Data analysis

The presence/absence of dermal melanophages in MPM images was determined by R.M.H based on visual assessment. Criteria for identification of melanophages included large size (>12 μm diameter)(Guitera et al., 2010; Ochoa et al., 2008), increased fluorescence brightness compared to fluorescence from surrounding elastin fibers(Majdzadeh et al., 2015), presence of a nucleus, shape, ill-defined cytoplasmic borders, and clustering distribution within the dermis(Pellacani et al., 2008).

For measuring the melanin volume fraction (MVF), we used the TPEF images acquired in the lesional and perilesional epidermal areas (880 nm excitation wavelength). The MVF corresponding to each acquired stack was calculated following image processing based on an algorithm described in a previous study(Saager et al., 2015). In short, each TPEF image was converted into a binary image using the same threshold for all images. The pixels of value “1” represented the melanin contribution and were used to calculate the area occupied by melanin in each image and subsequently, the volume occupied by melanin in each acquired stack (Fig. 1 C). The MVF was defined as the ratio between the melanin volume and the imaged volume.

The severity of solar elastosis (overproduction or clumping of elastin in the dermis)(Braverman and Fonferko, 1982) in the lesional and perilesional areas was evaluated semi-quantitatively as “mild”, “moderate” and “severe”, based on visual assessment, by the board certified dermatologist R.M.H who has extensive expertise with MPM image interpretation from his involvement in previous studies on MPM clinical skin imaging.(Balu et al., 2014; Balu et al., 2015b; Pouli et al., 2016) We correlated this evaluation with a more rigorous quantitative assessment based on the TPEF and SHG signals from elastin and collagen fibers, respectively. For each patient, we selected about 5 frames corresponding to superficial dermis in each stack. We measured the mean intensity of the TPEF signal from elastin fibers and of the SHG signal from collagen fibers in each frame. We determined the elastosis severity in the superficial dermis by defining a metric as a/(a+b), where a and b represented the mean intensity of the TPEF and SHG signals, respectively. Based on its definition, this metric increased with the increase of elastosis. We used a pair t-test to determine whether the metric value was significantly different in the lesion from perilesion of each patient.