Ring Chromosomes in Patients with Myeloid Neoplasms Are Associated with a Poor Response to Therapy Open Access

Ring chromosomes (RCs) are circular chromosomal structures associated with loss or gain of genetic material, which are present in <10% of patients with myeloid neoplasms (MNs) [1, 2]. Specific mechanisms underlying RC formation are elusive; however, several possibilities include two breaks in both chromosome arms, one broken end of a chromosome fusing with the opposite telomere region, or fusion of sub-telomeric sequences and telomere-telomere fusion without deletion [3, 4]. RCs can also result from genetic damage due to exposure to mutagens like radiation and radiotherapy. Structurally, RCs are divided into two groups: (1) complete rings without loss of genetic material by telomere-to-telomere fusion and (2) incomplete rings with distal or interstitial deletion and/or duplications by one or multiple fusion events [4].

Assessment of the prognosis of patients with MN, including individuals with acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and myeloproliferative neoplasms (MPN), usually involves the implementation of a variety of disease-specific prognostic scoring systems. Therapeutic options for patients with MN are largely determined depending on the assessment of their disease severity using such prognostic tools. In AML, the European Leukemia Network 2022 guidelines distinguish favorable, intermediate, and adverse risk disease based on characteristic cytogenetic and molecular aberrations [5]. Patients with adverse risk AML typically have poor outcomes with conventional induction chemotherapy [6, 7]. Recent studies have shown that despite the therapeutic advances in the treatment of MN with the use of hypomethylating agents (HMA) and venetoclax (VEN), this approach has limited efficacy in specific high-risk groups, including those with multi-hit loss of TP53 and/or a complex karyotype [8‒11].

Due to the rarity of MN patients with RC and lack of understanding of its prognostic implications, this abnormality has not been specifically incorporated in current risk stratification tools that include karyotypic abnormalities [12‒14]. RCs have been previously associated with a poor prognosis and correlate with other adverse cytogenomic markers in separate multi-institutional cohorts of AML and MDS patients [15, 16]. Furthermore, the implications of the presence of RC on response rates to therapy in patients with MN have not been previously explored. Therefore, we sought to compare clinical outcomes and pathologic correlates in a cohort of AML, MDS, and MPN patients found to harbor RC at our institution.

We conducted retrospective outcomes and descriptive study of consecutive patients found to have RC during chromosomal analysis with a diagnosis of AML, MDS, or MPN at the Mount Sinai Health System between October 2017 and May 2024. All study procedures were conducted under an approved protocol by the Institutional Review Board of the Program for the Protection of Human Subjects at the Icahn School of Medicine at Mount Sinai. Clinical-pathologic data including demographic, treatment, molecular, cytogenomic, and patient responses to a variety of treatment options were collected by manual chart review of the electronic medical record. Chromosomal identification of the RC was achieved by conventional bone marrow cytogenetics and confirmed using metaphase, interphase, or multi-color fluorescence in situ hybridization (FISH). In those specimens where there was additional DNA and the ring origin could not be identified with FISH, we used array comparative genomic hybridization plus single nucleotide polymorphism (aCGH+SNP) to elucidate the origin of ring chromosomes [17]. Myeloid gene mutational status for individual patients was collected from next-generation myeloid malignancy gene sequencing panels provided by a commercial vendor (NeoGenomics Laboratories, Fort Myers, FL). The Kaplan-Meier method was used to estimate median overall survival (mOS) and progression-free survival (PFS). Records with incomplete data were censored at the time of the last follow-up. Statistical analysis was performed in SPSS (IBM) and figures were generated using GraphPad (Dotmatics). Study data were collected and managed using REDCap electronic data capture tools hosted at the Icahn School of Medicine at Mount Sinai [18, 19].

In total, 1980 records of patients with AML (n = 570), MDS (n = 565), or MPN (n = 845) were screened for RC. We identified 39 patients with RC (1.97% of total) with accompanying clinical data available for 31 patients. Demographic and treatment data are presented in Table 1. Nineteen patients were males and twelve were females. The median age was 73.2 years (69.3–77.4). The MN included AML (n = 15), MDS (n = 10), and MPN (n = 6), of which 4 were in blast phase. Of AML and MDS cases, 4 each were considered therapy related to all receiving prior chemotherapy except for 1 case of MDS who received prior radiation. The most frequently observed rings were r(7) and r(11), in 7 and 6 cases respectively, followed by r(8) and r(21) in 3 cases each, and r(1q), r(5), r(6), r(17), and r(19) in 2 cases each. Multiple RCs were detected in 6/31 (19%) cases. Representative RCs are depicted in Figure 1.

Myeloid malignancy gene mutational profiling was available for all 31 cases (Fig. 2). Mutations in TP53 were the most frequently identified mutation in the presence of RC (52%). The next most frequent mutations involved DNMT3A (n = 8) and TET2 (n = 7).

Full cytogenomic data are provided in Table 2. A complex karyotype, defined by the presence of greater than or equal to 3 cytogenetic abnormalities [5], was identified in 25 patients (81%), including 12/15 (80%) of patients with AML, 8/10 (80%) with MDS, and 5/6 (83%) with MPN. Loss of 17p was detected in 13 (42%) cases and accompanied by a TP53 mutation in 10 (32%) patients which led to multi-hit loss of TP53. AML was diagnosed in 7/10 cases with multi-hit loss of TP53, with the remaining 3 patients having MDS. The percentage of metaphases which included RC ranged between 20 and 100%, with values for each case reported in Table 2. Notably, in 22 (71%) patients, the detection of RC was not preceded by earlier cytotoxic or radiation therapy, suggesting de novo formation of the RC (AML = 11, MDS = 6, MPN = 5).

Large-scale chromosomal rearrangements such as chromothripsis (massive chromosomal breakage and random rearrangement) were seen in 3 cases and chromoanasynthesis (gene copy alterations due to replication-based errors) was detected in 4 cases [20]. Conventional cytogenetics identified 9 cases with homogenously staining regions (HSRs), indicative of genomic amplification, involving chromosomes 17 (n = 3), 19p (n = 2), and 21 (n = 3) [21].

Ring (11) was seen at the time of MN diagnosis in 5 of 6 cases and was part of a complex karyotype including chromothripsis, amplification, and HSRs involving histone-lysine-specific N-methyltransferase 2A (KMT2A). In contrast, in 5 out of 6 cases with r(7), the RC was the sole cytogenomic abnormality.

Chromosomes 5, 7, or 21 were involved in 23 cases (74%), including 17/22 (85%), where the RC was identified initially at the time of diagnosis. These abnormalities include −5/del(5q), −7(del(7q), and HSR of chromosome 21.

We assessed the impact of the therapy received after detection of RC on clinical outcomes. The most common first-line treatment regimen was HMA/VEN (13), followed by supportive care alone (6), HMA alone (6), and induction chemotherapy (3). Only 4 patients underwent allogeneic hematopoietic stem cell transplantation (allo-SCT), which followed therapy with HMA/VEN (n = 3) and induction chemotherapy (n = 1).

A Kaplan-Meier analysis of mOS of patients with MN and RC revealed that patients who received supportive care had significantly shorter mOS from the time of RC detection than patients who received HMA, HMA/VEN, or induction chemotherapy (log-rank test, p = 0.001) (Fig. 3a). We did not observe a significant difference in mOS between patients who received HMA alone versus HMA/VEN versus induction chemotherapy (11.4 vs. 8.1 vs. 7.3 months, p = 0.86) (Fig. 3a). Likewise, PFS was significantly prolonged after receipt of any therapeutic option as compared with supportive care alone (log-rank test, p = 0.001), but a significant difference in PFS was not observed with any of the treatment modalities employed (Fig. 3b). An independent effect on mOS for TP53 allele status (Fig. 3c), the presence of a complex karyotype (Fig. 3d), history of a therapy-related neoplasm, or for pair-wise interactions between those categories and individual therapies was not observed (p > 0.05 for all tested associations).

Long-term follow-up was available for 1 patient with r(7) treated with HMA/VEN followed by allo-SCT, who was alive at last follow-up at 8.9 months at which time the marrow revealed a normal karyotype without RC. An additional patient with r(17) in the context of newly diagnosed AML with complex karyotype and bi-allelic TP53 loss was treated with HMA/VEN and had a complete response after 2 cycles with reversion to a normal karyotype and mutational profile; however, longer-term follow-up data is unavailable.

RCs are rare cytogenomic findings and their utility in predicting responses of MN patients to therapy and prognosticating survival outcomes is unknown. Small case studies and multi-institutional retrospective studies have indicated the negative prognostic impact of RC in AML and MDS [15, 22]. Recently, RCs have been shown to be associated with high rates of TP53 mutations [16]. Our study further supports these findings and is the first to investigate the impact of therapy in this patient population.

We identified RC in 31 consecutive and analyzable cases of MN from our institution, representing <2% of all cases, with AML being the most common diagnosis (48.3%). The most frequent myeloid gene mutations observed in our cohort involved TP53 in 58% of cases, consistent with earlier studies [16]. We likewise observed a high prevalence of chromosomal loss of 17p, suggesting that genome-wide instability might be associated with RC formation, as postulated by others [15, 16]. In total, 20/31 cases (64.5%) RC was associated with either a TP53 mutation, chromosomal loss of 17p, or both. In contrast, the rate of TP53 mutations in MN irrespective of RC has been reported to be only 5–10% [23‒25]. While we lacked adequate longitudinal data to establish temporal relationships regarding the sequential acquisition of TP53 abnormalities and RC formation, in three cases where prior samples were available, one had a TP53 mutation that preceded RC detection, and all had intact 17p. Despite these findings, 71% of patients did not have prior exposure to cytotoxic or radiation therapy, and 32% of all patients lacked TP53 mutations or chromosomal p53 loss, indicating a high rate of de novo RC formation without a prior genome-disrupting event. Especially notable was the high proportion of RC involving chromosomes 5, 7, or 21 at diagnosis. These abnormalities, in addition to those involving TP53, are frequently associated with cytotoxic therapy-related disease signatures and could suggest a similar pathogenesis for RC independent of geno-toxic therapy [5].

Comparing responses to therapy and clinical outcomes, we observed that the mOS in patients with RC receiving supportive care alone was significantly inferior compared with any of the therapeutic options pursued. This is not surprising given the aggressive nature of these diseases but should be taken in the context that our retrospective study was not designed to control for individual patient factors. Age was another potential confounder influencing outcomes, as the median age of the total cohort was 73.2 years. However, the differences in age between treatment groups were small and overall representative of the older ages at which diagnoses of MN are typically made and in which RCs have been previously reported [5, 15]. Of relevance to the older MN patient population where HMA/VEN is favored over intensive chemotherapy, the addition of VEN to HMA did not positively influence mOS or PFS compared to HMA alone or chemotherapy. In younger patients, only 3/9 aged ≤65 years received allo-SCT owing to the high early mortality seen in this group. Notably, we observed eradication of the RC in 2 patients after only two cycles of HMA/VEN each, which permitted 1 patient to proceed to allo-SCT with ongoing disease remission after nearly 9 months of follow-up.

Despite our data being biased by cases where therapy was not given due to poor performance status or where goals of care were not consistent with aggressive management, these findings are consistent with previously published reports regarding the dampened efficacy of HMA/VEN in patients with TP53 mutations and/or complex karyotypes [26]. Additionally, there was no relationship between known markers of adverse-risk disease and type of therapy on mOS, and the benefit of adding VEN to HMA-based therapy in these high-risk individuals was not observed. Even with the small numbers of patients studied in each cohort, our findings indicate that RC may be associated with poor response rates to HMA/VEN therapy.

Other adverse risk markers including complex karyotype, TP53 allele status, or therapy-related disease, did not have any independent impact on mOS in patients with RC. This suggests that RC formation may be an important independent adverse prognostic biomarker; however, our study was not designed to evaluate these outcomes with RC as an independent variable. We were also limited by inadequate statistical power to perform prognostic analyses for individual types of RC and different therapies. Nevertheless, it is notable that r(11), frequently identified at the time of diagnosis, was commonly observed with other high-risk cytogenetic markers.

Our cohort is among the largest and highly clinically annotated collection of cases in this unique and rare population. Despite this, the major limitations of our study are the small numbers included in our analyses, which limits the statistical power of our testing, and the lack of non-RC control cases. This is due in large part to the rarity of RC and additionally to our single-institutional retrospective design. Nevertheless, our cohort is unique in having comprehensive clinical, treatment, and molecular data available for a rare patient population. Future investigation should involve multi-institutional collaborative efforts to pool larger cohorts of patients with these rare cytogenetic abnormalities to better understand their response to various therapies. Additionally, as our data show inferior outcomes of patients with RC and MN in response to conventional therapies, functional studies could be considered to elucidate how the RCs impact various cell-pathways and resistance mechanisms.

The authors acknowledge the contribution of the administrative staff and laboratory technologists of the Tumor Cytogenomics Laboratory for their service to our patients and their generation of the primary data used in this manuscript.

This study protocol was reviewed and approved by the Institutional Review Board of the Program for the Protection of Human Subjects at the Icahn School of Medicine at Mount Sinai (STUDY-22-00852) with a waiver of consent requirement.

D.I.N., R.H., D.A., L.S., and V.N. have no disclosures to report. J.M. reports research funding paid to his institution from Incyte, Novartis, AbbVie, BMS, Sobi, Geron, Karyopharm, Kartos, Disc, and PharmaEssentia and consulting fees from AbbVie, Sobi, Incyte, Novartis, BMS, GSK, Kartos, Keros, Karyopharm, PharmaEssentia, Geron, Merck, Pfizer, MorphoSys, and Sumitomo. B.K.M. reports consulting fees from Cellarity. D.T. reports contracted research funding paid to his institution from Sobi, Sumitomo, Cogent Biosciences, and Gilead and consulting fees from Sobi, AbbVie, PharmaEssentia, Sierra Oncology, GSK, and Cogent Biosciences. J.F. reports research funding from Orzyon Genomics, Syros Pharmaceuticals, and Taiho Oncology. All authors report no conflicts of interest with regard to this manuscript.

Research reported in this manuscript was supported by the National Cancer Institute of the National Institutes of Health under award number T32 CA225617 and The Tisch Cancer Institute Paul Calabresi K12 Career Development Award for Clinical Oncology under award number K12 CA270375 to D.I.N. This work was supported in part through the computational and data resources and staff expertise provided by Scientific Computing and Data at the Icahn School of Medicine at Mount Sinai and supported by the Clinical and Translational Science Awards (CTSA) grant UL1TR004419 from the National Center for Advancing Translational Sciences.

D.I.N.: writing – original draft, conceptualization, investigation, data curation, and formal analysis. R.H. and J.M.: writing – review and editing and supervision. D.A.: investigation. B.K.M., D.T., and L.R.S.: writing – review and editing. V.N.: writing – review and editing, conceptualization, investigation, data curation, and supervision. J.F.: writing – review and editing, conceptualization, investigation, and supervision.

Vesna Najfeld and Jonathan Feld contributed equally to this work.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from D.I.N. upon reasonable request.