MRI/US Fusion Improves Prostate Cancer Detection
The approach can help identify suitable candidates for active surveillance, but false positives and cost effectiveness are concerns.
AMELIA ISLAND, FL—Targeted biopsy techniques combining magnetic resonance imaging (MRI) and ultrasound can improve detection of significant prostate cancer (PCa) and could influence development of future criteria for active surveillance, according to a researcher. Significant prostate tumors still are often found only on template biopsies, so both targeted and template biopsy is optimal.
At the Florida Urological Society's 68th annual meeting, Sanoj Punnen, MD, a urological oncologist at University of Miami Hospital, discussed benefits of MRI/ultrasound fusion (MRI/US) in targeted biopsy, the revised MRI Prostate Imaging Reporting and Data System (PI-RADS) image scoring to reduce false positive predictions, and the potential to adjust active surveillance criteria in light of technological advancements.
“Current active surveillance criteria, I think, were never really made with this type of technology in mind, Dr. Punnen told attendees. “We may need to revisit some of these [criteria] to see how we can refine the criteria to be more appropriate to how we're moving forward in the way we do biopsies right now.”
MRI/US fusion is the latest technology in targeted biopsy, fusing MRI with real-time ultrasound to aim a biopsy needle into an area of concern with precision. Diffusion-weighted imaging shows water molecule diffusion through tissue, indicating how “thick” the tissue is T2-weighted imaging gives a high-resolution image of prostate tissues, while a third image, dynamic contrast, shows blood flow into areas of tissue. The 3 image sequences provide independent and complementary data to create a 3D map of the prostate, showing specific areas that may be at risk of significant disease.
A National Cancer Institute prospective study conducted from 2007 to 2014 tested a cohort of 1,003 men—80% of whom had a previous negative biopsy result—with concurrent targeted and systematic (random sampling) biopsy. Both methods detected PCa similarly, but targeted biopsy diagnosed 30% more high-risk (Gleason score ≥4+3) and 17% less low-risk (Gleason score 3 + 3 or low-volume 3 + 4) cases, according to a report in the Journal of the American Medical Association (2015;313:390-397).
PI-RADS scores each prostate lesion 1 to 5, with 1 indicating that clinically significant cancer (Gleason score 7 or higher, volume greater than 0.5 cc, and/or extraprostatic extension) is highly unlikely to be present and 5 indicating the highest probability that it is present. Between October 2014 and July 2015, Dr. Punnen and colleagues studied a prospective cohort of patients undergoing MRI-US fusion biopsy at their institution and compared the detection of cancer in the peripheral zone (PZ) versus transitional zone (TZ).(http://meeting.ascopubs.org/cgi/content/abstract/34/2_suppl/56).
The study included 133 men with elevated PSA or positive digital rectal examination findings. The men underwent MRI/US fusion biopsy with an average of 2 cores taken per target. Target location was recorded and categorized as TZ or PZ. Radiologists grouped the targeted lesions according to level of suspicion as probably benign (1–2), indeterminate (3), and probably malignant (4–5). Dr Punnen and colleagues identified 143 lesions in the PZ and 82 in the TZ. The researchers found that 44.7% of lesions in the PZ were probably malignant compared with 57.3% in the TZ. Cancer was diagnosed in 23% of lesions in the PZ versus 9.7% in the TZ. The investigators identified a non-significant trend toward higher detection of significant cancer in the PZ compared with TZ (13.3% vs. 6.1%). In addition, PZ lesions with PI-RADS scores higher than 4 were associated with a significant 3-fold increase in the odds of detecting cancer versus PZ lesions with score below 4, whereas in the TZ, no increased risk of cancer was observed with higher PI-RADS scores.
In the first version of PI-RADS criteria, the findings associated with a score of 1–2 were 86% benign, 14% any cancer, and 9% significant cancer in the PZ and 100% benign in the TZ. For a PI-RADS score of 3, the findings were 85% benign, 15% any cancer, and 8% significant cancer in the PZ and 88% benign, 12% any cancer, and 6% significant cancer in the TZ. For a score of 4–5, the findings were 62% benign, 38% any cancer, and 22% significant cancer in the PZ and 86% benign, 14% any cancer, and 10% significant cancer in the TZ.
In the TZ, PI-RADS 4-5 scores produced false positive results frequently, identifying benign tumors as likely to be clinically significant 74% of the time, Dr Punnen said. The PI-RADS 4-5 in the PZ were benign in 46% of cases.
The PI-RADS criteria were updated in August 2015, making diffusion-weighted imaging the major data source for PZ lesions and prioritizing T2-weighted imaging for the TZ. This updated version decreased the proportion of false positives for PI-RADS 4–5 in both zones. Using this second version of PI-RADS, the findings associated with a score of 4–5 were 36% benign, 64% any cancer, and 46% significant cancer in the PZ and 62% benign, 38% any cancer, and 22% significant cancer in the TZ. While some of this improvement may be due to the revision in PIRADS, a significant amount may be due to the experience and learning curve of MRI interpretation.
Though targeted biopsy is more accurate, Dr Punnen indicated that false positives, cost effectiveness, and increased core samples are major concerns. His team reviewed a cohort of about 100 men eligible for active surveillance based on the results of their random biopsy. Taking 2 additional targeted cores from every significant lesion disqualified 10%–40% of men from some active surveillance criteria, particularly those with an absolute maximum of 2 cores.
“Most of the current active surveillance criteria published in the literature were based on template biopsy of the prostate,” Dr Punnen said. “When applied to a cohort of men undergoing targeted biopsy of the prostate, they result in a significant exclusion of potentially good candidates from active surveillance of their low risk tumors. Future guidelines to reflect the current technology being employed in practice may facilitate improved decision-making regarding the management of prostate cancer.”