DCE-MRI of the prostate: what's an adequate temporal resolution?

J Magn Reson Imaging. 2016 Sep 20. doi: 10.1002/jmri.25451. [Epub ahead of print]

Short summary 

Investigation on temporal resolution needed in DCE-imaging of the prostate.

Conclusion

Temporal resolutions of ≤10s add no benefit in tumor detection or quantitative analysis of pharmacokinetic parameters. 

Patient cohort

Retrospective study of 58 consecutive patients undergoing MRI prior to prostatectomy.

The New and the Good

• First study since the simplification of DCE-Interpretation in PIRADS v2 (categorization in a binary fashion based on visual interpretation of early enhancement) which investigates the effect of temporal resolution on tumor detection.
• Histopathologic correlation of prostate cancer foci available.
• Due to the GRASP framework used, post-hoc reconstruction of different temporal resolutions within the same patient was possible. 

Limitations

• The used GRASP framework is a technology not yet widely available. It remains unclear whether the same cut offs for temporal resolutions may be applied when using conventional DCE-sequences. 
• Study might be underpowered to show significance of slight differences in diagnostic performance/pharmacokinetic parameters.
• Selection bias (all patients underwent prostatectomy; tumors without DCE-abnormalities were excluded in pre-reading).

Possible consequences for clinical practice

• DCE-Protocols may have to be re-analyzed and the trade-off between temporal resolution and spatial resolution may have to be shifted away from temporal resolution towards spatial resolution.

Possible implications for PIRADS v3

• The current recommendation of a "preferred temporal resolution" of ≤7s may be omitted in the next version of the guidelines.
• In view of the results that even a temporal resolution of 14.9s did not show a significant decline in diagnostic performance in the present study, the recommendations for the temporal resolution may even be lowered to 15s. 

Future study ideas derived from this paper

• A study investigating the ALARA (As Low As Reasonably Achievable)-question might be interesting to find the lowest temporal resolution possible without impairing the diagnosis of prostate cancer. The gained time may then be invested in increasing spatial resolution or SNR.  

Personal comment

• In our institution, we use a temporal resolution of about 8 seconds for DCE and have opted for a reasonably high spatial resolution and SNR despite the PIRADS v2-recommendation to go below 7 seconds. Considering the results of this paper, we might go back to an even slower temporal resolution (15s), however. What temporal resolutions do you use at your institutions?