Proposed Adjustments to PI-RADS v2 may increase diagnostic performance in detection of significant prostate cancer

/
Rosenkrantz AB, Babb JS, Taneja SS, Ream JM. Proposed Adjustments to PI-RADS Version 2 Decision Rules: Impact on Prostate Cancer Detection. Radiology. 2016 Oct 26. PubMed PMID: 27783538.

Rosenkrantz AB, Babb JS, Taneja SS, Ream JM. Proposed Adjustments to PI-RADS Version 2 Decision Rules: Impact on Prostate Cancer Detection. Radiology. 2016 Oct 26. PubMed PMID: 27783538.

Short summary 

Adjustments to the existing PI-RADS v2 algorithm are proposed to increase the detection of prostate cancer with a GS ≥ 7.

Conclusion

While the existing PI-RADS v2 algorithm and its decision rules showed a reasonably high diagnostic performance, three proposed adjustments to the decision rules were additionally associated with GS ≥ 7 tumor in 20% or more of upgraded cases:

  • Upgrading lesions in the TZ from 3→4 based on a DWI score of 4 (instead of 5).

  • Upgrading lesions in the TZ from 3→4 based on a (newly introduced) DCE score of

    "positive".

  • Upgrading lesions in the PZ and TZ from 4→5 based on a size ranging of 10-14mm (instead of ≥15mm).

Patient cohort

Retrospective study of 343 patients undergoing MRI prior to MR/US Fusion targeted biopsy.

The New and the Good

  • First study to propose detailed adjustments to the PI-RADS v2 decision algorithm.
  • The proposed adjustments do not complicate the existing algorithm and keep it simple and applicable.
  • Reasonably high patient number.
  • Incorporation of morphologic criteria to DCE evaluation (sheetlike enhancement). 
  • Incorporation of T2w-scores in the PZ did NOT add value to the existing algorithm.

Limitations

  • Proposed adjustments are focused mainly on "tie-braking" rules (upgrade from 3→4).
  • While the overall patient number (n=343) was reasonably high, the sample sizes for evaluating the individual proposed adjustments (n=1-49) and patients having GS ≥ 7 tumor (n=90) were quite small.

Possible consequences for clinical practice

  • To early to apply the proposed adjustments into clinical practice yet, however, one may be a bit less strict in measuring the lesions ;-)

Possible implications for PIRADS v3

  • The current size cut-off of 15mm for upgrading from 4→5 in PI-RADS may have to be re-evaluated and lowered to 10mm.
  • The current upgrade cut-off in the TZ of DWI-score of 5 may may have to be re-evaluated and lowered to a DWI-score of ≥ 4.
  • In view of the reasonably high percentage (33.3%-57.1%) of GS ≥ 7 in the TZ in patients with a T2-score of 3 and positive DCE, the introduction of an "adapted" DCE-score in the TZ may be re-evaluated. 

Future study ideas derived from this paper

  • Validation of these proposed adjustments should be performed in larger and preferably multi-centric studies.  

Personal comment

  • In my personal opinion,  lowering the size threshold to be able to call a lesion a PI-RADS 5 lesion is helpful, as it is sometimes disappointing to have to call an obvious cancer a "4" only because of it not making the size-cutoff.