Patients consecutively diagnosed as EBV-positive LAHS from January, 2010 to December, 2021 at West China Hospital, Sichuan University were enrolled (16, 17). EBV positivity was assessed by EBER in situ hybridization. Besides antiviral treatment, induction treatment was divided into anti-HLH treatment, anti-lymphoma treatment, and the combination of both. Anti-HLH treatment consisted of HLH-1994 regimen, HLH-2004 regimen, and GED regimen (gemcitabine-etoposide-dexamethasone). Anti-lymphoma treatment was chemotherapy depending on different subtypes of lymphoma. Treatment response was evaluated according to the following conditions. Complete response (CR) required disappearance of all HLH-related symptoms, and HLH-related lab findings returning to normal range, including ferritin, soluble CD25 (sCD25), complete blood count, triglyceride, alanine aminotransferase (ALT), etc. Consciousness should recover to normal if central nervous system was involved. Partial response (PR) required normal body temperature and improvement by over 25% in at least two symptoms or lab findings, including sCD25 decrease by at least one third, ferritin and triglyceride decrease by at least 25%, ALT decrease by at least 50% if formerly over 400U/L, and neutrophil increase by 100% without transfusion. If neutrophil counts were formerly less than 0.5×10⁹/L, they should exceed 0.5×10⁹/L after treatment. If neutrophil counts were formerly over 0.5×10⁹/L but less than 2×10⁹/L, they should exceed 2×10⁹/L after treatment. No response (NR) was defined as being unable to reach CR or PR. Overall response rate (ORR) was calculated as (CR+PR)/total patients × 100%. OS was defined as the interval between HLH diagnosis and all-cause death. The patients were followed up until December, 2021. This study was conducted with the approval of the West China Hospital ethics committee, according to Declaration of Helsinki.

Statistical analysis was conducted with SPSS 23.0 and GraphPad Prism 7.0 software. Measurable data with normal distribution and skewed distribution were evaluated with t test and U test, respectively. Enumeration data were evaluated with χ² test. Survival analysis was assessed with Kaplan-Meier methods, and compared with Log rank test. Univariate and multivariate analysis of prognostic predictors was conducted with Cox proportional hazard model. It was regarded as statistically significant with P<0.05.

Fifty-one EBV-positive LAHS patients were enrolled in this study, including 35 (68.6%) male patients and 16 (31.4%) female patients. Median age at HLH diagnosis was 37 (13-64) years. Three (5.9%) patients were newly diagnosed as lymphoma, while 48 (94.1%) patients were relapsed/refractory lymphoma cases. Clinical characteristics of these patients at baseline were listed in Table 1 . The total cohort consisted of 44 (86.3%) T/NK cell lymphoma cases, five (9.8%) B cell lymphoma cases, and two (3.9%) HL cases. Regarding T/NK cell lymphoma, extranodal NK/T-cell lymphoma (ENKL), nasal type (22/51, 43.1%) and aggressive natural killer cell leukemia (ANKL) (13/51, 25.5%) were the dominant subtypes. Besides, PTCL, not otherwise specified (PTCL, NOS) (4/51, 7.8%), angioimmunoblastic T cell lymphoma (AITL) (3/51, 5.9%), anaplastic large cell lymphoma (ALCL) (1/51, 2.0%), and systemic EBV-positive T-cell lymphoma of childhood (1/51, 2.0%) were also discovered. B cell lymphoma group included four (7.8%) DLBCL cases and one (2.0%) lymphoplasmacytic lymphoma (LPL) case. The majority of (94.1%) EBV-positive LAHS patients were at Ann Arbor stage III-IV. In terms of clinical manifestations, fever was present in all patients. Splenomegaly, serous effusion, hepatomegaly, edema, and jaundice were seen in 42 (82.4%), 28 (54.9%), 22 (43.1%), 12 (23.5%), and 10 (19.6%) patients, respectively. Hemophagocytosis phenomena were discovered in 24 (47.1%) patients.

Baseline characteristics of EBV-positive T/NK cell LAHS patients were compared to those of EBV-positive B cell LAHS patients. T/NK cell LAHS patients were significantly younger than B cell LAHS patients upon HLH diagnosis (34 v.s. 47 years, P=0.033). Fibrinogen level and C-reactive protein (CRP) level were significantly lower in T/NK cell LAHS patients than B cell LAHS patients (P=0.000 and P=0.004, respectively). Plasma EBV DNA level tended to be higher in B cell LAHS patients than in T/NK cell LAHS patients, but the difference was not significant. There was no significant difference in sex, IPI score, ECOG score, clinical manifestations, complete blood count, liver and kidney function, triglyceride, ferritin, sCD25, hemophagocytosis phenomenon, and the occurrence order of lymphoma and HLH between EBV-positive T/NK cell LAHS patients and EBV-positive B cell LAHS patients ( Table 2 ).

We also compared EBV-positive LAHS patients and EBV-negative LAHS patients who were consecutively diagnosed in our department during the same period of time. EBV-positive LAHS patients (median 37 years, range 13-64 years) were significantly younger at HLH diagnosis than EBV-negative LAHS patients (median 42 years, range 13-90 years, P=0.049). EBV-positive LAHS patients displayed significantly higher hemoglobin and triglyceride levels at baseline than EBV-negative LAHS patients (P=0.008, and P=0.007, respectively). T/NK-cell lymphoma tended to account for a higher proportion of EBV-positive LAHS patients (86.3%) than EBV-negative LAHS patients (69.8%), but the difference was insignificant ( Table S1 ).

Of the 51 patients, three patients only received antiviral treatment in our department, and two patients initiated HLH-1994 regimen but died in three days. After excluding these five patients, the remaining 46 patients were evaluable for treatment response. Treatment was divided into combined treatment group, anti-HLH treatment group, anti-lymphoma treatment group, and glucocorticoid treatment group ( Figure 1 ). Combined treatment of anti-HLH regimen and anti-lymphoma chemotherapy was conducted in 24 (52.2%) patients. Nineteen (41.3%) patients first underwent anti-HLH treatment of 3 (1-9) weeks, and then turned to anti-lymphoma chemotherapy. HLH-1994 regimen was conducted in 12 (26.1) patients. GED regimen was given to four (8.7%) patients, and HLH-2004 regimen was performed in three (6.5%) patients. Median number of chemotherapy cycles was 2 (1-8). The detailed anti-lymphoma chemotherapy was listed in Table 3 . Another five (10.9%) patients first received anti-lymphoma chemotherapy and turned to HLH-1994 regimen subsequently. These patients were all ENKL patients, and the detailed chemotherapy was also listed in Table 3 . The median number of chemotherapy cycles was 1 (1-3). The median weeks of HLH-1994 treatment was 4 (2-4).

In the 18 (39.1%) patients who only received anti-HLH treatment, the median weeks of induction treatment was 4 (1-8). The majority of (12/18, 66.7%) patients received HLH-1994 regimen. HLH-2004 regimen (3/18, 16.7%) and GED regimen (3/18, 16.7%) were conducted in three patients each. Three (6.5%) patients only received anti-lymphoma chemotherapy. One ENKL patient and one ANKL patient received GLIDE regimen, and one DLBCL patient received chidamide-RCHOP regimen. Median number of chemotherapy cycles was 3 (2-8). One (2.2%) patient only received glucocorticoid treatment.

For patients who were at higher risk or failed to achieve CR with a poor control of EBV infection, consolidation treatment was conducted, which included PD-1 monoclonal antibody treatment, chidamide treatment, and autologous or allogenic stem cell transplantation. For patients whose biopsy revealed positive PD-L1 staining of tumor cells, especially HL patients and DLBCL patients, PD-1 monoclonal antibody treatment was conducted. For patients who displayed aberrant epigenetic alterations, especially PTCL patients and double expression DLBCL patients, chidamide treatment was conducted. For young, fit, and high-risk patients with sufficient stem cells collected who were willing to receive transplantation, autologous or allogenic stem cell transplantation was conducted. Overall, seven (15.2%) patients underwent consolidation therapy. Three (6.5%) patients (one HL case, one DLBCL case, and one ENKL case) received PD-1 monoclonal antibody treatment, of whom the DLBCL patient also received autologous stem cell transplantation after PD-1 monoclonal antibody treatment. Two (4.3%) patients (one ENKL case and the DLBCL patient who received PD-1 monoclonal antibody treatment) received autologous stem cell transplantation. One (2.2%) ANKL patient received allogenic stem cell transplantation. Two (4.3%) AITL patients received chidamide treatment.

Of the 46 evaluable patients, CR was achieved in four (8.7%) patients, and PR was achieved in 18 (39.1%) patients. Twenty-four (52.2%) patients remained NR. ORR was 47.8%. In the combined treatment group, CR was reached in two (8.3%) patients, and PR was reached in 12 (50%) patients. NR was present in 10 (41.7%) patients. ORR was 58.3%. For patients who first received anti-HLH treatment and turned to anti-lymphoma chemotherapy subsequently, CR was achieved in two (10.5%) patients, and PR was achieved in 10 (52.6%) patients. Seven (36.8%) patients remained NR. ORR was 63.2%. For those who first received anti-lymphoma chemotherapy and turned to anti-HLH treatment subsequently, PR was achieved in two (40%) patients, and three (60%) patients remained NR. ORR was 40%. In anti-HLH treatment group, six (33.3%) patients achieved PR and twelve (66.7%) patients remained NR. ORR was 33.3% for the total anti-HLH treatment group, the HLH-1994 regimen subgroup, the HLH-2004 regimen subgroup, and the GED regimen subgroup. In anti-lymphoma treatment group, CR was achieved in two (66.7%) patients while one (33.3%) patient remained NR. ORR was 66.7%. One patient only received glucocorticoid treatment but remained NR. Patients who first received anti-HLH treatment and turned to anti-lymphoma treatment subsequently tended to show a higher ORR than those who received anti-HLH treatment only, but the difference was not significant (63.2% v.s. 33.3%, P=0.103). There was no significant difference in ORR among different treatment groups ( Figure 2 ).

At baseline, plasma EBV-DNA of all patients was positive (median 145000 copies/mL, range 647-744000000 copies/mL). After induction treatment, plasma EBV-DNA decreased significantly in all patients but one, who was a ANKL patient only receiving anti-HLH treatment but remaining NR (median 643.5 copies/mL, range 0-88800 copies/mL). Plasma EBV-DNA of ten patients after treatment was negative. Another three patients had a plasma EBV-DNA level lower than 50 copies/mL after treatment. There was no significant difference in plasma EBV-DNA after treatment between T/NK-cell LAHS (median 623 copies/mL, range 0-88800 copies/mL) and B-cell LAHS patients (median 14835 copies/mL, range 0-32300 copies/mL, P=0.540). In terms of treatment group, median plasma EBV-DNA of patients after treatment in combined treatment group and anti-HLH group was 272.5 copies/mL (range 0-64800 copies/mL), and 3130 copies/mL (range 0-33800 copies/mL), respectively. For the three patients in anti-lymphoma group, median plasma EBV-DNA of patients after treatment was 0 copies/mL (range 0-50 copies/mL). For the only patient in glucocorticoid group, his plasma EBV-DNA level decreased from 158000 copies/mL to 88800 copies/mL after treatment. There was no significant difference in plasma EBV-DNA level after treatment among treatment groups.

Regarding EBV-positive T/NK cell LAHS patients, three patients only received antiviral treatment in our department, and one patient initiated HLH-1994 regimen but died in three days. After excluding these patients, 40 patients were evaluable for treatment response. Twenty-one (52.5%) patients received combined treatment. Sixteen (40%) patients first underwent anti-HLH treatment of 3.5 (1-9) weeks, and then turned to anti-lymphoma chemotherapy. Median number of chemotherapy cycles was 1 (1-8). Another five (12.5%) patients first received anti-lymphoma chemotherapy and turned to HLH-1994 regimen subsequently. The median number of chemotherapy cycles was 1 (1-3). The median weeks of HLH-1994 treatment was 4 (2-4). In the 16 (40%) patients who only received anti-HLH treatment, the median weeks of induction treatment was 4 (1-8). Two (5%) patients only received anti-lymphoma chemotherapy. Median number of chemotherapy cycles was 2.5 (2-3). One (2.5%) patient only received glucocorticoid treatment.

Of the 40 evaluable EBV-positive T/NK cell LAHS patients, CR was achieved in three (7.5%) patients, and PR was achieved in 14 (35%) patients. Twenty-three (57.5%) patients remained NR. ORR was 42.5%. In the combined treatment group, CR was reached in two (9.5%) patients, and PR was reached in nine (42.9%) patients. NR was present in 10 (47.6%) patients. ORR was 52.4%. For patients who first received anti-HLH treatment and turned to anti-lymphoma chemotherapy subsequently, CR was achieved in two (12.5%) patients, and PR was achieved in seven (43.75%) patients. Seven (43.75%) patients remained NR. ORR was 56.25%. For those who first received anti-lymphoma chemotherapy and turned to anti-HLH treatment subsequently, PR was achieved in two (40%) patients, and three (60%) patients remained NR. ORR was 40%. In anti-HLH treatment group, five (31.25%) patients achieved PR and 11 (68.75%) patients remained NR. ORR was 31.25%. In anti-lymphoma treatment group, CR was achieved in one (50%) patient while one (50%) patient remained NR. ORR was 50%. One patient only received glucocorticoid treatment but remained NR. There was no significant difference in ORR among different treatment groups.

We also compared treatment and response of EBV-positive LAHS patients and EBV-negative LAHS patients in our department. Treatment choice for these patients was significantly different, with a higher proportion of EBV-positive LAHS patients (52.2%) receiving combined treatment than EBV-negative LAHS patients (31.4%, P=0.033, Table S2 ). The ORR of EBV-positive LAHS patients (47.8%) was similar to that of EBV-negative LAHS patients (51.4%). The ORR of EBV-negative LAHS patients who received combined treatment (72.7%) tended to be higher than that of EBV-positive LAHS patients (58.3%), especially for those who first underwent anti-HLH treatment and turned to anti-lymphoma treatment subsequently (80% v.s. 63.2%), but the difference was insignificant ( Table S3 ).

With a median follow-up of 79.5 (10-605) days, the median OS of the total cohort was 61 (95% CI 47.9-74.1) days. The median OS of EBV-positive NK/T cell LAHS patients was 65 (95% CI 52.7-77.3) days, while that of EBV-positive B cell LAHS or EBV-positive HL LAHS was not evaluable due to limited sample size. Patients whose baseline plasma EBV DNA exceeded 10⁵ copies/mL (median 51 days, 95%CI 30.6-71.4 days) tended to display worse OS compared to patients whose plasma EBV DNA was no more than 10⁵ copies/mL (median 188 days, 95%CI 151.0-225.0 days), but the difference was not significant (P=0.159) ( Figure 3A ). ENKL or ANKL patients (median 56 days, 95%CI 42.8-69.2 days) tended to have a poorer OS than patients with other subtypes of lymphoma (not reached), but the difference was not significant (P=0.060) ( Figure 3B ). Patients whose baseline ALT exceeded 80 U/L (median 42 days, 95%CI 17.6-66.4 days) had a significantly worse OS than patients with baseline ALT no more than 80 U/L (median 305 days, 95%CI 95.1-514.9 days, P=0.041) ( Figure 3C ). EBV-positive LAHS patients who received combined treatment (median 78 days, 95%CI 0.1-230.5 days) had a significantly improved OS compared to patients who only received anti-HLH treatment (median 31 days, 95%CI 18.1-43.9 days) (P=0.003) ( Figure 3D ). Specifically, for patients who received combined treatment, those who first received anti-HLH treatment and turned to anti-lymphoma treatment subsequently (median 174 days, 95%CI 0.1-371.9 days) displayed a significantly better OS than patients who only received anti-HLH treatment (median 31 days, 95%CI 18.1-43.9 days, P=0.003) ( Figure 3E ). Regarding treatment response, EBV-positive LAHS patients who achieved PR or above (median 305 days, 95%CI 54.2-555.8 days) had a significantly improved OS compared to those who remained NR (median 41 days, 95%CI 31.1-50.9 days, P=0.006) ( Figure 3F ). There was no significant influence of other baseline characteristics on OS. For patients who did not receive consolidation treatment, median OS was 56 days (95%CI 28.9-83.1 days). For patients who underwent consolidation treatment, median OS was not reached. Patients who underwent consolidation treatment had a significantly better OS than those without consolidation treatment (HR 0.203, 95%CI 0.048-0.865, P=0.031, Figure S1 ). Of the three patients who received PD-1 monoclonal antibody consolidation treatment, one ENKL patient died at 55 days after HLH diagnosis. A DLBCL patient and a HL patient were still alive after a follow-up of 673 days and 1484 days, respectively.

In univariate analysis, lower baseline ALT level, combined treatment of first anti-HLH regimen and subsequent anti-lymphoma regimen, achieving CR or PR were significantly related to improved OS (P=0.001, P=0.004, and P=0.009, respectively). ENKL or ANKL tended to associate with worse OS (P=0.069). There was no significant association between plasma EBV-DNA level after treatment and OS. There was no significant relationship between other baseline characteristics and OS. When combining baseline ALT level, combined treatment of first anti-HLH regimen and subsequent anti-lymphoma regimen, treatment response, and ENKL or ANKL subtype in multivariate analysis, elevated baseline ALT level was the independent risk factor of OS (P=0.025). Combined treatment of first anti-HLH regimen and subsequent anti-lymphoma regimen and achieving CR or PR were independent protective factors of OS (P=0.001, and P=0.025, respectively) ( Table 4 ).

Regarding EBV-positive T/NK cell LAHS patients, univariate analysis revealed lower ALT level, lower IL6 level, higher PT level, combined treatment of first anti-HLH regimen and subsequent anti-lymphoma regimen, achieving CR or PR were significantly related to improved OS (P=0.001, P=0.016, P=0.044, P=0.017, and P=0.013, respectively). When combining these factors in multivariate analysis, lower ALT level, combined treatment of first anti-HLH regimen and subsequent anti-lymphoma regimen, and achieving CR or PR were independent protective factors of OS (P=0.038, P=0.040, and P=0.041, respectively).

We also compared prognosis of EBV-positive LAHS patients and EBV-negative LAHS patients in our department. OS of EBV-positive LAHS patients was similar to that of EBV-negative LAHS patients (61 v.s. 42 days, P=0.997, Figure S2 ). In univariate analysis of EBV-negative LAHS patients, baseline ferritin more than 2000 ng/mL (HR=4.222, 95%CI 1.423-12.528, P=0.009), and achieving CR or PR were significantly related to improved OS (HR=0.168, 95%CI 0.053-0.536, P=0.003). When combining both factors in multivariate analysis, only achieving CR or PR remained significantly related to improved OS (HR=0.228, 95%CI 0.057-0.917, P=0.037), serving as protective factor of prognosis.