Abstract
Objective
Extra nodal natural killer/T-cell lymphoma (ENKTL), nasal type is a rare and highly aggressive type of non-Hodgkin's lymphoma (NHL) commonly presented in the nasal cavity or lymphatic system. However, the common causes of mortality in ENKTL remain unclear. We conducted a retrospective population-based cohort study to elucidate the different causes of mortality in ENKTL and illustrate the main causal and associated risk factors leading to death.
Methods
The study included patients diagnosed with ENKTL from 1987 to 2014 in the Surveillance, Epidemiology, and End Results (SEER) program. Univariate survival analysis was conducted using Kaplan-Meier analysis, and multivariate analyses were performed using Cox proportional hazards regression model. Competing-risks regression model was applied to estimate specific risks associated with mortality.
Results
The analysis demonstrated increased mortality in males and patients diagnosed at older age and higher disease stage. NHL was the most common cause of mortality in patients with ENKTL, accounting for 74.13% of deaths in the cohort, followed by other malignant cancers, heart diseases, and infection. However, NHL-specific death events were fewer in patients diagnosed with advanced disease stage compared with incidences of death by other causes such as disease of heart and infections. Significant difference was seen between patients diagnosed earlier than 2000, who showed a higher probability of dying from NHL, and those diagnosed later, who showed propensity to die from other malignant tumors and infection. No differences were found when comparing sex or age at diagnosis.
Conclusion
The most common cause of mortality in cases with ENKTL-NT is NHL. The female sex, diagnosis at young age and early stage are associated with improved prognosis. Further, the classification of Ann Arbor stage and year of diagnosis can provide references of specific causes of death, which might help decrease the mortality rate.
Introduction
Extra nodal natural killer/T-cell lymphoma, nasal type (ENKTL-NT) is a comparatively rare and highly invasive type of non-Hodgkin's lymphoma (NHL) that originates from NK cells or T cells and is commonly presented in the nasal cavity or lymphatic system.[1–6] The overall incidence of ENKTL-NT from 2004 to 2009 was reported to be 0.03 in SEER dataset from National Cancer Institute (NCI), USA (per 100,000 population).[6] Further, studies have shown that Epstein-Barr virus (EBV) can be a risk factor in the development of ENKTL-NT,[7–9] leading to a poor outcome with regular relapse, rapid progress, or resistance to treatment. In addition, standard guidelines of therapy for ENKTL patients have not yet been established.[1, 10, 11] Although curative therapies, including radiotherapy, chemotherapy, immunotherapy, and autologous peripheral blood stem cell transplantation (auto-PBSCT), have been around for years; the prognosis of ENKTL patients is disappointing; the overall survival (OS) was reported to be around 42% while the 5-year progression-free survival (PFS) was only 29%.[11, 12] Thus, it is necessary to gain an understanding of the causes and associated risk factors leading to mortality in cases with ENKTL-NT for improving prognosis and curbing the disease progression.
Bluhm reported that the causes of mortality in cases diagnosed with NHL were NHL, leukemia, other malignant tumors, cardiovascular diseases and infections.[13] Further, organ transplantations, HIV infections and immunodeficiency syndromes have been reported as other risk factors leading to death.[14] However, the cause of death in cases with ENKTL-NT remains unclear. Therefore, this study aimed at conducting a retrospective population-based cohort study to analyze the various causes of death and risk factors associated with ENKTL-NT and estimated the most common causes, in order to provide references for clinical practice and early prognosis.
Materials and methods
Surveillance, Epidemiology, and End Results (SEER) Program
Information regarding the participants was downloaded from the Surveillance, Epidemiology, and End Results (SEER) Program [15], which was carried out by the National Cancer Institute (NCI), USA. The program included 18 registries and covered over 28% of the US population. We identified patients who were diagnosed with ENKTL-NT from 1987 to 2014 in SEER program. The data were selected and analyzed through the SEER*stat software version 8.3.4 (www.seer.cancer.gov/seerstat). SEER is supported by the Surveillance Research Program (SRP) in NCI's Division of Cancer Control and Population Sciences (DCCPS) (www.seer.cancer.gov).
Patients
Cases were selected based on the classification of lymphoid neoplasms published by International Lymphoma Epidemiology Consortium (InterLymph) Pathology Working Group, according to the WHO classification of lymphoid neoplasms and the International Classification of Diseases-Oncology, Third Edition (ICD-O-3). Cases diagnosed with ENKTL-NT with adequate available information were selected for this study. No restricting criteria were added for age, sex, and race. A total of 163 patients participated in the program from diagnosis till the end of a 5-year follow-up or death, whichever occurred first. However, since we aimed to analyze the cause of death, all patients who survived past the follow-up period were excluded from our study. In addition, all patients with death certificate only or autopsy only were excluded, as were patients whose survival time could not be ascertained. The patients included were those diagnosed with ENKTL as the first primary tumor. The study focused on analyzing the primary cause leading to death. Finally, based on our inclusion and exclusion criteria, 107 patients were analyzed for the cause of death and divided into two groups: NHL-specific mortality and mortality by other causes. The latter was further classified in four sub-groups: (1) other malignant tumors, (2) diseases of heart, (3) infections, and (4) other causes.
Variables
Variables analyzed included sex (male or female), calendar year of diagnosis (1987–2000, 2001–2007, or 2008–2014), race (Caucasian, African American, others), Ann Arbor Stage of the disease (stage I and II or stage III and IV) and age at diagnosis (< 60 or ≥ 60).
Statistical analysis
Distributions of survival time were assessed grouped by basic information and Ann Arbor stage of disease in patients with ENKTL, nasal type. The ICD-10 codes for different causes of death were used in the study to standardize the analysis. The overall survival (OS) was defined as the date from diagnosis to the date of death or the end of the follow-up.
Descriptive statistics, including median and confidence intervals (CI), were provided for continuous variables. Frequencies and percentages were used to summarize categorical variables. For comparisons of OS between subgroups classified by sex, Ann Arbor Stage, race and age at diagnosis of patients, survival curves were drawn using the Kaplan-Meier method (Graphpad Prism 6 Software, USA). Univariate analyses were performed by the log-rank test while multivariate analyses were performed using the Cox proportional hazards regression model. The pie nest chart was drawn using Echarts (https://ecomfe.github.io/echarts-doc/public/en/index.html). All statistical analyses were performed using IBM SPSS Statistics for Windows, version 21.
Owing to the nature of retrospective studies, a potential confounding bias may exist. Therefore, the competing-risks regression model was applied using software STATA (version 15) to analyze the competing events. When calculating NHL-specific mortality, death by NHL was the event of interest and other-causes mortality was the competing event, and vice versa. Similarly, when considering the sub-group ‘other causes,’ the four groups were competing events against each other. The analysis of each cause was further grouped by sex, Ann Arbor Stage, calendar year of diagnosis and age at diagnosis of patients. The cause-specific sub-hazard ratios (SHRs) and cumulative incidence function (CIF) were calculated to demonstrate competing risks. We calculated the CIFs for stage, age group, and calendar year of diagnosis. All statistically significant differences were defined as p < 0.05.
Results
Distributions of patients
This study included 163 participants with adequate information during the 5-year follow up from the SEER database. Among these participants, 116 died at a median age of 58, with male and female patients accounting for 67% and 33% of the deaths, respectively. The number of patients diagnosed in early years was much smaller than that of patients diagnosed in recent years. Since the data was collected in the U.S., the cohort mostly comprised of White Americans. All patients were classified by the Ann Arbor staging system used for classifying lymphomas (1983+). The number of patients in stage I and II (n = 111, 68.1%) were more than twice compared with that in the advanced stage III and IV (n = 43, 26.38%). Detailed distributions of all cases classified by sex, race, stage, calendar year of diagnosis, and age at diagnosis are presented in Table 1.
Table 1. Distributions of characteristics of patients with extra nodal natural killer/T-cell lymphoma, nasal type.
| Characteristics | No. | (%) | MS (months) | CI (months) | P-value |
|---|---|---|---|---|---|
| Sex | |||||
| Male | 109 | (66.87) | 11 | (6.288, 15.712) | P = 0.038 |
| Female | 54 | (33.13) | 15 | (0.000, 36.604) | |
| Race | |||||
| Caucasian | 98 | (60.12) | 12 | (4.062, 19.938) | P = 0.100 |
| African American | 9 | (5.52) | 5 | (2.078, 7.922) | |
| Others | 56 | (34.36) | 14 | (3.523, 24.477) | |
| Ann Arbor Stage | |||||
| I and II | 111 | (68.10) | 23 | (9.499, 36.501) | P<0.001 |
| III and IV | 43 | (26.38) | 6 | (3.247, 8.753) | |
| Unknown | 9 | (5.52) | 11 | (1.298, 20.702) | |
| Age at diagnosis | |||||
| <60 | 96 | (58.90) | 23 | (10.782, 35.218) | P<0.001 |
| ≥60 | 67 | (41.10) | 7 | (5.234, 8.766) | |
| Calendar year of diagnosis | |||||
| 1987–2000 | 29 | (17.79) | 28 | (0.000, 68.436) | - |
| 2001–2007 | 70 | (42.94) | 23 | (9.340, 36.660) | |
| 2008–2014 | 64 | (39.26) | - | - |
MS: Median Survival Time, CI: confidence intervals.
Survival analysis
Kaplan-Meier analysis of the cohort demonstrated increased mortality in males, and diagnosis at older age and with higher disease stage. Female patients showed better OS, with median survival time of 15 months, compared with males, who showed a median survival time of 11 months (P = 0.038; Fig 1A). The OS rate was higher in the < 60-year-old age group compared with that in the older age group (P < 0.001). The median survival duration was 23 months (for patients < 60-year-old) and 7 months (for patients ≥ 60-year-old) (Fig 1B). Further, the patients in stage III and IV displayed a shorter lifetime and higher mortality (P < 0.001; Fig 1C). Caucasian Americans and others, including American Indian-Alaska Natives, Asians, and Pacific Islanders showed better OS, with a median of 12 months or 14 months respectively, compared with African American patients, who had a median survival time of 5 (p = 0.100; Fig 1D). However, this could be attributable to the small sample size for African Americans compared with that of other races.
Fig 1.
Kaplan-Meier curves for comparison of patients diagnosed with extra nodal natural killer/T-cell lymphoma, nasal type according to (A) Sex, (B) Age at diagnosis, (C) Ann Arbor Stage and (D) Race.
The survival rates for 2-year, 3-year and 10-year follow up were 39.9%, 33.7%, and 28.8%, respectively (Table 2). The survival rate dropped quickly in males, other races (including American Indian-Alaska Native, Asian and Pacific Islander), and patients with older age.
Table 2. Survival of 2-, 3- and 5-year follow-up in patients with extra nodal natural killer/T-cell lymphoma, nasal type in 1987–2014.
| Characteristics | Total number |
2-year survival | 3-year survival | 5-year survival | ||||
|---|---|---|---|---|---|---|---|---|
| Frequency | Rate | Frequency | Rate | Frequency | Rate | |||
| Sex | ||||||||
| Male | 109 | 39 | 35.8 | 31 | 28.4 | 25 | 22.9 | |
| Female | 54 | 26 | 48.1 | 24 | 44.4 | 22 | 40.7 | |
| Race | ||||||||
| Caucasian | 98 | 41 | 41.8 | 36 | 36.7 | 32 | 32.7 | |
| African American | 9 | 1 | 11.1 | 1 | 11.1 | 1 | 11.1 | |
| Others | 56 | 23 | 41.1 | 18 | 32.1 | 14 | 25.0 | |
| Ann Arbor Stage | ||||||||
| I and II | 111 | 54 | 48.6 | 46 | 41.4 | 40 | 36.0 | |
| III and IV | 43 | 8 | 18.6 | 6 | 14.0 | 4 | 9.3 | |
| Unknown | 9 | 3 | 33.3 | 3 | 33.3 | 3 | 33.3 | |
| Age at diagnosis | ||||||||
| <60 | 96 | 46 | 47.9 | 39 | 40.6 | 37 | 38.5 | |
| ≥60 | 67 | 19 | 28.4 | 16 | 23.9 | 10 | 14.9 | |
| Total | 163 | 65 | 39.9 | 55 | 33.7 | 47 | 28.8 | |
Consistently, Cox proportional hazards regression showed that diagnosis at older age (P < 0.001) was associated with worse survival (Table 3). In addition, African Americans were associated with worse survival, compared with Caucasians. However, there was no significant difference between sex and stage.
Table 3. Hazard ratios and 95% confidence intervals of Cox regression for Mortality Among Patients With extra nodal natural killer/T-cell lymphoma, nasal type Diagnosed Among US Veterans From 1987 to 2014.
| Characteristics | Hazard Ratios | P | 95.0% confidence intervals | |
|---|---|---|---|---|
| Sex | ||||
| Male | 1.00 (reference) | |||
| Female | 1.476 | 0.075 | 0.962 | 2.264 |
| Race | ||||
| Caucasian | 1.00 (reference) | |||
| African American | 0.455 | 0.045 | 0.211 | 0.984 |
| Others | 0.575 | 0.170 | 0.261 | 1.267 |
| Ann Arbor Stage | ||||
| I and II | 1.00 (reference) | |||
| III and IV | 0.943 | 0.893 | 0.397 | 2.237 |
| Unknown | 1.843 | 0.183 | 0.750 | 4.527 |
| Age at diagnosis | ||||
| <60 | 1.00 (reference) | |||
| ≥60 | 0.502 | <0.001 | 0.345 | 0.731 |
Cause of death
Among all the causes listed in Table 4, NHL was the most common cause of death, accounting for 74.13% of all deaths, followed by other malignant cancers. Several of the deaths were also accounted for by infectious and parasitic diseases and diseases of the heart. Apart from these, chronic obstructive pulmonary disease and allied conditions, complications of pregnancy, childbirth, puerperium, congenital anomalies, accidents, suicide, and self-Inflicted injury were also listed as causes of mortality, at a relatively low rate.
Table 4. Distribution of causes of death of follow-up in patients with extra nodal natural killer/T-cell lymphoma, nasal type, 1987–2014.
| Mortality status | No. | (%) |
| Total | 163 | (100.00) |
| Alive throughout 5-year follow-up | 47 | (28.83) |
| Death | 116 | (71.17) |
| Non-Hodgkin lymphoma | 86 | (52.76) |
| Other malignant cancers | 6 | (3.68) |
| Nasopharynx | 1 | (0.61) |
| Nose, nasal cavity and middle ear | 1 | (0.61) |
| Lung and bronchus | 1 | (0.61) |
| Myeloma | 1 | (0.61) |
| Miscellaneous malignant cancer | 2 | (1.23) |
| Pneumonia and influenza | 1 | (0.61) |
| Septicemia | 1 | (0.61) |
| Other infectious and parasitic diseases including HIV | 5 | (3.07) |
| Diseases of heart | 4 | (2.45) |
| Chronic obstructive pulmonary disease and allied cond | 1 | (0.61) |
| Complications of pregnancy, childbirth, puerperium | 1 | (0.61) |
| Congenital anomalies | 1 | (0.61) |
| Accidents and adverse effects | 1 | (0.61) |
| Suicide and self-Inflicted injury | 1 | (0.61) |
| Other cause of death | 5 | (3.07) |
| State DC not available or state DC available but no COD | 3 | (1.84) |
DC: death certificate. COD: cause of death.
Competing risk regression
Ordinary Cox regression may give possibly biased estimates of cause-specific risk of death by failing to account for competing causes. [15] Therefore, we conducted competing risk regression for cause-specific purpose. The regression model to show the CIF for competing specific risks was first proposed by Fine and Gray and developed further by many researches in various statistical analysis software.[16, 17] The different causes of death were competing events in this study, and a series of regression models were utilized for comparison and risk determination.
Although this study included 116 participants who died within the follow-up period, six of them did not have records on Ann Arbor Stage and three did not have State DC (death certificate) or COD (cause of death). Only 107 had adequate information on COD, sex, Ann Arbor Stage, age at diagnosis, and year of diagnosis. Since all death cases in the African American group could be attributable to NHL, and not to any other cause, we considered these samples to be inadequate to conduct competing risk method analysis.
Tables 5 and 6 show the distribution of the 107 cases based on the cause of mortality. Results of the competing-risks regression model demonstrated that for patients of stage III and IV, NHL-specific death events were much fewer, while incidence of death by other causes was higher, such as by disease of heart and infections (Table 6). Significant difference was seen between the patients diagnosed earlier than 2000, who showed a higher probability to die from NHL (Fig 2), compared with those diagnosed at a later calendar year, who showed a higher rate of mortality from other malignant tumors and infection. We found no differences when comparing the sex or age at diagnosis.
Table 5. Sub-hazard ratios of cause-specific death by sex, Ann Arbor Stage, calendar year of diagnosis and age at diagnosis by competing-risks regression.
| Characteristic | NHL-specific mortality | Other-cause mortality | ||||
|---|---|---|---|---|---|---|
| SHR | (95% CI) | P-value | SHR | (95% CI) | P-value | |
| Sex | ||||||
| Male | 1.00(reference) | 1.00(reference) | ||||
| Female | 0.824 | (0.275, 2.471) | 0.729 | 0.975 | (0.615, 1.547) | 0.916 |
| Ann Arbor Stage | ||||||
| I and II | 1.00(reference) | 1.00(reference) | ||||
| III and IV | 0.714 | (0.520, 0.980) | 0.037 | 1.281 | (1.117, 1.468) | 0.001 |
| Calendar year of diagnosis | ||||||
| 1987–2000 | 1.00(reference) | 1.00(reference) | ||||
| 2001–2008 | 0.197 | (0.073, 0.532) | 0.001 | 3.104 | (1.384, 6.960) | 0.006 |
| 2009–2014 | 0.481 | (0.200, 1.156) | 0.102 | 1.961 | (0.875, 4.394) | 0.102 |
| Age at diagnosis | ||||||
| <60 | 1.00(reference) | 1.00(reference) | ||||
| ≥60 | 0.757 | (0.496, 1.157) | 0.199 | 1.308 | (1.033, 1.656) | 0.026 |
NHL: non-Hodgkin’s lymphoma SHR: Sub-hazard ratios. CI: confidence intervals.
Table 6. Sub-hazard ratios with 95% confidence intervals of other causes rather than non-Hodgkin’s lymphoma by sex, Ann Arbor Stage, calendar year of diagnosis and age at diagnosis by competing-risks regression.
| Characteristic | Other malignant tumors | Diseases of heart | Infections | Other causes | ||||
|---|---|---|---|---|---|---|---|---|
| SHR (95% CI) | P-value | SHR (95% CI) | P-value | SHR (95% CI) | P-value | SHR (95% CI) | P-value | |
| Sex | ||||||||
| Male | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | ||||
| Female | 1.440 (0.964, 2.151) | 0.075 | 1.205 (0.799, 1.818) | 0.374 | 0.913 (0.586, 1.423) | 0.688 | 0.920 (0.592, 1.430) | 0.711 |
| Ann Arbor Stage | ||||||||
| I and II | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | ||||
| III and IV | 1.140 (0.996, 1.305) | 0.058 | 1.217 (1.075, 1.378) | 0.002 | 1.230 (1.089, 1.388) | 0.001 | 1.224 (1.089, 1.421) | 0.001 |
| Calendar year of diagnosis | ||||||||
| 1987–2000 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | ||||
| 2001–2008 | 1.657 (1.025, 2.680) | 0.039 | 1.866 (0.974, 3.577) | 0.060 | 2.962 (1.484, 5.914) | 0.002 | 1.566 (0.891, 2.754) | 0.119 |
| 2009–2014 | 1.244 (0.720, 2.150) | 0.434 | 1.886 (0.981, 3.627) | 0.057 | 2.437 (1.218, 4.878) | 0.012 | 1.279 (0.694, 2.359) | 0.430 |
| Age | ||||||||
| <60 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | ||||
| ≥60 | 0.941(0.757, 1.170) | 0.583 | 0.958 (0.774, 1.185) | 0.690 | 1.184 (0.962, 1.456) | 0.111 | 1.197 (0.965, 1.484) | 0.102 |
SHR: Sub-hazard ratios. CI: confidence intervals.
Fig 2. Distribution of causes of death of follow-up in patients with extra nodal natural killer/T-cell lymphoma, nasal type, 1987–2014.
Figs 3 and 4 present cumulative incidence curves for NHL-specific mortality and other causes of mortality over the 5-year follow-up based on age at diagnosis, stage, and calendar year. In accordance with sub-hazard ratios of cause-specific death by competing-risks regression, the cumulative incidence of NHL-specific mortality increased more rapidly in cases diagnosed at an early calendar year and stage I and II. The cumulative incidences of other causes of mortality (Fig 5) showed the opposite result. In addition, the cumulative incidence of older age was higher in other-cause mortality.
Fig 3.
Cumulative incidence plot comparing (A) Ann Arbor stage and (B) calendar year of diagnosis in NHL-specific cause mortality.
Fig 4.
Cumulative incidence plot comparing (A) Ann Arbor stage, (B) calendar year of diagnosis (C) age at diagnosis in other cause mortality.
Fig 5.
Cumulative incidence plot comparing (A) calendar year of diagnosis attribute to other tumors, (B) Ann Arbor stage of death attribute to disease of heart, (C) Ann Arbor stage of death attribute to infection and (D) calendar year of diagnosis attribute to infection.
Discussion
Extra nodal natural killer/T-cell lymphoma, nasal type (ENKTL-NT), is a comparatively infrequent and highly aggressive lymphoma that currently lacks optimal treatment and shows poor outcome. [18] In our study, female patients, diagnosed at younger age and early stage showed a longer survival time than that reported in previous studies. [19–21]Some studies have illustrated that females have a better survival rate in patients diagnosed with ENKTL-NT, consistent with other NHL,[22–24] probably because male patients are more likely to suffer cardiovascular diseases after being diagnosed with NHL[3] and are more likely to use alcohol and tobacco that increase the risk.[25] Advanced age (defined as > 60 years) has been included in the International Prognostic Index as an unfavorable factor for ENKTL-NT disease outcome.[18, 26] Ann Arbor stage is another important prognostic factor in NHL.[18, 22] Patients in advanced stages may receive sequential and multi-cycle chemotherapy or combined radiotherapy that might bring about severe drug toxicity.[18] Indeed, toxic deaths (63%) have been reported, mostly occurring after the first cycle of chemotherapy.[27] Patients in stage IV face the higher risk of acquiring natural killer/T-cell lymphoma-associated hemophagocytic syndrome (HPS),[28] which is scarce, but life-threatening.[29, 30] Moreover, tumor size, which was excluded in staging, also significantly influences prognosis of the disease.[31] In addition, some researchers claim that intra-cavity disease had a better prognosis than extra-cavity disease.[6, 22]
The current retrospective cohort study illustrated that although the most common cause of death with ENKTL-NT was NHL, there is considerable risk of other malignant cancers, including the cancers of nasopharynx, nose, nasal cavity and middle ear, lung and bronchus, myeloma and miscellaneous malignant cancer. Additionally, several factors, including age, sex, stage and time of diagnosis, could be risk factors affecting cause-specific mortality.
The competing-risks regression model was applied in this study to evaluate the relationship of variables to cause-specific failures, [15] while the cause-specific sub-distribution hazard function was applied to indicate cumulative probability (incidence) of the causes.
Death rate attributed to causes other than NHL was much higher for patients in advanced stage of the disease. This could be because patients in the advanced stage are more likely to face aggressive therapies, leading to potential risk of hematologic toxicity. For instance, SMILE (dexamethasone, methotrexate, ifosfamide, L-asparaginase, and etoposide) regimen, which is the current standard to treat advanced-stage ENKTL, may be associated with more severe hematological toxicity compared with other L-asparaginase combinations and gemcitabine-based or CCRT (concurrent chemoradiotherapy) based regimens. [32] SMILE, which leads to the most toxic hematological outcome, has been associated with grades 3–4 neutropenia in 72% of treated patients. [32] After SMILE treatment, 45% patients had grade 3 and 16% patients had grade 4 infection, including 2 patients who died as a result of infection.[33]
Patients diagnosed in this century showed a higher survival rate than those diagnosed in the last century, partly attributed to the standardization of suitable and effective therapies, the development of HSCT,[34] and immune- and cellular therapy.[35] Patients of advanced age (defined as > 60 years) were more likely to face causes other than NHL, because other effects of aging such as a decrease in the baseline cardiac, immune, renal, and hepatic function. [3] No differences were found in variables of sex and age at diagnosis during the cause-specific analysis. However, previous studies have reported that female sex was associated with risks of mortality other than NHL among childhood NHL survivors.[13, 36] This risk factor remains to be further validated in case of ENKTL-NT.
The current study has some limitations. First, some confounding factors may be included due to the retrospective nature of the study. Since we evaluated only the data collected from SEER, other potential risk factors may be missed. Second, a selection bias exists since patients with incomplete information were excluded. Third, there was a lack of consistency in death certificates obtained from different registries. In addition, since ENKTL-NT is a rare subtype of lymphoma, it is difficult to get larger sample pool to decrease the sampling error.
Currently, IPI (patient age, tumor stage, serum lactate dehydrogenase [LDH] concentration, performance status, and number of extranodal disease sites) has been most widely used in NHL prognosis since more than twenty years. [37] However, IPI cannot provide context regarding the patient’s cancer-specific mortality risks in the presence of competing risks. [38] Predicting ENKTL-NT outcome and stratifying patients for different competing risk events are a challenge for physicians, making it important and urgent to identify novel biomarkers for better prognosis.
Conclusion
Our study indicates that the female sex, diagnosis at young age and with early stage of the disease are factors associated with better prognosis for ENKTL-NT. This population-based cohort study revealed that the most common cause of mortality in cases with ENKTL-NT is NHL, rather than other malignant cancers, diseases of heart, infections or other causes. Our study emphasizes that more effort should be made to identify patients in early stages so as to implement comprehensive treatment and prevent the occurrence of high-risk factors. Furthermore, the Ann Arbor staging and calendar year of diagnosis provide references for cause-specific death, and might help during the clinical procedures to decrease the overall and cause-specific mortality rate.
Availability of data and materials
The datasets analyzed are available in the Surveillance, Epidemiology, and End Results (SEER) program (www.seer.cancer.gov) which is supported by the Surveillance Research Program (SRP) in NCI's Division of Cancer Control and Population Sciences (DCCPS).Because of the sensitive nature of the data, SEER Data Use Agreement (https://seer.cancer.gov/seertrack/data/request/) is requested for everyone before accessing to the database through SEER*Stat's Client-Server Mode (https://seer.cancer.gov/seerstat/).
Ethics approval and consent to participate
Since the database used was a group of public sets, it was exempted from institutional review boards. The participants were requested using the National Cancer Institute SEER*stat software version 8.3.4 (www.seer.cancer.gov/seerstat).
Acknowledgments
We are grateful to all the study participants, doctors and coordinators in SEER program.
Data Availability
The data used in this manuscript can be obtained in the same manner the authors obtained it. It is convenient to obtain the data through SEER*Stat's Client-Server Mode. After submitting a request for access to the data (https://seer.cancer.gov/seertrack/data/request/), a personalized SEER Research Data Agreement will be created and send back by email. Then the user will be asked to print the data user agreement from, sign the agreement and send the form back to the SEER program by fax to 301-680-9571 or e-mail a scanned form to seertrack@imsweb.com. The SEER program will process the request and send back account number and password within 2 business days of receiving signed agreement. Then download the SEER*Stat and login in with the account number and password.
Funding Statement
This work was supported by the National Natural Scientific Foundation of China (No. 81570203).
References
- 1.Huang Y, Jia B, Jiang S, Zhou S, Yang J, Liu P, et al. Different clinical characteristics and treatment strategies for patients with localized sinonasal diffuse large B cell lymphoma and extranodal NK/T cell lymphoma. Journal of Hematology & Oncology. 2017; 10(1): 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Zeng LS, Huang WT, Qiu T, Shan L, Guo L, Ying JM, et al. Correlation between the clinicopathological features and prognosis in patients with extranodal natural killer/T cell lymphoma. Chronic Diseases & Translational Medicine. 2017; 3(4): 252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kamel MG, Elqushayri AE, Thach TQ, Huy NT. Cardiovascular mortality trends in non-Hodgkin's lymphoma: a population-based cohort study. Expert Rev Anticancer Ther. 2018; 18(1): 91–100. 10.1080/14737140.2018.1409626 [DOI] [PubMed] [Google Scholar]
- 4.Jhuang J, Chang S, Weng S, Pan S, Chu P, Hsieh P, et al. Extranodal natural killer/T-cell lymphoma, nasal type in Taiwan: a relatively higher frequency of T-cell lineage and poor survival for extranasal tumors. Human Pathology. 2015; 46(2): 313–321. 10.1016/j.humpath.2014.11.008 [DOI] [PubMed] [Google Scholar]
- 5.Al-Hakeem DA, Fedele S, Carlos R, Porter S. Extranodal NK/T-cell lymphoma, nasal type. Oral oncology. 2007; 43(1): 4–14. Epub 2006/10/27. 10.1016/j.oraloncology.2006.03.011 . [DOI] [PubMed] [Google Scholar]
- 6.Vazquez A, Khan MN, Blake DM, Sanghvi S, Baredes S, Eloy JA. Extranodal natural killer/T-Cell lymphoma: A population-based comparison of sinonasal and extranasal disease. The Laryngoscope. 2014; 124(4): 888–895. Epub 2013/10/12. 10.1002/lary.24371. [DOI] [PubMed] [Google Scholar]
- 7.Fei Q, Tian XK, Wu J, Zhu HM, Wang Y, Peng FY, et al. Prognostic significance of Epstein-Barr virus DNA in NK/T-cell lymphoma: a meta-analysis. OncoTargets and therapy. 2018; 11: 997–1004. Epub 2018/03/10. 10.2147/ott.s153942 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ito T, Kawazu H, Murata T, Iwata S, Arakawa S, Sato Y, et al. Role of latent membrane protein 1 in chronic active Epstein-Barr virus infection-derived T/NK-cell proliferation. Cancer medicine. 2014; 3(4): 787–795. Epub 2014/05/07. 10.1002/cam4.256 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Takada H, Imadome KI, Shibayama H, Yoshimori M, Wang L, Saitoh Y, et al. EBV induces persistent NF-kappaB activation and contributes to survival of EBV-positive neoplastic T- or NK-cells. PloS one. 2017; 12(3): e0174136 Epub 2017/03/28. 10.1371/journal.pone.0174136 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Li X, Cheng Y, Zhang M, Yan J, Li L, Fu X, et al. Activity of pembrolizumab in relapsed/refractory NK/T-cell lymphoma. Journal of Hematology & Oncology. 2018; 11(1): 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.William BM, Armitage JO. International analysis of the frequency and outcomes of NK/T-cell lymphomas. Best Pract Res Clin Haematol. 2013; 26(1): 23–32. 10.1016/j.beha.2013.04.003 [DOI] [PubMed] [Google Scholar]
- 12.Yamaguchi M. Current and future management of NK/T-cell lymphoma based on clinical trials. International Journal of Hematology. 2012; 96(5): 562–571. 10.1007/s12185-012-1189-4 [DOI] [PubMed] [Google Scholar]
- 13.Bluhm EC, Ronckers C, Hayashi RJ, Neglia JP, Mertens AC, Stovall M, et al. Cause-specific mortality and second cancer incidence after non-Hodgkin lymphoma: a report from the Childhood Cancer Survivor Study. Blood. 2008; 111(8): 4014 10.1182/blood-2007-08-106021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shankland KR, Armitage JO, Hancock BW. Non-Hodgkin lymphoma. Lancet. 2012; 380(9844): 848–857. 10.1016/S0140-6736(12)60605-9 [DOI] [PubMed] [Google Scholar]
- 15.Dignam JJ, Zhang Q, Kocherginsky MN. The Use and Interpretion of Competing Risks Regression Models. Clinical Cancer Research An Official Journal of the American Association for Cancer Research. 2012; 18(8): 2301 10.1158/1078-0432.CCR-11-2097 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fine JP, Gray RJ. A Proportional Hazards Model for the Subdistribution of a Competing Risk. Publications of the American Statistical Association. 1999; 94(446): 496–509. [Google Scholar]
- 17.Scrucca L, Santucci A, Aversa F. Regression modeling of competing risk using R: an in depth guide for clinicians. Bone Marrow Transplantation. 2010; 45(9): 1388 10.1038/bmt.2009.359 [DOI] [PubMed] [Google Scholar]
- 18.Liang R, Wang Z, Bai QX, Gao GX, Yang L, Zhang T, et al. Natural Killer/T Cell Lymphoma, Nasal Type: A Retrospective Clinical Analysis in North-Western China. Oncology Research & Treatment. 2016; 39(1–2): 45. [DOI] [PubMed] [Google Scholar]
- 19.Wang H, Wang L, Wuxiao Z, Huang H, Jiang W, Li Z, et al. Increased serum levels of interleukin-10 predict poor prognosis in extranodal natural killer/T-cell lymphoma patients receiving asparaginase-based chemotherapy. Onco Targets Ther. 2015; 8: 2589–2599. Epub 2015/09/24. 10.2147/ott.s91077 . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hong SA, Tajudeen BA, Choi S, Husain IA. Epidemiology and prognostic indicators in laryngeal lymphoma: A population-based analysis. The Laryngoscope. 2018; 128(9): 2044–2049. Epub 2018/02/03. 10.1002/lary.27074 . [DOI] [PubMed] [Google Scholar]
- 21.Au WY, Weisenburger DD, Intragumtornchai T, Nakamura S, Kim WS, Sng I, et al. Clinical differences between nasal and extranasal natural killer/T-cell lymphoma: a study of 136 cases from the International Peripheral T-Cell Lymphoma Project. Blood. 2009; 113(17): 3931–3937. Epub 2008/11/26. 10.1182/blood-2008-10-185256 . [DOI] [PubMed] [Google Scholar]
- 22.Pulte D, Gondos A, Brenner H. Ongoing Improvement in Outcomes for Patients Diagnosed as Having Non-Hodgkin Lymphoma From the 1990s to the Early 21st Century. Archives of Internal Medicine. 2008; 168(5): 469–476. 10.1001/archinternmed.2007.125 [DOI] [PubMed] [Google Scholar]
- 23.Sant M, Allemani C, De AR, Carbone A, De SS, Gianni AM, et al. Influence of morphology on survival for non-Hodgkin lymphoma in Europe and the United States. European Journal of Cancer. 2008; 44(4): 579–587. 10.1016/j.ejca.2007.12.016 [DOI] [PubMed] [Google Scholar]
- 24.Wu X, Li P, Zhao J, Yang X, Wang F, Yang YQ, et al. A clinical study of 115 patients with extranodal natural killer/T-cell lymphoma, nasal type. Clin Oncol. 2008; 20(8): 619–625. [DOI] [PubMed] [Google Scholar]
- 25.Battaglioli T, Gorini G, Costantini AS, Crosignani P, Miligi L, Nanni O, et al. Cigarette smoking and alcohol consumption as determinants of survival in non-Hodgkin's lymphoma: a population-based study. Annals of Oncology Official Journal of the European Society for Medical Oncology. 2006; 17(8): 1283 10.1093/annonc/mdl096 [DOI] [PubMed] [Google Scholar]
- 26.Vose JM, Armitage JO, Weisenburger DD, Bierman PJ, Sorensen S, Hutchins M, et al. The importance of age in survival of patients treated with chemotherapy for aggressive non-Hodgkin's lymphoma. Journal of Clinical Oncology Official Journal of the American Society of Clinical Oncology. 1988; 6(12): 1838–1844. 10.1200/JCO.1988.6.12.1838 [DOI] [PubMed] [Google Scholar]
- 27.Bairey O, Bar-Natan M, Shpilberg O. Early death in patients diagnosed with non-Hodgkin's lymphoma. Annals of Hematology. 2013; 92(3): 345 10.1007/s00277-012-1623-y [DOI] [PubMed] [Google Scholar]
- 28.Han L, Zhang M, Li L, Zhang L, Wu J, Li X, et al. Natural killer/T-cell lymphoma-associated hemophagocytic syndrome: A case report. Oncology Letters. 2014; 8(2): 886–890. 10.3892/ol.2014.2202 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Jin Z, Wang Y, Wang J, Wu L, Pei R, Lai W, et al. Multivariate analysis of prognosis for patients with natural killer/T cell lymphoma-associated hemophagocytic lymphohistiocytosis. Hematology. 2017; (1): 1–7. [DOI] [PubMed] [Google Scholar]
- 30.Xing Y, Yang J, Lian G, Chen S, Chen L, Li F. Chronic active Epstein–Barr virus infection associated with hemophagocytic syndrome and extra-nodal natural killer/T-cell lymphoma in an 18-year-old girl: A case report. Medicine. 2017; 96(19): e6845 10.1097/MD.0000000000006845 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wilder RB, Rodriguez MA, Ha CS, Pro B, Hess MA, Cabanillas F, et al. Bulky disease is an adverse prognostic factor in patients treated with chemotherapy comprised of cyclophosphamide, doxorubicin, vincristine, and prednisone with or without radiotherapy for aggressive lymphoma. Cancer. 2001; 91(12): 2440 [PubMed] [Google Scholar]
- 32.Pokrovsky VS, Vinnikov D. Defining the toxicity of current regimens for extranodal NK/T cell lymphoma: a systematic review and metaproportion. Expert Rev Anticancer Ther. 2018: 1–12. Epub 2018/11/20. 10.1080/14737140.2019.1549992 . [DOI] [PubMed] [Google Scholar]
- 33.Yamaguchi M, Kwong YL, Kim WS, Maeda Y, Hashimoto C, Suh C, et al. Phase II study of SMILE chemotherapy for newly diagnosed stage IV, relapsed, or refractory extranodal natural killer (NK)/T-cell lymphoma, nasal type: the NK-Cell Tumor Study Group study. Journal of clinical oncology: official journal of the American Society of Clinical Oncology. 2011; 29(33): 4410–4416. Epub 2011/10/13. 10.1200/jco.2011.35.6287 . [DOI] [PubMed] [Google Scholar]
- 34.Suzuki R. Pathogenesis and treatment of extranodal natural killer/T-cell lymphoma. Seminars in hematology. 2014; 51(1): 42–51. Epub 2014/01/29. 10.1053/j.seminhematol.2013.11.007 . [DOI] [PubMed] [Google Scholar]
- 35.Jo JC, Kim M, Choi Y, Kim HJ, Kim JE, Chae SW, et al. Expression of programmed cell death 1 and programmed cell death ligand 1 in extranodal NK/T-cell lymphoma, nasal type. Annals of hematology. 2017; 96(1): 25–31. Epub 2016/10/04. 10.1007/s00277-016-2818-4 . [DOI] [PubMed] [Google Scholar]
- 36.Möller TR, Garwicz S, Barlow L, Winther JF, Glattre E, Olafsdottir G, et al. Decreasing Late Mortality Among Five-Year Survivors of Cancer in Childhood and Adolescence: A Population-Based Study in the Nordic Countries. Journal of Clinical Oncology Official Journal of the American Society of Clinical Oncology. 2001; 19(13): 3173–3181. 10.1200/JCO.2001.19.13.3173 [DOI] [PubMed] [Google Scholar]
- 37.Project. IN-HsLPF. A predictive model for aggressive non-Hodgkin's lymphoma. The New England journal of medicine. 1993; 329(14): 987–994. Epub 1993/09/30. 10.1056/nejm199309303291402 . [DOI] [PubMed] [Google Scholar]
- 38.Hester LL, Park SI, Wood WA, Sturmer T, Brookhart MA, Lund JL. Cause-specific mortality among medicare beneficiaries with newly diagnosed non-Hodgkin lymphoma subtypes. Cancer. 2018. Epub 2018/12/15. 10.1002/cncr.31821 . [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data used in this manuscript can be obtained in the same manner the authors obtained it. It is convenient to obtain the data through SEER*Stat's Client-Server Mode. After submitting a request for access to the data (https://seer.cancer.gov/seertrack/data/request/), a personalized SEER Research Data Agreement will be created and send back by email. Then the user will be asked to print the data user agreement from, sign the agreement and send the form back to the SEER program by fax to 301-680-9571 or e-mail a scanned form to seertrack@imsweb.com. The SEER program will process the request and send back account number and password within 2 business days of receiving signed agreement. Then download the SEER*Stat and login in with the account number and password.
The datasets analyzed are available in the Surveillance, Epidemiology, and End Results (SEER) program (www.seer.cancer.gov) which is supported by the Surveillance Research Program (SRP) in NCI's Division of Cancer Control and Population Sciences (DCCPS).Because of the sensitive nature of the data, SEER Data Use Agreement (https://seer.cancer.gov/seertrack/data/request/) is requested for everyone before accessing to the database through SEER*Stat's Client-Server Mode (https://seer.cancer.gov/seerstat/).