Neuroendocrine tumors (NET) occur throughout the body but are not commonly suspected in breast, prostate or colorectal cancers. Morphologic evidence of neuroendocrine differentiation may prompt testing for neuroendocrine markers, but such testing is not routine. Now that NET-specific therapies can increase patient survival, distinction of NET from non-NET is essential. We developed a multiplexed mass spectrometry-based screening tool to measure tumor expression of 3 common neuroendocrine (NE) proteins. We tested and validated this NET panel in clinical biopsies of NET and non-NET.
Formalin-fixed-paraffin-embedded (FFPE) NET were microdissected and solubilized to tryptic peptides for mass spectrometric analysis using selected reaction monitoring. Synthetic versions of chromogranin A (CHGA), synaptophysin (SYP) and CD56 peptides were used to develop the assay. Using mass spectrometry with stable isotope labeled internal standards, these 3 NE proteins were quantitated in FFPE tumor biopsies of NET and non-NET.
In the test set, 20 of 20 previously diagnosed NET (of the lung and gastroenteropancreatic tract) expressed ≥2 of the 3 NE protein markers (positive predictive value=100%), and 47 of 50 non-NET (non-small cell lung cancer) expressed none of the markers, with only one sample expressing ≥2 markers (negative predictive value=98%.). NET positivity was therefore defined as expression of ≥ 2 markers. In a validation set of 16 NET, the proteomic panel confirmed 13 cases. Of the 3 discordant cases, one of these was a small-cell lung cancer with mixed NE and squamous histology. When used to screen 614 consecutive clinical samples of multiple tumor types, the panel found 16 tumors that unexpectedly expressed ≥2 NE markers. Upon pathology review, 5 of these were confirmed as NET, thus revealing new treatment options for 5 patients.
A mass spectrometry-based screening tool can identify NET with sensitivity and specificity similar to that of immunohistochemistry. Such proteomic testing can identify NE proteins simultaneously with dozens of therapeutically relevant biomarkers (eg, HER2, EGFR) to inform treatment decision making without the need for additional FFPE sections.
Clinical trial identification
Legal entity responsible for the study
S. Thyparambil, E. An, S. Sellappan, E.R. Wertheimer, F. Cecchi, R. Heaton, T. Hembrough: Employee: NantOmics. All other authors have declared no conflicts of interest.