Cancer Cytopathology

FISH detection of t(14;18) in follicular lymphoma on Papanicolaou-stained archival cytology slides

Authors


Abstract

BACKGROUND

The t(14;18)(q32;q21) translocation is present in about 85% of follicular lymphomas (FL) and can be identified using fluorescence in situ hybridization (FISH). In the diagnostic laboratory setting, the cytologic archival material consists of stained slides, and only rarely is material saved for molecular testing. The authors proposed FISH for FL using Papanicolaou-stained archival cytology material as a practical ancillary technique for diagnosing FL.

METHODS

Cases included 35 FL, 6 small lymphocytic lymphomas/chronic lymphocytic leukemias (SLL/CLL), 4 mantle cell lymphomas (MCL), 4 marginal zone lymphomas (MZL), 1 lymphoplasmacytic lymphoma (LPL), and 10 reactive lymphoid tissues (RLT). FISH was performed on Papanicolaou-stained archival cytology slides using probes for immunoglobulin heavy chain (IGH) on chromosome 14 and BCL2 on chromosome 18.

RESULTS

In all, 25 of 32 (81%) FL cases exhibited the t(14;18) translocation, whereas 7 of 32 (19%) lacked the translocation. No cases of non-FL were positive for t(14;18). This series shows a sensitivity of 81% and specificity of 100% for detecting the t(14;18) translocation as a diagnostic tool in FL.

CONCLUSIONS

When performed on Papanicolaou-stained cytology slides, FISH for t(14;18) is relatively sensitive and quite specific for FL. These findings are similar to those reported on other specimens, such as paraffin-embedded tissue and unstained cytology slides. The authors proposed that their technique would allow the pathologist and clinician the flexibility to utilize previously stained fine-needle aspiration slides for FISH evaluation. Cancer (Cancer Cytopathol) 2006. © 2006 American Cancer Society.

Follicular lymphoma (FL) is characterized by the translocation t(14;18)(q32;q21) and molecular rearrangement of the immunoglobulin heavy chain (IGH) gene on chromosome 14 with the BCL2 gene on chromosome 18. This translocation juxtaposes these 2 genes, with BCL2 coming under the control of the IGH promotor. This leads to overexpression of the BCL2 protein, which is normally involved in prevention of apoptosis. The t(14;18) translocation is detected in approximately 85% of FL and the identification of this genetic aberration can be useful in confirming the diagnosis.1 This translocation may be detected using a variety of techniques, including Southern blot analysis, conventional cytogenetics, polymerase chain reaction (PCR), or fluorescence in situ hybridization (FISH).1–4 The hematopathologist typically utilizes a combination of morphologic examination with or without immunohistochemistry, flow cytometry, and molecular identification of t(14;18) to reach a diagnosis of FL. Depending on the difficulty of the individual case, the amount and quality of material, the practice environment, and the clinical scenario, 1 or more ancillary techniques may be employed to aid in making the diagnosis. In our institution, we have found that the need to perform ancillary techniques on a cytology case may not arise until after the specimen is completely exhausted. In such cases, if the morphologic and/or flow cytometric features are less than diagnostic, additional material must be obtained from the patient.

FISH is a well-established molecular technique that can identify specific genetic aberrations in the context of cytomorphology.5 A wide variety of probe kits are available and have become more cost effective over time. In addition, automated electronic counting platforms have been utilized, making FISH a practical option in the diagnostic laboratory.6–9 Cytologic preparations are ideal for FISH examination due to the single-cell suspension offering optimum visualization of individual cells that enables clear signal detection and straightforward counting. With adequate spacing between cells, morphologic features are better appreciated, which is an advantage of FISH over other molecular techniques.

It has been shown that FISH can be effectively performed on cytologic material.10–17 Studies performing FISH for FL-associated t(14;18) translocation on exclusively Papanicolaou-stained cytology slides have not previously been performed. We feel that this is a practical approach, given that most cytology slides are stained with either Papanicolaou (Pap) or Giemsa for cytological evaluation. We chose to perform FISH for t(14;18) on Pap-stained cytology slides of FL, other non-Hodgkin lymphomas (NHL), and normal lymphoid tissue in an attempt to establish this method as a sensitive and specific ancillary technique in the diagnosis of FL.

MATERIALS AND METHODS

From the cytology files, 88 fine-needle aspiration (FNA) cases of NHL and 19 cases of benign reactive lymphoid tissue (RLT) were obtained and the slides were reviewed for diagnosis, quality of cellular material, amount of material on each slide, and number of slides in each case. Cases containing a moderate amount of diagnostic material on 2 or more Pap-stained slides were selected for study. For each case that met inclusion criteria, a single slide was selected for FISH analysis. In total, 60 ethanol-fixed, Pap-stained cytology slides were selected for study, including 35 cases of FL, 6 cases of small lymphocytic lymphoma/chronic lymphocytic leukemia (SLL/CLL), 4 cases of mantle cell lymphoma (MCL), 4 cases of marginal zone lymphoma (MZL), 1 case of lymphoplasmacytic lymphoma (LPL), and 10 cases of RLT. Patient records were searched for surgical pathology excision diagnosis and flow cytometry data for the corresponding lymph node.

Target areas on the Pap-stained slides were marked with a diamond-tipped pen on the underside of each slide. Coverslips were removed by slide immersion in xylene for 36 hours and slides were cleared of xylene by 3 10-minute 100% ethanol washes. Air-dried slides were pretreated with 50 μL (100 μg/mL) ribonuclease A solution at 37°C for 1 hour to aid nonspecific signal reduction. Slides were washed in 2× saline sodium citrate (SSC) for 5 minutes, dehydrated through an ascending ethanol series at 1-minute intervals, and allowed to air-dry for 20 minutes. After immersion in 2× SSC at 73°C for 2 minutes slides were treated with 0.1 mg/mL pepsin / 0.85% NaCl, pH 2.0, at 37°C for 10 minutes. Slides were rinsed in phosphate-buffered saline (PBS), postfixed in 1% paraformaldehyde for 5 minutes, rinsed for 5 minutes with PBS, dehydrated through serial ethanol solutions, and left to air-dry for 20 minutes.

The LSI IGH/BCL2 Dual Color, Dual Fusion Translocation Probe (Vysis, Downers Grove, IL), designed to detect the juxtaposition of immunoglobulin heavy chain (IGH) locus and BCL gene sequences, was used for this study. This probe is a mixture of the LSI IGH probe (green), spanning approximately 1.5 Mb and containing sequences homologous to essentially the entire IGH locus, as well as sequences extending about 300 kb beyond the 3′ end of the IGH locus. The LSI BCL2 probe (orange) covers an approximate 750-kb region, including the entire BCL2 gene, with additional sequences extending approximately 250 kb both distal and proximal to the gene. Using this probe the expected pattern in a normal cell would be 2 orange and 2 green signals. In a cell with t(14;18), one would expect 1 orange signal, 1 green signal, and 2 orange-green (yellow) fusion signals representing the reciprocal translocation. According to the manufacturer, patterns other than this may be observed in some abnormal cells, including instances of nuclei containing more than 2 fusion signals, as seen in this study (see Results).

Four microliters of the LSI IGH/BCL2 Dual Color, Dual Fusion Translocation Probe were applied to target tissues, overlaid with a 22 × 22 mm glass coverslip, and sealed with rubber cement. This probe was chosen to view the presence of a translocation involving IGH at 14q32 and BCL2 at 18q21, t(14;18)(q32;q21). Slides were placed in a HYBrite Denaturation and Hybridization System (Vysis) programmed at 73°C, for 2 minutes (to denaturation DNA) followed by 12 hours at 37°C (for hybridization). The rubber cement was removed from the slides and coverslips were floated off in 2XSSC/0.1%NP-40. Stringency treatment consisted of washes in 2XSSC/0.3%NP-40 buffer at 73°C. All posthybridization steps were conducted in a light-protected environment to avoid photobleaching of the fluorescent probe.

After air drying, slides were counterstained with 10 μL DAPI (Vysis) and viewed with an Olympus BX50 fluorescent microscope using 3 filters: red excitation 568, filter cube emission 605DF32, green excitation 488, filter cube emission 522DF32, and UV excitation 330, filter cube emission 420DF32. Images were captured using an Olympus U-PMTVC digital camera using Magnafire software.

The FISH slides were examined on the fluorescent microscope and 50 cells were counted on each slide. Cells were categorized as positive for t(14;18), normal, noninterpretable, or “other pattern,” as described below. Cells were counted from at least 5 different 1000× fields on each slide. All cells showing a fluorescent signal were counted, including cells with normal IGH and BCL2 genes, cells with t(14;18) translocation, and cells with other signal patterns (e.g., multiple copies of the chromosome 14 or 18 locus). Cases with distinct signals for IGH and BCL2, without evidence of translocation, were regarded as normal. Cases showing only 1 cell with t(14;18) were also regarded as “normal,” and cases with 2 or more cells harboring t(14;18) were classified as “positive” (see Discussion). Cells containing recognizable trisomies or additional copies of chromosome 14 and/or 18 were assigned to the “other pattern” category. Cells with unrecognizable signal patterns were regarded as “noninterpretable.” Such cells showed haphazardly arranged fluorescence, making visualization of a distinct signal(s) impossible. The “noninterpretable” cells were counted so as to reflect the number of such cells one may encounter when performing this method in a diagnostic laboratory.

RESULTS

Fifty-nine out of 60 cases (98%) had either a subsequent surgical excision with surgical pathology (27%) and/or flow cytometry (88%) of the corresponding lymph node (Table 1). Among exclusively follicular lymphoma cases, 35 out of 35 (100%) had a supportive diagnosis by surgical pathology (34%) and/or flow cytometry (91%) on the corresponding lymph node. Among the 60 cases studied with FISH, 9 failed to produce adequate signal for counting, including 3 cases of FL and 6 cases of other NHL and RLT. Among the remaining cases, 26 of 32 cases of FL (81%) were positive for the t(14;18) translocation (Fig. 1), and the remaining 6 cases (19%) were negative for the translocation (Table 1). Among the 26 FISH-positive cases, the average number of cells containing a t(14;18) translocation was 18.6, with a range of 2 to 44 (data not shown). Among the 20 cases of non-FL NHL and RLT that were evaluated, no cases harbored the t(14;18) translocation (Fig. 2). This series yields a sensitivity of 81% and a specificity of 100% for FISH detection of t(14;18) as a diagnostic tool in cases of FL using Pap-stained archival cytology slides. Twenty-two out of 51 cases (43%) contained a few cells with a noninterpretable signal pattern, with a range of 1-4 cells per 50. Twelve out of 51 cases (24%) contained cells with identifiable numerical aberrations (such as trisomy 14 or trisomy 18).

Table 1. Pathologic Diagnosis, FISH Interpretations Using Subjective Screening (Noncounting) and Counting Methods, and Cell Counting Data for Each Case; Numbers Indicate Number of Cells Out of 50 Counted for Each Case
Case no.Cytopathologic diagnosisSurgical pathology diagnosisFlow cytometryFISH interpretationt(14;18)NormalNoninterpretableOther patterns*
  • FISH: fluorescence in situ hybridization; FL: follicular lymphoma; MAR: marginal zone lymphoma; MAN: mantle cell lymphoma; SLL/CLL: small lymphocytic lymphoma/chronic lymphocytic lymphoma; LPL: lymphoplasmocytic lymphoma; RLT: reactive lymphoid tissue.

  • *

    Other patterns included trisomy 14, trisomy 18, and multiple copies of chromosomes 14 and/or 18.

  • FISH signal was present for subjective screening, but signal subsequently became absent or too faint for counting to be performed.

1FLNACD20/CD10Positive341501
2FLFLNAPositive222440
3FLNACD20/CD10Positive2413211
4RLTNANon-DxNegative04910
5SLL/CLLNANANegative04910
6SLL/CLLNACD19/CD5/CD23Negative05000
7FLNACD20/CD10Positive162509
8MANNACD20+/CD5+/CD10−/CD23−Negative04901
9FLNACD20/CD10Negative05000
10RLTRLNNon-DxNo signalNANANANA
11MANNACD20+/CD5+/CD10−/CD23−Broken slideNANANANA
12FLFLCD20/CD10Positive123701
13FLNACD20/CD10Positive24800
14SLL/CLLNACD19/CD5/CD23Negative05000
15FLNACD20/CD10Positive84200
16MARNACD20+/CD5−/ CD23−Negative04802
17FLNACD20/C10Positive143510
18FLFLCD20/CD10Negative04811
19FLFLCD20/CD10+−Positive163310
20RLTNANon-DxNegative04910
21FLNACD20/CD10Negative114035
22RLTNANon-DxNegative05000
23MARNACD20+/CD5−/ CD23−Negative05000
24FLNACD20/CD10Positive93812
25FLFLNAPositive84011
26MARNACD20+/CD5−/ CD23−No signalNANANANA
27RLTNANon-DxNegative05000
28RLTNANon-DxNegative05000
29FLNACD20/CD10Positive242600
30FLNACD20/CD10Positive271535
31RLTRLNNANegative05000
32RLTNANon-DxNegative05000
33RLTNANon-DxNegative05000
34RLTNANon-DxNegative05000
35FLNACD20/CD10Positive133520
36FLNACD20/CD10Positive261932
37FLNACD20/CD10No signalNANANANA
38FLFLCD20/CD10Positive212621
39LPLDLBCL transformed from LPLCD20Negative05000
40FLNACD20/CD10Positive44510
41FLNACD20/CD10No signalNANANANA
42FLFLCD20/CD10Positive93623
43FLFLCD20/CD10Positive391010
45FLNACD20/CD10Negative04523
46FLNACD20/CD10Positive183200
48FLNACD20/CD10Positive162716
49FLNACD20/CD10Positive133304
50FLFLCD20/CD10Positive291911
51FLFLNAPositive222611
52FLNACD20/CD10Positive271841
53FLFLNon-DxNegative04901
47FLFLNAWeak signalNANANANA
54FLNACD20/CD10No signalNANANANA
55FLNACD20/CD10No signalNANANANA
56MARNACD20+/CD5−/ CD23−Weak signalNANANANA
57SLL/CLLNACD20/CD5/CD23Negative05000
58MANMANNANegative05000
59SLL/CLLNACD20/CD5/CD23Negative05000
60SLL/CLLNACD20/CD5/CD23Negative05000
Figure 1.

(A) Medium-power view of several normal lymphocytes in a case of reactive lymphoid tissue. (B) High-power view of a normal lymphocyte in a case of reactive lymphoid tissue, showing 2 spatially distinct signals for the IGH 14q32 probe (green) and BCL2 18q21 probe (red). Original magnification ×400 (A); ×1000 (B).

Figure 2.

(A) Medium-power view of several malignant lymphocytes in a case of follicular lymphoma. (B) High-power view of a malignant lymphocyte in a case of follicular lymphoma, showing the t(14;18) translocation, characterized by the juxtaposition of a green IGH 14q32 signal with a red BCL2 18q21 signal. Original magnification × 400 (A); ×1000 (B).

DISCUSSION

The presence of t(14;18) in the malignant lymphocytes of FL, the juxtaposition of the IHC and BCL2 genes, the resulting overexpression of BCL2, and the subsequent proliferation of clonal lymphocytes, provides a pathophysiologic mechanism for FL.1 Currently, with the aid of molecular genetic techniques such as FISH, this chromosomal rearrangement can be exploited for the purposes of aiding in the diagnosis of FL. Performing FISH on cytologic material has been described,10–17 but studies have not been previously performed evaluating t(14;18) on Papanicolaou-stained cytology slides of FL. We felt that showing the technical feasibility of this technique would be a helpful contribution to diagnostic cytopathology and molecular cytogenetics, and could especially be useful in diagnostically challenging cases where flow cytometry is not available. Although we performed this study in the context of lymphoid lesions, we would expect this technique to be applicable in other diseases that are 1) amenable to cytologic examination, and 2) contain recurrent genetic aberrations associated with the lesion.

Our study shows that FISH for t(14;18), performed on Pap-stained cytology material, is 81% sensitive and 100% specific for FL. Our rates are similar to those seen in FL using paraffin-embedded tissue and unstained cytology material.2, 3, 10, 11, 18–22 Considering that the pathogenetic mechanism for FL involves t(14;18) and the corresponding overexpression of BCL2, it is interesting that the translocation is not detected in approximately 10% to 20% of FL cases in many studies employing a range of diagnostic techniques, including standard cytogenetics, molecular genetics, and FISH. In view of this, the question is raised of how many cells need to be counted to confidently exclude a diagnosis of FL? This issue is not limited to the context of lymphoma, but is also applicable to other chromosomally aberrant lesions. The malignant clonal proliferation may be outnumbered by the background polyclonal population, yielding a low percentage of translocation-positive cells on FISH analysis. Sampling is also an issue. Depending on where within the lesion the needle is placed, variable amounts of malignant cells may be recovered. In selecting a reasonable cell-number threshold for calling a FISH case “positive,” a balance must be struck between the possibility of a small malignant clone and the appearance of a translocation in benign cells.

The t(14;18) translocation has been reported in reactive lymphoid tissue.23, 24 Disease-associated translocations are those that occur in a specific sequence, becoming incorporated into a proliferating clone.1 A single lymphocyte may develop a t(14;18) translocation, but this may not reflect malignancy. Therefore, one must interpret a translocation present in extremely low levels with caution. In addition, random overlap of fluorescent probe could give the appearance of a translocation. In this series, we regarded cases showing only 1 translocated cell as “negative” for FL. Among our cases, there was 1 case in which t(14;18) was identified in only 1 cell (Case 21, Table 1). This case was diagnosed as FL by cytology and flow cytometry; however, the FISH findings were unusual, with 35 cells showing various combinations of trisomy 18 and trisomy 14 (see below). The authors felt that this case, with its unusual FISH findings, was supportive evidence for interpreting cases with a fusion signal in 1 out of 50 cells as FISH-“negative.” Another case of FL (Case 13, Table 1), diagnosed by cytology and flow cytometry, was found to have 2 cells with t(14;18). The remaining 48 cells counted in this case showed normal chromosome 14 and 18 probe signals, and lacked any noninterpretable or other identifiable signal patterns. Based on these cases, we set our threshold for classifying cases as “positive” for t(14;18) at a minimum of 2 cells containing a fusion signal. Noteworthy was the fact that we did not see any non-FL cases with t(14;18)-positive cells in this series, further validating our selected threshold value.

The importance of cytomorphology is critically important in considering a positive FISH result. The t(14;18) translocation is not specific to FL, with t(14;18) known to occur in about 15% of diffuse large B-cell lymphomas (DLBCL).1 Because of this, DLBCL could cause a false-positive when doing FISH for FL. Therefore, clinicopathological correlation is crucial in interpreting FISH analysis in NHL. The broad category of lymphomas considered in this study included those of the small B-cell type, and DLBCL cases were therefore excluded. One situation where there may be considerable morphological and FISH overlap is in the differential diagnosis of DLBCL and Grade III FL. This distinction, although a possible use for t(14;18), was not evaluated in this study.

In an attempt to provide practical information for those wishing to utilize this method in a diagnostic laboratory, we chose not to limit our counting to cells with a t(14;18) fusion signal and cells with normal signals, but also those showing other signal patterns. The presence of “noninterpretable” cells did not preclude the ability to reach a conclusion in our cases, with a low number of noninterpretable cells seen in some cases (43%, range, 1-4 cells). Identifiable numerical aberrations, such as trisomy 14 or 18, was seen in 24% of cases, but usually at a low level that did not interfere with the interpretation. One case of FL (Case 21, Table 1), however, contained 35 out of 50 cells with numerical aberrations, including 27 trisomy cells and 8 cells with both trisomy 18 and trisomy 14 (data not shown). Although interesting, an explanation for this finding is unclear in this case.

From a FISH interpretation standpoint, the issue of slide heterogeneity is also important to consider when making FISH observations. When observing or counting cells on a slide, we found it useful to count from several different areas. For example, in this study the observer counted no more than 10-12 cells from a single area of the slide. By doing this, the possibility of counting only from an area overrepresented by 1 cell type, benign or malignant, will be minimized. This method of using the whole slide gives the observer a better view of the cell types represented on the slide and will provide a more accurate assessment of the case.

In a diagnostic laboratory, the proposed method could be useful in some cases where the other ancillary techniques are unavailable or noncontributory. Ideally, an unstained slide could be made onsite at the FNA procedure for possible FISH analysis if necessary. However, that was not done in this study. At our institution an onsite cytopathologist is generally present to provide an indication of adequacy and a rapid interpretation when possible. Perhaps this would be an appropriate time to decide whether to prepare a slide for FISH. However, we have found that most of the onsite effort goes into obtaining optimal material for cytopathology and flow cytometry and making extra blank slides is often not feasible when material and time are limited. This underscores the practicality of utilizing previously stained slides, as demonstrated in this study. Cytopathologists and cytotechnologists choosing this technique will not have to change their current mode of onsite operation. Rather, the request for FISH can be an afterthought, with minimal waste of valuable cellular material.

When performed on Pap-stained cytology slides, FISH for t(14;18) is relatively sensitive and quite specific for FL. This technique may be useful for challenging cytology cases, when the morphological features and/or immunophenotypic data are not fully diagnostic. We feel that Pap-stained cytology slides are comparable to other types of material, such as paraffin-embedded tissue, in demonstrating the presence of t(14;18) by FISH in FL. We propose that this technique will allow the pathologist and clinician the flexibility to utilize previously stained cytology slides for FISH evaluation, expanding the availability of material for molecular cytogenetic analysis.

Ancillary