Follicular lymphoma

* Context.-The World Health Organization classification recommends categorizing grade 3 follicular lymphomas based on the presence of centrocytes (grade 3A) or of sheets of centroblasts (grade 3B). The clinical significance of this practice is not known.

Objective.-To determine whether grade 3 follicular lymphoma subtype is associated with prognosis.

Design.-Multi-institutional retrospective case series.

Main Outcome Measure.-Overall survival.

Results.-Forty-five cases of grade 3 follicular lymphoma without diffuse large B-cell lymphoma were studied (35 cases of grade 3A, 10 cases of grade 3B) from 21 men and 24 women (median age, 67 years; mean age, 63.8 years; range, 26-86 years). Follow-up information from the time of diagnosis was available in all patients, with a median follow-up time of 24 months (mean, 34 months; range, 2-115 months). Treatment information was available in 40 patients. There was no difference in age (P = .45, Wilcoxon test) or stage (P = .76, Fisher exact test) between patients with follicular lymphoma of grade 3A or grade 3B. Furthermore, the Cochran-Armitage test for trends showed no evidence that the proportion of patients with follicular lymphoma grade 3A or 3B increased or decreased with increasing stage at presentation. Kaplan-Meier analysis showed a median overall survival of 44 months from the time grade 3 follicular lymphoma was diagnosed, with no significant difference between cases diagnosed as grade 3A or grade 3B (P = .14, log-rank test). Univariable Cox proportional hazards modeling showed no evidence that an anthracycline-containing chemotherapy regimen or history of lower grade follicular lymphoma affected overall survival.

Conclusions.-In this retrospective series, subclassification of grade 3 follicular lymphoma into type 3A and 3B categories had limited clinical and prognostic significance. However, the study was limited by lack of statistical power. Since morphology often provides clues for progress in defining biologic differences, subtyping may still be useful, particularly in the setting of prospective clinical studies.

(Arch Pathol Lab Med. 2004;128:863-868)

Follicular lymphoma (FL) represents 25% of all adult non-Hodgkin lymphomas in the Western world.1 It is a clinically indolent lymphoma that generally presents at high clinical stage, and the expected median survival is approximately 8 to 10 years, a figure that has not changed substantially in the last 30 years.2 The putative cell of origin is the germinal center B cell. Central to the pathogenesis of this lymphoma is t(14;18)(q32;q21).1 This translocation results in overexpression of the bcl-2 gene, causing resistance to apoptosis.4 Histologically, FL is a neoplasm of centrocytes and centroblasts showing at least partial follicular architecture. The tumor cells express monotypic surface immunoglobulin, B-cell antigens CD9 19 and CD20, CDlO, and Bcl-6. The proportion of centrocytes and centroblasts can vary considerably, and over the years cytologie grading has been shown to be of prognostic and therapeutic significance. For example, the International Working Formulation (IWF) showed outcome differences based on the proportion of large cells present in the follicles. Follicular large cell lymphoma was placed into the intermediate-grade category, along with diffuse large cell lymphoma, whereas follicular small cleaved and follicular mixed small cleaved and large cell lymphomas were considered low-grade lymphomas.5

Grading of FL has, however, been fraught with irreproducibility problems. The system of Mann and Berard was shown to be reproducible and was recommended by the International Lymphoma Study Group in the Revised European-American Classification of Lymphoid Neoplasms.6,7 In this system, FL grades are assigned according to the number of large cells (centroblasts) in 10 different representative neoplastic follicles not selected for those with the most numerous large cells. The count is based on a high-power field (HPF) of 0.959 mm^sup 2^, which is defined as a field resulting from use of an ocular with an 18-mm^sup 2^ field of view at ×lO magnification and ×40 objective (with appropriate correction factors to compensate if different field oculars are used). Follicular lymphomas with O to 5 centroblasts per HPF are classified as a grade 1, with 6 to 15 centroblasts per HPF as grade 2, and with more than 15 centroblasts per HPF as grade 3. The recent World Health Organization (WHO) classification further proposes to subclassify grade 3 into grades 3A and 3B, based on the presence or absence of centrocytes in the neoplastic follicles.8

It is uncertain whether there is any biologic or clinical utility to subdividing grade 3 FLs, as suggested by the WHO classification. Therefore, we decided to retrospectively evaluate the importance of morphology-based subclassification of FL grade 3 into FL grades 3A and 3B with respect to overall survival in a series of 45 cases of grade 3FL.

MATERIALS AND METHODS

Excisional biopsy specimens of grade 3 FL without a diffuse large B-cell component were retrieved from the pathology files of the Cleveland Clinic Foundation (Cleveland, Ohio; n = 21) and Cross Cancer Center (Edmonton, Alberta; n = 24). The cases were categorized as FL grade 3A or 3B according to WHO criteria after a panel of 3 hematopathologists reached consensus. Observers were blinded as to outcome. Olympus BX-40 or BH-2 microscopes with ×40 objectives and ×lO oculars (20-mm field of view) were used, with correction for expanded field of view in centroblast counting when necessary. Follicular lymphoma grade 3B follicles were essentially composed of sheets of centroblasts, while FL grade 3A had remaining cleaved cells present, interspersed among the centroblasts. Only cases showing pure follicular pattern on hematoxylin-eosin-stained sections without diffuse areas were included in this study. Clinical data, such as demographics, Ann Arbor stage at presentation, history of previous diagnosis of lower grade FL, and use of anthracycline in treatment, and follow-up data were evaluated by chart review.

Statistical analysis was performed at the Collaborative Biostatistics Center at the Cleveland Clinic Foundation using the SAS (Gary, NC) software package. Histologie subtypes were compared for age using a Wilcoxon rank sum test, and categorical values were compared using a [chi]^sup 2^ or Fisher exact test, as appropriate. In addition, clinical stage was compared between the 2 subtypes using the Cochran-Armitage test for trend. Kaplan-Meier survival estimates were generated to compare the differences in survival time between patients with FL grades 3A and 3B. Bootstrapped confidence limits on the median survival times were also calculated. In addition, univariable Cox proportional hazards regression models were created to estimate the hazard ratio associated with each of the following factors: FL grade 3A versus FL grade 3B, previous FL diagnosis, and anthracycline use.

RESULTS

Clinical Findings

The 45 patients included 21 men and 24 women with a mean age of 63.8 years (median, 67 years; range, 26-86 years). Lymph node biopsy sites were as follows: cervical (n = 14), axillary (n = 12), supraclavicular (n = 5), inguinal (n = 5), abdominal/retroperitoneal (n = 3), preparotid (n = 1), occipital (n = 1), iliac (n = 1), submental (n = 1), and epitrochlear (n = 1). One biopsy was from the thyroid gland. Review of patient records showed that 6 patients had a prior history of low-grade FL, all of which were diagnosed as FL grade 3A. Stage information was available in 42 patients, and most were deemed to have high-stage disease (stage III or IV); 3 were diagnosed with stage I, 8 with stage II, 6 with stage III, and 25 with stage IV.

Initial treatment information was available for 40 of the patients and was divided into those who received anthracycline-containing regimens and those who did not. Clinical follow-up with regard to overall survival was available in all patients. There was a mean follow-up time of 34 months (median, 24 months; range, 2-115 months). Details of individual patient data are shown in Table 1.

Histopathologic Findings

Review of hematoxylin-eosin-stained sections and classification of purely follicular lymphomas resulted in 35 FL grade 3A and 10 FL grade 3B cases. The FL grade 3A cases showed numerous centroblasts, but also contained easily identifiable centrocytes, whereas, by definition, FL grade 3B cases demonstrated follicles composed of sheets of centroblasts (Figure 1). The latter cases were generally easily recognizable at intermediate magnification (×lO or ×20 objective).

Outcome and Statistical Analysis

The patient groups (FL grade 3A and FL grade 3B) appeared similar with respect to age at presentation and stage. When comparing patients with FL grade 3A and FL grade 3B, there appeared to be no difference in age (P = .45, Wilcoxon rank sum) or clinical stage (P = .81, Cochran-Armitage test for trend). At last follow-up, 21 deaths had occurred in the entire cohort, with a median overall survival of 44 months.

To address the question of whether there were overall survival differences between FL grades 3A and 3B, we compared the Kaplan-Meier estimates from the 2 groups using the log-rank test. When broken down by histologie subtype (FL grade 3A vs FL grade 3B), the median survival estimates were 46 months (95% CI, 32-73 months) and 22 months (95% CI, 18-35 months), respectively (Table 2, Figure 2). The log-rank test showed no significant difference in survival between FL grades 3A and 3B (P = .14). The survival at 3 years was 64.9% (95% CI, 46.4%83.5%) in the FL grade 3A group versus 22.0% (95% CI, 0.00%-58.7%) in the FL grade 3B group. Univariable Cox proportional hazards modeling for FL grade 3 subtype, history of previous FL, and anthracycline-containing therapy showed no significant differences (Table 3). Thus, we can also conclude that prior FL diagnosis and anthracycline-based therapy did not influence the relationship between FL grade 3 subtype and survival. Indeed when cases with a prior history of FL were excluded from the analysis, there was no difference in overall survival between patients with FL grade 3A and FL grade 3B (P = .16, logrank test).

COMMENT

Follicular lymphoma is one of the most common types of non-Hodgkin lymphomas, comprising 22% of all non-Hodgkin lymphomas in a recent survey of more than 1400 lymphomas." A lymphoma primarily occurring in adults, FL is considered a clinically indolent lymphoma with a prolonged clinical course and with multiple relapses and recurrences. Cytologie grading of FL has long been suggested to be an important prognostic factor and has been a characteristic feature of many lymphoma classification systems over the years.5,10-13 The IWF demonstrated successively shorter survival in patients with FL stratified into 3 groups, based on the proportion of small cleaved and large cells.5 Although considered a single entity, the Revised European-American Classification of Lymphoid Neoplasms recommended a similar 3-tiered grading system6 .6 6 Notably, not all recent series have demonstrated survival differences between cytologie grades.14,15 The WHO classification recommends a similar cytologie breakdown, but also proposes subtyping grade 3 FL into A and B types, with the latter composed of sheets of centroblasts for investigative purposes.8 We retrospectively reviewed our experience with FL grade 3 to determine whether any differences in survival could be discerned that would support this practice.

Grade 3 FLs make up only a minority of all FLs. In 2 recent large series, grade 3 FLs represented 17.8% and 28.9% of FLs.9,15 This represents a relatively higher proportion than some older series (11.2% for the IWF, group D), perhaps due to different criteria for cytologie stratification.5 There are few data regarding the proportion of FL grade 3A in relation to grade 3B. In our series, which excluded diffuse large B-cell lymphomas (DLBCLs) with FL grade 3, 22% of cases of grade 3 FLs were grade 3B. Thus, FL grade 3B is an uncommon subtype of FL, perhaps representing roughly 5% of FLs. Ott and colleagues6 16 6 reported a series of FL that, when excluding DLBCL with grade 3 FL, contained 19% (15/77) FL grade 3. Of these 15 grade 3 FLs, 4 (26.7%) were FL grade 3B, a proportion quite similar to the current series.

The clinical features of our cohort of patients with grade 3 FL do not seem to differ substantially from those of even older published series of FL. However, there may be small differences, particularly with age of the patients. Looking at the IWF data for group D (follicular large cell), we see that our median patient age of 67 years is slightly older than the 55 years of the IWF series.5 Our male-female ratio was lower (0.9 vs 1.8), and the median survival was similar (3.7 years in our series vs 3.0 years). Most (86%) of our patients had high-stage disease (stage III or IV), again similar to the IWF data (73%).s These data are also similar to a more recent series. Rodriguez and colleagues17 reported a series of 62 patients with follicular large cell lymphoma with a slightly older median age of 57 years (male-female ratio, 1.1) and a longer median overall survival of 5.1 years; 76% were high stage.

When comparing patients with FL grades 3A and 3B, the FL grade 3B group appeared to have a slightly shorter overall survival, although this difference was not statistically significant. Of note, there appeared to be no plateau in the curves, similar to lower grade FL, suggesting that pure FL grade 3 may not be curable with current therapies. Prospective trials with uniform treatment protocols and larger patient populations to add statistical power would, of course, be required to substantiate these observations. When excluding the few patients with a prior history of lower grade FL, there was still no difference. Furthermore, whether or not anthracycline-containing regimens were used did not appear to confound the analysis. Thus, within the limitations of this study, our data suggest that cytologie subtype in grade 3 FL may not correlate well with overall survival.

Grade 3 FL has certain biologic features that distinguish it from lower grades. For example, lower grade FLs appear to have a higher percentage of cases that express CDlO, with more than 80% of grade 1 FL cases being CDlO-positive compared to less than 20% of grade 3 FLs.18 The frequency of Bcl-2 expression as well as presence of the t(14;18) also decrease with increasing cytologie grade.1"2 While our data suggest that there is presently no added prognostic value to FL grade 3 subtyping, this does not obviate the need to do so for investigative purposes. In fact, there is some evidence to suggest that FL grade 3B is biologically distinct from FL grade 3A. Ott and colleagueslh recently presented evidence to suggest biologic differences between FL grades 3A and 3B, with or without a DLBCL component. Although there were too few pure FL grade 3B cases to compare separately, FL grade 3B with or without a DLBCL component appeared to have a lower expression of CDlO and cytoplasmic immunoglobulins than cases of FL grade 3A. There also appeared to be some cytogenetic differences, with FL grade 3B, with or without a DLBCL component, having a lower proportion of cases with the t(14;18) and more cases showing +Iq than seen in FL grade 3A (12.5% vs 73% and 33% vs 0%, respectively).16 Bosga-Bouwer and colleagues21 further supported this concept by recently proposing cytogenetic subgroups of FL grade 3B based on presence of the t(14; 18)(q32;q21) and abnormalities of 3q27, the locus of bcl-6. It may be that one should consider FL grade 3A, pure FL grade 3B, and DLBCL with FL grade 3 (A or B) as separable variants of FL, each with their own biologic features that reflect underlying genetic progression events. Alternatively, like DLBCL,24 grade 3 FL may also be a heterogeneous disorder that might better be molecularly subtyped, rather than morphologically subtyped. Because of the retrospective nature of this study and lack of sufficient tissue blocks, we were unable to characterize phenotypic or molecular genetic features further.

Despite the proposed genetic differences, it is still uncertain whether there are clinically apparent differences. Recently, during the preparation of this article, an investigation involving a larger cohort of patients than used in our study also was unable to demonstrate outcome differences between cases of FL grade 3A and grade 3B, although diffuse areas appeared to be significant.25 It should be emphasized that our data are retrospective, and prospective studies should also be done to confirm this finding and to examine other features, such as relapse-free survival time. We recognize that our number of cases is small and lacks statistical power. Given the limited sample size, an optimistic post-hoc power analysis of these data suggests a power of 0.25 to detect the observed 24-month difference. However, these are uncommon lymphomas and when considering grade 3 FLs without diffuse components, 45 cases is in actuality among the largest numbers in recent series.16,25

In summary, we present a series of pure FL grade 3. Subtyping into FL grades 3A and 3B subtypes, as proposed by the WHO classification, did not appear to be useful as a prognostic factor for overall survival. Previous lower grade FL or treatment with anthracycline-containing regimens did not appear to confound the analysis. Despite our findings, it may be that important biologic differences between these subtypes exist. Since morphologic clues often provide leads for further investigation, it is premature to argue against the continued practice of subtyping. Further characterization of such cases at the genetic and protein-expression levels may yield clinically important factors that do predict outcome.

References

1. Armitage JO, Weisenburger DD, tor the Non-Hodgkin's Lymphoma Classification Project. New approach to classifying non-Hodgkin's lymphomas: clinical features of the major histologie subtypes. J Clin Oncol. 1998;16:2780-2795.

2. Horning SJ. Natural history of and therapy for the indolent non-Hodgkin's lymphomas. Semin Oncol. 2003:20:75-88.

3. Tsujimoto Y, Cossman J, Jaffe E, Croce CM. Involvement of the bcl-2 gene in human follicular lymphoma. Science. 1985:228:1440-1443.

4. Hockenbery D, Nunez G, Milliman C, Schreiber RD, Korsmeyer SJ. Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature. 1990:348:334-336.

5. The Non-Hodgkin's Lymphoma Pathologic Classification Project. National Cancer Institute sponsored study of classifications of non-Hodgkin's lymphomas: summary and description of a working formulation for clinical usage. Cancer. 1982:49:2112-2135.

6. Harris NL, Jaffe ES, Stein H, et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood. 1994:84:1361-1392.

7. Mann RB, Berard CW. Criteria for the cytologie subclassification of follicular lymphomas: a proposed alternative method. Hematol Oncol. 1983:1:187-192.

8. Jaffe ES, Harris NL, Stein H, Vardiman J, eds. Tumours ot Haemjtopoeitic and Lymphoid Tissues. Lyon, France: IARC Press; 2001. World Health Organization Classification of Tumours; vol 3.

9. The Non-Hodgkin's Lymphoma Classification Project. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin's lymphoma. Blood 1997;89:3909-3918.

10. Dorfman RF. Classification of non-Hodgkin's lymphomas lletterl. Lancet. 1974;2:96l-962.

11. Dorfman RF. The non-Hodgkin's lymphomas. In: The Reticuloendothelial System. Baltimore, Md: Williams & Wilkins; 1975:262-281.

12. Lukes RJ, Collins RD. Immunologie characterization of human malignant lymphomas. Cancer. 1974;34(suppl):503.

13. Rappaport H. Tumors of the Hematopoietic System. Washington, DC: Armed Forces Institute of Pathology; 1966.

14. Miller TP, LeBlanc M, Grogan TM, Fisher Rl. Follicular lymphomas: do histologie subtypes predict outcome? Hematol Oncol CUn North Am. 1997; 11 : 893-900.

15. Pittaluga S, Bijnens L, Teodorovic I, et al. Clinical analysis of 670 cases in two trials of the European Organization for the Research and Treatment of Cancer Lymphoma Cooperative Group subtyped according to the Revised European American Classification of Lymphoid Neoplasms: a comparison with the Working Formulation. Blood. 1996:87:4358-4367.

16. Ott C, KatzenbergerT, Lohr A, etal. Cytomorphologic, immunohistochemical, and cytogenetic profiles of follicular lymphoma: 2 types of follicular lymphoma grade 3. Blood. 2002:99:3806-3812.

17. Rodriguez ), Mclaughlin P, Fayad L, et al. Follicular large cell lymphoma: long-term follow-up of 62 patients treated between 1973-1981. Ann Oncol. 2000:11:1551-1556.

18. Eshoa C, Perkins S, Kampalath B, Shidham V, juckett M, Chang CC. Decreased CDlO expression in grade III and in interfollicular infiltrates of follicular lymphomas. Am I Clin Pathol. 2001:115:862-867.

19. Lai R, Arber DA, Chang KL, Wilson CS, Weiss LM. Frequency of bcl-2 expression in non-Hodgkin's lymphoma: a study of 778 cases with comparison of marginal zone lymphoma and monocytoid B-cell hyperplasia. Mod Pathol. 1998:11:864-869.

20. Weisenburger DD, Gascoyne RD, Bierman PJ, et al. Clinical significance of the t(14;18) and BCL2 overexpression in follicular large cell lymphoma. Leuk Lymphoma. 2000;36:513-523.

21. Yunis JJ, Frizzera C, Oken MM, McKenna J, Theologides A, Arnesen M. Multiple recurrent genomic defects in follicular lymphoma: a possible model for cancer. N Engl J Med. 1987;316:79-84.

22. Offit K, Jhanwar SC, Ladanyi M, Filippa DA, Chaganti RS. Cytogenetic analysis of 434 consecutively ascertained specimens of non-Hodgkin's lymphoma: correlations between recurrent aberrations, histology, and exposure to cytotoxic treatment. Genes Chromosomes Cancer. 1991;3:189-201.

23. Bosga-Bouwer AC, van lmhoff CW, Boonstra R, et al. Follicular lymphoma grade 3B includes 3 cytogenetically defined subgroups with primary t(14;18), 3q27, or other translocations: t(14;18) and 3q27 are mutually exclusive. Blood. 2003:101:1149-1154.

24. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000:403:503-511.

25. Hans CP. A significant diffuse component predicts tor inferior survival in grade 3 follicular lymphoma, but cytologie subtypes do not predict survival. Blood. 2003;101:2363-2367.

Eric D. Hsi, MD; lmran Mirza, MD; Gerard Lozanski, MD; John Hill, MD; Brad Pohlman, MD; Matthew T. Karafa, PhD; Robert Coupland, MD

Accepted for publication April 15, 2004.

From the Departments of Clinical Pathology (Drs Hsi and Lozanski), Hematology and Medical Oncology (Drs Hill and Pohlman), and Biostatistics and Epidemiology (Dr Karat'a), Cleveland Clinic Foundation, Cleveland, Ohio; and the Department of Pathology, Cross Cancer Center, Edmonton, Alberta, Canada (Drs Mirza and Coupland).

Presented in abstract form at the annual meeting of the United States and Canadian International Academy of Pathology, Washington, DC, March 24, 2003.

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

Reprints: Eric D. Hsi, MD, Department of Clinical Pathology, Cleveland Clinic Foundation, 9500 Euclid Ave, Cleveland, OH 44195 (e-mail: hsie@ccf.org).

Copyright College of American Pathologists Aug 2004
Provided by ProQuest Information and Learning Company. All rights Reserved