Abstract
Background/Aim: Diffuse large B-cell lymphoma (DLBCL) is the most common form of non-Hodgkin lymphoma. A systematic review to evaluate the association between Epstein-Barr Virus (EBV) and programmed death ligand-1 (PD-L1) in DLBCL biopsy was conducted. Materials and Methods: Only studies comparing EBV+ and EBV− groups were eligible following database search. Prevalence ratios were calculated for results comparison. The EBV impact on PD-L1 positivity in tumour cells and its microenvironment was analysed. Results: With 270 records screened, eleven cross-sectional studies were identified for final review. Eight studies investigated PD-L1 expression in tumour cells and found an EBV trend unlikely, while four studies found an increase in its expression in the tumour microenvironment. Nine studies showed that EBV+ cases were more commonly of non-germinal centre B-cell origin. Four studies examined genetic aberrations, but no definite consensus was reached. Conclusion: A non-EBV related mechanism is likely related to increased PD-L1 expression, with relevance to the cell of origin.
- EBV
- diffuse large b-cell lymphoma
- Epstein-Barr virus positive diffuse large b-cell lymphoma
- programmed death-ligand 1
- tumour immunology
- non-Hodgkin lymphoma
- review
Diffuse large B-cell lymphoma (DLBCL) is the most common form of Non-Hodgkin Lymphoma (NHL) and represents 25-45% of NHL globally (1). With annual incidence of 5-7 per 100,000 people, age is a significant factor as incidence rises to around 30 per 100,000 in those aged 65-84 years (2). DLBCL, not otherwise specified (DLBCL, NOS) accounts for 80-85% of all DLBCL cases and its diagnosis is made via exclusion of DLBCL subtypes with a distinct morphology or immunophenotype (3). Herein, DLBCL and DLBCL, NOS are interchangeable terms in literature unless a distinct subtype of DLBCL is specified.
DLBCL typically arises at a lymph node, although extra-nodal manifestations are also present (4). Its aetiology remains poorly understood, but changes in gene expression and mutations which promote malignant behaviour have been noted (5). DLBCL is diagnosed via biopsy, where tumours derived from germinal centre B-cells (GCB) have a better prognosis than those of non-GCB (N-GCB) subtype, also known as Activated B-cells (ABC) (6). Despite advancements, approximately 40% of patients relapse or are refractory (7) to standard R-CHOP treatment (Rituximab, cyclophosphamide, hydroxydoxorubicin, oncovin and prednisolone/prednisone), where the presence of Epstein-Barr Virus (EBV) appears to coincide with more severe outcomes (8).
DLBCL has undergone a major revision in the World Health Organisation (WHO) 2016 classification of lymphoid neoplasms, including renaming of “EBV-positive DLBCL of the elderly” subtype to “EBV-positive DLBCL, NOS” as younger individuals can also be affected (9). Approximately 90-95% of the world's population sustains a life-long, asymptomatic infection with EBV through saliva (10), following a primary lytic infection, where the virus avoids immune response by acquiring various latency types (11). About 5-15% of all DLBCL cases are EBV positive (EBV+) (12) and higher incidence rates are found in developing countries (13) where both immunocompetent and immunocompromised patients are affected (14).
EBV association has been increasingly analysed in articles exploring programmed cell death protein-1 (PD-1) (2q37.3 locus) and programmed death-ligand 1 (PD-L1) (9p24.1 locus). PD-1 is normally expressed on the surface of immune cells and it has the ability to negatively regulate the immune response. In cancerous state, PD-L1 present on a cancer cell binds to PD-1, thus reducing T-cell function and preventing immune response (15) leading to immune evasion. PD-1 blockade has been established as a therapy for some cancers (16), and PD-L1 has been found to be over-expressed in various DLBCL tumours, prolonging tumour progression and survival (17). The precise mechanism(s) involved in EBV− driven DLBCL carcinogenesis remain unknown (18) and up to date EBV trends related to DLBCL PD-L1 expression have not been summarised. Accordingly, a systematic review was carried out to evaluate available literature to determine whether EBV positivity has an impact on PD-L1 expression in DLBCL tumour biopsy.
Materials and Methods
This systematic review was registered with PROSPERO, an international prospective register of systematic reviews (registration ID: CRD42020183091).
Search strategy. A search of the published literature was carried out on 25th of April 2020 using Embase, LILACS, Web of Science, CINAHL, Ovid, National Cancer Institute, Cochrane Central Register of Controlled Trials (CENTRAL), EU Clinical Trials Register and ClinicalTrials.gov databases. The search terms relating to EBV (“EBV” OR “Epstein Barr Virus” OR “HHV4” OR “EBER” OR “LMP” or “EBNA”), DLBCL (“lymphoma” OR “b-cell lymphoma” OR “DLBCL” OR “diffuse large b-cell lymphoma”) and PD-L1 (“PD-L1” OR “PDL1” OR “CD274” OR “programmed death-ligand 1” or OR “b7-h1” OR “PDCD1LG1”) were adapted to each database, and limited to publication date of 2015 to present. In addition, article reference lists were reviewed to ensure thorough search of the literature.
Eligibility criteria. The EPICOT framework (evidence, population, intervention, comparison, outcome, timestamp) (19), was used to develop the research question and to formulate inclusion criteria to select appropriate studies. Studies were selected if they met the following requirements: 1) Discuss EBV+DLBCL and/or their extranodal manifestations, include DLBCL classified within lymphoproliferative disorder, post-transplant or immunocompromised patients. Provide analysis relevant to PD-L1 expression or therapy, whether in the tumour itself or in the tumour microenvironment; 2) Studies including minimum of ≥10 participants, 5 of which were in the EBV+ DLBCL subgroup and 5 in the control group; 3) Articles discussing tumour biopsy, whether archival or with sample obtained during study. Specific treatment(s) not required, but any intervention or study looking at genetic alterations, molecular application or drug therapies which affect PD-L1 expression to be noted; 4) Have a control group of EBV- DLBCL patients; 5) Describe EBV status identification or confirmation, accept expression and/or genetic identification of EBER, EBNA or LMP. PD-L1 identification explicitly stated and justification for positivity status provided. Offer a comparison between EBV+ and EBV− subgroups; 6) Papers with publication date 2015 onwards, with first author known, full-text article available and any conflict of interest is explicitly stated. Exclusion criteria included the following: 1) Papers discussing other distinct DLBCL subtypes or distinct clinical issues including chronic inflammation associated DLBCL or primary lymphoma of the central nervous system, addressing EBV-related B-cell neoplasms classed as: Burkitt, Hodgkin, plasmablastic or primary effusion lymphomas, discussing “gray zone” lymphomas, T-cell lymphomas or leukaemias. Papers discussing PD-1 expression only, without addressing PD-L1. Any data relating to cell culture, animal studies, case studies, editorials, abstracts; 2) Papers including patients who's EBV status was not known or where data was mixed with other population subgroups where DLBCL relevant data cannot be extracted; 3) Research describing the already established standard R-CHOP treatment for DLBCL, unless new context is added with relevance to effect on PD-L1 expression; 4) Presence of EBV+ or EBV− DLBCL group only; 5) EBV or PD-L1 Identification methods or positivity criteria not stated; 6) The articles were in non-English language.
Study selection and data extraction. Two independent reviewers were involved in the study selection and any disagreement was resolved through conference with a 3rd reviewer. The results were recorded in a PRISMA flow diagram (20), alongside a summary of reasons for exclusion.
An Excel data extraction tool was devised to facilitate comparisons and to aid data collection consistency. The following items were extracted: study design, ethical approval, sample type, sample size, methods of identification and measurement of EBV and PD-L1 including dilution factor and antibody source, positivity thresholds, expression values and supporting statements, significance values and statistical analysis methods. Obtained population characteristics included: age, gender, cell of origin, Lymphoma International Prognostic Index (IPI) and chemotherapy treatment.
Quality assessment. The critical Appraisal Tool for Cross-sectional Studies (AXIS) (21) was used to assess the quality of the included studies by the two reviewers. Each item was graded as “low risk” or “high risk”, while insufficient information was classified as “uncertain”.
Data synthesis. A narrative synthesis was carried out using tables summarising study characteristics and results of individual studies. The principal summary measures included data relevant to PD-L1 expression in tumour and tumour microenvironment alongside data related to genetic alterations. Where applicable, prevalence ratio (PR) was calculated with 95% confidence intervals to give general commonality to results, particularly where statistical analysis was not present, in order to evaluate EBV association with PD-L1 expression in DLBCL.
Results
Study selection. The database search identified 345 records for screening alongside 21 additional records (Figure 1). Following duplicate removal, 270 studies underwent title and abstract screening, where 105 records were excluded. A total of 165 studies underwent full text screening and eleven studies met the inclusion criteria for review.
Characteristics of included studies. The characteristics of the included studies are summarised in Table I. All eleven articles (22-32) were cross-sectional studies published between 2015 and 2019. Four articles originated from Japan (24, 28-30), two from China (23, 31), and five articles originated from each of the following countries: United States (22), Australia (25), Sweden (26), Spain (27), and South Korea (32). Although all papers addressed DLBCL, there was a variation in its assortment, with two papers referring to ‘DLBCL’ only (22, 24) while others further specified the DLBCL sample type. Two papers (31, 32) referred to ‘DLBCL of the elderly’ due to publication date of 2015, which occurred before the WHO 2016 classification update. All studies used existing records to retrospectively select DLBCL participants. Out of 2,277 cases described with patient characteristics, 2,044 samples were PD-L1 tested including 272 EBV+ and 1,772 EBV− cases. Sample size varied between 27-1181 participants. Ten studies addressed EBV+ cohort with n=7-30 patients, and one had an n=90 cohort (30). The EBV− cohort ranged from n=16-1091, with two studies having an n>100 control cohort (25, 28).
Patient characteristics were described to varied degree, and most were not equally distributed within and between studies. Three articles (25, 28-29) addressed all relevant characteristics relating to age, gender, cell of origin, IPI and chemotherapy treatment, none were provided within one study (24), and in one (30) the characteristics were mixed with other lymphomas, hence, were not included in this review. When cell of origin proportions were combined from nine studies addressing this parameter (22-23, 25-29, 31, 32) where 850 samples were addressed, the overall proportion for EBV+ cases was 0.19 GCB and 0.81 N-GCB, while EBV− cases was 0.50 GCB and 0.50 N-GCB. Seven studies addressed age (23, 25, 27-29, 31, 32), and the median age for EBV+ groups ranged from 56 to 74 while for EBV− cases it ranged from 54 to 67; patients with an age of <60 years were included in four studies (25, 28-29, 31). Gender was addressed by six studies (23, 25, 28-29, 31, 32) where 805 samples were distinguished (466 males and 339 females), with overall male to female ratio (M:F) of 1.37. Of the six studies addressing IPI (23, 25, 28-29, 31, 32), EBV+ cases were of higher proportion in four out of six studies (25, 29, 31, 32) looking at IPI 3-5, with one study noting marginal difference in proportion between EBV+ and EBV− cases (28). Four studies specified chemotherapy approach (27, 29-31) relating to the patient characteristics and therapies were predominantly of Rituximab-containing nature.
EBV-encoded small RNAs (EBER) in situ hybridization (ISH) was applied in eleven studies, while two also applied genetic analysis (24-25). Overall, seven studies examined PD-L1+ expression via immunohistochemistry (IHC) staining (22-23, 26-30), four papers applied genetic analysis (24-25, 27, 32) and one utilised flow cytometry (31). A combination of IHC and genetic analysis was used in one study (27).
Risk of bias assessment. Quality assessment was based on the AXIS tool (21) with exclusion of questions 3, 13 and 14 which did not apply to the participants as non-responders were not associated with any of the studies.
The quality and clarity of reporting relating to various parameters of AXIS including design, measures and conclusions were uniform across multiple studies. Study design was appropriate in all articles, however, due to relative rarity of EBV+ DLBCL, sample size justification was not provided in seven studies (22, 24, 27-29, 31, 32), while four stated small sample size (23, 25-26, 30). All studies defined their population and had an appropriate sample frame. One study did not address its aims/objectives (23) while the participant selection process was not explained in sufficient detail in three studies (22, 31, 32), omitting details such as dates of retrospective record inclusion. Appropriate measures were incorporated in all studies and they were all measured correctly, with adequate data and methods, whether in main text or supplementary data. The statistical significance criterion was not applied to two studies (26, 30) as statistical analysis relevant to DLBCL was mixed with other lymphoma types, while another two (25, 27) did not address the reasons behind missing results. Study limitations were not addressed in three studies (24, 31, 32), whereas only three stated no funding or conflict of interest (28-30). Overall, all studies included ethical approval and were judged as low risk for selective outcome reporting as links between funding sources and flaws in reporting were not identified.
Results of individual studies. Table of results was compiled before undertaking thematic analysis. The analysis incorporated data relevant to PD-L1 expression in tumour cell, tumour microenvironment and genetic alterations. Alongside this, a trend regarding cell of origin type in EBV+ and EBV− cases was investigated.
EBV status and PD-L1 positivity
Tumour cells. Eight studies addressed PD-L1 expression on tumour cells (22-23, 26-31) (Table II) where two provided statistical analysis (23, 29). Seven articles (22-23, 26-27, 29-31) found that EBV+ showed an increase in PD-L1, and one detected no PD-L1 expression (28). Importantly, in six papers where PR was applied, only three had PR>1 with 95%CI (22, 27, 30), while three papers included PR=1 in CI (26, 29, 31).
Four articles (26-29) utilised 5% threshold for positivity, and three found an increase in PD-L1+ on EBV+ biopsies. One paper provided p-value: 0.072 (29), although this included PR=1 and a statement that PD-L1 and EBV were independent factors. Two other papers (26, 27) also showed PR>1 where one included PR=1 in CI (29), while another found that PD-L1 was not expressed on any EBV+/− tumour cells (28). Another study (30) utilised percentage expression with >30% threshold and found that PD-L1+ was expressed in EBV+ cases with PR 1.75 (1.04, 2.93).
Two articles (22, 23) utilised a modified score based on expression values, where IHC score (23) showed higher detection in EBV+ cases with a statically significant value. The other papers (22) identified PR 1.45 (1.05, 2.02) for EBV+ irrespective of the staining score, where EBV latency had higher %PD-L1+ expression and higher staining intensity in EBNA2+ cases, p-values: 0.0125 and 0.004, respectively. Using flow cytometry, another study (31) stated that PD-L1+ expression was higher in EBV+ cases, PR 1.78 (0.88, 3.64), but specified EBV and cell of origin as separate factors.
Tumour microenvironment. Five papers examined PD-L1 expression in relation to the tumour microenvironment (26-30) and four (27-30) found an EBV association with PR>1, whilst one paper (31) identified no PD-L1 positive cases in the microenvironment of either subgroup (Table III).
All papers adopted a 20% threshold for positivity. Of those, four analysed total tissue cellularity (26, 28-30), where two which did not specify the cellularity further found an association in EBV+ cases, at p-values of 0.09 (29) and 0.006 (28), respectively. One paper further specified the total tissue cellularity to non-malignant cells only (30), finding PR 3.16 (2.29, 4.37) for EBV+ cases. This article considered samples which were already established as PD-L1- in tumour cells, an approach also undertaken by another study (26). However, the latter identified no patients with PD-L1 positivity, and also commented on the lack of association between PD-L1 and intratumoural regulatory T-cells irrespective of the EBV status.
One paper (30) remarked on a significantly higher rate of PD-L1 detection in microenvironmental immune cells without further evidence. Another study (27) examined this in relation to tumour associated macrophages (TAM), noting an EBV+ related increase in PD-L1+ at PR 1.90 (1.16, 3.14).
Cell of origin subtype. Overall, nine studies characterised cell of origin (22-23, 25-29, 31, 32) (Table IV). Six papers (22-23, 26-29) noted EBV+ N-GCB majority, one study found N-GCB was higher in EBV− cases (25), one did not identify any cases with GCB origin (32), and one did not report EBV+ cell origin (31). Five papers reported origin without further analysis (23, 25-26, 28, 32) and among those, three reported ≥90% of N-GCB in EBV+ cases (23, 26, 32), whilst another had a 67% majority (28). There was only one study where N-GCB was a minority at 31% (25).
In four papers reporting origin in detail (22, 27, 29, 31), two found that EBV+ DLBCL cases were more commonly N-GCB, at 94% with p=0.001 (27) and 90%, respectively (29). EBV+ cases were exclusively N-GCB in one study (22), showing increased PD-L1+ cells (91% vs. 28%). The fourth paper (31) reported that PD-L1 expression appeared more commonly in N-GCB subtype and EBV+ cases, but EBV+ cell of origin was not reported. When the overall cell origin proportions were combined, EBV+ cases were 0.19:0.81 GCB:N-GCB, while EBV− cases were 0.5:0.5 GCB:N-GCB.
In relation to PD-L1, five papers (22-23, 26-27, 29) reported an increase in PD-L1 in tumour cell when EBV+ N-GCB cases were greater in number, although two of those included PR=1 (26, 29). One of the two papers (29) showed PR>1 increase in environment, consistent with two other papers reporting on microenvironment and cell type (27, 28).
EBV status and PD-L1 genetic aberrations. Four studies (24-25, 27, 32) examined genetic aberrations relating to PD-L1 and 9q24.1 (Table V). All studies noted an increase, but only two concluded that EBV+ harbours significant change (24, 25). One paper (25) showed an EBV-related increase in frequency of PD-L1 gene count, where latent membrane protein 1 (LMP1) showed significant correlations. Another paper (24) described an increase in frequency of PD-L1/PD-L2 structural variation (SV) in EBV+ cases, although SV type was not specified. Interestingly, it noted that EBV+ cases did not show differences in other genetic aberrations irrespective of the PD-L1/PD-L2 alteration, while another study (32) stated that other genetic aberrations where lower in EBV+ cases.
In contrast, despite noting that PD-L1 copy number alterations (CNAs) were detected in more EBV+ cases, another study (27) found PD-L1 CNAs in variable proportion of cells regardless of EBV status. In addition, no significant difference was found between disomies of EBV+/− cases, although it was noted that with high percentage of residual disomies, EBV may play a role in PD-L1 up-regulation in post-transplant lymphoproliferative disorder (PTLD). One study (32) found a 9q24.1 copy number gain in EBV+, where it was reported that the PD-L2 gene (PDCD1LG2) located on 9q24.1 was also overexpressed in EBV+ cases.
Discussion
The systematic review included eleven cross-sectional studies and to the authors' knowledge, this was the first systematic review which aimed to determine whether EBV impacts PD-L1 expression in DLBCL tumour biopsy. There was considerable variability in the methodologies of published literature and although staining thresholds seemed more consistent, it is apparent that investigation of PD-L1 expression requires a degree of standardisation to distinguish definite DLBCL trends. However, the review was able to identify several research avenues to explore.
EBV+ is unlikely to have an impact on PD-L1+ in tumour cells as despite seven out of eight papers finding an apparent increase (22-23, 27-29, 31, 32), PR overlap at 95%CI was noted. Particularly in cases where 5% threshold was applied giving commonality to results (26-29), values indicated that there may be no difference as some PR>1 had CI close to 1.0. Importantly, as this is an arbitrary cut-off value where higher thresholds may show definite trends (34), and as EBV-related PD-L1 expression has been identified in epithelial malignancies (35), this result should not be negated, and should be further investigated using higher positivity thresholds. Due to small sample sizes for EBV+ cohorts, it is likely that patients experience PD-L1 related changes, however, this does not appear to be an exclusive feature of EBV+ patients. As such, there is a need to identify the non-EBV mechanism of this phenomenon, although EBV latency may be a contributing factor in disease progression.
One of the most interesting trends indicated that EBV+ is associated with increased PD-L1 expression in tumour microenvironment. Although taken with caution due to limited number of studies, the PR values suggest that this may be a common finding (36). However, as the 20% threshold was applied to various parameters of the total tissue cellularity, it is possible that this result arose due to selection or confirmation bias (34). Moreover, the use of different antibodies and dilution factors could account for the observed patterns due to false-positives as selected dilutions were not justified (37). Taking into consideration that EBV relationship to PD-L1 in tumour cell is unlikely, it is apparent that selected patients might benefit from targeted anti-PD-1 therapies as the 20% threshold relating PD-L1 to tumour microenvironment appears to show a more predominant EBV link to PD-L1 expression. However, as this feature is not solely found in EBV+ patients, this finding further supports the notion that a non-EBV related mechanism operates leading to PD-L1 increase.
It is well understood that N-GCB DLBCL is associated with worse outcomes (38). A pattern regarding PD-L1 increase in N-GCB was found across several studies (39) suggesting that those of N-GCB cell type may benefit from introduction of PD-L1 therapy irrespective of EBV status. However, EBV has been linked to increased pSTAT3 expression in N-GCB (40), highlighting that this behaviour is likely cell-type dependant, where EBV+ may have an enhancing effect on PD-L1. EBV has been stipulated to up-regulate PD-L1 via a direct effect on JUN-B signalling cascade via LMP-1 (41-43) or via indirect up-regulation of inflammatory cytokines (44-46), with effects on JAK/STAT signalling pathway in Hodgkin and NHL (47). As five articles in this review related N-GCB to an increase in PD-L1 expression (22-23, 26-27, 29), although with variable PR, this result indicates that EBV+ patients may have greater likelihood of PD-L1+ and may need treatment revaluation. This includes possible addition of anti-PD-L1 medication such as Pembrolizumab (48) to selected patients considering that various trials involving R-CHOP combinations have not yielded positive effects (49). The analysis of cell type with focus on N-GCB and its impact on PD-L1 should thus be prioritised in future research to enable appropriate immunotherapeutic strategy and risk stratification in DLBCL. Moreover, more studies should utilise genetic approaches in addition to methods such as IHC, thus enabling the assessment of whether PD-L1+ and EBV+ are independent prognostic factors considering the direct and indirect involvement of EBV, particularly in relation to N-GCB potency and worse prognostic outcomes. Regarding patient characteristics, it has been reassuring to see inclusion of younger patients following the WHO classification update, where aside from N-GCB, no other correlations were identified.
Considering genetic aberrations, findings similar to those identified in this review were noted in PTLD-DLBCL (50). However, due to mixed methods and commentary, no consensus can be appreciated despite PD-L1 aberrations appearing to increase in EBV+. Various studies link 9p24.1 locus alteration (43, 47, 51) and the JAK/STAT pathway (42, 52) as a contributing factor to PD-L1 overexpression irrespective of the EBV status, as identified in this review. This supports the notion that EBV is not a sole deterministic factor (53), but likely an enhancer of PD-L1. As such, it is important to determine the level of contribution of EBV to PD-L1 genetic aberrations and JAK/STAT through the indirect mechanisms. Of note, EBV+ may also explain the relatively small gain of genetic aberrations identified in DLBCL in some studies (54), as often EBV status is not addressed and such cases may in fact be EBV+, affecting PD-L1 expression via the direct pathway.
Limitations. The review included a comprehensive search of literature, where stringent inclusion criteria ensured that studies which described relevant methodology in sufficient detail were included in the synthesis. However, a substantial amount of literature did not describe methodology to sufficient standard and multiple results were presented in conference abstracts. As these articles were not suitable for inclusion in this review and only English language papers were included, publication bias may have been introduced. Importantly, EBER status was considered sufficient for EBV identification irrespective of threshold, however experts agree that >50% of malignant cells should be EBER positive (55) and this could have impacted the data. Similarly, thresholds for PD-L1+ and mPD-L1+ need standardisation, as overall meta-analysis was not possible due to diversity of PD-L1 and EBV measures. Despite application of PR and CI to provide commonality to data comparison, variability in thresholds meant that overall, the predictive capability of this parameter was less reliable.
It is important to note that gene expression does not equate to molecular expression and studies where staining was not used to clarify results offer limited picture into impact of a genetic aberration. Furthermore, due to relative rarity of EBV+ DLBCL, sample sizes were not justified in multiple studies and this could have caused data inaccuracies.
Implications for future research. One of the most important factors identified from this review is the need to standardise reporting and thresholds, perhaps at higher levels (34), to enable meta-analysis of identified trends. More studies should utilise a dual approach with staining and genetics to clarify whether EBV-related changes affect PD-L1 expression and this must be related to cell of origin type, which appears to be more indicative of PD-L1 relationship.
Conclusion
With increased number of EBV+ cases identified as N-GCB, this systematic review indicates that cell of origin is of relevance to DLBCL, where a link or enhancement of PD-L1 pathway via EBV may be of significance. EBV+ does not seem to be the sole factor determining PD-L1+ despite noting increase in PD-L1 in tumour microenvironment as EBV− patients also show PD-L1+, and it is likely that another mechanism is involved. Although this needs to be further explored due to small sample size, it is apparent that the protocols require standardisation to confirm trends through large-scale studies before any anti-PD-L1 treatment application is considered.
Footnotes
Authors' Contributions
GA Barzyk (study concept and design, data collection, data analysis and interpretation, manuscript preparation and final approval), V Sheriff (study design, data collection, data analysis and interpretation, manuscript preparation and final approval).
This article is freely accessible online.
Conflicts of Interest
The Authors declare no potential conflicts of interest. The review has no funding source.
- Received August 5, 2020.
- Revision received August 24, 2020.
- Accepted August 25, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved