4
views
0
recommends
+1 Recommend
1 collections
    0
    shares
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      Utilidad de las redes neuronales artificiales en la predicción de cáncer de próstata en la biopsia transrectal Translated title: The utility of artificial neural networks in the prediction of prostate cancer ontransrectal biopsy

      research-article

      Read this article at

      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Abstract

          Objetivo: Determinar si el desarrollo de una red neuronal artificial (RNA) formada por variables clínicas permite predecir el resultado de la biopsia prostática (BP). Material y métodos: Pacientes (n=953) sometidos a BP en el Hospital Arquitecto Marcide, Ferrol, entre enero-2000 y junio-2005. Las variables estudiadas fueron edad, PSA, tacto rectal y volumen prostático, disponiendo de todos estos datos en 843 casos. Para determinar factores relacionados con el diagnóstico de cáncer de próstata (CP), se desarrollaron un análisis de regresión logística y una red neuronal "feed-forward", con tres nodos en su capa oculta y un nodo de salida, que representa la probabilidad de CP. Ambos modelos fueron construidos a partir de una muestra aleatoria de n=643 pacientes (set de derivación). La capacidad predictiva de ambos modelos se valoró con los 200 pacientes restantes (set de validación), mediante curvas ROC y su área bajo la curva (ABC). Resultados: Se detectó CP en 500 (59,3%) casos. Ajustando por edad, PSA, tacto rectal y volumen prostático, en un modelo de regresión logística multivariante, se observó que todas las variables predecían CP de forma independiente. Las ABC fueron de 0,693 para el PSA, 0,707 para el volumen prostático, 0,815 para la regresión logística y 0,819 para la RNA. La capacidad predictiva de la RNA fue significativamente superior a la del PSA (p=0,002) y volumen prostático (p<0,001) y similar a la de la regresión logística (p=0,760). Conclusiones: La RNA presenta una capacidad de predicción de CP significativamente superior a los métodos diagnósticos unimodales, y similar a la regresión logística.

          Translated abstract

          Objective: To determine whether the development of an artificial neural network (ANN) made up of clinical variables allows for the prediction of prostate biopsy (PB) outcome. Materials and methods: Patients (n=953) underwent PB at the Arquitecto Marcide Hospital in Ferrol (Spain), between january 2000 and june 2005. The variables studied were age, PSA, digital rectal examination (DRE) and prostate volume, data for all of which were available in 843 cases. In order to determine factors related to prostate cancer (PC) diagnosis, a logistic regression analysis and a feed-forward neural network were developed, including three hidden layer nodes and an output node, representing the probability of PC. Both models were constructed from a random sample of n=643 patients (derivation set). The predictive capacity was assessed with the remaining 200 patients (validation set), by means of ROC curves and the area under the curve (AUC). Results: PC was detected in 500 (59.3%) cases. Adjusting for age, PSA, digital rectal examination and prostate volume, in a multivariate logistic regression model it was observed that all the variables were independent predictors of PC. The AUC were 0.693 for PSA, 0.707 for prostate volume, 0.815 for logistic regression and 0.819 for ANN. The predictive capacity of the ANN was significantly higher than that of the PSA (p=0.002) and prostate volume (p<0,001) and similar to that of logistic regression (p=0.760). Conclusions: The ANN shows a PC prediction capacity that is significantly higher than unimodal diagnosis methods, and similar to that of logistic regression.

          Related collections

          Most cited references39

          • Record: found
          • Abstract: found
          • Article: not found

          Artificial neural networks: opening the black box.

          Artificial neural networks now are used in many fields. They have become well established as viable, multipurpose, robust computational methodologies with solid theoretic support and with strong potential to be effective in any discipline, especially medicine. For example, neural networks can extract new medical information from raw data, build computer models that are useful for medical decision-making, and aid in the distribution of medical expertise. Because many important neural network applications currently are emerging, the authors have prepared this article to bring a clearer understanding of these biologically inspired computing paradigms to anyone interested in exploring their use in medicine. They discuss the historical development of neural networks and provide the basic operational mathematics for the popular multilayered perceptron. The authors also describe good training, validation, and testing techniques, and discuss measurements of performance and reliability, including the use of bootstrap methods to obtain confidence intervals. Because it is possible to predict outcomes for individual patients with a neural network, the authors discuss the paradigm shift that is taking place from previous "bin-model" approaches, in which patient outcome and management is assumed from the statistical groups in which the patient fits. The authors explain that with neural networks it is possible to mediate predictions for individual patients with prevalence and misclassification cost considerations using receiver operating characteristic methodology. The authors illustrate their findings with examples that include prostate carcinoma detection, coronary heart disease risk prediction, and medication dosing. The authors identify and discuss obstacles to success, including the need for expanded databases and the need to establish multidisciplinary teams. The authors believe that these obstacles can be overcome and that neural networks have a very important role in future medical decision support and the patient management systems employed in routine medical practice. Copyright 2001 American Cancer Society.
            Bookmark
            • Record: found
            • Abstract: not found
            • Article: not found

            How Neural Networks Learn from Experience

              Bookmark
              • Record: found
              • Abstract: found
              • Article: not found

              Comparison of percent free PSA, PSA density, and age-specific PSA cutoffs for prostate cancer detection and staging.

              Various methods have been proposed to increase the specificity of prostate-specific antigen (PSA), including age-specific PSA reference ranges, PSA density (PSAD), and percent free PSA (%fPSA). In this multicenter study, we compared these methods for their utility in cancer detection and their ability to predict pathologic stage after radical prostatectomy in patients with clinically localized, Stage T1c cancer. Seven hundred seventy-three men (379 with prostate cancer, 394 with benign prostatic disease), 50 to 75 years old, from seven medical centers were enrolled in this prospective blinded study. All subjects had a palpably benign prostate, PSA 4.0 to 10.0 ng/mL, and a histologically confirmed diagnosis. Hybritech's Tandem PSA and free PSA assays were used. %fPSA and age-specific PSA cutoffs enhanced PSA specificity for cancer detection, but %fPSA maintained significantly higher sensitivities. Age-specific PSA cutoffs missed 20% to 60% of cancers in men older than 60 years of age. %fPSA and PSAD performed equally well for detection (95% sensitivity) if cutoffs of 25% fPSA or 0.078 PSAD were used. The commonly used PSAD cutoff of 0.15 detected only 59% of cancers. %fPSA and PSAD also produced similar results for prediction of the post-radical prostatectomy pathologic stage. Patients with cancer with higher %fPSA values (greater than 15%) or lower PSAD values (0.15 or less) tended to have less aggressive disease. The results of this study demonstrated that cancer detection (sensitivity) is significantly higher with %fPSA than with age-specific PSA reference ranges. %fPSA and PSAD provide comparable results, suggesting that %fPSA may be used in place of PSAD for biopsy decisions and in algorithms for prediction of less aggressive tumors since the determination of %fPSA does not require ultrasound.
                Bookmark

                Author and article information

                Contributors
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Role: ND
                Journal
                aue
                Actas Urológicas Españolas
                Actas Urol Esp
                Asociación Española de Urología (, , Spain )
                0210-4806
                January 2006
                : 30
                : 1
                : 18-24
                Affiliations
                [01] Ferrol orgnameHospital Arquitecto Marcide - Profesor Novoa Santos orgdiv1Servicio de Urología
                [02] La Coruña orgnameHospital Juan Canalejo orgdiv1Unidad de Epidemiología Clínica y Bioestadística
                Article
                S0210-48062006000100003
                10.4321/s0210-48062006000100003
                b11ca9bd-3563-47fb-8bd6-3ae95b0dd8de

                This work is licensed under a Creative Commons Attribution-NonCommercial 3.0 International License.

                History
                Page count
                Figures: 0, Tables: 0, Equations: 0, References: 24, Pages: 7
                Product

                SciELO Spain


                Próstata,Biopsia,Neoplasia prostática,Red neuronal artificial,Regresión logística,Diagnóstico,Prostate,Biopsy,Prostatic neoplasms,Neural networks,Logistic regression,Diagnosis

                Comments

                Comment on this article