Variation in prostate cancer growth rates in an MRI-based active surveillance cohort

Understanding tumour growth rates helps optimise screening and active surveillance (AS) schedules. We estimated prostate cancer growth rate accounting for individual variation in a longitudinal AS cohort.

We modelled tumour growth in 145 biopsy-confirmed prostate cancer patients undergoing MRI-based AS. Primary lesion volumes were measured longitudinally using planimetry. We compared three mixed-effects models (exponential, Gompertz, and logistic) and investigated relationships between growth rate and clinical characteristics. We estimated the natural trajectory of prostate cancer lesions starting at a single cell to clinical detectability (diameter ≈ 1 cm), with diameters estimated based on spherical volume.

All three models fit observed data well; however, only the Gompertz model provided reasonable estimates from a single cell to an MRI-detectable size (diameter ≈ 3 mm). The Gompertz growth parameter (mean = 0.07, range = 0.02-0.15), describing exponential growth deceleration, was positively correlated with: patient age; lesion volume at AS onset; prostate-specific antigen (PSA) level; and PSA density. Lesions with Gleason 3 + 4 had faster volume doubling times than Gleason 3 + 3 lesions (mean = 3.5 and 5.2 years, respectively). On average, it would take 17 years (95% CI [15, 19]) for a lesion to grow from a single cell to an MRI-detectable size and an additional 12 years to reach a clinically detectable size (95% CI [10, 13]). At age 50, 75% of lesions would remain undetectable by MRI.

Prostate cancer shows slow growth with large variation between patients, posing a challenge for early detection.

Question What is the population distribution of growth rates and the natural history of prostate cancer? Findings Prostate cancer typically grows slowly, with considerable variation between individuals. On average, lesions take 17 years to grow from initiation to an MRI-detectable size. Clinical relevance Small lesions undetectable on MRI may take many years to reach a clinically significant size, posing a challenge for early detection, as it increases the risk of detecting indolent lesions that may never cause harm.