Clonality Testing in the Diagnosis of Cutaneous T-Cell Lymphoma

Abstract

Cutaneous T-cell lymphomas (CTCL) are heterogeneous disorders including several different entities with variable morphology, phenotype, and clinical features. The diagnosis of CTCL is often challenging; the distinction between tumors and reactive, non-neoplastic conditions is sometimes elusive. Therefore, molecular testing for T-cell receptor gamma (TRG) clonality is often performed. In this study, we evaluated the accuracy of TRG testing protocol in 110 routinary cases and discussed the subject of clonality testing in light of novel technologies (namely next-generations sequencing advent). For TRG analysis, the BIOMED-2 protocol was adopted. Sensitivity, specificity, positive and negative predictive values were calculated using the Oxford CatMaker software following the Standards for Reporting of Diagnostic Accuracy Studies (STARD) requirements for diagnostic accuracy. We found that this approach was feasible in most cases (87%) despite the small sample dimensions and the fixation issues. In addition, we found that sensitivity and specificity were 90 and 84%, respectively; accordingly, positive predictive value (PPV) and negative predictive value (NPV) were 84 and 90%, respectively. Of note, the molecular test was somehow influential in 83% of cases, when histology and phenotyping were not conclusive. In conclusion, TRG analysis by BIOMED-2 protocols is feasible and effective in the routine diagnostics of CTCL. The integration of histological, phenotypical, molecular, and clinical data is mandatory for a correct diagnosis.

1. Introduction

Peripheral T-cell lymphomas (PTCLs) are clonal neoplasms deriving from mature T-lymphocytes at different stages of differentiation [1]. Clinical presentations are protean, and diagnosis involves a synthesis of clinical history and pathological data from morphological, immunophenotypic, cytogenetic, and molecular investigations. Based on such criteria, the recent WHO Classification of Tumors of Hematopoietic and Lymphoid Tissues codified many disease entities, identifying specified and not otherwise specified (NOS) forms, nodal and extranodal tumors [1]. Among the latter, cutaneous T-cell lymphomas (CTCL) are relatively common, including several distinct diseases [2]. The diagnosis of CTCL, as for other PTCLs, is challenging owing to the heterogeneous pathophysiology, complex taxonomy, and, especially, the lack of tumor-associated immunophenotypic or molecular markers. Furthermore, the clinical scenario becomes yet more complicated when the disease presents in the early phase or with the contemporaneous occurrence of a prominent reactive component overwhelming the malignant elements [2].

Nonetheless, importantly, the diagnosis of lymphoproliferative disease can benefit from the demonstration of clonality at a molecular level. Specifically, in the field of T-cell lymphomas/leukemias, the study of T-cell receptor gene (TCR) rearrangements has been used to assist in the determination of clonality for over two decades [3, 4]. The somatic assembly of TCR genes is an early event during T-cell development, occurring through a V(D)J recombination directed by site-specific recombinase (RAG1 and RAG2), which eventually leads to a diverse T-cell population [3, 5]. In 2003, the BIOMED-2 consortium established standardized protocols for multiplex PCR analysis, which improved the efficiency and reproducibility of clonality detection [6]. This protocol was validated by the BIOMED-2 group in 2007 in 188 cases of T-cell malignancy [7, 8]. Specifically, the combined analysis of TCR-gamma (TRG) and TCR-beta (TRB) led to sensitivity and specificity values above 90% [6, 7, 8].

Although the efficiency of BIOMED-2 protocols was defined, in both studies, the BIOMED-2 protocol was tested in a series of tissues with well-established histological diagnoses of T-cell malignancy, principally represented by fresh or frozen specimens [6, 7, 8]. By contrast, routine clinical diagnosis is predominantly based on formalin-fixed paraffin-embedded (FFPE) tissue blocks. FFPE blocks are convenient as they are small, easy to transport, and allow for retrospective analysis by immunohistochemistry and in situ hybridization (ISH/FISH). However, by comparison to fresh specimens, molecular analyses on FFPE blocks are more challenging owing principally to the more involved method required for extracting high-quality DNA. The interference of formalin fixation with DNA architecture and nucleoside fragmentation can lead to lower amplification yield and potentially a lower diagnostic utility [9, 10]. In this regard, small specimens, particularly those such as skin biopsies, represent a significant challenge in molecular testing. Practically, the two steps (TRG and TRB) can be performed simultaneously to save time, or consecutively, to save resources. In fact, the latter solution may allow the analysis to be limited to TRG in most instances. Noteworthy is that a formal, evidence-based diagnostic accuracy study has never been performed in this setting.

In order to properly quantify this possibility, in this study we performed a diagnostic accuracy study based on a retrospective analysis of our experience with the BIOMED-2 TRG protocol in 110 skin specimens aiming to: [1] assess the feasibility of the protocol in non-selected consecutive FFPE cases included in the routine practice; [2] assess sensitivity, specificity, PPV and NPV of the test in detecting T-cell malignancies in FFPE tissues; and [3] determine the rate and clinical significance of clonal rearrangements in cases deemed to be reactive or non-diagnosable by histopathological and morphological assessment (“TRG monoclonality of uncertain significance”).

Furthermore, we discuss the issue of clonality testing in light of the new available technologies, namely next-generation sequencing (NGS).

2. Materials and methods

2.1 Case selection

We included in this study 110 consecutive cases received over 3 years at the Molecular Pathology Laboratory, Hematopathology Unit, Institute of Hematology and Medical Oncology “L. & A. Seràgnoli.” Histological samples were classified by at least two experienced hematopathologists who were blinded to the TRG result using the morphological, immunohistochemical, and cytogenetic criteria of the most updated WHO classification [2]. The diagnoses were as follows: PTCL/CTCL (N = 48; 49%), B-cell lymphomas (N = 8; 8%), and inflammatory disorders (N = 42, 43%). We also studied a further 12 cases for which a definitive diagnosis was not reached based on morphology, phenotype, or clinical presentation, and where the differential between tumor and inflammatory disorders was not resolved. For all cases, a minimum of 12 months of clinical follow-up from the last PCR study was available. Patients’ characteristics are reported in Table A1.

All specimens were formalin-fixed and paraffin-embedded (FFPE) prior to PCR analysis.

Calculations of sensitivity (ST), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) were made by CATmaker software (Centre for Evidence-Based Medicine, Oxford University, http://www.cebm.net) using the blinded histopathological review as the reference test in our case series.

For the purposes of this analysis, a distinction was not drawn between primary cutaneous T-cell lymphomas and systemic lymphomas manifesting in the skin, as the analysis was designed to assess the capacity of the molecular investigation to detect any T-cell lymphoma in the skin. The STARD Statement for diagnostic accuracy studies was fulfilled (Table A2) [11].

The study design is summarized in Figure 1.

Figure 1.

Schematic representation of the diagnostic accuracy study design.

2.2 Molecular analysis

Molecular tests were performed and analyzed by at least two experienced molecular pathologists. Molecular testing was performed within 1 week of the histological evaluation. Histopathological evaluation and molecular tests were performed in a blinded manner.

Genomic DNA was extracted from paraffin-embedded tissue by using a Qiagen DNA mini kit according to the manufacturer’s instructions. PLZF was studied as a control gene in order to verify the DNA integrity. PCR analysis of TRG rearrangement was executed using multiplex PCR tubes (BIOMED-2 Concerted Action) [6]. Amplification reactions were performed in an automated thermocycler (mastercycler Eppendorf) according to the BIOMED-2 multiplex PCR protocol [6]. Each 50-microliter PCR reaction included 100 ng of DNA, 10 pm of each primer, 0,2 mmol/L dNTP, 5 microliter of 10X Gold buffer, 1.5 mmol/L of MgCl2 and 1 U of Ampli-Taq Gold polymerase (Applied Biosystems). The cycling parameters were as follows: pre-activation for 7 minutes at 95°C, 35 cycles of denaturation (95°C for 30 sec), an annealing step (60°C for 30 sec), an extension step (72°C for 60 sec), finally 10 minutes of extension at 72°C. Each sample was evaluated using heteroduplex analysis and GeneScanning to determine polyclonal or monoclonal character.

2.2.1 Heteroduplex analysis

In order to allow the spontaneous formation of homoduplex or heteroduplex from DNA fragments, amplicons were denatured by heating (94°C for 5 min) and cooled at low temperatures (4°C for 1 hour). PCR products were then visualized by electrophoresis on polyacrylamide gels being separated according to their length and conformation. According to BIOMED-2 protocols, samples were defined as monoclonal when a single band was identified within a predictable size range and polyclonal when only a smear was detected [6]. This technique has a well-recognized detection limit of ∼5% due to the frequency of polyclonal/reactive lymphocytes present in the tissue.

2.2.2 GeneScanning analysis

After amplification, 1 μ of PCR product with 0.5 μ of a standard molecular weight product (LIZ Applied Biosystems) was mixed with 12 μ of formamide to induce denaturation into single DNA strands (1 minute to 95C°). Subsequently, they were separated through a polymer capillary electrophoresis system and automatically detected by fluorescence reading with a laser system in an automatic DNA sequencer (ABI Prism 310 Applied Biosystems).

Tissue samples were considered to have a clonal T-cell population if 1 or 2 peaks of the amplified PCR product in question were obtained (monoclonal and biclonal/biallelic, respectively); detection of three to five peaks was counted as an oligoclonal result, while a Gaussian distribution of peaks was referred to as a polyclonal population [6]. The control gene minimum amplification requirement was 300 bps, and when this was not reached, or the target gene was not amplified, the sample was considered “not evaluable.”

3. Results

A total of 110 case samples were interrogated using the BIOMED-2 protocol for TRG analysis. Fourteen cases did not have an evaluable molecular result (14/110, 12.7%), owing to failure to amplify either the target gene product or the control gene, and were then excluded from the final analysis. 96/110 samples (87.3%) could be evaluated and were included in the final analysis.

Of the specimens that allowed a clear and definitive histopathological diagnosis, 84/96 had an evaluable molecular result (87.5%).

Based on TRG analysis, there was evidence of clonal rearrangement in 42/84 (50%) of cases, oligoclonal in 3/84 (4%), and a polyclonal pattern was detected in 39/84 (46%) cases, respectively.

Mono/biclonality was detected in 36/47 (77%) of the CTCL cases, 2/8 (25%) of the B-NHL cases, and 4/42 (10%) of the inflammatory reactions.

The ST and SP of the molecular test in discriminating malignant T-cell lymphoproliferation vs. other conditions (including B-NHL and inflammatory reactions) were 90.0 and 84%, respectively. Accordingly, PPV and NPV were 84.0 and 90%, respectively (Table 1).

Target: Differential diagnosis CTCL vs. Other
		Histopathological diagnosis
		CTCL	Other
TRG	Clonal vs. Oligo-polyclonal	36	7
		4	37
	Parameter	Value	95% CI
	ST	90%	81–99
	SP	84%	73–95
	Pre-test probability	48%	37–58
	PPV	84%	73–95
	NPV	90%	81–99
	LR+	5.66	2.85 to 11.25
	LR-	0.12	0.05–0.30
ST = sensitivity
SP = specificity
PPV = positive predictive value
NPV = negative predictive value
LR + = likelyhood ratio positive
LR- = likelyhood ratio negative

Table 1.

Calculation of sensitivity (ST), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV), basing on the histopathological analysis taken as diagnostic gold standard.

CATmaker software (Centre for Evidence-Based Medicine, Oxford University, http://www.cebm.net).

Subsequently, we evaluated the 12 cases for which a conclusive diagnosis was not made by histopathological review. Specifically, we detected monoclonal rearrangement in 4/12 cases, oligoclonal rearrangement in 3/12, and 5/12 presented with a polyclonal pattern. In the monoclonal group, 3/4 ultimately received a diagnosis of CTCL after the integration of all diagnostic procedures. Among oligoclonal/polyclonal cases, 1 was eventually diagnosed as CTCL, while the remaining 8 were diagnosed as inflammatory conditions. Overall, based on such data, the molecular result was influential on the final diagnosis in 10/12 (83%) of the cases when morphology, immunophenotype, and molecular data were integrated (Table 2).

	Number of cases	Diagnosis
		TCL	Non-TCL
TRG
Monoclonal	4	3 (75%)	1 (25%)
polyclonal/oligoclonal	8	1/0 (12.5%)	4/3 (87.5%)

Table 2.

Impact of the molecular result in cases with undefined diagnosis.

We then evaluated the molecular results in PTCLs according to the histological subtype in order to assess whether specific diseases are differentially associated with different degrees of clonality at TRG analysis (Table 3). We found that there were in fact no significant differences recorded, with TRG successful in detecting clonal rearrangements in the majority of cases (Table 4).

Lymphoma type	Number of cases	Evidence of clonality (number of cases)	Evidence of clonality (percentage of cases)
Mycosis fungoides	18	16	88.89
PTCL/NOS	15	13	86.67
CTCL/NOS	5	5	100.00
T-LGL	1	1	100.00
ALCL, ALK-	1	1	100.00
Total	40	36	90.00

Table 3.

TRG clonality according to lymphoma histotype.

	Sample	Tissue	Efficiency (% of evaluable cases)	Probe	ST (%)	SP (%)	PPV (%)	NPV (%)
Seràgnoli	FFPE	Skin	88	γ	73	80	75	78
Zhang	FFPE	Skin	n/a	γ	64	84	var	var
				β	64	84	var	var
				γ + β	78	74	var	var
Lukowsky	FFPE	Skin	n/a	γ	81	n/a	n/a	n/a
				β	78	n/a	n/a	n/a
				γ + β	87	n/a	n/a	n/a
Goeldel	Frozen	Skin	0	γ	77	84	84	79
Ponti	Fresh/frozen	Skin	n/a	γ	84	98	95	92
Biomed-2 Group	Fresh/frozen	Not specified	n/a	γ	89	n/a	n/a	n/a
				β	91	n/a	n/a	n/a
				γ + β	94	n/a	n/a	n/a

Table 4.

Comparison of sensitivity (ST), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) in different studies adopting the BIOMED-2 protocols.

n/a = not available.

var = variable PPV and NPV dependent on pre-test probability.

Finally, we investigated whether cases presenting evidence of clonal rearrangement but without clear evidence of PTCL had a higher risk of developing T-cell malignancies in the following months. Specifically, among cases that received a final diagnosis of inflammatory disorder, 4/42 (9.5%) demonstrated TRG clonality. These particular patients largely fulfilled the minimum requested follow-up of 12 months, the mean follow-up period being 35 months (range 33–36 months). Patients were then monitored by means of outpatient clinical follow-up and/or phone contact with the treating physician. No patient received a diagnosis of a T-cell neoplasm during the study period, to the best of our knowledge.

Matched histological and molecular analysis results are detailed in Table A3.

4. Discussion

In this study, we examined the sensitivity and specificity and, accordingly, PPV and NPV, of the BIOMED-2 protocol in detecting TRG clonal rearrangements in cutaneous T-cell lymphoproliferative disorders. We included 110 consecutive cases of histologically complex FFPE skin biopsies referred to our institution over a 3-year period.

In our hands, the BIOMED-2 TRG protocol was technically successful in 87% of cases, based on a control gene minimum amplification requirement of 300 bps. In particular, it demonstrated a sensitivity of 90.0% and a specificity of 84%. The sensitivity and specificity data using this same approach have been published in previous papers (Table 4), and the figures obtained in our institution are comparable to previously published data in FFPE specimens. This suggests that sensitivity in FFPE specimens can approach that of fresh frozen specimens. With regard to FFPE samples, Zhang et al. reported a sensitivity and specificity of 64 and 84%, respectively, in a case series looking exclusively at MF, while Lukowsky et al. reported a sensitivity of 81% [12, 13]. In fresh/frozen tissues, Goeldel et al. reported a sensitivity and specificity of 77 and 84%, respectively [9, 14], while Ponti et al. described slightly higher figures of 83.5 and 97.7%, respectively [15]. A direct comparison with the original data from the BIOMED-2 consortium does, however, pose some problems as that group only analyzed fresh and frozen specimens and included non-skin specimens as well (Table 4).

Our study differs from the previous one, being the first phase 3 diagnostic accuracy study specifically designed and conducted according to the STRAD requirements for evidence-based medicine [11]. In doing so, we evaluated, for the first time in the specific setting of FFPE samples, the PPV and NPV of the BIOMED-2 TRG protocol. Importantly, such variables are strictly dependent on the incidence of the disease in the population tested (pre-test probability). In our laboratory, the use of TRG rearrangement testing is typically reserved for those cases where diagnostic uncertainty persists after morphological assessment with immunohistochemical staining, in which the pre-test probability is moderate/high and in which a positive or negative result may influence the final diagnosis [16]. In this context, the addition of a molecular tool as a diagnostic aid becomes significant. In this series, the pre-test probability was quite balanced, being 48%. We calculated a PPV of 84% with an NPV of 90% and found that most histologically “uncertain” cases could be satisfactorily resolved with the addition of the PCR information. Recently, Zhang et al. created a model that demonstrated the effect of pre-test probability on the PPV and NPV and suggested that the molecular result should only be considered when the pre-test probability lies between 0.15 and 0.75 [12], and this is actually consistent with our practice. Zhang et al. further pointed out that at the extremes of the pre-test probability range, the molecular result should be ignored [12]. Practically, based on the same SP and ST recorded in our study, a pre-test probability of 8% would lead to a positive predictive value of as low as 34%. Indeed, this underlines the importance of patient selection based on morphology and phenotype before molecular testing, which cannot represent a reliable screening test.

Regarding fresh/frozen samples, Ponti et al. described a remarkable PPV and NPV of 95 and 92%, respectively. It should be noted, however, that PPV and PNV may vary in the different studies due to differences in the considered series (i.e., research vs. routine diagnostic). In fact, at times, the clinical imperative urges contemporaneous initiation of multiple investigative strategies, including PCR, not respecting the ideal logical order (histology, immunohistochemistry, and genetics), thus affecting the pre-test probability. This real-world environment may alter the PPV and the NPV of the analysis. Further, the addition of TRB and TRD analysis, not available in this series, would potentially certainly improve the efficiency of the BIOMED-2 protocol. In this regard, it should be noted that this study did not aim to assess the accuracy of the entire protocol but rather of the TRG step, which is adopted as the first and often only one in many laboratories. For sure, the addition of the subsequent steps would increase the overall sensitivity, though negatively affecting, on the other hand, the specificity.

As mentioned, in spite of these remarkable figures for PPV and NPV, the importance of clinicopathologic integration in making a diagnosis in such a complex clinical setting must be remembered. It should be kept in mind that clonality, per sè, does not always indicate malignancy, and it is well-recognized that clonality can also be seen in reactive processes and thus cannot be considered a “sine qua non” of cancer [14, 17, 18, 19, 20]. However, when clonality is detected without a definitive diagnosis of malignancy, a close clinical follow-up is mandatory, in order to ensure occult T-cell lymphoma is not missed [17, 21, 22]. In fact, progression to frank T-cell malignancy has been reported with variable frequency [23, 24]. Indeed, several studies have examined the fate of patients with an indeterminate diagnosis, with a variable percentage of patients who later on presented with a clear CTCL, ranging from 0 to 85% [15, 25, 26, 27]. With this in mind, we followed up on cases that demonstrated clonality, but the global picture was not indicative of lymphoma. However, in our experience, no one developed a lymphoma within the follow-up period. We cannot exclude that a later onset might occur.

Diagnostic algorithms for MF incorporating molecular testing with integrated clinicopathologic findings have been examined by a number of groups including most notably the ISCL [28, 29, 30, 31]. Our findings suggested that TRG performed in FFPE in this subgroup remains a valid method of supporting a diagnosis in histologically challenging cases. Particularly, we had evidence that TRG clonal rearrangements were clinically useful also when a clear diagnosis of lymphoma could not be established. The role of molecular clonality testing in influencing the diagnosis of cutaneous lymphoid pathology in cases that are histologically uncertain needs further investigation. Nevertheless, it is now considered part of the diagnostic repertoire available to the pathologist and, indeed, has formed part of the ISCL and EORTC guidelines for diagnosis of MF since 2007 [29]. In our study, the most common malignancy identified was MF, in which 16/18 cases (88.9%) demonstrated TRG clonality. This compares to 68.2–88% as reported in the literature [12, 30], possibly reflecting the remarkable selection of cases based on immunomorphology in dedicated and ultra-specialistic Hematopathology Units. It is conceivable that the addiction of miRNA analysis [32] as well as comprehensive genetic testing [33, 34, 35], will further increase the importance of molecular pathology in CTCL analysis.

Finally, recent studies showed the potential efficacy of a next-generation sequencing (NGS) approach to resolve clonality testing [36]. These methods are based on the preparation of a DNA samples library amplified by PCR, using appropriate forward and reverse primers (tagged with unique barcodes) which usually target, respectively, different portions of V and J regions of the TRG gene; the amplified library is subsequently sequenced through a reversible dye-terminators technique and analyzed with bioinformatics approaches that return the proportion of the given sequences and assign them a rearrangement identity based on an alignment score [37, 38, 39]. These technologies are complex and parallelized processes that permit the simultaneous analysis of several samples returning high-throughput data. The introduction of NGS approach for the clonality assessment in CTCL can lead to useful advantages with respect to the gold standard PCR-based ones, overcoming its main limits [40, 41]: first, the possibility of distinguishing between same-sized amplified sequences, given that sequencing separates them also based on their nucleotides composition and not only on their length; further, also the possibility to better interpreting ambiguous results, as non-uniform Gaussian distributions, and better resolving a polyclonal background. In addition, this approach necessitates a minor amount of DNA (10–20 ng vs. 100–500 ng) and has the ability to detect and quantify also recurrent minor clones with small numbers of circulating tumor cells: as a result, NGS can improve the monitoring of the disease progression and treatments response, especially regarding minimal residual disease (MRD) context [42]. Hence, these new techniques could lead to a better and more objective classification, stratification, and monitoring of lymphoid malignancies, such as CTCL, which are essential proprieties in molecular diagnostics.

5. Conclusions

In conclusion, the TRG BIOMED-2 protocol appeared to be a feasible and remarkably effective method for analyzing clonality from FFPE specimens in histologically challenging cases. However, as a slight overestimation of clonal restrictions is possible, analysis repetition and careful patient follow-up are prudent. Proper clinical comparison with NGS technologies [36] is now certainly warranted.

Acknowledgments

The authors are grateful to Dr. Francesco Bacci, Dr. Elena Sabattini, and Dr. Carlo Sagramoso-Sacchetti for their contribution to diagnostic assessments as well as to Prof. Stefano A. Pileri for the precious advice.

The work reported in this publication was funded by the Italian Ministry of Health, RC-2023-2778976 (Prof. Piccaluga) and FIRB Futura 2011 RBFR12D1CB.

Conflict of interest

“The authors declare no conflict of interest.”

Note

Prof. Pier Paolo Piccaluga is currently affiliated with the Jomo Kenyatta University of Agriculture and Technology (Nairobi, Kenya), The University of Nairobi (Nairobi, Kenya), and the University of Botswana (Gaborone, Botswana).

Appendix

Case number	Gender	Site of biopsy	Clinical History
1	M	skin	Skin lesion NOS
2	M	skin	Skin lesion NOS
3	F	skin	Skin lesion NOS
4	F	skin	Skin lesion in pregressed NHL
5	M	skin	Skin lesion NOS
6	M	skin	Skin lesion NOS
7	M	skin	Erythema
8	F	skin	Dermatosis
9	M	skin	Psoriasis + adenopathies
10	M	skin	Skin lesion NOS
11	F	skin	Tumoral skin lesion
12	F	skin	Hashimoto thyroiditis
13	M	skin	Abdominal erythema
14	M	skin	Skin lesion NOS
15	M	skin	Skin nodules at lower limbs
16	F	skin	Skin lesion NOS
17	M	skin	Skin lesion NOS
18	M	skin	Skin lesion NOS
19	M	skin	Suspected Lymphomatoid Papulosis
20	M	skin	Skin lesion NOS
21	F	skin	Scalp erythema
22	M	skin	Skin lesion in pregressed NHL
23	F	skin	Skin lesion NOS
24	F	skin	Skin lesion NOS
25	F	skin	Skin lesion NOS
26	F	skin	Skin lesion NOS
27	M	skin	Skin lesion NOS
28	M	skin	Skin lesion NOS
29	F	skin	Erythema
30	M	skin	Erythema
31	M	skin	Papulo-nodular skin lesions
32	M	skin	Suspected lymphoma
33	F	skin	Cutaneous pseudolymphoma
34	F	skin	Plaque
35	F	skin	Skin lesion NOS
36	M	skin	Mycosis fungoides
37	F	skin	Erythrodermic lesion
38	M	skin	Skin lesion NOS
39	M	skin	Skin lesion NOS
40	F	skin	Skin lesion NOS
41	M	skin	Skin lesion NOS
42	F	skin	Skin lesion NOS
43	M	skin	Skin lesion NOS
44	F	skin	Skin lesion NOS
45	M	skin	Erythema
46	M	skin	Lower limb ulcer
47	F	skin	Skin lesion NOS
48	F	skin	Suspected insect bite
49	M	skin	Skin lesion NOS
50	F	skin	Skin lesion NOS
51	M	skin	Skin lesion NOS
52	M	skin	Plaque
53	M	skin	Skin lesion + adenopathy
54	M	skin	Plaque
55	M	skin	Skin lesion NOS
56	F	skin	Skin lesion NOS
57	F	skin	Skin lesion NOS
58	F	skin	Tumoral skin lesion
59	F	skin	Skin lesion NOS
60	F	skin	Mycosis fungoides
61	M	skin	Skin lesions
62	M	skin	Tumoral skin lesion
63	M	skin	Mycosis fungoides
64	M	skin	Skin lesion NOS
65	M	skin	Skin lesion NOS
66	F	skin	HIV+ + skin lesions
67	M	skin	Skin lesion NOS
68	F	skin	Skin lesion NOS
69	M	skin	Skin lesion NOS
70	F	skin	Tumoral skin lesion
71	M	skin	Skin lesion NOS
72	M	skin	Skin lesion NOS
73	M	skin	Skin lesion NOS
74	M	skin	Skin lesion NOS
75	M	skin	Skin lesion NOS
76	M	skin	Skin lesion NOS
77	F	skin	Skin lesion NOS
78	F	skin	Skin lesion NOS
79	M	skin	Skin lesion NOS
80	F	skin	Tumoral skin lesion
81	F	skin	Skin lesion NOS
82	M	skin	Pregressed Cutaneous PTCL
83	F	skin	Mycosis fungoides
84	M	skin	Skin lesion NOS
85	F	skin	Plaque
86	M	skin	Skin lesion NOS
87	F	skin	Diffuse pruritis
88	F	skin	Skin lesion NOS
89	M	skin	Erythema + axillary adenopathy
90	M	skin	Skin lesion NOS
91	M	skin	Skin lesion NOS
92	F	skin	Skin nodules
93	M	skin	Skin lesion NOS
94	M	skin	Plaque
95	F	skin	Skin lesion NOS
96	M	skin	Papular skin lesions
97	F	skin	Skin lesion NOS
98	M	skin	Skin nodules
99	F	skin	Skin lesion NOS
100	M	skin	Skin lesion NOS
101	M	skin	HIV+, skin lesions, adenopathies
102	M	skin	Suspected follicular mucinosis
103	M	skin	Skin lesion NOS
104	M	skin	Erythema and pruritis
105	M	skin	Plaque
106	F	skin	Dermatosis
107	F	skin	Skin lesion NOS
108	M	skin	Skin lesion NOS
109	F	skin	Skin lesion NOS
110	M	skin	Plaque

Table A1.

Patients’ characteristics.

NOS = not otherwise specified.

Section and Topic	Item #		On page #
TITLE/ABSTRACT/KEYWORDS	1	Identify the article as a study of diagnostic accuracy	2
INTRODUCTION	2	State the research questions or study aims, such as estimating diagnostic accuracy or comparing accuracy between tests or across participant groups.	4
METHODS
Participants	3	Describe the study population: The inclusion and exclusion criteria, setting and locations where the data were collected.	5
	4	Describe participant recruitment: Was recruitment based on presenting symptoms, results from previous tests, or the fact that the participants had received the index tests or the reference standard?	Consecutive patients who received the index test (Figure 1)
	5	Describe participant sampling: Was the study population a consecutive series of participants defined by the selection criteria in items 3 and 4? If not, specify how participants were further selected.	Consecutive cases
	6	Describe data collection: Was data collection planned before the index test and reference standard were performed (prospective study) or after (retrospective study)?	Restrospective study
Test methods	7	Describe the reference standard and its rationale.	5
	8	Describe technical specifications of material and methods involved including how and when measurements were taken, and/or cite references for index tests and reference standard.	5-7
	9	Describe definition of and rationale for the units, cutoffs and/or categories of the results of the index tests and the reference standard.	5-7
	10	Describe the number, training and expertise of the persons executing and reading the index tests and the reference standard.	5-6
	11	Describe whether or not the readers of the index tests and reference standard were blind (masked) to the results of the other test and describe any other clinical information available to the readers.	5-7
Statistical methods	12	Describe methods for calculating or comparing measures of diagnostic accuracy, and the statistical methods used to quantify uncertainty (e.g. 95% confidence intervals).	5-6
	13	Describe methods for calculating test reproducibility, if done.	6-7
RESULTS
Participants	14	Report when study was done, including beginning and ending dates of recruitment.	5
	15	Report clinical and demographic characteristics of the study population (e.g. age, sex, spectrum of presenting symptoms, comorbidity, current treatments, recruitment centers).	Table A1
	16	Report the number of participants satisfying the criteria for inclusion that did or did not undergo the index tests and/or the reference standard; describe why participants failed to receive either test (a flow diagram is strongly recommended).	8
Test results	17	Report time interval from the index tests to the reference standard, and any treatment administered between.	6
	18	Report distribution of severity of disease (define criteria) in those with the target condition; other diagnoses in participants without the target condition.	8
	19	Report a cross tabulation of the results of the index tests (including indeterminate and missing results) by the results of the reference standard; for continuous results, the distribution of the test results by the results of the reference standard.	Table A3
	20	Report any adverse events from performing the index tests or the reference standard.	/
Estimates	21	Report estimates of diagnostic accuracy and measures of statistical uncertainty (e.g. 95% confidence intervals).	Tables 1 and 2
	22	Report how indeterminate results, missing responses and outliers of the index tests were handled.	6
	23	Report estimates of variability of diagnostic accuracy between subgroups of participants, readers or centers, if done.	/
	24	Report estimates of test reproducibility, if done.	/
DISCUSSION	25	Discuss the clinical applicability of the study findings.	8-13

Table A2.

STARD requirements for evidence based medicine diagnostic accuracy studies.

Case number	Gold standard	Index test: TCRG analysis
1	Inflammatory disorder	Polyclonal
2	MF/SS	Clonal
3	Inflammatory disorder	Clonal
4	Inflammatory disorder	Polyclonal
5	MF/SS	Clonal
6	Inflammatory disorder	Clonal
7	Inflammatory disorder	Polyclonal
8	MF/SS	Clonal
9	Inflammatory disorder	Oligoclonal
11	Inflammatory disorder	Polyclonal
12	non-diagnostic	Polyclonal
13	Inflammatory disorder	Polyclonal
14	PTCL/NOS	Clonal
16	B-NHL	Polyclonal
17	B-NHL	Clonal
18	MF/SS	Clonal
19	non-diagnostic	Clonal
20	PTCL/NOS	Clonal
21	Inflammatory disorder	Polyclonal
22	non-diagnostic	Polyclonal
24	Inflammatory disorder	Clonal
25	non-diagnostic	Clonal
26	Inflammatory disorder	Clonal
27	MF/SS	Clonal
28	non-diagnostic	Clonal
30	MF/SS	Clonal
32	B-NHL	Clonal
33	Inflammatory disorder	Polyclonal
34	Inflammatory disorder	Polyclonal
35	PTCL/NOS	Clonal
37	MF/SS	Clonal
38	Inflammatory disorder	Polyclonal
39	MF/SS	Clonal
40	PTCL/NOS	Clonal
41	B-NHL	Oligoclonal
42	PTCL/NOS	Clonal
43	Inflammatory disorder	Polyclonal
44	Inflammatory disorder	Polyclonal
45	PTCL/NOS	Clonal
46	T-LGL	Clonal
47	Inflammatory disorder	Polyclonal
48	Inflammatory disorder	Polyclonal
50	CTCL/NOS	Clonal
51	Inflammatory disorder	Polyclonal
52	MF/SS	Polyclonal
53	PTCL/NOS	Clonal
54	MF/SS	Clonal
55	non-diagnostic	Polyclonal
56	MF/SS	Clonal
57	Inflammatory disorder	Polyclonal
58	CTCL/NOS	Clonal
59	non-diagnostic	Oligoclonal
60	PTCL/NOS	Clonal
61	MF/SS	Polyclonal
62	Inflammatory disorder	Polyclonal
64	MF/SS	Clonal
65	non-diagnostic	Polyclonal
67	non-diagnostic	Polyclonal
68	non-diagnostic	Oligoclonal
69	ALK neg ALCL	Clonal
70	CTCL/NOS	Clonal
71	PTCL/NOS	Clonal
72	non-diagnostic	Clonal
73	Inflammatory disorder	Polyclonal
74	Inflammatory disorder	Polyclonal
75	Inflammatory disorder	Polyclonal
76	MF/SS	Clonal
77	Inflammatory disorder	Polyclonal
78	Inflammatory disorder	Polyclonal
79	B-NHL	Polyclonal
80	CTCL/NOS	Clonal
81	PTCL/NOS	Polyclonal
82	MF/SS	Clonal
83	Inflammatory disorder	Polyclonal
85	Inflammatory disorder	Polyclonal
86	PTCL/NOS	Polyclonal
87	Inflammatory disorder	Polyclonal
88	Inflammatory disorder	Polyclonal
91	Inflammatory disorder	Polyclonal
92	PTCL/NOS	Clonal
93	B-NHL	Oligoclonal
94	MF/SS	Clonal
95	non-diagnostic	Oligoclonal
96	Inflammatory disorder	Polyclonal
98	MF/SS	Clonal
99	PTCL/NOS	Clonal
101	PTCL/NOS	Clonal
102	MF/SS	Clonal
103	PTCL/NOS	Clonal
104	Inflammatory disorder	Polyclonal
105	B-NHL	Polyclonal
106	Inflammatory disorder	Polyclonal
107	CTCL/NOS	Clonal
108	Inflammatory disorder	Polyclonal
109	Inflammatory disorder	Polyclonal
110	B-NHL	Polyclonal

Table A3.

Comparison of Gold standard (histopathological diagnosis) and Index Test (TRG clonality testing).

B-NHL = B-non Hodgkin lymphoma; CTCL = cutaneus T-ceel lymphoma; DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; LGL = large granular lymphocyte leukemia; MF = mycosiss fungoides; MZL = marginal zone lymphoma; NOS = not otherwise specified; PTCL = peripheral T-cell lymphoma; SS = Sezary syndrome.