471P - Baseline full blood count parameters as predictors of response to PD-1 inhibition in advanced NSCLC

Date 18 December 2016
Event ESMO Asia 2016 Congress
Session Poster lunch
Topics Non-Small-Cell Lung Cancer, Metastatic
Presenter Megan Crumbaker
Citation Annals of Oncology (2016) 27 (suppl_9): ix139-ix156. 10.1093/annonc/mdw594
Authors M. Crumbaker, M. Carlino, R. Hui, A. Mersiades, H. Gurney, J. Todd, B. Gao, A. Nagrial
  • Medical Oncology, The Crown Princess Mary Cancer Centre, 2145 - Westmead/AU



Programmed cell death 1 (PD-1) inhibitors have demonstrated a survival advantage in advanced non-small cell lung cancer (NSCLC) with some patients achieving a durable response. However, a substantial portion of patients will not respond and, to date, a simple predictor of response has not been identified.


We conducted a retrospective review of all NSCLC patients treated with single-agent PD-1 inhibitors at our institution between June 2012 and June 2016. In addition to clinical data, pre-treatment full blood count results for cycles 1-4 were collected and correlated with response as assessed by RECIST 1.1 and irRECIST.


Sixty-two patients, with a median age of 67y (range 36-86y) were eligible. The majority (67.7%) were former smokers, 10 (16.1%) were non-smokers. Forty-eight (77.4%) were non-squamous, 4 had known EGFR mutations and 1 had an ALK translocation. Thirty-seven (59.7%) received pembrolizumab, the remainder nivolumab. Overall 20 (32.3%) were treated 1st line and 28 (45.2%) 2nd line. Fifty-seven (91.9%) had nodal/soft tissue involvement, 10 (16.1%) had previously treated CNS metastases and 8 (12.9%) had hepatic metastases at the start of anti PD-1 treatment. Twenty-one (33.9%) achieved a partial response. In a univariate analysis of age, gender, ECOG, ethnicity, smoking status, smoking pack years (SPY), BMI, histology, mutation status, sites of disease, line of treatment and baseline full blood count, only baseline absolute eosinophil count (AEC) and SPY were found to significantly correlate with response (p = 0.014, 0.0128 respectively). Median OS for AEC greater or less than 0.25 x109/L was 33m vs. 10m (p = 0.051). In the multivariate analysis using a logistic regression model, baseline AEC and SPY remained significant independent predictors of response; baseline AEC adjusted odds ratio (OR) 184.1 (95% CI 2.41-14088.27, p = 0.018), SPY OR 1.04 (95% CI 1.01-1.08, p = 0.016).


In this retrospective study, baseline AEC and smoking pack years predicted for response in advanced NSCLC patients treated with PD-1 inhibitors. There was a trend to improved survival with an elevated baseline AEC. Further analysis including additional patient numbers and PD-L1 status will be available at time of presentation.

Clinical trial indentification

Legal entity responsible for the study



Crown Princess Mary Cancer Centre, Westmead Hospital


M. Carlino: Advisory boards for MSD, BMS, Novartis, Amgen. R. Hui: Consultant - Merck Sharp & Dohme, Astra Zeneca Honoraria from above and Bristol-Myers Squibb. H. Gurney: Advisory boards for BMS, Pfizer. A. Nagrial: Speaking Fees from Bristol-Myers Squibb. All other authors have declared no conflicts of interest.