Dynamic cytogenetic evolution in multiple myeloma: prognostic implications from diagnosis to first relapse

Abstract

Background

Cytogenetic abnormalities (CAs) in multiple myeloma (MM) are traditionally assessed statically at diagnosis, but longitudinal monitoring reveals dynamic clonal evolution, particularly during relapse. In Asian populations, data on CA dynamics and their prognostic impact remain scarce.

Patients and Methods

This retrospective study included 106 MM patients (2014-2024) in whom CAs were evaluated at diagnosis and relapse via cIg-FISH. The prognostic impact of clonal burden, copy number alterations, and longitudinal evolution of cytogenetic complexity was comprehensively assessed.

Results

At first relapse, 33.0% of patients acquired new CAs, with 1q21+ and del(17p13) incidences significantly increasing (P = .032, P = .003). Longitudinal analysis showed 1q21 amplification (≥4 copies) expanded (27.4%-40.6%), while gain (3 copies) remained stable. Baseline 1q21 amplification predicted shorter 1st PFS (P = .002), whereas 1q21 gain at relapse correlated with inferior 2nd PFS (P = .023), often co-occurring with del(17p13) (28.1%). Persistent 1q21+ and de novo del(17p13) both predicted inferior OS (P < .05). Based on clonal dynamics, patients with >20% expansion exhibited significantly shorter 1st PFS (P = .045) compared to the stable (unchanged) CAs subgroup. Similarly, those acquiring new CAs had shorter 2nd PFS (P = .039) and a trend toward worse OS (P = .056).

Conclusion

Both acquisition of new CAs and expansion of existing clones at relapse are key prognostic indicators, highlighting the importance of dynamic CA monitoring in MM management.

Implications for Practice.

This study highlights the clinical significance of cytogenetic clonal evolution in Asian multiple myeloma patients. We demonstrate that the clonal expansion of 1q21 and de novo acquisition of del(17p13) at relapse significantly impair survival. These findings underscore the limitations of baseline-only risk assessment and advocate for serial cytogenetic monitoring at disease progression. Incorporating longitudinal profiling into routine practice is essential to capture the evolving risk landscape, thereby refining prognostication and guiding risk-adapted therapeutic strategies for relapsed disease.

Introduction

Multiple myeloma (MM) is a malignant disorder of terminally differentiated plasma cells characterized by clonal proliferation within the bone marrow. It is the second most common hematologic malignancy, accounting for approximately 1% of all cancers and 10% of hematologic cancers worldwide.^1–3 The past two decades have witnessed a paradigm shift in the management of MM, with the introduction of novel therapeutic classes, including proteasome inhibitors (PIs), immunomodulatory drugs (IMiDs), monoclonal antibodies, and autologous stem cell transplantation (ASCT). These agents have dramatically improved response rates and extended median overall survival.^4–10 Despite this progress, MM remains a largely incurable disease for most patients, typically following a relapsing-remitting course in which the duration of response shortens with successive relapses, ultimately leading to refractory disease and death.^11–13

The clinical and biological behavior of MM is driven by profound genomic complexity and heterogeneity. At its core, MM is a cytogenetic disease, with nearly all patients harboring numerical or structural cytogenetic abnormalities (CAs).^14–20 Recently, the International Myeloma Society/International Myeloma Working Group (IMS/IMWG) has systematically defined a set of high-risk CAs that inform disease prognosis.²¹ A hallmark of MM biology is extensive intra-clonal heterogeneity, wherein multiple genetically distinct subclones coexist at diagnosis.²² This diversity provides the substrate for Darwinian selection under therapeutic pressure: while sensitive plasma cell clones are eradicated, resistant subclones can survive and subsequently expand, leading to relapse.^16,22–26

Clonal evolution is a central driver of heterogeneity in MM, and accumulating evidence indicates that MM advances through a multistep process of sequential genetic lesion acquisition.^27,28 Recent studies employing quantitative multigene fluorescence in situ hybridization (QM-FISH) have delineated multiple evolutionary trajectories, with clonal stability correlating with more favorable survival compared to dynamic change.^22,29 Among the CAs involved in clonal evolution, 1q21+ and del(17p13) have emerged as the most clinically impactful lesions.³⁰ The emergence of these aberrations is often traced back to minor subclones that were present at diagnosis but were too small to be detected by standard clinical-grade fluorescence in situ hybridization (FISH).³¹ Studies have shown that minor clones harboring high-risk CAs like 1q21+ or del(17p13) can act as a reservoir for relapse, expanding under therapeutic pressure to drive disease progression.^31,32

Accumulating evidence supports the concept that risk in MM is dynamic, thereby necessitating cytogenetic re-evaluation at relapse for accurate prognostication and optimal therapeutic decision-making.²¹ The 2024 National Comprehensive Cancer Network (NCCN) Guidelines classify newly acquired genetic lesions at relapse as “dynamic high-risk” markers, emphasizing the clinical importance of serial genetic profiling for individualized management and risk refinement.

However, while the principle of clonal evolution is well established, large-scale systematic studies detailing the distinct modes of cytogenetic change and their comparative prognostic impact in the era of contemporary therapy are still limited. Most prior research has centered on single high-risk markers in isolation,^33,34 yet comprehensive categorizations of cytogenetic evolution—and clarification of their survival implications—remain underexplored, particularly in the Chinese population.

To address these gaps, we conducted a longitudinal study of 106 MM patients, comparing cytogenetic profiles at diagnosis and first relapse. Our objectives were threefold: (1) to delineate the incidence and dynamic patterns of cytogenetic changes during disease progression; (2) to categorize clonal evolution modes into prognostically distinct patterns; and (3) to evaluate the impact of specific acquired CAs and evolutionary trajectories on post-relapse survival. We aim to refine risk stratification at relapse and provide real-world evidence supporting cytogenetic-informed personalized clinical management in multiple myeloma.

Methods

Patients

We retrospectively analyzed patients with newly diagnosed multiple myeloma (NDMM) treated at the First Affiliated Hospital of Nanjing Medical University between April 2014 and March 2024. Both initial diagnosis and subsequent clinical relapse were defined according to the International Myeloma Working Group (IMWG) criteria.^35,36 To ensure the robustness of longitudinal cytogenetic analysis, the study cohort was restricted to patients with available paired FISH results at both time points (diagnosis and first relapse). Key exclusion criteria included inadequate induction therapy (<4 cycles) and isolated extramedullary relapse. A comprehensive flow diagram of the patient selection process is provided in Figure S1 (see online supplementary material for a color version of this figure). The study was conducted in accordance with the Declaration of Helsinki and was approved by the Medical Research Ethics Review Committee of the Jiangsu Province Hospital (No. 2022-SR-448).

Cytoplasmic light chain immunofluorescence with fluorescence in situ hybridization (cIg-FISH) and criteria

CIg-FISH³⁷ was performed on fresh bone-marrow aspirates obtained at diagnosis and first relapse. Mononuclear cells were processed for fluorescence in situ hybridization (FISH) using VYSIS dual-color probes targeting 1q21.2/1p32.3 (CKS1B/CDKN2C), 17p13.1/17p11.1-q11.1 (TP53/CEP17), 14q32.3 (IGH), t(4; 14)(p16.3; q32.3) (IGH/FGFR3), t(14; 16)(q32.3; q23) (IGH/MAF), t(11; 14)(q13.3; q32.3) (IGH/CCND1), and t(14; 20)(q32.3; q12) (IGH/MAFB). Following hybridization, cytoplasmic immunoglobulin (cIg) staining was performed to identify malignant plasma cells. In alignment with the post-2014 IMWG consensus,³⁸ del(13q) was excluded from the routine panel as it is no longer classified as a high-risk abnormality. A minimum of 100 marked plasma cells was scored; positivity cut-offs adhered to the European Myeloma Network criteria: ≥20% for 1q21+ (a cytogenetic abnormality defined as the gain of three copies or amplification of ≥4 copies of the 1q21), del(1p32), and del(17p13); and ≥10% for IGH translocations.³⁹

Criteria for determining dynamic changes of CAs during disease progression

Dynamic cytogenetic patterns were classified as follows:

  ① Loss: CAs present at diagnosis are absent at relapse.

  ② Stable (unchanged): retention of identical CAs with a ≤20% variation in clonal size.²⁹

  ③ Stable (expansion): presence of identical CAs with a >20% increase in clonal size.

  ④ Gain: acquisition of novel CAs identified exclusively at relapse.

Therapeutic interventions

Induction was initiated with IMiD-based regimens (56.6%), PI-based regimens (31.1%), or their combination (12.3%), followed by ASCT in 36.8% of cases. Maintenance therapy (n = 95) was predominantly monotherapy (86.3%). In the relapsed setting, salvage strategies (n = 58) utilized novel modalities, specifically daratumumab-based protocols (31.1%), chimeric antigen receptor (CAR) T-cell therapy (13.2%), and their combination (10.4%) (Table S1—see online supplementary material).

Follow-up

Follow-up (cut-off: 31 December 2024) was obtained from medical records and telephone contact. First progression-free survival (1st PFS) was defined as the interval from initial diagnosis to the first relapse. Second PFS (2nd PFS) was calculated from the date of first relapse to the occurrence of secondary relapse, death from any cause, or last follow-up. Overall survival (OS) was defined as the period from diagnosis to death or last follow-up.

Statistical methods

Statistical analyses were performed with SPSS 27.0 and GraphPad Prism 10.6.0. Categorical variables are presented as numbers (percentages) and compared by χ² or Fisher’s exact test. Continuous variables are expressed as medians (range) and compared using Student’s t-test or the Mann-Whitney U test. Survival curves were generated by the Kaplan–Meier method and compared with the log-rank test. A two-sided P < .05 was considered significant.

Results

Baseline data

One hundred and six patients with paired diagnostic and first-relapse cytogenetic data were analyzed. The cohort comprised 57 men (53.8%) and 49 women (46.2%), with a median age of 61 years (34-78). At first relapse, no significant shifts were observed in Durie-Salmon (DS), International Staging System (ISS), or Revised International Staging System (R-ISS) stages relative to baseline (P = .209, .210, and .876, respectively). Among all cytogenetic anomalies, only 1q21+ and del(17p13) exhibited a statistically significant increase between the two time points (P = .032 and .003, respectively) (Table 1). After a median follow-up of 51 months (7-137), median 1st PFS was 26 months (4-95), median 2nd PFS was 45 months (1-78), and median OS was 95 months (7-137). At the time of data cut off, 35.8% (38/106) of the patients had died, among whom only 2 had experienced a second relapse.

Table 1.

Baseline demographics and clinical characteristics stratified by disease status.

Characteristic, n (%)	Newly diagnosed (n = 106)	First relapsed (n = 106)	P-Value
Age, median (range), years	61(34-78)	63(36-80)	–
Sex			–
Male	57(53.8)
Female	49(46.2)
M-protein isotype			–
IgG	44(41.5)
IgA	26(24.5)
IgD	3(2.8)
IgM	1(0.9)
Light chain only	31(29.3)
Non-secretory	1(0.9)
Durie-Salmon stage			0.209
I/II	10(9.4)	16(15.1)
III	96(90.6)	90(84.9)
ISS stage			0.210
I/II	57(53.8)	66(62.3)
III	49(46.2)	40(37.7)
R-ISS stage			0.876
I/II	79(74.5)	78(73.6)
III	27(25.5)	28(26.4)
Cytogenetic abnormalities
1q21+	60(56.6)	75(70.8)	0.032
Del(1p32)	6(5.7)	7(6.6)	0.775
Del(17p13)	12(11.3)	29(27.4)	0.003
t(4; 14)	24(22.6)	28(26.4)	0.523
t(14; 16)	2(1.9)	4(3.8)	0.408
t(11; 14)	8(7.5)	10(9.4)	0.622
t(14; 20)	1(0.9)	1(0.9)	–

Temporal evolution of CAs between diagnosis and first relapse

CAs were identified in 90 of 106 patients (84.9%) at diagnosis and 98 (92.5%) at first relapse.

CAs at diagnosis: 46 patients (43.4%) harbored a single CA—1q21+ in 45.7%; 39 patients (36.8%) harbored two CAs—1q21+ plus IGH rearrangement in 82.1%; 5 patients (4.7%) harbored ≥3 CAs (all were 3 CAs) (Figure 1A).

Figure 1.

Comparative cytogenetic landscape of multiple myeloma at diagnosis and first relapse. A: Distribution of cytogenetic abnormalities (CAs) at diagnosis. B: Distribution of cytogenetic abnormalities (CAs) at first relapse. C: Comparative frequencies of CAs between diagnosis and first relapse.

CAs at first relapse: 36 patients (34.0%) showed a single CA—1q21+ in 50%; 45 patients (42.5%) showed two CAs—1q21+ plus IGH rearrangement in 77.5%; 17 patients (16.0%) showed ≥3 CAs (including two patients with 4 CAs), a significant increase versus diagnosis (P = .007) (Figure 1B). Comparative incidences are depicted in Figure 1C.

Thirty-five patients (33.0%) acquired ≥1 new CA at relapse: 26 gained 1 CA, and 9 gained ≥2 CAs. Conversely, six patients exhibited CA loss: 3 patients with disappearance of 1q21+, 2 with loss of del(17p13), and 1 with loss of del(1p32) (Figure 2A). By contrast, IGH rearrangements and their translocation partners remained largely stable throughout disease evolution.

Figure 2.

Longitudinal evolution trajectories of cytogenetic abnormalities from diagnosis to first relapse in paired patients. A: Evolution of the total number of cytogenetic abnormalities (CAs) from diagnosis to first relapse. B: Evolution of 1q21 copy number status from diagnosis to first relapse. C: Evolution of 1q21+ clonal burden from diagnosis to first relapse. D: Evolution of del(17p13) clonal burden from diagnosis to first relapse.

Prognostic impact of 1q21+ kinetics

Longitudinal analysis revealed that 1q21+ status was dynamically unstable, with its overall prevalence increasing from 60 (56.6%) at diagnosis to 75 (70.8%) at relapse (P = .032). Based on temporal changes, patients were stratified into four evolutionary groups: persistently negative, negative-to-positive, positive-to-negative, and persistently positive groups. Regarding clonal burden, a high load (≥50%) remained stable in the persistently positive group (51 vs. 53, 85.0% vs. 88.3%), whereas only 44.4% (8/18) of converters reached this threshold at relapse (Figure 2C). Survival differed across the four groups: median 1st PFS 35.5, 28, 30, and 24 months (P = .053); median 2nd PFS not reached (NR), 45, NR, and 24 months (P = .077); median OS NR, 95, NR, and 61 months (P = .015). Compared with sustained 1q21-negativity, persistent 1q21+ predicted inferior 1st PFS (P = .008), 2nd PFS (P = .038), and OS (P = .007) (Figure 3D-F).

Figure 3.

Longitudinal assessment of 1q21 copy number status and its evolving prognostic significance from diagnosis to first relapse. A: Schematic timeline of clinical milestones and longitudinal cytogenetic sampling. B: Impact of 1q21 copy number status at initial diagnosis on first progression-free survival (1st PFS). C: Impact of 1q21 copy number re-evaluation at first relapse on second progression-free survival (2nd PFS). D, E, F: Impact of evolving 1q21 patterns on 1st PFS, 2nd PFS, and overall survival (OS). *P < .05, **P < .01, ***P < .001.

Patients were further categorized into negative (2 copies), gain (3 copies), and amplification (≥4 copies) cohorts at both diagnosis and relapse. At diagnosis, 29.2% (31/106) exhibited gain and 27.4% (29/106) amplification. At relapse, the proportion of gain remained similar at 30.2% (32/106), whereas amplification increased to 40.6% (43/106) (Figure 2B). Notably, concurrent del(17p) was observed in 9 patients (28.1%) with 1q21 gain and in 9 (20.9%) with 1q21 amplification at relapse. In terms of survival, stratification by baseline copy number yielded median 1st PFS of 29, 28, and 19 months for the negative, gain, and amplification groups, respectively (P = .006); post-hoc analysis confirmed significantly shorter 1st PFS in the amplification group compared with the negative group (P = .002). When re-stratified by status at relapse, median 2nd PFS was NR, 28, and 42 months (P = .067), with the gain group exhibiting significantly inferior outcomes compared with the negative group (P = .023) (Figure 3B, C).

Prognostic impact of del(17p13) kinetics

Del(17p13) was dynamically unstable, rising from 12 patients (11.3%) at diagnosis to 29 (27.4%) at relapse (P = .003). Patients were stratified into persistently negative, negative-to-positive, positive-to-negative, and persistently positive groups. The persistently positive group showed ≥50% clonal burden in 10/10 (100.0%) at diagnosis and 6/10 (60.0%) at relapse, with ≥80% in 4/10 (40.0%) at diagnosis and 3/10 (30.0%) at relapse. Among 19 converters, 9 (47.4%) achieved ≥50% and 2 (10.5%) ≥80% at relapse (Figure 2D). Survival analyses revealed no difference in 1st PFS across the four groups (27 months vs. 25 months vs. 15.5 months vs. 32 months; P = .123), whereas the negative-to-positive subgroup exhibited markedly inferior 2nd PFS (NR vs. 14 months vs. NR vs. NR; P < .001) and OS (NR vs. 55 months vs. NR vs. NR; P = .006) (Figure 4A-C). Compared to patients who remained negative, acquisition of del(17p13) was significantly associated with adverse 2nd PFS and OS (both P < .001).

Figure 4.

Impact of evolving del(17p13) patterns on first progression-free survival (1st PFS), second progression-free survival (2nd PFS), and overall survival (OS). *P < .05, **P < .01, ***P < .001.

Prognostic implications of CAs’ evolutionary patterns

CA dynamics stratified patients into three groups: loss, stable, and gain. Corresponding median 1st PFS was 20.5, 28, and 25 months (P = .256); median 2nd PFS was NR, NR, and 31 months (P = .047); and median OS was NR, NR, and 74 months (P = .099), respectively (Figure 5A-C).

Figure 5.

Survival analysis of patients with multiple myeloma stratified by longitudinal evaluation of cytogenetic evolution and refined clonal dynamics. A, B, C: Classification based on directional changes of cytogenetic abnormalities (loss, stable, and gain) on 1st PFS, 2nd PFS, and OS; D, E, F: Refined stratification incorporating clonal proportion dynamics of stable cytogenetic abnormalities on 1st PFS, 2nd PFS, and OS. *P < .05, **P < .01, ***P < .001.

Patients were further divided into unchanged and expansion subgroups within the stable category. The expansion subgroup exhibited shorter 1st PFS (23 months vs. 29 months; P = .045), whereas the CA-gain group had inferior 2nd PFS compared to the stable (unchanged) subgroup (31 months vs. NR; P = .039) and a trend toward shorter OS (74 months vs. NR, P = .056) (Figure 5D-F).

Discussion

MM is a highly heterogeneous malignancy of clonal plasma cells, distinguished by pronounced inter- and intra-patient genomic heterogeneity that profoundly impacts clinical outcomes.^18,19 Substantial evidence indicates that clonal evolution of CAs is a critical determinant of prognosis.¹⁵ Such heterogeneity is apparent even in the earliest stages of the disease and tends to intensify as the illness progresses.⁴⁰ Furthermore, therapeutic interventions can selectively eradicate vulnerable subclones or alter clonal dominance,⁴¹ thereby fostering disease progression, relapse, and acquisition of drug resistance. Consequently, a comprehensive characterization of CAs’ dynamics and clonal evolutionary trajectories in MM is essential for refining risk stratification and therapeutic strategies.

The frequency of detectable CAs increased from 84.9% at diagnosis to 92.5% at first relapse. Notably, the proportion of patients with ≥3 concurrent CAs was significantly higher at relapse (P = .007), and two individuals harbored four distinct abnormalities, indicating a substantial escalation in genomic complexity after disease recurrence. Among the 106 patients analyzed, 33.0% (n = 35) acquired at least one new CA at relapse, and 25.7% (n = 9) of these accumulated two or more, reinforcing the concept that relapse constitutes a genetically distinct phase of MM. This evolutionary transition is likely driven by prior therapeutic exposure, which can eradicate drug-sensitive subclones and thereby unmask or facilitate the expansion of minor resistant clones,⁴¹ ultimately contributing to disease progression, subsequent relapses, and the development of drug resistance. Importantly, asymmetry in CA stability was observed. Primary immunoglobulin heavy chain (IGH) rearrangements remained conserved across disease stages, whereas secondary aberrations such as 1q21+ and del(17p13) demonstrated marked instability. These findings support the biological model in which IGH translocations act as initiating oncogenic events, while secondary CAs serve as principal drivers of disease progression.⁴² The 2025 National Comprehensive Cancer Network (NCCN) Guidelines (Version 1) designate newly acquired genetic lesions at relapse as “dynamic high-risk” markers, underscoring the clinical value of serial genetic profiling to enable individualized therapeutic planning and more precise prognostication.

In the current cohort, 1q21+ emerged as the predominant cytogenetic aberration, acting as a stable driver of disease progression with its frequency rising from 56.6% at baseline to 70.8% at relapse (P = .032). In light of evidence associating a high 1q+ clonal burden (≥50%) with poor survival outcomes,³¹ our finding that patients with persistent 1q21+ maintained a burden of ≥50% across time points underscores a pattern of stable clonal dominance that likely drives disease progression. Conversely, only half of those with newly acquired 1q21+ at relapse reached this threshold, indicating a relatively constrained capacity for de novo high-level acquisition. Survival analyses confirmed 1q21+ as an adverse prognostic marker. Persistent positivity was associated with markedly inferior 1st PFS, 2nd PFS, and OS compared with persistent negativity. Notably, patients converting to 1q21+ at relapse displayed a similarly unfavorable trend, suggesting that acquisition of 1q21+—whether present at diagnosis or emerging during disease course—portends poor clinical outcomes. Regarding copy number specificities, previous studies have substantiated that 1q21 amplification (≥4 copies) confers a more dismal prognosis than gain (3 copies) in newly diagnosed patients. A plausible explanation is that modern induction regimens, particularly when followed by autologous stem cell transplantation, may partially mitigate the adverse impact of 1q21 gain, whereas 1q21 amplification remains largely refractory to these interventions—a notion consistent with our 1st PFS findings. However, the selective pressure exerted by multi-agent therapy appears to drive further genomic instability and clonal evolution. Indeed, while the proportion of 1q21 gain remained stable from diagnosis to relapse (29.2% vs. 30.2%), amplification markedly increased from 27.4% to 40.6%. This selective expansion of the amplification clone, in contrast to the stable prevalence of gain, highlights the role of therapeutic pressure in driving clonal evolution toward a more aggressive phenotype. Mechanistically, 1q21+ is implicated in promoting chromosomal instability, potentially facilitating the development of additional high-risk cytogenetic lesions.^31,43–45 Consistent with this, our data revealed that patients with 1q21 gain at relapse exhibited a higher incidence of concurrent del(17p) acquisition compared with the amplification group (28.1% vs. 20.9%). This secondary enrichment of del(17p) likely contributes to the heightened susceptibility to subsequent relapse or death in the gain subgroup, suggesting that the prognostic weight of 1q21 copy number is modulated by the evolving complexity of the cytogenetic landscape.

Del(17p), which is present in up to 10% of patients at diagnosis and becomes more prevalent at relapse, is a well-established marker of poor prognosis.^21,46 In our cohort, the frequency of del(17p13) more than doubled from diagnosis to relapse, increasing from 11.3% (12 patients) to 27.4% (29 patients) (P = .003), indicative of marked genomic instability. Persistently positive patients displayed uniformly high clonal burden at diagnosis (≥50% in 100%, ≥80% in 40.0%), which subsequently declined at relapse (≥50% in 60%, ≥80% in 30%), potentially reflecting therapeutic selection pressure or shifts in interclonal dynamics. Similar to the pattern observed with 1q21+, only approximately half of the converters achieved high-level del(17p13) (≥50%) at relapse, indicating a relatively limited capacity for de novo acquisition of substantial clonal burden. Survival analyses further clarified the prognostic impact of dynamic del(17p13) acquisition. While there was no significant difference in 1st PFS across the groups, patients who acquired del(17p13) at relapse experienced markedly worse 2nd PFS and OS compared with those who remained negative. Moreover, these patients exhibited a trend toward inferior outcomes, even compared with the persistently positive group, with shorter median 1st PFS, 2nd PFS, and OS. Importantly, acquisition of del(17p13) at relapse emerged as a prognostic factor for adverse post-relapse outcomes, suggesting that late clonal emergence of this lesion may confer a more aggressive disease biology than its presence at baseline.³² This observation challenges the conventional view that baseline del(17p13) alone defines high-risk disease and instead supports the concept that dynamic cytogenetic evolution—particularly the de novo emergence of high-risk lesions—may serve as a more powerful determinant of prognosis after relapse.

Our single-center analysis revealed that patients in the CAs-increased cohort experienced a significantly shorter 2nd PFS compared with those in the stable group, along with a clear tendency toward reduced 1st PFS and OS. This deterioration likely reflects the cumulative impact of increasing cytogenetic aberrations, serving as a surrogate marker for intensified clonal evolution, thereby promoting therapeutic resistance and accelerating relapse kinetics. In contrast, patients with stable CAs had not reached median 2nd PFS or OS at the time of analysis, underscoring a more indolent disease course and a favorable prognosis. Due to the limited number of cases within the CAs-loss subgroup, the prognostic implications of this pattern warrant further investigation and validation in larger cohorts.

Notably, within the stable-type cohort characterized by an increasing clonal burden, the 1st PFS was significantly shortened (median, 23 months), suggesting that clonal expansion alone—without the acquisition of additional CAs—may accelerate first relapse. This finding highlights that alterations in intraclonal architecture, irrespective of net CA accrual, have the potential to potentiate therapeutic resistance. Moreover, the significantly shorter 2nd PFS observed in the CAs-increased group compared with the truly stable cohort (no change in both absolute CA number and relative clonal proportion) further underscores the detrimental effect of dynamic cytogenetic shifts on sustained disease control, implying an increased susceptibility to secondary relapse or death.

We acknowledge several limitations in our study. First, as a single-center experience with 106 paired cases, the sample size is relatively modest, limiting statistical power for comprehensive subgroup stratification. Second, the baseline cytogenetic profile warrants cautious interpretation. Since the study mandated paired diagnosis-relapse samples, this cohort constitutes a relapse-ascertained subset, likely enriching for high-risk abnormalities (eg, t(4; 14)) while underrepresenting indolent markers due to extended remission. Furthermore, the lower prevalence of IGH translocations (35%) at diagnosis reflects both this selection bias and the limitations of routine dual-fusion probes; nevertheless, the rigorous application of cIg-FISH ensured high specificity. Finally, treatment-related heterogeneity introduces potential confounding, as the extended 2nd PFS was primarily driven by the administration of novel salvage therapies (eg, daratumumab and CAR-T). Additionally, the analysis was restricted to two discrete time points with a relatively short follow-up duration. However, unlike highly selected populations in controlled clinical trials, this unselected cohort authentically mirrors the complex real-world clinical landscape, enhancing the generalizability of our findings. Future prospective, multi-center studies with larger datasets and multiple sequential sampling are warranted to comprehensively delineate clonal evolution, validate the prognostic relevance of dynamic high-risk aberrations, and ultimately advance precision medicine in multiple myeloma.

Conclusion

This study demonstrates that MM is characterized by pronounced clonal heterogeneity and dynamic, therapy-driven remodeling of CAs over the disease course. Relapse is frequently accompanied by an escalation in genomic complexity, with high-risk lesions such as 1q21+ and del(17p13) emerging de novo or undergoing clonal expansion, both scenarios conferring adverse prognostic implications. Notably, dynamic acquisition of del(17p13) at relapse was identified as a predictor of poor post-relapse survival, challenging the conventional paradigm that baseline detection alone defines high-risk status. Similarly, increases in absolute CA number or clonal proportion were associated with inferior survival, underscoring the detrimental impact of active cytogenetic evolution on long-term disease control. These findings highlight the clinical importance of timely, sequential cytogenetic reassessment—particularly for 1q21+ and del(17p13)—to accurately trace clonal evolutionary trajectories, refine risk stratification, and optimize individualized therapeutic strategies.

Author contribution

Qi Chen (Writing—original draft), Tian Tian (Writing—original draft), Yu Shi (Data curation), Xuxing Shen (Investigation), Hairong Qiu (Data curation), Yuanyuan Jin (Investigation), Lijuan Chen (Conceptualization, Funding acquisition, Writing—review & editing), and Rui Guo (Conceptualization, Writing—review & editing)

Supplementary material

Supplementary material is available at The Oncologist online.

Funding

This work was supported by the National Natural Science Foundation of China (82370205), Specialized Diseases Clinical Research Fund of Jiangsu Province Hospital (XB202405).

Conflicts of interest

The authors declare that they have no competing interests.

Data availability statement

The datasets presented in this article are not readily available because of ethical/privacy restrictions. Requests to access the datasets should be directed to the corresponding author.