• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • Moreover it is still uncertain if


    Moreover, it is still uncertain if the levels of in circulating tumor DNA (ctDNA) at osimertinib baseline and during treatment could predict the pattern of resistance. Oxnard et al. observed that patients who lost T790M had only a slightly lower relative T790M AF before osimertinib than patients who maintained T790M; moreover, the evaluation of levels demonstrated a larger decrease for patients who maintained T790M vs. T790M loss [11]. Therefore, the present study analysed the cfDNA of advanced NSCLC patients treated with osimertinib to confirm the correlation of PSB 1115 levels of and T790M with clinical outcome, and to predict the pattern of resistance at tumor progression.
    Materials and methods
    Discussion Our study aimed to investigate the role of plasma monitoring during osimertinib therapy in patients with advanced T790M-positive NSCLC resistant to prior EGFR-TKI. Forty patients were enrolled and treatment outcomes were prospectively recorded whilst plasma was collected before and during treatment until progression of disease. Overall, outcomes of our study population reflect data reported in literature in this setting [5]; response rate was slightly inferior to what expected due to the heterogeneity of resistance mechanisms (i.e. SCLC transformation) likely present together with T790M mutation, as previously reported [8]. Plasma samples were analysed using both ddPCR and Therascreen, allowing us to obtain several parameters with quantitative or qualitative significance to describe EGFR mutational status. The most reliable predictive factor was expressed at baseline. In particular, the quantification of the amount of activating mutations in copies/mL, performed with ddPCR, confirmed that higher levels are associated with poor outcome. This observation was strengthened when analyses were conducted considering median AF. Other authors described inverse relationship between activating mutations levels and outcome [17]. Recently, our group published results from a cohort of patients treated with osimertinib with and resistance mutations detectable in plasma and documented a significant improvement in clinical outcome in patients with lower levels of AF [9]. The present new analysis included all patients enrolled in the ASTRIS trial at the University Hospital of Parma, and allowed also a comparison between shedders vs. non-shedders, highlighting a trend favouring patients with undetectable mutations in plasma. The real-time PCR approach does not provide quantitative information about ctDNA and its sensitivity is inferior to ddPCR: in our study population, the median AF of activating mutations determined by ddPCR was 3.5% in patients resulted negative by real-time PCR vs. median AF 20.4% in patients positive by real-time PCR (pā€‰=ā€‰0.001). Therascreen allowed a selection of patients with highly-positive plasma samples with a statistically significant poorer disease control and survival, with respect to negative patients. Our findings are confirmed in other reports evaluating the prognostic impact of cfDNA plasma positivity [10,18]. A possible explanation could be found in the correlation between ctDNA quantification and tumor burden, historically considered a clinical feature of poor prognosis [19]. Taken together, the previous and present studies reinforce the prognostic significance of plasmatic mutational analysis in patients candidate to osimertinib, deserving further in depth analysis especially in first-line setting perspective. Our study underlines the poor reliability of T790M quantification by ddPCR for prognostic purpose and similar results have been reported by other authors [9,11,17]. A possible explanation can be argued considering the heterogeneous context determined by different resistant clones selected during first-line EGFR-TKI, that share the as molecular driver. In fact, we previously documented that the amount of T790M mutation is notably inferior than , suggesting that T790M determination can not reflect the molecular complexity of different sub-clones [20]. Moreover, in this study we observed that patients with SCLC transformation had a disappearance of T790M in plasma despite disease progression, again confirming that T790M level cannot represent a reliable marker for response. Our results are confirmed by current literature, which underscores the weakness of prognostic role of T790M alone [9,21,22].