Who Really Needs A Piece Of BML-190 ?

Figure 9, however, shows that there is actually a larger number of collagen outlying spectra associated with untreated samples (magenta circles) than treated samples (blue squares). The reason for this is that all Cyclopamine molecular weight of the ��high collagen�� outlying spectra evident in Figure 9 (data points within and near the dashed ellipse) were recorded from the deep surface of normal tissue samples, and a greater number of deep surface Raman measurements were performed on untreated samples than treated samples; treatment (i.e., partial laser ablation) was clearly not sufficient to fully remove the epidermis and expose the collagen-rich dermis on the superficial side of the samples. It is the difference in deep versus superficial surface sampling numbers and hence numbers of high collagen-containing spectra for the treated and untreated data groups that is likely the main cause of the better-than-random classification success rate in the present case. For SCC tumor samples, one might expect a much higher degree of tissue uniformity from the superficial through to the deep surface compared to normal skin. It is therefore not easy to predict whether Raman spectral changes will occur as a result of partial laser ablation of an SCC tumor. Again, for a sufficiently large data set it may be expected that the percentage of correct classifications using logistic regression between the two categories (treated vs. untreated) should be around PLK inhibitor 50% if there is no significant biochemical difference (on average) between treated and untreated samples. However, for SCC, spectra from treated samples were classified with 85% sensitivity and 83% specificity against untreated samples (Table 4), indicating quantifiable differences in the Raman spectra between the two cohorts. To further quantify this implied separation of treated and untreated PCA data, we performed one-way ANOVA (treated vs. untreated) on the scores for each of the first five PCs. For normal skin, the average PC1 scores for treated and untreated BML-190 cohorts were significantly different (P?