Neutrophil to lymphocyte ratio at diagnosis predicts venous thrombosis in prefibrotic primary myelofibrosis: results from a multicenter cooperative study

Abstract

Primary myelofibrosis (PMF) is a myeloproliferative neoplasm (MPN) characterized by splenomegaly, symptoms, cytopenias, and chronic inflammation. PMF has two stages: pre-fibrotic (prePMF) and overt PMF. PrePMF and essential thrombocythemia share a similar high thrombotic risk, but few studies have examined thrombosis risk factors in prePMF. The neutrophil-to-lymphocyte ratio (NLR), reflecting the imbalance between systemic inflammation and immunity, has emerged as a prognostic biomarker in various diseases. We investigated the predictive value of NLR for thrombotic risk in a multicenter cohort of 225 prePMF patients enrolled in the retro-prospective observational INFLA-ME (INFLAmmation in MyeloproliferativE disease) cooperative study. After a median follow-up of 5.9 years, 37 thrombotic events occurred in 31 patients (2.5 events/100 patients/year; 18 arterial, 19 venous). Multivariate analysis linked venous thrombosis risk to prior venous events (HR 4.46, p = 0.001) and NLR ≥ 6 (HR 3.82, p = 0.008). The patients with NLR ≥ 6 showed a shorter venous thrombosis-free survival (p = 0.003). NLR value had no significant association with total and arterial thrombotic events. In conclusion, NLR is an inexpensive and accessible prognostic biomarker of venous thrombosis in prePMF. The integration of NLR into conventional risk scores may allow for better identification of pre-PMF patients requiring cytoreduction.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00277-026-06763-1.

Introduction

Prefibrotic primary myelofibrosis (prePMF) represents a distinct entity within the spectrum of myelofibrosis (PMF), characterized by the absence of pronounced fibrosis in the bone marrow and a more benign clinical course compared to overt-PMF [1].

Despite its clinical relevance, thrombotic risk in prePMF is often underestimated and overshadowed by concerns about fibrotic progression and leukemic transformation, even though it significantly affects patient management. However, the incidence of thrombotic events in prePMF is comparable to that in essential thrombocythemia (ET): in one study, 25.4% of 180 prePMF patients experienced thrombotic complications, versus 21.5% of 891 ET patients, with annual major thrombosis rates of 1.9% and 1.7%, respectively [2]. Recently, Guglielmelli et al. reported the patients with prePMF can be stratified in different risk categories for cardiovascular events using the IPSET score as developed for ET [3].

Thrombotic risk stratification in prePMF could be improved by incorporating additional variables, such as inflammatory markers. Inflammatory processes, driven by cytokine dysregulation (including interleukin-1 and tumor necrosis factor-alpha), promote the expansion of abnormal stem cells and disrupt normal hematopoiesis, fostering disease progression and fibrotic transformation. Elevated levels of inflammatory cytokines are also linked to the symptomatic burden reported by patients [4, 5]. The inflammatory bone marrow microenvironment promotes also a prothrombotic milieu, a characteristic common to prePMF and other myeloproliferative neoplasms (MPN). Elevated inflammatory cytokine levels enhance platelet activation and the release of prothrombotic mediators, triggering endothelial cell and neutrophil granulocyte activation. Increasing evidence indicates that neutrophils are not merely bystanders but active mediators of thrombosis. Through the formation of neutrophil extracellular traps (NETs), they can initiate and amplify thromboinflammatory responses, contributing to the heightened thrombotic risk observed in MPNs [6, 7].

The clinical relationship between inflammation and thrombosis has been investigated in the MPN setting [8–10]. Several studies have shown that neutrophils circulate in an activated state in MPN, where they can suppress the cytotoxic activity of T-lymphocytes and natural killer (NK) cells, reflecting impaired immunity [11]. This may promote clonal progression and immune escape of the malignant clone. Therefore, the neutrophil-to-lymphocyte ratio (NLR) could serve as a valuable clinical biomarker, providing insights beyond the simple neutrophil and lymphocyte counts.

In recent years, an increasing body of research across various fields, including oncology and non-oncology, has highlighted the predictive value of NLR. It has been strongly linked to thrombotic events in cardiovascular disease [12]. Recently, its predictive ability for thrombotic events has also been demonstrated in studies involving patients with MPN: Carobbio et al. [13] showed that higher NLR value (> 5) was associated to increase of venous thrombosis in the setting of Polycythemia Vera (PV); Ripamonti et al. [14] highlighted that patients with ET and elevated NLR (> 4) had a higher risk of developing thrombotic events. However, the prognostic role of NLR in the prefibrotic phase of PMF remains largely unexplored.

In this study, we aimed to explore the relationship between NLR evaluated at diagnosis and the risk of thrombotic events during follow-up in prePMF patients.

Materials and methods

Patients and study design

This retro-prospective, cooperative, observational study enrolled 225 consecutive patients diagnosed with prePMF, followed at eight Italian Hematology Centers between June 1994 and December 2023. The list of participating Centers is provided in Supplementary Files (SL1). This study represents a sub-analysis of the INFLA-ME (INFLAmmation in myeloproliferativE disease) trial, which received approval from the Institutional Review Board. Each Center was responsible for entering data related to the enrolled prePMF patients into an electronic case report form. Patient anonymity was ensured using alphanumeric identification codes. Demographics, histopathological, clinical, molecular and laboratory features data were recorded at multiple time points, including at initial diagnosis of prePMF, at the initiation of cytoreductive and/or antiplatelet and/or anticoagulant treatment, and at the time of any thrombotic event.

At diagnosis, the following parameters were collected: age, sex, prognostic risk category, presence of palpable splenomegaly (> 5 cm from the costal arch), type of cardiovascular risk factors (e.g., active smoking, dyslipidemia, hypertension, diabetes), history of thrombosis, type of driver mutation (JAK2, MPL, or CALR), and blood counts (including absolute white blood cell differential, haemoglobin level, and platelet count).

Treatment decisions were made at the discretion of the treating physician and were not influenced by participation in the study. Following the initial data entry, follow-up information was validated during clinical data reviews, and specific queries were sent to participating centers to resolve any data inconsistencies. All patients were followed until death or until the data cut-off (December 2023), with a median follow-up duration of 5.9 years.

Definitions

Diagnosis of prePMF was made according to the 2022 World Health Organization (WHO) criteria [15]; all previous histological diagnosis were revised accordingly at local institutions, accordingly. Histologic examination was performed at local institutions; fibrosis was graded according to the European Consensus Grading System [16]. The risk prognostic category was assessed at diagnosis using the International Prognostic Scoring System (IPSS) for survival and the IPSET-Thrombosis score for thrombotic risk assessment [17, 18].

NLR value was calculated as the ratio between absolute neutrophil count (ANC) and lymphocyte count (ALC) at the time of prePMF diagnosis, before the start of any cytoreductive treatment. Splenomegaly was defined as the presence of a spleen palpable at least 5 cm from the costal arch.

Thrombotic complications, both at diagnosis and during follow-up, were recorded when objectively documented, whether symptomatic or incidental. These included arterial events (ischemic stroke, myocardial infarction, peripheral arterial disease, and retinal artery ischemia) and venous events (superficial and deep vein thrombosis, pulmonary embolism, and cerebral or splanchnic vein thrombosis).

Ethical aspects

The INFLA-ME study was conducted in accordance with the guidelines of the institutional review boards of the participating centers and to the principles outlined in the Declaration of Helsinki. All patients provided written informed consent. The study was promoted by the Hematology Division and Bone Marrow Unit of the Fondazione IRCCS San Gerardo dei Tintori, Monza. It received approval from the Monza Ethics Committee on June 23, 2022, as well as from the local ethics committees of all participating centers subsequently. The study received no commercial funding. The authors declare no conflicts of interest. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Endpoints

The primary endpoint of the study was to evaluate the predictive value of NLR, assessed at the time of diagnosis, for thrombotic events (considered both overall and separately as arterial and venous events). The secondary endpoint of the study was to investigate the association between demographics, histopathological, clinical, molecular and laboratory variables, and the occurrence of thrombosis.

Statistical methods

Statistical analysis was performed at the Bicocca Bioinformatics Biostatistics and Bioimaging Centre at Milano-Bicocca University. Descriptive statistical analyses were employed to characterize the patient’s sample. Continuous variables were described using the median and interquartile range (IQR), while categorical variables were analyzed using absolute frequency and relative percentage (%). Incidence rates of events were expressed as events per 100 patients per year (% p-y). Comparisons between two or more groups were performed with chi-square test for categorical variables, and the Mann-Whitney test for unpaired continuous variables. To assess the association between clinical and laboratory variables collected at time of diagnosis and thrombotic outcome (thrombosis overall, arterial and venous thrombosis), univariate and multivariate analysis were conducted using the hazard-ratio (HR) based Cox Regression model. Results with p-values (p) < 0.05 were considered statistically significant and 95% confidence intervals (CI) on the HR were reported.

Thrombosis-free survival (TFS) and OS were defined as the time from prePMF diagnosis to the first thrombotic event and to death (uncensored) or last contact (censored), respectively. TFS curves (overall, arterial, or venous thrombosis) and OS were estimated by Kaplan-Meier method; the log-rank test was used to compare curves.

The NLR cut-off (≥ 6) for thrombotic outcome was identified as the minimum value of NLR that guaranteed.

an increment of 10% in the cumulative incidence probability of venous thrombosis, given that NLR was greater than the cut-point (positive predictive value), with respect to the probability of venous thrombosis in the whole sample. This increment, indeed, corresponds to a reduction of 10% in the probability of surviving free from venous thrombosis. Of note, 13% of the patients have an NLR greater that this cut-point and thus the cut-point is not extreme. All analyses were conducted using the statistical software GraphPad Prism 10.2.1 and Stata version 18.

Results

Patients’ characteristics

A total of 225 patients were enrolled into this study; 114 were male (50.7%). At diagnosis, median age was 61.7 (IQR 24.8–85.6) years. Most patients (86%) were categorized as low-intermediate 1 IPSS risk; 58.7% were high risk according to IPSET-Thrombosis score. Splenomegaly was reported in 26 (11.6%) patients.

According to driver mutational status, 170 (75.6%) patients had a JAK2 V617F mutation, 33 (14.6%) were CALR mutated, while 4 (1.8%) had a MPL mutation; 18 (8%) were “triple negative”.

Sixty-seven patients had a thrombosis before the diagnosis of prePMF (27.8%), for a total of 69 thrombotic events: 30 (44.8%) venous and 39 (58.2%) arterial thrombosis; 2 (3%) patients experienced both venous and arterial events. 113 patients had at least one cardiovascular risk factor, mainly arterial hypertension (43.5%), dyslipidemia (28.4%), active smoking (7.6%), and diabetes (8.9%).

Median NLR value at diagnosis was 3 (IQR: 0.64–41.5). The main clinical and laboratory features are resumed into Table 1.

Table 1.

Main clinical and laboratory features at diagnosis of the 225 patients enrolled into the study

Patients characteristics
Sex, N° (%)
Male	114 (50.7%)
Female	111 (49.3%)
Driver mutation, N° (%)
JAK2	170 (75.6%)
CALR	33(14.7%)
MPL	4 (1.8%)
Triple negative	18 (8%)
Age, median (IQR)	61.7 (24.8–85.6)
Fibrosis degree, N° (%)
0	54 (24%)
1	171 (76%)
IPSS, N° (%)
Low	102 (45.3%)
INT-1	91 (40.4%)
INT-2	25 (11.1%)
High	7 (3.1%)
Splenomegaly at diagnosis, N°(%)	26 (11.6%)
Thrombotic event prior to diagnosis, N°(%)	69
Arterial	30 (43.5%)
Venous	39 (56.5%)
Hemorragic events prior to diagnosis, N°(%)	12 (5.3%)
Major bleeding	6 (2.7%)
Non major clinically relevant bleeding	6 (2.7%)
Antithrombotic therapy at diagnosis	25 (11.1%)
Cardiovascular risk factors, N° (%)
Active smoking	17 (7.6%)
Dyslypidemia	64 (28.4%)
Hypertension	98 (43.6%)
Diabetes	20 (8.9%)
IPSET thrombosis score, N°(%)
Low	41 (18.2%)
Intermediate	52 (23.1%)
High	132 (58.7%)
Neutrophil count, median (IQR), (*10^9/l)	5952 (1272–29549)
Lymphocyte count, median (IQR), (*10^9/l)	1984 (172–8229)
Neutrophil-to-lymphocyte ratio, median (IQR)	3 (0.64–79.64)

Thrombotic events

Over a cumulative observation period of 1,363 patient-years, a total of 37 thrombotic events were recorded in 31 patients, corresponding to an incidence rate of 2.5 events per 100 patient-years. Of these events, 18 (48.6%) were arterial and 19 (51.4%) were venous. Six patients experienced two thrombotic events: four patients had two venous thromboses, one had two arterial events, and one patient experienced both a venous and an arterial thrombosis.

The first thrombotic event occurred, on average, 4.0 years after diagnosis (interquartile range [IQR], 0.2–12.1), with a similar time to event observed for venous (median 4.0 years, IQR 0.2–10.0) and arterial (median 4.0 years, IQR 0.2–12.1) thromboses.

Characteristics of thrombotic events were resumed into Table 2.

Table 2.

Features of the 37 thrombotic events occurred during follow-up

Thrombosis in the follow-up, N° (%)	37 (100%)
Arterial	18 (48.6%)
Venous	19 (51.4%)
Type of venous thrombosis, N° (%)	19 (100%)
Deep venous thrombosis	7 (36.8%)
Pulmonary embolism	6 (31.6%)
Superficial thrombosis	1 (5.3%)
Splanchnic thrombosis	5 (26.3%)
Type of arterial thrombosis, N° (%)	18 (100%)
Ischemic stroke	8 (44.4%)
Myocardial infarction	5 (27.8%)
Obliterant arteriopathy	5 (27.8%)

Overall thrombotic risk

In univariate analysis, the rate of thrombotic events was associated with the presence of at least one cardiovascular risk factor among those included into the IPSET Thrombosis score (p = 0.006, HR 3.20, 95% CI: 1.40–7.31), a history of a previous thrombotic events before prePMF diagnosis (p = 0.043, HR 1.95, 95% CI: 1.02–3.73), a high score IPSET Thrombosis score (p = 0.004, HR 3.03, 95% CI: 1.43–6.43) and the presence of a JAK2 V617F mutation (p = 0.032, HR 4.77, 95% CI: 1.14–19.85).

The rate of thrombotic events was not associated to NLR value, considered as continuous variable. In multivariate analysis, only the presence of at least one cardiovascular risk factor was associated with a higher thrombotic risk (p = 0.004, HR 3.35, 95% CI: 1.47–7.67) (Table 1S).

Arterial thrombotic risk

A total of 18 arterial thrombotic events were observed in 17 patients, with one patient experiencing two arterial events, corresponding to an incidence rate of 1.25 events per 100 patient-years. In univariate analysis, the occurrence of arterial thrombosis was significantly associated with a prior history of arterial thrombotic events (p = 0.013; HR 3.50, 95% CI: 1.20–9.42) and the presence of at least one cardiovascular risk factor (p = 0.013; HR 13.09, 95% CI: 1.74–98.75). Specifically, active smoking (p = 0.008), dyslipidemia (p = 0.004), and hypertension (p = 0.004) were individually associated with increased risk.

The neutrophil-to-lymphocyte ratio (NLR) was not associated with a higher incidence of arterial thrombosis, whether analyzed as a continuous or dichotomous variable.

In multivariate analysis, active smoking (p = 0.001, HR 3.15, 95% CI: 1.57–6.30) and the presence of arterial hypertension (p = 0.000, HR 5.67, 95% CI: 2.32–13.86) remained independent factors associated with a higher rate of arterial thrombotic events. (Table 2S)

Venous thrombotic risk

A total of 19 venous thrombotic events was reported in 15 patients (4 patient had a dual venous event), for an incidence rate of 1.38 events/100 patients/year. In univariate analysis, previous venous thrombosis was associated with a significant higher rate of thrombotic venous events during follow-up (p = 0.001, HR 4.58, 95% CI: 1.84–11.39), as well as high IPSET score (p = 0.025, HR 3.53, 95% CI: 1.17–10.65); patients with an Hb level ≥ 10 g/dl at diagnosis showed a lower risk of developing venous thrombosis (p = 0.014, HR 0.21, 95% CI: 0.06–0.73). The rate of thrombotic events was not associated to NLR value, considered as continuous variable. Therefore, patients with NLR value ≥ 6 at diagnosis showed an increased rate of developing venous thrombotic events (p = 0.006, HR 3.95, 95% CI: 1.48–10.55) compared to patients with NLR < 6.

In multivariate analysis, we confirmed a significant association between rate of venous thrombosis and NLR value ≥ 6 (p = 0.008, HR 3.82, 95% CI: 1.42–10.29) as well as a history of venous thrombosis before the diagnosis (p = 0.001, HR 4.46, 95% CI: 1.79–11.11). (Table 3). In a similar way, we documented the patients with NLR value ≥ 6 had a significantly lower venous TFS compared to those with an NLR < 6 (p = 0.003) (Fig. 1).

Table 3.

Univariate and multivariate analysis on venous thrombotic risk – variables at diagnosis (HR: hazard ratio; hb: haemoglobin; WBC: white blood cells)

Univariate and multivariate analysis: venous thrombotic events
	Univariable analysis			Multivariable analysis
Variable	HR	p-value	95% CI	HR, p value- 95% CI
NLR (continuous variable)	1.02	0.157	0.99–1.06
NLR ≥ 6	3.95	0.006	1,48 − 10,55	HR 3.82, p = 0.008, (1.42–10.29)
Leukocytosis (WBC > 9000*10^9/l)	0.63	0.324	0.25–1.59
Neutrophils > 10,000*10^9/l	2.05	0.170	0.74–5.69
Lymphocytes < 2000*10^9/l	1.75	0.240	0.69–4.46
Sex (female)	0.64	0.336	0.26–1.59
Age	0.99	0.667	0.96–1.03
IPSS category rated HIGH/INTERMEDIATE-2	1.17	0.835	0.27–5.11
Thrombotic event prior to diagnosis	2.06	0.117	0.83–5.07
Venous thrombotic event prior to diagnosis	4.58	0.001	1.84–11.39	HR 4.46, p = 0.001, (1.79–11.11)
Arterial thrombotic event prior to diagnosis	0.57	0.450	0.13–2.47
At least one cardiovascular risk factor among those included into IPSET Thrombosis score (active smoking, hypertension, diabetes)	1.59	0.351	0.60–4.18
IPSET thrombosis score rated “HIGH”	3.53	0.025	1.17–10.65
Presence of JAK2 V617F mutation	-	1.000	0.00 -.
Hb value at diagnosis (g/dl) (≥ 10 g/dl)	0.21	0.014	0.06–0.73

Fig. 1.

Venous TFS (thrombosis-free-survival) according with NLR value (≥ 6 or < 6) at diagnosis

Discussion

The balance between neutrophils and lymphocytes plays a crucial role in maintaining immune homeostasis. Deviations from this equilibrium, as indicated by an elevated NLR value, have been associated with systemic inflammation [19]. Nowadays, the close interplay between inflammation and thrombosis is well recognized—so much so that the term “thromboinflammation” has been coined to describe the convergence of inflammatory and thrombotic processes, with thrombosis often emerging as an epiphenomenon of inflammation [20, 21].

As a marker of systemic inflammation, NLR may serve as a valuable prognostic indicator of vascular outcomes in patients with MPN. Recently, its prognostic significance has begun to receive attention in the MPN setting, particularly in patients with PV and ET [13, 14, 22].

In the setting of MF, the prognostic role of NLR remains largely unexplored. Laganà et al. evaluated the prognostic role of NLR value in a cohort of 140 patients with MF treated with Ruxolitinib [23]. Contrary to expectations, a low NLR value (< 2) was associated with a decreased survival and a higher rate of Ruxolitinib discontinuation in this population. However, in this case, the NLR value was not assessed at diagnosis but rather prior to the initiation of JAK inhibitor therapy, in a population already receiving cytoreductive treatment, a fact which that potentially influence NLR values. Additionally, the NLR cut-off used in this study is relatively low and closely approximates the median value reported in the general population by recent studies [24].

In this multicentre retrospective series of patients with prePMF, we aimed to investigate, for the first time, whether the NLR at diagnosis could serve as a reliable marker for thrombosis risk. Our findings indicate that patients with higher NLR values at diagnosis have a significantly increased risk of developing venous thrombosis.

Interestingly, while it is intuitive to consider systemic inflammation—of which NLR is a marker—as a driver of disease progression toward fibrosis, it is less straightforward to explain why it is specifically associated with an increased risk of venous thrombosis, but not with arterial or overall thrombotic risk.

It is important to remember that while the pathogenesis of venous thrombotic processes remains firmly rooted in the historically identified principles by Virchow, including venous stasis, hypercoagulability, and endothelial damage, facilitated by systemic inflammation, arterial thrombotic events mainly depend on atherosclerotic processes and by the development of plaques within arterial vessels [25]. Although atherosclerosis involves a chronic inflammatory process, its development is primarily driven by factors such as dyslipidemia and common cardiovascular risk factors, including smoking [26]. This could explain our differing results, where NLR was predictive of venous thrombosis but not arterial thrombosis, with arterial events being associated with the presence of ”classical” cardiovascular risk factors.

It is natural to question whether the predictive power of the NLR is primarily driven by the predictive strength of one of its two components—namely, the lymphocyte or neutrophil counts. In fact, it’s well established that elevated neutrophil counts can predict a higher incidence of venous thrombotic events in patients with MPNs, and similarly, leukocytosis is recognized as one of the major risk factors for thrombosis development. However, in our study, neither neutrophil nor lymphocyte counts demonstrated predictive value for recurrent thrombosis, whether analyzed as continuous variables or using arbitrary cut-offs derived from existing literature. The higher NLR observed in prePMF compared with other MPN subtypes may reflect the distinct inflammatory profile of this condition, characterized by elevated circulating cytokines, enhanced neutrophil activation, and a dysregulated immune environment. This inflammatory state may not only underlie the elevated NLR but also contribute to the increased thrombotic risk described in prePMF.

The primary limitation of this study lies in its partially retrospective design. It was not possible to completely eliminate potential selection biases, particularly due to the possible under-recognition of unidentified factors influencing thrombotic outcomes—especially in patients diagnosed before 2000 and in deceased individuals.

Among the study’s limitations, we acknowledge the inability to exclude the potential impact of ongoing antithrombotic therapy at the time of MPN diagnosis on thrombotic risk, as well as the lack of a centralized review of histopathological specimens, which may represent an additional limitation.

Nevertheless, the large cohort of patients, drawn from specialized hematology centers with expertise in myeloproliferative disorders, and equipped with highly experienced pathology services in the diagnosis of MPNs according to current WHO criteria, combined with a thorough review of each patient’s medical history and retrieval of missing data through electronic health records, likely mitigates these inherent limitations.

Lastly, recent data from the Spanish ET registry indicate that the IPSET-thrombosis and its revised version are able to predict arterial thrombosis but fail to stratify the risk of venous thrombosis. Therefore, the integration of NLR and IPSET-thrombosis might in the future enable the development of a more accurate predictive score for overall thrombotic risk in MPN patients, although prospective studies will be necessary to validate this hypothesis [27].

Conclusions

This multicentric observational study is the first to explore whether the NLR value at diagnosis can serve as an independent predictive marker for venous thrombosis in the context of prePMF. Patients with an NLR value greater than 6 showed a significantly higher risk of venous thrombosis and a shorter venous TFS. Our findings suggest that NLR assessment is a simple and cost-effective inflammatory marker that could be incorporated into conventional cardiovascular risk scores to more accurately identify patients who may benefit from cytoreductive therapy. While these results are promising, further validation through larger, prospective studies is needed.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

F. Cavalca: Conceptualization; Investigation; Writing - original draft; Writing - review & editing; Data curation; Resources; and Visualization. R. Latagliata: Conceptualization; Investigation; Writing - review & editing; Resources; and Data curation. N. Pugliese, G.A Palumbo, N. Polverelli, G. Benevolo, P. Musto F. Branzanti, A. Di Veroli, E. Accorsi Buttini, A. Ripamonti, I. Civettini, : Investigation; Writing - review & editing; and Resources. L. Montelisciani and L. Antolini: Writing - original draft; Writing - review & editing; Data curation; Formal analysis; and Visualization. C. Gambacorti-Passerini: Investigation; Writing - review & editing; Funding acquisition; and Resources; F. Palandri: Conceptualization, Investigation; Writing - original draft; Writing - review & editing; Resources; and Data curation. E.M. Elli: Conceptualization; Investigation; Writing - original draft; Writing - review & editing; Data curation; Resources; and Visualization.

Funding

No funds, grants, or other support was received.

Data availability

The data that support the findings of this study are available from the corresponding author Fabrizio Cavalca upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.