Background
Extranodal nasal-type NK/T-cell lymphoma is a rare and severe disease, occurring more frequently in Asia and South America than in Europe and North America [
1‐
4]. This type of disease represents a distinct entity among T-cell lymphomas according to the World Health Organization (WHO) classification, [
5] being found in both the nasal cavity and extranasal sites [
6‐
8]. The disease is characterized histologically by the local invasion and necrosis of natural killer (NK) cells or T-cells with an invariable Epstein-Barr virus (EBV) infection. In published studies involving adults with a median age of 50 years, 60-90 % of lymphomas were localized in the nasal and upper airway regions, with the remaining found in extranasal sites [
2,
9,
10]. Due to the low incidence of the disease, only a few randomized controlled trials have been undertaken [
11‐
13]. Previous studies showed that 5-year overall survival (OS) rates were less than 40 %, with progression usually occurring within 2 years [
2,
7,
9‐
15]. However, the early use of sequential radiotherapy (RT) and chemotherapy (CT) for localized nasal NK/T-cell [
13,
16] lymphoma was shown to be a successful therapy, which cured approximately half of patients [
1,
17]. In most patients with advanced disease (stage III/IV), the clinical course is highly aggressive, with frequent CT resistance and poor outcome. CHOP-based therapy [
18,
19] was often used, but associated without satisfactory results. However, recent combination therapies involving L-asparaginase have improved the outcome in high-risk, refractory, or relapsed patients [
20‐
22]. The optimal therapy for advanced-stage or relapsed and refractory disease is yet to be established, although the results of several recent prospective trials showed improved results using CT or CT + RT [
20,
23‐
25]. Considering the rarity of this lymphoma in Europe, we conducted a multicentric retrospective study on nasal-type NK/T cell lymphoma. To this end, we reviewed the clinical and biological characteristics as well as treatments of 36 patients, with outcomes analyzed according to disease responses in order to determine the optimal induction strategy and identify prognostic factors.
Discussion
Extranodal nasal-type NK/T-cell lymphoma is considered a distinct clinicopathological entity according to the WHO classification of lymphoid tumors, being more frequent in Asia and Central and South America than in Western countries [
2,
3,
5,
9,
17,
26]. In a recent international multicenter study reported by the International Peripheral T-cell Lymphoma Project, median survival time for patients with NK/T-cell lymphoma was 8 months, being the worst among all peripheral T-cell lymphomas included in the study [
17]. To date, only a few studies have compared the clinical and pathologic features of patients with nasal or extranasal disease [
7,
10,
17].
In our study, 36 French patients, all of European descent, observed over a 22-year period were diagnosed with nasal-type NK/T cell lymphoma..Main patient characteristics, such as age, sex ratio, B symptoms, and clinical aspects of the disease, were comparable to those previously reported [
17,
27,
28]. At diagnosis, 80 % of tumors involved the upper aerodigestive tract, including the nasal and oral cavity, as well as the naso-, oro- and hypopharynx, while 20 % were extranasal. Disease presentation in our population was aggressive, with 65 % of UNKTL patients presenting stage IV at diagnosis. Although extranasal cases presented more adverse clinical features, no statistical differences were found in terms of OS or PFS between UNKTL and EUNKTL patients. This absence in OS difference may be accounted for by the more advanced disease stage of the UNKTL population. Previous studies reported bone marrow and central nervous system involvement at presentation to be rare, affecting 3 % and 7 % of cases, respectively [
8,
29‐
31]. In our study, however, bone marrow and central nervous system involvement was observed in 22 % and 11 % of patients, respectively. However, regional node involvement was found to be more frequent, affecting 44 % of patients [
32,
33]. In line with previous studies, we identified several variables associated with poor survival, notably regional lymph node involvement [
34], local invasiveness [
34‐
37], elevated lactate deshydrogenase [
1,
34], poor performance status [
1], and B symptoms [
1,
37,
38].
Several staging systems have been proposed in respect to nasal NK/T-cell lymphoma in order to predict prognosis. The Ann-Arbor staging system, while designed mainly for Hodgkin lymphoma, is not always accurate in the case of NK-cell lymphomas. In a number of studies on nasal NK-cell lymphoma, the involvement of areas outside the nasal cavity, including the paranasal sinuses, nasopharynx, and orbits, was defined as stage IE disease. In this case, the Ann-Arbor staging system cannot be used to determine the extent of the disease. Thus, for our patients, we scored local invasiveness greater than or equal to T3 according to the TNM classification. Overall, 72 % of UNKTL patients had a local invasiveness > T3, which was associated with worse survival.
Several studies investigated the impact of IPI in patients with nasal NK/T-cell lymphoma [
7,
9,
13]. However, the use of this index has been controversial [
13,
34,
39,
40], as in a number of studies, only a small proportion of NK/T-cell lymphoma patients (up to 7 %) were categorized as high-risk [
1,
13,
34,
41,
42]. Assigning a conventional IPI score appears to be of limited value, since most cases of NK/T-cell lymphomas are localized and result in a low score, despite survival being poor.
Lee
et al. [
9] developed the KPI, a prognostic model based on four risk factors: B symptoms, advanced stage, elevated lactate dehydrogenase levels, and involvement of regional lymph nodes. When analyzed according to the number of risk factors, the 5-year OS rate was 81 % for patients with no risk factors (score 0) and 7-15 % for those with three to four (scores 3–4). Our data validated the prognostic impact of the KPI, with 5-year OS rates being 60 % for patients with score 0
versus 17 % for those with score 4. This index appears more accurate than the IPI score in distinguishing high-risk groups, potentially being the most appropriate prognostic scoring system for patients with nasal-type NK/T lymphoma [
43].
There are currently a few randomized trials evaluating the different therapeutic options for nasal-type NK/T-cell lymphomas. The majority of studies are retrospective in nature and almost all are conducted in the geographical areas where the tumor is prevalent. Most authors reported on the use of RT alone or combined with CT, with RT being associated with high remission rates and prolonged survival, mainly in the case of localized disease [
34,
40,
44,
45]. In a study of 82 patients with localized disease, early RT was shown to be the only independent prognostic factor, with 5-year OS being significantly better in patients receiving >54 Gy [
46]. There appears to be a consensus that the optimal dose is 50 Gy, to be delivered to both the nasal cavity and sinuses. Nevertheless, concurrent CT may improve both local and systemic disease control. When comparing CT alone to RT alone or CT + RT, CR rates were significantly higher in patients treated with CT + RT [
13,
28,
34,
44,
47]. In addition, two recent reports on CT + RT for localized disease (stage IE to IIE) showed improved results compared to the historical controls using RT alone [
24,
48].
In our study, the overall response rate was 48 %, with progressive disease observed in 52 % of patients and no partial response. Combined therapy comprising RT and anthracycline-based CT regimens was associated with higher CR rates and longer OS compared to CT or RT alone, even in advanced-stage disease (CR rate of 31 %
vs 50 % for CT and CT + RT, respectively). However, the CT + RT cohort has a better prognostic profile than the CT population and this difference will create a significant bias in favor of CT + RT group, especially for this high stage group. No CR was recorded for RT alone, although this treatment was only administered to one elderly patient with a non optimal dose of 40 Gy. In the two published metaanalyses, first-line CT regimens involved CHOP, CHOP-like schedules, etoposide, ifosfamide, or cisplatin. In our study, however, the disease did not respond well to these combinations. The introduction of L-asparaginase-containing regimens led to further improvements, as most studies using asparaginase-based regimens in a relapsed or refractory setting reported response rates of around 50 %, with 5-year OS being 65 % (86 % for localized disease and 38 % for advanced stage) [
20,
22,
25,
48‐
50]. In our study, the different CT regimens were compared, with no statistical difference (but low numbers of patients in each type) found in CR rates according to the type of regimen. The use of L-asparaginase-based regimens in induction therapy did not improve CR rates, although only four patients were administered this regimen. Data regarding the use of L-asparaginase as first-line therapy is therefore still needed.
For all patients, 5-year OS and PFS rates were 39 % and 33 %, respectively. In addition, the quality of response after first-line treatment was found to be crucial for survival, with 5-year OS being 80 % in CR patients compared to 0 % in progressive disease patients.
Competing interest
The author(s) indicated no potential conflicts of interest.
Authors’ contributions
Conception and design: AC, A-SM and BC Provision of study materials or patients: all of the co authors Collection and assembly of data: AC and A-SM. Data analysis and interpretation: AC, A-SM and BC Manuscript writing: AC, A-SM, GS and BC. Final approval of the manuscript: All the co authors.