MRI Tumor Stage Emerges as Powerful Predictor of Nodal Spread on PSMA PET in High-Risk Prostate Cancer

…

…

The distinction between T3a and T3b disease sits at the heart of the study’s most striking finding, and it deserves careful unpacking for readers less familiar with prostate cancer staging conventions. In the TNM system, T3a denotes tumour that has breached the prostatic capsule — the fibrous sheath surrounding the gland — while T3b signifies that the tumour has extended further, invading the seminal vesicles, the paired structures that sit immediately behind and above the prostate and contribute fluid to semen. Seminal vesicle invasion has long been recognised as a marker of more aggressive tumour biology, and the present study quantifies that clinical intuition in a modern imaging context: when MRI showed unequivocal contiguous invasion of the seminal vesicles, the odds of finding PSMA-avid nodal disease rose dramatically, whereas mere capsular breach without seminal vesicle involvement carried a weaker, statistically non-significant signal after adjustment for other variables. This gradient — T2 disease carrying very low risk, T3a occupying an intermediate and ambiguous zone, and T3b and T4 disease carrying substantial risk — mirrors the biological logic of tumour progression, in which each additional step of local invasion reflects greater tumour volume, higher grade potential, and correspondingly greater likelihood of lymphatic spread.

The technical rigour applied to MRI staging in this study merits attention, because MRI-derived T stage has historically been criticised for interobserver variability. The investigators required that ambiguous or borderline findings of extracapsular extension and seminal vesicle invasion be resolved by consensus among three subspecialty-trained radiologists, with tumours defaulting to T2 classification unless definite extension was agreed. This conservative approach likely reduced false upstaging, which is important because overcalling T3 disease could artificially inflate the apparent association between T stage and nodal positivity. The use of a single 3 Tesla scanner with a standardised multiparametric protocol and a 32-channel pelvic phased-array coil also reduced equipment-related heterogeneity, a common confounder in multi-scanner imaging studies. Readers should bear in mind, however, that these safeguards reflect the controlled conditions of a single-centre study, and the strength of the reported associations may be somewhat more modest in settings where MRI quality, scanner field strength, and reporting expertise vary more widely.

The finding that T3b retained independent predictive value within the Gleason 4 + 3 subgroup, with an odds ratio of approximately 23.6, is particularly informative given ongoing debates about how best to categorise intermediate-risk disease. Gleason 4 + 3 tumours — in which the pattern 4 component predominates over pattern 3 — are widely regarded as carrying substantially higher metastatic potential than their 3 + 4 counterparts, and many guidelines now group them with high-risk disease for management purposes. That seminal vesicle invasion on MRI added prognostic information on top of grade in this subgroup suggests that local extent and histological aggressiveness are not redundant markers but complementary dimensions of risk. A tumour that is both architecturally aggressive on biopsy and locally invasive on imaging appears to represent a particularly dangerous phenotype, one in which the probability of occult nodal disease is high enough that PSMA PET/CT findings are likely to materially alter treatment planning, whether by expanding the radiation field, prompting extended pelvic lymph node dissection, or steering patients toward systemic therapy in combination with local treatment.

Equally noteworthy is the zero rate of PSMA-positive nodal disease among patients who simultaneously had PSA below 20 ng/mL, MRI-derived T2 disease or lower, and Gleason score below 8. This triad of favourable features — modest PSA, organ-confined disease on MRI, and biopsy grade below the high-risk threshold — defined a subgroup in which nodal metastases were entirely absent in this cohort of 152 men. Such a finding, if replicated in larger and more diverse populations, could support a more selective approach to PSMA PET/CT referral, reserving the scan for patients whose clinical and imaging profile suggests meaningful pre-test probability of nodal disease. The logic is straightforward: the value of any diagnostic test depends on the prevalence of the target condition in the population tested. When pre-test probability approaches zero, even a highly accurate test contributes little, and the resources consumed — scanner time, radiopharmaceutical production, specialist reporting — may be better directed elsewhere. Conversely, patients with T3b or T4 disease on MRI represent the opposite end of the spectrum, where the high pre-test probability of nodal involvement makes PSMA PET/CT most informative.

The model’s discrimination, expressed as an area under the receiver operating characteristic curve of 0.863, indicates good but not perfect ability to separate patients with and without nodal disease. For context, an AUC of 0.5 corresponds to chance performance, values above 0.8 are generally considered acceptable for clinical prediction models, and values approaching 0.9 or higher are rare in medicine. The additional reporting of calibration plots and decision curve analysis reflects contemporary standards in prediction model research. Calibration plots assess whether predicted probabilities match observed frequencies across the risk spectrum — a model that systematically overestimates or underestimates risk can mislead even if its ranking of patients is accurate. Decision curve analysis goes further by quantifying net benefit across a range of threshold probabilities, effectively asking whether acting on the model’s predictions would lead to better decisions than default strategies such as treating all patients or treating none. The demonstration of clinical net benefit suggests the model could function as a practical triage instrument rather than merely a statistical exercise.

Several limitations inherent to the study design warrant consideration when interpreting these results. The retrospective, single-centre nature of the cohort means that referral patterns, imaging protocols, and reporting practices reflect one institution’s approach, and the 152 patients analysed were drawn from a larger pool of 199 staging scans after exclusions for incomplete records. Selection bias cannot be excluded: patients who underwent both MRI and PSMA PET/CT before treatment may differ systematically from those who did not. Furthermore, the definition of high-risk disease — PSA of at least 20 ng/mL, Gleason score of at least 8, or clinical stage of at least T3 — means the findings apply specifically to this population and cannot be extrapolated to men with lower-risk disease, in whom nodal metastases are rare and PSMA PET/CT yield is correspondingly low. The observation that some patients clinically staged as cT3 on digital rectal examination were downstaged to T2 or lower on MRI highlights the well-documented tendency of digital rectal examination to overestimate local extent, and it underscores why imaging-based staging has largely supplanted clinical examination for T-stage assignment in modern practice.

The biological rationale connecting local tumour extent to nodal spread is worth articulating. Prostate cancer spreads to lymph nodes through predictable lymphatic drainage pathways, and the probability of nodal metastasis correlates with tumour volume, grade, and the extent of local invasion. As tumour grows beyond the capsule and into adjacent structures, it encounters a richer network of lymphatic and vascular channels, increasing the opportunity for tumour cells to disseminate. Seminal vesicle invasion, in particular, has been associated with micrometastatic disease in pelvic nodes across decades of surgical-pathological studies, and the present study extends this association into the era of molecular imaging, where PSMA-targeted radiotracers can detect nodal deposits far smaller than the size thresholds that limited conventional CT and MRI. The convergence of an anatomical marker of local aggressiveness with a molecular imaging marker of metastatic disease provides a coherent picture of tumour behaviour that spans both local and systemic dimensions.

From a health systems perspective, the study’s framing around resource utilisation is timely. PSMA PET/CT has been incorporated into guidelines for the initial staging of high-risk prostate cancer in many jurisdictions, but scanner availability, radiopharmaceutical supply chains, and trained personnel remain constrained in numerous health systems, particularly outside major academic centres. A validated pre-test risk stratification tool built from variables already available at the time of referral — MRI stage, PSA, and biopsy grade — could allow centres to prioritise scans for patients most likely to benefit, while deferring or omitting imaging in patients with very low predicted probability of nodal disease. The authors’ decision to analyse Gleason 3 + 4 and 4 + 3 subgroups separately, rather than collapsing all Gleason 7 disease into a single category, reflects growing recognition that these patterns differ meaningfully in metastatic potential, and it provides clinicians with more granular guidance than a binary intermediate-risk label would allow.

Finally, the findings sit within a broader trajectory in prostate cancer imaging. PSMA-targeted agents have transformed not only staging but also the detection of recurrence, the selection of patients for radioligand therapy, and the planning of salvage treatment after local therapy failure. As molecular imaging becomes more accessible, questions shift from whether to scan toward whom to scan and when. Studies that identify simple, reproducible predictors of scan positivity — such as MRI-derived T stage — contribute to that evolving calculus. Future work will need to validate these associations in external cohorts, examine whether quantitative MRI features such as lesion volume or radiomic signatures add predictive value beyond categorical T stage, and determine whether risk-stratified scanning strategies improve outcomes without delaying the detection of genuinely nodal-positive disease. For now, the message for clinicians is that the MRI report, often treated primarily as a roadmap for biopsy and local treatment, also carries actionable information about the likelihood that a staging PSMA PET/CT will reveal nodal involvement — information that can be used to sequence investigations, counsel patients, and allocate imaging resources more rationally.

Subject of Research: …

Article Title: MRI-derived T stage predicts nodal positivity on PSMA PET/CT in high-risk primary prostate cancer

Article References: Courtney, M., Whiriskey, R., Johnston, C., & Sheehy, N. (2026). MRI-derived T stage predicts nodal positivity on PSMA PET/CT in high-risk primary prostate cancer. Frontiers in Nuclear Medicine, 6, Article 1854064. https://doi.org/10.3389/fnume.2026.1854064

Image Credits: AI Generated

DOI: 10.3389/fnume.2026.1854064

Keywords: prostate cancer, MRI, PSMA PET/CT, nodal metastasis, T stage, risk stratification, Gleason score, PSA, nuclear medicine, predictive model, MRI-derived, stage