Evaluation of a nomogram used to predict the pathologic stage of clinically localized prostate carcinoma

A Nomogram based on pretreatment prostate specific antigen (PSA) level, tumor grade, and clinical stage has recently been developed and distributed to physicians. It was distributed to aid physicians in making treatment recommendations by predicting the probability of the final pathologic stage of clinically localized prostate carcinoma. The Nomogram was based on data for one patient population, and the validity of its application in general urologic practices had not yet been evaluated.

In the current study, the authors tested the performance of the Nomogram against data from their series of 697 men who underwent radical prostatectomy during the PSA era. Predictions made with the Nomogram were applied to the authors' data set, and the predictions were compared with actual outcomes of the authors' patients. A localized least-squares regression smoothing technique was used to determine whether the Nomogram was calibrated accurately for the authors' data and whether it discriminated across a full spectrum of patient characteristics.

Many of the predicted probabilities of the Nomogram were accurate, but some were suboptimal when applied to the authors' data set. Although the Nomogram did discriminate quite well between organ-confined and nonconfined cancer, it had difficulty predicting high probabilities of seminal vesicle invasion and lymph node metastasis, which are the pathologic features with the most profound impact on prognosis.

The Nomogram predicted organ-confined disease accurately. However, because not all of its predictions were completely calibrated when applied to the authors' data set, the authors conclude that the Nomogram may not be totally applicable to general urologic practice until further validation and possible modifications are performed.