was not

 significant at the .05 level. It was, however, retained as a

 variable for further study because of other findings that

 suggest its importance as a prognostic indicator.

 The second regression analysis used as independent

 variables age, sex, race, stage, histology, initial treat-

 ment, and subsequent treatment and also included the follow-

 ing interactions: Age x Stage, Age x Histology, Age x

 Initial Treatment, Age x Subsequent Treatment, Race x Stage,

 Stage x Initial Treatment, and Age x Stage x Initial

 Treatment.

 The regression was significant with F (79, 141) = 1.83,

n = .0009. The multiple correlation coefficient was .71 and

R = .51. None of the partial regression coefficients for

the interaction effects were significant at the .05 level.

However, because considerable differences do exist between

the type of treatments offered patients depending on the

stage of their disease, and consequently differences do exist

in their survival, the interaction between stage and initial

treatment was retained for evaluation in a subsequent linear

model.