Skip to main content
  • Research article
  • Open access
  • Published:

bFGF could be a biomarker of malignancy in RS3PE syndrome: an ambispective single-center cohort analysis of 51 patients

Abstract

Objectives

Remitting seronegative symmetrical synovitis with pitting edema (RS3PE) is a rare inflammatory arthritis, with a higher incidence of malignancy. The aim of this study is to identify biomarkers for predicting malignancy in RS3PE.

Methods

A total of 51 patients with RS3PE from September 2007 to May 2019 were retrospectively reviewed and followed for up to 5 years, with 15 patients with osteoarthritis (OA) and 14 patients with elderly-onset rheumatoid arthritis (EORA) as disease controls. Serum levels of angiogenesis cytokines were measured by electrochemiluminescent immunoassay and Luminex Human Magnetic Assay. Clinical data and laboratory parameters were analyzed to identify risk factors for malignancy.

Results

A total of forty-eight RS3PE patients (94.1%) were available with follow-up data; 8 patients (16.7%) were diagnosed with malignancy, of which 6 patients were hematological tumor; and 2 patients were solid tumors. Serum levels of basic fibroblast growth factor (bFGF) were exclusively higher in RS3PE patients with malignancy [14.21 (7.52, 23.18) ng/mL] than RS3PE patients without malignancy [4.32 (2.88, 7.42) ng/mL], OA [3.20 (2.20, 5.30) ng/mL], and EORA [3.20 (2.20, 5.30) ng/mL]. The optimal cut-off value of bFGF for malignancy was 10ng/mL in RS3PE. Logistic regression analysis indicated that elevation of bFGF was a risk factor for malignancy in RS3PE.

Conclusions

This study indicated that bFGF was elevated in RS3PE patients with malignancy and could serve as a biomarker for predicting paraneoplastic RS3PE.

Introduction

Remitting seronegative symmetrical synovitis with pitting edema (RS3PE) is a rare elderly-onset inflammatory arthritis, characterized by symmetrical involvement of small joints and marked pitting edema on the dorsum of the hands and feet [1, 2]. In addition, a higher incidence of malignancy was reported in RS3PE after the first symptom onset or during follow-up [2,3,4]. However, no significant demographic or clinical differences were observed between idiopathic and paraneoplastic cases of RS3PE, which suggests the importance of investigating novel serum tumor markers.

Previous studies have found two angiogenesis cytokines, namely vascular endothelial growth factor (VEGF) and matrix metalloproteinase (MMP)-3, were involved in the pathogenesis of RS3PE along with malignancy [5,6,7,8]. But later findings indicated that elevated levels of VEGF were also characteristics of infections and organizing pneumonia in RS3PE [9, 10]. MMP-3 was elevated in RS3PE patients with solid malignancy [11], but it was also an indicator for active arthritis [12]. It still remains elusive whether there is a specific biomarker for identifying arthritis with malignancy.

Since angiogenesis plays an important role in the pathogenesis and progression of cancer, we simultaneously evaluated the serum levels of 12 angiogenesis cytokines in paraneoplastic RS3PE, comparing with idiopathic RS3PE, osteoarthritis (OA), and elderly-onset rheumatoid arthritis (EORA) in the current study. We aimed to discover some novel markers for predicting malignancy in RS3PE.

Methods

Patients enrolled

A single-center cohort study was performed in the Department of Rheumatology, Peking University People’s Hospital. Fifty-one patients diagnosed with RS3PE syndrome were consecutively enrolled from September 2007 to May 2019, fulfilling the following criteria: (1) bilateral pitting edema of dorsum of hands and/or feet, (2) abrupt onset of polyarthritis, (3) age > 50 years, and (4) seronegative for rheumatoid factor (RF) [13]. As disease controls, 15 patients with OA and 14 patients with EORA were also enrolled, with sex- and age-matched. All the participants in disease control groups were excluded from malignancy.

Study design and data collection

Patients with RS3PE were followed up for 5 years or monitored up to February 29, 2021, if they enrolled after February 28, 2016. The primary clinical outcome was the occurrence of malignancy. The baseline clinical and laboratory characteristics, coexistence of malignancy, and response to treatment were obtained from the medical records. If no follow-up data was available in our center, we contacted the family members to acquire the physical status (especially the occurrence of malignancy) confirmed by regular medical examination reports.

Measurement of angiogenesis cytokines

The serum samples were collected from 45 RS3PE patients and all disease controls at baseline and stored at −80°C in polypropylene microfuge tubes without thawing before the test. Serum levels of VEGF-A, VEGF-C, VEGF-D, FMS-like tyrosine kinase 1 (Flt-1), Tie-2, placental growth factor (PIGF), and basic fibroblast growth factor (bFGF) were measured via electrochemiluminescent immunoassay by V-PLEX Plus Angiogenesis Panel 1 Human Kit (Meso Scale Discovery). Serum levels of MMP-1, MMP-3, MMP-7, mesothelin, and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) were determined with Luminex Human Magnetic Assay (5-Plex) LXSAHM-05 (R&D).

Statistical analysis

Data analyses were performed using SPSS 23.0 for Windows. Continuous data with the normal distribution were expressed as the mean ± standard, and differences between groups were analyzed by one-way ANOVA. Continuous data with skewed distribution were expressed as median (P25, P75), and differences between groups were analyzed by Kruskal-Wallis test. Dichotomous variables were reported as frequency (percentages), and differences between groups were compared using the chi-square test (or Fisher’s exact test when appropriate). The cut-off value of bFGF in RS3PE patients with malignancy was determined by receiver operating characteristic (ROC) methods. Univariate and multivariate logistic regression analyses were adopted to identify risk factors of malignancy. The variables assessed in the univariate regression analysis were entered as independent variables in multivariate logistic regression analysis when P value <0.1. Two-sided P < 0.05 was considered statistically significant. The P value was adjusted by Bonferroni correction in multiple tests.

Results

Clinical and laboratory features of RS3PE patients

The clinical and laboratory features of overall RS3PE patients were shown in Table 1. Forty-eight patients (94.1%) were available with follow-up data, and a total of 26 patients (54.1%) completed a 5-year follow-up, and the follow-up time for the remaining 22 patients ranged from 1 month to 54 months. During the study period, eight of them (16.7%) were diagnosed with malignant tumors. Twenty-eight patients (54%) were male, and the average age at onset was 73.24±9.23 years. Pitting edema was seen in the hands of 45 patients (88.2%) and in the feet of 27 patients (52.9%). Weight loss was seen in 19 patients (37.3%). Patients had an elevated level of C-reactive protein (CRP) (43.9 [22.8, 82.0] mg/dL) and erythrocyte sedimentation rate (ESR) (55.37±34.12 mm/h). All patients had a normal level of carcinoembryonic antigen (<4.7 ng/mL). Eight patients (17.4%) had an elevated level of neuroenolase (>16.3ng/mL), 3 patients (5.8%) had an elevated level of cytokeratin 19 fragment (>3.3ng/mL), 2 patients (3.9%) had an elevated level of carbohydrate antigen 19-9 (>39U/mL), and only one patient (1.9%) had an elevated level of alpha-fetoprotein (>7ng/mL). Antinuclear antibody (ANA) was positive (≥1:80) in 6/51 (11.8%), and anti-Ro-52 was positive in 5/51 (9.8%). The median of the initial prednisolone dose was 15 mg/day, and 44/51 (89.8%) showed a good response to prednisolone.

Table 1 Characteristics of RS3PE patients with or without malignancy

Comparison between RS3PE patients with and without malignancy

The detailed clinical profiles of the eight RS3PE patients with malignancy were displayed in Table 2. The prevalence of malignancy was 16.7% (8/48), six were hematological tumors, and 2 were solid tumors. The time from the onset of arthritis to confirmation of malignancy was from 2 months to 3 years. In these 6 patients with hematological tumors, 4 patients were diagnosed within 6 months from arthritis onset, and 3 patients showed poor response to low-dose prednisolone. Both patients with solid tumors were diagnosed 2 years after arthritis onset, and one of them (50%) was resistant to low-dose prednisolone.

Table 2 Detailed clinical profiles of RS3PE patients with malignancy

We next compared the clinical and laboratory features between patients with or without malignancy (Table 1). Better response to prednisolone was found in patients without malignancy (n=38/40, 97.4%) than patients with malignancy (n=3/7, 42.8%). However, significant differences were not seen in demographic figures (age and gender), clinical features (patterns of edema and weight loss), and laboratory features (CRP, ESR, immunoglobulin, complement, and tumor markers).

Serum levels of angiogenesis cytokines among RS3PE with/without malignancy, OA, and EORA

Twelve angiogenesis cytokines were measured, and the results were demonstrated in Table 3. Serum levels of bFGF were exclusively higher in RS3PE patients with malignancy [14.21 (7.52, 23.18) ng/mL] than RS3PE patients without malignancy [4.32 (2.88, 7.42) ng/mL], OA [3.23 (1.96, 5.59) ng/mL], and EORA [3.20 (2.20, 5.30) ng/mL]. However, there were no significant differences in serum levels of VEGF-A, VEGF-C, VEGF-D, Flt-1, Tie-2, PIGF, MMP-1, MMP-3, MMP-7, mesothelin, and TRAIL among different groups. Figure 1 showed the ROC curve of bFGF with an AUC value of 0.817, and the optimal cut-off value was 10ng/mL; the sensitivity was 75% and the specificity was 89.5%.

Table 3 Comparison of angiogenesis cytokines among RS3PE with/without malignancy, OA, and EORA
Fig. 1
figure 1

ROC curve for bFGF in predicting malignancy in patients with RS3PE. ROC receiver operating characteristic, bFGF basic fibroblast growth factor, RS3PE remitting seronegative symmetrical synovitis with pitting edema

Risk factors for malignancy in RS3PE

As shown in Table 4, the results of univariate logistics models found that bilateral pitting edema of the hands (OR=0.074, 95%CI (0.006–0.968), P=0.047) and good response to prednisolone (OR=0.039, 95%CI (0.005–0.311), P=0.002) were negatively associated with malignancy in RS3PE, and elevation of bFGF (>10ng/ml) (OR=14.084, 95%CI (2.421–83.332), P=0.003) was positively associated with malignancy in RS3PE. Then, the multivariate logistic models showed elevation of bFGF was a unique risk factor for malignancy (OR=14.667, 95%CI (2.029–106.038), P=0.008).

Table 4 Risk factors for malignancy in RS3PE by logistic models

Discussion

In the present study, we reviewed the clinical and laboratory features and simultaneously analyzed multiple angiogenesis cytokines in RS3PE patients with malignancy. We found elevation of bFGF might be a useful predictor for malignancy in RS3PE.

Increased associated malignancy in RS3PE has been reported since 1985, including hematological malignancies and solid tumors [2, 14,15,16], and the average malignancy rate was estimated to 20% [2], which is similar to our study. Although hematological malignancies were the primary tumors in our study and most of them were diagnosed within the first 6 months, both two associated solid tumors were confirmed during the follow-up. Besides, a French study of six men with RS3PE demonstrated that all solid malignancy was discovered during a 5-year follow-up [17]. These findings indicate that solid tumors might be relatively insidious in RS3PE-related malignancies, reminding rheumatologists the importance of tumor screening during the follow-up.

Poor response to low-dose prednisolone is associated with malignancies in RS3PE in our study, and some reported cases of paraneoplastic RS3PE are also revealed poor response to glucocorticoid [2, 18]. However, rapid response to glucocorticoid therapy is also found for some paraneoplastic RS3PE, and there are no clinical variables for predicting malignancy in RS3PE [1, 3, 4, 19], which calls for more effective biomarkers.

Interestingly, our study discovered bFGF is the only angiogenesis cytokine which is elevated particularly in RS3PE-associated malignancy, and further multiple logistic regressions revealed elevation of bFGF may serve as a marker for predicting malignancy in RS3PE. bFGF, also known as fibroblast growth factor 2 (FGF-2), is one of the prototypes of the FGF family, which signals through FGF receptors (FGFRs) and promotes growth and differentiation of a broad spectrum of cell types, including dermal fibroblasts, keratinocytes, endothelial cells, and melanocytes [20,21,22]. In addition, bFGF also plays a critical role in promoting tumor angiogenesis and metastasis and has been shown to be involved in the invasion and progression of solid and hematological malignancies [21, 23,24,25,26].

Apart from tumor genesis, it has also been found that bFGF could stimulate osteoclastogenesis and promote bone absorption through binding to FGFRs, and is the only one of the bone-resorptive cytokines that are highly expressed in the synovial fluid of RA patients [27,28,29]. Thus, significantly higher serum bFGF in RS3PE may reflect the secretion of bFGF in situ of tumor tissues as well as synovium, suggesting that bFGF might play an important role in the pathogenesis of RS3PE. Besides, the titers of bFGF were relatively lower in the RS3PE-associated solid malignancy (confirmed after 2 years from arthritis onset) than RS3PE-associated hematological malignancy (confirmed within 1 year from arthritis onset), which might partially due to the late onset of solid tumor.

Previous researches have pointed out that RS3PE might be a VEGF-associated disorder and elevated serum level of VEGF was also found in paraneoplastic RS3PE [6, 7]. However, a recent study has found serum VEGF is elevated in many elderly patients with different rheumatic diseases, indicating that VEGF may not be a marker for predicting malignancy [30]. Tomoki et al. reported high-serum MMP-3 is a characteristic of RS3PE patients with neoplasm. Nevertheless, serum levels of MMP-3 are relatively lower in our paraneoplastic RS3PE patients compared with non-paraneoplastic RS3PE patients. This difference might be partially due to different kinds of malignancy. All of the cancers in Tomoki et al.’s study are solid tumors, and they merely compared the difference of serum MMP-3 between patients with and without malignancy [11]. Most of our paraneoplastic RS3PE patents are hematological, and we utilized multiple logistic models to fully confirm the relationship between bFGF and malignancy.

Limitations

Due to the rarity of RS3PE, the number of associated malignant cases is relatively less at a single center; therefore, a prospective cohort or multi-center studies are needed to confirm the clinical significance of bFGF in further studies.

Conclusion

Our study revealed the clinical significance of serum bFGF in RS3PE; thus, bFGF might be associated with malignancy in RS3PE. Further research might verify our findings by multi-center studies and explore the prognostic value for angiogenesis cytokines.

Availability of data and materials

Dr. Hua Ye and Dr. Yuzhou Gan had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis

Abbreviations

bFGF:

Basic fibroblast growth factor

EORA:

Elderly-onset rheumatoid arthritis

Flt-1:

FMS-like tyrosine kinase 1

MMP:

Matrix metalloproteinase

OA:

Osteoarthritis

PIGF:

Placental growth factor

RF:

Rheumatoid factor

ROC:

Receiver operating characteristic

RS3PE:

Remitting seronegative symmetrical synovitis with pitting edema

TRAIL:

Tumor necrosis factor-related apoptosis-inducing ligand

VEGF:

Vascular endothelial growth factor

References

  1. Manger B, Schett G. Paraneoplastic syndromes in rheumatology. Nat Rev Rheumatol. 2014;10(11):662–70.

    Article  CAS  Google Scholar 

  2. Li H, Altman RD, Yao Q. RS3PE: clinical and research development. Curr Rheumatol Rep. 2015;17(8):49.

    Article  CAS  Google Scholar 

  3. Karmacharya P, Donato AA, Aryal MR, et al. RS3PE revisited: a systematic review and meta-analysis of 331 cases. Clin Exp Rheumatol. 2016;34(3):404–15.

    PubMed  Google Scholar 

  4. Russell EB. Remitting seronegative symmetrical synovitis with pitting edema syndrome: followup for neoplasia. J Rheumatol. 2005;32(9):1760–1.

    PubMed  Google Scholar 

  5. Kenzaka T. The Relationship between remitting seronegative symmetrical synovitis with pitting edema and vascular endothelial growth factor and matrix metalloproteinase 3. Intern Med. 2020;59(8):1021–2.

    Article  Google Scholar 

  6. Tabeya T, Sugaya T, Suzuki C, et al. A case of angioimmunoblastic T-cell lymphoma with high serum VEGF preceded by RS3PE syndrome. Mod Rheumatol. 2016;26(2):281–5.

    Article  Google Scholar 

  7. Arima K, Origuchi T, Tamai M, et al. RS3PE syndrome presenting as vascular endothelial growth factor associated disorder. Ann Rheum Dis. 2005;64(11):1653–5.

    Article  CAS  Google Scholar 

  8. Kenzaka T, Goda K. Serum matrix metalloproteinase 3 in detecting remitting seronegative symmetrical synovitis with pitting edema syndrome: A case report. World J Clin Cases. 2018;6(5):84–7.

    Article  Google Scholar 

  9. Hosoda C, Ishiguro T, Morimoto Y, et al. Remitting seronegative symmetrical synovitis with pitting edema syndrome complicated with organizing pneumonia. Intern Med. 2020;59(8):1065–9.

    Article  Google Scholar 

  10. Drago F, Ciccarese G, Agnoletti AF, et al. Remitting seronegative symmetrical synovitis with pitting edema associated with parvovirus B19 infection: two new cases and review of the comorbidities. Int J Dermatol. 2015;54(10):e389–93.

    Article  Google Scholar 

  11. Origuchi T, Arima K, Kawashiri SY, et al. High serum matrix metalloproteinase 3 is characteristic of patients with paraneoplastic remitting seronegative symmetrical synovitis with pitting edema syndrome. Mod Rheumatol. 2012;22(4):584–8.

    Article  CAS  Google Scholar 

  12. Tuncer T, Kaya A, Gulkesen A, Kal GA, Kaman D, Akgol G. Matrix metalloproteinase-3 levels in relation to disease activity and radiological progression in rheumatoid arthritis. Adv Clin Exp Med. 2019;28(5):665–70.

    Article  Google Scholar 

  13. Olivé A, del Blanco J, Pons M, Vaquero M, Tena X. The clinical spectrum of remitting seronegative symmetrical synovitis with pitting edema. The Catalán Group for the Study of RS3PE. J Rheumatol. 1997;24(2):333–6.

    PubMed  Google Scholar 

  14. Gisserot O, Crémades S, Landais C, Leyral G, Bernard P, de Jauréguiberry JP. RS3PE revealing recurrent non-Hodgkin’s lymphoma. Joint Bone Spine. 2004;71(5):424–6.

    Article  Google Scholar 

  15. Tunc SE, Arslan C, Ayvacioglu NB, Sahin M, Akkus S, Yorgancigil H. Paraneoplastic remitting seronegative symmetrical synovitis with pitting edema (RS3PE syndrome): a report of two cases and review of the literature. Rheumatol Int. 2004;24(4):234–7.

    Article  Google Scholar 

  16. Chiappetta N, Gruber B. Remitting seronegative symmetrical synovitis with pitting edema associated with acute myeloid leukemia. J Rheumatol. 2005;32(8):1613–4.

    PubMed  Google Scholar 

  17. Sibilia J, Friess S, Schaeverbeke T, et al. Remitting seronegative symmetrical synovitis with pitting edema (RS3PE): a form of paraneoplastic polyarthritis. J Rheumatol. 1999;26(1):115–20.

    CAS  PubMed  Google Scholar 

  18. Paira S, Graf C, Roverano S, Rossini J. Remitting seronegative symmetrical synovitis with pitting oedema: a study of 12 cases. Clin Rheumatol. 2002;21(2):146–9.

    Article  CAS  Google Scholar 

  19. Lakhmalla M, Dahiya DS, Kichloo A, Fatima T, Edigin E, Wani F. Remitting seronegative symmetrical synovitis with pitting edema: a review. J Investig Med. 2021;69(1):86–90.

    Article  Google Scholar 

  20. Pallotta MT, Nickel W. FGF2 and IL-1β - explorers of unconventional secretory pathways at a glance. J Cell Sci. 2020;133(21):jcs250449.

    Article  CAS  Google Scholar 

  21. Labanca E, Vazquez ES, Corn PG, et al. Fibroblast growth factors signaling in bone metastasis. Endocr Relat Cancer. 2020;27(7):R255–65.

    Article  CAS  Google Scholar 

  22. Belov AA, Mohammadi M. Molecular mechanisms of fibroblast growth factor signaling in physiology and pathology. Cold Spring Harb Perspect Biol. 2013;5(6):a015958.

    Article  Google Scholar 

  23. Song MZ, Mao YM, Wu J, Pan HF, Ye QL. Increased circulating basic fibroblast growth factor levels in acute myeloid leukemia: a meta-analysis. Hematology. 2020;25(1):186–93.

    Article  CAS  Google Scholar 

  24. Ribatti D, Vacca A, Rusnati M, Presta M. The discovery of basic fibroblast growth factor/fibroblast growth factor-2 and its role in haematological malignancies. Cytokine Growth Factor Rev. 2007;18(3-4):327–34.

    Article  CAS  Google Scholar 

  25. Akl MR, Nagpal P, Ayoub NM, et al. Molecular and clinical significance of fibroblast growth factor 2 (FGF2 /bFGF) in malignancies of solid and hematological cancers for personalized therapies. Oncotarget. 2016;7(28):44735–62.

    Article  Google Scholar 

  26. Chua V, Orloff M, Teh JL, et al. Stromal fibroblast growth factor 2 reduces the efficacy of bromodomain inhibitors in uveal melanoma. EMBO Mol Med. 2019;11(2):e9081.

    Article  Google Scholar 

  27. Nakano K, Okada Y, Saito K, Tanaka Y. Induction of RANKL expression and osteoclast maturation by the binding of fibroblast growth factor 2 to heparan sulfate proteoglycan on rheumatoid synovial fibroblasts. Arthritis Rheum. 2004;50(8):2450–8.

    Article  CAS  Google Scholar 

  28. Yan D, Chen D, Cool SM, et al. Fibroblast growth factor receptor 1 is principally responsible for fibroblast growth factor 2-induced catabolic activities in human articular chondrocytes. Arthritis Res Ther. 2011;13(4):R130.

    Article  CAS  Google Scholar 

  29. Zhao S, Wang Y, Hou L, Wang Y, Xu N, Zhang N. Pentraxin 3 inhibits fibroblast growth factor 2 induced osteoclastogenesis in rheumatoid arthritis. Biomed Pharmacother. 2020;131:110628.

    Article  CAS  Google Scholar 

  30. Smets P, Devauchelle-Pensec V, Rouzaire PO, Pereira B, Andre M, Soubrier M. Vascular endothelial growth factor levels and rheumatic diseases of the elderly. Arthritis Res Ther. 2016;18(1):283.

    Article  Google Scholar 

Download references

Acknowledgements

The authors thank Dr. Huixin Liu for the support in the statistics of the project.

Funding

This work was supported by the National Natural Science Foundation of China (No. 81801615, and 81871289) and Beijing Municipal Science & Technology Commission (No. Z191100006619109).

Author information

Authors and Affiliations

Authors

Contributions

YG: data interpretation and analysis, manuscript drafting, review, and editing. HY: methodology, review, and editing. JC, YC, and XL: clinical data collecting and follow-up of the paticipants. YS and YW: laboratory analysis. The authors read and approved the final manuscript.

Corresponding author

Correspondence to Hua Ye.

Ethics declarations

Ethics approval and consent to participate

The study was approved by the ethics committee of Peking University People’s Hospital (2020PHB060) and the study complied with the Declaration of Helsinki guidelines. All the participants were given written informed consent.

Consent for publication

Not applicable

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

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

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/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gan, Y., Sun, Y., Jin, J. et al. bFGF could be a biomarker of malignancy in RS3PE syndrome: an ambispective single-center cohort analysis of 51 patients. Arthritis Res Ther 23, 261 (2021). https://doi.org/10.1186/s13075-021-02638-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13075-021-02638-0

Keywords