Confirmation by Early Oncologic Outcomes After Surgery of the Accuracy of Intermediate-risk Prostate Cancer Classification Based on Magnetic Resonance Imaging Staging and Targeted Biopsy.

Over the past decade, prostate cancer (PCa) diagnosis drastically evolved from systematic biopsies (SBs) to multiparametric magnetic resonance imaging (mpMRI) and targeted biopsy (TB), which have emerged as powerful imaging tools for diagnosis, staging, and preoperative planning. MRI and TB should now be widely adopted for assessing prognosis and be incorporated into predictive models. To date, the standard intermediate risk classification (IRC) defined unfavourable and favourable disease with clinical information and overall biopsy data. Roumiguie et al have proposed a new model based on mpMRI staging and grade group on TB and validated it using radical prostatectomy (RP) pathology (Urol Oncol 2020;38:386-92). The aim of our study was to validate the accuracy of this new IRC with early oncologic outcomes and biochemical recurrence (BCR) after RP. From a prospective database of RP patients with positive prebiopsy mpMRI (Prostate Imaging-Reporting and Data System score ≥3) followed by SB in combination with TB, 454 patients with intermediate-risk PCa were included. Median follow-up was 31.5 mo. The new IRC outperformed the standard IRC in predicting BCR (p =  0.007). The area under the curve was 0.613 for the new MRI- and TB-based IRC versus 0.575 for the standard IRC. This new IRC could optimise the prediction of recurrence risk before treatment decision-making. PATIENT
SUMMARY: Outcomes after surgery confirm the accuracy of the new classification of intermediate-risk prostate cancer based on magnetic resonance imaging (MRI) staging and targeted biopsy data. We found that this new classification outperformed the standard classification in predicting biochemical recurrence of cancer for men with positive MRI findings undergoing targeted biopsies.

It has been demonstrated that magnetic resonance imaging (MRI) is a valuable tool in improving the diagnosis of clinically significant prostate cancer (PCa) [1]. MRI-targeted biopsy (TB) improves PCa diagnosis by increasing the detection of significant PCa and decreasing the detection of insignificant PCa [2]. Localised PCa is a heterogeneous entity with different prognoses. The goal is to differentiate patients with poor prognosis from those with good prognosis to reinforce therapy in aggressive cases and to avoid unnecessary treatments in favourable ones. Efforts are being made to update the current predictive tools by incorporating MRI data and TB pathology findings [3].

The classification proposed by d’Amico et al [4] based on digital rectal examination (DRE), Gleason score and prostate-specific antigen (PSA) stratifies men into low-, intermediate-, and high-risk groups. However, intermediate-risk cases are a heterogeneous group, leading to subclassification between favourable and unfavourable cases. This validated intermediate risk classification (IRC) was introduced by Zumsteg et al [5] and is described in the National Comprehensive Cancer Network guidelines, with unfavourable disease defined as patients who have at least one of the following criteria: two or more intermediate risk factors, grade group (GG) ≥3, or more than half of biopsy cores showing cancer. Nevertheless, MRI and TB features are not taken into account, which could lead to inaccurate assessment for patients with positive MRI findings.

Roumiguie et al [6] recently used MRI criteria and TB data to improve this IRC. The favourable group was defined as men with the absence of extracapsular extension (ECE) on MRI and GG < 3 on TB; all other cases (any ECE on MRI and/or GG ≥ 3 on TB) were classified as unfavourable intermediate-risk disease. This new IRC significantly improved the prediction of final pathology and could reduce the risk of overtreatment for misclassified unfavourable intermediate-risk cases. The aim of our study was to validate the accuracy of this new IRC with early oncologic outcomes and biochemical recurrence (BCR) after RP.

We selected from our prospective database consecutive patients between January 2014 and July 2019 with intermediate-risk PCa who had positive prebiopsy mpMRI findings (Prostate Imaging-Reporting and Data System score ≥3) followed by systematic biopsy (SB) in combination with TB. SBs were performed randomly, and the number of targeted cores taken for each suspicious lesion on mpMRI was chosen at the physician’s discretion, with a median of four TBs per lesion. Overall, 454 patients with intermediate-risk PCa were included who had one or more intermediate risk factors (PSA 10–20 ng/ml, GG 2–3, stage T2b/T2c on DRE) and no high risk factor (GG 4–5, PSA > 2 ng/mL, or stage ≥ T3 on DRE). Biochemical follow-up was standardised with a PSA test at 6 wk, 3 mo, 6 mo, and then every 6 mo after surgery. No patient received any neoadjuvant or adjuvant treatment without BCR.

BCR was defined as any confirmed PSA value >0.2 ng/mL. The median follow-up was 31.5 mo. Overall, 7.9% of patients experienced BCR. The patient characteristics and pathology results are presented in Supplementary Table 1. The mean patient age was 64.7 yr. Mean PSA was 8.4 ng/mL (median 8) and mean PSA density was 0.19 ng/mL/g (median 0.18). ECE on MRI was reported for 11.7% of the patients and GG ≥ 3 on TB for 22.4%. ECE on MRI was confirmed on radical prostatectomy for 61.1%. Baseline characteristics and pathology outcomes for the favourable and unfavourable subgroups according to the standard IRC and new IRC are presented in Table 1. The standard IRC classed 35.9% of patients in the favourable group and 64.1% in the unfavourable group, whereas the corresponding proportions according to the new IRC were 70.1% and 29.1%. Some 48.8% of patients in the unfavourable group according to the standard IRC were misclassified and had favourable intermediate-risk disease (without ECE or lymph node invasion and with GG ≤ 2), whereas only 32.6% were misclassified according to the new IRC.

Table 1

Baseline characteristics and pathology outcomes for the favourable and unfavourable subgroups according to the standard IRC and new IRC.

	Standard IRC			New IRC
	Favourable	Unfavourable	p value	Favourable	Unfavourable	p value
Patients (%)	35.9	64.1		70.3	29.7
Clinical stage (%)T1bT1cT2	0.693.26.2	0.751.947.4	<0.001	0.668.530.9	0.862.436.8	0.46
PI-RADS score (%)PIRADS 3PIRADS 4PIRADS 5	36.247.916.0	14.152.633.3	<0.001	28.552.019.4	6.748.145.2	<0.001
3-T MRI (%)	4.9	15.5	0.001	0.0	34.1	<0.001
Biopsy GG for TB + SB (%)123	34.465.60.0	3.158.138.8	<0.001	20.475.54.1	0.025.974.1	<0.001
Biopsy GG on TB (%)01234	19.025.854.60.60.0	8.97.648.834.00.7	<0.001	17.920.162.10.00.0	0.00.024.474.11.5	<0.001
Biopsy GG on SB (%)01234	17.845.436.80.00.0	10.321.651.516.20.3	<0.001	13.834.847.63.40.3	11.119.343.026.70.0	<0.001
pT stage (%)pT2pT3apT3b–4	62.629.48.0	45.439.914.8	0.001	58.332.98.8	35.643.720.7	<0.001
pN stage (%)pN0pN1pNx	71.21.227.6	84.96.28.9	<0.001	77.71.620.7	85.211.13.7	0.001
GG on RP (%)12345	6.771.220.21.20.6	2.752.241.61.02.4	<0.001	5.369.023.50.91.3	1.535.658.51.53.0	<0.001
Positive margin (%)	19.0	18.2	0.9	16.9	22.2	0.18
Biochemical recurrence (%)	4.9	9.8	0.102	5.7	13.7	0.007

GG = grade group; IRC = intermediate risk classification; MRI = magnetic resonance imaging; PI-RADS = Prostate Imaging-Reporting and Data System; RP = radical prostatectomy; SB = systematic biopsy; TB = targeted biopsy.

Using the new IRC, 5.7% of men in the favourable subgroup and 13.7% in the unfavourable subgroup had BCR, while the corresponding rates using the standard IRC were 4.9% and 9.7%.

Survival curves assessed using the Kaplan-Meier method and the log-rank test showed a significant difference (p =  0.004) between the favourable and unfavourable groups according to the new IRC (Fig. 1). The area under the receiver operating characteristic curve was 0.613 for the new IRC versus 0.575 for the standard IRC (Supplementary Fig. 1). Importantly, the new IRC reclassified half of unfavourable cases in the standard IRC as favourable intermediate-risk disease, and these reclassified patients could avoid unnecessary aggressive treatment.

Fig. 1

Kaplan-Meier estimate of biochemical recurrence (BCR)-free survival in favourable and unfavourable subgroups according to (A) the standard (log rank p =  0.049) and (B) the new (log rank p =  0.004) intermediate risk classification.

mpMRI with TB improves the accuracy of models based on clinical data in predicting adverse pathology and should be incorporated into routine practice and novel models [7], [8]. Our study revealed that a predictive model that includes ECE on MRI and TB improved BCR prediction, in addition to final pathology prediction, in comparison to the standard IRC. In this population of patients, BCR has been linked to more advanced disease and higher rates of metastasis and prostate cancer–specific mortality [9].

Our study confirms that standard IRC is not optimal and that new tools such as mpMRI and TB should be used to improve the evaluation of intermediate-risk PCa. The new IRC could be improved but remains an easy-to-use model and allows for reclassification of 56% of unfavourable cases misclassified in the standard IRC (Supplementary Table 2). This represents a clinically relevant improvement, especially in reducing overtreatment. The rate of BCR in intermediate-risk PCa remains low at only 7.9% in our study, so a larger prospective study is needed to confirm our findings.

In the heterogeneous population of men with MRI-positive intermediate-risk PCa, the addition of MRI and TB features to the IRC significantly improved the prediction of final pathology and of early oncologic outcomes after RP. The use of this new IRC can optimise treatment decision-making by providing more accurate discrimination of favourable and unfavourable risk groups. This could result in a decrease in overtreatment (neoadjuvant therapy or long-term androgen deprivation associated with radiation).

: Cécile Manceau had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Ploussard, Manceau, Malavaud, Roumiguié.

Acquisition of data: Ploussard, Manceau, Beauval, Lesourd, Almeras, Gautier, Loison, Salin, Soulié, Tollon, Malavaud, Roumiguié.

Analysis and interpretation of data: Ploussard, Manceau.

Drafting of the manuscript: Ploussard, Manceau.

Critical revision of the manuscript for important intellectual content: Ploussard, Manceau.

Statistical analysis: Ploussard, Manceau.

Obtaining funding: None.

Administrative, technical, or material support: Ploussard, Manceau.

Supervision: Ploussard.

Other: None.

Cécile Manceau certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.

None.