Artificial neural network model to predict biochemical failure after radical prostatectomy

BACKGROUND

Biochemical failure, defined here as a rise in the serum prostate specific antigen (PSA) concentration to >0.3 ng/mL or the initiation of adjuvant therapy, is thought to be an adverse prognostic factor for men who undergo radical prostatectomy (RP) as definitive treatment for clinically localized cancer of the prostate (CAP). We have developed an artificial neural network (ANN) to predict biochemical failure that may benefit clinicians and patients choosing among the definitive treatment options for CAP.

MATERIALS AND METHODS

Clinical and pathologic data from 196 patients who had undergone RP at one institution between 1988 and 1999 were utilized. Twenty-one records were deleted because of missing outcome, Gleason sum, PSA, or clinical stage data. The variables from the 175 remaining records were analyzed for input variable selection using principal component analysis, decision tree analysis, and stepped logistic regression. The selected variables were age, PSA, primary and secondary Gleason grade, and Gleason sum. The records were randomized and split into three bootstrap training and validation sets of 140 records (80%) and 35 records (20%), respectively.

RESULTS

Forty-four percent of the patients suffered biochemical failure. The average duration of follow up was 2.5 years (range 0-11.5 years). Forty-two percent of the patients had pathologic evidence of non-organ-confined disease. The average area under the receiver operator characteristic (ROC) curve for the validation sets was 0.75 +/- 0.07. The ANN with the highest area under the ROC curve (0.80) was used for prediction and had a sensitivity of 0.74, a specificity of 0.78, a positive predictive value of 0.71, and a negative predictive value of 0.81.

CONCLUSION

These results suggest that ANN models can predict PSA failure using readily available preoperative variables. Such predictive models may offer assistance to patients and physicians deciding on definitive therapy for CaP.

Cystectomy for bladder cancer: a contemporary series

PURPOSE

To validate the current TNM staging system, we analyzed our contemporary experience with 300 cystectomies.

MATERIALS AND METHODS

The pathological material and medical records of 300 patients treated with cystectomy were reviewed, and the new TNM classification was adopted.

RESULTS

The median followup of patients with no evidence of disease was 65 months, and overall survival rate was 45% with a median survival of 50 months. In a Cox regression analysis only patient age, pT stage and neoadjuvant chemotherapy were significant factors for survival. The disease specific survival was 67% with a median survival of 94 months. In a multiple proportional hazards analysis only pT stage and previous chemotherapy were significant factors of disease specific survival. A significant difference was seen in the overall and disease specific survival between patients with organ confined and nonorgan confined tumors. We did not observe a difference in the survival rate among patients with pT4a to pT3 tumors. Significant differences were not seen in survival rates between sexes or among patients of different age groups. Transitional cell carcinoma was the predominant histological type, and no significant difference was found in patient outcome among the different histological subtypes.

CONCLUSIONS

Bladder cancer can be categorized into organ confined and nonorgan confined tumors. This dichotomous grouping is better suited for evaluating adjuvant clinical trials. The pT stage of the bladder and prostate should be prospectively analyzed together to better define the clinical implications of prostatic involvement. In our opinion the histological subtypes do not affect outcome.

Papillary carcinoma of the endometrium: morphometric predictors of survival

This study compares morphometric features of endometrial papillary carcinomas, serous and endometrioid types, in an effort identify characteristics that predict clinical outcome. Fifty-one consecutive patients with papillary carcinoma were identified at NYU Medical Center from January 1979 through December 1991 and were followed through 1994. Morphologic analysis was conducted by investigators (R.D. and K.M.) blinded as to original diagnosis, depth of invasion, stage and outcome. Of 25 variables analyzed, six were significantly associated with reduced survival, namely marked nuclear pleomorphism, multinucleated cells, hobnail cells, psammoma bodies, uneven papillary borders, and inflammation. Serous carcinomas showed a highly significant (p = 0.0001) association with reduced length of recurrence-free survival, with a median survival of 27 months (mean +/- SE, 38.9 +/- 7.0) versus 95 months (106.7 +/- 15.3) for patients with endometrioid types. Of six morphologic characteristics that correlated with reduced recurrence-free survival by univariate analyses, only one, nuclear pleomorphism, showed a significant association with an adverse outcome in a multivariate regression analysis (p = 0.0020), even after adjustment for the effect of tumor stage. Therefore, we believe that the presence of marked nuclear pleomorphism should serve as the major criterion for making a diagnosis of serous carcinoma. In tumors with moderate nuclear pleomorphism, a diagnosis of serous carcinoma is aided by the presence of multinucleated cells, uneven papillary borders, high nuclear/cytoplasmic ratio, apical location of the nucleus, and hobnail cells. These features were significantly associated with a diagnosis of serous carcinoma in our patients.