787P - Association of neutrophil-to-lymphocyte ratio (NLR) with survival (OS) in metastatic castration-resistant prostate cancer (mCRPC) patients receivin...

Date 27 September 2014
Event ESMO 2014
Session Poster Display session
Topics Anticancer Agents
Prostate Cancer
Translational Research
Basic Principles in the Management and Treatment (of cancer)
Biological Therapy
Presenter David Lorente
Citation Annals of Oncology (2014) 25 (suppl_4): iv255-iv279. 10.1093/annonc/mdu336
Authors D. Lorente1, A.J. Templeton2, O. Sartor3, A. Bahl4, Z. Su5, J. Devin6, J.S. De Bono1
  • 1Drug Development Unit, The Institute of Cancer Research & The Royal Marsden NHS Trust, SM2 5PT - Sutton/GB
  • 2Department Of Medical Oncology, Kantonsspital St. Gallen, St. Gallen/CH
  • 3Department Of Medicine: Section Of Hematology & Medical Oncology And Department Of Urology, Tulane University, 70112 - New Orleans/US
  • 4Bristol Haematology And Oncology Centre, University Hospitals Bristol NHS Foundation Trust, Bristol/GB
  • 5Oncology, Sanofi Oncology, Cambridge/US
  • 6Biostatistics, Sanofi, Bridgewater/US



NLR is a marker of systemic inflammation that is associated with response to abiraterone acetate in mCRPC. CBZ + P provided survival benefit over MTX + P in post-docetaxel mCRPC patients in the TROPIC trial. This retrospective analysis of the TROPIC study explores the prognostic value of NLR and its association with response to CBZ.


Baseline (BL) NLR (absolute neutrophil count divided by absolute lymphocyte count) was calculated. Univariate analysis (UVA) of the association of NLR >5 with OS and response was performed, followed by inclusion in a multivariate (MVA) model (Cox regression, log-rank test) with prognostic factors previously identified in the TROPIC trial (rising PSA at BL, CBZ vs MTX, time from first hormonal treatment, BL ALP, pain, time from docetaxel chemotherapy).


742 men with mCRPC had available BL NLR data. BL NLR was ≤5 in 543 and >5 in 199. No significant BL differences were found between groups. A greater risk of death in patients with NLR >5 was observed in UVA (HR 1.72, 95% CI 1.441–2.062) irrespective of treatment arm. NLR >5 remained significantly associated with OS in MVA (P = 0.0017). Prognostic factors previously identified in the TROPIC study remained significant (all P-values ≤0.0083) after the inclusion of NLR in the model. OS was significantly longer in CBZ versus MTX-treated patients with NLR ≤5 (P = 0.0002) (Table). CBZ also provided OS benefit in patients with NLR >5 (P = 0.0503).


NLR is a valuable independent prognostic factor for OS in mCRPC. CBZ was associated with an OS benefit over MTX in both NLR ≤5 and NLR >5 subgroups. Data are consistent with previous analyses, which demonstrate CBZ superiority over MTX across subgroups. Dr Lorente is supported by a Spanish Society of Medical Oncology grant ‘Beca SEOM para la Investigacion Traslacional en el Extranjero’.

OS by NLR subgroup in the TROPIC trial (intent-to-treat)

Median, months (95% CI)
Baseline NLR MTX (n = 371) CBZ (n = 371) HR (95% CI) CBZ vs. MTX P-value*
NLR ≤5 (n = 543) 14.0 (12.7–15.6) 16.9 (15.2–18.5) 0.687 (0.564–0.837) 0.0002
NLR >5 (n = 199) 9.17 (7.33–11.3) 11.2 (9.23–13.6) 0.741 (0.548–1.002) 0.0503
Overall (N = 742) 12.8 (11.6–13.7) 15.2 (14.1–16.7) 0.703 (0.596–0.829) <0.0001

*Log rank test.


O. Sartor: has a consultant/advisory relationship with Sanofi; A. Bahl: has a consultant/advisory relationship with Sanofi, and has received honoraria and research funding from Sanofi; Z. Su: is a Medical Director at Sanofi, and has stock/other ownership interests at Sanofi; J. Devin: is an employee of Sanofi; J. de Bono: has received research funding from the Institute of Cancer Research. All other authors have declared no conflicts of interest.