MRI Versus PSA in Prostate Cancer Screening

BACKGROUND AND RATONALE:

Prostate cancer is the most common non-cutaneous malignancy and the third leading cause of cancer death among men in Canada1. In 1987, prostate specific antigen was introduced for prostate cancer screening2. Widespread adoption of PSA screening resulted in a significantly increased number of incident cases and a significantly reduced number of cases of metastatic disease at presentation3. Coinciding with the introduction of PSA testing, prostate cancer mortality has decreased approximately 40% from an epidemiologic perspective4. Approximately 45-70% of the decline in mortality is attributable to PSA-based prostate cancer screening5. Several studies have examined whether screening for prostate cancer using the PSA test improves overall and prostate cancer mortality. In particular, two large randomized studies in the U.S. and Europe have been conducted to evaluate this. Recently, the U.S. Preventative Services Task Force (USPSTF) reviewed these and all studies to evaluate whether PSA should be used as a screening test.

Current Recommendations for PSA Screening from the USPSTF

The USPSTF makes recommendations regarding the effectiveness of screening tests for asymptomatic patients after assessing the evidence regarding benefits and harms of an intervention. The most recent USPSTF guidelines regarding prostate cancer were published in 20126.

The first component of the USPSTF assessment is an evaluation of the benefits of early detection and treatment of prostate cancer. The review panel drew on two large randomized controlled trials which have been conducted to assess the effect of PSA-based prostate cancer screening on prostate cancer mortality: the European Randomized Study of Screening for Prostate Cancer (ERSPC)7 and the U.S.-based Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial8. In their recommendation the guideline panel do not consider significant differences between the trials, in large part due to the fact that PSA had been widely adopted in the US during the study interval while it did not have such uptake in Europe. As a result, there are concerns that the trial did not compare screening to no screening and that the trial would be unlikely to find a benefit even if a significant one existed9. Regardless, the panel concluded that prostate cancer screening resulted in the avoidance of 0 to 1 prostate cancer deaths per 1000 men screened6, a minimal benefit.

The second component of the guideline panel's assessment focused on potential harms of early detection and treatment. They found evidence of significant harm due to false-positive PSA results which confer a risk of psychological harm in addition to medical evaluation including biopsy6. Further, they considered there to be, at minimum, a small harm associated with prostate biopsy due to pain, bleeding and infectious risk. The panel also concluded that there was significant evidence of at least moderate overdiagnosis and resultant overtreatment among patients undergoing PSA-based screening.

Due to a perceived lack of benefit and presence of significant harms, the USPSTF concluded, with moderate certainty, that the benefit of PSA-based screening did not outweigh the harms and thus recommended against PSA-based screening for prostate cancer6.

A major limitation of this task force was that they received no input from experts in prostate cancer in order to minimize bias. While this may improve the objectivity of the review, many aspects of understanding prostate cancer were lost. In response to these recommendations, the American Society of Clinical Oncology (ASCO) - the largest oncology association in North America, developed a consensus statement to address the USPSTF recommendations.

ASCO Provisional Clinical Opinion in Prostate Cancer Screening

The ASCO provisional clinical opinion on prostate cancer screening using PSA-testing was based on an Agency for Healthcare Research and Quality systematic review10. Despite drawing on the same data as the USPSTF, this panel concluded that men with a longer life expectancy (greater than 10 years) may benefit from prostate cancer screening using PSA-testing11. In men with a shorter life expectancy, PSA testing was discouraged. Unlike the USPSTF, the ASCO panel placed greater weight on the results of the ERSPC, demonstrating a significant reduction in the risk of prostate cancer death for men undergoing prostate cancer screening. They noted similar harms as the USPSTF, namely false-positive results and prostate biopsy complications.

Since the publication of this opinion, further data has become available which indicates that the relative reduction in prostate cancer death from PSA-based screening may be larger than previously estimated. The most mature data from the ERSPC has 13 years of follow up12. Based on these data, the absolute risk reduction in prostate cancer mortality from PSA screening was 0.11 per 1000 person years or 1.28 per 1000 men randomized. This risk reduction has increased with increasing duration of follow up. Additional analysis has shown an absolute risk reduction of metastatic disease was 3.1 per 1000 men randomized13. In a subgroup of the ERSPC with longer follow-up, the absolute risk reduction in prostate cancer mortality was 4.0 per 1000 men randomized14. This corresponds to a number needed to screen of 293 and number needed to diagnose of 12 in order to prevent one prostate cancer death.

Underdiagnosis of Prostate Cancer with PSA

In addition to harms of biopsy and intervention, the primary concern regarding PSA-based prostate cancer screening is the inability to distinguish between patients with indolent and aggressive forms of the disease. Prostate cancer screening programs have traditionally used a serum PSA cut-off of 4.0 ng/mL to indicate abnormality. However, many men with PSA values in excess of 4.0 ng/mL do not have prostate cancer, and even fewer have clinically significant prostate cancer. Further, up to 25% of men with PSA levels less than 4.0 ng/mL will be found to have high-grade prostate cancer, if subject to biopsy15. Thus PSA-based prostate cancer screening lacks both sensitivity and specificity to identify men with aggressive prostate cancer. As a result, in addition to significant overdiagnosis and overtreatment which has been well recognized, there is a risk for underdiagnosis.

Due to the USPSTF recommendations against PSA-based prostate cancer screening, there has been a significant decrease in PSA testing being performed by primary care physicians16. It is predicted that complete discontinuation of PSA-based screening would result in the prevention of overdiagnosis for 710,000 to 1,120,000 men in the United States over a 12 year period17. However, it would result in 36,000 to 57,000 preventable deaths due to prostate cancer over the same time period. Thus, improved methods for prostate cancer screening may allow for a diminishment in overdiagnosis while avoiding preventable deaths from prostate cancer.

Magnetic Resonance Imaging (MRI) of The Prostate

Multiparametric prostate magnetic resonance imaging (mpMRI) has an increasingly large role in the management of patients with clinically-localized prostate cancer. MpMRI was initially used as a staging test in patients with prostate cancer18. Despite the use of what would now be considered obsolete technologies, Bezzi et al. demonstrated in 1988 that MRI could identify nodal metastases with an accuracy of 88% and could distinguish disease localized to the prostate from that invading beyond the capsule with an accuracy of 78% among patients undergoing prostatectomy18. Since that time, there has been a migration in the use of MRI earlier in the disease process.

In the realm of prostate cancer treatment, MRI is used for the monitoring of patients on active surveillance following a prostate cancer diagnosis as a means of reducing prostate biopsies, with their incumbent risks19,20. When performed in the evaluation of patients with elevated PSA levels with previous negative prostate biopsy, mpMRI has been shown to identify clinically significant prostate cancers which would have been otherwise missed by routine systematic biopsy21. The use of MRI and ultrasound fusion imaging in the targeting of prostate biopsy has increased detection of clinically significant prostate cancer while limiting the diagnosis of clinically insignificant prostate cancer22.

Among men with an abnormal PSA who have never undergone a prostate biopsy, mpMRI demonstrated promise in both the detection and exclusion of prostate cancer, using an extensive prostate mapping biopsy (median 41 cores) as the referent23: the AUC was 0.27 (95% CI 0.65-0.79). In a multivariable analysis of an independent cohort including age, family history, prior 5-alpha reductase inhibitor use, digital rectal examination findings, PSA level, PSA density, and MRI score, only MRI score was predictive of clinically significant (Gleason score ≥7) prostate cancer among men without a history of previous prostate biopsy (adjusted OR 40.2, p=0.01)24.

While historically prostate MRI has required the use of an endorectal coil25, recent advances in MRI technology have obviated the need for the coil26, thus reducing the cost and burden of the imaging.

Pilot Study to Examining the Feasibility of MRI Prostate Cancer Screening

We recently conducted a pilot study assessing the feasibility of mpMRI as an initial prostate cancer screening test27. Following a newspaper based call for volunteers, we had 319 men present for possible inclusion in this study. Of these, 120 were eligible, 50 were enrolled due to limitations in funding, and 47 completed the study protocol. Serum PSA testing, mpMRI, digital rectal examination, and systematic (+/- targeted) prostate biopsies were performed on all men.

Prostate cancer was identified in 18 of 47 men (38.3%). MpMRI (AUC 0.81, 95% CI 0.67-0.94) significantly outperformed PSA (AUC 0.67, 95% CI 0.52-0.84) in the prediction of prostate cancer. In multivariable analyses including age, digital rectal examination findings, PSA and MRI score, mpMRI was the only significant predictor for the presence of prostate cancer (adjusted OR 2.7, 95% CI 1.4-5.4). These findings persisted when we sought to predict only clinically significant prostate cancer (Gleason ≥7; adjusted OR 3.5, 95% CI 1.5-8.3).

To better evaluate the use of mpMRI in prostate, a large, randomized study comparing the efficacy of mpMRI in diagnosing clinically significant, aggressive prostate cancer, compared to conventional PSA screening will be required.