Abstract 1045P
Background
An elevated derived neutrophil-to-lymphocyte ratio (dNLR) has been correlated with worse outcomes across several malignancies and is also a known prognostic factor for patients on checkpoint inhibitors (CPI). However, the dNLR cut-off ranges from 3 to 5 and it’s unclear if dNLR variation while on CPI affects patient outcomes.
Methods
Prospectively maintained pharmacy databases in our regional cancer centre were interrogated for patients who were prescribed CPI in the advanced setting between January 2017 - 2020. Baseline patient characteristics were collated including dNLR {neutrophil count/ [white blood cell count – neutrophil count]} upon CPI commencement and repeated six weeks later.
Results
There were 102 patients with a median age of 69 years (range 30 - 88). Median follow-up was 7.8 months (Range 0.3 – 55.4 months). Seventy-three patients received prior systemic therapy (0 -3 lines). The median PFS for patients with a baseline dNLR > 3 when compared with a dNLR of ≤3 was 4.1 vs 14.2 months (p=0.001) and a similar comparison of the median OS was 6.4 vs 30.2 months (p=0.001). There were 10 patients who died before the six week dNLR change could be captured. Table: 1045P
Patient characteristics
| Cancer | n | Previous treatment | Pembrolizumab | Nivolumab | Atezolizumab | dNLR > 3 | dNLR ≤3 |
| 102 | 73 | 61 | 26 | 15 | 22 | 80 | |
| Melanoma | 46 | 3 | 41 | 51 | 0 | 6 | 40 |
| Lung | 42 | 28 | 18 | 9 | 15 | 12 | 30 |
| Squamous | 20 | 13 | 9 | 3 | 8 | 7 | 13 |
| Adeno | 22 | 15 | 9 | 6 | 7 | 5 | 17 |
| Renal & Bladder | 6 | 6 | 1 | 5 | 0 | 0 | 6 |
| Head & Neck | 6 | 6 | 0 | 6 | 0 | 3 | 3 |
| Others | 2 | 2 | 1 | 1 | 0 | 1 | 1 |
1 Combination with ipilimumab. Patients in the dNLR > 3 cohort with a higher dNLR six weeks later had a significantly reduced median PFS when compared with decreased dNLR (3.6 vs 12.2 months, p=0.047). Similar analysis identified a significantly reduced median OS when higher dNLR was compared with decreased dNLR; 5.5 vs 16.3 months, p=0.008. In the dNLR ≤3 cohort, the two median PFS were statistically different (8.4 months vs not reached, p = 0.046). A similar analysis of the median OS identified a difference (24.2 months vs not reached) which approached significance (p=0.061).
Conclusions
We have a diverse patient population in our study and CPI reponse is likely to be multifactorial. However, patients with lower baseline dNLR have improved outcomes with CPI and therefore a dNLR < 3 could be used to select patients for CPI therapy.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.