Computer decision support systems for the diagnosis of prostate cancer

Nikitaev V.G., Pronichev A.N., Prilepskaya V.A., Kovylina M.V., Pushkar' D.Yu.

Introduction. Histological examination is crucial for the final conclusion in the diagnosis of prostate cancer. The result depends on the experience and skill of the doctor performing the analysis of the microscopic picture of specimen of the prostate tissue. In this regard, task of automation of such analysis is the actual for improving objectivity and accuracy of the analysis.

Materials and methods. The article presents a computerized decision support system for the histological diagnosis of prostate cancer. Microscopic images of prostate specimens were used for source data of a knowledge base system. Descriptions of the images were made by medical experts. Options benign conditions of the prostate tissue, prostatic intraepithelial neoplasia of high degree, acinar adenocarcinoma (from the second to fifth stage according to Gleason) were considered. The doctor obtains a sample of the histological images most similar to the investigated case in the result of iterative interaction with the system. The sample images are accompanied by relevant expert opinions from the knowledge base of the system.

Results. The doctor produces a diagnostic conclusion by using this experts opinions and results of images comparisons. If the received data are not sufficient for diagnosis, the system provides on-line remote consultation with experts at the virtual microscopic specimen. The system must be equipped with a robotic microscope with a remote control to do consultation.

Conclusion. The system allows the expert to navigate through the specimen and research necessary fragments of the specimen. Application of the system in difficult diagnosis cases will help the doctor to obtain additional grounds for the formation of a diagnostic conclusion. In addition, the system can be used as a means of training and advanced training of physicians.

Authors declare lack of the possible conflicts of interests.

Download157.44 KB
prostate cancer, automated diagnosis, decision support system, automated microscopic image processing