Improving beyond current prostate-specific antigen testing to predict prostate cancer

Due to the co-existing prostate-specific antigen (PSA) raising and lowering factors, the commonly used PSA screening test often lacks sufficient sensitivity or specificity to accurately predict prostate cancer (PCa).1 Spermine, a novel non-prostate massage urine biomarker, was recently shown to be capable of predicting prostate cancer (PCa) and high grade PCa (HGPCa) without the need of prior prostate massage.2 In another study, the sequential use of prostate health index (phi), then the multiparametric magnetic resonance imaging (mpMRI) with or without biopsy was shown to be an efficient and effective way to identify PCa patients.3 Together, these two novel strategies represent possible further approaches to better guide the clinical decision-making in the treatment of PCa.

The PSA test is a blood test that is primarily used to screen PCa by measuring blood PSA level, which is a protein produced by both cancerous and noncancerous tissue in the prostate.1 Where a high level of PSA may be indicative of PCa, other factors such as an enlarged or inflamed prostate may also increase blood PSA level. On the other hand, high doses of certain chemotherapy medications or being obese may lower PSA level and lead to misdiagnosis.1 An alternative predictor with better sensitivity or specificity would be required to better diagnose PCa.

Spermine, a novel non-prostate massage urine biomarker, was proposed to be a potential alternative biomarker to predict PCa.2 In a local study, 556 consecutive men from 2 hospitals were prospectively recruited before prostate biopsy. The first 30mL of urine without prior prostatic massage was saved immediately before a 10 core systematic prostate biopsy, and the performance of spermine in predicting PCa and HGPCa, defined as Gleason score of 7 or above, was evaluated.2 From the study, normalized urine spermine level was found to be inversely associated with the risk of PCa.2 In PCa patients, the mean spermine level was 2.27 versus 7.8 in non-PCa patients (p=0.003).2 Multivariate analyses further showed that age, PSA, prostate volume, and spermine were all independent predictors of PCa.2 When used to predict PCa, the multivariate spermine score was found to have negative predictive values of 91.3% and 96.7% for PCa and HGPCa, respectively.2 By avoiding 35.4% and 50.7% unnecessary biopsies in PCa and HGPCa prediction at 90% sensitivity, respectively, the multivariate spermine score can provide net clinical benefit over PSA or spermine alone in predicting PCa.2

In addition to PSA test, mpMRI is also commonly used in PCa diagnosis.3 While capable of targeting positive lesions and avoid biopsy in cases with negative imaging results, mpMRI is limited by its costs, accessibility, quality control, sustainability, and the ability to meet the needs of a growing population of aging that only some would harbor lethal diseases.3 In addition, mpMRI is also highly operator-dependent and its negative predictive value often varies between studies depending on the scanner type, sequences selected and experience of radiologist.3

In the PRIM study, 545 men with a median age of 66 years, PSA level of 8.0ng/ml, and phi score of 44 were recruited. Test statistics for PSA, PSA density (PSAd), and phi were assessed for detecting significant cancers using 2 definitions of ≥Grade Group (GG) 2 and ≥Cambridge Prognostic Groups (CPG) 3.3 Overall, mpMRI has detected 64% of any cancer, 47% of cancer with ≥GG2 and 32% of cancer with ≥CPG3.3 The overall area under the curve (AUC) for predicting ≥GG2 cancers was 0.70 for PSA and 0.82 for phi. For ≥CPG3 cancers, the AUC was found to be 0.81 and 0.87 for PSA and phi, respectively. The AUC values for phi did not differ between centres, suggesting that phi test is reliable across different diagnostic settings.3 Although other strategies such as mpMRI with biopsy or using PSAd as threshold can detect more PCa cases, the phi-based triage pathway had the lowest mean cost with the decision curve analysis indicating a net clinical benefit over other existing strategies.3 However, when used in the re-biopsy setting, a phi ≥35 threshold produced a rather poor sensitivity, suggesting that useful phi thresholds may differ depending on the detection context.3,4 Nonetheless, the phi test can refine and reduce the frequency of both mpMRI and biopsy when investigating suspected PCa cases without compromising the ability to detect significant PCa patients.3

To conclude, the multivariate spermine risk score was demonstrated to better predict PCa and HGPCa when compared to conventional PSA tests with the added benefit of not needing prior prostate massage.2 On the other hand, phi testing costs less, has a better predictive value when compared to PSA, and could reduce mpMRI or biopsy use without compromising the detection of prognostically important PCa cases.3 Together, these two novel strategies may improve current PSA testing strategies and better guide the future clinical decision-making in the treatment of PCa.

  1. Mayo Clinic. PSA test. Accessed 30 September 2020.
  2. Chiu PK et al. The novel urine Spermine test predicts prostate cancer: The first prospective evaluation. Eur Urol Open Sci. 2020;19(Suppl 2):e521-e522.
  3. Kim L et al. Clinical utility and cost modelling of the phi test to triage referrals into image-based diagnostic services for suspected prostate cancer: the PRIM (Phi to RefIne Mri) study. BMC Medicine. 2020;18:95:1-9.
  4. Mayo Clinic Laboratories. Prostate Health Index (phi), Serum. Accessed 26 September 2020.
Related Articles