Abstract 3735
Background
We previously validated a clinical risk classification model for resected SqCLC by combining clinicopathological predictors to discriminate patients' (pts) prognosis (Pilotto JTO 2015). Here we investigate the molecular portrait of prognostic outliers to identify differentially expressed, potentially druggable molecular targets (AIRCMFAG project no. 14282).
Methods
On the basis of the published 3-class model, 176 and 46 pts with good and poor prognosis, respectively, were identified. Next Generation Sequencing (NGS) analysis (Ion Proton system, Ion Ampliseq custom panel) evaluating Somatic Mutations (SM) and Copy Number Alternations (CNA) of 44 genes was performed; RNA expression, immunohistochemistry (IHC), immunofluorescence (FISH) were performed on Tissue Micro-Arrays (TMA). Descriptive statistics was adopted; continuous variables were dichotomized according to AUC or medians.
Results
The distribution of relevant SM and CNA analysis of 60 pts according to prognosis (good: 27; poor: 33) is reported in the table.
| Analysis (NGS) | Gene | Good: 27 pts [%] | Poor: 33 pts [%] | p-value |
|---|---|---|---|---|
| SM | PI3KCA | 0 | 3 [9.1] | 0.24 |
| NOTCH1 | 2 [7.4] | 0 | 0.19 | |
| CUL3 | 2 [7.4] | 0 | 0.19 | |
| DDR2 | 3 [11.1] | 0 | 0.085 | |
| CDH10 | 4 [14.8] | 1 [3.0] | 0.16 | |
| CDH1 | 3 [11.1] | 1 [3.0] | 0.3 | |
| CNA Gains | RICTOR | 1 [3.7] | 9 [27.3] | 0.017 |
| SOX2 | 20 [74.1] | 17 [51.5] | 0.11 | |
| CNA Losses | CDKN2A | 6 [22.2] | 1 [3.0] | 0.038 |
| PTEN | 11 [40.7] | 17 [51.5] | 0.57 | |
| RB1 | 8 [29.6] | 17 [51.5] | 0.12 | |
| SMAD4 | 9 [33.3] | 19 [57.6] | 0.074 |
No significant differences in terms of phospho-mTOR and PD-L1 IHC expression were found in the 2 different prognostic subgroups. Patients with concurrent high PD1, SNAI, and Vimentin RNA expression were significantly more likely to be at poor prognosis (p = 0.003).
Conclusions
Although performed on a limited number of pts, the approach to comprehensively analyze DNA, RNA and proteins, using different methodologies, strengthens the clinically relevance of RICTOR/PI3K/mTOR signaling cascade activation in determining the poor prognosis of SqCLC. The possibility to inhibit this pathway with selective agents is currently under investigation in in vitro preclinical models.
Clinical trial identification
Legal entity responsible for the study
University of Verona, Verona, Italy
Funding
Associazione Italiana per la Ricerca sul Cancro (AIRC): MFAG Project 14282.
Disclosure
All authors have declared no conflicts of interest.