MRI Combination Techniques Hone Accuracy of Prostate Cancer Diagnosis

Andre Luis de Castro Abreu, MD

Several studies presented at this year’s American Urological Association (AUA) meeting underscored the value of multiparametric magnetic resonance imaging (mpMRI) in the diagnosis and management of prostate cancer. According to the findings, the use of mpMRI in combination with other tools for diagnosis can reduce unnecessary prostate biopsies. They also indicated that use of mpMRI is on the rise, particularly in men aged 60 to 65.

A retrospective study of 401 patients with negative findings from mpMRI testing (PI-RADS score <3 or Likert score <3) sought to determine whether mpMRI could identify patients most likely to benefit from biopsy.¹

A negative mpMRI with no prior biopsies and a low prostate-specific antigen (PSA) density (PSAD) enabled investigators to rule out the presence of clinically significant (Gleason score ≥ 3+4) and high-grade (Gleason score ≥ 4+3) prostate cancer with a high degree of accuracy.

The combination of a negative mpMRI, negative biopsy history, and PSAD <0.15 ng/mL/cc was associated with a negative predictive value (NPV) of 95%. With a higher PSAD and a non-negative prostate biopsy history, NPV declined by more than 20%. The results led investigators to propose an MRI-based algorithm for determining the need for prostate biopsy.

“If the MRI suggests a clinically suspicious lesion, we perform a biopsy; if the MRI is negative, we look at the patient’s biopsy history,” said Andre Luis de Castro Abreu, MD, at the meeting, held May 18 to 21 in San Francisco, California. “If the patient never had prostate biopsy, it is probably better to perform a biopsy.

“If the patient had a prior negative biopsy, we then assess the markers, and if the levels are high or suspicious, the patient has a biopsy. If not, then the patient may avoid a prostate biopsy,” said Abreu, an assistant professor of clinical urology at the University of Southern California, Los Angeles.

The study population included 107 patients with no prior biopsy, 212 with prior negative biopsy, and 79 with prior positive biopsy. Biopsy results revealed cancer in 34% of the patients, including clinically significant prostate cancer in 11% and high-grade cancer in 5%. MRI alone had an NPV of 95% for highgrade prostate cancer, 89% for clinically significant cancer, and 66% for any prostate cancer.

The biomarker analysis showed that PSAD had significant discriminatory power for distinguishing clinically significant from insignificant cancer (P <.01) and from benign disease (P <.001). For a subgroup of patients with a prior negative biopsy (n = 128), pairing a negative mpMRI and PSAD cutoff of <0.15 ng/mL/cc produced an NPV of 95%. Further lowering of the PSAD cutoff further improved NPV but with a trade-off of excluding substantially more patients.

The study is the largest to date to evaluate the combination of a negative MRI and PSAD for ruling out clinically significant and high-grade disease. The results are in line with those of the previous research, which yielded NPV values of 89% to 100%, Abreu said.

Figure. PSA Test Compliance Rate Significantly Higher than Biopsy Compliance Rate⁶

Combining PHI and mpMRI

A recent review showed that previous studies of MRI to detect clinically significant prostate cancer yielded NPVs of 63% to 96%, as determined by biopsy or prostatectomy specimen.² Results of a large, multicenter study published last year suggested that triage by mpMRI could eliminate 25% of unnecessary prostate biopsies.³ However, NPV ranged between 72% and 89%, depending on the definition of clinically significant prostate cancer.The Prostate Health Index (PHI) and mpMRI have been used independently to predict prostate cancer grade reclassification (Gleason score >6) for patients on active surveillance (AS). Results from a study reported at the AUA meeting showed that combining these tools would have enabled more accurate prediction of grade reclassification at surveillance biopsy than use of mpMRI or PHI alone. Further, the combination of these tools would have eliminated the need for repeat biopsy in about 20% of cases.⁴

Biopsy requirements associated with AS are a major contributor to noncompliance (Figure),^5,6 and identifying methods that would safely reduce biopsy burden without missing opportunities to treat cancers that develop high-risk features could improve compliance with AS, said lead author Zeyad R. Schwen, MD, a urology resident at Johns Hopkins University in Baltimore, Maryland.

Schwen reported findings from a retrospective study of 253 men enrolled in AS who underwent mpMRI and PHI no more than 6 months apart, followed by prostate biopsy. Investigators evaluated PHI, PHI density (PHID), and PSAD across a range of PI-RADS scores to determine the predictive accuracy for prostate cancer grade reclassification, defined as an increase in cancer grade from the patient’s original biopsy to a Gleason score >6.

Biopsy results showed that 38 patients had an increase in Gleason score and that grade reclassification did not occur in the remaining 215 patients. PHI, PHID, and PSAD all differed significantly between men who had grade reclassification and those who did not (P = .037 to P = .001). PSA levels did not differ significantly between those with and without grade reclassification.

Additionally, PI-RADS scores differed significantly between men with and without grade reclassification, including the proportion of men with PI-RADS scores ≤3 or those with PI-RADS scores of 4 or 5 (P <.001). The proportion of men who met National Comprehensive Cancer Network (NCCN) criteria for low or very low risk was significantly different between the patients who had grade reclassification and those who did not (P = .001).

Comparing the ability of various parameters to rule out grade reclassification, Schwen and colleagues found that a PI-RADS score <3 was associated with an NPV of 91%, a PSAD <0.07 ng/ mL/cc with an NPV of 88%, a PHID <0.39 with an NPV of 86%, and a PHI <25.6 with an NPV of 95%.

Investigators then paired a PI-RADS score ≤3 with the various parameters and recalculated NPV. The results showed an NPV of 95% for PSAD <0.07, 92% for PHID <0.39, and 98% for PHI <25.6. Receiver operating characteristic analysis showed that the combination of a PI-RADS score ≤3 plus PHI <25.6 was associated with an area under the curve of 0.70, compared with 0.68 to 0.69 for the other pairings.

Increasing Use of Prostate MRI

“Using a PHI cutoff of less than 25.6 in combination with a PI-RADS less than or equal to 3, nearly one-fifth of surveillance biopsies could have been avoided at the cost of missing only 2.6% of clinically significant prostate cancers,” Schwen said. “Together, PHI and multiparametric MRI may be useful for decreasing the burden of surveillance prostate biopsies. This may result in fewer biopsy-associated complications, less cost to the healthcare system, and improved compliance with active surveillance.”The studies demonstrating the potential benefits of prostate mpMRI reflected the growing use of the imaging modality, as documented by a study reported as a poster presentation at the AUA meeting. Investigators examined national trends in the use of mpMRI for patients undergoing prostate biopsy, using data from a large national payer (OptumLabs), Simon Kim, MD, MPH, a urologic oncologist at UH Cleveland Medical Center, and colleagues reported.⁷

The analysis included men aged 40 to 80 who underwent PSA screening for prostate cancer from January 2010 through December 2016. Men enrolled had to have a diagnosis of enlarge prostate or elevated PSA within 14 days after the screening date. Investigators collected data on the use of mpMRI of the prostate and biopsies within the 30 days after diagnosis.

In addition to documenting the use of prostate MRI among men screened for prostate cancer and identifying patient characteristics associated with the use of MRI, investigators evaluated the frequency of diagnosis of localized prostate cancer within 3 months after biopsy.

The data showed that 42,428 men underwent prostate cancer screening, and 7799 (18.3%) had prostate MRI. Results from previous studies had shown a rate of 3.7%, the investigators noted. Over the 7 years studied, use of prostate MRI increased from 190.3 to 257.5 per 1000 biopsies in an unadjusted analysis and from 192.2 to 257.5 after adjustment.

Rates of MRI use increased in all age groups except men aged 40 to 49. Black and Asian patients were less likely to undergo prostate MRI, and Hispanic patients and men with Medicare Advantage coverage were more likely. Investigators said “the largest and only significant increase in use” was in men aged 60 to 65 (4.4 per 1000 in 2010 to 16.0 in 2016).

Patients who underwent prostate MRI were significantly more likely to have a diagnosis of localized prostate cancer than men who did not have MRI exams (OR, 1.77; P <.001). The observation supports use of MRI to optimize the detection of clinically significant prostate cancer, Kim and colleagues concluded.

References

1. Oishi M, Smyth TB, Shin T, et al. Negative MRI: which patients could safely avoid prostate biopsy? Results from multi-institutional study in 401 patients [AUA abstract MP57-08]. J Urol. 2018;199(4)(suppl):e767.
2. Fu&#776;tterer JJ, Briganti A, De Visschere P, et al. Can clinically significant prostate cancer be detected with multiparametric magnetic resonance imaging? A systematic review of the literature. Eur Urol. 2015;68(6):1045- 1053. doi: 10.1016/j.eururo.2015.01.013.
3. Ahmed HU, El-Shater Bosaily A, Brown LC, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS): a paired validating confirmatory study. Lancet. 2017;389(10071):815-822. doi: 10.1016/S0140-6736(16)32401-1.
4. Schwen ZR, Carter HB, Tosoian JJ, et al. Prostate health index and multiparametric MRI to predict prostate cancer grade reclassification in active surveillance. Presented at: American Urological Association 2018 Annual Meeting; May 18-21, 2018. San Francisco, CA.
5. Moore CM, Parker C. The evolution of active surveillance for prostate cancer. Eur Urol. 2015;68(5):822-823. doi: 10.1016/j.eururo.2015.07.020.
6. Bokhorst LP, Alberts AR, Rannikko A, et al. Compliance rates with the Prostate Cancer Research International Active Surveillance (PRIAS) protocol and disease reclassification in noncompliers. Eur Urol. 2015;68(5):814-821. doi 10.1016/j.eururo.2015.06.012.
7. Kim A, Karnes RJ, Houten HV, et al. Contemporary national trends in prostate MRI among patients undergoing a prostate biopsy: results from a privately insured patient population. Poster presented at: American Urological Association 2018 Annual Meeting; May 18-21, 2018. San Francisco, CA.