Grading of prostatic cancer (I): An analysis of the prognostic significance of single characteristics.

This paper contains the first part of an attempt to quantitate the impact on prognosis of various parameters used in grading of prostatic cancer. Out of 346 patients of Elmer Belts series, 113 were identified whose tumors showed homogeneity with respect to single characteristics of a total of 12 parameters applied for grading in Mostofi's system. By this procedure it was possible to eliminate the possible influence on prognosis of the presence of several tumor formations within the same tumor. By using overall and intercurrent death corrected survival as end points, the impact of each of the 12 parameters on prognosis was studied. Only the architecture of the tumor (the parameter "glands"), variation in size and shape of the nucleus (anaplasia), and grade significantly influenced overall survival. In addition, corrected survival was significantly dependent on the amount of tumor seen and on the presence of mitoses. Subsequently, an attempt was made to replace "grade" by single parameters which had been shown previously to be of prognostic significance. It turned out that this was not possible. Grade is largely dependent on architecture and nuclear pleomorphism, but neither one of these parameters alone can reproduce "grade." Multivariate analysis was next used to further determine the prognostic weight of the individual parameter, and, if possible, to construct a new, more efficient grading system. These results will be reported separately [8]. It is unknown at the present time what the impact of several architectural formations within the same tumor on prognosis may be. The number of different formations found ranges from 1 to 4 in this material; 668 different formations belonging to 346 tumors were graded. The results of this analysis will be reported in part two of this series of papers [7].