Comparison of qualitative and quantitative approach to prostate MR spectroscopy in peripheral zone cancer detection

• Published in: European journal of radiology Vol. 81; no. 3; pp. 411 - 416
• Main Authors: Klijn, Stijn, De Visschere, Pieter J, De Meerleer, Gert O, Villeirs, Geert M
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier Ireland Ltd 01-03-2012
• Subjects: Adult; Aged; Aged, 80 and over; Area Under Curve; Diagnostic techniques and procedures; Humans; Image interpretation; Image Interpretation, Computer-Assisted - methods; Magnetic resonance imaging; Magnetic resonance spectroscopy; Magnetic Resonance Spectroscopy - methods; Male; Middle Aged; Pattern Recognition, Automated; Prospective Studies; Prostatic neoplasms; Prostatic Neoplasms - diagnosis; Prostatic Neoplasms - pathology; Radiology; Retrospective Studies; ROC Curve; Sensitivity and Specificity; Software; Magnetic resonance spectroscopy; Prostatic neoplasms; Diagnostic techniques and procedures; Magnetic resonance imaging; Image interpretation
• ISSN: 0720-048X; 1872-7727
• DOI: 10.1016/j.ejrad.2010.12.017

Abstract Objective To compare the diagnostic performance of a qualitative (pattern recognition) and a quantitative (numerical assessment) approach to magnetic resonance spectroscopy (MRS) in the diagnosis of peripheral zone prostate cancer. Methods 185 patients (131 with histopathologically proven cancer, 54 normal/benign after at least 12 months follow-up) were prospectively evaluated with qualitative MRS using a 4-point scale between 3/2004 and 1/2008, and retrospectively reassessed using a prototype quantitative postprocessing software in April 2008. Based on pathology and follow-up data, diagnostic performance parameters were calculated. Results The qualitative and quantitative approaches were concordant in 78.9% (146/185) of cases. The difference between the areas under the ROC curve (0.791 versus 0.772, respectively) was not statistically significant. The sensitivity, specificity and accuracy were 55.7%, 94.4% and 67.0% for the qualitative approach, and 55.0%, 83.3% and 63.2% for the quantitative approach. The sensitivity for high grade tumours (Gleason 4 + 3 or higher) was 85.2% (23/27) for both approaches. All cancers missed on either one approach separately (31/31) and 91% of cancers missed on both approaches together (23/27) were of lower grade (Gleason 3 + 4 or lower). Conclusions Qualitative and quantitative approaches to MRS yield similar diagnostic results. Discordances in tumour detection only occurred in lower grade cancers.