Background
Methods
Patients and data acquisition
phi assay
mpMRI
Statistical analysis and decision modelling
Decision curve analysis
Results
Cohort description and between centre comparison
Cohort descriptors | N = 545 |
---|---|
Age (median) (inter-quartile range) | 66 (60,70) years |
PSA (median) (inter-quartile range) | 8 (6,13) ng/ml |
PHI (median) (inter-quartile range) | 44 (30,69) |
MRI (3–5) | 77% |
MRI (4–5) | 58% |
Any cancer detection | 64% |
≥ GG2 cancer detection | 47% |
≥ CPG3 cancer detection | 32% |
Performance of PSA, PSAd and phi in predicting prostate cancer at biopsy
≥ GG2 cancer detection | pvalue (vs. PSA) | ≥ CPG3 cancer detection | pvalue (vs.PSA) | |
---|---|---|---|---|
Whole cohort | ||||
PSA | 0.70 (0.66–0.74) | – | 0.81 (0.78–0.85) | – |
PSAd | 0.79 (0.75–0.83) | < 0.001 | 0.84 (0.80–0.87) | 0.12 |
phi | 0.82 (0.78–0.85) | < 0.001 | 0.87 (0.84–0.90) | < 0.001 |
MRI | 0.63 (0.59–0.66) | < 0.001* | 0.63 (0.60–0.66) | < 0.001* |
MRI + PSA | 0.76 (0.72–0.80) | < 0.001 | 0.85 (0.81–0.88) | 0.03 |
MRI + PSAd | 0.81 (0.77–0.84) | < 0.001 | 0.85 (0.81–0.88) | 0.08 |
MRI + phi | 0.81 (0.78–0.85) | < 0.001 | 0.86 (0.83–0.90) | 0.02 |
MRI negative | ||||
PSA | 0.64 (0.52–0.76) | – | 0.86 (0.76–0.97) | – |
PSAd | 0.76 (0.64–0.87) | 0.01 | 0.95 (0.91–0.99) | 0.08 |
phi | 0.78 (0.66–0.90) | 0.01 | 0.89 (0.74–1.00) | 0.76 |
phi thresholds in selecting men for referral to mpMRI and biopsy
≥ GG2 cancer detection | ≥ CPG3 cancer detection | ||||||||
---|---|---|---|---|---|---|---|---|---|
Threshold | Sensitivity | Specificity | NPV | % cancers missed* | Sensitivity | Specificity | NPV | % cancers missed** | |
phi | 20 | 0.99 | 0.10 | 0.90 | 1.1 | 1.00 | 0.08 | 1.00 | 0 |
25 | 0.96 | 0.25 | 0.87 | 4.2 | 0.99 | 0.22 | 0.99 | 0.05 | |
30 | 0.92 | 0.40 | 0.85 | 7.7 | 0.95 | 0.35 | 0.94 | 1.0 | |
35 | 0.87 | 0.55 | 0.83 | 12.8 | 0.93 | 0.49 | 0.93 | 7.3 | |
PSAd | 0.10 | 0.93 | 0.31 | 0.82 | 7.3 | 0.97 | 0.28 | 0.95 | 2.8 |
0.15 | 0.81 | 0.57 | 0.77 | 18.6 | 0.90 | 0.53 | 0.92 | 9.6 | |
0.20 | 0.69 | 0.77 | 0.73 | 31.3 | 0.80 | 0.72 | 0.88 | 20.0 |
Impact of using phi as a triaging test into an image-guided diagnostic pathway
Pathway | ||||||
---|---|---|---|---|---|---|
MRI + biopsy all | MRI all + biopsy if M3–5 | MRI all + biopsy if M3–5 or PSAd ≥ 0.15 | MRI all + biopsy if M3–5 or PSAd ≥ 0.10 | MRI + biopsy only ifphi ≥ 25 | MRI + biopsy only ifphi ≥ 30 | |
No. of MRI scans (% lower*) | 1000 | 1000 (0%) | 1000 (0%) | 1000 (0%) | 850 (−15%) | 750 (−25%) |
No. of biopsies needed (% lower*) | 1000 | 772 (−23%) | 884 (−12%) | 941 (−6%) | 850 (−15%) | 750 (−25%) |
Detection of ≥ GG2 cancers | ||||||
Cancers identified (% lower*) | 100% | 91% (−9%) | 98% (−2%) | 99% (−1%) | 96% (−4%) | 92% (−8%) |
Unnecessary biopsies (% lower*) | 527 | 343 (−35%) | 420 (−20%) | 471 (−11%) | 396 (−25%) | 314 (−40%) |
Detection of ≥ CPG3 cancers | ||||||
Cancers identified (% lower*) | 100% | 95% (−5%) | 100% (0%) | 100% (0%) | 99% (−1%) | 95% (−5%) |
Unnecessary biopsies (% lower*) | 677 | 466 (−31%) | 562 (−17%) | 618 (−9%) | 529 (−21%) | 442 (−35%) |
Cost analysis | ||||||
Cost/pt. (% lower*) | £965 | £796 (−18%) | £879 (−9%) | £921 (−5%) | £869 (−10%) | £774 (−20%) |