Medizinische Universität Graz Austria/Österreich - Forschungsportal - Medical University of Graz
Gewählte Publikation:
Mischinger, J.
Real-world interobserver agreement and risk stratification of PI-RADS v2 after combining 4-core targeted MRI-TRUS fusion and systematic 12-core TRUS prostate biopsy for the diagnosis of prostate cancer
Doktoratsstudium der Medizinischen Wissenschaft; Humanmedizin; [ Dissertation ] Medizinische Universität Graz; 2023. pp. 109
[OPEN ACCESS]
FullText
- Autor*innen der Med Uni Graz:
- Betreuer*innen:
- Augustin Herbert
- Pummer Karl
- Speicher Michael
- Altmetrics:
- Abstract:
- Purpose: The necessity of multiparametric (mp) magnetic resonance imaging (MRI) unblinded re-read in unbiopsied (UB) versus former negative prostate biopsy (FNB) patients on the interobserver variability of the Prostate Imaging Reporting and Data System version 2 (PI-RADS) remains unclear. In FNB cohorts with a suspicion of prostate cancer (PCa), the ‘mpMRI/TRUS-fusion lesion directed prostate biopsy’ (LDB) alone strategy for PI-RADS ≥3 lesions is frequently recommended, lacking evidence of the direct comparison of biopsy performances in UB versus FNB patients or assessment of the biopsy to radical prostatectomy specimen (RPS) correlation as ideal reference standard. Materials and Methods: PI-RADS reports and mpMRI-images generated in 15 qualified radiology institutes were intramurally reviewed for assessment of agreement and biopsy planning. We investigated the corresponding UB- and FNB PCa detection rates of the PI-RADS score substratified by the LDB, conventional systematic transrectal ultrasound guided prostate biopsy (CTB), or combination method (LDB&CTB). In addition, we evaluated the RPS to biopsy core histopathological PCa grade variability. The PSA density (PSAD) cut-off >0.15 ng/ml/cm3 was incorporated. Results: We observed UB- vs. FNB-PI-RADS 2-5 agreement, up- and downgrading in 83% vs. 71% (p<0.0001), 9% vs. 16% (p=0.006) and 8% vs. 13% (p=0.01), respectively. The inter-rater-coefficient was substantial for UB-patients (0.73; 95% CI: 0.68-0.78) and moderate for FNB men (0.57; 95% CI: 0.49-0.64). After comparing UB (n=499) vs. FNB (n=314) individuals significant (cs) PCa was discovered more frequently in UB patients with the LDB&CTB technique (62 vs. 43%) than with the LDB (54 vs. 34%) or CTB (57 vs. 34%) alone (all p<0.0001). We observed no difference for the rate of csPCa found by LDB and LDB&CTB according to reevaluated UB-PI-RADS-5 patients with a PSAD>0.15 ng/ml/cm3 (n=79 vs. n=81; p=0.2). When diagnosing csPCa, FNB patients had a higher rate of false-negative results using the LDB method (UB 11% vs. FNB 19%; p=0.02). The LDB&CTB approach showed a greater “biopsy-RPS” agreement in the entire and UB patient cohort but generally demonstrated substantially less upgrading. Conclusions: The PIRADS interobserver agreement is more accurate in UB situations. Both in PI-RADS 3, 4 or 5 lesions of UB and FNB males, the LDB&CTB method appears to discover a considerably larger number of csPCa. In addition, reevaluated UB-PI-RADS 5 patients with a PSAD >0.15 ng/ml/cm3 may be advised to omit a concurrent CTB without compromising the csPCa diagnostic accuracy. The exclusion of a concomitant CTB in FNB-men produces significantly more false negative targets among those patients who were csPCa positive. In the total- and UB patient cohorts, the LDB&CTB approach demonstrated the best RPS-grade prediction and consistently displayed the lowest upgrading rates. In general the LDB&CTB process should also be suggested to clarify the csPCa suspicion in FNB patients.