AUTHORS
Siddh van Oost, PhD, MSc, Sanne Tonino2, Ruben AL De Groen, PhD, MSc3, Lorraine M. de Haan, MD4, Fleur A. De Groot5, Susan Blommers6, Romée E.W. Jansen7, Esther J. Kret8, Rosanne E. van der Post, MSc9, Joost A.A. Scheffers, MD10, Julia J. Whien, MSc11, Vera I. Zwarst12, Marijn A. Gillissen, MD, PhD13, Dina Ruano, PhD, MSc14, Marieke E. IJsselsteijn, PhD, MSc15, Noel F.C.C. De Miranda, PhD, MSc16, Martin Dreyling, MD, PhD17, Kim M. Linton Sr, MBChB, MD18, Dana Chitu, PhD, MSc19, Daphne De Jong, MD, PhD20, Jeanette Doorduijn, MD21, Josée Zijlstra22, Marie Jose Kersten, MD23, Joost S.P. Vermaat, MD, PhD, MSc24, Tim J.A. Dekker, MD, PhD25
AFFILIATIONS
1 Leiden University Medical Center, Leiden, Netherlands
2 Amsterdam UMC, University of Amsterdam, Cancer Center Amsterdam, Amsterdam, Netherlands
3 Leiden University Medical Center, Leiden, Netherlands
4 Leiden University Medical Center, Leiden, Netherlands
5 Leiden University Medical Center, Leiden, Netherlands
6 Leiden University Medical Center, Leiden, Netherlands
7 Leiden University Medical Center, Leiden, Netherlands
8 Leiden University Medical Center, Leiden, Netherlands
9 Leiden University Medical Center, Leiden, Netherlands
10 Leiden University Medical Center, Leiden, Netherlands
11 Leiden University Medical Center, Leiden, Netherlands
12 Leiden University Medical Center, Leiden, Netherlands
13 Leiden University Medical Center, Leiden, Netherlands
14 Leiden University Medical Center, Leiden, Netherlands
15 Leiden University Medical Center, Leiden, Netherlands
16 Leiden University Medical Center, Leiden, Netherlands
17 LMU University Hospital, Munich, Germany
18 The Christie NHS Foundation Trust, Manchester, United Kingdom, University of Manchester, Manchester, United Kingdom
19 Erasmus University Medical Center, Rotterdam, Netherlands, HOVON Foundation, Rotterdam, Netherlands
20 HOVON Pathology Facility and Biobank, Amsterdam, Netherlands, Netherlands Cancer Institute, Amsterdam, Netherlands
21 Erasmus MC Cancer Institute, University Medical Center Rotterdam, Rotterdam, Netherlands
22 Amsterdam UMC, location VUmc, Cancer Center Amsterdam, Amsterdam, Netherlands
23 Amsterdam UMC, University of Amsterdam, Cancer Center Amsterdam, Amsterdam, Netherlands
24 Leiden University Medical Center, Leiden, Netherlands
25 Leiden University Medical Center, Leiden, Netherlands
ABSTRACT
Despite the clinical efficacy of rituximab and lenalidomide (R2) with or without bendamustine (R2B) in relapsed/refractory follicular lymphoma, the biological basis of heterogeneous clinical outcomes remains poorly understood. We therefore conducted a prospective translational study within the randomized HOVON110/ReBeL trial (PMID: 41988778), integrating spatial and transcriptomic profiling to identify immune phenotypes associated with treatment outcome.
Pre-enrollment tumor biopsies obtained at primary diagnosis, relapse, or both were available from 55 patients (R2, n=20; R2B, n=35). Imaging mass cytometry (IMC; n=42), targeted transcriptomic profiling (n=55), and targeted next-generation sequencing (tNGS; n=44) were performed in parallel. Baseline patient characteristics were balanced between treatment arms. Median age was 64 years, 59% of patients had Follicular Lymphoma International Prognostic Index (FLIPI) scores of 3-5 (high-risk), and 79% had received a single prior line of therapy, most commoly (R)-CVP. Complete metabolic response by PET/CT was achieved in 57% of patients. Estimated 2- and 5-year progression-free survival (PFS) were 70% and 45%, respectively.
To comprehensively characterize the follicular lymphoma tumor microenvironment, unsupervised clustering of IMC-derived lymphoid, myeloid, and stromal populations identified 4 biologically distinct immune phenotypes: B cell-dominant (n=16), enriched for B cells; immune-rich (n=11), enriched for CD4+/CD8+ T cells and antigen-presenting cell (APC)-like macrophages; stroma-rich (n=8), enriched for SPARC+ stromal cells and CD163+ macrophages; and germinal center (GC)-dominant (n=7), enriched for B cells together with follicular helper T cells (PD-1+CD4+) and regulatory T cells, alongside APC-like macrophages. The 4 IMC-defined immune phenotypes were significantly associated with 5-year PFS (log-rank p=0.031), with outcomes ranging from no progression events in the stroma-rich subgroup to a median PFS of 36 months in B cell-dominant tumors.
Independent transcriptomic profiling using established microenvironmental signatures (Kotlov et al. PMID: 33541860) identified 4 transcriptional immune phenotypes: immune-depleted, inflammatory, mesenchymal, and intermediate. These transcriptional phenotypes broadly corresponded to the IMC-derived immune phenotypes. Concordance between transcriptomic and IMC-based classifications was 58% (p=9.1 x 10^-5), with strongest agreement for the immune-depleted/B cell-dominant (78%), mesenchymal/stroma-rich (62%), and inflammatory/immune-rich (57%) phenotypes.
To translate the IMC-defined immune phenotypes into a transcriptome-based classifier, a generalized linear model with LASSO regularization was trained using paired IMC and transcriptomic data (n=42). This yielded a 31-gene signature comprising lineage-defining markers of B cells (FCRL5), T cells (CD3E), macrophages (CD163), NK cells (GNLY), and stromal cells (POSTN). The classifier was then applied to tumors lacking IMC data, expanding immune phenotype classification to the complete cohort (n=55), comprising 23 B cell-dominant, 11 GC-dominant, 12 immune-rich, and 9 stroma-rich tumors. The prognostic relevance of the 4 immune phenotypes was maintained following transcriptome-based assignment (log-rank p=0.029). B cell-dominant tumors continued to exhibit the poorest outcome (median PFS, 19 months) and remained independently associated with inferior 5-year PFS relative to immune-rich tumors after adjustment for treatment arm (HR 4.42, 95% CI 1.28–15.3, p=0.019).
Matched tNGS data from the same samples are currently being analyzed to identify the genetic determinants of immune phenotypes. The availability of paired pre-enrollment primary and relapse biopsies further provides a unique opportunity to investigate the interplay between tumor genetics, immune contexture, and disease progression.
Collectively, this prospective translational study identifies 4 reproducible immune phenotypes in follicular lymphoma that are independently captured by spatial and transcriptomic profiling and associated with clinical outcome following R2(B)-based therapy, highlighting tumor immune contexture as a potential determinant of disease progression.
DISCLOSURES
Siddh van Oost, PhD, MSc: No relevant conflicts of interest to declare.