Comparing tumour staging and grading systems: a case study and a review ofthe issues, using thymoma as a model

We consider the problem of comparing alternative cancer staging and grading systems. Statistical comparisons are on the basis of the ability to predic t survival, but more qualitative criteria, such as parsimony, and distincti ve prognostic separability of the categories are relevant also. Furthermore , some staging systems are clearly ordinal, while others are not. Three can didate statistical measures are studied and compared: explained variation; area under the ROC curve; and the probability of concordance of stage and s urvival. Each of these has individual strengths and weaknesses. A data set involving the staging of thymoma is analysed in detail to motivate the prob lem and illustrate the results. Copyright (C) 2000 John Wiley & Sons, Ltd.