Prognostic significance of peripheral monocyte count in patients with extranodal natural killer/T-cell lymphoma

This study was approved by the Institutional Review Board (IRB) of Sun Yat-Sen University Cancer Center. Study was performed in accord with the principles of the Declaration of Helsinki. All patients agreed to use their medical records for research.

Information regarding demographics, Eastern Cooperative Oncology Group (ECOG) performance status, physical examinations, systemic B symptoms, complete blood count (CBC), biochemical profiles, and serum lactate dehydrogenase (LDH) level were collected for analysis. The staging was based on the Ann Arbor staging system. Whole-body positron emission tomography/computed tomography (PET/CT) scans, CT scans or magnetic resonance imaging (MRI) scans of the involved sites, as well as thorax, abdomen, and pelvic CT scans, and bone marrow findings were used for staging. Patients with contiguous involvement extending to the adjacent tissues or organs were regarded as stage IE. Upper aerodigestive tract NK/T-cell lymphoma (UNKTL) and extra- upper aerodigestive tract NK/T-cell lymphoma (EUNKTL) were defined as previously described [8, 27]. The International Prognostic Index (IPI: stage, ECOG performance status, serum LDH, stage, extranodal sites) and the Korean Prognostic Index for NK/T-cell lymphoma with nasal disease (KPI: serum LDH level, B symptoms, stage, regional lymph nodes involvement) were evaluated as previously described [12, 28].

Patients received one of the following treatments: (i) chemotherapy followed by radiotherapy; (ii) chemotherapy alone; (iii) involved-field radiation alone; (iv) surgery followed by radiation; (v) best supportive care. The first-line chemotherapy regimens were: EPOCH (etoposide, doxorubicin, vincristine, cyclophosphamide, prednisone), CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) or CHOP-like therapy, alternating triple therapy (CHOP-B [cyclophosphamide, doxorubicin, vincristine, prednisone, bleomycin], IMVP-16 [ifosfamide, etoposide, methotrexate], and DHAP [dexamethasone, cytarabine, cisplatin]), GEMOX (gemcitabine, oxaliplatin) +/− L-asparaginase, and SMILE (dexamethasone, methotrexate, ifosfamide, L-asparaginase, etoposide). Radiotherapy for the involved-field was given in daily fractions of 2 Gy (five fractions per week) for a total of 36–68 Gy. The International Working Group Recommendations for Response Criteria for non-Hodgkin’s lymphoma were used to evaluate the treatment response [29].

Overall survival (OS) and progression-free survival (PFS) were assessed by the Kaplan–Meier method. OS was calculated from the date of diagnosis to the date of death from any cause, or date of the last follow-up. PFS was calculated from the date of diagnosis to the date of first lymphoma progression, death from any cause, or date of the last follow-up [29]. The absolute monocyte count (AMC) at diagnosis was analyzed as both continuous and dichotomized parameters. The optimal cut-off values of the AMC were determined using the receiver operating characteristics (ROC) method. In the ROC curve analysis, survival outcomes were dichotomized into death versus survival. The relationship between AMC (as a dichotomized variable) and clinical parameters was analyzed by Pearson's χ2 test. Survival curves were constructed using the Kaplan–Meier method. The prognostic influence of different parameters on survival was established by multivariate analysis using the Cox proportional hazards model. The two-tailed log-rank test method was used to determined statistical difference, and P value of less than 0.05 was regarded as statistically significant. Statistical analysis was carried out by SPSS 16.0 software.

The baseline characteristics of all 163 patients with ENKL are shown in Table 1. The median age at the time of diagnosis was 43 years (range: 17–80). There was a male predominance (113 men and 50 women). Eight-five patients (52.1%) presented systemic B symptoms, and 55 patients (33.7%) had elevated serum LDH level prior to treatment. Nearly 80% of the patients (128 cases) had localized disease (Ann Arbor stage: I-II). Upper aerodigestive tract NK/T-cell lymphoma (UNKTL) was diagnosed in 140 cases (85.9%) and extra-upper aerodigestive tract NK/T-cell lymphoma (EUNKTL) in 23 cases (14.1%). For those with EUNKTL, the most common sites of primary extranodal involvement were cutaneous or soft tissue (11 cases), followed by the gastrointestinal tract (5 cases). International Prognostic Index (IPI) assessment found that the majority of the patients (125 cases, 76.7%) were in the low-risk category (IPI score = 0–1). The Korean Prognostic Index (KPI) was assessable in 138 patients with nasal disease. In the KPI model, 84 patients (60.9%) had no or one adverse factor, and the remaining 54 patients (39.1%) had at least two adverse factors.

The AMC were obtained from the complete blood count (CBC) test prior to treatment. The median AMC at diagnosis, range, and the 25% and 75% quartiles were 0.60×10⁹/L, 0.05–2.10×10⁹/L, 0.44×10⁹/L, and 0.80×10⁹/L, respectively. The most discriminative cut-points of the AMC was 0.495×10⁹/L (area under the curve [AUC]: 0.630, 95% confidence interval: 0.534-0.726, P = 0.011), as determined by receiver operating characteristics (ROC) analysis. Therefore, AMC ≥0.50×10⁹/L which were close to the most discriminative cut-points, were selected as the optimal cut-off values. The most discriminative cut-points of the ALC were 1.145×10⁹/L (AUC: 0.601, 95% confidence interval: 0.512-0.690, P = 0.027), as determined by ROC analysis. Therefore, ALC ≤1.10×10⁹/L, which was close to the most discriminative cut-point, was selected as the optimal cut-off value. Patients with high AMC at diagnosis tended to have poorer performance status (P = 0.002) and IPI >1 (P = 0.022).

The initial treatment modalities were chemotherapy followed by radiotherapy (n = 102); chemotherapy alone (n = 50); surgery followed by radiation (n = 2); radiotherapy alone (n = 4); and supportive care (n = 5). The data on evaluation of treatment response to the initial therapy were available in 151 cases (92.6%). Complete remission (CR) was observed in 71 patients (47.0%). No statistical difference was found in the baseline clinical characteristics between patients with AMC ≥0.50×10⁹/L and those with AMC <0.50×10⁹/L.

The estimated overall survival (OS) and progression-free survival (PFS) at 5 years was 43.7% and 33.9%, respectively. Patients with higher AMC level (AMC ≥0.50×10⁹/L, n = 118) seemed to have shorter OS and PFS (OS: P = 0.023; PFS: P = 0.016; Figure 1A and B). Univariate analysis of AMC as a continuous variable also showed AMC to be related to inferior survival (OS: P <0.001; PFS: P = 0.049). As KPI and IPI are commonly used prognostic indexes for ENKL, they were included with AMC in multivariate analysis. As shown in Table 2 and Table 3, AMC retained their prognostic impact on OS and PFS in patients with ENKL, independent of IPI score. KPI score was not found to be an independent prognostic factor of survival in multivariate analysis.

Absolute lymphocyte count (ALC), as an indicator of immune status, was shown to be associated with OS and PFS in our previous study [19]. However, either AMC or ALC, as a single parameter, appeared to have limited ability to identify patients in the poor-risk category. Therefore, we combined the baseline AMC with the baseline ALC as dichotomized variables, to obtain a host immunity-related prognostic index of survival in patients with ENKL. Regarding the AMC/ALC prognostic index as a categorical variable, all 163 patients were stratified into the following three risk categories: group 1 (low risk), AMC <0.50×10⁹/L and ALC>1.10×10⁹/L; group 2 (intermediate risk), AMC ≥0.50×10⁹/L or ALC ≤1.10×10⁹/L; group 3 (high risk), AMC ≥0.50×10⁹/L and ALC ≤1.10×10⁹/L. The associations between patients’ clinical characteristics and AMC/ALC prognostic index were evaluated. Patients with high risk AMC/ALC index seemed to have a higher rate of B symptoms (P = 0.005). The OS and PFS based on the AMC/ALC prognostic index for patients with ENKL are shown in Figure 2. Univariate analysis showed that the AMC/ALC index allowed the discrimination of the three risk groups with a 5-year OS ranging from 74.2% to 23.2%, P = 0.001 and a 5-year PFS varying from 54.0% to 17.5%, P = 0.002. For patients who received chemotherapy followed by radiotherapy (102 cases), the three AMC/ALC index categories identified patients with significantly different survivals (OS: P = 0.006; PFS: P = 0.014; Figure 3). Moreover, when applied to the patients with low-risk IPI score (IPI = 0–1) or low-risk KPI score (KPI = 0–1), the AMC/ALC index identified significantly different prognostic categories in patients with low-risk KPI score (OS: P = 0.006; PFS: P <0.001; Figure 4) and patients with low-risk IPI score (OS: P = 0.014; PFS: P = 0.027; Figure 5). Further stratification by AMC/ALC index category identified 15 patients (17.9%) in the low-risk KPI score category and 25 patients (20%) in the low-risk IPI score category as high risk (5-year OS of 21.0% and 29.7%, respectively).