Diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma (NHL) and accounts for 30% to 40% of B-cell NHLs (B-NHL) (1). Gene expression profile has classified DLBCL into 3 molecular subtypes: germinal center B cell–like (GCB), activated B cell–like (ABC) and unclassifiable according to the cell of origin (COO) (2, 3). Alternatively, Hans algorithm is used to classify DLBCL into GCB and non-GCB subtypes based on approximation of immunohistochemistry (4). Studies have shown that the classification of DLBCL according to COO is an important prognostic factor in determining the treatment outcomes (5, 6). Non-GCB subtype showed significantly lower 5-year overall survival (OS) (54% vs. 78%) and progression-free survival (PFS) (48% vs. 76%) compared to GCB subtype (7).

Patients with DLBCL commonly present with a rapidly growing mass or enlarged lymph nodes at the nodal/extranodal site. Of all the cases of DLBCL, about one-third cases showed the primary extranodal sites of disease origin, whereas <70% showed involvement of at least one extranodal site. The most common extranodal sites of involvement include gastrointestinal tract, bones, testes, spleen, and central nervous system (CNS) (8–10). The presence of multiple extranodal involvements in DLBCL has been considered as an important factor in the International Prognostic Index (IPI) score because of associated poor prognosis (9).

In recent years, China has reported increased incidence of extranodal lymphomas, with non-GCB subtype DLBCL (70.7%) associated with significantly poor treatment response rates (75.5%) and overall survival rates (61.8%) as compared to single (84.6% and 80.4%) and no nodal involvement (89.3% and 81.3%) (11, 12). Rituximab in combination with CHOP regimen (R-CHOP) was used as a standard first-line treatment in DLBCL patients with a survival rate of 60% to 70%. However, 30% to 40% non-GCB DLBCL patients are either refractory or experience relapse in response to the first-line R-CHOP regimen and show poor prognosis and short OS rates (13–15). Therefore, newer treatment regimens with increased efficacy need to be developed for DLBCL patients non-responsive to routine first line treatment to achieve better remission and long-term survival.

There are several clinical trials to evaluate the effectiveness of new drugs alone or in combination with R-CHOP regimen in first-line treatment of DLBCL. However, currently only POLARIX study has shown PFS benefits with study drug compared to R-CHOP in first-line setting (16); in comparison with other studies which did not show any clear survival benefits. Nevertheless, different drugs were beneficial in different subgroups (Bcl-2 protein overexpression by immunohistochemistry, CD20+, ABC-type/non-GCB type DLBCL) in PFS, OS, or event-free survival (EFS) (17–19). Use of next-generation sequencing (NGS) in clinical studies to identify oncogenic mutations has enhanced our understanding about DLBCL heterogeneity which can facilitate precise diagnosis and effective treatment. In ABC subtype/non-GCB subtype of DLBCL, active B-cell receptor (BCR) signaling results in inhibition of apoptosis, thereby promoting uncontrolled cell division thus resulting in disease progression (PD), or can be responsible for insensitivity against chemotherapy (20–22). Non-GCB subtype tumors enriched for MCD (based on the co-occurrence of MYD88^L265P and CD79B mutations) and BN2 (based on BCL6 fusions and NOTCH2 mutations) subtypes can be treated using inhibitors targeting Bruton’s tyrosine kinase (BTK) present downstream to BCR and involved in activation of NF-κB activation (23, 24). In Phoenix study, the action of a first-in-class BTK inhibitor, ibrutinib, showed improved EFS and OS in in younger patients (aged <60 years) (19). According to the latest results, younger patients (aged <60 years) with BCR-dependent NF-κB pathway enriched MCD and N1 subtypes showed 100% 3-year EFS on treatment with ibrutinib plus R-CHOP (25). Thus, BTK inhibitors have shown promising results in the treatment of non-GCB DLBCL patients.

Zanubrutinib is a new type of small molecule oral BTK inhibitor, which can effectively inhibit BTK targets (26). In phase 1/2 clinical studies, zanubrutinib has shown promising results for safety and efficacy in patients with relapsed/refractory DLBCL (27). The present single-arm, phase 2, the prospective, single-center clinical trial has evaluated the efficacy and safety of the zanubrutinib in combination with R-CHOP in newly diagnosed non-GCB DLBCL with extranodal involvement (NCT04835870).

The present study reported a single-arm, prospective, single-center, phase 2 trial for newly diagnosed non-GCB DLBCL with extranodal involvement to evaluate the efficacy and safety of the ZR-CHOP treatment regimen. We have observed satisfactory therapeutic effects in patients of different ages, although without a large sample. This study showed an ORR of 91.3% (21/23), of which, CR was 82.61%, 1-year PFS and OS were 80.8% and 88.5% respectively while expected PFS and OS for 2-years are 74.0% and 88.5% respectively in ITT population with median follow-up of 16.7 months and the median DOR as of the last follow-up was 13.5 months (3.2~25.1months). Although the overall sample size may not be enough, we still observe a good benefit trend in these patients. The proportion of patients with grade ≥3 neutropenia in the R-CHOP cohort of GOYA study was 39.5%. While the present study observed a higher proportion of patients experiencing neutropenia (50%) than GOYA study, the proportion remained roughly the same for febrile neutropenia in both the studies (15.9% vs. 15.4%) (29). In the Phoenix study, more serious AEs (SAEs) were reported in the ibrutinib plus R-CHOP than in the placebo plus R-CHOP arm (19). Although almost no grade ≥3 neutropenia was observed in zanubrutinib monotherapy reported by Yang et al. (27), the combination of medications might have led to additional hematological toxicities. It has been reported that primary prophylaxis with pegfilgrastim tended to have a lower odds ratio for the occurrence of febrile neutropenia in patients aged 50-70 years and was significantly associated with reduced febrile neutropenia occurrence in patients aged 70-80 years (30). Therefore, the slight increase in the incidence of hematological toxicities might also be due to no use of prophylactic G-CSF. Crucially, these additional events were manageable using opportune symptomatic treatment such as anti-infection and G-CSF and did not translate into an increased risk of fatal AEs and death.

Non-GCB subtype of DLBCL with high Ki-67 index and overexpression of MYC and BCL2 proteins (double expression), insensitive to the standard first-line treatment show poor prognostic significance (31). Also, previous studies have shown positive correlation between high MYC and concurrent high MYC/BCL2 double-expression with individual markers of active BCR signaling, thereby showing enhanced BCR activation (32, 33). Similar results were seen in the present study. BCR and NF-κB pathway gene mutations were detected in 4 of the 8 patients with double-expression lymphoma, which showed effective response on zanubrutinib treatment. Form these 8 double-expression lymphoma patients, there were 7 patients with effective post-treatment response, 6 patients achieved CR, whereas all of them obtained continuous remission. Therefore, the use of BTK inhibitor has effectively helped in controlling increased BCR activation in these patients with BCR and NF-κB pathway gene mutations and thereby achieved satisfactory results. No incidence of increased risk of severe infection and hemorrhage was reported in patients on zanubrutinib regimen in combination with R-CHOP. Moreover, no serious AEs and fatal events were observed as well. The presence of zanubrutinib-related atrial fibrillation and skin and mucosal hemorrhage was manageable and did not lead to dose reduction or treatment discontinuation.

The BCR signaling pathway is widespread and plays an exceedingly vital role in the development of DLBCL, especially the ABC subtype of DLBCL relies on chronic active BCR signaling and the survival of ABC DLBCL cells depends on the BCR signaling pathway (21, 22). As zanubrutinib is a BTK inhibitor directly acting on BCR signaling, this treatment has been effective in a majority of patients. Previous studies have shown BCR-dependent NF-κB activation to be common in MCD and BN2 tumors. The frequency of MYD88 and CD79b mutations has been observed to be high in patients with primary extranodal involvement. All of these patients have shown a poor prognosis in response to the standard immunotherapy (12). To test the genetic predisposition of patients to treatment efficacy, oncogenic mutations closely related to DLBCL pathogenesis were assessed in 20 (76.9%) of 26 patients. TP53 mutations were detected in 5 patients. There was total 10 patients with BCR and NF-κB pathway gene mutations, involving MYD88, CD79B, CARD11, and TNFAIP3. Among these patients, 8 of 10 achieved a sustained response. Therefore, patients with these genetic lesions were effectively treated with BTK inhibitors and could achieve a better clinical effect. For other DLBCL patients with PD or relapse, their genetic mutations were involved in histone or DNA methylation, cell cycle, or p53 and not in BCR or NK-κB pathway. Since these mutations are nowhere linked with BCR signaling, BTK inhibitors may not be the best choice to effectively treat these patients.

Treatment responses have been analyzed by interim and end of treatment PET-CT response evaluation using Deauville scores. While the evaluation of interim PET-CT analysis using Deauville score has been observed to be useful in determining treatment prognosis in several clinical and retrospective studies (34, 35), there are studies reporting contradictory results as well (36–38). Kurch et al. have previously reported that lower Deauville score for the end of treatment PET-CT evaluation is associated with superior treatment response in patients with similar prognosis (37). In addition, ΔSUV_max is another parameter commonly used to evaluate treatment response with different implications involved in prediction of patient recurrence or survival at different thresholds (35, 36, 39, 40). Although the results from the present study do not show any unified conclusion, the trend observed in the results denotes that high ΔSUV_max could be a measure for better PFS and OS. The present study showed posttreatment median ΔSUV_max of 91.2% (67.6%-100%), which was higher than the median ΔSUV_max of 66% (41), 70% (41), even 82% (35) reported in the previous studies. This denotes relatively satisfactory (ideal) remission in patients after treatment. However, these results are not statistically significant, which may be due to lack of patients and follow-up time. This study hopes to get more predictive values from the medium-term PET-CT to make prompt adjustments in the treatment of patients with poor prognoses. Follow-up of these patients will be continued along with performing multicenter study for further verification.

In conclusion, this study demonstrated an encouraging clinical benefit and a tolerable safety of ZR-CHOP regimen in newly diagnosed DLBCL patients with extranodal involvement. Furthermore, this study provides evidence that ZR-CHOP regimen could enable more high-risk patients to achieve better efficacy at the end of the first-line treatment and may make it possible for some patients to get long-time remission in the first-line treatment. However, these results will need further confirmation through a large sample size randomized clinical trial.

The original contributions presented in the study are publicly available. This data can be found here: National Genomics Data Centre (NGDC), GSA, accession number HRA005297. Further inquires can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by ethics committee and institutional review board of the first affiliated hospital of Soochow University (Approval number: 2021035). The patients/participants provided their written informed consent to participate in this study.

YZ, JL, CL, and DW contributed to conception and design. XZ, YL, SL, JZ, and DW provided study material or patients. All authors participated in the collection and assembly of data. All authors analyzed and interpreted the data. HG wrote the manuscript. All authors gave final approval of the manuscript and agreed to be accountable for all aspects of the work.