AUTHORS
Lena Osswald, MD, Jan Boegeholz, MD2, Jurik Andreas Mutter, PhD3, Takeshi Sugio, MD, PhD4, Feng Tian, PhD5, Mari Olsen6, Chih Long Liu, PhD7, Inhye E Ahn, MD8, Adrian Wiestner, MD, PhD9, Eugen Tausch, MD, PhD10, Stephan Stilgenbauer, MD11, Sandra Robrecht, MD, PhD12, Barbara Eichhorst, MD13, Michael Hallek Jr, MD14, Maximilian Diehn, MD, PhD15, Othman Al-Sawaf, MD, PhD16, Ash A. Alizadeh, MD, PhD17
AFFILIATIONS
1 Stanford University, Stanford, California, United States
2 Stanford University, Stanford, California, United States
3 Stanford University, Stanford, California, United States
4 Stanford University, Stanford, California, United States
5 Stanford University, Stanford, California, United States
6 Stanford University, Stanford, California, United States
7 Stanford University, Stanford, California, United States
8 Dana-Farber Cancer Institute, Harvard Medical School, Boston, United States, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, United States
9 National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, United States
10 University of Ulm, Ulm, Germany
11 University of Ulm, Ulm, Germany
12 Univeristy of Cologne, Cologne, Germany
13 Univeristy of Cologne, Cologne, Germany
14 Univeristy of Cologne, Cologne, Germany
15 Stanford University, Stanford, United States
16 Univeristy of Cologne, Cologne, Germany
17 Stanford University, Stanford, California, United States
ABSTRACT
Background: Richter transformation (RT), the transformation of chronic lymphocytic leukemia (CLL) into an aggressive lymphoma, has poor outcomes with conventional chemotherapy. Bruton tyrosine kinase inhibitors (BTKi) show promising response rates (Jain et al. 2023, Al-Sawaf et al. 2024), however, in contrast to CLL, resistance mechanisms and dynamics remain poorly characterized. We performed longitudinal circulating tumor DNA (ctDNA) profiling for detection and real-time monitoring of BTKi resistance.
Methods: We analyzed 103 longitudinal plasma samples from 50 patients with RT enrolled in cohort 1 of the phase 2 RT1 trial (NCT04271956) of zanubrutinib (BTKi) plus tislelizumab (PD-1 inhibitor). Baseline ctDNA was sequenced by CAPPseq (Newman et al. 2016) for 43 patients with paired peripheral blood mononuclear cell (PBMC) collection; follow-up plasma was available for 35 patients. Growth coefficients (s, year-1) of BTK/PLCG2-mutant clones were estimated from serial variant allele frequencies using a logistic-growth approximation with depth-weighted 95% CIs derived by the delta method. Gene expression was inferred from ctDNA fragmentation patterns using EPICseq (Esfahani et al. 2022).
Results: ctDNA was detectable in all 43 patients at baseline, with mutations in TP53 (49%), SF3B1 (37%), IGLL5 (37%), and NOTCH1 (35%). When considering 13 genes covered by both assays, ctDNA captured 93% of PBMC mutations, yet most ctDNA variants were not detected in paired PBMC (83/152; 55%). Baseline BTK and/or PLCG2 mutations were present in ctDNA from 9/43 patients (21%), 7 of whom had prior BTKi exposure. Only 3/15 BTK/PLCG2 variants were detected in matched PBMCs, suggesting that ctDNA captures resistant subclones underrepresented or undetectable in circulating CLL cells.
During treatment, BTK/PLCG2 mutations were detected in 17/35 patients (49%). Among patients without baseline BTK mutation, median time to BTK mutation acquisition was 8.5 months in BTKi-naive patients (n=20, 5 events) and not reached in BTKi-pretreated patients (n=15, 3 events). This is markedly shorter than the 30-70 months reported in CLL cohorts (Brown et al. 2024, Woyach et al. 2023, Evans et al. 2025). Among patients with BTK mutations, C481S/R occurred in 72%, L528W in 39%, and T474I in 22%. 50% harbored multiple variants and 17% had a co-occurring PLCG2 mutation. Clonal growth rate estimation showed consistent positive selection of BTK mutant clones: 29/32 mutation intervals increased significantly (median s=9.5 year-1). For PLCG2 mutant clones, 6/12 mutation intervals increased (median s=3.7 year-1).
Detection of a BTK mutation was strongly associated with progression: 16/18 (89%) BTK-mutated patients developed progressive disease (PD), versus 11/25 (44%) of patients without detectable BTK mutation (Fisher's exact p=0.004). In a time-dependent Cox model, detection of a BTK mutation conferred a ~4x higher risk of progression (HR 4.17, 95% CI 1.98-8.80, p=0.0002). Median time from first BTK mutation detection to PD was only 2.9 months (95% CI 2.0-NR, n=13,11 events), shorter than the 7-11 months reported in CLL.
Inferred gene expression analysis revealed significant repression of ibrutinib-response signatures during RT1 therapy (n=13 pairs; Landau2017: NES −2.07, Rendeiro2020: NES −1.84; both FDR<0.03). This effect was limited to responders, and evident in paired and unpaired analyses. Between-group analyses showed baseline enrichment of E2F-target, G2M-checkpoint and MYC-target programs in nonresponders (n=14) versus responders (n=17; NES 1.60–2.42; all FDR<0.05). At follow-up, nonresponders (n=6) had higher expression of an ibrutinib-resistance signature than responders (n=21; NES 1.5, FDR<0.05) and increased TCF4 expression (log2FC 5.6, FDR<0.01), both previously implicated in epigenetic BTKi resistance in large B-cell lymphoma models (Shaffer et al. 2017).
Conclusions: This longitudinal ctDNA analysis of BTKi resistance in RT reveals highly accelerated mutation dynamics compared to the CLL context. BTK/PLCG2 mutations were detected in half of the patients on treatment and were strongly associated with progression, with a median interval of 2.9 months from first detection to PD. ctDNA captured resistant subclones frequently missed by PBMC genotyping and resolved resistant-clone growth in real time, while EPICseq identified complementary transcriptional programs. These findings support prospective evaluation of ctDNA-guided earlier treatment adaptation in RT.
DISCLOSURES
Lena Osswald, MD: No relevant conflicts of interest to declare.