Figure 3

Quantitative assessment of pre-PDT vascular density in normal skin, peri-tumoral BCC tissues and the four (nodular, nodular with necrotic core, superficial and sclerosing) BCC subtypes. (A) Averaged vascular density results for the various regions derived from en-face OCA images; asterisk (*) indicates a statistically significant difference between normal skin and nodular without necrotic core, superficial and sclerosing BCC subtypes (Bonferroni post-hoc test for multiple comparisons, with p < 0.001); double asterisk (**) indicates a statistically significant difference between nodular BCC with necrotic core from nodular without necrotic core, superficial and sclerosing BCC subtypes (p < 0.001). (B) Maximal distance to the nearest skeletonized vessel results for the various BCC subtypes derived from en-face OCA images; asterisk (*) indicates a statistically significant difference between nodular BCC without necrotic core and normal skin and all the other BCC subtypes (Bonferroni post-hoc test for multiple comparisons, with p < 0.001). The results are shown as mean ± SD. (C) ROC for differentiating normal face skin from nodular, superficial and sclerosing BCC subtypes based on OCA’s vascular density metric. (D) ROC for differentiating the nodular-with-necrotic-core BCC (more clinically challenging pathology) from the more treatable superficial, sclerosing and nodular subtypes based on OCA’s vascular density metric. Dots on the curve represent possible choices of the OCA vascular threshold value (Pth), yielding the listed false negative (FN), false positive (FP) and diagnostic accuracy (DA) values. The curves look stepwise because of limited sample size. The two sets of values on each curve represent possible choices of the vascular density threshold which provide similar diagnostic accuracies but different percentage of false positive and false negative outcomes. Bold font numbers indicate the choices of the vascular density Pth using a trade-off between the percentages of the FN and FP outcomes and the greatest DA. Additional abbreviations: ROC - receiver operating characteristic; AUC - area under ROC curve.