Comparison of Three Sequencing Panels Used for the Assessment of Tumor Mutational Burden in NSCLC Reveals Low Comparability

• Published in: Journal of thoracic oncology 2020-09, Vol.15 (9), p.1535-1540
• Main Authors: Heeke, Simon, Benzaquen, Jonathan, Hofman, Véronique, Long-Mira, Elodie, Lespinet, Virginie, Bordone, Olivier, Marquette, Charles-Hugo, Delingette, Hervé, Ilié, Marius, Hofman, Paul
• Format: Article
• Language: English
• Subjects: Biomarkers, Tumor - genetics; Carcinoma, Non-Small-Cell Lung - drug therapy; Carcinoma, Non-Small-Cell Lung - genetics; Computer Science; FoundationOne; High-Throughput Nucleotide Sequencing; Humans; Lung Neoplasms - drug therapy; Lung Neoplasms - genetics; Medical Imaging; Mutation; NSCLC; Oncomine TML; QiaSeq; Tumor mutational burden
• ISSN: 1556-0864; 1556-1380
• DOI: 10.1016/j.jtho.2020.05.013

Tumor mutational burden (TMB) has been proposed as a novel predictive biomarker for the stratification of patients with NSCLC undergoing immune checkpoint inhibitor (ICI) treatment. The assessment of TMB has recently been established using large targeted sequencing panels, and numerous studies are ongoing to harmonize TMB assessment. “Correlation” or the coefficient of determination has generally been used to evaluate the association between different panels. We hypothesized that these metrics might overestimate the comparability, especially for lower TMB values. A total of 30 samples from patients with NSCLC undergoing ICI treatment were consecutively sequenced using the following three large, targeted sequencing panels: FoundationOne, Oncomine TML, and QiaSeq TMB. The TMB values were compared in the whole patient population and in a subset of patients in which the TMB assessed by FoundationOne was between 5 and 25 mutations/Mb. Prediction of durable clinical benefit (>6 mo with no progression) was assessed using receiver operator characteristics, and optimal cutoff values were calculated using the Youden J statistic. Correlation between the three targeted sequencing panels was strong in the whole patient population (R2 > 0.79) but was dramatically reduced in the subset of patients with TMB of 5 to 25 mutations/Mb. The agreement assessed using the Bland-Altman method was also very low. All panels were able to predict durable clinical benefit in the TMB-high population. Assessment of TMB using the three targeted sequencing panels was possible and predictive of response to ICI treatment, but correlation was an inappropriate measurement to assess the association between the respective panels.