Ratio of Circulating CD8 ⁺ T Lymphocytes to M-MDSCs (CD8MMR): A Novel Prognostic Predictor for Treatment-Naïve DLBCL Patients

Introduction:

CD8 ⁺ T lymphocytes are crucial in contemporary cell-and-immunotherapies as they can be functionally enhanced or manipulated. Monocytic myeloid-derived suppressor cells (M-MDSCs) are known for their intense immunosuppressive ability, and higher percentages of circulating M-MDSCs among peripheral-blood monocytes or white cells are associated with a worse prognosis. Recently, a low lymphocyte-to-monocyte ratio (LMR) has been linked to a poor prognosis in lymphoma patients. Consequently, we hypothesized that the ratio of circulating CD8 ⁺ T lymphocytes to M-MDSCs (CD8MMR) could be a novel and more precise predictor for the clinical outcomes of DLBCL patients before initiating treatment.

Methods:

This prospective, observational study enrolled 136 adults with treatment-naïve DLBCL from May 2019 to March 2023 in a tertiary medical center in Taiwan. Patients with primary CNS lymphoma or primary mediastinal B-cell lymphoma were excluded. Fresh peripheral blood samples were obtained to calculate the absolute counts of CD8 ⁺ T lymphocytes and M-MDSCs, defined as alive singlets with CD45 ⁺ CD3 ⁺ CD8 ⁺ and CD45 ⁺ CD15 ^- CD14 ⁺ CD11b ⁺ CD33 ⁺ HLA-DR ^low/- by flow cytometry, respectively. ROC analysis and AUC determined the cut-off values for LMR and CD8MMR.

Results:

For 136 treatment-naïve DLBCL patients, the median age was 70, with 61.8% being male. Among them, 37.5% were in stage IV, and 27.9% were classified as the germinal-center B-cell (GCB) subtype by IHC staining according to the Hans algorithm. Additionally, 43.4% were diagnosed with double expressor lymphoma (DEL). The IPI risk score distribution was as follows: 7.4%, 18.4%, 17.6%, 24.3%, 19.9%, and 12.5% were in scores 0, 1, 2, 3, 4, and 5, respectively. Of note, the median level of LMR was 2.17, the median level of absolute counts of circulating CD8 ⁺ T lymphocytes was 385/µL, the median level of absolute counts of circulating M-MDSCs was 102/µL, and the median level of CD8MMR was 3.56.

DLBCL patients with high-risk IPI had a significantly lower CD8MMR than patients with intermediate or low-risk IPI (1.34 vs. 4.51 vs. 7.82, P_{H vs. L} = 0.001, P_{H vs. I} = 0.060, P_{I vs. L} = 0.082). DLBCL patients with the non-GCB subtype (2.75 vs. 5.63, P = 0.031) also demonstrated a significantly lower CD8MMR. Additionally, there were trends of lower CD8MMR in elderly patients > 70 (2.74 vs. 3.91, P = 0.209) and patients with DEL (3.33 vs. 4.65, P = 0.101).

After a median follow-up of 22.6 months for the 136 DLBCL patients, LMR was at first tested, and patients with LMR < 1.19 had significantly worse PFS (13.8 months vs. non-reach, Log Rank P = 0.004) and OS (non-reach vs. non-reach, Log Rank P = 0.019) than those with LMR ≥ 1.19. Next, we examined the CD8MMR, and patients with CD8MMR < 4.40 experienced significantly worse PFS (21.3 months vs. non-reach, Log Rank P < 0.001) and OS (non-reach vs. non-reach, Log Rank P < 0.001) compared to those with CD8MMR ≥ 4.40 [Figure].

We subsequently conducted a multivariate Cox regression analysis to validate whether CD8MMR < 4.40 could serve as an independent predictor of poor prognosis [Table]. The results showed that “CD8MMR < 4.40” was confirmed as an independent prognostic factor for both PFS (HR = 2.747; 95% CI = 1.275 to 5.918; P = 0.010) and OS (HR = 3.669; 95% CI = 1.462 to 9.211; P = 0.006) after adjusting for confounding factors, including age, sex, IPI risk scores, bulky mass > 10 cm, non-GCB type, DEL, and LMR < 1.19.

In this study cohort, 106 (77.9%) patients received R-CHOP-like regimens as induction therapies, 25 (18.4%) patients received an R-EPOCH regimen, two (1.5%) patients received an R-HyperCVAD regimen, and three (2.2%) patients died before completing the first cycle of chemoimmunotherapy. Among the 106 DLBCL patients receiving R-CHOP-like regimens, 57 patients showed CD8MMR < 4.40 and reported a significantly lower rate of complete remission (73.7% vs. 95.9%, P = 0.006) at the end of induction therapy compared to the 49 patients with CD8MMR ≥ 4.40. Furthermore, patients with CD8MMR < 4.40 had significantly worse PFS (17.4 months vs. non-reach, P < 0.001) and OS (non-reach vs. non-reach, P < 0.001) compared to those with CD8MMR ≥ 4.40.

Conclusion:

A low ratio of circulating CD8 ⁺ T lymphocytes to M-MDSCs (CD8MMR) serves as a poor prognostic factor for both PFS and OS in treatment-naïve DLBCL patients. This finding warrants further investigation and highlights the possibility of risk-adapted strategies when treating DLBCL patients.