Prognostic value of quantitative ctDNA levels in non small cell lung cancer patients

// Mariano Provencio 1 , María Torrente 1 , Virginia Calvo 1 , David Pérez-Callejo 1 , Lourdes Gutiérrez 1 , Fernando Franco 1 , Clara Pérez-Barrios 2 , Miguel Barquín 2 , Ana Royuela 3 , Francisco García-García 4 , Coralia Bueno 5 , Aranzazu Garcia-Grande 6 , Carlos Camps 7 , Bartomeu Massuti 8 , Eduardo Sotomayor 9 and Atocha Romero 1, 2 1 Medical Oncology Department, Hospital Universitario Puerta de Hierro-Majadahonda, Majadahonda, Spain 2 Liquid Biopsy Laboratory, Biomedical Sciences Research Institute Puerta de Hierro-Majadahonda University Hospital, Majadahonda, Spain 3 Biostatistics Department, Biomedical Sciences Research Institute Puerta de Hierro-Majadahonda University Hospital, Majadahonda, Spain 4 Computational Genomics, Centro de Investigación Príncipe Felipe, Valencia, Spain 5 Medical Oncology Department, Hospital Infanta Cristina, Parla, Spain 6 Cytometry Unit, Biomedical Sciences Research Institute Puerta de Hierro-Majadahonda, Spain 7 Medical Oncology Department, Hospital General de Valencia, CIBERONC Network, Valencia, Spain 8 Medical Oncology Department, Hospital Universitario de Alicante, Alicante, Spain 9 George Washington Cancer Center, Washington D.C., United States of America Correspondence to: Mariano Provencio, email: mariano.provencio@salud.madrid.org Atocha Romero, email: atocha10@hotmail.com Keywords: ctDNA, non-small cell lung cancer, tyrosine kinase inhibitor, EGFR Received: June 15, 2017 Accepted: September 30, 2017 Published: November 16, 2017 ABSTRACT Background: Circulating tumor DNA (ctDNA) levels correlate well with tumor bulk. In this paper we aim to estimate the prognostic value of the dynamic quantification of ctDNA levels. Materials and Methods: A total of 251 serial plasma samples from 41 non-small-cell lung cancer patients who carried an activating EGFR mutation were analysed by digital PCR. For survival analysis, ctDNA levels were computed as a time-dependent covariate. Results: Dynamic ctDNA measurements had prognostic significance (hazard ratio for overall survival and progression free survival according to p.T790M mutant allele frequency; 2.676 and 2.71 respectively; P < 0.05). In the same way, patients with p.T790M-negative or unchanging or decreasing plasma levels of sensitizing EGFR mutation were 12 and 4.8 times more likely to maintain response or stable disease, respectively, than patients in which the opposite occurred ( P < 0.05).On average, the p.T790M mutation was detected in plasma 51 days before the assessment of progression disease by CT-scan. Finally, ctDNA outperformed CTCs for assessing tumor progression ( P = 0.021). Conclusions: The appearance or increase in a unit of the p.T790M allele frequency almost triples the risk of death and progression. This information can be used to design clinical trials aiming to estimate whether T790M positive patients should start second line treatment based on molecular data rather than imaging data.