Background: The strongest known risk factor for melanoma is the presence of many melanocytic naevi, commonly known as “moles”. Prospective population-based studies of naevus counts in adults, especially older adults who carry highest melanoma risks, are lacking. This study aimed to understand the change in total naevus count over a 3-year period.
Methods: This prospective longitudinal study of naevi in Queensland recruited participants 18-70 years from the electoral roll. Participants underwent 3D total-body-imaging every six months for three years. To automate the process of identifying and counting naevi, a convolution neural networks-based algorithm was developed and validated. Changes in naevus count over time were described, and a multinomial model with change in total naevus count as three groups; decreased, no change and increased, was used to assess participant characteristics associated with these groups.
Results: Participants’ (n=156) baseline median total naevus count was 32 (IQR:16-62). Over the average 2.6 years observation period, naevi appeared and disappeared dynamically, resulting in a gross median increase of 2 naevi (IQR:0-5; p<0.001). Overall, 71 (46%) participants had both appearing and disappearing naevi, 59 (38%) only appearing, 11 (7%) only disappearing naevi, and 15 (10%) had stable naevus count. The gross naevus-count change and volatility was independent of gender or age, but associated with baseline counts and family history of melanoma.
Conclusion: Our results contradict the previous understanding that naevus counts decrease with age. We saw on average naevus counts increase even in older adults over time, and that naevus volatility is common, with naevi appearing and disappearing across all demographic age groups. Larger and longer studies should assess whether greater change is associated with melanoma risk.