X. Filella^1,*, L. Foj², R. Wijngaard¹, P. Luque³

Présenté par : Xavier FILLELA, Barcelone – Espagne

 

¹Department Of Biochemistry And Molecular Genetics, Hospital Clínic – Barcelona (Spain), ²Department Of Clinical Analysis, Hospital Universitari Arnau De Vilanova – Lleida (Spain), ³Department Of Urology, Hospital Clínic – Barcelona (Spain)

 

Contextualisation

Prostate Specific Antigen (PSA) is widely used to detect prostate cancer (PCa). However, the low positive predictive value of PSA results in large numbers of unnecessary prostate biopsies. Prostate Health Index (PHI) has been proposed to improve results obtained with PSA. PHI is a blood test combining total PSA (tPSA), free PSA (fPSA) and p2PSA.

 

Objectifs

To evaluate the value of PHI and PHI density (PHID) for the detection of PCa.

 

Méthodes

We measured tPSA, fPSA and p2PSA in 455 patients scheduled for biopsy, including 243 patients with PCa. The percentage of fPSA (%fPSA) was calculated as fPSA/tPSA and multiplying the result by 100. On the other hand, PHID and PSA density (PSAD) were calculated as tPSA/prostate volume and PHI/prostate volume, respectively.

D’Amico criteria (based on Gleason score, tPSA and clinical stage at diagnosis) were used to classify PCa patients in three groups related to risk of progression: low-risk (132 patients), intermediate-risk (79 patients), and high-risk (32 patients). Intermediate- and high-risk PCa were considered as aggressive PCa.

Prostate volume was assessed by transrectal ultrasound or magnetic resonance imaging, applying the ellipsoid formula. In 89 cases the prostate volume measured <36 cc, in 97 between 36 and 50 cc and in 269 it was >50 cc. Estimation of the prostate size by digital rectal exam was available for 406 patients, whereby patients were classified into three groups: small (n: 51), medium (n: 203) and large prostate (n: 152).

The predictive performance for detecting the presence of intermediate- or high-risk PCa of the different tests was evaluated by the area under the curve (AUC) of the receiver operating characteristic curve, assessing the statistical differences between the AUCs with the DeLong method.

 

Résultats obtenus ou attendus

The best AUC value obtained in the detection of aggressive PCa was achieved for PHI (0.766; 95% CI: 0.725-0.804) and PHID (0.760; 95% CI: 0.718-0.798), while the AUCs for the other tests were: 0.653 (95% CI: 0.607-0.696) for %fPSA; 0.746 (95% CI: 0.703-0.785) for PSAD; and 0,637 (95% CI: 0.591-0.681) for PSA. The AUC obtained for PHI was significantly higher than that observed for tPSA and %fPSA.

Furthermore, we found a relationship of the performance of by these tests with the calculated prostate volume or the estimated prostate size by digital rectal exam, obtaining the higher AUC in patients with a small prostate. Thus, the AUC for PHI was 0,843 for patients with small calculated prostate volume and 0,817 for patients with small estimated prostate size.

 

Discussion

Our results underline that PHI and PHID outperforms the efficacy obtained with tPSA and %fPSA in the detection of intermediate- and high-risk PCa. Substantial differences in their value in relation to prostate volume were found.

 

Conclusion et perspectives

Our results suggest that PHI is a useful an easy test to avoid biopsies in patients with suspected PCa, particularly in patients with small or medium prostate volume.