Search

Purpose: To clarify the value of targeted versus off-target biopsies in men with a suspicion of prostate cancer (PC) and a visible lesion in multi-parametric magnetic resonance imaging (mpMRI) using transperineal robot-assisted biopsy., Methods: Fifty-five consecutive men with one non-palpable suspicious lesion in mpMRI after negative 12-core transrectal ultrasound-guided biopsy were enrolled in 2014-2015. Lesions were scored using the Prostate Imaging Reporting and Data System. A robot-assisted system was utilized to collect four robot-assisted targeted transperineal biopsy cores (RA-TB) within the lesion using mpMRI-TRUS elastic fusion. Untargeted transperineal 14-core biopsy was performed only outside the lesion (RA-UB). Histological grade was compared in biopsies and available prostatectomy specimens., Results: Overall, 34 of 55 patients (62%) were diagnosed with PC based on biopsy. 85% of cancers were clinically significant PC (csPC) defined as GS ≥ 7. 85% of biopsy-proven cancers were detected with RA-TB alone. RA-UB identified only one additional patient with csPC and lead to upgrading in five biopsy cases (14.7%). Pathological evaluation of 14 prostatectomy specimens showed upgrading in 2 patients (14.3%), while all other patients were correctly classified by RA-TB without need of additional RA-UB. Mean procedure duration was 43 (±6) min, and only minor complications according to Clavien-Dindo were recorded during 30-day follow-up., Conclusions: This is the first report of transperineal robot-assisted elastic mpMRI-TRUS fusion biopsy. RA-TB of positive MR lesions enabled reliable detection of csPC, while RA-UB in MRI-negative regions is of minor importance.

Multiparametric medical imaging data can be large and are often complex. Machine learning algorithms can assist in image interpretation when reliable training data exist. In most cases, however, knowledge about ground truth (e.g. histology) and thus training data is limited, which makes application of machine learning algorithms difficult. The purpose of this study was to design and implement a learning algorithm for classification of multidimensional medical imaging data that is robust and accurate even with limited prior knowledge and that allows for generalization and application to unseen data. Local prostate cancer was chosen as a model for application and validation. 16 patients underwent combined simultaneous [(11) C]-choline positron emission tomography (PET)/MRI. The following imaging parameters were acquired: T2 signal intensities, apparent diffusion coefficients, parameters Ktrans and Kep from dynamic contrast-enhanced MRI, and PET standardized uptake values (SUVs). A spatially constrained fuzzy c-means algorithm (sFCM) was applied to the single datasets and the resulting labeled data were used for training of a support vector machine (SVM) classifier. Accuracy and false positive and false negative rates of the proposed algorithm were determined in comparison with manual tumor delineation. For five of the 16 patients rates were also determined in comparison with the histopathological standard of reference. The combined sFCM/SVM algorithm proposed in this study revealed reliable classification results consistent with the histopathological reference standard and comparable to those of manual tumor delineation. sFCM/SVM generally performed better than unsupervised sFCM alone. We observed an improvement in accuracy with increasing number of imaging parameters used for clustering and SVM training. In particular, including PET SUVs as an additional parameter markedly improved classification results. A variety of applications are conceivable, especially for imaging of tissues without easily available histopathological correlation., (Copyright © 2015 John Wiley & Sons, Ltd.)

Objective: To directly compare the diagnostic performance of targeted MRI-guided biopsy (MR-GB) and systematic transrectal ultrasound-guided biopsy (TRUS-GB)., Methods: Thirty-five patients with at least one negative TRUS-GB, persistently elevated or rising prostate-specific antigen and a lesion suspicious for prostate cancer (PC) on multiparametric MRI (mpMRI) scored by using the Prostate Imaging Reporting and Data System (PI-RADS) were included. A median of three targeted biopsies per lesion were obtained and systematic TRUS-GB was performed subsequently by an independent urologist without knowledge of the MRI findings. Definite pathology reports were analyzed for anatomical location and criteria of clinical significance., Results: The tumor detection rate was significantly higher with MR-GB compared with TRUS-GB (16/35, 46% and 8/35, 23%, respectively, p < 0.05). MR-GB detected PC in all patients with positive TRUS-GB. All tumors detected by MR-GB exhibited at least one criterion of clinical significance. PC lesions showed a significantly higher PI-RADS sum score compared with benign lesions., Conclusions: MR-GB is more effective compared with TRUS-GB in detecting clinically significant PC in men after previous negative TRUS-GB. PI-RADS scores give additional information and could be part of the decision-making process when considering retrial biopsy. Additional systematic biopsy can be omitted in patients undergoing targeted MR-GB., (2014 S. Karger AG, Basel)