W. H. Ward1, A. C. Esposito2, N. Goel1, K. J. Ruth3, E. R. Sigurdson1, J. E. Meyer5, C. S. Denlinger4, J. M. Farma1 1Fox Chase Cancer Center,Department Of Surgical Oncology,Philadelphia, PA, USA 2Temple University,Lewis Katz School Of Medicine,Philadelpha, PA, USA 3Fox Chase Cancer Center,Biostatistics And Bioinformatics Facility,Philadelphia, PA, USA 4Fox Chase Cancer Center,Department Of Hematology/Oncology,Philadelphia, PA, USA 5Fox Chase Cancer Center,Department Of Radiation Oncology,Philadelphia, PA, USA
Introduction:
Although advances in the multidisciplinary treatment of locally advanced rectal cancer have improved survival, there is variability in response to therapy. In addition to tumor biology, host factors and immunologic capacity may play a role. Given the morbidity of therapy and risk of recurrence, there is much interest in preoperative identification of predictive biomarkers. In colorectal cancer, recent data suggest the utility of the lymphocyte-to-monocyte ratio (LMR), neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) in predicting survival. The aim of our study was to examine these factors in our rectal cancer patients and determine whether any association exists between these ratios and overall survival.
Methods:
Using data from a prospectively maintained database at our tertiary referral center, a query was completed for all patients with clinical stage II to III rectal adenocarcinoma who underwent comprehensive treatment from 2002-2016. We included patients with full demographic, staging, treatment, and survival data who had a complete blood count collected prior to neoadjuvant chemoradiation (pre-CRT) and again prior to surgery (post-CRT). LMR, NLR, and PLR were calculated for the pre-CRT and post-CRT time points. Potential cutpoints associated with overall survival (OS) differences were determined using the maxstat R package, which identifies optimal cutpoints while controlling for repeat testing (p<0.05). Survival curves based on these cutpoints were compared using log-rank tests and were adjusted for age and stage using Cox regression.
Results:
A total of 146 patients were included. Cutpoints were significantly associated with OS for pre-CRT ratios but not for post-CRT ratios. Within the pre-treatment group, “low” (<2.86) LMR was associated with decreased OS (log-rank p=0.004, 5 year OS= 69% [95%CI 54%-80%]) compared to “high” (>2.86) LMR (5 year OS= 86% [95%CI 75%-93%]). In the same group, “high” NLR (>4.47) was associated with decreased OS (log-rank p<0.001), and “high” PLR (>203.6) was associated with decreased OS (log-rank p<0.001). With adjustment for age and final pathologic stage, the associations of NLR and PLR with OS retained their statistical significance (p=0.017 and p=0.005, respectively), and the association of LMR and OS had borderline statistical significance (p=0.075).
Conclusion:
If obtained prior to the start of neoadjuvant chemoradiation, LMR, NLR, and PLR values are accurate predictors of 5-year OS in patients with locally advanced rectal adenocarcinoma. Following the administration of neoadjuvant therapy, these ratios lose their predictive ability. Further confirmation of the value of these ratios in larger datasets will be important.