2022 年 69 巻 3 号 p. 291-297
Preoperative flow cytometry is recommended to prove the monoclonality and confirm the diagnosis of thyroid lymphoma. However, lymphoma cases without light chain restriction may also have monoclonality. The aim of our study was to identify a novel marker for thyroid lymphomas using aspirated materials for flow cytometry. We retrospectively analyzed 26 patients with primary thyroid lymphomas and 16 patients with benign lymphoproliferative lesions. The materials for flow cytometry were obtained by fine-needle aspiration cytology using a 22-gauge needle under ultrasound guidance. Light chain restriction was defined as a κ to λ ratio of less than 0.5 or more than 3.0. According to the light chain–positive rate, 25% or less and more than 25% were classified as the low and high light chain–positive rate groups, respectively. B-cell predominance was defined as a CD19 to CD4 ratio (B- to T-cell ratio) of more than 2.0. B-cell predominance was more frequently observed in lymphomas (88.5%) than in benign lymphoproliferative lesions (25.0%; p < 0.001). Light chain restriction based on the κ/λ ratio was detected in 69.2% of lymphomas, but not in benign lymphoproliferative lesions. Among lymphomas belonging to the low light chain–positive rate group, 88.9% did not exhibit light chain restriction and B-cell predominance was present. In contrast, benign lymphoproliferative lesions with B-cell predominance were not detected in the low light chain–positive rate group. B-cell predominance was a useful indicator for diagnosing thyroid lymphoma in the low light chain–positive rate group without light chain restriction.
PRIMARY THYROID LYMPHOMA is an uncommon neoplastic disease, accounting for 0.55% of all thyroid malignancies [1-3]. Thyroid lymphoma typically occurs in older women [2, 4-7] and is frequently accompanied by chronic thyroiditis [8-10]. Nearly all cases of thyroid lymphoma are B-cell-derived non-Hodgkin’s lymphoma, mainly diffuse large B-cell lymphoma (DLBCL) and mucosa-associated lymphoid tissue lymphoma (MALTL) [1, 11]. DLBCL may be transformed from MALTL or can occur as a de novo disease [12-14].
Preoperative diagnosis of primary thyroid lymphoma can usually be made by ultrasound examination and fine-needle aspiration cytology. However, the diagnostic accuracy is not satisfactory [15]. The distinction between MALTL and chronic thyroiditis with extensive lymphocytic proliferation is often difficult [16]. Recently, the ancillary technique, flow cytometry using aspirated materials and the presence of light chain restriction as an indicator of lymphoma, has been recommended for lymphoma-suspected nodules [16, 17]. Among the three diagnostic methods, flow cytometry had the highest specificity (88.4%) and sensitivity (75.0%) [17], and the analysis using aspirated materials was as reliable as that using resected materials [18]. Some reports have described B-cell lymphoma cases without the presence of light chain restriction [17, 19, 20]. Flow cytometry has many parameters other than light chains. Therefore, we decided to analyze the flow cytometry data in more detail. The aim of our study was to identify a novel marker for thyroid lymphoma in flow cytometry using aspirated materials.
The study protocol was reviewed and approved by the Institutional Review Board of Kuma Hospital (reference #170706-3) and was in accordance with the 1964 Declaration of Helsinki and its amendments, or comparable ethical standards. All study participants provided informed consent. We reviewed 52 patients who underwent ultrasound, aspiration cytology, and flow cytometry using aspirated materials at Kuma Hospital from January to December 2019. Of the 52 patients, 27 underwent total thyroidectomy, lobectomy, or open surgical biopsy for confirmation of the diagnosis. The diagnoses included 26 primary thyroid lymphomas and 1 thymoma. The diagnosis of primary thyroid lymphoma was determined by histological and immunohistochemical examinations using antibodies for CD20, UCHL-1, Bcl-2, CD10, CD23, κ light chain, λ light chain, and cytokeratin AE1/AE3, as well as flow cytometry results based on CD45 and side scatter-based gating, G-banding chromosomal examination, and immunoglobulin heavy chain JH DNA rearrangement analysis. Primary thyroid lymphoma cases included 14 MALTLs, 9 DLBCLs, 2 follicular lymphomas (FL; 1 grade 1–2 and 1 grade 3b), and 1 B-cell lymphoma, not otherwise specified. The remaining 25 patients were followed up by ultrasound examination for more than 6 months without therapy. The clinical diagnoses included lymphocytic thyroiditis, MALTL, and methotrexate-related lymphoproliferative disorders (16, 6, and 3 patients, respectively). As none of the 16 patients with a clinical diagnosis of lymphocytic thyroiditis showed progression of the lesions during follow-up, they were regarded as benign lymphoproliferative lesions (BLLs) in this study. Thus, 26 patients with primary thyroid lymphoma and 16 patients with BLL were included in this study.
Clinical data were obtained from the patients’ medical records at Kuma Hospital. Fine-needle aspiration cytology was performed using a 22-gauge needle under ultrasound guidance. Immediately after the aspirates were obtained, the materials were placed into a cell preservation liquid (5 mL, H00, Nissui Pharmaceutical, Tokyo, Japan) [21] and transported to the flow cytometry laboratory (SRL, Tokyo, Japan). The samples were analyzed using the FACSCantoTM II Cell Analyzer (BD Biosciences, San Jose, CA, USA), with antibodies for CD45, T-cell markers (CD2, CD3, CD4, CD5, CD7, and CD8), B-cell markers (CD10, CD19, CD20, CD23, surface κ, and surface λ), and other markers (CD11c, CD16, CD25, CD30, CD34, and CD56). No cases were excluded from being examined due to low cell counts.
Light chain restriction was defined as a κ to λ ratio of less than 0.5, or more than 3.0, according to the proposal by Stacchini et al. [22]. We classified the patients into two groups according to the light chain–positive rate (LCPR). A rate of 25% or less was classified as the low LCPR group; more than 25% was classified as the high LCPR group. B-cell to T-cell ratio (B/T cell ratio) was defined as the ratio of CD19 and CD4 showing the most significant differences between BLLs and lymphomas among the B-cell and T-cell groups, respectively. B-cell predominance was defined as a B/T cell ratio of more than 2.0.
We determined the statistical significance of the data using Student’s t-test and Fisher’s exact probability test. A p-value <.05 was considered statistically significant.
Table 1 shows the results of flow cytometry using aspirated materials in 16 BLLs and 26 thyroid lymphomas.
| BLL | Lymphoma | (MALTL) | (DLBCL) | (Others) | p-value* | ||
|---|---|---|---|---|---|---|---|
| n | 16 | 26 | 14 | 9 | 3 | ||
| Numbers of Counted Cells (range) | 4,993.6 (562–10,609) |
7,063.4 (1,216–16,067) |
6,438.9 (1,834–15,757) |
7,305.7 (1,216–16,067) |
8,107.3 (4,058–12,165) |
0.0809 | |
| T-Cell Markers (range [%]) | CD2 | 56.5 (14.2–96.8) |
21.3 (1.7–87.7) |
20.2 (4.0–53.3) |
25.7 (1.7–87.7) |
13.2 (11.1–14.7) |
<0.0001 |
| CD3 | 50.6 (7.6–83.7) |
18.4 (0.6–88.3) |
17.4 (2.6–48.8) |
22.9 (0.6–88.3) |
9.6 (3.6–14.7) |
<0.0001 | |
| CD4 | 37.3 (9.1–57.1) |
13.1 (0.2–54.2) |
14.1 (2.3–40.4) |
13.8 (0.2–54.2) |
6.5 (2.2–10.0) |
<0.0001 | |
| CD5 | 53.0 (7.6–82.1) |
21.3 (0.8–91.1) |
21.5 (4.8–48.2) |
24.3 (0.8–91.1) |
11.9 (6.5–15.1) |
<0.0001 | |
| CD7 | 51.4 (5.2–81.6) |
18.4 (1.0–76.4) |
16.6 (4.6–42.1) |
23.7 (1.0–76.4) |
11.3 (3.3–18.7) |
<0.0001 | |
| CD8 | 16.2 (2.3–41.4) |
7.4 (0.5–30.1) |
6.2 (1.5–17.0) |
10.1 (0.5–30.1) |
4.6 (1.3–9.1) |
<0.01 | |
| B-Cell Markers (range [%]) | CD10 | 7.0 (0.6–32.6) |
17.2 (1.1–87.8) |
4.7 (1.1–24.8) |
31.7 (2.8–87.8) |
31.4 (6.5–52.4) |
0.0594 |
| CD19 | 46.6 (3.9–93.9) |
82.3 (15.9–98.6) |
83.5 (51.4–97.4) |
79.8 (15.9–98.6) |
84.4 (72.1–97.9) |
<0.0001 | |
| CD20 | 43.1 (4.4–91.8) |
72.2 (10.6–99.7) |
78.8 (36.2–97.4) |
64.6 (10.6–99.7) |
64.4 (32.0–96.5) |
<0.001 | |
| CD23 | 19.3 (2.5–40.1) |
9.1 (0.8–24.2) |
12.9 (8.1–24.2) |
4.1 (0.8–9.9) |
6.4 (1.0–15.4) |
<0.01 | |
| Other Markers (range [%]) | CD11c | 15.7 (4.3–60.5) |
23.5 (0.6–84.0) |
31.8 (8.0–84.0) |
16.0 (0.6–68.8) |
7.6 (2.9–13.6) |
0.208 |
| CD16 | 4.2 (0–24.4) |
1.4 (0–6.8) |
0.8 (0–3.9) |
1.9 (0.2–6.8) |
2.2 (1.3–3.2) |
0.0885 | |
| CD25 | 7.1 (1.1–18.7) |
9.1 (0.1–32.1) |
6.2 (1.9–18.4) |
13.2 (0.1–32.1) |
10.4 (3.6–16.3) |
0.293 | |
| CD30 | 2.4 (0–9.5) |
1.8 (0–5.6) |
2.6 (0.7–5.6) |
1.0 (0–2.2) |
0.9 (0.2–1.4) |
0.422 | |
| CD34 | 1.6 (0.1–10.2) |
0.6 (0.1–1.8) |
0.6 (0.1–1.8) |
0.5 (0.2–1.2) |
0.5 (0.2–0.7) |
0.170 | |
| CD56 | 12.4 (0.9–79.5) |
9.7 (0.4–92.7) |
2.4 (0.4–9.4) |
22.3 (0.5–92.7) |
5.9 (2.0–12.2) |
0.690 | |
BLL, benign lymphoproliferative lesion; MALTL, mucosa-associated lymphoid tissue lymphoma; DLBCL, diffuse large B-cell lymphoma.
* Statistical analysis between BLLs and lymphomas.
The number of counted lymphoid cells varied from 562 to 16,067. The mean numbers of counted cells in cases of BLLs and lymphomas were 4,993.6 and 7,076.6, respectively. There were no significant differences between the two groups or between histological lymphoma subtypes.
T-cell markersThe proportions of CD2, CD3, CD4, CD5, CD7, and CD8 were significantly higher in BLLs than in lymphomas. All of them showed p-values less than 0.01, and the lowest was CD4 (p = 0.0000075). There were no significant differences in histological lymphoma subtypes.
B-cell markersThe proportions of CD19 and CD20 were significantly higher in lymphomas than in BLLs (p = 0.0000969 and 0.000623, respectively). In contrast, CD23-positive cells were more frequently detected in BLLs than in lymphomas (p < 0.01). The proportion of CD10 was significantly higher in DLBCLs than in MALTLs (p < 0.01). In two of the nine DLBCLs, CD10-positive cells accounted for more than 80% of the counted cells. The proportions of CD10-positive cells in two FLs were 6.5% and 35.2%, respectively. CD23-positive cells (mean: 12.9%) counted in MALTLs were significantly higher than those in DLBCLs (4.1%; p < 0.0001).
B/T cell ratioIn lymphomas, the B/T cell ratio varied from 0.1 to 493.0 (mean: 33.1), and B-cell predominance was seen in 88.5% of them (85.6% and 88.9% in MALTLs and DLBCLs, respectively). On the other hand, the B/T cell ratio in BLLs varied from 0.1 to 9.6 (mean: 1.2), with B-cell predominance in 25.0%. The incidences of B-cell predominance in MALTLs and DLBCLs were significantly higher than that in BLLs (p < 0.01), and there were no significant differences in the histological lymphoma subtypes.
Light chains Light chain restrictionLight chain restriction based on the κ/λ ratio was detected in 69.2% of lymphomas but was not demonstrated in BLLs (Table 2). The incidence of light chain restriction in MALTLs (78.6%) was higher than that in DLBCLs (44.4%). Of the 18 lymphomas with light chain restriction, 16 were κ-light chain predominant (88.9%). The remaining two lymphomas (both DLBCLs) showed λ-light chain predominance. Immunohistochemical examinations using light chain antibodies failed to detect light chain restriction because there were no lymphoma cases with plasma cell differentiation.
| BLL (n = 16) | Lymphoma (n = 26) | MALTL (n = 14) | DLBCL (n = 9) | Others (n = 3) | p-value | |
|---|---|---|---|---|---|---|
| κ (range) % | 22.3 (5.6–71.4) | 45.9 (2.4–92.4) | 57.5 (2.4–89.7) | 25.5 (2.5–92.4) | 52.8 (22.8–70.4) | <0.01* |
| λ (range) % | 17.3 (5.6–31.4) | 13.7 (0.7–94.3) | 7.9 (3.0–13.6) | 24.9 (0.7–94.3) | 7.2 (1.2–18.6) | 0.479* |
| κ to λ ratio | ||||||
| ≤0.5 | 0 (0%) | 2 (7.7%) | 0 (0%) | 2 (22.2%) | 0 (0%) | 0.517** |
| ≥3 | 0 (0%) | 16 (61.5%) | 11 (78.6%) | 2 (22.2%) | 3 (100%) | <0.0001** |
| ≤0.5 or ≥3 | 0 (0%) | 18 (69.2%) | 11 (78.6%) | 4 (44.4%) | 3 (100%) | <0.0001** |
BLL, benign lymphoproliferative lesion; MALTL, mucosa-associated lymphoid tissue lymphoma; DLBCL, diffuse large B-cell lymphoma.
* Statistical analysis between BLLs and lymphomas using Student’s t-test.
** Statistical analysis between BLLs and lymphomas using Fisher’s exact test.
All 17 lymphomas belonging to the high LCPR group showed light chain restriction (Fig. 1, Fig. 2A). In contrast, light chain restriction was not detected in eight of the nine lymphomas in the low LCPR group (88.9%; Fig. 2B). B-cell predominance was detected in 88.2% and 88.9% of lymphomas classified to the high and low LCPR groups, respectively, and there was no significant difference between them. No BLLs belonging to the low LCPR group showed B-cell predominance.
κ to λ ratio and the light chain–positive rate in 16 benign lymphoproliferative lesions and 26 lymphomas of the thyroid.
Two patterns of flow cytometry in thyroid lymphomas. A: Light chain restriction is present in case with a high light chain–positive rate. B: Light chain restriction is not detected in case with low light chain–positive rate.
Table 3 shows the diagnostic accuracy of B-cell predominance and light chain restriction for lymphomas in the low and high LCPR groups. Light chain restriction for the high LCPR cases showed the highest diagnostic accuracy. The sensitivity, specificity, positive predictive value, and negative predictive value were 94.1%, 100%, 100%, and 92.9%, respectively. Light chain restriction for the low LCPR cases had exceedingly low sensitivity (11.1%) and negative predictive values (27.3%). On the other hand, sensitivity (88.9%) and negative predictive values (75.0%) of B-cell predominance for the low LCPR cases were higher than those of light chain restriction. One lymphoma case without B-cell predominance or light chain restriction was cytologically interpreted as chronic thyroiditis.
| Light Chain-Positive Rate | Indicator | Sensitivity | Specificity | Positive Predictive Value | Negative Predictive Value |
|---|---|---|---|---|---|
| Low <25% | B-cell predominance* | 88.9% (8/9) | 100% (3/3) | 100% (8/8) | 75.0% (3/4) |
| Light chain restriction** | 11.1% (1/9) | 100% (3/3) | 100% (1/1) | 27.3% (3/11) | |
| High ≥25% | B-cell predominance* | 88.2% (15/17) | 69.2% (9/13) | 78.9% (15/19) | 81.8% (9/11) |
| Light chain restriction** | 94.1% (16/17) | 100% (13/13) | 100% (16/16) | 92.9% (13/14) |
* Defined as CD19 to CD4 ratio (B-cell to T-cell ratio) greater than 2.
** Defined as κ to λ ratio less than or equal to 0.5 or greater than 3.
When it is difficult to distinguish between MALTLs and chronic thyroiditis, the Bethesda System for Reporting Thyroid Cytopathology 2nd Edition recommends performing a repeat fine-needle aspiration to obtain materials for flow cytometry, which can better characterize the lymphocyte population [16]. Some reports have described the usefulness of flow cytometry in distinguishing thyroid lymphoma or chronic thyroiditis [17, 18, 22-24]. Light chain restriction is generally used to determine lymphoma monoclonality, and the criteria vary according to previous reports [17, 18, 22, 23]. Stacchini et al. defined it as a κ/λ ratio greater than 3 or less than 0.5 [22], whereas Hirokawa et al. defined it as a κ/λ ratio greater than 3 or less than 0.33 [17]. However, its use is hardly widespread. In Asian countries, only 1 of 10 institutions with extensive experience in thyroid lymphoma routinely perform flow cytometry for lymphoma-suspected lesions [15].
In flow cytometry, it is important to obtain as many cells as possible in appropriate proportions. Zeppa et al. described that a gate of 5,000 cells was considered sufficient for phenotypization [15]. However, we showed that the lymphoid cells obtained from a single aspirate varied from 562 to 16,067 and none of our samples were insufficient for flow cytometry analysis. We demonstrated that a single aspirate can obtain sufficient material for flow cytometry in cases of chronic thyroiditis or lymphoma. The diagnostic accuracy was comparable to that obtained using resected materials [18].
Light chain restriction has been reported as a reliable indicator of lymphoma, and thyroid lymphomas with κ-light chain predominance were reportedly five times more common than those with λ-light chain predominance [17]. In the present study, light chain restriction was observed in 69.2% of lymphomas but was not detected in BLL lesions. As previous researchers have stated, most lymphomas with light chain restriction show κ-light chain predominance. Our 2 lymphomas with λ-light chain predominance were both DLBCLs, but the significance was not clear.
There are also reports that focus on parameters other than light chain restriction. CD10 is a marker of germinal center B cells and is expressed characteristically in FLs [25]. In our study, however, one of the two FLs (grade 3b) showed a low percentage of CD10-positive cells (6.5%). Two DLBCLs were CD10-positive. Eshoa et al. reported that 83% of grade 3 FLs showed negative or weak expression of CD10 [26]. Therefore, we believe that CD10 cannot be used as a marker for FLs. CD10-positive DLBCLs may be transformed from FLs [27]. CD23 is a marker of germinal center B-cells and follicular dendritic cell mesh works and is negative in germinal centers or germinal center-like lymphoid aggregations seen in MALTLs and DLBCLs [28, 29]. The decrease in CD23-positive cells in lymphomas seemed to be due to the destruction of the germinal center or the progression of follicular colonization [25]. In fact, CD23-positive counted cells in lymphomas tended to be lower than those of BLLs in the present study. In addition, the finding that CD23-positive counted cells in DLBCLs were lower than those in MALTLs was reasonable.
Because almost all thyroid lymphomas are of B-cell origin, the lymphoid cells counted by flow cytometry should be B-cell predominant. However, few studies have focused on T-cell or B-cell predominance in flow cytometry analysis [22, 24, 30], and no quantitative study has been conducted. To the best of our knowledge, no study has focused on the B/T cell ratio in the differential diagnosis between BLLs and lymphomas in the thyroid. We demonstrated that CD19 and CD4 had the most significant differences between BLLs and lymphomas among the B-cell and T-cell groups, respectively. Thus, we used CD19 and CD4 as markers to evaluate the B/T cell ratio. CD19 is a pan-B-cell marker [31] and CD4 is a helper T-cell marker [32]. Because thyroid lymphoma is mostly B-cell type, and T-helper type 1 cells (Th1) are predominant in chronic thyroiditis [33], the combination of the two markers can be considered significant. We demonstrated that B-cell predominance was significantly more common in lymphomas than in BLLs, regardless of whether the lymphomas were MALTLs or DLBCLs.
Interestingly, we noticed that most lymphomas with low LCPR did not show light chain restriction. Zeppa et al. reported that chronic thyroiditis with low LCPR exhibited light chain restriction [34]. In cases with low LCPR, the evaluation of the κ/λ ratio might not be an accurate assessment. In contrast, B-cell predominance was identified even in lymphomas with low LCPR. Based on our results, we propose a diagnostic algorithm using flow cytometry for evaluating primary thyroid lymphoma (Fig. 3). When we evaluate the flow cytometry, the first thing we have to notice is the LCPR. In cases with low LCPR, B-cell predominance indicates lymphoma. In contrast, in cases with high LCPR, the presence of light chain restriction is useful in the diagnosis of lymphoma.
Diagnostic algorithm of flow cytometry for evaluating primary thyroid lymphoma.
Finally, we should consider the pitfalls of flow cytometry using aspirated materials. Because most thyroid lymphomas occur in the background of chronic thyroiditis [8-10], aspiration from the area of chronic thyroiditis adjacent to the lymphoma can lead to false-negative results. In fact, we had one DLBCL case with a false-negative result, and it was cytologically reported to be chronic thyroiditis. It is necessary to check not only the results of flow cytometry, but also the results of ultrasound and cytology to make a comprehensive judgment [17]. As this algorithm assumes the presence of B-cell lymphoma [1, 11], it does not work for T-cell lymphoma or Hodgkin’s lymphoma, but their frequency of occurrence is very low [11]. In addition, it is unclear whether it is applicable to cases of extra-thyroidal lymphoma.
We would like to thank Editage (www.editage.jp) for English language editing.
We declare that there are no conflicts of interest that could be perceived as prejudicing the impartiality of the research reported.
