Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study
Authors Ahmed HU, El-Shater Bosaily A, Brown LC, et al.
Review Date February 2017
Citation Lancet 2017; Jan 19. pii: S0140-6736(16)32401-1. doi: 10.1016/S0140-6736(16)32401-1. [Epub ahead of print]
The diagnostic pathway for prostate cancer usually involves an elevated serum prostate specific antigen test and a digital rectal examination combined with clinical history followed by a transrectal ultrasound-guided prostate biopsy (TRUS-biopsy). In addition to the side-effects of TRUS-biopsy, including bleeding, pain, and infection, many men are found to either not have cancer at all or clinically insignificant cancers are identified. Moreover, some clinically significant cancers are missed with TRUS-biopsy. Recently evidence has been accruing that the use of multi-parametric magnetic resonance imaging (mp-MRI) can enhance the detection of aggressive cancers and reduce over-detection and improve targeted biopsies. It is also being investigated for use in active surveillance protocols and to aid focal treatments for prostate cancer.
To investigate whether mp-MRI can discriminate between men with and without clinically significant prostate cancer based on template prostate mapping biopsy (TPM-biopsy); to compare the accuracy of mp-MRI with TRUS; and to determine whether mp-MRI could be used as a triage test to avoid unnecessary TRUS-biopsy.
PROMIS was a prospective multi-centre, paired-cohort, confirmatory study to test diagnostic accuracy of MP-MRI and TRUS-biopsy against a reference test (TPM-biopsy). Men with prostate-specific antigen concentrations up to 15 ng/mL, with no previous biopsy, underwent 1·5 Tesla MP-MRI followed by both TRUS-biopsy and TPM-biopsy. The conduct and reporting of each test was done blind to other test results. Clinically significant cancer was defined as Gleason score ≥4 + 3 or a maximum cancer core length 6 mm or longer.
Sensitivity, specificity, positive and negative predictive values were calculated and compared between mp-MRI and TRUS biopsy using McNemar tests and odds ratios (ORs). Ratios were presented as TRUS relative to mp-MRI so ratios greater than 1.0 favour TRUS and less than 1.0 favour mp-MRI.
Between May 17, 2012, and November 9, 2015, 740 men were enrolled in the study across 11 centres. 576 underwent 1·5 Tesla mp-MRI followed by both TRUS-biopsy and TPM-biopsy. On TPM-biopsy, 408 of 576 men (71%; 95% confidence interval (CI) 67%-75%) had cancer with 230 of 576 patients (40%; 95% CI 36%-44%) with clinically significant cancer.
For clinically significant cancer, mp-MRI was more sensitive (93%, 95% CI 88%–96%) than TRUS-biopsy (48%, 42%–55%; OR=0.52, p<0·0001) and had higher negative predictive value (89%, 95% CI 83%-94%) than TRUS-biopsy (74%, 95% CI 69%-78%; OR=0.34, p<0.0001). However, mp-MRI was less specific (41%, 95% CI 36%-46% than TRUS (96%, 95% CI 94%–98%; OR=2.34; p<0·0001) and had lower positive predictive value: mp-MRI = 51% (95% CI 46%-56%) vs TRUS = 90% (95% CI 83%-94%; OR=8.2; p<0.0001).
The authors considered the implications of using mp-MRI by comparing the standard strategy of of TRUS-biopsy for all men to 2 alternative strategies using mp-MRI as a triage test with only men with a suspicious mp-MRI going on to either TRUS-biopsy or MRI-directed TRUS-biopsy (as the latter weren’t done in the study the authors assumed similar accuracy to TPM-biopsy). 27% (n=158) of men would avoid a primary biopsy under the triage scenarios. In the ‘best-case’ scenario, using MRI-guided biopsy, there could be up to 5% more clinically insignificant cancers detected (121 out of 576; 21%) compared to TRUS-biopsy alone (90 out of 576; 16%) but 5% fewer cases (n=62, 11%) in the triage scenario with TRUS-biopsy following triage with mp-MRI (‘worst-case’ scenario). In the ‘best case’ scenario, the correct diagnosis of clinically significant cancer might increase to 213 (37%) compared to standard TRUS-biopsy (n=111, 19%).
44 (5·9%) of 740 patients reported serious adverse events, including eight cases of sepsis.
The use of mp-MRI to triage men might allow 27% of patients avoid a primary biopsy and the diagnosis of 5% fewer clinically insignificant cancers. If subsequent TRUS-biopsies were directed by mp-MRI findings, up to 18% more cases of clinically significant cancer might be detected compared with the standard pathway of TRUS-biopsy for all. mp-MRI, used as a triage test before first prostate biopsy, could reduce unnecessary biopsies by a quarter. mp-MRI can also reduce over-diagnosis of clinically insignificant prostate cancer and improve detection of clinically significant cancer.
Points to Note
- The is the first study to present blinded data from multiple centres on the diagnostic accuracy of both mp-MRI and TRUS-biopsy against an accurate reference test.
- With mp-MRI triage, up to 27% of men with suspicion of prostate cancer might avoid a biopsy.
- The high negative predictive value is reassuring that a negative mp-MRI indicates a high probability of no significant prostate cancer.
- However, the low positive predictive value suggests that men with suspicious mp-MRI findings still need a biopsy.
- Limitations of the study include: a 5 mm sampling frame of the whole prostate was used which is too invasive for routine clinical use and prostates over 100mL had to be excluded; a selected group of men participated in the study; doing TPM-biopsy followed by TRUS-biopsy may have affected the accuracy of the TRUS-biopsy due to swelling and tissue disruption; MRI-targeted biopsies were not done; an assessment of inter-observer variability between expert and not so experienced reporters was not done.
- Cost-effectiveness analyses of the PROMIS data are underway and will be important to guide recommendations regarding the use of mp-MRI. Presently in Australia use of mp-MRI in prostate cancer diagnosis or treatment is not covered by Medicare.