# A novel immune model predicts the prognosis of mantle cell lymphoma

## TL;DR

This study creates a new immune-based model to predict outcomes for patients with mantle cell lymphoma, using blood markers and immune cell levels.

## Contribution

The novel contribution is an XGBoost-based immune prognostic model that outperforms existing MCL prognosis tools.

## Key findings

- High SUVmax (≥11) was identified as an independent predictor of poor progression-free and overall survival in MCL patients.
- A new immune model using SUVmax, LDH, IL-2, TNF-α, and CD4+ T cells showed better discrimination than MIPI and MIPI-C models.
- The model successfully categorized patients into high-risk and low-risk groups for prognosis.

## Abstract

Mantle Cell Lymphoma (MCL) is a subtype of B-cell lymphoma characterized by varied clinical manifestations. The immune status is associated with MCL’s development and outcome. This study aims to evaluate the prognostic value of peripheral cytokines and blood lymphocyte subsets in MCL patients.

This retrospective study analyzed patients’ clinical characteristics, treatment strategies, progression-free survival (PFS), and overall survival (OS). Immune cell levels and cytokines were evaluated via peripheral blood flow cytometry. Prognostic models incorporating immune characteristics were developed using XGBoost algorithms.

The study involved 78 MCL patients with a median follow-up period of 40 months. The median PFS and median OS were 32 and 48 months. Univariate analysis linked poor PFS to factors including elevated β2-MG, Ann Arbor stage III-IV, low albumin levels (<35 g/L), high SUVmax (≥11), reduced T cells (<70.42%), reduced CD4 + T cells (<34.63%), and increased NK cells (≥8.37%). Factors linked to poor OS included the pleomorphic and blastoid subtypes, albumin below 35 g/L, and high SUVmax (≥11). Multivariate analysis identified high SUVmax (≥11) as an independent predictor of poor PFS and OS in MCL patients. XGBoost regression and classification models were developed to determine feature importance, highlighting five key features: SUVmax, LDH, IL-2, TNF-α, and CD4 + T cells. A prognostic model using these immune features was created to predict patients’ PFS, dividing them into high-risk and low-risk categories. This model showed superior discriminatory power compared to the MIPI and MIPI-C models and had comparable calibration ability.

This study developed an innovative immune prognostic model for evaluating the prognosis of MCL patients, integrating immune factors with existing clinical features to improve prognostic evaluation.