Flow cytometry for the assessment of blood tumour burden in cutaneous T-cell lymphoma: towards a standardized approach.

Mycosis fungoides (MF) and Sézary syndrome (SS) are the best-studied subtypes of cutaneous T-cell lymphoma, a rare non-Hodgkin lymphoma that primarily presents in the skin but can also involve blood, lymph nodes and viscera. The role of blood involvement in the assessment and staging of MF and SS has evolved in recent years from being classed as simply 'present' or 'absent', with no impact on staging, to full analysis of abnormal peripheral blood T cells using flow cytometry (FC) to detect and quantify aberrant T-cell phenotypes and polymerase chain reaction (PCR) to characterize T-cell receptor gene rearrangements. These sensitive peripheral blood assessments are replacing manual Sézary cell counts and have become an important part of clinical workup in MF and SS, providing the potential for more accurate prognosis and appropriate management. However, although international recommendations now include guidelines for FC analysis of peripheral blood markers for staging purposes, many clinics only perform these analyses in patients with advanced-stage lymphoma, if at all, and there is still a need for standardized use of validated markers. Standardization of a single effective multiparameter FC panel would allow for accurate identification and quantification of blood tumour burden for diagnosis, staging, assessment of therapeutic response, and monitoring of disease progression at all stages of disease. Once defined, validation of an MF/SS biomarker FC panel will enable uptake into clinical settings along with associated standardization of protocols and reagents. This review discusses the evolution of the role of FC in evaluating blood involvement in MF and SS, considers recently published international guidelines and identifies evidence gaps for future research that will allow for standardization of FC in MF and SS.

Introduction

Cutaneous T‐cell lymphomas (CTCL) are a heterogeneous group of non‐Hodgkin T‐cell lymphomas that primarily present in the skin but can also progress to involve blood, lymph nodes and viscera. The best‐studied CTCL subtypes are mycosis fungoides (MF) and Sézary syndrome (SS), together accounting for around two‐thirds of all CTCL cases. , Disease‐specific staging in MF and SS is the major predictor of prognosis. ,

MF is generally considered to be indolent with a variable type and extent of skin disease, and includes a subset of patients presenting with or developing extracutaneous disease. Previously thought to originate from mature tissue‐resident T cells expressing skin‐homing markers, multifocal skin presentation and failure of skin‐directed treatments to achieve cure or long‐term remissions, represent two aspects of the disease that disagree with this model of pathogenesis. Recent evidence suggests that neoplastic T‐cell clones are present in the peripheral blood of patients with MF, even in the early stages of the disease, and haematogenous spread of neoplastic cells may consecutively seed skin lesions. ,

Early‐stage MF (stage IA, IB, IIA) often has little impact on life expectancy; estimated median survival for stage IA disease is 35·5 years. , , , , However, up to 34% of patients with early‐stage disease progress to advanced‐stage disease (stage IIB to IVB); overall survival is estimated at 4·7 years for stage IIB compared with 15·8 years for stage IIA. ,

SS is a rare and aggressive CTCL subtype characterized by overt blood involvement (leukaemic cell count ≥ 1000 per mm3) at diagnosis, erythroderma and generalized lymphadenopathy, and is associated with a median survival of around 3 years. , ,

In 2007, blood classification was added to the tumour–node–metastasis (TNM) classification of MF/SS by the International Society for Cutaneous Lymphomas (ISCL) and the European Organisation for Research and Treatment of Cancer (EORTC), based on recognition of blood involvement as an independent negative prognostic factor, leading to the TNM–blood (TNMB) classification system. , In these recommendations, blood assessment was based on both morphology (manual Sézary cell counting) and flow cytometry (FC). It was acknowledged that FC offered a more objective tool for identifying and quantifying neoplastic lymphocytes in blood compared with manual cell counts, which are notoriously subjective.

In recent years, FC usage has expanded in this indication and has been demonstrated to be of value for disease staging, prognostic evaluation, guiding treatment approach, and therapeutic response assessment. However, there is still a need for standardization of FC methodologies to provide objective, consistent and quantifiable measures of blood tumour burden in MF/SS. , This review discusses the evolution of FC in evaluating blood involvement in MF/SS, considers recently published international guidelines, and identifies evidence gaps for future research that will allow for FC standardization in MF/SS.

History of the development of flow cytometry in cutaneous T‐cell lymphomas

Before the TNMB classification, blood involvement was considered to be either ‘absent’ (B0) or ‘present’ (B1) and was not included in clinical staging (Table 1). In 2007, the ISCL/EORTC revised MF/SS staging and classification guidelines defined the following three stages of blood involvement: B0, absence of significant blood involvement; B1, low blood tumour burden and B2, high blood tumour burden (Table 1). In recognition of the fact that proportional measures can vary depending on overall leucocyte count, the use of absolute counts was suggested in a 2011 ISCL/United States Cutaneous Lymphoma Consortium (USCLC)/EORTC update, with a subsequent update published in 2021. These definitions are still used in relevant MF/SS treatment guidelines, including those of the EORTC, the European Society for Medical Oncology (ESMO), and the National Comprehensive Cancer Network in the USA. , The 2018 ESMO guidelines recommend peripheral blood FC for all MF stages but state that the test may be more appropriate in patients with suspected SS, while the EORTC recommend FC for measuring blood involvement at all MF stages and in SS, as accurate blood class assignment is necessary for overall disease staging and consequent management. ,

Table 1

Evolution of blood involvement classification for mycosis fungoides and Sézary syndrome

TNM classification established 1	B0: No significant Sézary cell presence (< 5%)
	B1: Sézary cells present (> 5%); record total white blood cell count, total lymphocyte counts and number of Sézary cells per 100 lymphocytes
TNM classification revised to TNMB12	B0: Lack of significant blood involvement, indicated by ≤ 5% Sézary cells
	B0a: T‐cell clone absent
	B0b: T‐cell clone present
	B1: Low blood tumour burden, indicated by > 5% Sézary cells, but < 1000 per μL and/or absence of T‐cell clone
	B1a: T‐cell clone absent
	B1b: T‐cell clone present
	B2: High blood tumour burden, indicated by ≥ 1000 Sézary cells per μL with T‐cell clone present
	Alternatively, Sézary cell count may be substituted by CD4+ or CD3+ cells with CD4/CD8 ratio of ≥ 10
	or expanded CD4+ cells with abnormal immunophenotype including CD7 or CD26 loss (either ≥ 40% CD4+ CD7– or ≥ 30% CD4+ CD26–)
EORTC‐CLTF recommendation, 17 based upon parameters first presented by ISCL‐USCLC‐EORTC 19	Flow cytometry used to measure absolute CD4+ CD7– or CD4+ CD26– counts
	B0: Absolute count is < 250 per μL
	B1: Absolute count is from 250 per μL to < 1000 per μL
	B2: Absolute count is ≥ 1000 per μL with T‐cell clone present
ISCL‐USCLC‐EORTC modified staging 20	Flow cytometry used to measure absolute CD4+ CD7− or CD4+ CD26− counts or another aberrant lymphocyte population that has been identified by flow cytometry. Absolute counts should be determined by the percentage of aberrant lymphocytes identified by flow cytometry, multiplied by the total lymphocyte count of a complete blood count. Alternatively, the percentage of aberrant CD45+ leucocytes multiplied by the white blood cell count may be used
	B0: Absolute count is < 250 per μL
	B0A: T‐cell clone absent or equivocal
	B0B: T‐cell clone present and identical to skin
	B1: Absolute count is from 250 per μL to < 1000 per μL
	B1A: T‐cell clone absent or equivocal
	B1B: T‐cell clone present and identical to skin
	B2: Absolute count is ≥ 1000 per μL
	B2A: T‐cell clone absent or equivocal
	B2B: T‐cell clone present and identical to skin
	BX: Blood involvement cannot be quantified according to agreed guidelines
	BXA: T‐cell clone absent or equivocal
	BXB: T‐cell clone present and identical to skin

CLTF, Cutaneous Lymphoma Task Force; EORTC, European Organisation for Research and Treatment of Cancer; ISCL, International Society for Cutaneous Lymphoma; TNM, tumour–node–metastasis; TNMB, tumour–node–metastasis–blood; USCLC, United States Cutaneous Lymphoma Consortium.

The ISCL/EORTC criteria also provided subcategories for B0/B1 disease to account for the absence or presence of a T‐cell clone identified by analysis of the T‐cell receptor (TCR) gene (Table 1). Polymerase chain reaction (PCR) has replaced Southern blot analysis as the ‘gold standard’ for clonality testing and the development of standardized BIOMED‐2/EuroClonality multiplex PCR protocols has greatly supported T‐cell clonality assessment and diagnosis of lymphoid malignancies. , TCR gene rearrangement in peripheral blood T cells can occur early in MF and has prognostic significance if it is the same as that found in the skin. , , Based on these criteria, diagnosis of SS requires B2 blood involvement with a T‐cell clone in peripheral blood identical to that found in the skin. , , , B2‐level blood involvement can also occur rarely in patients with MF who have nonerythrodermic disease. ,

Historically, detection of Sézary cells has been performed using light‐microscopic analysis, a time‐consuming and subjective technique with high rates of interobserver variability. , , This has now largely been replaced by FC analysis, which offers immunophenotyping of T‐cell subsets and assessment of clonality. , , TCR gene sequencing in conjunction with FC can increase blood classification accuracy, particularly for patients with low tumour burden.

Current guidelines for peripheral blood staging and assessment of blood response

There was previously no objective definition of blood class using FC, and centres adopted different definitions in their publications. The potential use of FC to measure numbers of CD4+ CD26– and/or CD4+ CD7– T‐cell populations to quantify blood involvement was discussed in the 2011 ISCL/USCLC/EORTC consensus statement as a reasonable and quantifiable method for use in clinical trials. An EORTC Cutaneous Lymphoma Task Force Committee later published recommendations for blood classification and blood response criteria to bridge the transition from manual Sézary cell counting to a more standardized use of FC. FC analysis was recommended to count absolute numbers of abnormal (CD4+ CD26– and/or CD4+ CD7–) T cells, owing to the potential for proportional skewing based on overall lymphocyte counts, and recommended FC as the method to measure blood involvement in all stages of MF and SS.

The most recent guidelines update (ISCL/USCLC/EORTC) published in 2021, defines blood classifications B0, B1 and B2 by FC using an absolute count of CD4+ CD26–, CD4+ CD7– or other aberrant lymphocyte population identified by FC, with an additional BX classification where blood involvement cannot be quantified according to agreed guidelines (Table 1). T‐cell counts are based on flow gating using markers to select cells of interest. However, gating strategies vary and, therefore, a set of basic gating principles has been proposed.

This recommendation for FC evaluation of the immunophenotype of abnormal T cells in blood complements the analysis of skin biopsy samples in which immunohistochemical analysis of CD2, CD3, CD4, CD5, CD7, CD8, CD20 and CD30 expression and molecular analysis of TCR gene rearrangements are recommended.

Despite international recommendations, FC analysis and tracking of blood involvement is often performed only in patients with advanced‐stage disease. , Therefore, we present our consensus recommendations for cases where use of FC in MF/SS is especially appropriate, in centres where it may not be practical or possible to assess and track blood tumour burden for all patients (Table 2).

Table 2

Consensus recommendations for minimum use of flow cytometry for mycosis fungoides and Sézary syndrome when not performed at all stages

Flow cytometry use: consensus recommendations
Clinical flags for the use of flow cytometry	How often during follow‐up?
Patient with advanced‐stage disease (stage IIB and above)	Every 3 months in those patients with abnormal flow cytometry at baseline
Intractable pruritus	In the case of disease/stage progression
Generalized patches and/or plaques (T2A/T2B)	Upon development of any clinical flags presented in the left‐hand column
Erythroderma
Lymphocytosis on WBC
High serum LDH
Lack of response to treatment

LDH, lactate dehydrogenase; WBC, white blood cell count.

Biomarkers used for flow cytometry analysis of mycosis fungoides and Sézary syndrome peripheral blood involvement

Although the EORTC definitions are of value and utilized in clinical practice, there is still need for a widely used standardized and validated set of markers. , Recent efforts have focused on further identifying the most useful cell‐surface markers to immunophenotypically characterize Sézary cells and distinguish them from non‐neoplastic T cells. ,

The most common abnormalities in T cells of MF/SS are the loss of CD7 or CD26 followed by reduced or absent CD2 and/or CD3. , , Identification of higher numbers of T cells with abnormal expression of CD3, CD7, CD26 or TCR variants (TCRαβ, TCRγδ) has been correlated with more advanced stages of MF and shown to predict disease progression. , Recently, the six‐biomarker panel of CD3, CD4, CD7, CD8, CD26 and CD45 has been recommended as the minimum combination of antibodies to be included in a single six‐colour FC assay for MF/SS. , CD3 is associated with the TCR and has value for gating the FC analysis to target T cells. Surface expression of CD3 has also been shown to be diminished in 40–80% of MF/SS cases with blood involvement. CD4 and CD8 are T‐cell markers that identify the relevant T‐cell subset, and most cases involve CD4+ T cells. CD8 is included to allow for CD4+ CD8+ double‐positive cells to be gated out, as these are more commonly involved in reactive conditions and complicate the phenotypic evaluation of the abnormal T‐cell subset. While CD8− CD4− (‘double‐negative’) cases do exist, they are rare and only case studies/series are available. Therefore, these cases are the subject of ongoing identification, which further emphasizes the importance of standardized approaches to FC, in order to allow for meaningful interpretation of data generated from these cases. Loss of CD7 and/or CD26 is a common finding in MF/SS, identified in 50–80% and 70–100% of cases with blood involvement, respectively; however this is also often observed in reactive CD4 T‐cell conditions. When CD4+ CD7– cells comprised > 40% of the CD4+ T‐cell population, researchers were able to discriminate between SS and erythrodermic inflammatory dermatoses (EID) with a sensitivity of 54% and specificity of 100%. The same study also examined CD4+ CD26– cells and found that when these comprised ≥ 30% of the CD4+ T‐cell population, SS vs. EID could be distinguished with a sensitivity of 86% and specificity of 47%; when there were > 80% CD26– cells within the CD4+ population, the specificity was 100% and sensitivity was 66%. However, when considering the percentage value of the CD4+ CD26− subset, in the original paper by Bernengo et al., the positive predictive value was 100% for a cut‐off value of 30%; sensitivity was 97% and specificity was 100%. Both of these markers have been associated with disease stage and disease progression in MF/SS· , ,

However, the reference population (either total lymphocytes or CD4+ T cells) used for CD7– or CD26– percentages was not consistent between centres. In a prospective study that included 254 patients referred for testing for CTCL at initial diagnosis, the best specificity and sensitivity was obtained when total lymphocyte count was used as the reference subset at the thresholds indicated by EORTC/ISCL. Finally, CD45 inclusion is suggested as a way to identify haematopoietic (CD45+) cells and exclude CD45– nonhaematopoietic cells from the analysis.

The presence of cell populations with low (‘dim’) expression of certain markers on FC has been included as part of the diagnostic criteria for other haematological malignancies. SS has been commonly reported to include a ‘dim’ CD3, CD4 or CD2 population. , It has been suggested that the presence of ‘dim’ T‐cell marker expression, coupled with CD26 and/or TCR‐Vβ analysis, may represent a good tool for tumour burden assessment. Good resolution of a ‘dim’ population is reliant on fluorochrome choice, particularly in multicolour combinations. ,

While the CD3+ CD4+ CD7− and/or CD26− immunophenotype is characteristic of MF/SS, it is not specific, as intrapatient heterogeneity of MF/SS cell markers has been observed. , Healthy CD4+ T cells in patients with SS often display aberrant phenotypes, not only with regard to CD7 and CD26, but also in relation to the usual markers of naive/memory cells and exhaustion, similar to that seen in Sézary cells. , Analysis of CD4+ T cells by Roelens et al. found that CD4+ CD45RA− CD26− CD27+ CD28+ PD1+ T cells represented 58·5% of Sézary cells and 49·2% of healthy CD4+ T cells in a patient with SS, but only 3·8% of CD4+ T cells in a healthy donor. In addition, healthy and malignant CD4+ T cells often evolve in the same way under treatment. Given these common phenotypes, the reliability of extended phenotypes using only nonspecific markers for Sézary cells may be questionable.

Additional potentially useful markers for FC are also under evaluation, although no consensus has been reached. It is suggested that additional antigens should be combined with CD3, CD4, CD, and CD26, at the very least (Figure 1). , , ,

Figure 1

Markers with evidence for the Sézary cell phenotype and those recommended as diagnostic markers for FC. FC, flow cytometry; PD‐1, programmed cell death‐1. *Not limited to markers routinely detected by FC. [Colour figure can be viewed at wileyonlinelibrary.com]

Recent studies have investigated the expression of killer‐cell immunoglobulin‐like receptors (KIRs) (e.g. CD158a, CD158b and CD158k) and adhesion molecules (CD164); however, results have generally been inconclusive. The identification of KIR3DL2/CD158k on malignant cells in SS and advanced MF has greatly helped the detailed study of the malignant clone. Since it was first described by Bagot et al. in 2001, the reliability of this marker for detecting Sézary cells in blood and skin has been demonstrated by several teams. , , In a recent study, the presence of KIR3DL2+ Sézary cells ≥ 200 μL−1 correlated with an SS diagnosis at a sensitivity of 88·6% and specificity of 96·3%; this contrasts with Boonk et al. who found a similar specificity of 95%, but a sensitivity of only 33%. However, the difference in sensitivity may be related to the use of fresh vs. frozen samples in the Roelens and Boonk studies, respectively. Tracking Sézary cells with KIR3DL2/CD158k, which remains constant over time, allows discrimination from benign cells that may have the same aberrant phenotypes with defects in nonspecific marker expression. Indeed, the decrease of such normal T cells expressing CD4+ CD7− or CD4+ CD26− may not indicate neoplastic cell clearance and, conversely, the presence of these cells may underestimate the rate of blood response. Since 2014, KIR3DL2/CD158k analysis has formed part of the routine care of patients with erythroderma at Hospital Saint‐Louis in Paris, and, together with Bordeaux and Clermont‐Ferrand centres, the Moins‐Teisserenc team provides harmonized FC protocols to all French laboratories and standardized guidelines for analysis. Further international validation of this marker is required to confirm its usefulness for diagnostics, prognostics and evaluation of therapies.

As T‐cell antigenic variations are not always specific to MF/SS, evaluation of TCR gene rearrangements is also recommended to confirm clonality in diagnosis and monitoring, and to correlate with other molecular findings from skin or node biopsy analyses. This is critical, as peripheral blood nonpathogenic T‐clones can occur in healthy individuals, especially with increasing age, but these clones are not identical to those in skin or node biopsies. Identification of an identical T‐cell clone in the skin and in the blood has been shown to be associated with a poor prognosis, even in patients with a B0 blood classification.

The use of FC panels of TCR‐Vβ‐specific antibodies to screen for Vβ domain expression in T‐cell proliferations was originally proposed two decades ago, based on the assessment of large series of antibodies in normal and malignant T cells. , This method has shown efficacy for both initial blood staging and for assessment of therapeutic response, and can be used to confirm that the features of the T‐cell clone in blood are similar to those of the skin biopsy T cells. , , However, this TCR‐Vβ‐specific antibody panel does not identify all different TCR‐Vβ chain variables, and the expression of different TCR‐Vβ varies in healthy individuals, making interpretation of small clones challenging.

The future of flow cytometry in mycosis fungoides/Sézary syndrome

While progress has been made towards development of a standardized biomarker set to identify and track disease in MF/SS, further improvements are still required and protocols must be disseminated to encourage universal adoption of a single effective and validated standard. Standardization of a single effective multiparameter FC panel will allow for accurate identification and quantification of blood tumour burden for diagnosis, staging and assessment of therapeutic response, including the development of escape variants and monitoring of disease progression. Blood staging is not only relevant to initial diagnosis, but also as an indicator of response to therapy. Thus, FC needs to be evaluated as a criterion for determining blood response, and as a method for both detection of resistant clones to therapy and minimal residual disease assessment. Moreover, FC can be useful to identify the expression of specific cell markers to be recognized by monoclonal antibodies for treatment as targeted therapies. Current potential therapeutic targets and relevant therapies include CD30 (brentuximab vedotin, licensed in CD30+ CTCL), CCR4 (mogamulizumab), PD‐1 (nivolumab and pembrolizumab), CD158k/KIR3DL2 (lacutamab), CD47 (TTI‐621) and CD52 (alemtuzumab).

One example of a successful standardization method is the EuroFlow programme, established to develop standardized protocols for diagnostic use and monitoring treatment response by FC in various lymphomas and leukaemias. , , This programme is being used to design a marker set for MF/SS, with the six‐biomarker antibody set and TCR clonality markers as a starting point. For example, CD3, CD4, CD8 and CD45 would be included as the backbone markers, and CD7 and CD26 would be included to identify MF/SS cells. Other markers could then be tested in combination with this backbone set to identify and validate the most clinically useful final set to be used in conjunction with TCR clonality testing. Prospective multicentre studies and collaborations between EORTC and EuroFlow are under way to develop rapid, accurate, standardized protocols and validated cut‐off values for the assessment of blood involvement in CTCL. ,

Accurate quantification of blood involvement could also help to ascertain whether it would be clinically relevant to identify a further cut‐off, higher than B2, to characterize patients with a very high blood burden, potentially with a worse prognosis, who therefore would be candidates for a more aggressive therapeutic strategy.

There are a number of barriers to overcome in the use of FC for the assessment of blood tumour burden in CTCL. Firstly, markers may overlap with normal or reactive T‐cell populations. , , Secondly, clonality evaluation using TCR‐Vβ antibodies covers only about 70% of the TCR repertoire, is ineffective for those T cells that have lost TCR expression, and is expensive to perform. , Finally, standardized protocols, including the development of standard operating procedures for reagents, procedures and analysis, must be adhered to and technicians should be trained in the use of appropriate reagents.

Recent FC studies using EuroFlow‐based panels to analyse blood samples from 24 patients with SS demonstrated extensive inter‐ and intrapatient phenotypic heterogeneity with changes over time, which could be tracked by next‐generation sequencing of TCRs to identify a patient’s tumour‐specific ‘barcode’. , Multicentre follow‐up studies are currently being performed to develop a new standard for improved blood staging and disease monitoring.

Conclusions

This review focuses on the relevance of adequate blood monitoring in patients with MF/SS during the disease course in order to better tailor treatment strategies and define the clinical behaviour of the disease. Recent advances support the use of FC to quantify blood tumour burden, using absolute counts of abnormal T cells, but there is a need to test the clinical implications of this in prospective studies. Moreover, an adequate blood tumour burden definition could help in identifying the prognostic and/or predictive role played by blood involvement in early disease, and in understanding whether patients with a very high blood tumour burden could be classified in a separate stage. For this objective, and to better characterize the clonal evolution and potential heterogeneity of circulating T‐cell subclones, a better definition of the circulating clonal cells is needed, specifically in terms of FC markers that would be able to identify neoplastic cells.

Author contributions

Maarten H. Vermeer: Conceptualization (equal); validation (equal); writing – review and editing (equal). helene Moins‐Teisserenc: Conceptualization (equal); validation (equal); writing – review and editing (equal). Martine Bagot: Conceptualization (equal); validation (equal); writing – review and editing (equal). Pietro QUAGLINO: Conceptualization (equal); validation (equal); writing – review and editing (equal). Sean Whittaker: Conceptualization (equal); visualization (equal); writing – review and editing (equal).