Reliable Entity Subtyping in Non-small Cell Lung Cancer by Matrix-assisted Laser Desorption/Ionization Imaging Mass Spectrometry on Formalin-fixed Paraffin-embedded Tissue Specimens

Supplemental Data

• Supplementary Figure 1 (.pdf, 176 KB) - Spectra overlay between ADC (spectra in red) and SqCC (spectra in green) used to train the LDA classification model. CK7 (peak identified at m/z 1406.7) is present exclusively in ADC spectrum profiles; CK5 (peak identified at m/z 1410.7) and CK15 (peak identified at m/z 1821.9) are detected exclusively in SqCC.
• Supplementary Figure 2 (.pdf, 128 KB) - Average spectrum obtained from one representative tissue section before and after processing. The expanded view from m/z 1520-1620 highlights the data quality of the peaks observed between the two spectra.
• Supplementary Figure 3 (.pdf, 87 KB) - Correlation of the discriminatory peaks. Correlation matrix showing pairwise correlations of the statistical significant peaks discriminating ADC and SqCC. The correlation matrix was calculated over the two classes and has a color scheme for better visualization ranging from white (poor correlation) to red (perfect correlation).
• Supplementary Figure 4 (.pdf, 35 KB) - Venn-Diagram of IHC markers in SqCC: IHC of the validation set (n=60) revealed considerable overlap of the well known SqCC marker CK5/6 (yellow) and the new markers HSP27 (red), CK5/6 (yellow) and CK15 (green).
• Supplementary Table 1 (.pdf, 360 KB) - On tissue protein identification: Specifications for database research and results from MS/MS experiments are highlighted.
• Supplementary Table 2 (.pdf, 96 KB) - Antibodies used in this study