Impact of Large Cells and Ki-67% on Outcomes in Patients with Newly Diagnosed Marginal Zone Lymphoma – Results from a Multicenter Cohort Study

Introduction
Marginal zone lymphoma (MZL) is an indolent B-cell lymphoma primarily composed of small/medium cells with low proliferation. However, there are cases where large B-cells (LC) are increased on the biopsy or Ki-67 expression is higher. The significance of LC content and Ki-67 in MZL is uncertain compared to other lymphomas (e.g., follicular or mantle cell). Previous studies evaluating prognostic indices in MZL or the risk of histologic transformation (HT) to diffuse large B-cell lymphoma did not assess these factors. Clinicians and patients often have concerns about the meaning and implications of increased LC or Ki-67 expression. We analyzed a large dataset of US MZL patients to evaluate the impact of LC and Ki-67 on outcomes.

Methods
This retrospective, multicenter study involved adults with MZL treated at 10 US medical centers on or after 1/1/2010. We included patients with splenic (SMZL), nodal (NMZL), and extranodal MZL (EMZL) of the mucosa-associated lymphatic tissue (MALT). Cases with information on LC or Ki-67+ cell percentages determined by immunohistochemistry on the diagnostic biopsy were included. Patients with HT at or within 60 days after MZL diagnosis were excluded. The primary endpoint was progression-free survival (PFS) and secondary endpoints included overall survival (OS) and cumulative incidence of HT. We assessed PFS and OS from diagnosis and the start of first-line treatment for patients receiving systemic therapy. Cox models were adjusted for age, sex, performance status, MZL subtype, stage, hemoglobin, albumin, and lactate dehydrogenase (LDH), and used multiple imputation for missing data. Adjusted hazard ratios (aHR) with 95% confidence intervals (CI) are reported.

Results
The study included 783 MZL patients (median age: 63 years; 53% women; 60% EMZL, 20% NMZL, 20% SMZL). Of these, 91 (12%) had LC reported in the biopsy, and 44 (6%) had >10% LC content (Fig. 1A).

Ki-67 staining was performed on biopsies from 438 MZL patients. Of those, 235 (54%) had low Ki-67 expression (≤10%), and 98 (22%) had Ki-67 >20%, which we determined to be the best prognostic cutoff for high Ki-67 in MZL. Ki-67 >20% correlated with increased LDH (34% versus 20%, P=.017), but not with other clinical parameters or with the use of first-line RCHOP chemotherapy (P=.07). The Ki-67 >20% group had significantly worse PFS (aHR=1.69, 95%CI=1.12-2.55, Fig. 1B) with a trend for worse OS from diagnosis (aHR=2.00; 95%CI = 0.97-4.15). These associations were statistically significant for survival outcomes measured from the start of first-line systemic therapy: both PFS (aHR = 1.84, 95%CI = 1.15-2.94) and OS (aHR = 2.71, 95%CI = 1.11-6.60). There were no significant differences in PFS or OS within the Ki-67 >20% group based on first-line therapy (R-CHOP vs. BR/RCVP). The cumulative incidence of HT was significantly higher in the high Ki-67 group (5-year risk, 9.8% versus 3.8%, p=0.013).

LC presence was associated with higher Ki-67 but not other clinical characteristics. Patients with LC-containing MZL received R-CHOP as first-line chemotherapy (18%) more frequently than those without LC (5%; P=.004). Presence of LC was linked to shorter PFS from diagnosis (aHR = 1.56, 95%CI = 1.08-2.26), but not OS (aHR = 1.07; 95%CI = 0.53-2.18). Presence of LC did not affect PFS or OS from the start of first-line systemic therapy (n=523). Among patients with LC, there was no significant difference in PFS or OS according to the type of first-line therapy (R-CHOP vs. BR/RCVP). The cumulative incidence of HT (31 events, 5-year risk 3.7% [95%CI, 2.4-5.3%]), was higher in the group with LC (5-year risk 9.4% versus 2.9%, P=0.04). Comparing patients with >10% LC to others yielded similar associations without statistical significance for any endpoint (possibly due to few patients with >10% LC).

Discussion
In this large retrospective cohort of newly diagnosed MZL patients, Ki-67 staining >20% was an independent prognostic factor for worse PFS and OS. LC content in the tumor was not significantly associated with most outcomes. Anthracycline-based chemotherapy did not impact PFS or OS in the presence of high Ki-67 or LC. These findings, along with the increased risk of HT, suggest that MZL with Ki-67 >20% may represent a higher-risk subset to prioritize for experimental approaches. Patients with LC or high Ki-67 at diagnosis should undergo a repeat biopsy at relapse to assess for HT.