Context.—Bone marrow (BM) examination is part of the staging workup of lymphoma patients. Few studies have compared BM histologic findings with results of flow cytometric immunophenotyping analysis in follicular lymphoma (FL) patients.

Objective.—To correlate histologic findings with immunophenotypic data in staging BM biopsy and aspiration specimens of FL patients.

Design.—Bone marrow biopsy specimens of untreated FL patients were reviewed. Histologic findings were correlated with 3-color flow cytometric immunophenotyping results on corresponding BM aspirates.

Results.—Bone marrow biopsy specimens (with or without aspirates) of 114 patients with histologic evidence of FL in BM were reviewed. There were 76 bilateral and 38 unilateral biopsies performed, resulting in 190 specimens: 187 involved by FL and 3 negative (in patients with a positive contralateral specimen). The extent of BM involvement was <5% in 32 (17.1%), ≥5% and ≤25% in 102 (54.6%), >25% and ≤50% in 27 (14.4%), and >50% in 26 (13.9%) specimens. The pattern of involvement was purely paratrabecular in 81 (43.3%), mixed in 80 (42.8%), and purely nonparatrabecular in 26 (13.9%). Immunophenotyping was only performed unilaterally, on BM aspirates of 92 patients, and was positive for a monoclonal B-cell population in 53 (57.6%) patients. Immunophenotyping was more often negative when biopsy specimens showed FL with a purely paratrabecular pattern. For comparison, we assessed 163 FL patients without histologic evidence of FL in BM also analyzed by flow cytometric immunophenotyping. A monoclonal B-cell population was identified in 5 patients (3%).

Conclusions.—Our data suggest that 3-color flow cytometric immunophenotyping adds little information to the evaluation of staging BM specimens of FL patients.

Staging is an important aspect of the management of patients with non-Hodgkin lymphoma, useful for planning treatment, monitoring response to therapy, and evaluating for relapse.1,2 Most staging protocols include bone marrow (BM) aspiration and biopsy as a part of the standard workup.

For patients with follicular lymphoma (FL), the morphologic findings of BM involvement were described 20 to 30 years ago.3,4 These studies showed that histologic assessment of BM biopsy specimens is more sensitive than examination of BM aspirate smears and clot specimens. It is common for biopsy specimens to be positive for FL when aspirate smears and clot specimens are negative. By contrast, the converse is uncommon. These studies, however, were performed prior to the advent of routine multicolor flow cytometric immunophenotyping, a technique that can be easily applied to the assessment of BM aspirates.

Clinicians at our hospital often order flow cytometric immunophenotyping analysis on BM aspirate material as part of the staging workup of lymphoma patients. However, the value of flow cytometric immunophenotyping in this setting is not well defined.5–10 A number of studies have yielded contradictory results, and most of these studies have included patients with all histologic types of non-Hodgkin lymphoma, as well as previously treated patients at time of relapse. No study has focused specifically on the role of flow cytometric immunophenotyping in the evaluation of initial staging of BM specimens in patients with FL.

In this study, our goals were threefold. First, we reviewed the histologic findings in staging BM specimens of FL patients. In histologically positive BM biopsy specimens both pattern and extent of involvement were determined. Second, we correlated histologic findings with the results of flow cytometric immunophenotyping of BM aspirate material. Lastly, we assessed the contribution of flow cytometric immunophenotyping to the workup of FL patients with histologically negative BM biopsy specimens.

Case Selection

The patient specimens analyzed were derived from 3 separate treatment protocols for patients with untreated FL and were collected for this study retrospectively. These protocols did not accrue patients concurrently, and the patients were not consecutive. For this reason, the numbers of patients with positive and negative BM specimens are not a simple numerator and denominator, and these numbers cannot be used to generate a frequency of BM involvement in FL patients. In all patients, the diagnosis of FL was established on the basis of tissue (mostly lymph node) biopsy results. Most of these biopsies were performed at other institutions and the slides were reviewed at our hospital. In a smaller subset of patients, the diagnostic tissue biopsy was performed at our institution.

One treatment protocol was designed for FL patients with BM involvement (stage IV) and accrued patients from 1999 to 2001. Most of these patients were not known to be stage IV until after bone marrow staging was performed. The other 2 treatment protocols were designed for patients with stage I to III FL without BM involvement and accrued patients from 1997 to 2001. From the latter 2 groups, we only included patients who also had immunophenotyping performed on BM aspirates. For these protocols, bilateral BM biopsy was recommended but not mandatory, and flow cytometric immunophenotypic analysis was optional. In addition, in patients who underwent bilateral BM biopsy, immunophenotyping was performed only on BM aspirate material from 1 side.

For this study only the results of BM biopsy specimens were compared with immunophenotypic data, and hematoxylin-eosin–stained slides of the BM biopsy sections were reviewed. Although the morphologic findings of FL in BM are well known,3,4,11 criteria for a positive BM biopsy specimen are not precisely defined in the literature. In this study, to consider a BM biopsy specimen positive for FL we required the presence of lymphoid aggregates composed of atypical lymphoid cells, either small cleaved or a mixture of small and large cleaved, or large noncleaved. Regarding the size of the lymphoid aggregates, we accepted aggregates of any size as evidence of FL if they were paratrabecular or formed atypical follicles. However, if the lymphoid aggregates were purely nonparatrabecular they had to replace at least 5% to 10% of the BM medullary space to be considered positive.

The extent and pattern of involvement by FL in BM biopsy specimens was semiquantified into 4 groups: <5%, ≥5% and ≤25%, >25% and ≤50%, and >50%. The patterns of involvement were classified into 1 of 3 groups: paratrabecular only, nonparatrabecular only, and paratrabecular with other patterns mixed. We also noted the presence of a follicular pattern according to the definition of Torlakovic and colleagues12: presence of neoplastic follicles in BM, with less than 10% of the infiltrate paratrabecular.

BM Biopsy Specimen Size

Bone marrow biopsy specimen size (length) was measured grossly, after fixation and prior to processing in most cases. In a subset of cases without a gross measurement, however, size was measured on routinely stained glass slides. It is recognized that this approach underestimates size because the length of a BM biopsy specimen can shrink up to 25% during routine processing.13 For FL patients with bilateral BM biopsy specimens, the length of each specimen was added to create 1 combined size.

Immunophenotypic Methods

As this is a retrospective study, BM aspirates were assessed using the methods used in the laboratory at that time. All BM aspirates were analyzed using 3-color flow cytometric immunophenotyping and a FACScan (BD Biosciences, San Jose, Calif) instrument as described previously.14,15 Each tube contained antibodies specific for CD19 and CD45, with the addition of antibodies specific for either immunoglobulin κ or λ light chains (BD Biosciences). Lymphocytes were gated for analysis using CD45 expression and side scatter. The CD45 antibody used for gating was conjugated to peridinin chlorophyll alpha protein. The CD19 antibody was conjugated with phycoerythrin and the immunoglobulin light chain antibodies were conjugated with fluorescein isothiocyanate. Isotype-matched irrelevant fluorescein isothiocyanate- and phycoerythrin-conjugated antibodies were used as negative controls, and cursors were set to include more than 95% of events as negative. For each specimen that had an adequate number of CD19-positive events, a κ/λ ratio was calculated.

Other antibodies included in immunophenotypic analysis were used variably over time, mostly on cases with numerous cells for evaluation, and included 1 or more of the following: CD2, CD3, CD4, CD5, CD7, CD8, CD10, CD16, CD20, CD23, and CD56, and FMC-7 (Immunotech/Beckman-Coulter, Fullerton, Calif).

Statistical Analysis

The initial statistical analysis approach taken was to use the data generated in this study to calculate the sensitivity and specificity of different symmetrically paired κ/λ ratio threshold values for the groups histologically positive or histologically negative for FL (Table 1).

Table 1. 

Sensitivity and Specificity for Symmetric (Paired) κ/λ Ratio Thresholds*

Sensitivity and Specificity for Symmetric (Paired) κ/λ Ratio Thresholds*
Sensitivity and Specificity for Symmetric (Paired) κ/λ Ratio Thresholds*

Subsequently, a receiver operator curve (ROC) was generated, using these calculated sensitivity and specificity values, for different symmetric paired κ/λ ratio threshold values. The distance from the upper left corner of the ROC was also used to determine the optimum symmetrically paired threshold value for the κ/λ or λ/κ ratio. The next step was to calculate the sensitivity and specificity for various combinations of asymmetrically paired threshold values (Table 2). A range of asymmetrically paired threshold values was used, and the distance from the upper left corner of the ROC was calculated. This distance was then plotted as a contour surface, and the minimum point from the upper left corner was chosen as the optimum asymmetric set of threshold values. All analyses were performed using S-PLUS 6.1 for Windows (Insightful Inc, Seattle, Wash).

Table 2. 

Sensitivity and Specificity for Asymmetric (Unpaired) κ/λ Ratio Thresholds*

Sensitivity and Specificity for Asymmetric (Unpaired) κ/λ Ratio Thresholds*
Sensitivity and Specificity for Asymmetric (Unpaired) κ/λ Ratio Thresholds*

Histologic Findings

Histologically Positive Group

This group consisted of 114 patients, 76 (66.7%) of whom underwent bilateral and 38 (33.3%) of whom underwent unilateral BM biopsy. These procedures resulted in a total of 190 BM biopsy specimens: 187 histologically positive for FL and 3 negative. The 3 negative specimens were obtained as part of bilateral BM biopsies, and the contralateral BM biopsy specimen was histologically positive for FL.

Bone marrow biopsy specimen size in this group ranged from 0.8 to 4.8 cm, with a median of 2.8 cm. Bone marrow biopsy specimen size was 0.8 cm in 1 patient, 1.0 to 1.1 cm in 5 patients, and 1.2 cm or greater in the remaining 108 patients.

The extent of BM involvement by FL was <5% in 32 (17.1%), ≥5% and ≤25% in 102 (54.6%), >25% and ≤50% in 27 (14.4%), and >50% in 26 (13.9%) specimens. The pattern of involvement was purely paratrabecular in 81 (43.3%), mixed (paratrabecular associated with other patterns) in 80 (42.8%), and purely nonparatrabecular in 26 (13.9%). In the latter group, in 6 BM biopsy specimens a purely follicular pattern was identified.

Histologically Negative Control Group

This group comprised 163 patients, 135 (82.8%) of whom underwent bilateral and 28 (17.2%) of whom underwent unilateral BM biopsy. These procedures resulted in a total of 298 BM biopsy specimens, all histologically negative for FL.

Bone marrow biopsy specimen size in this group ranged from 0.6 to 6.0 cm, with a median of 3.0 cm. Bone marrow biopsy specimen size was 0.6 cm in 1 patient, 1.1 cm in 1 patient, and 1.2 cm or greater in 157 patients. In 4 patients, BM biopsy specimen size was not available in the pathology reports, and these slides could not be located for review and assessment of size.

Flow Cytometric Immunophenotyping Results

In the group histologically positive for FL, immunophenotyping was performed on 92 BM aspirate specimens. In 58 specimens, the number of CD19-positive events was adequate to calculate a κ/λ ratio. In the remaining 34 specimens, too few CD19-positive events were identified to reliably calculate this ratio.

For the 58 BM aspirates on which a κ/λ ratio was calculated, 53 were monoclonal. Monoclonality in this study was defined as a κ/λ ratio of ≥4 or ≤0.3 (see “Statistical Determination of Threshold Values for κ/λ Ratios”). Thirty-six cases expressed κ with a κ/λ ratio ranging from 4.5 to 96.3 with a median of 14.0. Seventeen cases expressed λ with a κ/λ ratio ranging from 0.01 to 0.3 with a median of 0.1. The 5 remaining cases had κ/λ ratios that did not meet the specified threshold values and were therefore considered polyclonal. In this group the median κ/λ ratio was 1.35 (range, 0.7–2.6).

In the group histologically negative for FL, immunophenotyping was performed on 163 BM aspirate specimens, with an adequate number of CD19-positive events to calculate a reliable κ/λ ratio in 90 specimens. In the remaining 73 BM aspirates, too few CD19-positive events were present to reliably calculate this ratio.

For the 90 BM aspirates on which a κ/λ ratio was calculated, 5 were monoclonal as defined in this study. Three specimens had a κ/λ ratio ≥4, and 2 specimens had a κ/λ ratio ≤0.3. The remaining 85 cases had κ/λ ratios that did not meet the specified threshold values and were therefore considered polyclonal. In this group the median κ/λ ratio was 1.4 (range, 0.4–3.8).

BM Biopsy Histologic Findings Correlated With Immunophenotypic Results

Extent of Involvement by FL

As indicated in Table 3, immunophenotypic analysis detected a monoclonal B-cell population in 13 (59%) of 22, 20 (43.4%) of 46, 8 (66.6%) of 12, and 12 (100%) of 12 patients in whom BM biopsy specimens showed <5%, ≥5% and ≤25%, >25% and ≤50%, and >50% histologic involvement by FL, respectively. The difference between BM biopsy specimens with >50% involvement, 12 (100%) of 12, compared with BM biopsy specimens showing lesser involvement, 41 (44.5%) of 92, was statistically significant (P < .001; chi-square test). There was no significant difference between the <5%, ≥5% and ≤25%, and >25% and ≤50% groups.

Table 3. 

Extent and Pattern of Follicular Lymphoma Involving Bone Marrow Biopsy Specimens With Flow Cytometry Immunophenotypic Data

Extent and Pattern of Follicular Lymphoma Involving Bone Marrow Biopsy Specimens With Flow Cytometry Immunophenotypic Data
Extent and Pattern of Follicular Lymphoma Involving Bone Marrow Biopsy Specimens With Flow Cytometry Immunophenotypic Data

Pattern of Involvement by FL

Flow cytometric immunophenotyping detected a monoclonal B-cell population in 16 (41%) of 39, 10 (71.4%) of 14, and 27 (69.2%) of 39 BM biopsy specimens histologically involved by FL with a purely paratrabecular pattern, a purely nonparatrabecular pattern, and a mixed pattern, respectively. The difference in positivity rates by immunophenotyping between BM biopsy specimens with a purely paratrabecular pattern, 16 (41%) of 39, compared with BM biopsy specimens with other patterns, 37 (58.7%) of 53, was statistically significant (P = .001; chi-square test). There was no significant difference between the groups with a purely nonparatrabecular versus mixed pattern.

Statistical Determination of Threshold Values for κ/λ Ratios

Using the data set of this study, the ROC generated for the symmetric paired κ/λ threshold values identified the minimum distance to the upper left corner as 0.1161, corresponding to κ/λ ratio threshold values of ≥2.9 or ≤0.34 (1/2.9), with a sensitivity of 91.4% and a specificity of 92.2%. The ROC curve generated for the asymmetric paired κ/λ threshold values identified the minimum distance to the upper left corner as 0.0924, corresponding to κ/λ threshold values of ≥4.4 or ≤0.33 (1/3) to 0.4 (1/2.6), with a sensitivity of 91.4% and a specificity of 96.7%.

Therefore, for this study we selected a κ/λ ratio of ≥4.0 or ≤0.3 (1/3) as optimum threshold values for evidence of monoclonality. Using a κ/λ ratio ≥4.0 or ≤0.3 corresponds to a distance of 0.1026 to the upper left corner on the ROC and has a sensitivity of 91.4% and a specificity of 94.4% (Table 2). One potential drawback to our statistical approach is that we have not validated the threshold values using an independent data set.

Others have described the frequency of involvement and typical paratrabecular pattern of FL in BM biopsy specimens.3,4,11,16 Most of these studies were performed more than 20 years ago, however, prior to the advent of widely available flow cytometric immunophenotyping. At our institution, clinicians often order flow cytometry immunophenotypic studies on BM aspirate material, presumably to supplement histologic evaluation. As far as we are aware, few studies have focused on patients with FL, systematically correlating histologic findings in staging BM biopsy specimens with flow cytometric immunophenotyping results derived from BM aspirates.

Determining the presence or absence of a monoclonal B-cell population by flow cytometric immunophenotyping is a useful means of confirming the presence or absence of FL in staging BM specimens. This is typically done by assessing BM aspirate material for immunoglobulin light chain expression by B cells and calculating a κ/λ ratio.7–10,17,18 A major issue we addressed in this study is the definition of monoclonality. What κ/λ ratio should be used to conclude that a B-cell population is monoclonal? A number of κ/λ ratio threshold values, ranging from 1.4 to 6, have been proposed as evidence of monoclonality.7–9,17–24 We concluded, based on our statistical methods, that the optimum κ/λ threshold ratios were >4.0 or ≤0.3. These threshold values differ slightly from the most popular κ/λ ratios used by others in the literature, >3.0 or <0.5.7,20,25 However, in support of the threshold values we have chosen, Reichard and colleagues26 reported that germinal center B cells of reactive lymph nodes can have a κ/λ ratio as high as 3.07.

As the data in this study show, 3-color flow cytometric immunophenotyping of BM aspirate specimens often can be negative in patients with histologically positive BM biopsy specimens. In 92 patients with histologically positive BM biopsy specimens, immunophenotyping showed a monoclonal B-cell population in only 53 (57.6%) patients. As one might predict, immunophenotyping detected a monoclonal B-cell population in patients who had more extensive involvement of the BM biopsy specimen. Immunophenotyping was more often positive in cases in which the BM biopsy specimen had >50% involvement compared with ≤50% involvement, and this was statistically significant. However, we observed no statistical difference in detection rate between cases with <5%, ≥5% and ≤25%, and >25% and ≤50% histologic involvement by FL. The inability to show a difference in sensitivity among the 3 groups with lesser involvement suggests that the extent of involvement by FL is less important than other factors. It may be that FL cells are often not aspirated, even when relatively numerous in the BM biopsy specimen. One possible explanation for this is that increased reticulin fibrosis surrounds FL cells, preventing their aspiration. Reticulin fibrosis is known to be associated with lymphoma infiltration in the BM and is often relatively pronounced in association with FL.27 

We hypothesized that the pattern of histologic involvement in the BM biopsy specimens might correlate with the results of immunophenotypic analysis and therefore we assessed this feature. As has been reported by others,3,4,11,16 follicular lymphoma commonly involves the BM in a paratrabecular pattern, and this was true in this study. Approximately 85% of BM biopsy specimens with histologic involvement by FL had a paratrabecular pattern, with a purely paratrabecular pattern in 43.3%. This frequency is close to the 36% frequency of a purely paratrabecular pattern reported by Canioni and colleagues.28 Flow cytometric immunophenotyping was more often negative in cases with a purely paratrabecular pattern, approximately 40% positive, compared with other patterns, approximately 70% positive, and this was statistically significant (Table 3). Follicular lymphoma aggregates in a paratrabecular location may not be accessible to the aspiration needle or may resist aspiration caused by reticulin fibrosis.

In a control group of 163 patients with BM biopsy specimens histologically negative for FL, 5 patients had a demonstrable monoclonal B-cell population by flow cytometric immunophenotyping, according to the threshold values determined by our statistical analysis. Of this group, in 90 patients a κ/λ ratio could be calculated and in 73 patients too few CD19-positive events were identified to calculate a κ/λ ratio. These 5 patients represent 3.1% of all 163 patients in whom BM aspirates were analyzed by flow cytometric immunophenotyping, or 5.6% of the 90 patients in which a κ/λ ratio was calculated. If we had used the lower κ/λ threshold values of >3 or <0.5, threshold values with identical sensitivity (Table 2), a total of 9 patients without histologic evidence of FL in BM would have been considered positive by flow cytometric immunophenotyping.

The percentage of immunophenotypically positive cases in this study using our threshold values is higher than the 1.3% previously reported by Naughton and colleagues.8 Nevertheless, our results confirm earlier studies that concluded that flow cytometric immunophenotyping is of low yield in the workup of BM specimens with no morphologic evidence of lymphoma.7–10 Furthermore, the clinical significance of detecting a small monoclonal B-cell population by flow cytometric immunophenotyping in a staging BM specimen histologically negative for lymphoma is uncertain at this time. Cheson and colleagues29 have recommended that these cases be considered negative for lymphoma for staging purposes.

We considered the possibility that inadequate BM sampling could be a possible explanation for the results of this study. The issue of BM adequacy is controversial in the literature. It has been suggested in one major textbook that an adequate BM biopsy specimen should contain at least 5 to 6 intertrabecular spaces and, after processing, should be at least 2 to 3 cm in length.30 Others have considered a BM biopsy specimen of 1.5 to 2 cm to be of an acceptable length.31,32 Statements regarding BM biopsy specimen length, however, may be arbitrary because the amount of assessable hematopoietic tissue in a BM biopsy specimen is likely to be of more importance than total length. Nonetheless, Bishop and colleagues13 have shown a strong correlation between BM biopsy specimen length and adequacy. In their study, they demonstrated that the likelihood of detecting a metastatic tumor in a BM biopsy specimen increased as the length of the BM biopsy specimen increased from 0 to 0.04 mm to >2 cm. However, their study also showed that there was little further gain in sensitivity above a BM biopsy specimen length of >1.2 cm.13 

According to the data of Bishop and colleagues,13 most of the BM biopsy specimens in the current study were adequate; very few were <1.2 cm, and our histologic examination showed adequate BM medullary space in these specimens for diagnostic purposes. There was no significant difference in BM biopsy specimen size between the groups shown to be histologically positive or negative for FL, nor did biopsy specimen size correlate with frequency of detection of a monoclonal B-cell population by immunophenotypic analysis.

We considered the possibility that obtaining unilateral versus bilateral BM biopsy specimens could account for a lower yield of detecting a monoclonal B-cell population by flow cytometry. However, the gain in sensitivity of detection of FL by performing bilateral BM biopsies may not be great. In 3 studies that reported more than 500 patients with all histologic types of non-Hodgkin lymphoma who underwent staging BM examination,33–35 the lymphoma involved bilateral and unilateral BM biopsy specimens in 25% and 10% of patients, respectively. In the group of patients with unilateral lymphoma, as has been stated by Bain,2 there is most likely an equal chance that the first of 2 BM biopsy specimens will be positive or negative for lymphoma. Thus, unilateral BM biopsy may miss lymphoma in approximately 5% of all patients. However, it is likely that bilateral BM biopsy is more valuable for certain histologic types of lymphoma than others.

Among the 114 patients with BM biopsy specimens histologically positive for FL in this study, 76 underwent bilateral BM biopsies resulting in 152 biopsy specimens. Only 3 patients underwent bilateral BM biopsy and had only 1 side that was histologically involved by FL. Our experience in this study does not support the conclusions of older studies advocating the value of bilateral BM biopsies in staging patients with non-Hodgkin lymphomas.33–36 The fact that we focused specifically on FL, whereas the older studies included all histologic types of non-Hodgkin lymphoma, may explain this discrepancy. The frequency of bilateral versus unilateral BM biopsy was also not significantly different between the groups shown to be histologically positive or negative for FL in this study.

We believe the problem of falsely negative immunophenotypic results, defined for the purpose of this discussion as absence of a monoclonal B-cell population in a patient with histologic evidence of FL in the BM biopsy specimen, is likely to be lessened by improvements in staining, acquisition, and gating strategies that were not employed in this study. For instance, by including a CD10 antibody and gating specifically on CD10-positive cells, or on cells positive for both CD10 and CD19, a κ/λ ratio can be determined on a more informative subset of cells than could be done in this study. In many cases increased numbers of cells also can be acquired for analysis (eg, 100 000 instead of the 10 000 to 20 000 often acquired routinely) without altering the workflow of the clinical laboratory. The increased detection rate for very small monoclonal B-cell populations could thereby reduce the number of cases that show falsely negative immunophenotypic results because too few B cells are present to assess. Nevertheless, because our data suggest that inadequate aspiration of FL cells is a significant cause of false-negative immunophenotypic results, in addition to simply lesser extent of involvement by FL, we suspect that the false-negative problem cannot be completely overcome by improved flow cytometry techniques.

In summary, we have reviewed the histologic findings of FL in BM biopsy specimens and correlated these findings with flow cytometry immunophenotypic data derived from corresponding BM aspirates. Our results show that 3-color flow cytometry immunophenotyping can be negative for a monoclonal B-cell population in an appreciable subset of FL patients with histologically positive BM biopsy specimens. Patients with a purely paratrabecular pattern of involvement, or to a lesser degree, patients with less tumor in the BM, are most likely to be negative by immunophenotyping. In addition, flow cytometric immunophenotyping detects evidence of a monoclonal B-cell population in only a small subset of patients with histologically negative BM biopsy specimens. The limitations of 3-color flow cytometric immunophenotyping in this clinical setting, along with the uncertain clinical implications of detecting a monoclonal B-cell population in the absence of histologic evidence of FL, lead us to suggest that this technique, in most cases, does not contribute to the evaluation of staging BM aspiration and biopsy specimens obtained from FL patients.

Zinzani
,
P. L.
Lymphoma: diagnosis, staging, natural history, and treatment strategies.
Semin Oncol
2005
.
32
:(
suppl 1
).
S4
S10
.
Bain
,
B. J.
Bone marrow trephine biopsy.
J Clin Pathol
2001
.
54
:
737
742
.
McKenna
,
R. W.
,
C. D.
Bloomfield
, and
R. D.
Brunning
.
Nodular lymphoma: bone marrow and blood manifestations.
Cancer
1975
.
36
:
428
440
.
Foucar
,
K.
,
R. W.
McKenna
, and
G.
Frizzera
.
et al
.
Bone marrow and blood involvement by lymphoma in relationship to the Lukes-Collins classification.
Cancer
1982
.
49
:
888
897
.
Sandhaus
,
L. M.
,
K. V.
Voelkerding
, and
J.
Dougherty
.
et al
.
Combined utility of gene rearrangement analysis and flow cytometry in the diagnosis of lymphoproliferative disease in the bone marrow.
Hematol Pathol
1990
.
4
:
135
148
.
Wells
,
D. A.
,
M. C.
Hall
, and
H. M.
Shulman
.
et al
.
Occult B-cell malignancies can be detected by three-color flow cytometry in patients with cytopenias.
Leukemia
1998
.
12
:
2015
2023
.
Dunphy
,
C. H.
Combining morphology and flow cytometric immunophenotyping to evaluate bone marrow specimens for B-cell malignant neoplasms.
Am J Clin Pathol
1998
.
109
:
625
630
.
Naughton
,
M. J.
,
J. L.
Hess
, and
M. M.
Zutter
.
et al
.
Bone marrow staging in patients with non-Hodgkin's lymphoma: is flow cytometry a useful test?
Cancer
1998
.
82
:
1154
1159
.
Hanson
,
C. A.
,
P. J.
Kurtin
, and
J. A.
Katzmann
.
et al
.
Immunophenotypic analysis of peripheral blood and bone marrow in the staging of B-cell malignant lymphoma.
Blood
1999
.
94
:
3889
3896
.
Duggan
,
P. R.
,
D.
Easton
, and
J.
Luider
.
et al
.
Bone marrow staging of patients with non-Hodgkin's lymphoma by flow cytometry: correlation with morphology.
Cancer
2000
.
88
:
894
899
.
Dick
,
F.
,
C. D.
Bloomfield
, and
R. D.
Brunning
.
Incidence cytology, and histopathology of non-Hodgkin's lymphomas in the bone marrow.
Cancer
1974
.
33
:
1382
1398
.
Torlakovic
,
E.
,
G.
Torlakovic
, and
R. D.
Brunning
.
Follicular pattern of bone marrow involvement by follicular lymphoma.
Am J Clin Pathol
2002
.
118
:
780
786
.
Bishop
,
P. W.
,
K.
McNally
, and
M.
Harris
.
Audit of bone marrow trephines.
J Clin Pathol
1992
.
45
:
1105
1108
.
Schlette
,
E.
,
L. J.
Medeiros
,
M.
Keating
, and
R.
Lai
.
CD79b expression in chronic lymphocytic leukemia: association with trisomy 12 and atypical immunophenotype.
Arch Pathol Lab Med
2003
.
127
:
561
566
.
Konoplev
,
S.
,
L. J.
Medeiros
, and
C. E.
Bueso-Ramos
.
et al
.
Immunophenotypic analysis of lymphoplasmacytic lymphoma/Waldenstrom macroglobulinemia.
Am J Clin Pathol
2005
.
124
:
414
420
.
Baroni
,
C. D.
,
L.
Manente
, and
M.
Occhionero
.
et al
.
Involvement of the bone marrow by non-Hodgkin's lymphomas: incidence, histology and pathology correlations.
Tumori
1981
.
67
:
191
196
.
Samoszuk
,
M. K.
,
M.
Krailo
, and
Q. H.
Yan
.
et al
.
Limitations of numerical ratios for defining monoclonality of immunoglobulin light chains in B-cell lymphomas.
Diagn Immunol
1985
.
3
:
133
138
.
Witzig
,
T. E.
,
P. M.
Banks
, and
M. J.
Stenson
.
et al
.
Rapid immunophenotyping of B-cell non-Hodgkin's lymphomas by flow cytometry: a comparison with the standard frozen section method.
Am J Clin Pathol
1990
.
94
:
280
286
.
Geary
,
W. A.
,
H. F.
Frierson
, and
D. J.
Innes
.
et al
.
Quantitative criteria for clonality in the diagnosis of B-cell non-Hodgkin's lymphoma by flow cytometry.
Mod Pathol
1993
.
6
:
155
161
.
Morse
,
E. E.
,
H. T.
Yamase
, and
B. R.
Greenberg
.
et al
.
The role of flow cytometry in the diagnosis of lymphoma: a critical analysis.
Ann Clin Lab Sci
1994
.
24
:
6
11
.
Fukushima
,
P. I.
,
P. K. T.
Nguyen
, and
P.
O'Grady
.
et al
.
Flow cytometric analysis of kappa and lambda light chain expression in evaluation of specimens for B-cell neoplasia.
Cytometry
1996
.
26
:
243
253
.
Zardawi
,
I. M.
,
S.
Jain
, and
G.
Bennet
.
Flow-cytometric algorithm on fine-needle aspirates for the clinical workup of patients with lymphadenopathy.
Diagn Cytopathol
1998
.
19
:
274
278
.
Davidson
,
B.
,
B.
Risberg
, and
A.
Berner
.
et al
.
Evaluation of lymphoid cell populations in cytology specimens using flow cytometry and polymerase chain reaction.
Diagn Mol Pathol
1999
.
8
:
183
188
.
Chizuka
,
A.
,
Y.
Kanda
, and
Y.
Nannya
.
et al
.
The diagnostic value of kappa/lambda ratios determined by flow cytometric analysis of biopsy specimens in B-cell lymphoma.
Clin Lab Haematol
2002
.
24
:
33
36
.
Kaleem
,
Z.
,
R. T.
Vollmer
, and
G.
White
.
Clonality evaluation in B-cell lymphoproliferative disorders by flow cytometric immunophenotyping [abstract].
Mod Pathol
2001
.
14
:
152A
.
Reichard
,
K. K.
,
R. W.
McKenna
, and
S. H.
Kroft
.
Comparative analysis of light chain expression in germinal center cells and mantle cells of reactive lymphoid tissues.
Am J Clin Pathol
2003
.
119
:
130
136
.
Vega
,
F.
,
L. J.
Medeiros
, and
W. H.
Lang
.
et al
.
The stromal composition of malignant lymphoid aggregates in bone marrow: variations in architecture and phenotype in different B-cell tumours.
Br J Haematol
2002
.
117
:
569
576
.
Canioni
,
D.
,
P.
Brice
, and
E.
Lepage
.
et al
.
Bone marrow histological patterns can predict survival of patients with grade 1 or 2 follicular lymphoma: a study from the Groupe d'Etude des Lymphomes Folliculaires.
Br J Haematol
2004
.
126
:
364
371
.
Cheson
,
B. D.
,
S. J.
Horning
, and
B.
Coiffier
.
et al
.
Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas: NCI Sponsored International Working Group.
J Clin Oncol
1999
.
17
:
1244
1253
.
Bain
,
B. J.
,
D. C.
Clark
, and
I. A.
Lampert
.
et al
.
Bone marrow pathology, 3rd ed.
Oxford, United Kingdom: Blackwell Science; 2001
.
Brynes
,
R. K.
,
R. W.
McKenna
, and
R. D.
Sundberg
.
Bone marrow aspiration and trephine biopsy: an approach to a thorough study.
Am J Clin Pathol
1978
.
70
:
753
759
.
Islam
,
A.
Manual of Bone Marrow Examination.
Amsterdam, the Netherlands: Overseas Publishers Association; 1997
.
Brunning
,
R. D.
,
C. D.
Bloomfield
, and
R. W.
McKenna
.
et al
.
Bilateral trephine bone marrow biopsies in lymphoma and other neoplastic diseases.
Ann Intern Med
1975
.
82
:
365
366
.
Ebie
,
N.
,
J. M.
Loew
, and
S. A.
Gregory
.
Bilateral trephine bone marrow biopsy for staging non-Hodgkin's lymphoma: a second look.
Hematol Pathol
1989
.
3
:
29
33
.
Juneja
,
S. K.
,
M. M.
Wolf
, and
I. A.
Cooper
.
Value of bilateral bone marrow biopsy specimens in non-Hodgkin's lymphoma.
J Clin Pathol
1990
.
43
:
630
632
.
Haddy
,
T. B.
,
R. I.
Parker
, and
I. T.
Magrath
.
Bone marrow involvement in young patients with non-Hodgkin's lymphoma: the importance of multiple bone marrow samples for accurate staging.
Med Pediatr Oncol
1989
.
17
:
418
423
.

The authors have no relevant financial interest in the products or companies described in this article.

Author notes

Reprints: L. Jeffrey Medeiros, MD, The University of Texas M. D. Anderson Cancer Center, Department of Hematopathology, Box 72, 1515 Holcombe Blvd, Houston, TX 77030 ([email protected])