Simultaneous Assessment of Total Metabolic Tumor Volume and Intratumoral CD4+ T-Cell Infiltration Identifies Early Progression of Follicular Lymphoma on Watch-and-Wait Approach

Introduction

Follicular lymphoma (FL) is an indolent lymphoma with heterogeneous clinical behavior. To refine risk stratification, several biomarkers have been explored. Total metabolic tumor volume (TMTV) derived from positron emission tomography/computed tomography (PET/CT) data has been shown to serve as an accurate tumor burden surrogate and be associated with outcomes in FL (Cottereau et al. Ann Oncol 2024). Importantly, a major caveat of TMTV is that it cannot purely represent the volume of malignant B-cells and can be influenced by intratumoral immune cells (Nath et al. Blood Advances 2021). However, there is very little data on the simultaneous assessment of radiomic parameters and tumor composition integrated with clinical outcomes in FL, especially on watch-and-wait (W/W) approach.

Methods

We retrospectively conducted a correlation analysis between quantitative PET/CT metrics and specific cell proportions within biopsied tumors obtained from FL patients. Among 223 patients with newly diagnosed FL between 2006 and 2023 at Kameda Medical Center, 136 were excluded solely due to insufficient biopsied samples. The maximum of standardized uptake value (SUVmax) and TMTV were semi-automatically calculated using the open-source software Metavol (Hokkaido University, Sapporo, Japan). TMTV was defined as sum of the volume of lymphoma-associated lesions with an SUV of ≧ 2.5 as the absolute threshold. Concerning the under- or overestimation during the processing of flow cytometry (FCM) (Xu et al. Br J Haematol 2001), we applied a whole-slide imaging (WSI) approach to quantify the cellular components across lymph node architecture. Excisional biopsy specimens were analyzed to count cells positively stained with CD20, CD3, CD4, CD8, and CD68 using the positive cell function of QuPath software (University of Edinburgh, Edinburgh, Scotland). Next, time to first treatment (TTFT) from diagnosis in W/W FL patients was analyzed using the Kaplan-Meier estimate and cox-proportional hazard model.

Results

In total, 87 patients were included. First, the variability of cellular population within tumors between individuals was observed, with the median % of CD20+, CD3+, CD4+, CD8+, and CD68+ cells being 57% (range 6.4%–84%), 44% (range 15%–74%), 31% (range 2.3%–64%), 13% (range 2.2%–35%), and 5.7% (range 0.3%–35%), respectively. There were no significant differences in the distribution of specific cell types according to pathological FL grading (grade 1 or 2 vs. 3a in the WHO 2018 classification). As anticipated, the quantified proportions of CD4+ cells were discordant between the WSI-based and FCM-based methods (Spearman’s correlation rank test, r = 0.49, P = 0.1), although a significant but modest correlation was observed for CD20+ cells (r = 0.6, P = 0.02). Regarding radiomic markers, the median SUVmax and TMTV were 10.3 (range: 4.2–42.4) and 262.6 mL (range: 2.6–4269.6), respectively. Remarkably, none of the specific cell types, even in CD20+ cells that largely represents tumorigenic B-cells, did not significantly correlated with the levels of SUVmax and TMTV as well as SUV at the biopsy site (all P > 0.05).

Finally, whether TTFT was affected by tumor burden and specific intratumoral T-cell infiltration (TIL) was examined in 28 FL patients initially under observation. Reflecting the fulfillment of the GELF criteria, these subgroups had a lower tumor burden with the median TMTV of 74.8 mL. When dichotomized by each median value, patients with high TMTV and low CD4+ TIL (< 36%) were extremely shorter TTFT than those with low TMTV and high CD4+ TIL (the median of 2.7 months vs. not reached, P = 0.00054). This categorization retained significant impact on TTFT (hazard ratio 2.33 [95% confidence interval 1.33–4.08], P = 0.002) after adjustment for a Follicular Lymphoma International Prognostic Index score of 3-5 (n=9/28, 32.1%).

Conclusions

We showed that intratumoral immune cells accurately quantified by WSI-based approach had minimal influence on the pretreatment PET/CT parameters in FL. Moreover, for the first time, we showed that the simultaneous assessment of tumor volume and composition can improve the identification of early progressors under observation.