AUTHORS
Joseph Peter Novak, MS, Saurabh Zanwar, MBBS, MD2, Surendra Dasari, Associate Professor3, Vaishali Bhardwaj, PhD4, Dongni Yi, MD5, Jonas Paludo, MD6, Yucai Wang, MD, PhD7, Moritz Binder, MD, MPH8, Shaji Kumar, MD9, Anne J. Novak, PhD10, Prashant Kapoor, MD11, Stephen Maxted Ansell, MD, PhD12, Patrizia Mondello, MD, PhD13, Jithma Abeykoon, MD14
AFFILIATIONS
1 Mayo Clinic, Rochester, Minnesota, United States
2 Mayo Clinic, Rochester, Minnesota, United States
3 Mayo Clinic, Rochester, Minnesota, United States
4 Mayo Clinic, Rochester, Minnesota, United States
5 Mayo Clinic, Rochester, Minnesota, United States
6 Mayo Clinic, Rochester, Minnesota, United States
7 Mayo Clinic, Rochester, Minnesota, United States
8 Mayo Clinic, Rochester, Minnesota, United States
9 Mayo Clinic, Rochester, Minnesota, United States
10 Mayo Clinic, Rochester, Minnesota, United States
11 Mayo Clinic, Rochester, Minnesota, United States
12 Mayo Clinic, Rochester, Minnesota, United States
13 Mayo Clinic, Rochester, Minnesota, United States
14 Mayo Clinic, Rochester, Minnesota, United States
ABSTRACT
Introduction: Waldenström macroglobulinemia (WM) is a B-cell lymphoplasmacytic malignancy in which recurrent alterations involving MYD88, CXCR4, del(6q), and additional less frequent events influence disease biology and clinical outcome. We previously described three transcriptional states in WM (Mondello et al. Clin Cancer Res 2023). This taxonomy remains to be independently validated and the mechanisms by which mutations and transcriptional programs work in tandem to promote disease lymphomagenesis remain incompletely understood. To address these gaps, we extended our previous work and applied the molecular clustering model to a new cohort of bulk-RNAseq profiles with matched whole-exome sequencing (WES).
Methods: RNAseq and WES profiles were obtained from Mayo Clinic patients diagnosed with either smoldering or treatment-naïve (tnWM) WM adhering to all IRB guidelines. Four previously unclassified tumor profiles from the initial study were included. Transcriptional state scores were measured per-sample using the previously defined molecular signatures. State scores were transformed into confidence values using a softmax function and samples were classified using the highest value. Genomic correlates to transcriptional state were evaluated using Fisher’s exact test. Chi square was used for analysis of contingency tables. Follow up was measured as PFS and OS at 10 years and compared using the log-rank test. Statistical analysis was performed using R.
Results: Projection of the transcriptional states onto a new set of WM tumor tissues (n=36) recovered all three previously described molecular states with C1 representing the predominant group (n=27, 75%), followed by C3 (n=6, 16.7%) and a smaller C2 population (n=3, 8.3%), all of which was not explainable by diagnosis (p=0.46). Median confidence scores were 0.72, 0.66, and 0.55 respectively (median=0.67, IQR: 0.54 - 0.86; range 0.35-1.00). In a small subset of tumors with matched tissue to the initial study, we recaptured the original cluster assignment (C1 n=4/4, 100%) averaging a confidence of 0.82. Genomic integration revealed that C1 demonstrated higher frequencies of del(6q) (n=7/27, 25.9%) and CXCR4 mutations (n=7/27, 25.9%) relative to C3 (n=1/6, 16.7%; n=1/6, 16.7%). Although no feature reached statistical significance, several were suggestive. C1 harbored mutations in genes which have been shown to be enriched in relapse disease: ARID1A (5/27, 18.5%), MAP3K14 (5/27, 18.5%), TP53 (3/27, 11%) all otherwise devoid from C3. Survival analyses among tnWM showed that the molecular clusters separated patients by clinical outcome, with C3 associated with inferior OS (p<0.001), and trended toward inferior PFS.
Conclusion: Application of clusters onto a new set of untreated samples reveal a composition reminiscent of the original findings while providing additional insight into state-specific pathogenesis. C1, was marked by genomic complexity. C3, conferred a worse OS. C2, the IgM MGUS cluster, was sparse in this study. Together, these results suggest a potential avenue for molecular identification in WM as the molecular characteristics of transcriptional states are retained in a subsequent cohort.
DISCLOSURES
Joseph Peter Novak, MS: No relevant conflicts of interest to declare.