Tobias Penzkofer MD, Presenter: Nothing to Disclose

Andriy Fedorov PhD, Abstract Co-Author: Nothing to Disclose

Kemal Tuncali MD, Abstract Co-Author: Research Grant, Canon Inc

Fiona Mary Fennessy MD, PhD, Abstract Co-Author: Nothing to Disclose

Junichi Tokuda PhD, Abstract Co-Author: Research Grant, Siemens AG

Clare M. C. Tempany-Afdhal MD, Abstract Co-Author: Research Grant, InSightec Ltd Research Consultant, Profound Medical Inc

To evaluate the impact of multireader review prior to MRI-guided prostate biopsy. We hypothesized that (a) multireader review increases the positive yield MRI-guided prostate biopsy and (b) high grade lesions are reliably chosen by a high number of readers.

The study was approved by the Institutional Review Board and is compliant with HIPAA regulations. 59 patients (66.9 ± 6.5 years, PSA 12.3 ± 10.4 ng/ml) underwent MRI-guided transperineal prostate biopsy for suspicious prostate lesions (3.9 ± 1.8 per prostate) chosen through independent review by three radiologists with at least 10 years of prostate MRI experience. The set of targets was consolidated by one reader prior to biopsy, when multiple readers selected the same lesion. The number of readers per lesion was tabulated against the histopathological outcome and the grade of the lesion and compared. 

Thirty-eight of the 59 patients (64.4%) had positive biopsies. There were 73 (31.7%) cancer cores of 230 samples. 117 of the lesions were selected by one reader, 62 were chosen by two readers, and 51 by all three readers. The lesions selected by three readers were positive in 64.7% (33/51), lesions selected by two readers in 33.9% (21/62) and lesions selected by one reader in 16.2% of the cases (19/117). Lesions chosen by three radiologists had a significantly higher percentage of malignancy than the lesions selected by 2 or only one reader (chi², p≤0.0021). Among the 38 patients with a positive diagnosis, a highest grade lesion was selected in 25 (65.8%) by 3 readers, in 10 cases (26.3%) by 2 readers and 3 (7.9%) only by one reader. 

Multireader review of pre-biopsy MRI prior to MRI-guided prostate biopsy revealed a higher probability of malignancy in lesions selected by all readers, compared to lesions chosen by fewer readers. A high level of agreement between readers could be a marker for determining the highest grade lesion, although there were a number of lesions not detected by all of the three readers.

Multireader protocol in pre-biopsy target selection has a potential to either reduce the number of targets which need to be biopsied, or to increase the yield of MRI guided biopsy. Further evaluation of such protocols, similar to the BI-RADS multireader protocol, is warranted.

Penzkofer, T, Fedorov, A, Tuncali, K, Fennessy, F, Tokuda, J, Tempany-Afdhal, C, Multireader Review in Prostate Magnetic Resonance Imaging: Investigation of the Effect on Detection Rate and Sample Grade.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14015892.html