Design: Prospective method-comparison study
Population: 328 men with suspected prostate cancer undergoing 3 T multiparametric MRI
Primary outcome: Agreement of ellipsoid and segmented prostate volume
Effect: ICC 0.985; mean absolute difference 5.5 cc; PSA density sensitivity 0.86–0.95, specificity 0.88–0.97

A prospective study of 328 men with suspected prostate cancer compared prostate volume on 3 T MRI by the PI-RADS ellipsoid formula with manual segmentation of every axial slice. Published in European Radiology on 28 September 2026.

Agreement was excellent (intraclass correlation 0.985; mean absolute difference 5.5 cc). The ellipsoid tended to underestimate glands larger than 80 cc. Adjusting the multiplier for gland size did not meaningfully improve accuracy. At PSA density thresholds of 0.10, 0.15 and 0.20 ng/mL/cc, the ellipsoid classification matched segmentation with sensitivities of 0.86 to 0.95 and specificities of 0.88 to 0.97.

The practical message is reassuring: the quick three-axis measurement is adequate for routine reports. Take care with very large glands, where underestimation raises PSA density and could push a borderline case over a biopsy threshold.

Use the PI-RADS ellipsoid formula for routine prostate volume on MRI
Take care with glands above 80 cc, where the formula underestimates volume
Recheck PSA density near a decision threshold in large glands
Measure all three axes on the correct planes: sagittal and axial T2

Why it matters

It removes the case for time-consuming manual segmentation in everyday prostate MRI.

The statistics, in plain English

An intraclass correlation of 0.985 means the two methods gave almost identical volumes across patients. Underestimating a large gland makes PSA density look higher than it is, which matters only near a threshold.