Standardized assessment to enhance the diagnostic value of prostate volume; Part I: Morphometry in patients with lower urinary tract symptoms.

BACKGROUND: The diagnostic value of prostate volume results has been evaluated in patients with prostate problems of benign cause. METHODS: For 247 patients, automated volume results were compared to manual results of planimetric reference volume and of the classical ellipsoid formula. Also, transition zone volume was estimated and growth curves of the prostate and prostate dimensions over age were investigated. RESULTS: Application of automated volume determination gives accurate results compared to the reference volume (Pearson correlation, R = 0.938). The ellipsoid volume results were slightly less correlated (R = 0.921). Average growth of the entire prostate was 1.7% per year, for the transition zone the growth was 4.3%. Compared to growth rates for a community-based population, comparable growth rates were found for our group that had higher mean prostate volume. CONCLUSIONS: The results indicate that the age of onset of volume growth is the determining factor in developing benign prostate enlargement not a change in growth rate.

Standardized assessment to enhance the diagnostic value of prostate volume; Part II: Correlation with prostate-specific antigen levels.

BACKGROUND: Standardized estimations of prostate volumes are used for interpretation of prostate specific antigen (PSA) levels. METHODS: In 243 patients with clinically benign diagnosis, automated and reference prostate volumes and transition zone volumes are correlated to PSA levels. Besides, growth curves of PSA level when aging are determined. RESULTS: The strongest correlation was found with multiple regression analysis between PSA and transition zone volume and rest volume (R = 0.854). Mean PSA density was 0.092 ng/mL2, its labelling quality for benign disease was 91% (threshold = 0.15 ng/mL2). An average growth factor for PSA levels of 2.9% per year was obtained. CONCLUSIONS: The contribution per unit tissue to PSA level was for the transition zone 1.9 times higher than for the rest volume. Average growth of PSA per year is in consonance with the increase in normal levels of age specific PSA ranges, although only weak correlations were found between PSA and age and PSA density and age.

Construction and application of hierarchical decision tree for classification of ultrasonographic prostate images.

A non-parametric algorithm is described for the construction of a binary decision tree classifier. This tree is used to correlate textural features, computed from ultrasonographic prostate images, with the histopathology of the imaged tissue. The algorithm consists of two parts; growing and pruning. In the growing phase an optimal tree is grown, based on the concept of mutual information. After growing, the tree is pruned by an alternating interaction of two data sets. Moreover, the structure and performance of the constructed tree are compared to the results using a slightly modified corresponding growing and pruning algorithm. The modified algorithm provides better retrospective and prospective classification results than the original algorithm. The use of the tree for automated cancer detection in ultrasonographic prostate images results in retrospective and prospective accuracy of 77.9% and 72.3%, respectively. Using this tissue characterisation, a supporting tool is provided for the interpretation of transrectal ultrasonographic images.

Computer analysis of transrectal ultrasound images of the prostate for the detection of carcinoma: a prospective study in radical prostatectomy specimens.

PURPOSE: The reliability and objectivity of computer assisted transrectal ultrasound are examined. MATERIALS AND METHODS: Pathological examination of radical prostatectomy specimens was compared prospectively to automated cancer detection in corresponding transrectal ultrasound images. RESULTS: For automated cancer detection, a sensitivity of 0.75 and a specificity of 0.78 were obtained. Moreover, 74% of human interpretation of the percentage of malignancy in the analyzed images was equal to the actual calculated percentage (Pearson's product moment correlation coefficient 0.85). CONCLUSIONS: Comparing these results to those obtained with normal transrectal ultrasound, automated analysis provides additional information in the interpretation of transrectal ultrasound images by color coding them in an objective manner according to the probability of malignancy.

Planimetric volumetry of the prostate: how accurate is it?

Planimetric volumetry is used in clinical practice when accurate volume determination of the prostate is needed. The prostate volume is determined by discretization of the 3D prostate shape. The are of the prostate is calculated in consecutive ultrasonographic cross-sections. This area is multiplied by the distance between the cross-sections and the total volume is determined by summation of all contributions. Besides the quality of the automated outlining, the accuracy of this method depends on this intersection distance and on the angle of the scan plane with the probe. Also, the location of the first cross-section is of influence. This paper describes the influences of these parameters on the accuracy of the volume determination using a simple prostate model. This theoretical influence is compared to clinical volume determinations using automated planimetric volumetry with different step sizes. From our data, it is concluded that a step size of 4 mm for planimetric prostate volume determination is a good compromise between investigation time and accuracy in a clinical setting.

Automated prostate volume determination with ultrasonographic imaging.

A method for automated determination of the prostate volume based on planimetric volumetry was developed to overcome subjectivity in prostate volume measurements using ultrasonography. An edge detection method is applied to locate the prostate boundary in sequential ultrasonographic images. During regular examinations, the volume was assessed automatically in 55 patients. The Pearson correlation coefficient of the automated volume compared to the reference prostate volume was 0.93. The estimated ratio of the automated volume to the reference volume was 0.92. This method overcomes the individual subjectivity when interpreting ultrasonographic images, and provides accurate and objective results for followup of benign prostatic hyperplasia treatments and for comparison of multicenter trials. Also, this method is a useful tool for objective determination of prostate specific antigen in proportion to the volume (prostate specific antigen density).

Computerized analysis of transrectal ultrasonography images in the detection of prostate carcinoma.

OBJECTIVE: To report on the use of automated image analysis in the interpretation of transrectal ultrasonographic images of the prostate. PATIENTS AND METHODS: During transrectal ultrasonography, images were recorded from biopsies performed in 127 patients. Subsequently in the image, the puncture place was marked and analysed. Analysis of the images was performed with the Automated Urologic Diagnostic Expert (AUDEX) system, consisting of a personal computer connected to the ultrasound machine. From the images collected, parameters can be calculated for image classification. The parameters obtained with this procedure were correlated with the histological result. RESULTS: Evaluation showed a sensitivity of 84.8% and specificity of 87.5%. The positive and negative predictive values, to predict prostate carcinoma, were 84.8% and 87.5%, respectively. CONCLUSION: Automated image analysis can help in the diagnosis of prostate carcinoma. In patients with non-palpable lesions or with poorly visualized tumours, image analysis is superior to the standard current diagnostic techniques.

Automated analysis and interpretation of transrectal ultrasonography images in patients with prostatitis.

Transrectal ultrasound (TRUS) offers a valuable complement to digital rectal examination (DRE) in diagnosing prostate diseases. However, in case of prostatitis syndromes, contradictions are found with regard to characteristic ultrasound features in these patients. Therefore we sought for better imaging techniques. This paper describes a study on the automated analysis of ultrasonographic prostate images (AUDEX). With image processing, in the present study, tissue characterization is performed to predict the probability of the presence of inflammated prostate tissue. This technique already proved its validity in patients with prostate cancer. During prostate examinations, images were recorded with clear indication of the puncture position. Only patients with an unambiguously inflamed or noninflamed benign histologic result were included. Evaluation showed that a sensitivity of 90.6% and a specificity of 64.2% was reached. Finally, the prospective positive and negative predictive value for prostatitis was 50.0% and 94.6%, respectively. This means that AUDEX predicts the diagnosis 'prostatitis' in a large number of patients with no infection. In case of prostatitis, however, this prediction is almost always right.

Imaging in BPH patients.

Benign prostatic hyperplasia is one of the most common pathological processes to afflict men. Strikingly, there is a large variety of methods of evaluation and therapeutic strategies for BPH in various countries. Before adequate treatment is possible, the process of benign adenomatous hyperplasia has to be distinguished from malignant or inflammatory prostatic disease. Furthermore, secondary alterations of the bladder and upper urinary tract have to be evaluated. A wide range of imaging techniques of the lower and upper urinary tract have evolved over the years. The specific benefits and restrictions of each of them are discussed. Finally, the recent developments and possibilities of computer-aided tissue characterisation and volume determination are outlined. The development of imaging techniques for BPH has just begun...

The reliability of computer analysis of ultrasonographic prostate images: the influence of inconsistent histopathology.

This article describes a method to investigate the influence of inconsistent histopathology during the development of tissue discrimination algorithms. Review of the pathology is performed on the biopsies used as training set of a computer system for cancer detection in ultrasonographic prostate images. The influence of the discrepancies found between independent pathologists on the discriminating power of the system is investigated. A high diagnostic consistency in histopathology concerning only the categories malignant and nonmalignant is found. Therefore, review of the pathology does not significantly influence the results of tissue discrimination algorithms for cancer detection. However a high interobserver variability is obtained in the differentiation between more histology classes.

A practical clinical method for contour determination in ultrasonographic prostate images.

This paper describes a practical method for automated determination of the contour of the prostate in ultrasonographic images. In this method, we use specific edge detection techniques, based on nonlinear Laplace filtering. Possible edges are located at zero-crossings of the second derivative of the image. The strength of the edge is reflected by the value of the gradient of the image at that location. Combining the information about location and strength, an edge intensity image is constructed from the initial ultrasonographic image. In our method, edge enhancement techniques are performed on the edge intensity image. Edges that actually represent a boundary are selected and linked: interpolation techniques are used to fill the gaps between detected boundary edges. The method for contour determination in ultrasonographic images is used for accurate volume measurements in an everyday clinical environment. Its computer implementation is fast, accurate (mean difference within 6% of exact volume) and easy to use.

Analysis of ultrasonographic prostate images for the detection of prostatic carcinoma: the automated urologic diagnostic expert system.

This paper describes a study on the automated analysis of ultrasonographic prostate images. With image processing, tissue characterization in the prostate was performed to assess the probability of malignancy. During prostate examinations, images were recorded at the positions where biopsies were taken. The used samples were divided into three groups. Two of them were used for the construction of a classification tree, and the third was used for the evaluation of this classification. A sensitivity of 80.6% and specificity of 77.1% were reached retrospectively. In a prospective way, these results were 80.0% and 88.2%, respectively. The prospective predictive value for cancer detection was 85.7%. The presented prospective value for image analysis was almost twice as high as the values normally found for prostate examination.