Author’s Reply

Abstract

We thank Dr. Mezquita and colleagues for their interest in our study and their comments.1Mezquita L. Auclin E. Besse B. Letter to the editor.J Thorac Oncol. 2019; 14: e209Scopus (5) Google Scholar Mezquita et al.1Mezquita L. Auclin E. Besse B. Letter to the editor.J Thorac Oncol. 2019; 14: e209Scopus (5) Google Scholar enquire whether lung immune prognostic index (LIPI) remains a prognostic factor if the analysis was restricted to patients that received chemotherapy and did not subsequently use immunotherapy during follow-up. In the pooled OAK and POPLAR studies, there were 687 patients who were treated with docetaxel and had a known LIPI value. As reported for this cohort, compared to the good prognosis LIPI group the hazard ratios for overall survival (OS) were 1.55 (95% confidence interval [CI]: 1.28–1.89) and 2.84 (95% CI: 2.10–3.83) for the intermediate and poor LIPI groups, respectively.2Sorich M.J. Rowland A. Karapetis C.S. et al.Evaluation of the lung immune prognostic index for prediction of survival and response in patients treated with atezolizumab for non-small cell lung cancer: pooled analysis of clinical trials.J Thorac Oncol. 2019; https://doi.org/10.1016/j.jtho.2019.04.006Abstract Full Text Full Text PDF Scopus (60) Google Scholar In this cohort, 135 (19.7%) patients were recorded to have subsequent treatment with an immunotherapy agent, predominantly nivolumab. Based on analysis restricted to the 552 patients with no recorded use of immunotherapy during follow-up, LIPI was significantly associated with OS (p < 0.001). Specifically, the hazard ratios (reference: good LIPI group) for OS were 1.66 (95% CI: 1.34–2.05) and 3.07 (95% CI: 2.23–4.24) for the intermediate and poor LIPI groups, respectively. Mezquita et al.1Mezquita L. Auclin E. Besse B. Letter to the editor.J Thorac Oncol. 2019; 14: e209Scopus (5) Google Scholar also enquire regarding the power to detect treatment effect modification by LIPI groups in the pooled analysis of the OAK and POPLAR trials. We concur that individual clinical trials are typically not well powered to detect treatment effect modification, particularly for small subgroups such as the poor LIPI group. The lack of statistical significance for the test of LIPI-by-treatment statistical interaction indicates that random error alone may plausibly result in the differences in OS treatment effect observed between LIPI groups — hence the need to cautiously interpret any apparent variations in treatment benefit between LIPI groups. However, the lack of statistical significance does not rule out that treatment benefit from immunotherapy may differ by LIPI group. Evaluation of LIPI in other randomised clinical trials of immunotherapy for NSCLC would be a valuable direction of future research to improve the power to detect potential differences in treatment benefit for the poor LIPI group. Letter to the Editor about Sorich et al.Journal of Thoracic OncologyVol. 14Issue 9PreviewWe have read with great interest the recently published article by Sorich et al.1 The authors assessed the prognostic value of the Lung Immune Prognostic Index (LIPI) in a pooled cohort of patients enrolled in a clinical trial of atezolizumab for advanced NSCLC. The LIPI was associated with overall survival (OS) and progression-free survival in patients receiving atezolizumab (n = 1489); similar prognostic value was observed in the chemotherapy (CT) arm with docetaxel (n = 687). Full-Text PDF Open Archive