Year : 2009 | Volume
: 52 | Issue : 3 | Page : 332--338
Bone marrow biopsy in non-Hodgkin lymphoma: A morphological study
Suneet Kumar, Aarathi R Rau, Ramadas Naik, Hema Kini, Alka M Mathai, Muktha R Pai, Urmila N Khadilkar
Department of Pathology, Kasturba Medical College, Mangalore, India
Department of Pathology, Kasturba Medical College, P.O. Box 53, Mangalore - 575 001, Karnataka
Context: Bone marrow (BM) biopsy is an integral part of staging work-up for non-Hodgkin lymphoma (NHL). Aims: To study the characteristics of BM involvement in NHL with respect to incidence, histologic pattern and morphology of infiltration and its discordance with the histology of primary anatomic site. Settings and Design: Forty-nine cases of NHL in which BM biopsy was performed for staging were included in this study, the primary site being classified according to the WHO classification for NHL. Materials and Methods: A prospective study of 49 cases was conducted. Bilateral BM biopsy was obtained from the posterior superior iliac spine. The biopsies were fixed in 10% buffered formalin solution and decalcified using 10% formal - formic acid for 4 - 6 h followed by routine processing. The serial sections were stained by hematoxylin and eosin and reticulin stains. Results: BM biopsy showed involvement by lymphoma in 27 cases (55.10%). Unilateral positivity was found in four cases (14.81% cases). The overall incidence of marrow involvement by NHL was 55.1%. The incidence of involvement was higher in T-cell lymphomas when compared with B-cell lymphomas and predominant pattern of involvement was mixed. Diffuse large B-cell lymphomas had the lowest incidence in all the B-cell lymphomas. A discordant histology between BM and primary anatomic site was found in 29.63% (8/27) of the cases, where it was seen more in follicular lymphomas and diffuse large B-cell lymphomas. Conclusions: Critical examination of BM biopsies can increase the diagnostic accuracy, thereby contributing to the prognosis and appropriate treatment modalities.
|How to cite this article:|
Kumar S, Rau AR, Naik R, Kini H, Mathai AM, Pai MR, Khadilkar UN. Bone marrow biopsy in non-Hodgkin lymphoma: A morphological study.Indian J Pathol Microbiol 2009;52:332-338
|How to cite this URL:|
Kumar S, Rau AR, Naik R, Kini H, Mathai AM, Pai MR, Khadilkar UN. Bone marrow biopsy in non-Hodgkin lymphoma: A morphological study. Indian J Pathol Microbiol [serial online] 2009 [cited 2020 Feb 25 ];52:332-338
Available from: http://www.ijpmonline.org/text.asp?2009/52/3/332/54987
Bone marrow (BM) examination is considered essential in evaluation and staging of non-Hodgkin lymphoma (NHL) at the time of initial diagnosis as well as after therapy. Bilateral iliac crest BM biopsies and step sections for morphological view are advocated for optimal specimen and detection of small focal lesions accompanied by fibrosis that may not be readily detected in aspirate smear preparations, thereby increasing the yield of possible diagnosis and becoming superior to aspiration study.  Several studies regarding the BM involvement in NHL have been reported in relation to the older Rappaport classification, Lukes-Collins classification, Working formulation and Revised European-American Lymphoma classification. , We herein describe in detail the incidence and morphologic and histopathologic patterns of marrow involvement of NHL, which were classified according to the World Health Organization (WHO) classification.
Materials and Methods
Forty-nine cases of NHL in which BM biopsy was performed for staging were included in this study, the primary site being classified according to the WHO classification for NHL. Bilateral BM biopsies, performed under local anesthesia, were obtained using the conventional technique with a Jamshidi needle from the posterior superior iliac spines, fixed in 10% formalin solution and decalcified using 10% formal - formic acid for 4 - 6 h followed by routine processing and paraffin embedding.  In pediatric population, the marrow biopsies were carried out under general anesthesia. Serial sections of 4 - 6 Ám thickness were cut and stained by hematoxylin and eosin (H and E) and reticulin stains.  The pertinent patient details regarding the age, clinical details, diagnosis, blood counts and peripheral blood smears were also studied.
Sections with at least five well preserved marrow spaces were studied for cellularity, normal hematological elements, presence of infiltration if any, the extent, histologic pattern and morphology of infiltration, reticulin fibrosis and other secondary changes. Reticulin was graded from 0 to 4.  The morphology and histologic pattern of infiltration was categorized as diffuse, interstitial, focal (non-paratrabecular), paratrabecular and mixed patterns. Diffuse pattern was defined when there was extensive replacement of the marrow elements, both hematopoietic tissue and fat, so that the marrow architecture was effaced; interstitial pattern, when individual neoplastic cells were interspersed between hematopoietic cells and fat; focal, where nodular aggregates were seen separated by normal hematopoietic marrow; paratrabecular, when lymphoma aggregates were seen immediately adjacent to bony trabeculae. In two cases, immunohistochemistry was performed on the BM.
The histological classification of NHL was based on the WHO classification and the pattern of infiltrate in BM was compared with that of the primary anatomic site.  Thirty-five imprint smears and eight aspiration smears were available and correlated with the biopsy findings. Discordant lymphomas were defined when two or more distinct areas of NHL showed different histologic pattern when compared with the primary anatomic site.
The 49 cases of NHL comprised of 33 male and 16 female patients (M:F = 2:1). The age varied from 2 to 86 years. There were 38 cases of B-cell and 11 cases of T-cell lineage. The overall incidence of marrow involvement by NHL was 55.1% (27/49). Of the B-cell lymphomas, 52.26% cases (21/38), and 54.54% cases (6/11) of the T-cell lymphomas showed BM involvement. Among all the B-cell lymphomas, the lowest incidence of marrow involvement (47.37%) was shown by diffuse large B-cell lymphoma. The incidence of BM involvement in all the sub-types of NHL in relation to WHO classification is shown in [Table 1]. Four cases had unilateral involvement and the remaining 23 had bilateral involvement. Peripheral blood involvement was seen in 7.4% (two cases) of the cases.
The predominant histological pattern of involvement by a lymphomatous infiltrate was mixed (51.85%, 14/27), followed by focal non-paratrabecular [22.22%, 6/27; [Figure 1]], paratrabecular (11.11%,3/27), diffuse and interstitial [7.4%, 2/27 each; [Figure 2] and [Figure 3]. Of the mixed pattern, 13 of the 14 cases had predominantly interstitial involvement and comprised mainly of diffuse large B-cell lymphoma (five cases), follicular lymphoma (three cases) and small lymphocytic lymphoma (SLL) (two cases). Three of the five cases of follicular lymphoma showed predilection for paratrabecular area with paratrabecular only pattern in one case. The other lymphomas showing paratrabecular pattern was peripheral T-cell (PTC) lymphoma, unspecified, enteropathy-associated T-cell lymphoma and mantle cell lymphoma [Figure 4]a-c. In none of the cases of PTC, unspecified, mixed patterns were observed. Diffuse, interstitial and focal patterns were observed in T-cell neoplasms. Focal nodular pattern was found in most of the lymphomas, either as a distinct pattern or in a mixed pattern. Diffuse pattern was mainly concurrent with interstitial pattern. Distribution of different patterns of involvement in various histologic sub-types is given in [Table 2]. A discordant histology between BM and primary anatomic site was found in 29.63% (8/27) of the cases, where it was seen more in follicular lymphomas and diffuse large B-cell lymphomas.
All the involved marrows were normocellular to hypercellular except for one, which was hypoplastic. The normal hematopoietic elements were decreased in cases with diffuse and interstitial involvement. Chemotherapy-induced changes of increased density of blood vessels were noted in one of the cases. Reactive changes of multi-vacuolated fat spaces were present in one case and two cases showed aspiration-induced artefact. An increase in the reticulin (Grade 4) was noted in all the 27 positive cases. The stain highlighted reticulin fibers around the malignant cells. A case of PTC lymphoma, unspecified, showed granuloma formation. Lymphoma cells were evident on 37.14% (13/35) imprint smears and 25% (2/8) of aspiration smears, correlating with the positive marrow biopsy.
Immunohistochemistry of the BM showed one B-cell lymphoma with a monomorphic population while one case was of the T-cell type. The patterns of involvement were highlighted by immunohistochemistry. The T-cell lymphoma was a case of enteropathy-associated T-cell lymphoma. The scant neoplastic infiltrate was initially suspicious on H and E but was distinctly visible with CD 3 immunopositivity. In a case of the B-cell lymphoma, mantle cell lymphoma, a focal nodular pattern of lymphoma cells was seen on H and E-stained sections, but CD 5 and CD 20 positivity clearly showed an additional focal paratrabecular arrangement.
BM involvement by malignant lymphoma indicates stage IV disease and generous bilateral BM biopsy had been recommended, trephine biopsy being preferred to marrow aspiration for detecting marrow infiltration. , The incidence and clinical importance of marrow involvement varies according to various histologic sub-types. We used the WHO classification to classify the histology on the primary anatomic site for correlating the marrow histology.
In this study, BM was involved by lymphoma in 55.1% cases. Other studies have described a considerable variation in the BM involvement in lymphoproliferative disorders, ranging between 27.6 and 53%. [1,,,, The variation in previous studies and the present one can be attributed to inclusion of unequal proportions of patients with early and advanced disease. In addition, the inclusion of different proportions of various histological sub-types of lymphomas in the different studies may account for this variation as the incidence of BM pathology varies greatly according to the histological sub-types.  In the present study, the incidence of the T-cell lymphomas was greater than that of B-cell lymphomas, which was also observed by Jeong et al . 
BM morphology still continues to be the gold standard in evaluating marrow pathology in lymphoma, immunohistochemistry and molecular studies being supplementary techniques. , In the present study, the size of the infiltrating neoplastic cells was considered small when they were slightly larger than the small lymphocytes, having coarsely clumped nuclear chromatin, insignificant nucleoli and scant cytoplasm. The cells were considered large when their nuclear size was larger than that of a histiocyte nucleus. They were designated as of intermediate size if their nuclear size was comparable to that of the histiocyte. The nodular aggregates were difficult to differentiate from the reactive lymphoid nodules, which are generally small with well-defined margins surrounded by a mantle zone and having a polymorphous cell population, made up of predominantly small lymphocytes and smaller number of immunoblasts, macrophages and plasma cells. A neoplastic nodule is larger, with ill-defined margins, often extending around the fat cells, and has a relatively homogenous cellular composition.
The predominant histological pattern of involvement by a lymphomatous infiltrate was mixed, which was similar to that of Arber and George.  However, this varied from the findings of Foucar et al. ,  who found the marrow involvement to be predominantly focal (70%).  The difference could be attributed to the larger number of cases studied by them and the fact that there were more follicular lymphomas in their study.
Of the nine cases of diffuse large B-cell lymphoma, four cases showed focal nodular only pattern and five cases showed a mixed pattern, comprising of mixed interstitial, focal and diffuse pattern. It is important to distinguish the focal nodular pattern of the large cell lymphoma from that of an occult follicular lymphoma, wherein the prognosis of the former is dismal and does not affect the latter unless it is a morphologic discordance.  Incidence of marrow involvement of diffuse large B-cell lymphoma was low when compared with other B-cell lymphomas, which was also observed in the present study. Marrow infiltration suggests a poor prognosis, especially the diffuse pattern that was seen in two of the cases.
One of the two cases of B-lymphoblastic lymphoma showed marrow involvement that was mixed focal and interstitial. The differential diagnosis of precursor B-cell lymphoma includes acute myeloid leukemia, T-lymphoblastic lymphoma and mature B-cell lymphomas, which comprises of Burkitt lymphoma, large cell lymphomas and blastoid variant of mantle cell lymphoma. The lymphoblasts have a high nucleocytoplasmic ratio and are more regular than myeloblasts. The high mitotic count of Burkitt lymphoma, large pleomorphic cells of large cell lymphomas and focal infiltration of the low-grade lymphomas are some of the differentiating features. The reticulin fibers are increased in the areas of lymphoblastic infiltration and slowly regress following remission of the leukemia.
SLL had a high incidence of marrow involvement (3/3), with focal nodular only pattern in one and mixed pattern in others with no paratrabecular involvement in any, as observed by Foucar et al .  However, no prognostic significance has been found in BM histology in SLL.
Lymphoplasmacytic lymphoma, previously categorized as lymphoplasmacytoid lymphoma, is characterized by diffuse, focal, interstitial, paratrabecular and mixed patterns of infiltration by small mature lymphocyte tumor cells with plasma cell differentiation. Reticulin fibers are frequently increased in the areas of infiltration. Both cytological features and patterns of infiltration are related to prognosis, diffuse pattern having worst prognosis and nodular with best prognosis with the mixed pattern in between. Mixed focal and interstitial patterns were observed in the present case. A follicular lymphoma at a distant anatomic site, splenic marginal lymphomas and chronic lymphocytic leukemia has to be considered in the differential diagnosis. Lack of proliferation centers, paratrabecular infiltration, mature plasma cell component and a significant mast cell infiltration distinguish a lymphoplasmacytic lymphoma from a chronic lymphocytic leukemia.
A discordant histology between the BM and the primary anatomic site was encountered in 29.63% of the cases, especially in follicular lymphomas and diffuse large B-cell lymphomas, which is similar to 16-40% reported in the literature with this same histologic subtype of lymphomas. , Few cases have shown a more aggressive histologic pattern in the primary site than in BM. Reticulin fibers are significantly increased in infiltrated areas of follicular lymphoma. The paratrabecular infiltration is usually associated with follicular lymphoma but has also been described in marginal zone lymphoma and mantle zone lymphoma, which was also observed in the present study. However, marginal zone lymphoma and mantle zone lymphoma show follicle formation with a benign germinal center.  CD10 and bcl-2 positivity in the follicles rule out a marginal zone lymphoma and CD5 and cyclin D1 negativity, a mantle cell lymphoma. It is important to observe the follicular and paratrabecular pattern of follicular lymphoma because small lymphoid aggregates in the interstitium in the absence of paratrabecular involvement favor a benign nature. In the present study, the follicular and paratrabecular pattern was present concurrently with diffuse and interstitial patterns in three of the cases. The diffuse pattern observed in other two cases could be due to fusion of the follicles.
Mantle cell lymphoma, termed as centrocytic lymphoma in updated Kiel classification, has increased frequency for marrow involvement with small lymphoid cells, some with round and others with indented or clefted nuclei. Reticulin fibers are increased in infiltrated areas. The morphology closely resembles other small B-cell lymphoproliferative diseases, such as chronic lymphocytic leukemia, prolymphocytic leukemia, follicular lymphoma and marginal zone lymphoma. Accurate diagnosis is possible only with immunophenotyping.
One case of Burkitt-like lymphoma showed interstitial infiltration. Correctly recognizing the medium-sized cells with round nuclear outline and numerous mitotic figures are helpful in making the diagnosis. A few cases of large B-cell lymphoma can have a similar morphology; however, a high proliferative rate favors the former.
No specific pattern of involvement was observed in T-cell lymphomas. The T-cell lymphoma infiltration in nodular pattern comprise nodules with ill-defined margins, differing from that of B-cell and reactive follicles. The reactive changes are more frequent in T-cell lymphomas, which include eosinophilia, plasmacytosis, macrophage clusters, vascular proliferation, hemophagocytosis, granuloma formation and reactive follicle hyperplasia.
Adult T-cell leukemia/lymphoma, a separate entity in WHO classification, has a high frequency of marrow involvement. When the extent of involvement is less, morphology alone is difficult to recognize; CD3 immunoreactivity may be helpful in identification. The marrow infiltration of PTC lymphoma, unspecified, has to be carefully evaluated with immunophenotyping of the tumor cells to rule out a Hodgkin disease and T-cell-rich variant of large B-cell lymphoma. The reticulin fibers are increased in areas of infiltration and also mildly increased in non-infiltrated areas. It was difficult to comment on the incidence of anaplastic large cell lymphoma, T-cell type, angioimmunoblastic lymphoma and cutaneous T-cell lymphoma on marrow involvement due to the small sample size.
An increase in the reticulin (Grade 4) was noted in all the 27 positive cases. The stain highlighted reticulin fibers around the malignant cells. Increased reticulin deposition, restricted to the area of marrow infiltration, is common in lymphoma.  Collagen fibrosis is less common. This finding was consistently seen in all the cases showing infiltration in the present series; hence, elucidating the importance of reticulin stain on the marrow sections. We observed that the reticulin stain helped in detection of scanty focal or interstitial infiltrate, which may not be apparent in routine H and E stain.
Peripheral blood involvement was seen in 7.4% (two cases) of the cases; one case of adult T-cell leukemia and a PTC lymphoma, unspecified. The presence of atypical lymphoid cells was accompanied by increased total leukocyte count and percentage of lymphocytes, as observed by Arber and George.  The peripheral blood involvement was lower when compared with that of Jeong et al. and Lee et al. ,,
Immunohistochemistry can aid in increasing the diagnostic accuracy in cases where the material is scanty or crushed. Talaulikar et al.  recently reported the role of immunohistochemistry on routine BM trephine biopsies in improving the detection of occult lymphomas, especially diffuse large B-cell lymphomas, immunostaining itself being a stronger predictor of survival than routine histology.  This was evident in two of our cases, where marrow involvement was diagnosed using immunohistochemistry in a case of enteropathy-associated T-cell lymphoma and an additional paratrabecular pattern highlighted using immunohistochemistry in the marrow of mantle cell lymphoma. Immunophenotyping using flow cytometry and polymerase chain reaction (PCR)-based clonality on BM are the other ancillary tests preferred to confirm lymphomatous infiltrate in the marrow. ,
Chemotherapy-induced changes of increased density of blood vessels were noted in one of the cases. Reactive changes of multivacoulated fat spaces were present in one case and two other cases showed aspiration-induced artefact. Lymphoma cells were evident and better appreciated on 37.14% (13/35) imprint smears and 25% (2/8) of aspiration smears, correlating with the positive marrow biopsy. In cases of focal infiltration and fibrosis, the positivity rate of imprint smears and aspiration smears are relatively low. However, imprint smears may be useful as they can be made available at no additional cost and effort, thereby aiding in marrow interpretation as far as the cytologic features are concerned.
Four cases were diagnosed unilaterally and the remaining 23 on bilateral biopsies, bilateral biopsies increasing the positivity rate in most sub-types of NHL, as observed by Jung et al. ,  Brunning et al . However, Campbell et al.  recommended that a unilateral core biopsy of 20 mm can eliminate the need of bilateral biopsy. 
Because cytologic features depend greatly on tissue processing, minor cytologic atypia may not be appreciated on inadequately processed marrows. Numerous problems and pitfalls exist in the evaluation of lymphomatous infiltration of the marrow. Interstitial infiltration can occur in neoplastic and reactive conditions. Focal nodular infiltrate may resemble a reactive nodular hyperplasia. However, paratabecular infiltration or an extensive diffuse infiltration suggests neoplasia. Immunohistochemical staining too has limited value wherein nodule containing a monotonous population of B cells may be neoplastic but a mixed population of T- and B cells may be either neoplastic or reactive. A 3-color flow cytometry immunophenotyping can give false-negative results with patients with follicular lymphoma positive in marrow biopsy. Patients with only paratrabecular involvement and or with less tumor burden in the marrow are likely to be negative by immunophenotyping.  Kang et al.,  in 170 cases of NHL, reported that PCR-based clonality does not have prognostic significance or support the marrow involvement of NHL. 
To conclude, this study highlights the incidence and different patterns of involvement in the marrow, with discordance between BM morphology and primary site histology. BM trephine biopsy showed involvement by lymphoma in 55.10% cases, with unilateral positivity in only four cases. Therefore, bilateral and generous BM biopsy should be performed routinely for adequacy and to identify tiny focal involvement, thereby increasing the yield of possible diagnosis and for prognostic significance. The incidence of involvement was higher in T-cell lymphomas when compared with B-cell lymphomas and the predominant pattern of involvement was mixed. Diffuse large B-cell lymphomas had the lowest incidence in all the B-cell lymphomas. A discordant histology between BM and primary anatomic site was found in 29.63% of the cases, where it was seen more in follicular lymphomas and diffuse large B-cell lymphomas. Reticulin stain carried out for all BM trephines highlighted scant interstitial infiltrates, which could be detected with ease if a correlation between the reticulin and the H and E sections was performed. Adjuvant immunohistochemistry, when performed in marrows with scant cellularity or crush artifact, increased the diagnostic accuracy by unmasking obscured patterns and morphology.
|1||Brunning RD, Bloomfield CD, McKenna RW, Peterson L. Bilateral trephine biopsies in lymphoma and other neoplastic conditions. Ann Intern Med 1975;82:365-6.|
|2||Foucar K, McKenna RW, Frizzera G, Brunning RD. Bone marrow and blood involvement by lymphoma in relationship to the Luke-Collins classification. Cancer 1982;49:888-97.|
|3||Jeong SY, Chang YH, Lee JK, Hong YJ, Hong SI, Lee SS. Incidence and histologic patterns of bone marrow involvement of malignant lymphoma based on the World Health Organization classification-a single institution study. Korean J Lab Med 2007;27:383-7.|
|4||Culling CF, Allison RT, Barr WT. Connective tissue. In: Cellular Pathology Techniques. 4 th ed. London: Butterworth and Co. ltd; 1985. p. 172-3.|
|5||Gordon H, Sweets HH. A simple method of the silver impregnation of the reticulum. Am J Pathol 1936;12:545. [Cited in Bancroft JD, Cook CH. Manual of Histology Techniques and their Diagnostic Applications. Singapore: Churchill Livingstone; 1994.|
|6||Bauermeister DE. Quantification of bone marrow reticulin-a normal range. Am J Clin Pathol 1971;56:24-31. |
|7|| Aster JC. Diseases of white blood cells, lymph nodes, spleen and thymus. In: Kumar V, Abbas AK, Fausto N, editors. Robbins and Cotran Pathologic basis of disease. 7 th ed. Philadelphia: Saunders; 2004. p. 661-709.|
|8||Bartl R, Frisch B, Burkhardt R, Jδger K, Pappenberger R, Hoffmann-Fezer G. Lymphoproliferations in bone marrow: i0 dentification and evaluation, classification and staging. J Clin Pathol 1984;37:233-54.|
|9||Bain BJ, Clark DM, Lampert IA, Wilkins BS, editors. Bone marrow pathology. 3 rd ed. UK: Blackwell Science Ltd; 2001.|
|10||Varma N, Dash S, Sarode R, Marwah N. Relative efficacy of bone marrow trephine biopsy sections as compared to trephine imprints and aspiration smears in routine hematological practice. Indian J Pathol Microbiol 1993;36:215-26.|
|11||Schmid C, Isaacson PG. Bone marrow trephine biopsy in lymphoproliferative diseases. J Clin Pathol 1992;45:745-50.|
|12||Arber DA, George TI. Bone marrow biopsy involvement by non- Hodgkin's lymphoma: f0 requency of lymphoma types, patterns, blood involvement, and discordance with other sites in 450 specimens. Am J Surg Pathol 2005;29:1549-57.|
|13||Juneja SK, Wolf MM, Cooper IA. Value of bilateral bone marrow biopsy specimens in non-Hodgkin's lymphoma. J Clin Pathol 1990;43:630-2.|
|14||Lee WI, Lee JH, Kim IS, Lee KN, Kim SH. Bone marrow involvement by non-Hodgkin's lymphoma. J Korean Med Sci 1994;9:402-8.|
|15||Kang YH, Park CJ, Seo EJ, Huh J, Klm SB, Kang YK, et al . Polymerase chain reaction-based diagnosis of bone marrow involvement in 170 cases of non-Hodgkin lymphoma. Cancer 2002;94:3073-82.|
|16||Torlakovic E, Torlakovic G, Brunning RD. Follicular pattern of bone marrow involvement by follicular lymphoma. Am J Clin Pathol 2002;118:780-6.|
|17||Talaulikar D, Dalhstrom JE, Shadbolt B, Broomfield A, McDonald A. Role of immunohistochemistry in staging diffuse large B-cell lymphoma. J Histochem Cytochem 2008;56:893-900.|
|18||Iancu D, Hao S, Lin P, Andersen K, Jorgensen JL, McLaughlin P, et al . Follicular lymphoma in staging bone marrow specimens: c0 orrelation of histologic findings with the results of flow cytometry immunophenotypic analysis. Arch Pathol Lab Med 2007;131:282-7.|
|19||Campbell JK, Matthews JP, Seymour JF, Wolf MM, Juneja SK; Australasian Leukaemia Lymphoma Group. Optimum trephine length in the assessment of bone marrow involvement in patients with diffuse large cell lymphoma. Ann Oncol 2003;14:273-6.|