Skip to main content
Download PDF

Detection of early relapse in multiple myeloma patients

Cell Division volume 20, Article number: 4 (2025) Cite this article

Abstract

Background

Multiple myeloma (MM) represents the second most common hematological malignancy characterized by the infiltration of the bone marrow by plasma cells that produce monoclonal immunoglobulin. While the quality and length of life of MM patients have significantly increased, MM remains a hard-to-treat disease; almost all patients relapse. As MM is highly heterogenous, patients relapse at different times. It is currently not possible to predict when relapse will occur; numerous studies investigating the dysregulation of non-coding RNA molecules in cancer suggest that microRNAs could be good markers of relapse.

Results

Using small RNA sequencing, we profiled microRNA expression in peripheral blood in three groups of MM patients who relapsed at different intervals. In total, 24 microRNAs were significantly dysregulated among analyzed subgroups. Independent validation by RT-qPCR confirmed changed levels of miR-598-3p in MM patients with different times to relapse. At the same time, differences in the mass spectra between groups were identified using matrix-assisted laser desorption/ionization time of flight mass spectrometry. All results were analyzed by machine learning.

Conclusion

Mass spectrometry coupled with machine learning shows potential as a reliable, rapid, and cost-effective preliminary screening technique to supplement current diagnostics.

Background

Multiple myeloma (MM) is a complex hematological malignancy classified as a monoclonal gammopathy, with an estimated global incidence of 160,000 new cases per year [1, 2]. It is characterized by the uncontrolled proliferation of malignant plasma cells (PCs) and excessive secretion of monoclonal immunoglobulin (Ig) [2]. For MM diagnosis, two criteria are required: the presence of 10% or more clonal PCs in the bone marrow (BM) or a biopsy-proven plasmacytoma. The second criterion is the presence of at least one myeloma-defining events, which include so-called CRAB symptoms (hypercalcemia, renal insufficiency, anemia, and bone lesions) and three specific biomarkers: the presence of > 60% clonal PCs in BM, serum free light chain (FLC) ratio ≥ 100 and the detection of more than one focal lesion on MRI [3, 4]. MM treatment options have improved noticeably in the past few decades. New therapies, including monoclonal and bispecific antibodies and CAR-T cell therapy, have extended the overall survival of patients and improved their quality of life [5,6,7].

Nonetheless, most MM patients relapse [8]. In MM, relapse is defined as ≥ 25% increase from the lowest confirmed serum and/or urine monoclonal Ig value. In unmeasurable cases, it is defined by an increase in BM PC mass of ≥ 10% regardless of baseline. Additional criteria may include a new lesion, an increase in the number of bone lesions, etc [9,10,11]. Relapse poses a clinical challenge due to MM heterogeneity, new mutations, and drug resistance, making it difficult to determine its timing and severity [12, 13]. Accurate and minimally invasive biomarkers that enable more frequent monitoring of disease become increasingly important.

MicroRNAs (miRNAs), short non-coding RNAs involved in post-transcriptional gene regulation, play essential roles in various physiological processes, including proliferation, apoptosis, and differentiation, all of which are dysregulated in MM [13, 14]. MiRNA dysregulation was implicated in MM pathogenesis, disease progression, and drug resistance [15,16,17]. MiRNAs are also actively released from cells into body fluids as so-called circulating miRNAs [18, 19]. Unlike other RNA species, circulating miRNAs are stable and resistant to enzymatic degradation; thus, they can be used as biomarkers [20]. Circulating miRNAs from PB could represent a less invasive approach than BM biopsy.

Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) allows high-throughput, sensitive detection of molecular patterns [21,22,23]. This method has been utilized in a combined approach with machine learning (ML) algorithms to study monoclonal gammopathies, revealing molecular signatures relevant to diagnosis, prognosis, and treatment response [24,25,26,27,28,29].

In this study, we investigated biomarkers for early relapse in MM. While small RNA seq showed the best results, using MALDI-TOF MS and clinical data analyzed by predictive ML models provides economically sound and accurate alternatives for early relapse detection.

Results

Small RNA sequencing of microRNAs

Small RNA seq was performed using PB serum samples of 8 MM patients from group A, 8 MM patients from group B, and 8 MM patients from group C. Two samples (one from group A and one from group B) were excluded due to a small number of reads. Sequencing data analysis identified 748 miRNAs. Further analysis showed 360 miRNAs with more than 1 read per million in at least 7 samples.

In total, 8 different miRNAs (miR-16-2-3p, miR-148a-3p, miR-185-5p, miR-335-3p, miR-485-3p, miR-598-3p, miR-4433b-5p, miR-5010-5p) were dysregulated among all three analyzed subgroups (P < 0.1). However, the groups did not cluster properly (data not shown). When comparing groups A vs. B, and groups B vs. C, no statistically significant miRNAs were found (data not shown).

The largest differences were found between group A and group C: 24 miRNAs were identified as differentially expressed with P < 0.1 (Fig. 1); out of these, miR-16-2-3p and miR-598-3p were the most significantly differentially expressed (P < 0.05). Three miRNAs (miR-16-2-3p, miR-92b-3p, miR-598-3p) were chosen for validation in the second step of the study based on the corresponding log-fold change of expression (log-FC) and P-value (Table 1).

Fig. 1
figure 1

Clustergram and heatmap visualizing 24 miRNAs differentially expressed among 16 multiple myeloma patients, including 7 from group A (relapse within 6 months) (blue) and 8 from group C (relapse after more than 5 years) (red) (adjusted P < 0.1) in the exploration phase of the study

Full size image
Table 1 Three most significantly dysregulated miRNAs identified in the exploration phase of the study
Full size table

Consecutively, the study employed four ML algorithms to predict early relapse in MM patients, namely: PLS-DA (partial least squares discriminant analysis), k-NN (k-nearest neighbors), RF (random forest), and ANN (artificial neural network). These analyses were performed using 8 most dysregulated miRNAs between all 3 groups, 24 mostly dysregulated miRNAs between groups A and C, and 360 miRNAs with more than 1 read per million in at least 7 samples. Evaluation of the models was based on their overall accuracy in predicting the outcome. The best performance was observed when using 24 miRNAs dysregulated between groups A and C. The PLS-DA and RF algorithms exhibited the best performance when comparing groups A and C: PLS-DA: 97.7% (93.3–100%), RF: 92.8% (88.4–97.3%) (Fig. 2) as well as all 3 groups: PLS-DA: 96.8% (94.1–99.5%), RF: 84.8% (79.7–89.9%) (Fig. 3).

Fig. 2
figure 2

Comparison of 24 differentially expressed miRNAs based on accuracy. Comparing group A (relapse within 6 months) vs. group C (relapse after more than 5 years). The boxes represent the interquartile accuracy range, with black dots indicating the mean value. Whiskers show the accuracy range excluding outliers, which are plotted as blue dots

Full size image
Fig. 3
figure 3

Comparison of 24 differentially expressed miRNAs based on accuracy. Comparing group A (relapse within 6 months) vs. group B (relapse between 12 to 30 months) vs. group C (relapse after more than 5 years). The boxes represent the interquartile accuracy range, with black dots indicating the mean value. Whiskers show the accuracy range excluding outliers, which are plotted as blue dots

Full size image

Validation of microRNAs dysregulated among the diagnostic subgroups

Three miRNAs, which were the most dysregulated between group A and group C, were selected for the validation phase of the study. Based on the results of the Mann-Whitney test, significant differences in expression were observed in the case of miR-598-3p (P = 0.0145, Fig. 4). Table 2 summarizes the statistical results of the normalized expression values. The expression levels of miR-16-2-3p and miR-92b-3p did not differ between the analyzed subgroups (P = 0.2268 and P = 0.7108, respectively).

Table 2 RT-qPCR validation of miRNAs dysregulated between groups a and C
Full size table
Fig. 4
figure 4

Significantly dysregulated miR-598-3p between group A (relapse within 6 months) (blue) and C (relapse after more than 5 years) (red) (P = 0.0145)

Full size image

Based on the results of ROC analysis, miR-598-3p (AUC = 0.8636, sensitivity 64%, specificity 100%, cut off: 0.03724; Fig. 5) distinguished groups A and C with high specificity.

Fig. 5
figure 5

ROC analysis of miR-598-3p (AUC = 0.8636, sensitivity 64%, specificity 100%, cut-off: 0.03724)

Full size image

Correlation analysis was performed between miRNA levels and quantitative clinical characteristics of MM patients. Elevated levels of β2-microglobulin were observed in patients with upregulation of miR-16-2-3p and miR-92b-3p. Other correlation analyses did not show any significant results (Supplementary Table 4, Additional file 1).

MALDI-TOF MS

MALDI-TOF MS analysis was performed on a total of 29 PB serum samples, including 10 samples from group A and 19 samples from group BC (combined groups B and C). The mass spectral data were processed using the OPLS-DA method to evaluate whether molecular profiles could distinguish between patient groups based on relapse timing. The optimal number of components to be input to OPLS-DA was determined based on the R2Y and Q2Y values to prevent the overfitting of the model. Consecutively, four ML algorithms (PLS-DA, k-NN, RF, and ANN) were applied to the mass spectral data to verify whether the classification of the selected sample groups can be predicted solely based on computationally processed data (Fig. 6).

Fig. 6
figure 6

OPLS-DA score plot (left) of analysis of mass spectral data from serum samples. Comparison of established classifiers based on accuracy (right). The boxes represent the interquartile accuracy range, with black dots indicating the mean value. Whiskers show the accuracy range excluding outliers, which are plotted as blue dots

Full size image

To improve the model predictive value, clinical parameters were added to the dataset; the analyses were performed again, resulting in better-separated clusters and higher accuracy with which the two groups can be distinguished (Fig. 7). The PLS-DA and RF algorithms exhibited the best performance when comparing MALDI data alone: PLS-DA: 62.5% (57.7–67.4%), RF: 66.3% (63.2–69.5%), as well as when clinical parameters were employed: PLS-DA: 74.1% (69.1–79.0%), RF: 65.6% (62.6–68.6%).

Fig. 7
figure 7

OPLS-DA score plot (left) of analysis of mass spectral data combined with clinical parameters from serum samples. Comparison of established classifiers based on accuracy (right). The boxes represent the interquartile accuracy range, with black dots indicating the mean value. Whiskers show the accuracy range excluding outliers, which are plotted as blue dots

Full size image

Discussion

MM is a heterogeneous plasma cell dyscrasia, accounting for 1.8% of all cancer cases. Nearly all MM patients secrete monoclonal immunoglobulin, which is produced by malignant plasma cells [30, 31]. Over the last two decades, the median overall survival of MM patients has improved substantially [32, 33] due to advanced treatment options. Despite these advances, MM remains a hard-to-treat disease; most patients eventually relapse [5, 34, 35]. MM relapse is defined as the recurrence of the disease following prior treatment [11]. Due to MM heterogeneity, patients relapse from several months to years [12, 36]. Predicting the exact timing of relapse remains challenging, yet a growing body of research suggests that miRNAs have prognostic value, potentially aiding in relapse prediction [12, 34, 36].

Within the scope of this study, patients were classified into three groups (A, B, and C) based on their time to relapse. The first objective of this study was to identify differentially expressed miRNAs in PB serum among the three MM patient groups at the time of diagnosis using small RNA seq. Initial analysis of sequencing data compared group A vs. group B and group B vs. group C, revealing no statistical differences. An analysis comparing the most distant groups, A and C, revealed significant miRNA expression differences. In this comparison, 24 miRNAs were identified as differentially expressed with an adjusted P < 0.1; 2 miRNAs (miR-16-2-3p and miR-598-3p) showed an adjusted P < 0.05.

Three miRNAs (miR-16-2-3p, miR-598-3p, and miR-92b-3p), exhibited increased expression in group A compared to group C, making these two groups ideal candidates for further validation. These three miRNAs were selected for validation, confirming significant expression differences between groups A and C only for miR-598-3p (P = 0.0145).

The gene for miR-598-3p is located at 8p23.1. MiR-598-3p regulates gene expression associated with cellular communication, metabolic processes, and nervous system development [37]. Dysregulated expression of this miRNA is implicated in several diseases, including cancers; miR-598-3p is involved in tumor cell proliferation, apoptosis, and invasion [38]. This miRNA is overexpressed in colorectal cancer but downregulated in breast cancer or acute T lymphoblastic leukemia (T-ALL) [37, 39,40,41]. In T-ALL, miR-598-3p was shown to target the DEPTOR gene (encoding the mTOR-interacting protein with the DEP domain). Previous experiments highlighted the diverse functions of DEPTOR in cancer development. In T-ALL patients, DEPTOR promotes proliferation [41, 42]. In MM, DEPTOR was also differentially expressed; MM patients with low DEPTOR expression had a significantly shorter progression-free survival [43]. These studies suggest that the elevated expression of miR-598-3p, observed in group A in this study, reduces DEPTOR protein levels, potentially leading to an earlier relapse in MM patients.

The miR-15b/16 − 2 cluster is located on chromosome 3, specifically within the SMC4 gene, which encodes structural maintenance of chromosomes 4 protein, regulating chromosome stability, assembly, and segregation through a complex with SMC2 [44]. The role of the miR-15b/16 − 2 cluster on chromosome 3 is not as well understood as the miR-15a/16 − 1 cluster, which is well-characterized in MM. Several studies suggest that the miR-15b/16 − 2 cluster plays a key role in cellular proliferation and apoptosis [44, 45]. In osteoporosis studies, miR-16-2-3p was found to target and regulate WNT5A expression. Protein Wnt-5a is part of the Wnt pathway, essential for bone formation and osteoblast differentiation. Studies demonstrated that increased miR-16-2-3p levels block the Wnt pathway through Wnt-5a, resulting in inhibited osteoblast differentiation, which has also been highlighted in MM studies [46,47,48]. Several studies have investigated this miRNA in MM patients. A study by Robak et al. on circulating serum miRNAs in MM found elevated miR-16-2-3p expression in bortezomib-resistant patients compared to bortezomib-sensitive patients [49]. Another MM study associated high miR-16-2-3p expression with poor overall survival [50]. These studies indicate that excessive miR-16-2-3p expression in MM correlates with disease aggressiveness, aligning with our findings, as miR-16-2-3p expression was elevated in group A. Dysregulation of this miRNA may be predictive for MM relapse.

MiR-92b-3p was investigated in different solid tumors and has shown context-dependent roles. It acts as a tumor-suppressive miRNA in pancreatic cancer but has an oncogenic role in colorectal, renal, breast, and prostate cancer. Its increased expression promotes cell proliferation, migration, invasion, and resistance through target gene suppression [51,52,53,54,55,56]. Increased miR-92b-3p expression predicted poor prognosis in prostate cancer patients, who exhibited shorter overall survival [56]. Breast cancer patients with high miR-92b-3p expression also showed significantly shorter relapse-free survival and overall survival [51, 54]. Since miR-92b-3p showed increased levels in group A in this study, it appears to be an oncogenic miRNA in MM and could be suitable for predicting disease relapse. Further research is needed to understand its specific role in MM.

ML algorithms have shown immense potential in enhancing the diagnosis and prognosis in MM [28, 29, 57,58,59,60]. In this study, ML models, such as RF and PLS-DA, effectively differentiated patients based on miRNA profiles, demonstrating their capability to uncover subtle patterns within complex datasets. By leveraging these algorithms, we achieved higher accuracy in distinguishing between groups A and BC. This underscores the ability of ML to augment traditional bioinformatics tools by incorporating high-dimensional data, improving predictive modeling.

The next objective of this study was to compare the mass spectra of two MM patient groups with different relapse intervals using MALDI-TOF MS; it has been used to detect peptides and proteins in PB serum in various cancers [61,62,63]. Several studies were conducted in MM, where MS identified potential diagnostic and classification markers [28, 29, 64,65,66]. For instance, our previous study showed that healthy donors were differentiated from MM patients using MALDI-TOF MS combined with artificial neural networks [24].

In this study, the OPLS-DA method assessed the data discriminatory potential, demonstrating partial separation between the groups. Four ML algorithms, namely PLS-DA, k-NN, RF, and ANN, were applied to the mass spectral data to enhance classification performance. Upon integration of clinical parameters into the dataset, significant improvements were observed in both clustering and classification accuracy. Based on the results of our study, which point to differences in the mass spectra of patients with varying relapse times in MM, comparing mass spectra and clinical parameters using MALDI-TOF MS could potentially be suitable for predicting disease relapse.

Compared to standard diagnostic tools like monoclonal protein quantification, BM biopsy or imaging, our approach uses serum-based biomarkers, enabling more frequent and minimally invasive monitoring. MALDI-TOF MS offers rapid and scalable molecular profiling, while miRNA analysis provides high specificity for early relapse prediction. Together, these methods could complement current diagnostics, addressing limitations in sensitivity and invasiveness. Analysis of circulating molecular species has become increasingly important in recent years. Our study demonstrates that while small RNA seq achieves the highest accuracy in distinguishing between groups of MM patients with different relapse timing, its routine clinical application is hindered by higher time requirements, operational costs, and technical challenges compared to MALDI-TOF MS analysis. Integrating MALDI-TOF MS analysis with clinical data emerges as a more feasible and efficient approach for relapse prediction, balancing accuracy with usability and cost-effectiveness. Nonetheless, the sample sizes for miRNA profiling and MALDI-TOF MS analyses were limited, necessitating caution in interpreting the results. Larger patient cohorts are critical for validating the identified biomarkers and ensuring their generalizability.

Conclusions

This study investigates biomarkers for predicting early relapse in MM by analyzing circulating miRNAs and small molecular species in peripheral blood serum. Small RNA seq and subsequent RT-qPCR validation identified miR-598-3p as a specific biomarker with elevated expression in early relapse patients with high specificity. While small RNA sequencing offers high specificity in distinguishing MM relapse groups, its clinical adoption is limited due to high costs, technical demands, and processing time. MALDI-TOF MS, being fast and cost-effective, serves as a viable alternative for routine use. Integrating clinical parameters with MALDI-TOF MS spectra provides a scalable approach for rapid detection, with partial least squares discriminant analysis (PLS-DA) achieving an accuracy of 71.4%. These findings highlight the potential of liquid biopsies for early relapse prediction, with multidimensional datasets further enhancing their utility and possibly enabling timely intervention. However, the relatively small sample size used for both miRNA and MALDI-TOF MS analyses highlights the need for further investigation on larger patient cohorts to validate these findings.

Materials and methods

Patients’ characteristics

This study included 74 MM patients diagnosed at the University Hospital Brno, Czech Republic, between 2006 and 2021. The patients were categorized into three groups (A, B, C) based on the time of relapse after diagnosis: group A (relapse within 6 months), group B (relapse between 12 and 30 months), and group C (relapse after more than 5 years). MM patients were classified into three groups based on clinically observed relapse patterns. Group A (relapse within 6 months) reflects aggressive disease. Group C (relapse after more than 5 years) represents indolent MM with slower progression, while Group B (relapse between 12 and 30 months) encompasses an intermediate phenotype. These thresholds allowed the investigation of distinct biological mechanisms underlying relapse timing. MM samples analyzed in this study were collected at MM diagnosis. All MM patients included in this study provided informed consent, and the study was approved by the hospital’s Ethics Committee in compliance with the Declaration of Helsinki. Clinical data were obtained from the Registry of Monoclonal Gammopathies (RMG) and are summarized in Supplementary Table 1, Additional file 1. Patients received therapies per standard guidelines, with bortezomib-based regimens being predominant. Detailed therapy specifications are included in Supplementary Table 2, Additional file 1.

Sample preparation

PB and BM samples were collected from MM patients at diagnosis. PB serum was collected as described previously [67, 68]. CD138 + PCs were isolated from BM as described previously [16].

RNA isolation from peripheral blood serum

After thawing, PB serum samples were prepared as described before; total RNA enriched for small RNAs, including miRNAs, was isolated from 200 µl of each PB serum sample [67, 68].

Small RNA sequencing

Small RNA seq was performed on 8 samples from each patient group (A, B, and C) to profile miRNA expression. QIAseq miRNA Library Kit (Qiagen, Germany) was used to prepare cDNA libraries, library concentration was then assessed using Qubit dsDNA HS Assay (Thermo Fisher Scientific, USA) followed by electrophoresis measurement of their size range with High Sensitivity D1000 ScreenTape Assay for Agilent 2200 TapeStation (both Agilent Technologies, USA). Equimolar amounts of cDNA libraries were pooled at a final concentration of 4 nmol·l− 1. Sequencing was performed using the NextSeq 500 Reagent kit v2 (Illumina, USA).

RT-qPCR validation

To validate small RNA seq results, 7 MM samples from group A and 13 MM samples from group C were analyzed by RT-qPCR. Reverse transcription was performed with the TaqMan Advanced miRNA cDNA Synthesis Kit (Applied Biosystems, USA) using 2 µL of RNA according to the manufacturer’s instructions. RT-qPCR was performed using TaqMan Advanced miRNA Assays (Applied Biosystems, USA) as previously described [16]. The miRNA assay IDs are listed in Supplementary Table 3, Additional file 1.

Bioinformatics and statistical analyses

The miRge3 sequencing analysis pipeline, which supports Unique Molecular Identifiers (UMI), was applied for data processing [69]. MirBase database was used for miRNA read alignment. Count-based miRNA expression data were further analyzed by R/Bioconductor libraries [70]. Only miRNAs with at least 1 read per million in at least 7 samples were analyzed. The read counts were pre-normalized by adding normalization factors within edgeR library [71, 72] and further between-sample normalized by the voom function in LIMMA library [73]. After the normalized expression levels were determined, the differentially expressed miRNAs among studied sample groups were screened applying linear model fitting and Bayes approach. The obtained P values were adjusted for multiple testing using the Benjamini–Hochberg method.

Acquired small RNA seq data were utilized to develop a predictive model for early relapse detection. Four supervised machine learning (ML) algorithms were employed: PLS-DA (partial least squares discriminant analysis), k-NN (k-nearest neighbors), RF (random forest), and ANN (artificial neural network). The caret R library was used to train and optimize the predictive model as described before [28, 29, 74].

To assess the relative expression of miRNAs, the threshold cycle values were obtained using QuantStudio 3 software (Thermo Fisher Scientific, USA). Relative miRNA expression was normalized to miR-191-5p, based on its stable expression in small RNA seq results and manufacturer’s recommendations. The relative expression was calculated by the 2−ΔCT method. Normalized expression levels were analyzed using a nonparametric Mann-Whitney U test, with model accuracy assessed by Receiver Operating Characteristic (ROC) analysis.

Correlations between continuous variables were evaluated using Spearmans’ correlation coefficient. These analyses were conducted in GraphPad Prism 8.0.1 for Windows (GraphPad Software, Boston, Massachusetts USA, www.graphpad.com). Statistical significance was set at p < 0.05.

Matrix-assisted laser desorption/ionization mass time-of-flight mass spectrometry (MALDI-TOF MS) sample preparation

A total of 29 PB serum samples were used for the MALDI-TOF MS analysis: 10 samples from group A and 19 samples from group BC (10 samples from group B and 9 samples from group C). Plasma samples were thawed on ice and centrifuged to remove cellular debris. Extraction involved two steps: first, 50 µL of ACN was mixed with 25 µL of plasma, sonicated, and centrifuged; the supernatant was discarded. Next, 50 µL of 50% ACN with 0.1% TFA was added to the precipitate, followed by sonication and centrifugation. The collected extract was mixed with the SA matrix in a 1:1 ratio (20 mg/mL dissolved in 50% ACN supplemented with 2.5% TFA). Immediately after, 2 µL of the homogenized sample were transferred to the metal target in five technical replicates as described previously [28].

Acquisition of mass spectra

Mass spectra were recorded using MALDI-7090-TOF-TOF mass spectrometer (Shimadzu, Japan) equipped with a 2 kHz ultrafast solid-state UV laser (Nd: YAG: 355 nm). Mass spectra were recorded in the linear positive ion mode, in the mass region of 2–20 kDa; pulsed extraction was set to 7.5 kDa, frequency of the laser was 1 kHz, and the laser diameter was 100 μm. In total, 5 profiles from 1000 points were accumulated to record 1 mass spectrum.

Processing of mass spectra

Raw mass spectra in the mzML format were preprocessed using R programming language (4.0.4) to detect differentially expressed species among mass spectra. MALDIquant library, and subsequently analysis using several R libraries enabling multivariate statistical modeling were used. The spectral preprocessing workflow followed standard procedures adopted from the MALDIquant: quality control, transformation, and smoothing, baseline correction, intensity calibration, spectra alignment, trimming (2–10 kDa), and peak detection [28, 75, 76]. The feature matrix was constructed only from the detected peaks presented in at least 20% of the total mass. An established matrix was employed in further multivariate statistical analysis and to develop a predictive ML model.

Results are presented using the orthogonal projections to latent structures discriminant analysis (OPLS-DA) using ropls library [28, 29]. The R2X and R2Y coefficients indicate the proportion of variance in the x and y variables that can be explained by the model. The Q2Y parameter provides an estimation of the predictive performance of the model through the 5-fold cross-validation (CV) [77].

Acquired MALDI-TOF MS data were utilized to develop a predictive ML model able to categorize patients into classes of early relapse (group A) and late-onset relapse (groups B and C combined). Furthermore, clinical parameters were added to the model to improve its predictive value. Four supervised ML algorithms were employed as described [28, 29].

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

  1. Ludwig H, Novis Durie S, Meckl A, Hinke A, Durie B. Multiple myeloma incidence and Mortality around the Globe; interrelations between Health Access and Quality, Economic resources, and patient empowerment. Oncologist. 2020;25(9):e1406–13.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Padala SA, Barsouk A, Barsouk A, Rawla P, Vakiti A, Kolhe R, et al. Epidemiology, staging, and management of multiple myeloma. Med Sci Basel Switz. 2021;9(1):3.

    CAS  Google Scholar 

  3. Rajkumar SV. Multiple myeloma: 2022 update on diagnosis, risk stratification, and management. Am J Hematol. 2022;97(8):1086–107.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Rajkumar SV, Dimopoulos MA, Palumbo A, Blade J, Merlini G, Mateos MV, et al. International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. Lancet Oncol. 2014;15(12):e538–548.

    Article  PubMed  Google Scholar 

  5. Cowan AJ, Green DJ, Kwok M, Lee S, Coffey DG, Holmberg LA et al. Diagnosis and Management of Multiple Myeloma: A Review. JAMA [Internet]. 2022 Feb 1 [cited 2024 Jul 9];327(5):464. Available from: https://jamanetwork.com/journals/jama/fullarticle/2788522

  6. Ludwig H, Terpos E, Van De Donk N, Mateos MV, Moreau P, Dimopoulos MA et al. Prevention and management of adverse events during treatment with bispecific antibodies and CAR T cells in multiple myeloma: a consensus report of the European Myeloma Network. Lancet Oncol [Internet]. 2023 Jun [cited 2024 Oct 10];24(6):e255–69. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1470204523001596

  7. Rajkumar SV. Multiple myeloma: 2020 update on diagnosis, risk-stratification and management. Am J Hematol [Internet]. 2020 May [cited 2024 Jun 3];95(5):548–67. Available from: https://onlinelibrary.wiley.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1002/ajh.25791

  8. Laubach J, Garderet L, Mahindra A, Gahrton G, Caers J, Sezer O et al. Management of relapsed multiple myeloma: recommendations of the International Myeloma Working Group. Leukemia [Internet]. 2016 May [cited 2024 Nov 5];30(5):1005–17. Available from: https://www.nature.com/articles/leu2015356

  9. Kumar S, Paiva B, Anderson KC, Durie B, Landgren O, Moreau P et al. International Myeloma Working Group consensus criteria for response and minimal residual disease assessment in multiple myeloma. Lancet Oncol [Internet]. 2016 Aug [cited 2024 Oct 15];17(8):e328–46. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1470204516302066

  10. Lee JH, Kim SH. Treatment of relapsed and refractory multiple myeloma. BLOOD Res [Internet]. 2020 Jul 31 [cited 2024 Oct 15];55(S1):S43–53. Available from: http://www.bloodresearch.or.kr/journal/view.html?doi=10.5045/br.2020.S008

  11. Rajkumar SV, Harousseau JL, Durie B, Anderson KC, Dimopoulos M, Kyle R et al. Consensus recommendations for the uniform reporting of clinical trials: report of the International Myeloma Workshop Consensus Panel 1. Blood [Internet]. 2011 May 5 [cited 2024 Oct 15];117(18):4691–5. Available from: https://ashpublications.org/blood/article/117/18/4691/21394/Consensus-recommendations-for-the-uniform

  12. Bhatt P, Kloock C, Comenzo R. Relapsed/Refractory Multiple Myeloma: A Review of Available Therapies and Clinical Scenarios Encountered in Myeloma Relapse. Curr Oncol [Internet]. 2023 Feb 15 [cited 2024 Oct 10];30(2):2322–47. Available from: https://www.mdpi.com/1718-7729/30/2/179

  13. Lu Q, Yang D, Li H, Niu T, Tong A. Multiple myeloma: signaling pathways and targeted therapy. Mol Biomed [Internet]. 2024 Jul 4 [cited 2024 Jul 9];5(1):25. Available from: https://link.springer.com/https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s43556-024-00188-w

  14. Saliminejad K, Khorram Khorshid HR, Soleymani Fard S, Ghaffari SH. An overview of microRNAs: Biology, functions, therapeutics, and analysis methods. J Cell Physiol. 2019;234(5):5451–65.

    Article  CAS  PubMed  Google Scholar 

  15. Gregorova J, Vychytilova-Faltejskova P, Sevcikova S. Epigenetic Regulation of MicroRNA Clusters and Families during Tumor Development. Cancers [Internet]. 2021 Mar 16 [cited 2024 Oct 10];13(6):1333. Available from: https://www.mdpi.com/2072-6694/13/6/1333

  16. Vlachová M, Gregorová J, Vychytilová-Faltejsková P, Gabło NA, Radová L, Pospíšilová L et al. Involvement of Small Non-Coding RNA and Cell Antigens in Pathogenesis of Extramedullary Multiple Myeloma. Int J Mol Sci [Internet]. 2022 Nov 25 [cited 2024 Jun 3];23(23):14765. Available from: https://www.mdpi.com/1422-0067/23/23/14765

  17. Wang W, Corrigan-Cummins M, Barber EA, Saleh LM, Zingone A, Ghafoor A et al. Aberrant Levels of miRNAs in Bone Marrow Microenvironment and Peripheral Blood of Myeloma Patients and Disease Progression. J Mol Diagn [Internet]. 2015 Nov [cited 2024 Oct 10];17(6):669–78. Available from: https://linkinghub.elsevier.com/retrieve/pii/S152515781500152X

  18. Papadimitriou MA, Soureas K, Papanota AM, Tsiakanikas P, Adamopoulos PG, Ntanasis-Stathopoulos I et al. miRNA-seq identification and clinical validation of CD138 + and circulating miR-25 in treatment response of multiple myeloma. J Transl Med [Internet]. 2023 Apr 6 [cited 2024 Oct 10];21(1):245. Available from: https://biomedcentral-translational-medicine.publicaciones.saludcastillayleon.es/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12967-023-04034-5

  19. Zhao G, Jing X, Li Z, Wu X, Gao Z, Ma R. The diagnostic and prognostic values of circulating miRNA-1246 in multiple myeloma. Hematology [Internet]. 2022 Dec 31 [cited 2024 Oct 10];27(1):778–84. Available from: https://www.tandfonline.com/doi/full/https://doiorg.publicaciones.saludcastillayleon.es/10.1080/16078454.2022.2095890

  20. Kupec T, Bleilevens A, Iborra S, Najjari L, Wittenborn J, Maurer J, PLOS ONE [Internet]. Stability of circulating microRNAs in serum. Roemer K, editor. 2022 Aug 31 [cited 2024 Oct 15];17(8):e0268958. Available from: https://dx.plos.org/10.1371/journal.pone.0268958

  21. Liu S, Cheng L, Fu Y, Liu BF, Liu X. Characterization of IgG N-glycome profile in colorectal cancer progression by MALDI-TOF-MS. J Proteomics [Internet]. 2018 Jun [cited 2024 Oct 15];181:225–37. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1874391918301854

  22. Teunissen C, Koel-Simmelink M, Pham T, Knol J, Khalil M, Trentini A et al. Identification of biomarkers for diagnosis and progression of MS by MALDI-TOF mass spectrometry. Mult Scler J [Internet]. 2011 Jul [cited 2024 Oct 15];17(7):838–50. Available from: https://journals.sagepub.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1177/1352458511399614

  23. Duncan M, DeMarco ML. MALDI-MS: Emerging roles in pathology and laboratory medicine. Clin Mass Spectrom [Internet]. 2019 Aug [cited 2024 Oct 15];13:1–4. Available from: https://linkinghub.elsevier.com/retrieve/pii/S237699981930008X

  24. Deulofeu M, Kolářová L, Salvadó V, María Peña-Méndez E, Almáši M, Štork M et al. Rapid discrimination of multiple myeloma patients by artificial neural networks coupled with mass spectrometry of peripheral blood plasma. Sci Rep [Internet]. 2019 May 28 [cited 2024 Oct 15];9(1):7975. Available from: https://www.nature.com/articles/s41598-019-44215-1

  25. Eveillard M, Korde N, Ciardiello A, Diamond B, Lesokhin A, Mailankody S et al. Using MALDI-TOF mass spectrometry in peripheral blood for the follow up of newly diagnosed multiple myeloma patients treated with daratumumab-based combination therapy. Clin Chim Acta [Internet]. 2021 May [cited 2024 Oct 15];516:136–41. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0009898121000371

  26. Gregorova J, Vychytilova-Faltejskova P, Kramarova T, Knechtova Z, Almasi M, Stork M et al. Proteomic analysis of the bone marrow microenvironment in extramedullary multiple myeloma patients. Neoplasma [Internet]. 2022 [cited 2024 Jun 3];69(02):412–24. Available from: http://www.elis.sk/index.php?page=shop.product_details&flypage=flypage.tpl&product_id=7523&category_id=180&option=com_virtuemart

  27. Murray DL, Puig N, Kristinsson S, Usmani SZ, Dispenzieri A, Bianchi G, International Myeloma Working Group Mass Spectrometry Committee Report. Mass spectrometry for the evaluation of monoclonal proteins in multiple myeloma and related disorders: an. Blood Cancer J [Internet]. 2021 Feb 1 [cited 2024 Oct 15];11(2):24. Available from: https://www.nature.com/articles/s41408-021-00408-4

  28. Pečinka L, Vlachová M, Moráň L, Gregorová J, Porokh V, Kovačovicová P et al. Improved Screening of Monoclonal Gammopathy Patients by MALDI-TOF Mass Spectrometry. J Am Soc Mass Spectrom [Internet]. 2023 Dec 6 [cited 2024 Oct 15];34(12):2646–53. Available from: https://pubs.acs.org/doi/10.1021/jasms.3c00166

  29. Vlachová M, Pečinka L, Gregorová J, Moráň L, Růžičková T, Kovačovicová P et al. Liquid biopsy of peripheral blood using mass spectrometry detects primary extramedullary disease in multiple myeloma patients. Sci Rep [Internet]. 2024 Aug 13 [cited 2024 Oct 15];14(1):18777. Available from: https://www.nature.com/articles/s41598-024-69408-1

  30. Dimopoulos MA, Moreau P, Terpos E, Mateos MV, Zweegman S, Cook G et al. Multiple myeloma: EHA-ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up†. Ann Oncol [Internet]. 2021 Mar [cited 2024 Oct 30];32(3):309–22. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0923753420431692

  31. Kumar SK, Rajkumar V, Kyle RA, Van Duin M, Sonneveld P, Mateos MV et al. Multiple myeloma. Nat Rev Dis Primer [Internet]. 2017 Jul 20 [cited 2024 Jun 3];3(1):17046. Available from: https://www.nature.com/articles/nrdp201746

  32. Cliff ERS, Mohyuddin GR. Overall survival as a primary end point in multiple myeloma trials. Nat Rev Clin Oncol [Internet]. 2022 Sep [cited 2024 Oct 30];19(9):565–6. Available from: https://www.nature.com/articles/s41571-022-00665-7

  33. Marinac CR, Ghobrial IM, Birmann BM, Soiffer J, Rebbeck TR. Dissecting racial disparities in multiple myeloma. Blood Cancer J [Internet]. 2020 Feb 17 [cited 2024 Jul 15];10(2):19. Available from: https://www.nature.com/articles/s41408-020-0284-7

  34. Puła A, Robak P, Jarych D, Mikulski D, Misiewicz M, Drozdz I et al. The Relationship between Serum miRNAs and Early Mortality in Multiple Myeloma Patients Treated with Bortezomib-Based Regimens. Int J Mol Sci [Internet]. 2023 Feb 2 [cited 2024 Oct 30];24(3):2938. Available from: https://www.mdpi.com/1422-0067/24/3/2938

  35. Sonneveld P, Broijl A. Treatment of relapsed and refractory multiple myeloma. Haematologica [Internet]. 2016 Apr [cited 2024 Oct 30];101(4):396–406. Available from: http://www.haematologica.org/lookup/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.3324/haematol.2015.129189

  36. Chim CS, Kumar SK, Orlowski RZ, Cook G, Richardson PG, Gertz MA et al. Management of relapsed and refractory multiple myeloma: novel agents, antibodies, immunotherapies and beyond. Leukemia [Internet]. 2018 Feb [cited 2024 Oct 30];32(2):252–62. Available from: https://www.nature.com/articles/leu2017329

  37. Insenser M, Quintero A, De Lope S, Álvarez-Blasco F, Martínez-García MÁ, Luque-Ramírez M et al. Validation of circulating microRNAs miR-142-3p and miR-598-3p in women with polycystic ovary syndrome as potential diagnostic markers. Hum Reprod [Internet]. 2023 May 2 [cited 2024 Oct 30];38(5):951–60. Available from: https://academic.oup.com/humrep/article/38/5/951/7080093

  38. Lin X, Huang X, Wang L, Liu W. The long noncoding RNA MALAT1/microRNA-598-3p axis regulates the proliferation and apoptosis of retinoblastoma cells through the PI3K/AKT pathway. Mol Vis. 2022;28:269–79.

    PubMed  PubMed Central  Google Scholar 

  39. Fu L, Li Z, Zhu J, Wang P, Fan G, Dai Y et al. Serum expression levels of microRNA-382-3p, – 598-3p, – 1246 and – 184 in breast cancer patients. Oncol Lett [Internet]. 2016 Jul [cited 2024 Oct 30];12(1):269–74. Available from: https://www.spandidos-publications.com/https://doiorg.publicaciones.saludcastillayleon.es/10.3892/ol.2016.4582

  40. Li K, Fang Y, Liao J, Duan J, Feng L, Luo X et al. Upregulation of miR–598 promotes cell proliferation and cell cycle progression in human colorectal carcinoma by suppressing INPP5E expression. Mol Med Rep [Internet]. 2017 Dec 6 [cited 2024 Oct 30]; Available from: http://www.spandidos-publications.com/https://doiorg.publicaciones.saludcastillayleon.es/10.3892/mmr.2017.8207

  41. Qiang Z, Feng J, Wang C, Zheng M, Wen Z. MiR-598-3p functions as a tumor suppressor in pediatric T-cell acute lymphoblastic leukemia. Trop J Pharm Res [Internet]. 2022 Jan 17 [cited 2024 Oct 30];20(3):497–503. Available from: https://www.ajol.info/index.php/tjpr/article/view/220126

  42. Chen X, Xiong X, Cui D, Yang F, Wei D, Li H et al. DEPTOR is an in vivo tumor suppressor that inhibits prostate tumorigenesis via the inactivation of mTORC1/2 signals. Oncogene [Internet]. 2020 Feb 13 [cited 2024 Oct 30];39(7):1557–71. Available from: https://www.nature.com/articles/s41388-019-1085-y

  43. Quwaider D, Corchete LA, Misiewicz-Krzeminska I, Sarasquete ME, Pérez JJ, Krzeminski P et al. DEPTOR maintains plasma cell differentiation and favorably affects prognosis in multiple myeloma. J Hematol OncolJ Hematol Oncol [Internet]. 2017 Dec [cited 2024 Oct 30];10(1):92. Available from: https://biomedcentral-jhoonline.publicaciones.saludcastillayleon.es/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13045-017-0461-8

  44. Kallingal A, Thankachan S, Venkatesh T, Kabbekodu SP, Suresh PS. Role of miR-15b/16–2 cluster network in endometrial cancer: An in silico pathway and prognostic analysis. Meta Gene [Internet]. 2022 Feb [cited 2024 Oct 30];31:101018. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2214540022000093

  45. Lovat F, Fassan M, Gasparini P, Rizzotto L, Cascione L, Pizzi M et al. miR-15b/16 – 2 deletion promotes B-cell malignancies. Proc Natl Acad Sci [Internet]. 2015 Sep 15 [cited 2024 Oct 30];112(37):11636–41. Available from: https://pnas.doi/full/10.1073/pnas.1514954112

  46. Duan L, Zhao H, Xiong Y, Tang X, Yang Y, Hu Z et al. miR-16-2* Interferes with WNT5A to Regulate Osteogenesis of Mesenchymal Stem Cells. Cell Physiol Biochem [Internet]. 2018 [cited 2024 Oct 30];51(3):1087–102. Available from: https://karger.com/CPB/article/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1159/000495489

  47. Frenquelli M, Caridi N, Antonini E, Storti F, Viganò V, Gaviraghi M et al. The WNT receptor ROR2 drives the interaction of multiple myeloma cells with the microenvironment through AKT activation. Leukemia [Internet]. 2020 Jan [cited 2024 Oct 30];34(1):257–70. Available from: https://www.nature.com/articles/s41375-019-0486-9

  48. Qiang YW, Chen Y, Brown N, Hu B, Epstein J, Barlogie B, et al. Characterization of Wnt/beta-catenin signalling in osteoclasts in multiple myeloma. Br J Haematol. 2010;148(5):726–38.

    Article  CAS  PubMed  Google Scholar 

  49. Robak P, Dróżdż I, Jarych D, Mikulski D, Węgłowska E, Siemieniuk-Ryś M et al. The Value of Serum MicroRNA Expression Signature in Predicting Refractoriness to Bortezomib-Based Therapy in Multiple Myeloma Patients. Cancers [Internet]. 2020 Sep 9 [cited 2024 Oct 30];12(9):2569. Available from: https://www.mdpi.com/2072-6694/12/9/2569

  50. Katiyar A, Kaur G, Rani L, Jena L, Singh H, Kumar L et al. Genome-wide identification of potential biomarkers in multiple myeloma using meta-analysis of mRNA and miRNA expression data. Sci Rep [Internet]. 2021 May 26 [cited 2024 Oct 30];11(1):10957. Available from: https://www.nature.com/articles/s41598-021-90424-y

  51. Du Y, Miao Z, Wang K, Lv Y, Qiu L, Guo L. Expression levels and clinical values of miR-92b-3p in breast cancer. World J Surg Oncol [Internet]. 2021 Dec [cited 2024 Oct 30];19(1):239. Available from: https://biomedcentral-wjso.publicaciones.saludcastillayleon.es/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12957-021-02347-7

  52. Gong L, Ren M, Lv Z, Yang Y, Wang Z. miR-92b-3p Promotes Colorectal Carcinoma Cell Proliferation, Invasion, and Migration by Inhibiting FBXW7 In Vitro and In Vivo. DNA Cell Biol [Internet]. 2018 May [cited 2024 Oct 30];37(5):501–11. Available from: http://www.liebertpub.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1089/dna.2017.4080

  53. Long M, Zhan M, Xu S, Yang R, Chen W, Zhang S et al. miR-92b-3p acts as a tumor suppressor by targeting Gabra3 in pancreatic cancer. Mol Cancer [Internet]. 2017 Dec [cited 2024 Oct 30];16(1):167. Available from: https://biomedcentral-molecular-cancer.publicaciones.saludcastillayleon.es/articles/https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12943-017-0723-7

  54. Sun Y, Feng Y, Zhang G, Xu Y. The endonuclease APE1 processes miR-92b formation, thereby regulating expression of the tumor suppressor LDLR in cervical cancer cells. Ther Adv Med Oncol [Internet]. 2019 Jan [cited 2024 Oct 30];11:1758835919855859. Available from: https://journals.sagepub.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1177/1758835919855859

  55. Wang C, Uemura M, Tomiyama E, Matsushita M, Koh Y, Nakano K et al. MicroRNA-92b‐3p is a prognostic oncomiR that targets TSC1 in clear cell renal cell carcinoma. Cancer Sci [Internet]. 2020 Apr [cited 2024 Oct 30];111(4):1146–55. Available from: https://onlinelibrary.wiley.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1111/cas.14325

  56. Wang G, Cheng B, Jia R, Tan B, Liu W. Altered expression of microRNA–92b–3p predicts survival outcomes of patients with prostate cancer and functions as an oncogene in tumor progression. Oncol Lett [Internet]. 2020 Nov 3 [cited 2024 Oct 30];21(1):1–1. Available from: http://www.spandidos-publications.com/https://doiorg.publicaciones.saludcastillayleon.es/10.3892/ol.2020.12265

  57. Chen X, Zhang Y, Li X, Yang Z, Liu A, Yu X. Diagnosis and staging of multiple myeloma using serum-based laser-induced breakdown spectroscopy combined with machine learning methods. Biomed Opt Express. 2021;12(6):3584–96.

    Article  PubMed  PubMed Central  Google Scholar 

  58. Mosquera Orgueira A, González Pérez MS, Díaz Arias JÁ, Antelo Rodríguez B, Alonso Vence N, Bendaña López Á, et al. Survival prediction and treatment optimization of multiple myeloma patients using machine-learning models based on clinical and gene expression data. Leukemia. 2021;35(10):2924–35.

    Article  CAS  PubMed  Google Scholar 

  59. Xiong X, Wang J, Hu S, Dai Y, Zhang Y, Hu C. Differentiating between multiple myeloma and metastasis subtypes of lumbar vertebra lesions using machine learning-based Radiomics. Front Oncol. 2021;11:601699.

    Article  PubMed  PubMed Central  Google Scholar 

  60. Yan W, Shi H, He T, Chen J, Wang C, Liao A, et al. Employment of Artificial Intelligence Based on Routine Laboratory Results for the early diagnosis of multiple myeloma. Front Oncol. 2021;11:608191.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  61. Han X, Li D, Wang S, Lin Y, Liu Y, Lin L et al. Serum amino acids quantification by plasmonic colloidosome-coupled MALDI-TOF MS for triple-negative breast cancer diagnosis. Mater Today Bio [Internet]. 2022 Dec [cited 2024 Dec 4];17:100486. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2590006422002848

  62. Lai X, Guo K, Huang W, Su Y, Chen S, Li Q et al. Combining MALDI-MS with machine learning for metabolomic characterization of lung cancer patient sera. Anal Methods [Internet]. 2022 [cited 2024 Dec 4];14(5):499–507. Available from: https://xlink.rsc.org/?DOI=D1AY01940F

  63. Lee JW, Lee K, Ahn SH, Son BH, Ko BS, Kim HJ et al. Potential of MALDI-TOF-based serum N-glycan analysis for the diagnosis and surveillance of breast cancer. Sci Rep [Internet]. 2020 Nov 5 [cited 2024 Dec 4];10(1):19136. Available from: https://www.nature.com/articles/s41598-020-76195-y

  64. Bhattacharyya S, Epstein J, Suva LJ. Biomarkers that Discriminate Multiple Myeloma Patients with or without Skeletal Involvement Detected Using SELDI-TOF Mass Spectrometry and Statistical and Machine Learning Tools. Dis Markers [Internet]. 2006 Jan [cited 2024 Oct 30];22(4):245–55. Available from: https://onlinelibrary.wiley.com/doi/10.1155/2006/728296

  65. He A, Bai J, Huang C, Yang J, Zhang W, Wang J et al. Detection of serum tumor markers in multiple myeloma using the CLINPROT system. Int J Hematol [Internet]. 2012 Jun [cited 2024 Oct 30];95(6):668–74. Available from: http://link.springer.com/https://doiorg.publicaciones.saludcastillayleon.es/10.1007/s12185-012-1080-3

  66. Wang Q, Li Y, Liang Y, Hu C, Zhai Y, Zhao G et al. Construction of A Multiple Myeloma Diagnostic Model by Magnetic Bead-Based MALDI‐TOF Mass Spectrometry of Serum and Pattern Recognition Software. Anat Rec [Internet]. 2009 Apr [cited 2024 Oct 30];292(4):604–10. Available from: https://anatomypubs.onlinelibrary.wiley.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1002/ar.20871

  67. Kubiczková L, Kryukov F, Slabý O, Dementyeva E, Jarkovský J, Nekvindová J, et al. Circulating serum microRNAs as novel diagnostic and prognostic biomarkers for multiple myeloma and monoclonal gammopathy of undetermined significance. Haematologica. 2014;99(3):511–8.

    Article  PubMed  PubMed Central  Google Scholar 

  68. Sedlarikova L, Bollova B, Radova L, Brozova L, Jarkovsky J, Almasi M et al. Circulating exosomal long noncoding RNA PRINS—First findings in monoclonal gammopathies. Hematol Oncol [Internet]. 2018 Dec [cited 2024 Oct 29];36(5):786–91. Available from: https://onlinelibrary.wiley.com/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1002/hon.2554

  69. Patil AH, Halushka MK. miRge3.0: a comprehensive microRNA and tRF sequencing analysis pipeline. NAR Genomics Bioinforma [Internet]. 2021 Jun 23 [cited 2024 Oct 21];3(3):lqab068. Available from: https://academic.oup.com/nargab/article/doi/https://doiorg.publicaciones.saludcastillayleon.es/10.1093/nargab/lqab068/6325159

  70. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing [Internet]. Vienna, Austria. 2020. Available from: https://www.R-project.org/

  71. McCarthy DJ, Chen Y, Smyth GK. Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Res [Internet]. 2012 May 1 [cited 2024 Oct 21];40(10):4288–97. Available from: https://academic.oup.com/nar/article/40/10/4288/2411520

  72. Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics [Internet]. 2010 Jan 1 [cited 2024 Oct 21];26(1):139–40. Available from: https://academic.oup.com/bioinformatics/article/26/1/139/182458

  73. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res [Internet]. 2015 Apr 20 [cited 2024 Oct 21];43(7):e47–e47. Available from: http://academic.oup.com/nar/article/43/7/e47/2414268/limma-powers-differential-expression-analyses-for

  74. Kuhn M. Building Predictive Models in R Using the caret Package. J Stat Softw [Internet]. 2008 [cited 2024 Dec 9];28(5). Available from: http://www.jstatsoft.org/v28/i05/

  75. Gibb S, Strimmer K. MALDIquant: a versatile R package for the analysis of mass spectrometry data. Bioinformatics [Internet]. 2012 Sep 1 [cited 2024 Dec 9];28(17):2270–1. Available from: https://academic.oup.com/bioinformatics/article/28/17/2270/246552

  76. Pečinka L, Moráň L, Kovačovicová P, Meloni F, Havel J, Pivetta T et al. Intact cell mass spectrometry coupled with machine learning reveals minute changes induced by single gene silencing. Heliyon [Internet]. 2024 May [cited 2024 Dec 4];10(9):e29936. Available from: https://linkinghub.elsevier.com/retrieve/pii/S240584402405967X

  77. Thévenot EA, Roux A, Xu Y, Ezan E, Junot C. Analysis of the Human Adult Urinary Metabolome Variations with Age, Body Mass Index, and Gender by Implementing a Comprehensive Workflow for Univariate and OPLS Statistical Analyses. J Proteome Res [Internet]. 2015 Aug 7 [cited 2024 Dec 9];14(8):3322–35. Available from: https://pubs.acs.doi/10.1021/acs.jproteome.5b00354

Download references

Acknowledgements

The authors would like to thank all patients and their caregivers for participating in this study.

Funding

This study was supported by the internal grant of Masaryk University MUNI/A/1587/2023, by grant of the Ministry of Health, Czech Republic Conceptual Development of Research Organization, FNBr, 65269705, by grant of Ministry of Health AZV NU21-03-00076 and by the project National Institute for Cancer Research (Programme EXCELES, ID Project No. LX22NPO5102) - Funded by the European Union – Next Generation EU.

Author information

Authors and Affiliations

  1. Babak Myeloma Group, Department of Pathophysiology, Faculty of Medicine, Masaryk University, Brno, Czech Republic

    Tereza Růžičková, Monika Vlachová, Monika Brychtová, Simona Ševčíková & Sabina Ševčíková

  2. Department of Internal Medicine, Hematology and Oncology, University Hospital Brno, Brno, Czech Republic

    Tereza Růžičková, Marie Jarošová & Luděk Pour

  3. Research Centre for Applied Molecular Oncology (RECAMO), Masaryk Memorial Cancer Institute, Brno, Czech Republic

    Lukáš Pečinka

  4. International Clinical Research Center, St. Anne’s University Hospital Brno, Brno, Czech Republic

    Lukáš Pečinka

  5. Centre for Molecular Medicine, Central European Institute of Technology, Masaryk University, Brno, Czech Republic

    Marek Večeřa & Lenka Radová

  6. Department of Chemistry, Faculty of Science, Masaryk University, Brno, Czech Republic

    Josef Havel

  7. Department of Clinical Hematology, University Hospital Brno, Brno, Czech Republic

    Sabina Ševčíková

Authors
  1. Tereza Růžičková
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Monika Vlachová
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Lukáš Pečinka
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Monika Brychtová
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Marek Večeřa
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Lenka Radová
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Simona Ševčíková
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Marie Jarošová
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Josef Havel
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Luděk Pour
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Sabina Ševčíková
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Conceptualization: MVl, LPe, MB and SaŠ; investigation: LPe, MB, TR, MJ, and Lpo; formal analysis: Lpe, MB, TR and LR; funding acquisition: SaŠ; methodology: MVl, Lpe and SaŠ; writing-original draft: TR; writing-review and editing: TR, MVl, LPe, MB, MVe, SiŠ, MJ, LPo, JH and SaŠ. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Sabina Ševčíková.

Ethics declarations

Ethics approval and consent to participate

The study was approved by the University Hospital Brno Ethics Committee. All patients included in this study signed an informed consent form approved by the hospital’s Ethics committee in accordance with the current Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Růžičková, T., Vlachová, M., Pečinka, L. et al. Detection of early relapse in multiple myeloma patients. Cell Div 20, 4 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13008-025-00143-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13008-025-00143-3

Keywords

Cell Division

ISSN: 1747-1028

Contact us