Pre-treatment neutrophil/lymphocyte ratio and platelet/ lymphocyte ratio are prognostic of progression in early stage classical Hodgkin lymphoma

Transcription

1 short report Pre-treatment neutrophil/lymphocyte ratio and platelet/ lymphocyte ratio are prognostic of progression in early stage classical Hodgkin lymphoma Jay P. Reddy, 1 Mike Hernandez, 2 Jillian R. Gunther, 1 Bouthaina S. Dabaja, 1 Geoffrey V. Martin, 1 Wen Jiang, 1 Mani Akhtari, 1,3 Pamela K. Allen, 1 Bradley J. Atkinson, 5 Grace L. Smith, 1 Chelsea C. Pinnix, 1 Sarah A. Milgrom, 1 Zeinab Abou Yehia, 1 Eleanor M. Osborne, 1 Yasuhiro Oki, 4 Hun Lee, 4 Fredrick Hagemeister 4 and Michelle A. Fanale 4 1 Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 2 Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, 3 Department of Radiation Oncology, The University of Texas Medical Branch at Galveston, Galveston,, 4 Department of Lymphoma Medical Oncology, Summary To determine whether pre-treatment neutrophil/lymphocyte (NLR) or platelet/lymphocyte ratios (PLR) are predictive for progression in early-stage classical Hodgkin lymphoma (chl), we derived NLR and PLR values for 338 stage I/II chl patients and appropriate cut-off point values to define progression. Two-year freedom from progression (FFP) for patients with NLR 64 was 822% vs. 957% with NLR <64 (P < 0001). Similarly, 2-year FFP was 843% for patients with PLR 2662 vs. 961% with PLR <2662 (P = 0003). On univariate analysis, both NLR and PLR were significantly associated with worse FFP (P = 0001). On multivariate analysis, PLR remained a significant, independent prognostic factor (P < 0001). Keywords: neutrophil, platelet, lymphocyte, relapse, refractory, Hodgkin lymphoma. The University of Texas MD Anderson Cancer Center, Houston, and 5 Department of Clinical Pharmacy Services University of Texas MD Anderson Cancer Center, Houston, TX, USA Received 2 August 2017; accepted for publication 11 October 2017 Correspondence: Michelle A. Fanale, Department of Lymphoma/Myeloma Medical Oncology, Unit 0429, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA. mfanale@mdanderson.org In the era of treatment de-escalation for stage I and II classical Hodgkin lymphoma (chl), appropriate risk-stratification of those patients with high relapse risk is critical to identify those for whom treatment de-escalation is inappropriate. Positron emission tomography (PET)-based responseadapted therapy is promising; however, the recent UK RAPID and EORTC/LYSA/FIL H10 trials demonstrated the inferiority of omitting radiation in early stage patients who develop a negative PET after ABVD (doxorubicin, bleomycin, vincristine; dacarbazine) chemotherapy (Radford et al, 2015; Andre et al, 2017), suggesting that PET alone may not be sufficient risk stratification. Consequently, there is interest in identifying simple, low cost laboratory assessments that improve current risk stratification approaches in early-stage chl. Recently, the neutrophil/lymphocyte (NLR) and platelet/lymphocyte (PLR) ratios have been demonstrated to be prognostic in some solid malignancies. NLR has been shown to be a prognostic factor in diffuse large B-cell lymphoma (DLBCL) (Porrata et al, 2010; Ho et al, 2015; Keam et al, 2015), while PLR was recently shown to be prognostic in Natural Killer/T-cell lymphoma (Wang et al, 2014). However, few studies have examined the potential prognostic value of NLR and PLR in chl. The goal of this study was to ª 2017 John Wiley & Sons Ltd First published online 21 December 2017 doi: /bjh.15054

2 examine our institutional experience to determine whether NLR and PLR can serve as prognostic indicators for early stage chl and should be included in future risk-stratification approaches. Methods and materials Study cohort and NLR/PLR assessment This retrospective analysis of early-stage chl patients was approved by the MD Anderson institutional review board. We identified 338 consecutively treated patients with stage I or II chl treated between 2002 and For study inclusion, patients must have had pre-treatment complete blood count (CBC) with differential, from which absolute neutrophil (ANC), lymphocyte (ALC), and platelet counts prior to chemotherapy were acquired. These values were subsequently used to calculate NLR and PLR values (Roxburgh & McMillan, 2010; Proctor et al, 2011). The NLR was derived by dividing the ANC by the ALC at diagnosis. The PLR was similarly derived by dividing the absolute platelet count by the ALC at diagnosis. Endpoints and statistical analysis Optimal cut-off points for both NLR and PLR were determined by both receiver operative curve (ROC) analysis and the method of Liu (2012) (Data S1). The primary endpoint was freedom from progression (FFP), which was defined as relapsed or refractory disease. Relapse was defined as disease recurrence found on restaging PET/computed tomography (CT) following a minimum disease-free interval of 3 months. Refractory disease was designated by the treating medical oncologist following a restaging PET/CT within 3 months following the end of treatment. Time to event was calculated from date of pathological diagnosis to the date of progression. Patients who survived were censored at their last follow-up date. Cox regression was used to identify variables associated with FFP. All P-values were two-sided, with values of 005 considered to be significant. Results and discussion Baseline characteristics of the 338 stage I/II patient cohort are shown in Table S1. Median age at diagnosis was 32 years [interquartile range (IQR) 24 42] with a median follow-up Table I. Pre-treatment characteristics stratified on NLR and PLR values. NLR strata PLR strata <64 64 < N = 239 (%) N = 99 (%) P-value N = 214 (%) N = 124 (%) P-value Age, years Median (IQR) 32 (25 44) 30 (24 41) 0246* 32 (24 43) 325 (24 41) 0920* Sex Female 134 (561) 52 (525) (523) 74 (597) 0191 Male 105 (439) 47 (475) 102 (477) 50 (403) Race White 187 (782) 70 (707) (790) 88 (710) 0051 Hispanic 31 (130) 10 (101) 27 (126) 14 (113) Black 7 (29) 10 (101) 6 (28) 11 (89) Other 14 (59) 9 (91) 12 (56) 11 (89) Ann Arbor Stage I 48 (201) 4 (40) < (192) 11 (89) 0012 II 191 (799) 95 (960) 173 (808) 113 (911) GHSG class Early Favourable 82 (343) 9 (91) (350) 16 (129) 0678 Early Unfavourable 148 (619) 76 (768) 132 (617) 92 (742) Advanced 9 (38) 14 (141) 7 (33) 16 (129) B Symptoms No 194 (812) 59 (596) < (808) 80 (645) 0001 Yes 45 (188) 40 (404) 41 (192) 44 (355) Bulky No 172 (720) 39 (394) < (766) 47 (379) <0001 Yes 67 (280) 60 (606) 50 (234) 77 (621) GHSG, German Hodgkin Study Group; IQR, inter-quartile range; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio. *P-value based on a Wilcoxon rank-sum test. P-value based on a chi-square test. P-value based on a Cochran Armitage trend test. 546 ª 2017 John Wiley & Sons Ltd

3 of 5 years (IQR 29 78). Most patients (n = 286, 846%) had Ann Arbor stage II disease and were classified as earlyunfavourable (n = 224, 663%) using the German Hodgkin Study Group criteria (Engert et al, 2010). Two-year FFP and OS for the entire study cohort was 917% and 968%, respectively. The median NLR was 41 (IQR 25 71), and the median PLR was 2142 (IQR , Table S1). Both NLR and PLR were significantly associated with disease progression on simple logistic regression (P < 001). Both ROC analysis and the method of Liu (2012) were in agreement in determining optimal cut-off points for NLR and PLR that predicts for disease progression (Data S1). The cut-off point for NLR was 64 [95% confidence interval (CI): ] and 2662 (95% CI: ) for PLR. Upon dichotomizing the cohort based on these cut-off points, we confirmed that elevated NLR and PLR values were associated with worse FFP. Two-year FFP for patients with an NLR 64 was 822% compared to 957% in patients with an NLR < 64 (P < 0001, Figure S1a). Similarly, 2-year FFP was 843% for patients with a PLR 2662 and 961% in patients with a PLR < 2662 (P = 0003, Figure S1b). We next examined whether NLR and PLR values stratified on the above cut-off points are significantly associated with clinical factors (Table I). Patients with elevated NLR and PLR levels were more likely to have adverse pre-treatment factors, e.g. stage II disease, B symptoms, or bulky disease (Table I), suggesting that NLR and PLR can serve as biomarkers for adverse features. Furthermore, we found that both NLR [Hazard ratio (HR): 119, 95% CI: , P = 0001) and PLR (HR: 105, 95% CI: , P < 0001) were significantly associated with worse FFP on univariate analysis (Table II). We next performed multivariate modelling to determine whether NLR and PLR can serve as independent factors associated with worse FFP. Given the significant multi-collinearity between these assessments in that both incorporate ALC values (Pearson coefficient = 082), both NLR and PLR were not retained in the final model. In the final model, only bulky disease (HR: 420, 95% CI: , P < 0001) and PLR (HR: 104, 95% CI: , P = 0017) continued to be significantly associated with worse FFP (Table II). NLR did not remain significantly associated with FFP. Collectively, these findings suggest that while both NLR and PLR are significantly associated with worse FFP on univariate analysis, only PLR is an independent prognostic factor of relapse or refractory disease. We next sought to determine whether post-treatment NLR and PLR values could reliably predict relapse. After excluding patients with refractory disease and those without laboratory values following frontline therapy, 278 patients were available for analysis. However, there were very few relapse events in this subgroup (n = 12, 43%). On simple logistic regression, neither post-treatment NLR nor PLR were significantly associated with relapse (P = 061 for NLR and P = 075 for PLR). Table II. Cox proportional hazards regression analysis of factors associated with FFP. Univariable analysis HR (95% CI) P- value Multivariable analysis* HR (95% CI) P- value Age 5 year 100 (088, 113) 0992 Gender Male Female 089 (043, 181) 0744 Race White Hispanic 073 (022, 243) 0613 Black 118 (028, 500) 0818 Other 044 (006, 326) 0423 Ann Arbor Stage I II 572 (078, 420) 0086 GHSG class Early Favourable Early 321 (096, 107) 0057 Unfavourable Advanced 681 (163, 285) 0009 B Symptoms No Yes 197 (096, 406) 0066 Bulky No Yes 513 (229, 115) < (182, 971) 0001 NLR 5 unit 119 (107, 132) 0001 PLR 50 unit 105 (103, 108) < (101, 106) % CI, 95% confidence interval; FFP, freedom from progression; GHSG, German Hodgkin Study Group; HR, Hazard ratio; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio. *Backward elimination was used to isolate predictors in the multivariable model. Both NLR and PLR were not retained in the final model due multicollinearity (Pearson s correlation coefficient = 082). Significant P values are in bold. These findings have potentially important clinical and research implications as future endeavours seek to further de-intensity treatment without compromising excellent outcomes in chl. This is the first study to demonstrate the prognostic value of both pre-treatment NLR and PLR in predicting FFP in early-stage chl, a critical population for whom treatment de-escalation in selected patients may be appropriate. Therefore, further investigation and validation in additional data sets is warranted. There is accumulating evidence that NLR and PLR are associated worse prognosis in multiple solid malignancies (Guthrie ª 2017 John Wiley & Sons Ltd 547

4 et al, 2013; Templeton et al, 2014) and non-hodgkin lymphoma (Porrata et al, 2010; Keam et al, 2015) (Wang et al, 2014). Recently, two studies have examined the influence of NLR on outcomes in HL. Koh et al (2012) found that NLR values >43 were associated with worse OS in advanced stage chl subgroups but not in early stage chl patients. In contrast Marcheselli et al (2016) found that NLR >6 was associated with worse PFS and OS in both early and advanced stage chl. Importantly, our study differs from these prior analyses in that it focuses on early stage patients, examines both NLR and PLR, and finds that PLR, rather than NLR, is an independent prognostic factor. Our study has several important limitations. First, this study is retrospective and, therefore, subject to inherent limitations and biases. Second, our study population had only 31 events of disease progression, which limited the complexity of our final multivariate model. Consequently, validation of the utility of NLR and PLR in a second data set is needed. Despite these limitations, our experience is unique in that it is a relatively large single institutional experience focusing on early stage chl and factors associated with disease progression. In conclusion, pre-treatment NLR and PLR are significantly associated with increased risk of relapse or refractory disease in stage I and II chl. Elevated levels of these parameters are associated with worse FFP on univariate analysis. However, only PLR remained significantly associated with worse FFP after correcting for the presence of other factors on multivariate analysis. Pre-treatment NLR and PLR represent low-cost, effective supplements to future risk stratification schema that merit further investigation. Author contributions JPR, JRG, GLS, BSD, CCP, MAF: conception and design. JPR, GVM, JRG, WJ, MA, BJA, ZAY, EMO: collection and assembly of data. JPR, MH, BSD, PKA, GLS, MAF: data analysis and interpretation. JPR: manuscript writing. JPR, MH, JRG, BSD, GVM, WJ, MA, PKA, BJA, GLS, CCP, SAM, ZAY, EMO, YO, HL, FH, MAF: final approval of manuscript. Conflict of interest statement The authors declare no conflicts of interest. Supporting Information Additional Supporting Information may be found in the online version of this article: Data S1. Supplemental methods. Fig S1. (A) Freedom From Progression by NLR Strata (P < 0.001). (B) Freedom From Progression by PLR Strata (P = 0.003). Table S1. Baseline patient and treatment characteristics (n = 338) s Andre, M.P.E., Girinsky, T., Federico, M., Reman, O., Fortpied, C., Gotti, M., Casasnovas, O., Brice, P., van der Maazen, R., Re, A., Edeline, V., Ferme, C., van Imhoff, G., Merli, F., Bouabdallah, R., Sebban, C., Specht, L., Stamatoullas, A., Delarue, R., Fiaccadori, V., Bellei, M., Raveloarivahy, T., Versari, A., Hutchings, M., Meignan, M. & Raemaekers, J. (2017) Early positron emission tomography response-adapted treatment in stage I and II Hodgkin lymphoma: final results of the randomized EORTC/LYSA/ FIL H10 Trial. Journal of Clinical Oncology, 35, Engert, A., Plutschow, A., Eich, H.T., Lohri, A., Dorken, B., Borchmann, P., Berger, B., Greil, R., Willborn, K.C., Wilhelm, M., Debus, J., Eble, M.J., Sokler, M., Ho, A., Rank, A., Ganser, A., Trumper, L., Bokemeyer, C., Kirchner, H., Schubert, J., Kral, Z., Fuchs, M., Muller-Hermelink, H.K., Muller, R.P. & Diehl, V. (2010) Reduced treatment intensity in patients with early-stage Hodgkin s lymphoma. New England Journal of Medicine, 363, Guthrie, G.J., Charles, K.A., Roxburgh, C.S., Horgan, P.G., McMillan, D.C. & Clarke, S.J. (2013) The systemic inflammation-based neutrophillymphocyte ratio: experience in patients with cancer. Critical Reviews in Oncology Hematology, 88, Ho, C.L., Lu, C.S., Chen, J.H., Chen, Y.G., Huang, T.C. & Wu, Y.Y. (2015) Neutrophil/Lymphocyte ratio, Lymphocyte/Monocyte ratio, and Absolute Lymphocyte Count/Absolute monocyte count prognostic score in diffuse large B-cell lymphoma: useful prognostic tools in the rituximab era. Medicine (Baltimore), 94, e993. Keam, B., Ha, H., Kim, T.M., Jeon, Y.K., Lee, S.H., Kim, D.W., Kim, C.W. & Heo, D.S. (2015) Neutrophil to lymphocyte ratio improves prognostic prediction of International Prognostic Index for patients with diffuse large B-cell lymphoma treated with rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone. Leukaemia & Lymphoma, 56, Koh, Y.W., Kang, H.J., Park, C., Yoon, D.H., Kim, S., Suh, C., Kim, J.E., Kim, C.W. & Huh, J. (2012) Prognostic significance of the ratio of absolute neutrophil count to absolute lymphocyte count in classic Hodgkin lymphoma. American Journal of Clinical Pathology, 138, Liu, X. (2012) Classification accuracy and cut point selection. Statistics in Medicine, 31, Marcheselli, R., Bari, A., Tadmor, T., Marcheselli, L., Cox, M.C., Pozzi, S., Ferrari, A., Baldini, L., Gobbi, P., Aviv, A., Pugliese, G., Federico, M., Polliack, A. & Sacchi, S. (2016) Neutrophil-lymphocyte ratio at diagnosis is an independent prognostic factor in patients with nodular sclerosis Hodgkin lymphoma: results of a large multicenter study involving 990 patients. Hematological Oncology, 2016, /hon.2359 Porrata, L.F., Ristow, K., Habermann, T., Inwards, D.J., Micallef, I.N. & Markovic, S.N. (2010) Predicting survival for diffuse large B-cell lymphoma patients using baseline neutrophil/ lymphocyte ratio. American Journal of Hematology, 85, Proctor, M.J., Morrison, D.S., Talwar, D., Balmer, S.M., Fletcher, C.D., O Reilly, D.S., Foulis, A.K., Horgan, P.G. & McMillan, D.C. (2011) A comparison of inflammation-based prognostic scores in patients with cancer. A Glasgow Inflammation Outcome Study. European Journal of Cancer, 47, Radford, J., Illidge, T., Counsell, N., Hancock, B., Pettengell, R., Johnson, P., Wimperis, J., Culligan, D., Popova, B., Smith, P., McMillan, A., Brownell, A., Kruger, A., Lister, A., Hoskin, P., O Doherty, M. & Barrington, S. (2015) Results of a trial of PET-directed therapy for early-stage Hodgkin s lymphoma. New England Journal of Medicine, 372, Roxburgh, C.S. & McMillan, D.C. (2010) Role of systemic inflammatory response in predicting survival in patients with primary operable cancer. Future Oncology, 6, Templeton, A.J., Ace, O., McNamara, M.G., Al- Mubarak, M., Vera-Badillo, F.E., Hermanns, 548 ª 2017 John Wiley & Sons Ltd

5 T., Seruga, B., Ocana, A., Tannock, I.F. & Amir, E. (2014) Prognostic role of platelet to lymphocyte ratio in solid tumors: a systematic review and meta-analysis. Cancer Epidemiology, Biomarkers & Prevention, 23, Wang, K.F., Chang, B.Y., Chen, X.Q., Liu, P.P., Wuxiao, Z.J., Wang, Z.H., Li, S., Jiang, W.Q. & Xia, Z.J. (2014) A prognostic model based on pretreatment platelet lymphocyte ratio for stage IE/IIE upper aerodigestive tract extranodal NK/T cell lymphoma, nasal type. Medical Oncology, 31, 318. ª 2017 John Wiley & Sons Ltd 549