Evaluation of interphase fluorescence in situ hybridization for the t(14;18)(q32;q21) translocation in the diagnosis of follicular lymphoma on fine-needle aspirates

A comparison with flow cytometry immunophenotyping




Diagnosing lymphoproliferative disorders on fine-needle aspiration (FNA) can be challenging due to variable cellularity and lack of architecture. Ancillary studies often are required for diagnosis. Follicular lymphoma (FL) is characterized by a monoclonal B-cell proliferation with coexpression of CD19/CD10 and a t(14;18)(q32;q21) reciprocal translocation, resulting in the immunoglobulin heavy chain/BCL-2 fusion gene. These features also can be found, with much lower frequency, in diffuse large B-cell lymphoma (DLBCL) of follicle center cell origin. The objective of the current study was to compare the accuracy in detecting FL and DLBCL of follicle center cell origin by interphase fluorescence in situ hybridization (I-FISH) versus flow cytometry immunophenotyping (FCM) on FNAs.


Concurrent testing by FISH for t(14;18)(q32;q21) and FCM was performed on 84 FNAs, including 40 FLs and 44 non-FLs (de novo DLBCLs, mantle cell lymphomas, small lymphocytic lymphomas/chronic lymphocytic leukemias [SLLs/CLLs], small B-cell lymphomas, and reactive lymphoid hyperplasias). The final diagnosis was rendered based on the combined information from cytomorphology, FCM, FISH, immunocytochemical staining for Ki-67, monoclonality for κ and λ light chains, and, if available, corresponding tissue biopsy, cytogenetic analysis, and polymerase chain reaction analysis.


Among 40 FLs, FISH produced positive results for the t(14;18) translocation in 85.0%, negative results in 7.5%, and insufficient results in 7.5%; whereas, with FCM, 75% of cases exhibited a CD19-positive (CD19+)/CD10+ population (28 monoclonal, 2 nonclonal), 12.5% of cases exhibited a CD19+/CD10-negative population (3 monoclonal, 2 nonclonal), and 12.5% of cases were insufficient. All of nonclonal results from FCM and all of the insufficient results from FCM analysis exhibited unequivocal t(14;18) translocation by FISH. In contrast, the three negative results and the three insufficient results from FISH were monoclonal and CD19+/CD10+ on FCM. The results from FISH and FCM were concordant in 75% cases. Of 44 non-FLs, FISH produced positive results for the t(14;18) translocation in 5 DLBCLs and 2 SLLs/CLLs. The latter showed single fusion signals just above the cutoff level. All cases in the non-FL group that failed to show clonality or had insufficient results from FCM were DLBCLs. Among 17 DLBCLs, FISH detected a t(14;18) translocation in 29.4%, whereas FCM demonstrated a CD19+/CD10+ population in 23.5%.


I-FISH for the t(14;18)(q32;q21) translocation provided high overall accuracy in detecting FLs on FNAs. This test can be used for diagnosing or monitoring FL on FNAs when cellularity is limited or when FCM results are noncontributory. For detecting a follicle center cell origin in DLBCLs, I-FISH for the t(14;18) translocation appeared to be slightly more sensitive than FCM for the CD19+/CD10+ immunophenotype. Cancer (Cancer Cytopathol) 2003;99:385–93. © 2003 American Cancer Society.

Follicular lymphoma (FL) is a common type of non-Hodgkin lymphoma (NHL). It can be clinically and morphologically difficult to distinguish FL from other types of small B-cell NHLs, such as mantle cell lymphoma (MCL) and small lymphocytic lymphoma/chronic lymphocytic leukemia (SLL/CLL), which differ in terms of prognosis and treatment.1, 2 Low-grade FL morphologically also may resemble reactive follicular hyperplasia.3 Fine-needle aspiration (FNA) cytopathology is a rapid, minimally invasive procedure that has been used widely in the diagnosis, follow-up, and staging of lymphomas.4, 5 However, certain intrinsic limitations of FNA specimens, such as relative lack of architectural features and small sample size, can make the diagnosis challenging.4, 6 Although flow cytometry immunophenotyping (FCM) is a useful ancillary study7, 8 and can support a diagnosis of FL by detecting the characteristic monoclonal B-cell proliferation with coexpression of CD19 and CD10,9 at least 2 million cells are required for a diagnostic panel. When cellularity is limited because of fibrosis, extensive necrosis, or cellular damage, FCM results can be inconclusive. In addition, although CD10 is a very useful marker in distinguishing FL from other low-grade NHLs, approximately 20% of FLs may not be CD10 positive (CD10+).10, 11

A reciprocal t(14;18)(q32;q21) translocation is the hallmark cytogenetic abnormality for FL, resulting in fusion of the immunoglobulin heavy-chain (IgH) and BCL-2 genes,12–14 although the same translocation can be found in diffuse large B-cell lymphoma (DLBCL) of follicle center cell origin.15, 16 The majority of the breakpoints on chromosome 14 are in the joining region (J) on IgH.14, 17 For the BCL-2 gene, 60% of breakpoints occur in the major breakpoint region (mbr) located in the 3′-untranslated region of exon 3,18 whereas 25% occur in a minor cluster region (mcr) approximately 30 kilobases (kb) 3′ of the BCL-2 gene.18, 19 The remaining breakpoints are scattered widely over the other regions of the BCL-2 gene region.20 Several methods, including conventional cytogenetics, polymerase chain reaction (PCR) analysis, and Southern blot analysis, can be used to demonstrate this rearrangement, but they are of limited value. Conventional cytogenetic analysis is cumbersome, requiring special cell culture techniques, and is successful in only 55–86% of FLs.21, 22 Southern blot and PCR analyses fail to detect breakpoints outside the mbr and mcr regions and thus also have low detection rates.23 A newly developed long-range PCR amplification method may have greater efficiency in detecting an IgH/BCL-2 rearrangements that occur outside the mbr and mcr regions24, 25; however, it is technically demanding, and it is not performed routinely.

Interphase fluorescence in situ hybridization (I-FISH) is a desirable alternative, because it is fast and convenient for detecting specific chromosomal translocations associated with different subtypes of NHL at the cellular level.26 It is particularly advantageous in evaluating cytologic material, because it requires only a small number of cells. Unlike the application of FISH to tissue sections, in which the truncation artifact of variable numbers of nuclei can interfere with interpretation of the probe reaction, I-FISH uses cells that are disaggregated and monolayered by cytospin preparation, which facilitates excellent technical hybridization.27, 28 Moreover, FISH results can be obtained with retention of cellular morphology, which permits simultaneous evaluation of morphology and chromosomal alterations. In the current study, we compared I-FISH results with FCM findings to determine the value of I-FISH as an adjunctive diagnostic test for morphology in FLs on FNA material. Its utility in detecting a follicle center cell origin in DLBCL also was investigated.



Between June 2000 and September 2002 at The University of Texas M. D. Anderson Cancer Center, FNA specimens of lymph nodes from 84 patients with clinical findings suggestive of malignant lymphoma were analyzed. The study was conducted with the approval of the M. D. Anderson Cancer Center's Institutional Review Committee (Protocol RCR03-0105). Specimens were obtained either by cytopathologists from palpable masses or by radiologists from deep-seated masses. Each specimen was triaged for morphologic, I-FISH for t(14;18) (q32;q21), and FCM studies. In some specimens, a portion was submitted for karytoping by cytogenetics or PCR to detect the t(14;18)(q32;21) translocation.

Direct smears were stained with Papanicolaou and Diff-Quik techniques (StatLab, Lewisville, TX) and were assessed immediately to determine the quality of the aspirates and cytomorphology. A portion of each sample (2–5 million cells) was submitted for immunophenotyping by FCM. The remaining cells collected in RPMI medium were processed over a Ficoll–Hypaque gradient (Pharmacia, Peapack, NJ) to purify the mononuclear cells. One hundred microliters of mononuclear cells, at a concentration of 1 × 106 cells/mL, were centrifuged onto silane-coated slides (1 × 105 cells per slide). For each specimen, 1 air-dried cytospin was stained with Diff-Quik to check morphology, and another cytospin that was fixed in Carnoy fixative (3:1 absolute methanol:glacial acetic acid) for 20 minutes was used for I-FISH. A third air-dried cytospin was immunostained for Ki-67. The remaining cells were stored, as cytospins, in a tumor bank at −70 °C.

Seventy-seven patients had previous, concurrent, or subsequent tissue biopsies, mostly within 6 months of the current aspiration. The cytologic diagnosis was rendered based on combined information of clinical setting, cytomorphology, histomorphology, immunophenotypic findings, I-FISH, and, if available, cytogenetic analysis and PCR. When a corresponding tissue biopsy was available, the grading was correlated with that of histology (World Health Organization criteria29).

DNA Probes and the FISH Method

A commercially available premixed probe (Vysis, Downers Grove, IL) was used to detect the t(14;18)(q32;21) translocation. The 18q21 locus–specific probe was labeled with SpectrumOrange-dUTP. This probe was 750 kb long, including the 250 kb bcl-2 gene and 250 kb flanking sequences at each side. The IgH probe was labeled with SpectrumGreen-dUTP and was approximately 1.5 megabases long. This probe contained the entire IgH locus as well as a 300 kb sequence flanking the 3′ end of the IgH locus. For all specimens that were suspicious for MCL, the 11q13 (BCL-1) probe and the 14q32 (IgH) probe (Vysis) were labeled with SpectrumOrange-dUTP and SpectrumGreen-dUTP, respectively.

Slides were pretreated in 2× sodium chloride citrate (SSC) at 37 °C for 30 minutes and then dehydrated in a series of cold alcohol solutions (70%, 85%, 95%, and 100%). A mixture of 0.6 μL of probe, 1.2 μL of distilled water, and 4.2 μL of hybridization buffer was then applied to the slides; coverslipped; and sealed with rubber cement. Samples and probes were codenatured in Hybrite (Vysis) at 74 °C for 1 minute. Washings after overnight hybridization at 37 °C were as follows: 0.4 × SSC/0.3% Nonidet P-40 (Vysis) at 74 °C for 2 minutes, and 2 × SSC/0.1% Nonidet P-40 at room temperature for 1 minute. Interphase nuclei were counterstained with 10 μg/mL 4′,6-diamidino-2-phenylindole (DAPI)-containing antifade solution (Vysis).

Visualization and Scoring of FISH Signals

Hybridization sites were analyzed using a Leica DMLB fluorescence microscope (Leica, Bensheim, Germany) equipped with appropriate filter sets for visualizing spectrum green or orange and DAPI fluorescence signals. Slides were analyzed only if 80% of the nuclei in the field of view were interpretable. At least 200 nuclei per case and at least 2 different areas on the same slide were scored. To avoid misinterpretation due to sister chromatids in the S-phase or G2-phase, two signals were counted as one (fusion) if they were situated extremely close to each other (a distance of less than one signal size). Only nonoverlapping, intact nuclei were scored. Each nucleus was scored for the number of green and orange signals simultaneously. Nuclei were counted only if at least one bright-green signal and one bright-orange signal were present to avoid misinterpretation due to false monosomies or deletions due to insufficient hybridization efficiency. A FISH preparation was considered insufficient for analysis when the hybridization signal was too weak, the background (noise) was too high, or the cellularity was too low. Appropriate positive controls were included. In some normal cells, double signals may be expected because of random colocalization of the BCL-2 and IgH regions. The frequency of such false-positive cells has to be determined carefully to define the cutoff level. We used normal lymphocyte nuclei as a negative control to assess hybridization efficiency. The cutoff level for positivity was set at the mean (%) ± 3 standard deviations, which was > 10% single fusion signals and/or > 1.5% double fusion signals.30

Quantitative Flow Cytometry Immunophenotyping

Flow cytometry immunophenotyping was performed on a Becton Dickinson flow cytometer (San Jose, CA) using a four-color technique. A panel of monoclonal antibodies (Becton Dickinson) specific for the following lymphocyte surface marker antigens was used: immunoglobulin light chains (κ and λ), CD19, CD20, CD10, CD5, CD23, CD11c, CD22, CD79a, FMC7, and CD3. Monoclonality was defined by a κ:λ ratio > 4.0 or < 0.5.

Immunocytochemical Staining

Immunocytochemical (ICC) staining was carried out on cytospin preparations that had been kept at −70 °C. After fixation for 10 minutes in absolute acetone at 4 °C, specimens were incubated with mouse monoclonal antibodies. If Ki-67 (Dako Corporation, Carpinteria, CA) immunoreactivity was demonstrated in > 26% of lymphoid cells, then the lesion was considered to have a high proliferation index.31 In specimens in which the FCM results were noncontributory or were not available, ICC for λ and λ light chains (Dako Corporation) was performed on cytospin preparations, and monoclonality was considered when the κ:λ ratio was > 6.0 or < 0.5.


The 84 lymph node FNA biopsies comprised 40 FLs and 44 non-FLs, including specimens of reactive lymphoid hyperplasia (Table 1). Results of I-FISH t(14;18) in normal lymphoid cells (the negative control) and FL are illustrated in Figures 1 and 2, respectively. Thirty-nine of the 40 FLs diagnosed by FNA had histologic confirmation. Of 40 FLs, t(14;18) translocation by FISH was positive in 34 FLs (85.0%), negative in 3 FLs (7.5%), and insufficient in 3 FLs (7.5%), whereas by FCM, 30 FLs (75%) had a CD19+/CD10+ population (28 monoclonal, 2 nonclonal), 5 FLs (12.5%) had a CD19+/CD10-negative (CD10) population (3 monoclonal, 2 nonclonal), and 5 FLs (12.5%) were insufficient (Tables 2, 3). All four FLs that were nonclonal by FCM had a prior or concurrent tissue-based diagnosis of FL and exhibited unequivocal t(14;18) translocation as determined by FISH. One of these four FLs was identified as monoclonal by ICC. All five insufficient FLs by FCM were positive for the t(14;18) translocation by FISH and exhibited clonality as determined by ICC. In contrast, the 3 negative FLs and 3 insufficient FLs for FISH were monoclonal and exhibited CD19/CD10 coexpression on FCM. Table 4 shows the relation between FISH and FCM results in FLs; 24 of 40 FLs (60%) showed both the t(14;18) translocation and CD19/CD10 coexpression. When the insufficient FLs are disregarded, the results from FISH and FCM were concordant in 24 of 32 FLs (75%).

Table 1. Summary of Non-Hodgkin Lymphomas Studied
DiagnosisNo of aspiratesNo of tissue biopsies
  • SLL/CLL: small lymphocytic lymphoma/chronic lymphocytic leukemia; NOS: not otherwise specified.

  • a

    Immunophenotype or histology was compatible with low-grade B-cell lymphoma, NOS. Four of the five patients had the corresponding histology: Three lymphomas were diagnosed as low-grade B-cell lymphoma, NOS, and 1 lymphoma was diagnosed as lymphoplasmacytic lymphoma.

  • b

    There was no evidence of concomitant or antecedent follicular lymphoma. Lymphoblastic and Burkitt lymphomas were excluded.

Follicular lymphoma4039
 Grade 177
 Grade 21918
 Grade 31414
Nonfollicular lymphoma4438
 Mantle cell lymphoma65
 B-cell SLL/CLL109
 B-cell lymphoma, NOSa54
 Reactive lymphoid hyperplasia65
 Diffuse large B-cell lymphomab1715
Figure 1.

Normal lymphocytes showing 2 red and 2 green fluorescence signals (Vysis probe; original magnification ×1000).

Figure 2.

The t(14;18)(q32;q21) translocation, demonstrated by yellow fusion signals (arrows) in a representative case of follicular lymphoma (Vysis probe; original magnification ×1000).

Table 2. Results of Fluorescence in Situ Hybridization t(14;18)(q32;q21) and Flow Cytometry for Follicular Lymphoma by Grade
GradeNo. of patientsFISH t(14;18)(q32;q21) (%)Flow cytometry immunophenotyping (%)
  • FISH: fluorescence in situ hybridization; +: positive; −: negative.

  • a

    For five follicular lymphomas with insufficient flow cytometric results, monoclonality was established by immunocytochemistry after unsuccessful flow cytometry.

175 (71.4)1 (14.3)1 (14.3)6 (85.7)1 (14.3)0 (0)
21916 (84.2)1 (5.3)2 (10.5)13 (68.4)3 (15.8)3 (15.8)
31413 (92.9)1 (7.1)0 (0.0)11 (78.6)1 (7.1)2 (14.3)
Total4034 (85.0)3 (7.5)3 (7.5)30 (75.0)5 (12.5)5 (12.5)a
Table 3. Results of Interphase Fluorescence in Situ Hybridization, Immunophenotyping, Polymerase Chain Reaction, and Cytogenetic Studies in Follicular Lymphoma and Diffuse Large B-Cell Lymphoma
Patient no.DiagnosisFISH t(14;18)(q32;q21)ImmunophenotypingPCR t(14;18)(q32;q21)Cytogenetics t(14;18)
  1. FISH: fluorescence in situ hybridization; FCM: flow cytometry; ICC: immunocytochemistry; PCR: polymerase chain reaction; FL1–FL3: follicular lymphoma Grades 1–3; −: negative; +: positive; Insuff: insufficient for either FISH or FCM analysis; DLBCL: diffuse large B-cell lymphoma.

1FL1Monoclonal, CD10+  
2FL1InsuffMonoclonal, CD10+  
3FL1+Monoclonal, CD10+   
4FL1+Monoclonal, CD10+  +
5FL1+Monoclonal, CD10+   
6FL1+Monoclonal, CD10−   
7FL1+Nonclonal, CD10+Monoclonal  
8FL2Monoclonal, CD10+ Insuff 
9FL2+Monoclonal, CD10+   
11FL2+Monoclonal, CD10+   
12FL2+Monoclonal, CD10+  
14FL2InsuffMonoclonal, CD10+   
15FL2+Monoclonal, CD10+   
16FL2+Monoclonal, CD10+   
18FL2+Monoclonal, CD10+  
19FL2+Monoclonal, CD10−  
20FL2+Monoclonal, CD10+  
21FL2+Monoclonal, CD10+   
22FL2+Nonclonal, CD10+Nonclonal  
23FL2+Nonclonal, CD10−Nonclonalmbr+ 
24FL2InsuffMonoclonal, CD10+   
25FL2+Monoclonal, CD10+   
26FL2+Nonclonal, CD10−Nonclonal  
27FL3+Monoclonal, CD10+  +
28FL3+Monoclonal, CD10+   
29FL3+Monoclonal, CD10+   
31FL3+Monoclonal, CD10+  
33FL3+Monoclonal, CD10−   
34FL3Monoclonal, CD10+   
35FL3+Monoclonal, CD10+  
36FL3+Monoclonal, CD10+  
37FL3+Monoclonal, CD10+ mbr+ 
38FL3+Monoclonal, CD10+   
39FL3+Monoclonal, CD10+   
40FL3+Monoclonal, CD10+   
44DLBCLNonclonal, CD10+Nonclonal  
45DLBCLNonclonal, CD10+Nonclonal  
46DLBCLMonoclonal, CD10−   
47DLBCLInsuffMonoclonal, CD10−   
48DLBCLMonoclonal, CD10−  
50DLBCLMonoclonal, CD10−   
51DLBCLMonoclonal, CD10−   
52DLBCLInsuffMonoclonal, CD10+   
53DLBCLNonclonal, CD10−Nonclonal 
54DLBCL+Monoclonal, CD10−   
56DLBCLMonoclonal, CD10−   
57DLBCL+Monoclonal, CD10+   
Table 4. Relation between t(14;18)(q32;q21) Interphase Fluorescence in Situ Hybridization and Flow Cytometry Results in Patients with Follicular Lymphoma
Flow cytometryFISH t(14;18)(q32;q21) (%)
  1. FISH: fluorescence in situ hybridization; +: positive; −: negative.

CD19+/CD10+24 (60.0)3 (7.5)3 (7.5)
CD19+/CD105 (12.5)0 (0.0)0 (0.0)
Insufficient5 (12.5)0 (0.0)0 (0.0)

The 2 specimens with positive t(14;18) translocation results on cytogenetic analysis also were positive for the translocation on FISH. The other FLs did not undergo cytogenetic analysis (Table 3). In addition, in 12 FL specimens for which PCR results for t(14;18)(q32;q21) were available, FISH detected the translocation not only in FLs with positive PCR results (n = 2) but also in FLs with negative PCR results (n = 8) (Table 3).

Among the 44 non-FLs, 38 had histologic confirmation (Table 1). Except for 3 DLBCLs which failed to exhibit clonality on FCM, all other non-FLs had FCM findings consistent with their diagnostic category (Table 1). Among these 3 DLBCLS, 2 expressed CD19/CD10 and 1 was CD19+/CD10. In addition, the 6 MCLs showed the hallmark t(11;14)(q13;q32) translocation on FISH. The t(14;18) translocation was detected in 7 of 44 non-FLs by FISH: 5 in DLBCLs and 2 in SLLs/CLLs. The values of fusion signals in the 2 SLL/CLLs were just above the cutoff level, and both had only single fusions. The insufficient rate for FISH and FCM analysis was 20.5% (9 of 44 non-FLs) and 11.4% (5 of 44 non-FLs), respectively (Tables 3, 5.) It is noteworthy that all specimens that were insufficient for FCM and specimens in which FCM failed to detect clonality in the non-FL group occurred in DLBCLs. However, three of five DLBCLs insufficient for FCM were demonstrated as monoclonal by ICC. For detecting features associated with follicular center cell origin in DLBCLs, FISH detected t(14;18) translocation in 5 of 17 cases (29.4%) and FCM demonstrated CD19+/CD10+ in 4 of 17 cases (23.5%).

Table 5. Comparison of Fluorescence In Situ Hybridization and Flow Cytometry Results among Patients with Nonfollicular Lymphoma
Lymphoma typeNo. of patientsFISH of t(14;18)(q32;q21) (%)Flow cytometry (%)
  • RLH: reactive lymphoid hyperplasia; BCL: B-cell lymphoma; NOS: not otherwise specified; MCL: mantle cell lymphoma; SLL/CLL: small lymphocytic lymphoma/chronic lymphocytic leukemia; DLBCL: diffuse large BCL.

  • a

    Both patients who had fluorescence in situ hybridization–positive SLL/CLL had classic clinical presentation, bone marrow and lymph node findings, and immunophenotyping on flow cytometry.

  • b

    A monoclonal B-cell population was found by immunohistochemistry in three of five DLBCL specimens with insufficient flow cytometry findings.

RLH60 (0)3 (50.0)3 (50.0)0 (0)6 (100.0)0 (0)
BCL, NOS50 (0)4 (80.0)1 (20.0)0 (0)5 (100.0)0 (0)
MCL60 (0)6 (100.0)0 (0)0 (0)6 (100.0)0 (0)
SLL/CLL102 (20.0)a5 (50.0)3 (30.0)0 (0)10 (100.0)0 (0)
DLBCL175 (29.4)10 (58.8)2 (11.8)4 (23.5)8 (47.1)5 (29.4)b
Total447 (15.9)28 (63.6)9 (20.5)4 (9.1)35 (79.5)5 (11.4)

Using non-FLs as negative controls, but excluding de novo DLBCLs, the sensitivity, specificity, and predictive values of I-FISH and FCM in FL diagnosis were calculated (Table 6). The sensitivity and negative predictive value of FISH were higher compared with FCM. However, the specificity and positive predictive value of FISH were less compared with FCM due to two SLL/CLLs that exhibited false-positive results for the t(14;18) translocation.

Table 6. Sensitivity, Specificity, and Predictive Values of Interphase Fluorescence In Situ Hybridization t(14;18)(q32;q21) and Flow Cytometry in the Diagnosis of Follicular Lymphoma
ValueI-FISH t(14;18)(q32;q21)Flow cytometry
  • I-FISH: interphase fluorescence in situ hybridization; +: positive; −: negative; FLs: follicular lymphomas; non-FLs: nonfollicular lymphomas.

  • a

    Non-FLs excluded de novo diffuse large B-cell lymphomas and lymphomas with insufficient specimens.

No. of FLs343305
No. of non-FLsa218027
Sensitivity (%)91.985.7
Specificity (%)90.0100.0
Positive predictive value (%)94.4100.0
Negative predictive value (%)85.784.4


Although typical cases may be diagnosed by cytomorphologic findings and FCM immunophenotyping, the diagnosis may be very difficult in atypical cases or in samples with suboptimal cellularity. The t(14;18)(q32;q21) chromosomal translocation is regarded as the hallmark of FL and is found in up to 95% of cases.12–14 It also can be found, with much lower frequency, in DLBCLs.15, 16 It has been shown that FISH is a sensitive and relatively simple technique for detecting this translocation, and it has greater detection efficiency compared with many other methods.17, 22, 30 Its suitability for small samples, such as FNA material, makes the FISH technique even more attractive. However, little is known about the value of I-FISH for t(14;18)(q32;q21) in the diagnosis of FL on FNA-derived materials compared with FCM. The data presented here showed that I-FISH could detect the t(14;18) translocation in 85% of FLs, whereas FCM could detect the typical CD19+/CD10+ immunophenotype in only 75% of FLs. We also found that the sensitivity and the negative predictive value of FISH were greater compared with FCM. The underlying reasons for the differences are multiple. The number of insufficient specimens is an important factor that can affect a technique's detection rate significantly. Compared with the FISH results, the insufficient rate of specimens for FCM in the FL group was higher and resulted mainly from insufficient cellularity. In addition, FISH appeared to be more sensitive than FCM, because it detected not only FLs that had the typical CD19+/CD10+ immunophenotype but also FLs that were monoclonal but CD10 or nonclonal but CD19+.

The specificity and the positive predictive value of the FISH technique were decreased due to 2 specimens of SLL/CLL that showed a t(14;18) translocation. The presence of this translocation in SLL/CLL is unusual, with only eight cases reported in the literature.32 The 2 patients in our series were older men, ages 78 years and 67 years, with peripheral blood, bone marrow, and concurrent tissue section findings characteristic of CLL. In addition, CD10 was negative, and CD5 and CD23 were positive by FCM. We believe that these two specimens most likely represent false-positive FISH results due to chance colocalization in the interphase nuclei of the two studied DNA regions. This also is supported by the finding that the frequencies of the colocalized signals in both specimens were similar to the cutoff value. Nevertheless, the overall positive predictive value and negative predictive value by FISH in the diagnosis of FL remain good, at 94.4% and 85.7%, respectively.

Studies have shown that the t(14;18) translocation is seen more frequently in FLs of lower histologic grade.17, 33 In our data, however, the highest detection rate occurred among the Grade 3 FLs, followed by Grade 2 FLs and Grade 1 FLs. This can be ascribed in part to the number of insufficient specimens in each subtype as well as the significant difference in the numbers of specimens among each subtype. All Grade 3 specimens were sufficient for analysis, whereas only 1 Grade 1 specimen and 2 Grade 2 specimens were insufficient, which may significantly decrease the detection frequency, especially when the total specimen number for that grade is small.

Similar to the findings reported in other studies,12, 22, 30 our findings show that FISH appears to be a more sensitive technique for detecting the t(14;18) translocation than PCR, although the number of specimens that had the PCR result was relatively small. The reported frequency of IgH/BCL-2 by PCR in FL has been highly variable in published studies, probably reflecting variation in patient selection, the size of series, and the probes used (major, minor, or variable breakpoint regions). Using cytogenetics as the gold standard, Horsman et al. found that PCR failed to detect rearrangements in approximately 25% of FLs.12 The low detection rate of PCR is associated with the fact that the current PCR method is not able to detect some breakpoints mapped outside the mbr and mcr regions.12, 23 The FISH technique circumvents this limitation by using a probe that covers a large region of the bcl-2 gene.

DLBCL is a heterogeneous group of neoplasms with different clinical, immunohistochemical, cytogenetic, and molecular characteristics. Using a microarray gene-expression profiling technique in a large series of DLBCLs, Alizadeh et al.34 identified two molecularly distinct forms of DLBCL: germinal center B-like DLBCL and activated B-like DLBCL. Those authors found that patients who had germinal center B-like DLBCL had a significantly better overall survival compared with patients who had activated B-like DLBCL. The gene expression program that distinguished germinal center B-like DLBCLs included several known markers of germinal center differentiation, including genes encoding CD10 and BCL-6. The t(14;18) translocation has been detected variably in 10–40% of DLBCLs15, 16 and was found to be correlated with immunohistochemical expression of CD10 and BCL-6.35, 36 A recent study found that the t(14;18) translocation occurred exclusively in germinal center B-like DLBCL.37 In the current study, the t(14;18) translocation was identified by FISH in 29% of DLBCLs, whereas a CD10+ monoclonal population was detected by FCM in only 23% of DLBCLs. Factors that contribute to the low detection frequency by FCM include fibrosis, necrosis, and poor cellular integrity associated with large cell lymphoma.38, 39 In a small number of DLBCLs, the malignant clone was obscured by reactive lymphoid cells and resulted in a nonclonal phenotype. These results suggest that FISH may be superior to FCM in detecting DLBCL of germinal center cell origin.

An advantage of I-FISH is its utility in cytology specimens, because it requires only 1.5 × 105 cells from a single cytospin. Furthermore, a cytospin is an optimal preparation for I-FISH, because the monolayer facilitates excellent hybridization results.27

Although FISH is a sensitive technique for detecting the t(14;18) translocation, there is a residual insufficient-specimen rate that could be improved. Notably, in the majority of specimens that were insufficient for FISH analysis, the insufficiency was due not to low cellularity but to poor hybridization results, either because the hybridization signal was too weak or the background was too high. It appears that more sophisticated specimen handling is required to yield optimal hybridization signals. It is critical to avoid delays in specimen processing to maintain high cellular viability before fixation, because degradation of the target DNA may occur due to improper preservation or cell preparation.26, 27

In conclusion, I-FISH analysis for the t(14;18) translocation in the diagnosis of FL in FNA specimens is reliable and accurate. This validates its value as an adjunct in the diagnosis of FL and for monitoring FL, especially when the cellularity is limited or FCM results are equivocal. For detecting a germinal center cell origin in DLBCL, I-FISH for the t(14;18) translocation appears to be slightly more sensitive than FCM for CD19/CD10 coexpression and thus may be more helpful from a therapeutic and prognostic point of view.