Timing of chemotherapy-induced neutropenia is a prognostic factor in patients with diffuse large B-cell lymphoma: a retrospective study

Chemotherapy-induced neutropenia (CIN) has been shown to be associated with improved clinical outcomes in patients with various solid tumors. The aim of this study was to investigate the relationship between the timing and degree of chemo-induced neutropenia (CIN) and short-term ecacy and survival in newly diagnosed patients with diffuse large B-cell lymphoma (DLBCL). A retrospective study was conducted on 236 newly diagnosed DLBCL patients who received at least 6 cycles of R-CHOP (like) or CHOP (like) between January 2012 and December 2018. According to the occurrence time of CIN, subjects were divided into CIN-absent group, early-onset CIN group and late-onset CIN group. According to the degree of CIN, they were divided into CIN-absent group, mild (grade 1-2) CIN group, and severe (grade 3-4) CIN group. Short-term ecacy was evaluated after 4 cycles of treatment. The Kaplan-Meier method was used to draw the survival curve, and the Cox proportional hazards model was applied to determine the correlation between the timing and extent of CIN and clinical features, short-term ecacy, progression-free survival (PFS) and overall survival (OS). overall survival rate 3 64%, and grade for the differences in results follows: 1) the cut-off points for evaluating CIN degree between the two studies were different. The cut-off point for Michael et al.’s study was the CIN degree after the rst treatment cycle, while our study was based on the third cycle. If CIN occurred during the rst three cycles, the lowest CIN degree of the previous three cycles was taken as the cut-off point. Otherwise, the cut-off point was the lowest CIN degree that occurred in the following cycle. 2) The total number of treatment cycles was different. There was no requirement for the number of treatment cycles in Michael et al.’s study, but at least 6 treatment cycles were required in this study, which may have excluded certain patients who could not complete 6 cycles of chemotherapy due to age, poor physique, severe myelosuppression and other reasons.

Background DLBCL is the most common subtype of adult non-Hodgkin's lymphoma [1,2], and its morbidity and mortality rates have been steadily increasing in China [3,4]. The current rst-line standard treatment for DLBCL is a combination of rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R-CHOP) or the latter 4 without rituximab (CHOP) [5], both of which can cure about 50-60% of patients. However, for patients who had failed prior treatment or relapsed after remission, outcomes were particularly poor [6]. In the past, when exploring better initial treatment options, other treatment schemes have failed to show a better effect than CHOP or R-CHOP; therefore, improving the therapeutic effect of CHOP or R-CHOP and improving DLBCL prognosis are areas worth exploring.
DLBCL is a chemotherapy-sensitive cancer, and the curative effect is proportional to the drug dose to a certain extent; therefore, it is worth studying individualised therapeutic doses to improve the curative effect. Through evaluations before and during treatment, an individualised chemotherapy plan can be made by adjusting the dose in time to determine the ideal amount for each patient. In R-CHOP or CHOP treatment regimens, the common adverse reaction is neutropenia [7,8], a major factor that leads to the reduction of subsequent doses and additional treatment costs [9,10]. However, the occurrence of CIN seems to represent dose intensity, and may be related to DLBCL prognosis. Michael et al. [11]studied the relationship between CIN severity and prognosis in 965 DLBCL patients after the rst cycle of R-CHOP treatment, and found that grade 1-2 CIN had the best prognosis (Grade 0 patients had a 67% 5-year overall survival rate grade 1-2: 78%, grade 3: 64%, and grade 4: 57%.). Similarly, for gastric cancer [12], lung cancer [13], pancreatic cancer [14], breast cancer [15], colorectal cancer [16] and other diseases, it has been shown that the occurrence of CIN is related to prognosis. Additionally, CIN timing also seems to be associated with survival, although previous research has shown that the dividing line between early and late occurrence differs. At present, studies have shown that the early onset of CIN in gastric cancer [17], lung cancer [18,19], pancreatic cancer [20] and colon cancer [21] indicates a longer survival time. However, the relationship between CIN timing and DLBCL prognosis is not clear.
The purpose of this study was to examine the relationship between the timing and degree of CIN in newly diagnosed DLBCL patients treated with R-CHOP and the short-term e cacy and prognosis, in order to improve the initial treatment e cacy of DLBCL and achieve a better prognosis.

Patients
Patients diagnosed with DLBCL and admitted to Shandong Cancer Hospital from November 2011 to December 2018 were enrolled in this retrospective study. Approval was granted by the Ethics Committee of Shandong Cancer Hospital, and all clinical records came from the electronic medical record database of Shandong Cancer Hospital.
The study criteria were as follows: 1) con rmation of DLBCL by pathological diagnosis with no prior treatment; 2) at least 6 cycles of CHOP or CDOP (liposome doxorubicin instead of epirubicin) or R-CHOP or R-CDOP regimen after diagnosis and a standard dose initial treatment regimen; 3) no bone marrow in ltration. The exclusion criteria were as follows: 1) incomplete bone marrow toxicity records; 2) lack of follow-up; 3) primary diffuse large B-cell lymphoma of the central nervous system or transformed large B-cell lymphoma of the central nervous system; 4) primary treatment in other hospitals; 5) presence of second or multiple cancers. According to the inclusion and exclusion criteria, 236 patients with DLBCL were identi ed and chosen for this study. The speci c selection criteria are shown in Figure 1.

Assessment of neutropenia
Routine blood tests were performed every 3-5 days from the rst day of chemotherapy until the following cycle. In this study, CIN severity was determined by the minimum absolute neutrophil count (ANC) in the peripheral blood tested during this period. According to National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events was regarded as the bound, early-onset was CIN occurring during weeks 1-3, late-onset was CIN occurring during the fourth cycle or later, and absence of CIN was no CIN occurrence during the entire treatment. Among them, the late-onset group and CIN-absent group were collectively referred to as the non-early-onset group. E cacy and survival assessment E cacy was evaluated according to imaging remission (CT/MRI) or metabolic remission (PET/CT) [22] as per the revised 2014 Lugano criteria. E cacy was evaluated every two cycles after treatment and divided into complete response (CR), partial response (PR), stable disease (SD) and progression disease (PD). The objective response rate (ORR) was the proportion of patients with CR and PR.
The follow-up end points were OS and PFS, and the follow-up deadline was January 30, 2021. OS is de ned as the duration from the beginning of treatment to date of death for any cause. PFS is de ned as the duration from the beginning of treatment to time of tumor progression or death of any cause.
Statistical analysis IBM SPSS Statistics software (version 22.0) and GraphPad Prism software (version 5.0) were used for the statistical analysis and to generate the graphs. Descriptive statistics was used to describe the baseline characteristics of patients, and a chi-square test was used to compare the baseline characteristics. The Kaplan-Meier method (logarithmic rank test) was used to construct the survival curve. Prognostic factors of OS and PFS were determined through univariate and multivariate Cox regression (enter method) analyses. Variables with P<0.1 from the univariate analysis were entered into multivariate analysis. P < 0.05 was de ned as a statistically signi cant difference.

Results
Demographics A total of 236 eligible patients were included in this study. The ratio of males to females was approximately 1:1.3, and the median age was 54.9 years (18-80 years Figure 2 and Figure 3, respectively. As shown in Figure 2, when using Kaplan-Meier analysis, every survival curve was meaningful regardless of the treatment regimen used, and the OS and PFS of the early-onset group were longer than those of the non-early-onset group.
To evaluate the prognostic values of CIN timing and degree, we performed univariate and multivariate COX regression analysis. The analysis results are shown in Table 3(See the table at  However, no independent correlation was observed between CIN degree and PFS and OS.

Discussion
Although CIN is the most common adverse reaction in cancer patients post-treatment, its occurrence and severity does not represent a poor prognosis. This study is the rst to report on the relationship between CIN timing and degree and survival in DLBCL patients. Studies have shown that CIN timing is an independent prognostic factor for patients with DLBCL. We found that the PFS and OS in the early-onset CIN group were signi cantly longer than those of in the nonearly-onset group (late-onset and absent CIN), which is consistent with the ndings of previous studies on other types of cancers.
In several different cancers, the early occurrence of CIN represents a good prognosis. He et al. [23] studied the relationship between CIN timing and survival after 6 cycles of carboplatin combined with paclitaxel in patients with primary serous ovarian cancer. They found that the median PFS of the early-onset group VS non-early-onset group was 23 months VS 9 months (P < 0.001), and the median OS of the early-onset group VS non-early-onset group was 55 months VS 24 months (P < 0.001), and concluded that the early onset of CIN can signi cantly improve PFS and OS.
Similar results have also been found in studies on gastric cancer [17], colorectal cancer [21] and other diseases.
Although these ndings are consistent with our results, it is worth noting that our results also showed no difference in PFS P=0.645 and OS P=0.500 between patients with late-onset CIN and absent CIN in the non-early-onset group.
The speci c mechanism of the good survival of patients with early-onset CIN is not clear, which may be related to the following points: 1) Early-onset CIN may indicate that patients are sensitive to drugs. Kvinnsland [24]stated that the sensitivity of treatment is a re ection of genetic susceptibility, and the sensitivity of neutrophils and cancer cells to drugs is theoretically the same. Non-early-onset CIN indicates less chemosensitivity and drug resistance, which possibly explains why the short-term effect seen in the early-onset group was better than that of the late-onset group and absent group. 2) CIN re ects the chemotherapy dose and pharmacokinetics of patients. The standard dose of R-CHOP or CHOP is calculated based on body surface area (BSA). However, due to individual differences in pharmacokinetics and pharmacodynamics, the drug dose calculated by BSA may lead to drug dose de ciency in certain patients [25][26][27]. CIN occurrence may indicate that a patient received an adequate treatment dose. 3) Neutrophils are related to the progression or prognosis of tumours. Studies have shown that neutrophils play an important role in cancer progression by interacting with cancer cells and immune cells in the blood and tumour microenvironment (TME). They are involved in a variety of functions that promote tumour growth, including proliferation, invasiveness and spread, and immunosuppression [28][29][30]. In addition, tumour-related neutrophils (TANs) are closely related to a patient's prognosis. Manfroi et al. [31] analysed the relationship between TANs and prognosis in 233 DLBCL patients using the R-CHOP regimen, and used the neutrophil elastase (ELANE) gene as a marker of TAN in ltration. The results showed that the high expression of ELANE was associated with decreased overall survival (HR=2.3, 95% CI: 1.2-4.3, P=0.01), which suggests that DLBCL patients with increased TANs may have a poor prognosis.
Additionally, our study shows that CIN occurrence is not an independent prognostic factor of PFS and OS, which differs from the results of Michael et al. [11], who found that grade 1-2 CIN leads to the best prognosis (5-year overall survival rate of patients with grade 0 was 67%, grade 1-2 was 78%, grade 3 was 64%, and grade 4 was 57%). The reasons for the differences in results may be as follows: 1) the cut-off points for evaluating CIN degree between the two studies were different. The cut-off point for Michael et al.'s study was the CIN degree after the rst treatment cycle, while our study was based on the third cycle. If CIN occurred during the rst three cycles, the lowest CIN degree of the previous three cycles was taken as the cut-off point. Otherwise, the cut-off point was the lowest CIN degree that occurred in the following cycle. 2) The total number of treatment cycles was different. There was no requirement for the number of treatment cycles in Michael et al.'s study, but at least 6 treatment cycles were required in this study, which may have excluded certain patients who could not complete 6 cycles of chemotherapy due to age, poor physique, severe myelosuppression and other reasons.
However, there were several limitations in this study. 1) This study was a single-center retrospective study, and the sample size was limited.
2) The follow-up time was relatively short, and no end point event occurred in many patients. Despite these limitations, our study still suggests that it seems possible to adjust the dose according to the presence of CIN in 1-3 cycles and choose the appropriate dose for each patient to achieve a better therapeutic e cacy of DLBCL. Of course, this needs to be veri ed by prospective tests.

Conclusion
The results of the present study suggest that CIN timing onset is a potential prognostic marker in newly diagnosed DLBCL patients who receive 6 cycles of R-CHOP or CHOP regimens, and early onset CIN leads to a better prognosis.