Search

e21627 Background: 15–40% of NSCLC adenocarcinoma patients harbor EGFR sensitizing mutations. Tyrosine kinase inhibitors (TKI) provide significant clinical benefit in this population, yet all patients will develop resistance. Liquid biopsy has been demonstrated to reliably identify tumor associated somatic EGFR mutations. Quantitative assessment of mutated EGFR driven tumors could potentially be used to monitor disease progression, to assess therapeutic response, and to identify resistance mechanisms. Methods: 106 longitudinal plasma samples from 16 NSCLC patients who were treated with osimertinib as either first line or second line therapy were collected. A series of plasma samples collected during treatment and at the time of disease progression were analyzed with the AVENIO ctDNA Surveillance kit*. Mutations at each time point were identified and reported by the AVENIO software v2.0*. The mutation profile of each patient at different timepoints along with the treatment journey was examined in combination with clinical outcome data. Results: EGFR sensitizing mutations were detected in all plasma samples by sequencing except in 3 cases. Patients responsive to anti-EGFR therapy showed a rapid decrease of EGFR driver mutations to non-detectable levels. Meanwhile, patients who had stable disease or rapid disease progression had stable or slightly decreasing ctDNA levels after receiving the treatment. One patient had a MET amplification, FBXL7 SNV, and EGFR T790M detected at the time of disease progression which were not detected at baseline. One patient had both EGFR L858R and T790M mutations. This patient progressed very quickly on erlotinib. Detection of the T790M mutation decreased upon osimertinib administration, however, the L858R mutation level stayed the same. TP53 mutations were elevated in 3 patients at the time of progression, and could potentially be related to anti-EGFR resistance. Conclusions: This study clearly demonstrated that liquid biopsy could identify resistance mutations beyond EGFR prior to clinical progression. Plasma samples collected prior to or at disease progression could facilitate identification of novel resistant mutations to TKI therapy. Further studies to demonstrate the clinical utility of serial blood EGFR testing in NSCLC management are necessary. *For Research Use Only. Not for use in diagnostic procedures.

e20701 Background: Quantifying circulating tumor DNA (ctDNA) is an emerging method to non-invasively assess treatment effect for solid tumors. Despite disease heterogeneity in NSCLC, we set out to identify a broadly applicable ctDNA-based method for disease monitoring. By employing plasma taken during early treatment cycles, we tested whether early response assessed by ctDNA level could predict treatment effect. Methods: Using a 197-gene NGS assay, the AVENIO ctDNA Surveillance Kit (For Research Use Only, not for use in diagnostic procedures), we measured ctDNA levels in post-treatment plasma samples based on variants identified at baseline. We used samples from an observational German Lung Cancer Multi-Marker Study. In a cohort of 83 stage IV lung adenocarcinoma treated with first-line chemo or chemoradiation therapies, we evaluated the association between survival and ctDNA levels in the first available post-treatment plasma sample (median number of days after start of treatment = 23). We used a ctDNA-based monitoring algorithm, and applied it to an independent set of 22 late stage lung squamous cell carcinoma (SCC) that also underwent chemo or chemoradiation therapies to further evaluate the algorithm in different histology subtypes. Results: We divided the 83 adenocarcinoma cohort into training (n = 53) and test (n = 30) sets. We found that subjects with longer progression free survival (PFS) had mean allele fraction (AF) < 1% in the training set. We applied the classifier to our validation set and found that subjects with mean AF < 1% had longer PFS (HR 0.35; 95% CI 0.12 - 0.93; log-rank P = 0.028) and overall survival (OS) (HR 0.29; 95% CI 0.09 - 0.89; log-rank P = 0.021). Using cutoffs identified in adenocarcinoma, we applied the same algorithms to the SCC cohort. Subjects with mean AF < 1% had longer PFS (HR 0.26; 95% CI 0.10 - 0.71; log-rank P = 0.005) and OS (HR 0.12; 95% CI 0.05 - 0.51; log-rank P = 0.001). Conclusions: Even in heterogeneous diseases such as NSCLC, changes in ctDNA levels in response to treatment may prove to be a valuable way of identifying subjects who may not benefit, which is earlier than current standard of care methods like computed tomography (CT) scan. Future prospective studies to confirm these results are warranted.

12077Background: Longitudinal ctDNA monitoring using variants identified in tissue biopsies is an emerging method for disease management. However, tissue biopsy is often inaccessible for many late ...

12088Background: Despite routine use of chemotherapy in advanced non-small-cell lung cancer (NSCLC) patients, the knowledge of optimal prognostic methods is still limited to imaging-based methods. ...

11115 Background: Tumor-stroma interaction plays a significant role for malignant growth. Results from prostate and breast cancer rodent models show cancerogenic properties of tumor-associated and genetically altered stromal cells (SC) when combined with initiated or normal epithelium (Olumi et al., Cancer Res 1999, Kuperwasser et al., PNAS 2004). However, data on the mechanisms and sequels of tumor-stroma interaction in lung cancer are scanty. Methods: Here, we analyzed the functional and molecular sequels of cross-talk between the non-small cell lung cancer (NSCLC) cell lines A549, H23, and H1703 and primary stromal cells (SC) derived from matched normal lung tissue and tumors of newly diagnosed NSCLC patients. Tumor cells were kept in a non-contact co-culture system with SC and analyzed for alterations in proliferation, colony formation, migration, adhesion, invasion, chemosensitivity and gene expression by Affymetrix HG U133 Plus 2.0 arrays. Results: Exposure to SC altered cellular functions and gene expression profiles related to tumor growth, metastasis and response to therapy. Each cell line showed individual alterations that were hierarchically governed by the (1) type of tumor cell, (2) the SC donor and his histology (3) and the local origin of the SC (normal lung tissue vs. tumor-associated). Conclusions: This in vitro model demonstrates an individual pattern of tumor-stroma interaction in NSCLC that is determined by both, the properties of the tumor cells and those of the stromal environment. Thus, biomarker programs in NSCLC should also consider the stromal compartment. No significant financial relationships to disclose.