Abstract:
Prostate cancer has become the most common malign tumor in males. Up to now there ist no reliable imaging technique that gives an exact picture of the tumor areas, which makes diagnosis and staging difficult. Magnetic resonance imaging with endorectal coil (eMRI) is currently dicussed as a new promising method. Several studies evaluate the ability of this method to detect prostate cancer. However, information about the accuracy of this method and its ability to detect tumors in the different parts of the prostate is sparse.
This study evaluates variations of local sensitivity, specificity and accuracy achieved by eMRI in different areas of the prostate gland. Whole-mount-series from prostate specimens of 69 consecutive patients were used, the prostate gland was divided into areas according to clinical zoning. The results of magnetic resonance imaging were then compared to the results of histology in every single area, allowing to detect area-related variations of sensitivity, specificity and accuracy.
Subdivision of the prostate gland in an anterior and a posterior part resulted in a local sensitivity of 65% vs. 88%, a local specificity of 87% vs. 100% and a local accuracy of 70% vs. 88%. Subdivision into apex, middle and base showed a rise of local sensitivity and of accuracy from apex to base: apex (54% and 59%), middle (76% and 77%), base (84% and 83%). Subdivision of the prostate in 12 symmetrical areas (anterior/posterior, right/left, apex/middle/base) showed a rise of local sensitivity from anteroapical to the posterobasal area, while local specificity decreased in the same direction.
This study shows that the quality of detedtion of prostate cancer by eMRI is dependent on the exact location of the tumor in the prostate gland; its data can therefore help to improve the interpretation of eMRI results.