MRI-based grading scheme eases prostate cancer evaluation

Researchers from the U.S. National Cancer Institute (NCI) and National Institutes of Health (NIH) have developed a standardized grading system using multiparametric MRI results and certain clinical markers to simplify the assessment of prostate cancer spread, according to a study published online January 22 in Radiology.

When cancer extends beyond the prostate -- a condition known as extraprostatic extension (EPE) -- patients face a greater risk of cancer recurrence and metastatic disease and a reduced chance of survival even after radical prostatectomy. With this MRI-based grading system, the researchers hope that early detection and treatment for EPE could improve a patient's quality of life.

"The system adds additional diagnostic value to clinical parameters and provides a graded quantifiable risk assessment of pathologic EPE," wrote first author Dr. Sherif Mehralivand and colleagues. "It is based on only a few imaging features, making it easy to teach, and it should be relatively easy to implement."

Various multiparametric MRI biomarkers are associated with EPE, such as curvilinear contact length, capsular irregularity and bulge, and seminal vesicle invasion, according to the authors. However, there is limited research validating the use of these markers, and a lack of standardization for the use of MRI for EPE has caused significant interreader variability.

"Hence, the purpose of our study was to evaluate the diagnostic value of six well-defined multiparametric MRI features for the prediction of pathologic EPE," they wrote. "Based on these results, we also introduce a standardized grading system that predicts the likelihood of pathologic EPE at multiparametric MRI."

The researchers prospectively analyzed 553 men (mean age, 60 years; range: 38-76 years) who underwent 3-tesla MRI to determine their EPE status; 125 (22%) had EPE status confirmed through radical prostatectomy.

Mehralivand and colleagues found that MRI-visible EPE, curvilinear contact length of more than 1.5 cm, and capsular bulge and irregularity were the main features associated with a higher risk of pathologic EPE. The researchers then incorporated the best-performing EPE markers to create their four-point grading system. Grade 0 meant no suspicion of pathologic EPE, grade 1 indicated curvilinear contact length or capsular bulge and irregularity, grade 2 included both features, and grade 3 meant EPE seen on MRI or tumor expansion into other structures, such as the rectum, bladder, or pelvic wall.

As one might expect, the grading system showed incremental increases in EPE detection with higher grades.

Overall, combining the grading system with clinical features proved best at predicting EPE. The area under the curve (AUC) for the combination was 0.81, compared with 0.77 for MRI grade alone (p < 0.001).

The researchers are conducting further validation and multireader studies to verify the generalizability of these results.