Magnetic resonance imaging (MRI) is increasingly used for local staging in prostate cancer, yet its value relative to digital rectal examination (DRE) staging remains uncertain.
To compare the performance of MRI-based vs DRE-based clinical T staging in men undergoing radical prostatectomy (RP) and to assess the outcome of substituting MRI staging in established prognostic classification systems.
This multicenter, retrospective case-control study (iTNM Consortium) was conducted from January 1, 2015, to December 31, 2021, among consecutive men 18 years or older with clinically localized or locally advanced prostate cancer who underwent RP and preoperative multiparametric MRI and systematic biopsies with or without targeted biopsies at 31 referral centers across 6 European countries. Patients receiving prior prostate cancer treatment or missing key variables were excluded. Data analysis was performed from August to October 2025.
MRI-derived T stage (iT) vs DRE-derived T stage (cT) incorporated into D’Amico, European Association of Urology, National Comprehensive Cancer Network, and International Staging Collaboration for Prostate Cancer risk groups to generate MRI-adapted classification systems.
Primary outcome was distant metastasis–free survival. Secondary outcomes were biochemical recurrence–free survival and overall survival. Discrimination was assessed using a C index and time-dependent area under the receiver operating characteristic curves.
A total of 4425 men were included (median [IQR] age, 66 [61-70] years; median [IQR] prostate-specific antigen, 7.4 [5.5-10.5] ng/mL). Median (95% CI) follow-up was 52 (51-53) months. MRI-based staging showed slightly higher discriminatory ability than DRE staging for biochemical recurrence–free survival (C index, 0.62 [95% CI, 0.61-0.64] vs 0.59 [95% CI, 0.57-0.61]) and distant metastasis–free survival (0.67 [95% CI, 0.64-0.70] vs 0.65 [95% CI, 0.62-0.68]). However, across all 4 risk classification systems, MRI-based and DRE-based versions demonstrated comparable discrimination with overlapping 95% CIs. Time-dependent area under the receiver operating characteristic curves showed no significant difference during follow-up. Subgroup analyses of patients with cT1 disease and those with PSA levels less than 10 ng/mL yielded similar results.
In this case-control study of patients undergoing radical prostatectomy, MRI-based staging demonstrated discrimination similar to traditional DRE-based staging for biochemical recurrence, distant metastasis, and overall survival. These findings suggested that MRI-derived staging could be safely integrated into contemporary risk systems without compromising accuracy.
Commentary by Dr. Jennifer Le Guévelou
MRI versus digital rectal examination: Should modern imaging redefine localised prostate cancer risk stratification?
Accurate clinical staging of localised prostate cancer (PCa) is essential for reliable risk stratification, informed treatment selection, and meaningful prognostic counselling. For several decades, digital rectal examination has served as the universal cornerstone of clinical local tumour staging, forming the foundational basis of widely utilised prognostic classification systems including those of D’Amico, the European Association of Urology, the National Comprehensive Cancer Network, and the International Staging Collaboration for Prostate Cancer. Nevertheless, digital rectal examination is inherently limited by subjective interpretation, significant inter-examiner variability, and a very restricted capacity to detect subtle extracapsular extension or seminal vesicle invasion.
In parallel, the widespread clinical integration of multiparametric magnetic resonance imaging (mpMRI) has profoundly transformed the modern PCa diagnostic pathway by improving tumour detection and enabling targeted biopsies. Because MRI provides detailed anatomical resolution of local tumour extent, it has increasingly supplanted digital rectal examination in routine clinical practice for local staging prior to radical treatment. However, a critical clinical dilemma remained unresolved, namely whether replacing clinical examination staging with MRI staging genuinely improves the estimation of oncologic outcomes or merely induces stage migration without adding real prognostic discrimination.
Material and methods
To definitively address this question, the international iTNM Consortium conducted a rigorous retrospective multicentre study encompassing 4,425 consecutive adult patients who underwent radical prostatectomy between January 2015 and December 2021 across 31 tertiary academic referral centres in six European countries, including France, Belgium, Denmark, Germany, Italy, and Switzerland. All included patients presented with histologically confirmed, clinically localised or locally advanced PCa and underwent standardised preoperative mpMRI on 1.5-Tesla or 3.0-Tesla scanners reported according to the Prostate Imaging Reporting and Data System, systematic prostate biopsies with or without targeted cores, and baseline digital rectal examination.
Clinical local staging determined by digital rectal examination according to the eighth edition of the TNM classification was directly compared with MRI -derived staging criteria. MRI-derived categories were subsequently incorporated into the four major risk classification systems to generate MRI-adapted models, while holding all other clinicobiological variables identical. The primary study outcome was distant metastasis-free survival (MFS), defined as the time from radical prostatectomy to radiologically confirmed distant metastasis. Secondary study outcomes included biochemical recurrence-free survival, defined as the occurrence of two consecutive post-prostatectomy prostate-specific antigen values exceeding 0.2 ng/mL, and overall survival.
Study results
Among the 4,425 men included in the analysis, the median age at the time of radical surgery was 66 years, the median preoperative prostate-specific antigen level was 7.4 nanograms per millilitre, and the median post-surgical follow-up was 52 months. Preoperative MRI reclassified a substantially higher proportion of patients into locally advanced tumour stages compared with digital rectal examination, identifying radiological extraprostatic extension in 17.0% of cases compared to only 3.3% detected on clinical examination, while final pathological examination revealed extraprostatic extension in 34.0 percent of radical prostatectomy specimens. Over the observation period, 1,111 patients experienced biochemical disease recurrence, and 356 patients developed distant metastatic progression.
When evaluating standalone local tumour staging, magnetic resonance imaging yielded a marginally higher concordance index than digital rectal examination for biochemical recurrence-free survival at 0.62 compared with 0.59, and for distant MFS at 0.67 compared with 0.65. Crucially, when incorporated into established multivariable risk systems, MRI-adapted models demonstrated prognostic discrimination that was virtually identical to traditional clinical examination-based systems, with extensively overlapping 95 percent confidence intervals across all four evaluated frameworks. Specifically, for biochemical recurrence-free survival, concordance indices were 0.63 versus 0.64 in the D’Amico classification, 0.65 versus 0.66 in the EAU classification, and 0.68 versus 0.68 in the STAR-CAP classification. Similarly, for distant MFS, concordance indices reached 0.70 versus 0.74 in the EAU model and 0.75 versus 0.76 in the STAR-CAP model. Furthermore, longitudinal time-dependent area under the curve analyses demonstrated completely parallel trajectories without significant statistical divergence at any point during follow-up, and subgroup analyses in patients with favourable baseline characteristics confirmed equivalent performance.
What does the study add?
The findings established by the iTNM Consortium provide definitive empirical evidence that reshapes our understanding of clinical staging in localised PCa. The primary insight emerging from this investigation is that anatomical resolution must be conceptually decoupled from biological aggressiveness. Although mpMRI clearly surpasses digital rectal examination in detecting microscopic capsular abutment and architectural disruption, these morphological features (and especially the presence of an extracapsular extension) do not automatically translate into a heightened metastatic potential. The core drivers of systemic disease progression remain intrinsic tumour grade and prostate-specific antigen kinetics, which cannot be superseded by morphological imaging alone.
An equally compelling explanation for the comparable predictive accuracy observed between the two modalities lies in the clinical definition of organ-confined disease. A palpable tumour detected on digital rectal examination typically represents a dense, bulky lesion that enriches for a biologically higher-risk subset of disease, whereas MRI captures an expansive, highly heterogeneous spectrum of small, anterior, and indolent lesions. This clinical selection bias inherent to physical palpation effectively offsets its anatomical insensitivity when evaluated across broad patient populations.
From a practical clinical perspective, these results provide strong and reassuring justification for the ongoing evolution of urological practice. Because digital rectal examination is increasingly recognised as having minimal value for population screening and is frequently perceived as uncomfortable by patients, substituting MRI-derived staging into contemporary risk algorithms can be performed safely without sacrificing prognostic accuracy. At the same time, MRI retains its indispensable and unique value for surgical planning, particularly in guiding personalised nerve-sparing decisions during radical prostatectomy.
Finally, the modest discriminative performance observed across all evaluated models underscores the inherent ceiling of staging systems that rely exclusively on traditional clinicoradiological variables. To achieve a true paradigm shift in personalised oncologic counselling, future risk stratification tools must move beyond anatomical tumour stage and integrate tissue-based genomic signatures, quantitative radiomics, and artificial intelligence-driven multimodal algorithms into routine clinical decision-making. While this study is retrospective and therefore might be associated with limitations regarding mpMRI interpretations and digital rectal examination findings, these findings require confirmation with data coming from prospective cohorts with imaging centrally reviewed.