Multiparametric MRI demonstrates poor sensitivity in staging for prostate cancer

Multiparametric magnetic resonance imaging (mpMRI) was found to demonstrate poor sensitivity, positive predictive value (PPV), and negative predictive value (NPV) in staging for prostate cancer, according to findings from a recent study published in Urologic Oncology.¹

No association was observed between the accuracy of mpMRI and MRI institution, MRI strength, or MRI timing.

"Multiparametric MRI has become a mainstay of prostate cancer evaluation and is routinely used in practice for local staging. However, our data from a multi-institutional collaborative demonstrates that, in the real-world setting, mpMRI should be used cautiously for local staging prior to radical prostatectomy. While it remains an important tool in our arsenal and is helpful in the diagnostic pathway, it is not a perfect local staging imaging modality," said lead author Thenappan Chandrasekar, MD, in correspondance with Urology Times®. Chandrasekar is an associate professor of urology at the University of California, Davis Medical Center in Sacramento, California.

For the study, investigators collected institutional data from a single academic medical center, as well as from the Pennsylvania Urologic Regional Collaborative (PURC) on men who underwent radical prostatectomy in 2021. The ability of mpMRI to predict pT2N0 organ confined disease (OCD), extracapsular extension (ECE), seminal vesicle invasion (SVI), lymph node involvement (LNI), and bladder neck invasion (BNI) was assessed in each dataset.

Institutional data showed that among men who underwent preoperative MRI, specificity was high for all outcomes measured. However, sensitivity (2.9%), PPV (40% to 100%), and NPV (56.3% to 94.3%) were generally low between all outcomes measured.

The investigators then stratified the data by American Urological Association (AUA) risk stratification (RS) groups, finding an association between AUA RS and MRI’s ability to predict SVI (P < .001). No significant associations were found between AUA RS and MRI’s ability to predict ECE (P = .947), BNI (P = .463), or LNI (P = .530).

Data obtained from PURC revealed similar trends to those seen in the institutional data. Specificity was high for all outcomes measured, at 93.9% for ECE, 98.3% for SVI, 98.6% for LNI, and 92.6% for OCD. PURC data did not capture BNI. Further, sensitivity (16.8% to 24.4%), PPV (35.8% to 68.2%), and NPV (68.4% to 96.2%) were generally low among outcomes measured.

When the data was stratified by AUA risk group, a significant association was seen between AUA RS and the ability of MRI to predict ECE, SVI, and LNI (P < .001).

AUA RS also revealed associations in MRI-A, sensitivity, specificity, PPV, and NPV. Preoperative mpMRI in both the institutional data and the data obtained from PURC was found to have higher sensitivity and PPV among patients in higher risk categories. Conversely, higher NPV of mpMRI was seen among patients in lower risk groups. Specificity was similar across all risk groups, ranging from 72.5% to 100.0% in the institutional data and 87.3% to 99.3% in the PURC data.

No association was observed between the accuracy of mpMRI and MRI institution, MRI strength, or MRI timing.

The authors concluded, “Our results support the notion that mpMRI is an inadequate screening tool for locally advanced [prostate cancer] and should not be used as the sole means of presurgical [prostate cancer] staging. Rather, mpMRI findings should be carefully considered in addition to other important clinical factors which include individual patient history and clinical risk stratification and used in concert with other staging techniques.”

Reference

1. Chandrasekar T, Denisenko A, Mico V, et al. Multiparametric MRI is not sufficient for prostate cancer staging: A single institutional experience validated by a multi-institutional regional collaborative. Urol Oncol. Published online June 23, 2023. Accessed June 26, 2023. doi:10.1016/j.urolonc.2023.05.004