High-resolution diffusion tensor imaging of prostate cancer using a reduced FOV technique.

2011 
Abstract Objective Diffusion tensor imaging (DTI) offers the promise of improved tumor localization in prostate cancer but the technique suffers from susceptibility-induced artifacts that limit the achievable resolution. The present work employs a reduced field-of-view technique that enables high-resolution DTI of the prostate at 3 T. Feasibility of the approach is demonstrated in a clinical study including 26 patients and 14 controls. Materials and methods Reduced field-of-view acquisition was established by non-coplanar application of the excitation and the refocusing pulse in conjunction with outer volume suppression. Accuracy for cancer detection of apparent diffusion coefficient (ADC) mapping and T 2 -weighted imaging was calculated and compared with reference to the findings of trans-rectal ultrasound-guided octant biopsy. Mean ADCs and fractional anisotropy (FA) values in the patients with positive and negative biopsies were compared to each other and to the controls. Results Fine anatomical details were successfully depicted on the ADC maps with sub-millimeter resolution. Accuracy for prostate cancer detection was 73.5% for ADC maps and 71% for T 2 -weighted images, respectively. Cohen's kappa ( κ  = 0.48) indicated moderate agreement of the two methods. The mean ADCs were significantly lower, the FA values higher, in the patients with positive biopsy than in the patients with negative biopsy and the controls. Monte Carlo simulations showed that the FA values, but not the ADCs, were slightly overestimated. Bootstrap analysis revealed that the ADC, but not the FA value, is a highly repeatable marker. Conclusion In conclusion, the present work introduces a new approach for high-resolution DTI of the prostate enabling a more accurate detection of focal tumors especially useful in screening populations or as a potential navigator for image-guided biopsy.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    20
    References
    46
    Citations
    NaN
    KQI
    []