Originally published 12/8/2016
It is just over 5 years since we posted our blog on better prostate cancer (PCa) risk prediction using multiparametric MRI (mpMRI) than a nomogram called the Partin tables. These tables, a mathematical model incorporating multiple PCa risk factors, were a pioneering system for classifying risk in order to match treatment to the level of disease aggression. The Partin tables were succeeded by the D’Amico classification system, as well as the online risk calculator tool developed by Memorial Sloan Kettering Cancer Center (MSKCC) and the CAPRA (Cancer of the Prostate Risk Assessment) Score developed by University of California San Francisco. We found two very good studies that update and confirm our previous post:
a) A 2018 study compared mpMRI vs. the Partin tables and the MSKCC calculator in their accuracy for stratifying PCa risk prior to external beam radiation. There were 80 patients in the study, all of whom underwent 3T mpMRI using 3 sequences (T2 weighted, diffusion weighted, and dynamic contrast enhanced MRI). Based on their clinical factors, the other two nomograms were applied, and the results compared with what imaging showed. Not surprisingly, mpMRI was better at detecting higher risk levels, including PCa extending beyond the prostate. The authors concluded that mpMRI “should refine or replace nomograms in risk classification before EBRT.”
b) A 2019 study compared 3T mpMRI vs. the Partin tables, the MSKCC calculator, and the CAPRA score in predicting PCa extending beyond the prostate capsule before prostatectomy. 73 patients were involved, and there were three MRI readers who staged PCa based on imaging, and all results (imaging, Partin tables, MSKCC and CAPRA) were correlated. Statistical analysis showed that “mpMRI was more accurate than clinical models and mpMRI plus clinical models in predicting stage ≥pT3, except for the combination of mpMRI and CAPRA in two out of three readers.”[i] As in the 2018 study, these results suggest that mpMRI independently can either outperform other clinically predictive models or—as in the case with CAPRA—enhance and validate the prediction.
In the fall of 2015, I posted a blog on prostate cancer risk calculators, or nomograms, and how multiparametric MRI (mpMRI) improves risk assessment. I wrote, “A nomogram is a set of scales that can be used to calculate an unknown value, and when adapted for medicine, they act as a statistical modeling tool. The classic nomogram used to predict the chances that PCa has left the gland at the time of treatment is called the Partin tables.” [ii] The Partin tables are periodically updated to keep in step with the growing body of clinical evidence. However, they only offer probability, not certainty.
To illustrate, let’s take the example of a 61-year old Caucasian man named Will. Five years ago, Will’s PSA was 2.3, but recently it has been rising. His GP does a digital rectal exam. It’s normal, but the doc refers Will to a urologist for a biopsy. After undergoing a standard 12-core TRUS biopsy, 2 needles turned out to be positive for prostate cancer: one shows Gleason 3+3, and the other Gleason 3+4. The doctor diagnoses Will with Stage T1c cancer. Based on this information, what are the chances that the cancer has escaped the gland? Will finds two online patient calculators based on the Partin tables, and enters his clinical factors (age, PSA, stage, Gleason score). Here’s what he discovers:
- org (http://www.prostatecalculator.org) produces a single result, a forecast of the likelihood that the cancer has left the gland at the time of treatment. It predicts that Will has a 51.7% chance that cancer will have spread outside the gland at the time of treatment – basically, it’s 50/50. The urologist recommended robotic prostatectomy for Will because of the Gleason 3+4. With 50/50 odds, Will wonders if the chances for cure are good enough to risk incontinence or ED. He wonders if he should discuss Active Surveillance with the doctor. It’s a tough call.
- The Johns Hopkins “New Partin Nomogram” (http://urology.jhu.edu/support/partinTables.php) produces four probabilities for him to weigh: Organ confined (OC), Extracapsular extension (ECE, meaning the tumor extends beyond the prostate capsule into the prostate bed), Seminal vesicle invasion (SVI, meaning the tumor has invaded the seminal vesicles) and Lymph node invasion (LNI). Will thinks this looks a little more promising:
- Organ-confined 2%(60.4%-68%) (Maybe Active Surveillance is okay after all!)
- Extraprostatic extension 8%(25.1%-32%) (Hmm, one chance in four of early spread)
- Seminal vesicle invasion 8%(4.2%-8%)
- Lymph node invasion 1%(0.6%-2%)
How much can Will trust the 2012 updated Partin tables for making his decision? A new international multi-center study explored how valid the tables are. The research team led by Dr. Leyh-Bannurah analyzed a records database of 25,254 North American patients treated in community institutions with prostatectomy and lymph node dissection from 2010-13. The authors compared each patient’s pre-surgery Partin table predictions with the post-surgery pathology findings.[iii] They published their analysis in terms of average accuracy for the whole population and subgroups:
- Predicting OC – 70.4% accurate
- Predicting ECE – 59.9% accurate
- Predicting SVI – 72.9% accurate
- Predicting LNI – 77.1% accurate
For elderly and African American patients, accuracy was lower across all four endpoints. Thus while average Partin table accuracy for predicting SVI and LNI (locally advanced prostate cancer) was considered excellent, it was merely “good” for OC and “poor” for ECE. Not very helpful for Will.
In our mpMRI experience at the Sperling Prostate Cancer, a single scan is worth a thousand probability calculations. Research backs this up. A Duke University/Johns Hopkins team designed a study “to test our hypothesis that multiparametric magnetic resonance imaging (mpMRI) may have a higher prognostic accuracy than the Partin tables in predicting organ-confined (OC) prostate cancer and extracapsular extension (ECE) after radical prostatectomy (RP).”[iv] Both the Leyh-Bannurah and Duke teams used the 2012 updated Partin tables for pre-surgery predictions and compared those with post-surgery pathology – so the two studies are worth comparing despite the small sample size (60 patients) at Duke. The Duke team found that mpMRI performed significantly better than the Partin tables in predicting OC disease (91.2% accuracy in identifying OC prostate cancer) and almost 90% accuracy in ruling out ECE (89.7% negative predicting value for ECE).
- Predicting OC – 70.4% accurate
- Predicting ECE – 59.9% accurate
- Predicting SVI – 72.9% accurate
- Predicting LNI – 77.1% accurate
They concluded that “mpMRI should be considered when planning prostate cancer treatment in addition to readily available clinical parameters.”
Let’s return to Will. Based on the Partin tables, he is torn between radical treatment and Active Surveillance. However, his online research leads him to information about mpMRI, and real-time MRI guided tumor ablation. He likes what he reads, and hopes that there’s a middle-ground resolution to his dilemma of RP vs. AS. He comes to us for an mpMRI. It shows a single focus of prostate cancer and no evidence of ECE. To rule out more aggressive cells, we do an in-bore MRI-guided targeted biopsy that confirms an area of Gleason 3+4 but no higher. Now, with his previous records, we have a portrait of his tumor. We discuss all his treatment options, including the latest advances in Focal Laser Ablation, Exablate (MRI-guided Focused Ultrasound) and TULSA-PRO. What does Will decide?
Whatever treatment he chooses, he has far more confidence in his decision than he would have had with the Partin tables alone. mpMRI gives him the information he needs to feel very good about his choice. The online prostate risk calculators can give an early but incomplete peek at the odds of prostate cancer leaving the gland. Only mpMRI, which outperforms all nomograms, supplies the detailed 3-D images that are essential for matching treatment choices with each person’s prostate cancer.
NOTE: This content is solely for purposes of information and does not substitute for diagnostic or medical advice. Talk to your doctor if you are experiencing pelvic pain, or have any other health concerns or questions of a personal medical nature.
References
[i] Zanelli E, Giannarini G, Cereser L, Zuiani C et al. Head-to-head comparison between multiparametric MRI, the partin tables, memorial sloan kettering cancer center nomogram, and CAPRA score in predicting extraprostatic cancer in patients undergoing radical prostatectomy. J Magn Reson Imaging. 2019 Nov;50(5):1604-1613.
[ii] https://sperlingprostatecenter.com/mpmri-improves-the-accuracy-of-the-partin-tables-and-other-nomograms/
[iii] Leyh-Bannurah SR, Gazdovich S, Budäus L, Zaffuto E et al. Population-Based External Validation of the Updated 2012 Partin Tables in Contemporary North American Prostate Cancer Patients. Prostate. 2016 Sep 29. doi: 10.1002/pros.23253. [Epub ahead of print]
[iv] Gupta RT, Faridi KF, Singh AA, Passoni NM et al. Comparing 3-T multiparametric MRI and the Partin tables to predict organ-confined prostate cancer after radical prostatectomy. Urol Oncol. 2014 Nov;32(8):1292-9. doi: 10.1016/j.urolonc.2014.04.017. Epub 2014 May 23.
About Dr. Dan Sperling
Dan Sperling, MD, DABR, is a board certified radiologist who is globally recognized as a leader in multiparametric MRI for the detection and diagnosis of a range of disease conditions. As Medical Director of the Sperling Prostate Center, Sperling Medical Group and Sperling Neurosurgery Associates, he and his team are on the leading edge of significant change in medical practice. He is the co-author of the new patient book Redefining Prostate Cancer, and is a contributing author on over 25 published studies. For more information, contact the Sperling Prostate Center.
