Angel Borque

The use of neural networks and logistic regression analysis for predicting pathological stage in men undergoing radical prostatectomy: a population based study.

PURPOSE Clinical under staging occurs in 40% to 60% of patients who undergo radical prostatectomy for prostate cancer. To decrease under staging several methods of predicting pathological stage preoperatively have been developed based on statistical logistic regression analysis and neural networks. To our knowledge none has been validated in our homogeneous regional patient population to date. We created logistic regression and neural network models, and implemented and adapted them into our practice. We also compared the 2 methods to determine their value and practicality in daily clinical practice. We present the results of our novel approach for predicting pathological staging of prostate adenocarcinoma. MATERIALS AND METHODS Between 1986 and 1999, 600 white men from the Aragon region of Spain underwent surgery for prostate cancer; of whom 468 were selected for study. Predictive study variables included patient age, clinical stage, biopsy Gleason score and preoperative prostate specific antigen (PSA). The predicted result included in analysis was organ confined or nonorgan confined disease. Data were analyzed by multivariate logistic regression and a supervised neural network (multilayer perceptron and radial basis function). Results were compared by comparing the areas under the receiver operating characteristics curves. RESULTS We generated 5 logistic regression models. The model created with clinical staging, Gleason biopsy score and PSA distributed in 5 categories (p <0.001) with an area under the receiver operating characteristics curve of 0.840 proved to be most predictive of pathological stage. Similarly of the 6 neural network models evaluated the radial basis function model, which included age, clinical stage, Gleason biopsy score and preoperative PSA distributed in 5 categories with an area under the curve of 0.882, proved the most predictive but not superior to the logistic regression model. The difference in the area under the curves in the 2 chosen models was 0.042 (p = 0.1). CONCLUSIONS It is possible to generate useful predictive models of organ confined disease using logistic regression or neural networks with high indexes of clinical and statistical validity. However, using these variables neural networks did not prove to be better than logistic regression analysis. Therefore, better predictive variables must be identified, preferably nonlinear characteristics with respect to the probability of organ confined tumor, to generate better predictive models using neural networks.

Improved Prediction of Biochemical Recurrence After Radical Prostatectomy by Genetic Polymorphisms

PURPOSE Single nucleotide polymorphisms are inherited genetic variations that can predispose or protect individuals against clinical events. We hypothesized that single nucleotide polymorphism profiling may improve the prediction of biochemical recurrence after radical prostatectomy. MATERIALS AND METHODS We performed a retrospective, multi-institutional study of 703 patients treated with radical prostatectomy for clinically localized prostate cancer who had at least 5 years of followup after surgery. All patients were genotyped for 83 prostate cancer related single nucleotide polymorphisms using a low density oligonucleotide microarray. Baseline clinicopathological variables and single nucleotide polymorphisms were analyzed to predict biochemical recurrence within 5 years using stepwise logistic regression. Discrimination was measured by ROC curve AUC, specificity, sensitivity, predictive values, net reclassification improvement and integrated discrimination index. RESULTS The overall biochemical recurrence rate was 35%. The model with the best fit combined 8 covariates, including the 5 clinicopathological variables prostate specific antigen, Gleason score, pathological stage, lymph node involvement and margin status, and 3 single nucleotide polymorphisms at the KLK2, SULT1A1 and TLR4 genes. Model predictive power was defined by 80% positive predictive value, 74% negative predictive value and an AUC of 0.78. The model based on clinicopathological variables plus single nucleotide polymorphisms showed significant improvement over the model without single nucleotide polymorphisms, as indicated by 23.3% net reclassification improvement (p = 0.003), integrated discrimination index (p <0.001) and likelihood ratio test (p <0.001). Internal validation proved model robustness (bootstrap corrected AUC 0.78, range 0.74 to 0.82). The calibration plot showed close agreement between biochemical recurrence observed and predicted probabilities. CONCLUSIONS Predicting biochemical recurrence after radical prostatectomy based on clinicopathological data can be significantly improved by including patient genetic information.

Genetic predisposition to early recurrence in clinically localized prostate cancer.

Currently available nomograms to predict preoperative risk of early biochemical recurrence (EBCR) after radical prostatectomy are solely based on classic clinicopathological variables. Despite providing useful predictions, these models are not perfect. Indeed, most researchers agree that nomograms can be improved by incorporating novel biomarkers. In the last few years, several single nucleotide polymorphisms (SNPs) have been associated with prostate cancer, but little is known about their impact on disease recurrence. We have identified four SNPs associated with EBCR. The addition of SNPs to classic nomograms resulted in a significant improvement in terms of discrimination and calibration. The new nomogram, which combines clinicopathological and genetic variables, will help to improve prediction of prostate cancer recurrence.

Malignant mesothelioma of the tunica vaginalis. Report of a case without risk factors and review of the literature.

Malignant mesothelioma of the tunica vaginalistestis is an aggressive tumour with localrecurrence being distant metastases the mainfeature of the clinical course. Usually appearsover the fourth decade, having a strongrelationship with occupational exposure toasbestos and long lasting hydrocele. Weintroduce a case of a 78-year-old caucasianmale who developed a malignant mesotheliomawithout personal history of hydrocele orexposure to asbestos. A revision of the currentliterature is performed to summarize the recenttherapeutic options as well as new diagnostictools.

Atypical small acinar proliferation Review of a series of 64 patients

Objective. To study the evolution of 64 patients initially diagnosed with ASAP (atypical small acinar proliferation). Material and methods. Between 1998 and the end of 2003, 64 patients were diagnosed at our centre with ASAP. Results. The mean age of the patients assessed was 69 years (SD 6.4 years), the median prostate-specific antigen (PSA) level was 7.1 ng/ml (range 2–39 ng/ml) and 25% of the patients had a suspicious rectal examination. These 64 patients were subjected to re-biopsy. At re-biopsy, we diagnosed 27 patients (42%) with prostate adenocarcinoma. We classified patients into two groups depending on whether they did (n=27) or did not (n=37) have tumours. There were no significant differences in median PSA level between the two groups. The rectal examination was suspicious in 14% of patients without tumours and in 39% with tumours. Radical prostatectomy was applied to 20/28 patients (71%) diagnosed with prostate cancer. In these 20 patients, the median tumour volume was 0.4 cm3 (range 0.1–3.2 cm3) and 44% of the tumours were significant. The 37 patients with an unsuspicious histology were subjected to follow-up for a median of 12 months (range 1–30 months). The median PSA level in these patients was 5.7 ng/ml (range 0.8–28 ng/ml). A third biopsy was performed in three of these patients in view of an elevated PSA level, and one result was positive. Conclusions. In our experience, a pathological result of ASAP is associated with a definitive diagnosis of prostate cancer in 42% of cases. Moreover, a significant cancer was found in 44% of patients subjected to radical prostatectomy. We therefore systematically perform repeat biopsies on all patients with a histological result of ASAP.

Implementing the use of nomograms by choosing threshold points in predictive models: 2012 updated Partin Tables vs a European predictive nomogram for organ-confined disease in prostate cancer.

To implement the use of nomograms in clinical practice showing how to choose thresholds in nomograms’ predictions to select risk groups. To validate and compare the predictive ability and clinical utility of the Hospital Universitario ‘Miguel Servet’ (HUMS) and the updated Partin Tables 2012 (PT‐2012) nomograms to predict organ‐confined disease (OCD) after radical prostatectomy (RP).

Significance of Protein p53 Overexpression in the Clinical Course of High-Risk Superficial Bladder Cancer

Objectives: This is a retrospective study in which the long-term biological behavior of 67 ‘high-risk’ superficial bladder tumors and the prognostic relevance (prediction of disease recurrence and progression) of the determination of the p53 phenotype in these cases were studied. Material and Methods: 67 tumors with a ‘high risk’ of progression were selected from the 1,103 transurethral resections for bladder cancer carried out in 640 patients in this center between 1987 and 1992. These included 39 T1G3, 14 Tis (isolated or associated with Ta-T1, non-G3 tumors), and 14 Ta-T1, non-G3 tumors with submucosal lymphatic affection (L+). The median follow-up of these cases was 69.7 months. An immunohistochemical technique with monoclonal antibodies (DO-7) was used to detect the p53 phenotype in paraffin-fixed material. Results: Tumor recurrence occurred in 31 patients (46.3%) and local or distant progression in 14 (20.9%). Radical cystectomy was carried out in 16 (23.9%) cases. p53 overexpression of ≧20% (‘p53+’) was detected in 40 tumors (59.7%). The rate of recurrence and progression, the disease and progression-free intervals, cancer-specific survival, disease-free survival and progression-free survival were similar in the 3 tumor groups (in all cases, p > 0.05). There were no significant differences in the overexpression of protein p53, using the standard cutoff point of 20% stained nuclei, on comparing the same variables in the whole group of 67 patients (in all cases, p > 0.05). Conclusion: The detection of protein p53 was not found to be of use in the retrospective prediction of disease progression or survival in ‘high-risk’ superficial bladder cancer.

A step-by-step algorithm for combining diagnostic tests

Combining data of several tests or markers for the classification of patients according to their health status for assigning better treatments is a major issue in the study of diseases such as cancer. In order to tackle this problem, several approaches have been proposed in the literature. In this paper, a step-by-step algorithm for estimating the parameters of a linear classifier that combines several measures is considered. The optimization criterion is to maximize the area under the receiver operating characteristic curve. The algorithm is applied to different simulated data sets and its performance is evaluated. Finally, the method is illustrated with a prostate cancer staging database.

Comparison of fetal weight distribution improved by paternal height by Spanish standard versus Intergrowth 21st standard

Abstract Objective: Our main objective was to study the influence on birth and ultrasound fetal weight of traditional factors in combination with non-traditionally explored predictors such as paternal height to provide a new customized in utero growth model. We also have compared it in our population with other customized and non-customized models. Methods: We collected 5243 cases of singleton pregnancies. An integrated study of the different variables was performed in a multivariate model to predict the fetus birthweight and customized growth curves were created following the Gardosi procedure. Results: Gestational age (P<0.001), parity (P<0.001), maternal age (P<0.001), maternal body mass index (P<0.001), maternal height (P<0.001), parental height (P<0.001), pregnancy-associated plasma protein A (PAPP-A) (P<0.001), free-beta human chorionic gonadotropin (FBHCG) (P<0.013), single umbilical artery (SUA) (P<0.009), region of origin (P<0.001), fetal sex (P<0.001), smoking (P<0.001) and pre-gestational diabetes (P<0.001) showed statistical significance. We created two growth customized models (simple and advance) that have shown good performance in predicting fetal weight at delivery and estimated by ultrasounds. The percentage of small for gestational age (SGA) cases (P10) predicted by the two models at birth were 9.9% and 9%, and for large gestational ages (LGA) (P90) we obtained values of 90.1% and 90.3%. Also, using the fetal weights measured by ultrasounds, we obtained P10 adjusted predictions, 9.2% and 9.4%, for the simpler and advance models, respectively, which were more adjusted than the 0.4, 4.6 and 10.6 obtained using the other compared models. For an easy use of models an app and a nomogram is provided. Conclusion: Using new predictor variables we implemented new growth in utero model, with predictions more adjusted to our population than Spanish customized or Intergrowth 21st models with better performance for birth and ultrasound fetal weights. We propose using a prediction model that includes parental height.

MP53-05 TESTOSTERONE RECOVERY AFTER LONG TIME DEPRIVATION THERAPY: PREDICITIVE FACTORS AND MODELS (NOMOGRAMS)

was 27.9 months (range: 3.3-114.6). Median PSA prior to initiation of ADT was 18 ng/mL (range: 0.61-2940). 72.3% of patients achieved a 1year mean T < 20 ng/dl; 18.6% achieved 20-32 ng/dl; 5.4% achieved 32-50 ng/dl; and 3.6% achieved > 50 ng/dl. There was no statistically significant difference in progression-free survival between patients with different levels of 1-year mean testosterone values (log-rank p1⁄40.813). CONCLUSIONS: The results suggest that there may not be a significant impact of strict testosterone control beyond what is considered the traditional castrate-level testosterone. However, only a small proportion of patients had 1-year testosterone > 32 ng/dl (9.0%). A larger study may reveal a beneficial role of strict testosterone reduction in the management of advanced prostate cancer.