Table 1
Histopathological findings and clinical follow-up data.
| NC | LG | IG | HG | T1 | T2 | T3 | T4 | No | Yes | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Radical prostatectomy | 68 | 1 | 5 | 33 | 29 | 0 | 44 | 23 | 0 | 3 | 65 |
| Radiation therapy | 91 | 0 | 18 | 26 | 47 | 28 | 40 | 20 | 3 | 14 | 77 |
| Active surveillance > 2 years | 13 | 0 | 11 | 2 | 0 | 13 | 0 | 0 | 0 | 11 | 2 |
| >2 years of cancer free follow-up | 73 | 73 | / | / | / | / | / | / | / | 73 | 0 |
| Total | 245 | 74 | 34 | 61 | 76 | 41 | 84 | 43 | 3 | 101 | 144 |
[i] NC: No prostate cancer.
LG: Low Grade Prostate Cancer, defined as prostate cancer with Gleason score 3+3 or lower.
IG: Intermediate Grade Prostate Cancer, defined as prostate cancer with Gleason score 3+4.
HG: High Grade Prostate Cancer, defined as any prostate cancer with primary Gleason grade 4 or any Gleason grade 5 (including tertiary patterns).
ClinsigPC: Gleason score ≥7 (including 3+4 with prominent but not predominant Gleason 4 component), and/or volume ≥0.5cc, and/or extraprostatic extension.
Table 2
Overall assessment scores and probability of clinically significant prostate cancer.
| PI-RADSv1 | PI-RADsv2 | P-value | PI-RADSv1 | PI-RADSv2 | P-value | |
|---|---|---|---|---|---|---|
| 1 | 4.9% (12/245) | 4.9% (12/245) | NS | 33.3% (4/12) | 50.0% (6/12) | NS |
| 2 | 28.6% (70/245) | 35.1% (86/245) | 0.02 | 18.6% (13/70) | 27.9% (24/86) | NS |
| 3 | 17.6% (43/245) | 10.2% (25/245) | 0.005 | 58.1% (25/43) | 40.0% (10/25) | NS |
| 4 | 27.8% (68/245) | 26.5% (65/245) | NS | 80.9% (55/68) | 78.5% (51/65) | NS |
| 5 | 21.2% (52/245) | 23.3% (57/245) | NS | 90.4% (47/52) | 93.0% (53/57) | NS |
Figure 1
ROC curves of the overall assessment scores of PI-RADSv1 (green line, AUC 0.82 [95% CI 0.76–0.87]) and PI-RADSv2 (blue line, AUC 0.79 [CI 0.73–0.85]) using the presence of ClinsigPC on a per-patient basis as the reference.
Table 3
AUC values of the ROC analysis for the overall assessment score and the different individual modalities.
| PI-RADSv1 | PI-RADSv2 | ||
|---|---|---|---|
| Overall assessment score | 0.82 (0.76-0.87) | 0.79 (0.73-0.85) | NS |
| T2-WI in PZ | 0.78 (0.72-0.84) | 0.78 (0.72-0.84) | NS |
| T2-WI in TZ | 0.52 (0.45-0.60) | 0.53 (0.45-0.60) | NS |
| DWI | 0.82 (0.77-0.88) | 0.81 (0.76-0.87) | NS |
| DCE | 0.81 (0.75-0.87) | 0.76 (0.70-0.83) | NS |
| MRSI | 0,65 (0.58-0.72) | / | / |
Table 4
Discrepancies between PI-RADSv1 and PI-RADSv2.
| N | PI-RADSv1 | PI-RADSv2 | ||||
|---|---|---|---|---|---|---|
| Missed ClinsigPC on PI-RADSv1 but true positive on PI-RADSv2 | 3,3% (1/30) | 100% (1/1) | 2 | 4 | The ClinsigPC was a large Gleason 3+3 PC | T2-WI and DWI suspicious, but DCE and MRSI normal |
| True negative on PI-RADSv1 but false positive on PI-RADSv2 | 20,0% (6/30) | 83,3% (5/6) | 2 | 3 | No ClinsigPC: no cancer in 4 cases, but a small Gleason 3+3 PC in 1 case | In all cases T2-WI and DWI suspicious, but DCE and MRSI normal |
| 16,7% (1/6) | 2 | 4 | No ClinsigPC although there was a very small Gleason 3+4 PC | |||
| False positive on PI-RADSv1 but true negative on PI-RADSv2 | 30,0% (9/30) | 11,1% (1/9) | 3 | 1 | No cancer | In 5 cases DCE suspicious but all other modalities normal and in 4 cases T2-WI suspicious but DWI normal in PZ |
| 77,8% (7/9) | 3 | 2 | No ClinsigPC although there was 1 case of small Gleason 3+3 PC, 2 cases of HGPIN & BCH and 1 case of ASAP | |||
| 11,1% (1/9) | 4 | 2 | No cancer | |||
| True positive on PI-RADsv1 but missed ClinsigPC on PI-RADSv2 | 46,7% (14/30) | 7,1% (1/14) | 3 | 1 | The ClinsigPC was a Gleason 4+3 PC | In 3 cases DCE suspicious but all other modalities normal, in 10 cases T2-WI suspicious but DWI normal in PZ and in 1 case only DWI suspicious in TZ |
| 78,6% (11/14) | 3 | 2 | The ClinsigPC were 1 Gleason 3+3 pT2c, 6 Gleason 3+4 PC and 4 Gleason 4+3 PC or higher | |||
| 7,1% (1/14) | 4 | 2 | The ClinsigPC was a Gleason 3+4 PC | |||
| 7,1% (1/14) | 5 | 2 | The ClinsigPC was a Gleason 3+4 PC |
[i] 1After dichotomization with considering an overall assessment score 1 and 2 as ‘negative’ and 3, 4 or 5 as ‘positive’ mpMRI.
Figure 2
A 63-year-old man with a PSA of 6.3 μg/L. On T2-WI (a) in the PZ an ill-defined low signal intensity focus is present posteriorly on the left side in the prostate base (white oval) (PI-RADSv1 score 4; PI-RADSv2 score 4). On DWI the signal intensity on ADC is high (b) and on high-b-value image isointense/mildly hyperintense (c) (PI-RADSv1 score 1; PI-RADSv2 score 1). On DCE (4) the area shows strong contrast enhancement (white oval) (PI-RADSv1 score 4; PI-RADSv2 negative). With the PI-RADSv1 scoring system this patient was assigned an overall assessment score 4, based on a subjective impression of the findings on T2-WI and DCE. With the PI-RADSv2 scoring system, this patient was assigned an overall assessment score of 1 since DWI is the dominant modality in the PZ. A prostate biopsy followed by radical prostatectomy in this patient showed however a Gleason 3+4 prostate cancer on the left side in the PZ of the prostate base.
Figure 3
A 66-year-old man with PSA 9.8μg/L. On T2-WI (a) the PZ shows diffuse mild hypointensity, with indistinct margin (PI-RADSv1 score3; PI-RADSv2 score 3). On DWI the PZ shows high ADC values (b) and is isointense on high-b-value image (c) (PI-RADSv1 score 1; PI-RADSv2 score 1). On DCE there is a focal enhancing lesion posterolateral on the right side in the PZ (black star) (PI-RADSv1 score 4; PI-RADSv2 positive). With the PI-RADSv1 scoring system, this patient was assigned an overall assessment score 4, based on a subjective overall impression of the findings in all modalities. With the PI-RADSv2 scoring system this patient was assigned an overall assessment score 1 because DWI was scored 1 and this is the dominant modality in PZ. Prostate biopsy followed by radical prostatectomy in this patient showed a Gleason 3+4 prostate cancer on the right side in the PZ.
Figure 4
A 56-year-old man with PSA of 11.2μg/L. On T2-WI (a) the PZ shows a homogenous high signal intensity (PI-RADSv1 score 1; PI-RADSv2 score 1) and in the TZ a well-margined homogenous hypointense area is noted (PI-RADSv1 score 2; PI-RADSv2 score 2). On DWI this area in the TZ shows low ADC value (b) (white star) and is isointense on high-b-value image (c) (PI-RADSv1 score 4; PI-RADSv2 score 4). On DCE this area shows strong contrast enhancement (black star) (PI-RADSv1 score 4; PI-RADSv2 positive). With the PI-RADSv1 scoring system this patient was assigned an overall assessment score 4, based on a subjective overall impression of the findings in all modalities. With the PI-RADSv2 scoring system, T2-WI in the TZ was scored 2, and since T2-WI is the dominant modality in the TZ the overall assessment score was also 2. Prostate biopsy followed by radical prostatectomy in this patient showed a Gleason 3+4 prostate cancer in the TZ.
Figure 5
A 63-year-old man with PSA 8.35 μg/L. In the PZ on the left side there is a low signal intensity focus on T2-WI (PI-RADSv1 score 5; PI-RADSv2 score 4) (a) (white star) with restricted diffusion (markedly hypointense on ADC (b) and hyperintense on high-b-value (c)) (PI-RADSv1 score 5, PI-RADSv2 score 4). The size of the lesion is <1,5 cm, it has no broad contact with the prostate capsule and shows no definite extraprostatic extension. On DCE (d) it shows strong contrast enhancement (black star) (PI-RADSv1 score 5; PI-RADSv2 positive). The overall assessment score is 5 when using PI-RADSv1 based on a subjective impression of the findings of all modalities and is 4 in PI-RADSv2 based on the DWI-only which should be given a score 4. Despite the different overall assessment category between PI-RADSv1 and PI-RADSv2 the message to the clinician is the same, i.e. targeted prostate biopsy is warranted. Prostate biopsy followed by radical prostatectomy confirmed a PC with Gleason score 3+4 = 7 (with tertiary pattern 5).