Send to:

Choose Destination
See comment in PubMed Commons below
Acta Radiol. 2007 Jul;48(6):695-703.

Evaluation of the potential of diffusion-weighted imaging in prostate cancer detection.

Author information

  • 1Clinical Magnetic Resonance Group, Royal Marsden NHS Foundation Trust, Institute of Cancer Research, Department of Statistics, Royal Marsden NHS Foundation Trust Sutton, Surrey, United Kingdom.



Conventional T2-weighted (T2W) imaging alone has a poor sensitivity for prostate cancer detection.


To evaluate combined T2W and diffusion-weighted magnetic resonance imaging (DW-MRI) versus T2W MRI alone for identifying tumor in patients with prostate cancer.


Fifty-four consecutive patients with prostate cancer (46 stage 1 and 2, 8 stage 3) and sextant biopsies within the previous 3 months were studied. Endorectal MR images were analyzed by two radiologists (1 experienced, 1 trainee) blinded to patient information and histopathology. T2W images were scored first, followed by combined T2W and isotropic apparent diffusion coefficient (ADC) maps calculated from DW-MRI (b = 0, 300, 500, and 800 s/mm(2)). Gland apex, middle, and base for each side were scored negative, indeterminate, or positive for tumor. Imaging data for each sextant were compared with histology. Sensitivity, specificity, and interobserver agreement were calculated.


Sensitivity and specificity for tumor identification significantly improved from 50% and 79.6% (T2W alone, experienced observer) to 73.2% and 80.8% (P<0.001), respectively. For the trainee observer, there was no improvement (44.3% and 72% T2W alone vs. 45.1% and 69.2% T2W plus ADC maps). Interobserver agreement was moderate for T2W imaging alone (kappa 0.51) and fair for T2W plus ADC maps (kappa 0.33).


In an experienced observer, DW-MRI together with T2W imaging can significantly improve tumor identification in prostate cancer.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for HighWire
    Loading ...
    Write to the Help Desk