Introduction
Basic Concepts
Follicular Lymphomas
Small Lymphocytic Lymphomas
Mantle-Cell Lymphoma
Therapy for Indolent Lymphomas
Histologic Progression and Clinical Transformation
Future Directions
References

Introduction

The indolent non-Hodgkin's lymphomas constitute a heterogeneous group of lymphoproliferative disorders usually associated with relatively prolonged survival. They are categorized based on pathologic and cytologic features, and, with few exceptions [1], they are almost exclusively of B-cell origin. The indolent lymphomas include follicular small cleaved, follicular mixed small- and large-cell, small lymphocytic, immunosecretory (Waldenström's), marginal-zone, and some cases of mantle-cell lymphoma (MCL). Marginal lymphoma includes mucosa-associated lymphoid tissue (MALT) and splenic (monocytoid) B-cell lymphomas (Table 1).

In the Working Formulation (WF) [2], most indolent lymphomas are classified as low-grade lymphomas, which include diffuse small lymphocytic, follicular small cleaved lymphomas (FSCL), and follicular mixed lymphomas (FML). Some of the new clinicopathologic entities, such as mantle-cell and marginal-zone lymphomas, do not easily fit within the confines of the subtypes of the WF, and this can be a source of confusion. Another source of confusion comes from the variable histology sometimes found in the same patient. This can happen in the same pathologic specimen, whereby two histologic subtypes are seen—a phenomenon referred to as composite lymphoma. More commonly, two distinct histologies are seen in specimens from two different sites—a discordant lymphoma.

Synchronous discordant lymphomas can be seen in 20% to 30% of newly diagnosed patients when more than one lymph-node biopsy is obtained simultaneously [3]. Another fairly common occurrence is the synchronous discordant pattern between the bone marrow, usually showing small cleaved cells, and a lymph node with diffuse large cells. By far, however, the most commonly recognized form of discordant lymphoma is asynchronous and is clinically referred to as transformation. This typically refers to an intermediate or high-grade lymphoma arising in a patient with prior low-grade histology.

Basic Concepts

Lymphocytes arise in the bone marrow from a pluripotent hematopoietic stem cell. For simplicity, their subsequent differentiation can be divided into two distinct stages: early, antigen-independent and late, antigen-driven. While early phases of T-cell differentiation occur in a specialized organ—the thymus—early B-cells undergo differentiation in the bone marrow.

The earliest sign of B-lineage commitment is the rearrangement of the immunoglobulin (Ig) heavy-chain locus on chromosome 14q32. This starts by the approximation of one of more than 20 D segments with a JH segment, creating a DJ region. This initial rearrangement occurs on both chromosomes. Subsequently, the DJ segment on one of the alleles is approximated with one of potentially several hundred V segments. The resulting VDJ segment then joins Cµ, and the rearranged heavy-chain gene is now ready to generate the heavy-chain protein. The successful production of the µ chain in the cytoplasm is the hallmark of pre-B-cell. Only if the rearrangement on the first chromosome was unsuccessful, will the second allele rearrange beyond the DJ stage (allelic exclusion). A failed rearrangement on the first allele can have potential pathogenic significance, as is the case with the t(14,18) translocation.

Once the heavy-chain locus successfully rearranges on either one of the two alleles, kappa light-chain gene rearrangement follows on chromosome 2(p11). The light-chain gene loci lack the D segment so that one of the variable genes will be directly approximated to one of five Jkappa regions. If the kappa gene rearrangement is unsuccessful, the other allele will rearrange, and only if that is unsuccessful will the lambda genes rearrange on chromosome 22(q11). Once a functional light chain is produced, it will bind to the µ heavy chain, producing a complete Ig molecule that will be expressed on the cell surface. The expression of surface immunoglobulin (sIg) is the hallmark of a mature B-cell.

The process of heavy-chain gene rearrangement requires an enzyme that joins the approximated splices by randomly adding nucleotides independent of a DNA template (N regions). These random processes add to the diversity of the Ig specificity and are mediated by the enzyme terminal deoxyribonucleotide transferase (TdT).

As the pre-B-cells express sIgM, they gradually change from the large, rapidly dividing cells (the “small non cleaved” cells) to the small resting ones and concomitantly lose TdT and CD10 reactivity. Called naive or virgin mature B-cells, they leave the bone marrow and circulate briefly before homing to the perifollicular lymphoid tissue or the splenic marginal zone. A fraction of these cells will express CD5 antigen. In humans, up to 17% of circulating B-cells are CD5 positive [4]. While in murine models there is evidence to suggest that these cells may arise from a separate ontogeny [5], such evidence is lacking in humans, and CD5 expression may simply represent activation of naive B-cells upon antigen exposure.

Most mature B-cells entering a germinal center will undergo apoptosis or programmed cell death. For B-cells to survive and mature into postfollicle stages, they need a dual signal: The first comes through antigen engagement with surface immunoglobulin (sIg) receptor, and the second is mediated by T-cell help that follows antigen presentation. One molecular consequence of T-cell help is the interaction between CD40 on the antigen-presenting follicular B-cells and CD40 ligand, expressed on activated helper T-cells. In vitro data suggest that dual signalling via surface Ig receptor and CD40 rescues B-cells from apoptosis. For example, CD40 stimulation was shown to protect cells from Ig cross-linking-induced apoptosis in murine WEHI 231 cells [6]. Similarly, murine B-cells stimulated by CD40 alone were very sensitive to Fas-induced apoptosis, whereas those simultaneously stimulated by CD40 and anti-IgM were not [7]. Surviving cells undergo isotype switching and the progeny memory B-cells will circulate in blood.

One model of lymphomagenesis suggests that clonal B-cells carrying the t(14;18) translocation are subject to the same regulatory mechanisms as normal cells upon entry into the follicles, at least in the early phases of the disease. However, upon antigen exposure, these cells may behave differently, failing to differentiate further or undergo apoptosis. The outcome is the follicular lymphoma, a disease initiated by a specific chromosomal translocation, possibly promoted by an antigen-driven process [8], and demonstrating progression with additional genetic abnormalities.

Follicular Lymphomas

The follicular lymphomas are the most common human B-cell neoplasms in the Western hemisphere but are less common in the non-Western world; they constitute approximately 45% of all non-Hodgkin's lymphomas [2] and 80% of all indolent lymphomas. Follicular lymphomas are characterized by a relatively indolent course, with a median patient survival of approximately 8 to 10 years. They occur exclusively in adults and equally among males and females.

Approximately 80% to 90% of all patients present with advanced-stage disease (III or IV), with generalized adenopathy and a high incidence of bone marrow involvement. A characteristic of follicular lymphomas (as well as of diffuse small lymphocytic lymphomas) is the phenomenon of spontaneous regression, which occurred in up to 30% of patients in one series [9]. Such regressions, however, are usually partial and typically short-lived (1 to 2 years).

Cytolologic and Pathologic Features

The follicular lymphomas usually grow in a nodular pattern, probably secondary to the expression of surface adhesion molecules, allowing homotypic adhesion or adhesion with dendritic reticulum cells. Among the prime molecular candidates to be involved (among many others) are the integrins LFA-1 (CD11a/18) and its ligand ICAM-1 (CD54). The concomitant expression of LFA-1 and ICAM-1 on neoplastic cells is shown to correlate with nodular growth pattern, whereas the lack of one or both molecules is associated with a diffuse growth pattern. Similarly, the lack of ICAM-1 expression is associated with a leukemic phase [10].

The neoplastic nodules are generally of uniform size and can result in the total effacement of the normal nodal architecture. These nodules constitute homogeneous clumps of neoplastic cells, unlike normal lymphoid follicles, which display functional polarization with germinal centers and lymphoid cuff. The neoplastic cells resemble normal counterparts present in the normal germinal center; hence, the name “follicular center-cell lymphoma” in the Lukes-Collins classification. Scattered within the follicles is a dense meshwork of dendritic reticulum cells that are invariably present in follicular lymphomas. In the normal lymphoid follicle, the cells evolve through different stages with distinct cytologic features: small cleaved, large cleaved, and large noncleaved cells. The last cell type is more proliferative and, thus, is identified as a centroblast, as opposed to the small cleaved centrocyte under the Kiel classification.

The cytologic subtyping of follicular lymphomas can be difficult [11,12], but they generally form a continuum from small cleaved predominance to mixed to large-cell predominance. The subtypes are divided based on the proportion of large cells in the nodules. The follicular small cleaved lymphoma generally should have no more than five large noncleaved cells easily identified per high-power field (HPF), while the FML has at least five large noncleaved cells per HPF [12,13]. The cutoff number between FML and follicular large-cell (FLC) lymphoma is somewhat arbitrary and may vary among different pathologists. Follicular large-cell lymphoma, constituting less than 10% of all follicular lymphomas, is typically included under intermediate-grade lymphoma of the WF because of its more aggressive clinical course. It typically consists of follicular large noncleaved cells as opposed to the less common category of follicular large cleaved; the latter probably maintains its indolent behavior and can be considered a low-grade lymphoma.

Under the Kiel classification, all three categories are lumped under centroblastic/centrocytic without identifying subgroups, recognizing the fact that each subtype should, by definition, have both cell types, albeit in different proportions. More recently, the International Lymphoma Study Group proposed a Revised European American Lymphoma (REAL) classification. In the proposed schema, follicular lymphomas were categorized under “follicle center lymphomas” with cytologic grades referring to the proportion of large cells in the follicle (Table 1). Thus, grades I, II, and III form a continuum from follicular small cleaved-cell to large-cell predominance, without specific recommendations being made about cutoff criteria between grades.

A Southwest Oncology Group (SWOG) study recently suggested that Ki-67 expression may represent an objective reproducible method of delineating the subtypes of follicular lymphomas. Ki-67 is a nuclear protein detected throughout all phases of the cell cycle but not the G0 phase [14]. It is, thus, a reliable marker of proliferation. In the SWOG study, the follicular small cleaved-cell lymphomas had a 5% proliferative rate, compared with 29% for follicular mixed lymphomas [15].

Biology, Cytogenetics, and Immunophenotypic Characteristics

Follicular lymphomas are, by definition, derived from follicular center cells [16]. Immunologically, the neoplastic cells carry the characteristics of mature B-cells and express CD19, CD20, CD22, and sIg (mostly IgM/IgD, but possibly also IgG or IgA). Being of follicular center-cell origin, the cells characteristically express CD10 and almost never express CD5. All subtypes of follicular lymphomas demonstrate heavy- and light-chain gene rearrangements [17,18], whereas the TCR genes are almost never rearranged. A potential T-B-cell interaction may play a role in lymphomagenesis.

In one study, purified follicular lymphoma cells underwent vigorous in vitro proliferation when cultured with a CD4+ T-cell clone that recognized an alloantigen expressed by the lymphoma cells. As seen in normal T-cell/B-cell interactions, the lymphoma cell proliferation was MHC class II dependent. In the absence of T-cells, the lymphoma cells did not respond to lymphokines [19]. This suggests that the T-cells commonly seen infiltrating follicular lymphomas may contribute to neoplastic B-cell proliferation. The paradoxical observation that T-cell infiltration is more commonly associated with spontaneous regression may indicate an early phase of the disease when the B-cells are still T-cell dependent and not autonomous.

With conventional cytogenetic techniques, 65% to 75% of patients will be successfully karyotyped. This contrasts with the higher success rate usually obtained in more aggressive lymphomas [20,21]. Most of the successfully karyotyped specimens will show several chromosomal abnormalities with complex karyotypes, while up to 16% will be normal. The nature of karyotypic abnormality is somewhat histology dependent. Follicular small cleaved-cell lymphomas tend to be associated with hyperdiploidy, with a modal number of 47 or 48, whereas small lymphocytic lymphoma tends to be more commonly pseudodiploid [22].

A variety of primary and secondary chromosomal aberrations have been described, with t(14;18)(q32;q13) being by far the single most common. This translocation, involving the bcl-2 gene, is most commonly seen in the follicular small cleaved histology, but its frequency declines as the proportion of large cells increases. It is almost always associated with other structural or numerical chromosomal aberrations [23]. Regional variation in the incidence of bcl-2 translocation has been described. In Japan, follicular B-cell lymphomas have been associated with a lower incidence of bcl-2 rearrangement [24].

Interestingly, tonsils and lymph nodes with follicular hyperplasia and, more recently, peripheral blood have been shown to have bcl-2-Ig translocation in approximately half the cases in some series [25-28]. This suggests that the t(14;18) translocation may be a common event in normal lymphocyte physiology and that cells carrying the translocation are not necessarily committed to evolving into lymphoma.

Several cytogenetic abnormalities are associated with specific histologic subtypes. For example, while follicular small cleaved lymphoma is associated with t(14;18), follicular mixed lymphomas also show association with trisomy 8 and follicular large-cell with trisomy 7 and breaks in 17q21-q25 [22]. Trisomy 12, on the other hand, is frequently accompanied by t(14;18) in diffuse large-cell lymphomas (DLCL), while as a solitary primary chromosomal abnormality, it is characteristically seen in chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL) and not follicular lymphomas [21,29].

The loss of genetic material is a rather common secondary abnormality in low-grade lymphoma. A number of nonrandom deletions of chromosomal material have been described, including 1q, 3p, 6q, 11q, 14q, and 16q, areas that are speculated to carry as yet unidentified tumor-suppressor genes [23]. No major role for p53 mutations has been described in the early phases of the development of low-grade lymphomas [30,31]. However, p53 abnormalities do appear to play a role in some cases of histologic transformation.

Using molecular analysis, about 85% of all follicular lymphomas are shown to have the t(14;18)(q32;q21) translocation juxtaposing the bcl-2 gene from 18q21.3 with the Ig heavy-chain locus on 14q32.2 [32]. The reciprocal event also happens such that the D segment, instead of joining the J segment on 14q32, is transposed to chromosome 18. While approximately 70% of the breakpoints occur at the 3´ untranslated region of bcl-2 gene (the major breakpoint region, MBR), about 20% will occur in a region 20 kb 3´ to the gene (the minor cluster region, MCR) [33,34]. Oligonucleotide primers that span both breakpoint regions are being utilized in polymerase chain reaction (PCR) techniques in an attempt to monitor minimal residual disease following therapy.

Of interest, the t(14;18) translocation, like others in lymphomas, uses the same enzyme machinery that normally catalyzes Ig gene rearrangement, including recombinase, which joins the oligonucleotide segments, and TdT, which inserts N-regions between the joining segments. Therefore, it may be that while the neoplastic cells in follicular lymphoma have a mature B-cell phenotype, the translocation occurs at the pre-B-cell stage, probably in the bone marrow.

Bcl-2 is a mammalian homolog of the Caenorhabditis elegans ced-9 gene [35]. It is formed by 3 exons, the first of which is untranslated. Exons 2 and 3 are separated by an intron more than 200 kb long. In t(14;18), the bcl-2 gene is juxtaposed with an enhancer element in the Ig heavy-chain locus on 14q32, with the resultant constitutive overexpression of hybrid mRNA of two different sizes. The accumulating Bcl-2 protein, however, has normal size and function, with a molecular mass of 26 kDa and 239 amino acids. The carboxyl terminal end of the molecule is hydrophobic and serves as an integral membrane anchor [36,37]. The protein will associate with several subcellular membranes including the nuclear membrane, the endoplasmic reticulum, and mitochondrial membranes [33,36].

The Bcl-2 protein accumulation, however, is not specific for follicular lymphomas. Indeed, a wide variety of lymphomas of both B- and T-cell origin, as well as normal mantle zone lymphocytes, overexpress the protein, while follicular hyperplasia and normal germinal centers do not [38,39]. The overexpression of Bcl-2 plays a critical role in blocking apoptosis, or programmed cell death, independent of affecting proliferation [36,40]. Normally, cells dying by apoptosis demonstrate membrane blebbing, volume contraction, nuclear condensation, and activation of a Ca(2+)-dependent endonuclease, cleaving the DNA into nucleosomal length segments. This is reflected by the characteristic ladder pattern on DNA electrophoresis.

In vitro data suggest that deregulated bcl-2 alone is insufficient to confer tumorigenicity to Epstein-Barr virus (EBV)-induced lymphoblastoid B-cells [41]. However, introducing bcl-2 into a selected number of cell lines has an apoptosis-sparing effect after growth-factor withdrawal. This included cell lines dependent on interleukin (IL)-3, granulocyte-macrophage colony-stimulating factor (GM-CSF, sargramostim(Drug information on sargramostim) [Leukine]), IL-4, and IL-6 [40,42]. Instead of undergoing apoptotic changes, the growth-factor-deprived cells simply enter G0 and do not die. Cells could still be rescued with growth factor up to 30 days after factor deprivation.

Similarly, transfection of bcl-2 protects a human pre-B-cell leukemia cell line by inhibiting apoptosis induced by several chemotherapeutic agents, including cytarabine, cisplatin (Platinol), etoposide (VePesid), and dexamethasone(Drug information on dexamethasone) [43]. Although these agents inhibit proliferation in the bcl-2-transfected cell line, unlike the wild cell line, the cells demonstrate growth arrest but do not undergo apoptosis [44]. On the other hand, antisense-mediated reduction in bcl-2 expression results in accelerated apoptosis in the setting of growth factor withdrawal [45].

The above data indicate that whatever the promoting signal for apoptosis is (growth-factor deprivation, exposure to cytotoxic agents), Bcl-2 is a critical inhibitor of a final common pathway mediating the process [46]. This may explain why patients with the t(14;18) translocation are generally not cured with chemotherapy.

Transgenic mice bearing the bcl-2/Ig minigene will display “polyclonal” follicular proliferation that expands the IgM/IgD-expressing B-cell population. Upon activation, the cells will readily enter the cell cycle and display protracted proliferation [47,48]. Splenic cells from these mice demonstrate a remarkable survival advantage over B-cells from other mice; similar to the transfected cell lines, these cells mostly reside in G0 but still proliferate in response to anti-IgM or lipopolysaccharide stimulation. Over time, these transgenic mice will progress from “polyclonal follicular hyperplasia” into a diffuse large-cell immunoblastic lymphoma [49]. As predicted, the transformation will invariably be associated with secondary genetic abnormalities. In this model, c-myc approximation with the immunoglobulin heavy-chain locus was observed as a secondary event in 50% of the mice.

Normal B-cells maintain high expression of Bcl-2 all through the pre-B phase and up to IgM/IgD expressing B-cell homing into the mantle zone. Subsequently, cells destined to die (eg, due to lack of T-cell help or failure to encounter their specific antigen within the follicle) maintain small cleaved morphology (centrocytes), downregulate Bcl-2 and undergo apoptosis within the follicle center. Even normal cells destined to survive will still partially downregulate Bcl-2 (compared with mantle zone cell levels) as differentiation proceeds.

Cells carrying t(14;18) entering the follicles with constitutive overexpression of Bcl-2 may not necessarily lead to follicular lymphoma. In fact, such cells are now commonly detected in the peripheral blood of normal individuals by PCR techniques, with higher titers correlating with advancing age [28]. Such aberrant expression of Bcl-2 by B-cells may lead to their prolonged survival within the follicle compared with normal counterparts, as suggested above. Antigen exposure [8], T-cell help [19], or further genetic abnormalities may add steps that permit lymphomagenesis. Subsequently, surviving cells will accumulate but, unlike their normal counterparts, will fail to differentiate to postfollicle stages like the immunoblast, plasma cell, and the memory B-cell [42]. This explains the original Rappaport terminology of “poorly differentiated lymphocytic lymphomas.”

The accumulating cells will maintain some responsiveness to surrounding stimuli and will be liable for secondary genetic abnormalities that will mount an accelerated pace to the neoplastic process. Such genetic abnormalities may include the loss of tumor suppressor genes (like p53) or the activation of proliferation oncogenes. As detailed elsewhere, the clinical counterparts of genetic progression include refractory relapses, accelerated growth, and finally, frank transformation [50]. In fact, it is now clear that the transformed cells arise directly from a subclone of the follicular lymphoma, sharing with them the same idiotypic specificity and immunoglobulin heavy- and light-chain genes [51].

The exact molecular mechanism of blocking apoptosis by Bcl-2 is not well understood and is beyond the scope of this review. Several areas of investigation have added to our knowledge in this regard, including the potentially important role of Bcl-2 in intracellular calcium partitioning, the newly described Bcl-2-related proteins like Bcl-x and Mcl-1, and the evolving crucial role of Bcl-2/Bax heterodimers in inhibiting the apoptotic effect of Bax homodimers [52]. In fact, the ratio between the two proteins may be more important than the absolute level of either.

Clinical Features, Diagnostic Workup, and Staging

Initial patient evaluation for follicular lymphoma involves history taking and physical examination. Patients should be questioned closely for the presence or absence of B symptoms. These symptoms occur in no more than 15% of patients with indolent lymphomas and include fever (38ºC or above), drenching night sweats, and significant weight loss of more than 10% of baseline weight within 6 months. Patients may develop symptoms related to lymph-node enlargement, especially with bulky masses in the neck or the retroperitoneum.

Splenomegaly is more common in CLL/SLL than in follicular lymphomas. It may be the only sign of disease, such as in patients with primary splenic lymphoma. Splenic enlargement may lead to left upper quadrant discomfort and early satiety. Mediastinal lymph-node involvement may occur, but the occurrence of direct pressure symptoms (eg, superior vena cava syndrome) is extremely unlikely. Patients may have symptoms of anemia and, less commonly, thrombocytopenia. Symptoms of gastrointestinal tract involvement are nonspecific and include abdominal pain or discomfort, change in bowel habits, and gastrointestinal bleeding. The physical examination should include all lymph-node sites, including epitrochlear and postauricular lymph nodes. The abdominal examination may show a mass, splenomegaly, and, less commonly, hepatomegaly.

Excisional lymph-node biopsy is crucial to establishing the diagnosis. Fine-needle aspiration (FNA) is inadequate since it does not preserve the nodal architecture. Bilateral bone marrow biopsies are needed in the staging workup due to the patchy nature of involvement. The bone marrow characteristically shows paratrabecular infiltration with small cleaved cells.

Laboratory data may show anemia or thrombocytopenia. The cytopenias may result from direct marrow involvement, hypersplenism, or may be autoimmune in nature. The latter is more commonly seen in CLL/SLL than in follicular lymphomas, while hypersplenism may be more commonly seen in primary splenic (marginal) lymphoma and hairy-cell leukemia. Patients with anemia, however, should have a direct Coomb's test.

Platelet-associated antibodies may be positive in immune thrombocytopenia. The peripheral blood smear may be helpful if immune cytopenias are suspected. This smear may also show circulating neoplastic cells, especially in CLL where mature-looking lymphocytes are by definition increased. The serum lactate dehydrogenase (LDH) and beta2-microglobulin levels may be elevated and are of prognostic significance.

Imaging studies should include a chest x-ray and a chest computed tomography (CT) scan if the x-ray result is suspicious. Abdominal and pelvic CT scans are essential, with their high sensitivity in detecting mesenteric lymphadenopathy. Lymphangiography is falling out of favor but remains a sensitive test to detect pelvic and para-aortic lymph node involvement, especially when lymph nodes are normal sized but have architecture abnormalities. It also provides an easy and accurate means to follow response to therapy with a plain abdominal film. A gallium scan is usually not indicated unless transformation is suspected.

The Ann Arbor Staging system is shown in Table 2. It was originally designed for the contiguously spreading Hodgkin's disease, as opposed to non-Hodgkin's lymphoma (NHL), which often spreads discontiguously. Thus, this staging system has drawbacks when applied to NHL—namely, the fact that most patients with indolent NHL have advanced disease (stage-III or -IV) at presentation and that even those with limited disease will have neoplastic cells detected in marrow and peripheral blood by sensitive molecular techniques.

Prognostic Factors

The importance of defining reliable prognostic factors in low-grade lymphomas becomes especially obvious in light of the great impact that patient selection can have on interpreting and comparing data among different trials. While histology is a major predictor in distinguishing clinical course between the low-grade and intermediate-grade lymphomas, most investigators found no clear difference in long-term survival among the different subtypes of low-grade histology (FSCL, FML, and SLL) [53,54].

Some data, however, suggest that patients with follicular mixed lymphomas have a more prolonged initial remission than those with follicular small cleaved lymphoma with potential curability [55]. Such data were contradicted by a different trial, making it difficult to reach any firm conclusions [56]. These differences may be due to the inconsistency in the criteria used by different pathologists in classifying follicular lymphomas.

 A higher degree of nodularity has been associated in some reports with improved outcome [57], but again, the issue remains controversial [53,57-59]. Other pathologic variables that may be associated with a favorable prognosis include the presence of interfollicular fibrosis [53] and the extent of helper T-cell infiltration [60,61]. In fact, the latter criterion was also associated with a higher rate of spontaneous regression [61].

Laboratory criteria correlating with a poor prognosis include elevation of serum LDH [62] and beta2-microglobulin levels [63] as well as increased expression of the nuclear proliferation antigen Ki-67, and the increased percentage of cells in S phase as determined by flow cytometry [64]. Some studies suggest that the presence of normal metaphases or the absence of abnormal ones correlates with a prolonged survival [65,66], although this could not be confirmed by subsequent studies [67].

The presence of structural breaks in either the short or long arm of chromosome 17 was shown by several groups to be a predictor of poor outcome, and seems to be an independent prognostic factor by multivariate analysis [66,68]. Patients with follicular small cleaved lymphoma with t(14;18) as a solitary abnormality typically have an indolent course, whereas those carrying additional karyotypic aberrations almost always have a more adverse outcome [69]. The ability to detect t(14;18) translocation by itself, on the other hand, has no impact on survival in patients with follicular lymphoma [70].

Recently, there has been a growing interest in the detection of clonal cells in the peripheral blood in limited-stage disease as a potential prognostic factor. Early studies focused on detecting clonal excess (CE) by surface Ig staining with fluorescent monoclonal antibody (MoAb) against human kappa or lambda chains [71,72]. Several studies reported a higher incidence of CE in patients with low-grade histology as opposed to histologies of more aggressive disease [73-75].

In the study by Johnson et al [73], 27 patients with early-stage, low-grade lymphoma in remission following involved-field (IF) radiation therapy were examined for CE. After a short median follow-up of 34 months, two of five patients with CE had relapsed, as opposed to none of 22 without CE (P < .0001). More sensitive techniques have been developed to detect clonal cells in peripheral blood, including Southern blot analysis [76] and, more recently, PCR techniques. PCR is a far more sensitive technique than the other two and is currently being investigated extensively as a potential predictor of relapse after high-dose and conventional chemotherapy.

Clinically, several variables have been shown to correlate with survival in follicular lymphoma. These include tumor burden, host factors, and response to therapy. Tumor burden can be defined in a variety of ways, utilizing variables including stage of disease, size of nodal disease, bone marrow involvement, beta2-microglobulin level, and number of extranodal sites. Limited-stage disease is clearly associated with favorable outcome [54,77,78], while the presence of two or more sites of extranodal involvement correlates with poor prognosis [79]. Adverse host factors include advanced age, B symptoms [80], low hemoglobin level, male gender, and poor performance status [77-79,81].

Attempts to design predictive models have been made [83,79], but unlike the case in intermediate-grade lymphoma, no single system has gained wide acceptance. Recently, Lopez-Guillermo et al [82] examined the prognostic value of the International Index in predicting outcome in low-grade lymphoma. The International Index was devised for aggressive lymphomas and consists of five variables: age, performance status, Ann Arbor stage, extranodal involvement, and LDH level [83]. Lopez-Guillermo categorized 125 patients with low-grade lymphoma into three groups of low, intermediate, and high risk. While the International Index had no predictive value for response to therapy, there was a significant correlation with survival. The overall 10-year survival was about 75% for the low-risk group, 0% for the high-risk group, and approximately 50% for the intermediate-risk group. One important limitation of this system is that only 11% of the patients fell into the high-risk group, and most of these patients will probably have poor performance status and be poor candidates for aggressive therapy.

As shown by several series, response to initial therapy remains a powerful predictor of survival [54], with more than 80% of complete responders living at 7 years compared with a median survival of 2 years among those failing to achieve a complete response (CR) in one study [84].

At the time of relapse, favorable predictors for survival include having achieved a CR with initial treatment, a durable response to initial therapy of more than 1 year, and an age less than 60 years [85].

Small Lymphocytic Lymphomas

This category of the WF includes the classic nodal SLL, which is immunophenotypically and morphologically identical to chronic lymphocytic leukemia. It also has similar clinical features but lacks the characteristic absolute lymphocytosis. In addition, it also includes more recently characterized entities not specifically identified in the WF. These are the low-grade B-cell lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma), monocytoid B-cell lymphoma, and extranodal SLL. The first two entities share similar pathologic and immunophenotypic characteristics and are thus encompassed together under the term marginal B-cell lymphomas [13].

Indolent Lymphomas Arising From Mucosa-Associated Lymphoid Tissue

This group of indolent lymphomas has distinct clinical and pathologic characteristics. It probably includes many of the “pseudolymphomas” described in the old literature [86]. It is characterized by its localized nature, prolonged history, and good response to local therapy. It has been described at several extranodal sites including the gastrointestinal tract, lungs [87], salivary glands, thyroid, thymus, breast, orbit, and conjunctiva [88,89]. Its potential for transformation into a high-grade histology is suggested by the significant number of high-grade gastric lymphomas that demonstrate a low-grade component in the background [90]. Dissemination is said to be rare and to occur late in the course.

Although most MALT lymphomas may be best categorized under SLL of the WF [91], a recent study found the majority of MALT lymphoma patients to have previously been classified as having diffuse or follicular small cleaved lymphoma (30% each), while only a minority of the patients (5%) were described as having SLL [91]. Sometimes, within the same specimen, different cell types are grouped together and not intermingled, comprising a multiphasic histology [88]. Plasmacytoid differentiation is seen in one-third of the patients.

With the exception of the Peyer's patches of the small intestines, MALT does not normally exist in any of the tissues in which MALT lymphomas arise. However, in response to an immune stimulus, MALT can arise as an ectopic tissue. The stimulus can be infectious in origin, like Helicobacter pylori in the gastric mucosa, or autoimmune in nature, as in Sjögren's syndrome [92] or Hashimoto's thyroiditis [93]. Whatever the underlying mechanism, the common pathophysiologic features of MALT are antigen recognition, T-cell help, and B-cell proliferation. If the proliferating B-cells show clonality, a MALT lymphoma is said to arise.

Both MALT and MALT lymphomas share constant pathologic features. They always have reactive lymphoid follicles. When lymphoma develops, tumor cells reside mostly in the area surrounding the mantle zone of the follicles; hence, the name marginal zone lymphoma. The marginal zone expands with diffuse cellular infiltrate, and the cells are of small lymphocyte or centrocyte morphology, explaining their WF classification. On occasion, neoplastic cells will “colonize” and even disrupt the follicles [94], resulting in a pseudofollicular pattern to the tumor as a whole.

A second pathologic characteristic shared by both MALT and MALT lymphomas is lymphoepithelial invasion, in which the cells of the marginal zone are shown to invade into the overlying epithelium. Immunophenotypically, the neoplastic cells do not express CD5, CD10, or CD23 but mostly express surface IgM. Neither bcl-1 nor bcl-2 are rearranged, even when a pseudofollicular pattern is seen [95]. In the stomach, the pathologic recognition of MALT lymphoma may not be easy since the process constitutes a continuum along a spectrum from acute gastritis to MALT to MALT lymphoma. In one study [96], the presence of dense lymphoid infiltrates, prominent lymphoepithelial lesions, moderate cytologic atypia, or Dutcher bodies were found to be highly suggestive of MALT lymphoma. Patients with gastric lymphoma arising in MALT may present with a previous history of gastritis or peptic ulcer disease. Other MALT-associated lymphomas will cause symptoms related to the anatomic site of the disease.

Recently there has been a great deal of interest in gastric lymphomas because of their association with Helicobacter pylori. Early studies correlated H pylori with gastric adenocarcinoma and gastric large-cell lymphoma [97,98]. A geographic correlation was also noted between H pylori infection and the incidence of gastric non-Hodgkin's lymphoma by several investigators [97,99,100]. In one study [97], a comparison was made between the incidence of H pylori infection in gastric vs nongastric large-cell lymphomas; the odds ratio for the association of gastric lymphoma with H pylori was 6.3 as opposed to 1.2 for nongastric lymphoma, suggesting that H pylori may play a role in gastric lymphomagenesis.

Recently, more emphasis was put on studying the association between H pylori and MALT lymphoma, which constitutes a minority of primary gastric lymphomas. Wotherspoon et al [101] described MALT in 125 of 450 patients with H pylori-associated gastritis. Eight of the 125 had more pronounced lymphoepithelial invasion suggestive of MALT lymphoma. In a separate cohort of 110 patients with known gastric MALT, 101 of 110 patients had evidence of H pylori infection. While this suggests an association without proving a causal relationship, the observation that therapy for H pylori with antibiotics results in a dramatic reduction in gastritis-associated MALT was highly instructive [101].

More recently, the same European group provided more convincing evidence of the causal relationship between H pylori and MALT lymphoma. In this study, Wotherspoon et al noted the complete disappearance of MALT lymphoma in five of six patients treated for H pylori infection [102]. In a recent update of their data, two more patients responded with regression of tumor, and remission was maintained up to 22 months after antibiotic therapy [103]. Stole et al [104] had less impressive results treating 32 patients with MALT lymphoma. Nineteen of the 32 had regression, 6 had full eradication, and 4 had persistence of the lymphoma.

Such clinical evidence was recently supported by laboratory investigation. Hussell et al [105] recently showed that cells from low-grade MALT lymphoma were stimulated by heat-killed H pylori in a strain-specific T-cell-dependent fasion in vitro. In contrast, extragastric MALT or gastric DLCL showed no such response. While the T-cells in this model demonstrate H pylori specificity, neoplastic B-cells did not, and in two cases there was a tissue autoantigen reactivity. This suggests that the B-cells may be autoreactive bystanders stimulated by nearby activated H pylori specific T-cells [105,106].

Other gastrointestinal MALT lymphomas include a “Mediterranean” type referred to as immunoproliferative small intestinal disease (IPSID), which may share with gastric MALT a similar lymphomagenesis mechanism, being associated with gastrointestinal bacterial infections [107], and demonstrating clinical responses to antibiotic therapy with tetracycline(Drug information on tetracycline) in the early phases of the disease.

Monocytoid B-cell lymphoma is the lymph-node counterpart of MALT lymphomas. The tumor consists of clear cells with reniform or oval nuclei. Instead of the lymphoepithelial lesions typical of MALT, these lymphomas have lymph-node growth pattern, with the neoplastic cells accumulating in confluent sinuses.

Early-stage marginal-zone lymphomas have an excellent prognosis with radiation therapy to the involved field. Advanced-stage marginal-zone lymphoma (MZL), on the other hand, has a comparable outcome to the other low-grade lymphomas. Among 43 patients with advanced-stage marginal zone lymphoma treated with CHOP and reported recently by Fisher et al [91], the 10-year overall and failure-free survival was not significantly different from that of other categories of the WF studied, including SLL, FSC, FML, FLC, and diffuse small-cell (DSC) categories. When 19 patients with advanced-stage MALT lymphoma were compared with 21 patients with advanced monocytoid lymphoma, there were significantly worse overall (21% vs 53%, P = .007) and failure-free (21% vs 46%, P = .009) survivals among the MALT lymphoma patients at 10 years. In fact, advanced-stage MALT lymphoma fared worse than other low-grade categories of the WF studied. Therefore, in sharp contrast with early-stage MALT, the limited literature that exists suggests that advanced stages of MALT lymphoma may be worse than many other advanced-stage low-grade lymphomas.

Mantle-Cell Lymphoma

Mantle-cell lymphoma is a relatively newly recognized clinicopathologic entity first recognized in the 1970s under the previously described categories “centrocytic lymphoma” of the Kiel classification and “intermediate differentiated lymphoma,” or IDL, of the modified Rappaport classification. The neoplastic cells arise from the mantle zone of secondary follicles [108,109] and usually give rise to a diffuse pattern of lymph-node involvement. Less commonly, a vaguely nodular pattern may be present (the mantle-zone lymphoma). As previously described, the WF does not recognize MCL as a specific entity, but it has most commonly been classified as diffuse small cleaved-cell lymphoma.

In a recent study, about 60% of MCLs were previously categorized as diffuse small cleaved, 25% as FSC, and the rest as SLL [91]. Presently, this lymphoma can be distinguished from lymphomas of follicular center-cell origin both morphologically and immunophenotypically. Mantle cells appear similar to the small cleaved lymphocytes with their scant cytoplasm and irregular nuclear contours, explaining their categorization under centrocytic lymphomas in the Kiel classification. But unlike the follicle-center lymphomas, which always contain large cells (centroblasts), mantle-cell lymphomas are more homogeneous morphologically. They also lack the dense organized meshwork of dendritic reticulum cells that characterizes follicle center-cell lymphomas [110].

Immunophenotypically, mantle-cell lymphomas are almost always CD5-positive, like CLL and normal follicular mantle-zone cells, reflecting a prefollicle maturation with possible activation and distinguishing them from true follicle center-cell lymphomas that are virtually always CD5-negative. Otherwise, the cells typically express mature B-cell phenotype with CD19, CD20, CD22, and sIg, usually IgM, with a notable unexplained preference for lambda light chain. The cells also express CD43 but not CD10 or CD23. Such specific immunophenotypic features help distinguish this lymphoma from small lymphocytic lymphoma (CD23-positive) and follicle center-cell lymphomas (CD10-positive and CD43-negative). The absence of large or transformed cells, together with the mantle-zone phenotypic features of the neoplastic cells, suggests that this tumor is a distinct entity unrelated to the spectrum of follicle center-cell lymphomas [111].

The characteristic cytogenetic lesion of mantle-cell lymphoma is the t(11;14)(q13;q32) translocation, seen in about 73% of cases [112-114]. As a consequence of the translocation, a newly identified gene, PRAD-1, on 11q13 is juxtaposed to the Ig heavy-chain joining region on 14q32. The major breakpoint region on chromosome 11 is located approximately 110 to 120 kb centromeric to PRAD-1 and is designated bcl-1. Unlike bcl-2, bcl-1 simply represents the breakpoint region and does not represent an oncogene. Molecular probes to bcl-1 can detect approximately 50% of all mantle-cell lymphomas, whereas probes to minor breakpoints will be needed to detect the rest [115]. As expected, the translocation will result in overexpression of PRAD-1 [116], a member of the cyclin family of proteins—hence, the other name of the protein, cyclin D1 [117]. In fact, nuclear cyclin D1 can be detected in virtually all mantle-cell lymphomas using polyclonal antibody on paraffin(Drug information on paraffin)-embedded sections [118,119].

The role of cyclin D1 in oncogenesis is far from clear. The protein normally forms complexes with p21(waf1) and cyclin-dependent kinases and plays a role in cell-cycle progression. PRAD-1 was originally described in benign parathyroid adenomas as the oncogene rearranged to the parathyroid hormone locus [119a]. In addition to mantle-cell lymphoma, the gene is overexpressed in a considerable number of patients with breast cancer [119b] and head and neck squamous-cell carcinomas [119c]. Transgenic mice overexpressing the gene under the control of an immunoglobulin enhancer showed somewhat fewer mature T and B lymphocytes, albeit with normal cell-cycle activity and spontaneous lymphomas observed [119d].

Mantle-cell lymphoma constitutes up to 10% of all NHL [91]. Patients are usually males older than 55 years who present with advanced-stage disease and generalized lymphadenopathy, bone marrow involvement, and a leukemic phase in up to 38% of cases. The liver, spleen, and Waldeyer's ring are frequently involved. Gastrointestinal tract involvement is seen in up to 20% of cases, with infiltration of the submucosa giving rise to multiple lymphomatous polyposis. This is distinct from lymphoepithelial involvement commonly seen in MALT lymphomas. Unlike other CD5-positive lymphoproliferative disorders, no autoimmune phenomena have been described in these patients.

The clinical course is heterogeneous, with some patients having very aggressive disease while other cases behave more like indolent lymphomas. In a small series of 23 patients at the National Cancer Institute (NCI), all patients presented with stage-III or -IV disease [111], and liver involvement was an especially poor prognostic indicator. The pattern of growth may be prognostically important: A nodular pattern with residual germinal centers (mantle-zone lymphoma) appears to be indolent, while effacement of germinal centers or the entire node suggests a more aggressive behavior [120,121]. Likewise, blastic cytologic features or high Ki-67 expression may be adverse features, but all these observations need to be validated in large groups of uniformly treated patients. The usual prognostic factors (elevated LDH or beta2-microglobulin levels, advanced stage, advanced age, and poor performance status) seem to be applicable to mantle-cell lymphomas [122]. Although the survival curves show patterns similar to those of low-grade lymphomas in that there is no plateau, the median survival is significantly shorter, ranging from 31 to 61 months [115]. The 10-year disease-free and overall survivals were 6% and 8% in one series, three to four times worse than corresponding survivals in indolent lymphomas reviewed [91].

Therapy for Indolent Lymphomas

The vast majority of patients with advanced-stage low-grade NHL will respond to initial chemotherapy, with CR observed in one-half to two-thirds of patients. However, virtually all patients will ultimately relapse and most will die of their disease. While obvious responses and clinical effects may be seen using various therapeutic modalities, the hallmarks of advanced-stage disease (namely, its continuous recurring nature and incurability) remain unchanged [111].

Assessment of new therapies for low-grade lymphomas can be difficult. First, the disease is heterogeneous, and differences in patient selection criteria can make comparison of data from different clinical trials difficult. Second, the disease has a long natural history, requiring long-term follow-up of patients before any final interpretation of data can be made. Surrogate biological markers that could reflect early therapeutic success are needed. PCR for bcl-2 may be such a marker, although further research is needed on this issue. Third, the frequency and stringency of restaging can affect the comparability of relapse-free survival in different series. Finally, the sequential utilization of multiple therapies in an individual patient confounds the analysis of each particular treatment intervention.

Limited-Stage Follicular Lymphoma

At time of presentation, about 15% to 20% of patients have limited-stage disease, and about half of these patients may be curable. Several series have reported that approximately half of stage I-II patients achieve long-term disease-free survival following treatment with IF radiation [123,124]. Table 3 summarizes survivorship data from selected trials in patients with early-stage disease treated with radiation therapy alone or in combination with chemotherapy. In a recent update of the Stanford data [123], the long-term failure-free survival (FFS) after definitive radiotherapy is 40% [125]. A subset of patients who received total-lymphoid irradiation (TLI) appeared more likely to be relapse-free than those treated with involved-field or extended-field radiation. However, when laparotomy-staged patients were excluded from analysis, such a difference was no longer seen, suggesting that a disproportionate number of patients receiving TLI underwent laparotomy.

Recent insights perhaps make a systemic approach to early-stage disease appealing, given that bcl-2 gene rearrangement can be detected in marrow and peripheral blood with high frequency in both limited and advanced-stage disease [126]. The M.D. Anderson Cancer Center reported 76 patients with stages I-II disease who were treated with different modalities. Fifty patients received IF radiation alone, while 19 received combined modality with IF radiation and chemotherapy and the rest received chemotherapy alone. The overall 5-year failure-free survival was 47%. Those receiving combined modality therapy had a 5-year failure-free survival of 64% as opposed to 37% for those treated with IF radiation therapy alone, although no overall survival difference was seen [127]. A second retrospective analysis of 51 patients with stages I-II low-grade lymphoma treated at St. Bartholomew's hospital over a 13-year period also showed improved disease-free survival among those receiving combined-modality therapy compared with radiation alone, but, again, no difference in overall survival was noted [128].

Based on the encouraging results with combined-modality therapy in the retrospective studies, investigators at M.D. Anderson prospectively treated patients with stage I-II disease with 10 cycles of COP-Bleo (cyclophosphamide [Cytoxan, Neosar], vincristine [Oncovin], prednisone(Drug information on prednisone), bleomycin [Blenoxane]) or CHOP-Bleo (COP-Bleo plus doxorubicin [Adriamycin, Rubex]) with radiation to involved sites “sandwiched” after the third cycle. The 5-year disease-free survival was 77%, an apparent improvement over results with radiotherapy alone [129]. A prospective randomized trial also examined the efficacy of adding single-agent chlorambucil(Drug information on chlorambucil) (Leukeran) for 6 months after radiation therapy with no improvement in disease-free or overall survival after up to 18 years of follow-up [130].

Experience is limited in the treatment of limited-stage MALT lymphoma and depends on the primary site. In general, long-term survival is attained with IF radiation with or without combination chemotherapy. As discussed elsewhere, antibiotic combinations have shown efficacy in the eradication of gastric MALT lymphoma in a large number of patients.

In summary, patients with limited-stage disease appear potentially curable, with overall long-term disease-free survival of approximately 50%. The role of radiation therapy in these patients is well established, and IF radiation remains the standard treatment. Total lymphoid irradiation may offer longer disease-free survival, although this is controversial and the subject of current investigation. The combined-modality approach has been intensively investigated, as discussed above. While adjuvant therapy with single agents did not improve outcome when added to radiation therapy, the addition of adjuvant or neoadjuvant combination chemotherapy may have an impact on disease-free survival, but large prospective trials to address this question are lacking.

Advanced-Stage Follicular Lymphoma

In patients with stage-III and -IV follicular lymphoma, the overall response rate with different chemotherapy programs is as high as 80% to 90%. The CR rate, however, varies widely, between 23% and 83% in various studies [131-134]. This is mostly explained by differences in patient selection criteria, techniques, and diligence used to assess response and the definition of response.

Recently, an attempt was made to standardize definitions of response and progression in Hodgkin's disease [135], and the recommendations are generally applicable in non-Hodgkin's lymphoma, as well. Complete response is defined as no clinical, radiologic, or other evidence of disease. Partial response (PR) is a decrease by at least 50% of the sum of the products of the largest perpendicular diameters of all measurable lesions. Progression is considered an increase in size of one or more measurable lesion by 25% or more. A new definition—CRu (unconfirmed/uncertain)—was introduced, referring to patients who have normal health but demonstrate some radiologic abnormality at the site of previous disease that is not consistent with the effect of previous therapy. This subset of patients has doubtlessly been classified in the past as CR in some trials and PR in others. The designation of CRu will hopefully improve the reliability and comparability of the data among different studies.

The impact of techniques used to assess response is best exemplified by the routine use of abdominal CT scan, which markedly increases the sensitivity of detecting mesenteric and upper abdominal lymph nodes. Thus, studies using such modalities to assess response may define CR more stringently and may have lower CR rates.

Because low-grade lymphomas are sensitive to radiation, radiation has been incorporated in primary therapy for some advanced stages of the disease. This has been best studied in patients with stage-III disease. In updated results of a treatment program with central lymphatic irradiation initially reported by Cox et al [136], the Milwaukee group demonstrated FFS rates of 40% at 15 years in stage-III patients [137]. At Stanford, the FFS rates at 5 and 10 years were 60% and 40%, respectively, in 66 patients with stage-III disease treated with TLI with or without chemotherapy. In a small subset of 16 patients prospectively randomized to receive TLI with or without CVP (cyclophosphamide, vincristine, prednisone), chemotherapy showed no significant advantage in survival or freedom from relapse [138]. Stage-III patients at M.D. Anderson treated with CHOP-Bleo plus radiation to involved sites had a 5-year disease-free survival of 52% [62]. All these data strongly support a primary role for radiation therapy in stage-III disease.

The role of radiation in patients with stage-IV disease is less well understood. In one trial, 118 patients with advanced stage (-III and -IV) disease were randomized to receive combined modality therapy (TLI plus CVP or IF plus CVP) vs chemotherapy alone (CVP) [139]. The 7-year relapse-free survival for patients in the combined modality groups (71% and 66% for TLI and IF, respectively) was significantly better (P < .01) than in patients receiving chemotherapy alone (33%). The overall survival was similarly improved, indicating a benefit for radiation therapy in advanced nodular lymphomas. This trial included both stage-III and -IV patients, however, and patients were not stratified according to stage when survival was examined. Other groups' experience with advanced-stage disease have failed to show any advantage for radiation therapy compared with or added to chemotherapy.

In a small randomized study by Hoppe et al, fractionated whole-body irradiation (with or without CVP) resulted in similar CR and overall survivals to those achieved with CVP or oral alkylating agents alone [131]. Whether radiation therapy will add to the effectiveness of more aggressive and dose-intensive regimens in stage-IV patients remains to be seen. The Vancouver group is examining the efficacy of such an aggressive regimen, BP-VACOP (bleomycin, cisplatin, etoposide, doxorubicin(Drug information on doxorubicin), cyclophosphamide(Drug information on cyclophosphamide), vincristine, prednisone) with TLI in patients with bulky stage-II, or stage-III and -IV disease [140]. Therefore, unlike stage-III disease, the role of radiation therapy in stage-IV follicular lymphoma is less well defined, and chemotherapy remains the mainstay of therapy in these patients.

Single-agent therapy with chlorambucil or cyclophosphamide (with or without prednisone) has long been considered a primary standard therapy in advanced low-grade lymphomas. Combination chemotherapy may result in more rapid responses than single-agent chlorambucil or cyclophosphamide [131], but there is no clear evidence that it improves overall survival over single agents. Since response rates have been shown to correlate with survival, there has been emphasis on the CR rates of various chemotherapy programs. CVP was among the earliest combinations described and has gained wide acceptance [141]. The role of adding doxorubicin to the chemotherapy regimen remains controversial. CR rates and survivals seen in CHOP-treated patients were superior to those of a historical group of COP-treated patients at M.D. Anderson Cancer Center [142]. However, a large randomized study conducted by SWOG, reported no difference in outcome between CHOP-Bleo and COP-Bleo in indolent lymphomas [143]. Retrospective analysis of survival among 415 patients treated with CHOP in three SWOG trials also showed no advantage over results with less aggressive programs [144]. Similarly, CHOP has been compared with chlorambucil/prednisone [145], and while more responses and shorter induction times were noted in the CHOP group, no survival difference was observed between the two groups.

In addition to CHOP and its variants (see Table 4), new programs incorporating different drugs have been developed. For example, procarbazine (Matulane)-containing regimens may be associated with more durable remissions in follicular mixed [55] and follicular small cleaved lymphoma [146]. Mitoxantrone(Drug information on mitoxantrone) (Novantrone) as a single agent had a high response rate of 95% in previously untreated [147] and up to 67% in relapsed patients [148]. Because of the synergism previously described between cytarabine(Drug information on cytarabine) and cisplatin(Drug information on cisplatin) in lymphoma [149], the combination of etoposide(Drug information on etoposide), methylprednisolone(Drug information on methylprednisolone), high-dose cytarabine, and cisplatin (ESHAP) was tested and shown to be effective in patients with relapsed low-grade lymphoma, with a CR rate of 35% and a PR rate of 40% [150].

In keeping with the Goldie-Coldman hypothesis, 138 newly diagnosed patients were treated at M.D. Anderson with a sequential three-combination chemotherapy program (CHOD-B [CHOP-Bleo with dexamethasone instead of prednisone]/ESHAP/NOPP (mitoxantrone, vincristine, procarbazine, prednisone) for a total of 12 cycles. The overall CR rate was 65% with a projected 4-year survival of 94% and an FFS of 67% [151]. Though promising, the median follow-up of 27 months is too short to allow firm conclusions about this intensive regimen, but the follow-up PCR data for bcl-2 gene rearrangement is provocative and encouraging. Several other aggressive combination chemotherapy and radiotherapy programs have been reported or are under investigation. To date, no single regimen has emerged as a standard, with 5-year failure-free survival ranging between 25% to 35% with most regimens (see Table 4).

Because low-grade lymphomas have a long indolent course and since no therapy has clearly impacted on the continuous recurring nature of the disease, several investigators looked into the effect of withholding therapy in relatively asymptomatic patients until their symptoms warrant treatment—the so called “watch and wait” policy. Several reports in the literature suggest that this approach had no significant adverse impact on survival [9,152,153], and some investigators recommend this as a standard approach in asymptomatic patients [152].

To examine this concept further, the NCI conducted a study randomizing patients to receive intensive therapy with ProMACE-MOPP (prednisone, methotrexate, doxorubicin, cyclophosphamide, etoposide, mechlorethamine [Mustargen], vincristine, procarbazine) at the time of diagnosis or to defer therapy until the disease had progressed to a degree that could not be managed by radiation therapy alone in the watch-and-wait group [154]. The CR rate among those randomized to immediate therapy was higher than among those initially randomized to no therapy and later requiring treatment (75% vs 43%). Those who could not be randomized because they required intensive therapy at the time of diagnosis also had a significantly worse CR rate of 57%. The lower CR rate in the watch-and-wait patients suggests that by the time the patients are symptomatic with bulky disease, a sufficient number of secondary genetic abnormalities may have accumulated to entail resistance to treatment. But with a median follow-up of 4 years, there was no survival difference among any of the 3 treatment groups. So far, therefore, it cannot be firmly concluded that early intensive chemotherapy will improve outcome. However, one of the preliminary conclusions of this trial has been a quality of life argument that, ironically, favors early intensive therapy: Those treated early enjoyed more time in remission off therapy than those managed with the palliative approach. An update of this trial is awaited.

Patients with follicular mixed lymphoma deserve special attention since several investigators have documented long-term disease-free survival in such patients treated with various regimens including C-MOPP (cyclophosphamide, vincristine, procarbazine, prednisone) [122], ProMACE-MOPP [155], and CHOP-Bleo [156], with long-term disease-free survival of up to 75%. This may be an argument for using early intensive therapy in this subgroup of patients. Glick et al [157], on the other hand, failed to confirm such results using COPP.

A new class of drugs demonstrating remarkable activity in indolent lymphoproliferative disorders are the purine analogs. In particular, fludarabine (Fludara) and cladribine(Drug information on cladribine), or 2-chlorodeoxyadenosine (2-CdA, Leustatin) have been extensively studied and found to have significant activity. In addition to being antiproliferative, these agents can induce apoptosis [158]. This may explain their distinct effectiveness in indolent lymphoproliferative disorders as opposed to the more aggressive lymphomas. In addition, their resistance to the effect of adenine deaminase contributes to their efficacy compared with other nucleoside analogs.

Several phase-II trials of fludarabine in low-grade lymphomas have been published. In general, when fludarabine is administered at a dose of 18 to 25 mg/m²/d for 5 days repeated every 3 to 4 weeks, response is noted in one-half to two-thirds of previously treated patients with low-grade lymphoma. Those with intermediate-grade lymphomas, on the other hand, tend to show poor responses [159,160]. Redman et al treated 38 patients with relapsed or refractory low-grade lymphoma. The overall response rate was 55%; of the 21 patients with follicular small cleaved-cell lymphoma treated in this trial, four had a CR, and nine had a PR. Other pathologic subtypes had less remarkable responses.

Similarly, in an Eastern Cooperative Oncology (ECOG) trial, 27 previously treated patients with low-grade lymphoma received fludarabine, with an overall response rate of 52% and CR noted in 5 of 27 patients [160]. In 16 previously untreated patients, Pigaditou et al demonstrated better overall and complete response rates of 75% and 60%, respectively [161]. The major dose-limiting toxicity of fludarabine in these trials was myelosuppression, mainly neutropenia, occurring in one-third to one-half of patients, with better tolerance noted among those without previous exposure to chemotherapy [161].

Early trials with high doses of fludarabine (up to 125 mg/m²/d for 7 days) for acute myelogenous leukemia (AML) were associated with significant neurotoxicity, but at the lower doses currently used for low-grade lymphomas and chronic lymphocytic leukemia (CLL), neurotoxicity is rare. However, Johnson et al recently described an unusual neurologic syndrome in five patients receiving fludarabine at standard dose for low-grade lymphoma. This included severe headache, hemiparesis, sensory abnormalities, and, in one patient, documented multiple brain infarcts. Therefore, patients receiving the drug should be carefully evaluated for any neurologic complications and the drug discontinued if necessary [162].

With such a degree of activity as a single agent, fludarabine has been combined with other active agents. In a phase I [163] and a subsequent phase II trial, McLaughlin et al treated patients with recurrent low-grade or follicular large-cell lymphoma with the combination of fludarabine, mitoxantrone, and dexamethasone; in the phase-II trial, the CR rate was 43% and PR rate was 51%, with durable responses [164]. The combination of fludarabine and chlorambucil was studied in phase-I trials [165] and is currently being studied in a randomized trial in untreated CLL, comparing chlorambucil alone, fludarabine alone, and the combination of the two drugs. Other fludarabine-containing regimens are currently being evaluated with various drugs including cyclophosphamide, cytarabine, interferon [166], and paclitaxel(Drug information on paclitaxel) (Taxol).

The combined immunosuppressive effects of steroids and fludarabine may be of concern since the combination resulted in a higher incidence of serious infections including listeriosis in one report of CLL patients [167]. Cytomegalovirus and Pneumocystis carinii pneumonia (PCP) have also been reported [168]. Therefore, in more recent trials that combine fludarabine with steroids, prophylactic trimethoprim(Drug information on trimethoprim)-sulfamethoxazole has been added, as was successfully done when the ProMACE-CytaBOM regimen was initially associated with a high incidence of PCP [169].

Cladribine (2-CdA) has been less extensively studied in low-grade lymphomas. The standard dose is 0.1 mg/kg/d by continuous infusion for 7 days. Kay et al [170] demonstrated a response rate of 43% in patients with relapsed disease. Hoffmann et al demonstrated similar results, but there was a high incidence of myelotoxicity and infections [171]. Hickish et al [172] reported a somewhat better overall response of 75% in eight previously treated patients. As with fludarabine, previously untreated patients generally demonstrate better response rates. Emanuele et al [173] reported an impressive overall response rate of 82% in previously untreated patients. Liliemark et al [174] treated 20 patients with newly diagnosed low-grade lymphoma at a dose of 5 mg/m² as a 2-hour infusion for 5 days with cycles repeated every 28 days; there was a 60% overall response rate with 20% CR. Grade III/IV neutropenia was noted in up to 50% of patients.

Two other trials of cladribine reported similar overall response rates (50% and 66%) and toxicity profiles [175,176]. Responses to the drug have generally been brief. Cladribine is also being investigated in several combination regimens that include alkylating agents [177] or mitoxantrone. Important limitations of therapy with purine analogs include cumulative neutropenia and thrombocytopenia and an inversion in the CD4:CD8 ratio [178], factors that limit the total number of cycles that can be delivered.

In summary, the purine analogs represent a highly promising group of compounds for therapy of indolent lymphomas. Their role in CLL is better established, but the data clearly show well-defined efficacy, especially for fludarabine, in the salvage of previously treated patients with indolent lymphomas. The role of purine analogs in various salvage combinations, as well as in primary therapy, is currently being extensively investigated. Their associated immune suppression and cumulative myelotoxicity may require dose or schedule modifications and prophylactic antibiotic therapy for PCP.

Given the recurring incurable nature of low-grade lymphoma, maintenance therapy has been an attractive modality to investigate. Although the initial remission duration may be prolonged, there is no evidence to suggest that either protracted single agents or maintenance with combinations of chemotherapy will prolong overall survival [179,180]. A biological agent, interferon (IFN), also may have an impact on remission duration in these patients.

Interferon

Interest in interferon therapy emerges from its unique mechanism of action and relatively mild self-limiting toxicity. In addition, it demonstrates activity in chemotherapy-resistant patients. The drug has been administered either as part of induction therapy or as maintenance after chemotherapy. It has been administered in a variety of doses and schedules with no clear superiority of one schedule over others. Its mechanism of action is discussed elsewhere.

As a single induction agent, the activity of IFN is substantial, both at diagnosis and relapse, with an overall response rate of about 50%, but the CR rate is only about 10% [181-183,151]. IFN has also been used in conjunction with chemotherapy as part of induction treatment. The ECOG conducted a trial showing that while response rates were similar among those receiving COPA (CHOP) alone vs COPA plus IFN, patients in the IFN-containing induction arm had a significantly longer remission duration, compared with COPA alone (P < .001) [185]. At 5 years, in 81% of those receiving COPA, disease had progressed, as opposed to 60% in the group receiving COPA and IFN (P < .001). A study by the Groupe d'Etude des Lymphomes Folliculaires (GELF) [186] supported these findings and also noted a survival advantage at 3 years in the IFN arm (86% vs 69%, P = .02). On the other hand, a randomized Cancer and Leukemia Group B (CALGB)-ECOG study has shown no benefit to date of adding IFN to single-agent cyclophosphamide in improving response rate, remission duration, or overall survival [133]. The latter study, however, treated patients at the time of initial diagnosis, unlike the ECOG and GELF studies, in which therapy was started after an initial “watch and wait” period.

Clinical studies have also investigated IFN in maintenance therapy. McLaughlin et al treated patients with stage-IV low-grade lymphoma with CHOP-Bleo followed by IFN maintenance. At 5 years, 48% of patients treated with CHOP-Bleo and maintained on IFN for 2 years remain failure-free, as opposed to 28% of historical controls treated with CHOP-Bleo alone (P = .01). However, similar to most other studies, there was no plateau in the relapse-free survival curve and no improvement in overall survival [187]. A trial by the European Organization for Research and Treatment of Cancer (EORTC) randomized patients into IFN maintenance therapy vs no therapy following eight courses of CVP and radiation therapy. There was a progression-free survival advantage noted in the IFN maintenance arm [188]. A British trial using chlorambucil with or without IFN [189] showed a significantly longer remission duration in the IFN maintenance arm without survival advantage.

In summary, the vast majority of programs that have integrated IFN into therapy for low-grade lymphoma have been associated with a benefit in prolonging remission duration, but the effect on survival is less convincing. Further studies and longer follow-up of ongoing trials will be required to establish the exact role of IFN in combination with various induction regimens and in different schedules as maintenance.

Mantle-Cell Lymphoma

To date, there is no accepted satisfactory treatment for mantle-cell lymphoma. As detailed above, the overall and disease free survival in patients treated with CHOP is rather poor [91]. Several reports have shown a fair frequency of achieving complete remission with standard CHOP-like therapy, averaging 40% to 50% [190,191], but relapses are the rule and long-term survival is uncommon [192]. A recent retrospective review of 26 patients with centrocytic lymphoma suggested that the inclusion of doxorubicin was beneficial [191]. However, a German report [192] of 63 patients with centrocytic lymphomas who were randomized to receive COP or CHOP showed no significant improvement in survival in the CHOP-treated group.

Some of the discrepancies among studies may relate to differences in the pathologists' criteria in identifying the morphology of this relatively newly defined entity. Given their generally worse outcome, patients with mantle-cell lymphoma are being treated at several institutions with more intensive therapy with or without bone marrow transplantation (BMT). The efficacy of this strategy remains to be seen.

High-Dose Chemotherapy With Autologous Bone Marrow Support

Given the incurability of low-grade lymphoma with conventional-dose therapy, high-dose therapy has been investigated. Prolonged disease-free survival is attainable in a small fraction of patients with advanced indolent lymphomas. To date, however, there is no proof that any category of patients is cured, since survival curves continue to decline, with approximately 5% of patients dying each year. In general, more than 50% of the patients will ultimately relapse after autologous BMT [193].

Bone marrow transplantation has mostly been reserved for young patients who have failed previous systemic therapy. The optimal high-dose regimen has not been identified. The sensitivity of follicular lymphomas to radiation therapy explains why most regimens have incorporated TBI as part of the preparative regimen. At the Dana-Farber Cancer Institute and St. Bartholomew's Hospital, patients with low-grade lymphoma mostly in second or subsequent remissions received high-dose cyclophosphamide and fractionated TBI, with autologous bone marrow rescue. The marrow was purged with MoAb and rabbit complement [194]. Recurrence rates were lower than expected at a median follow-up of 3.5 years, with only one-third of patients relapsing [195,196], when compared with a similar group of patients treated conventionally, but there was no improvement in survival and only those in second remission (as opposed to > 3 remissions) seemed to have improvement in freedom from progression. A European group reported 52% failure-free survival at 5 years in patients with chemosensitive disease in CR or good PR. The median follow-up of the study is still short at 19 months [197].

High-dose chemotherapy with autologous BMT has also been investigated in patients with transformed disease. Investigators from the University of Nebraska reported extremely poor outcome in patients transplanted after transformation. Most of their patients, however, had bulky advanced disease resistant to chemotherapy [198]. On the other hand, investigators from the Dana-Farber Cancer Institute did not report important differences in survival between those transplanted before or after transformation [199]. Their transformed disease patients, however, were sensitive to standard chemotherapy given before transplantation. Patient selection criteria, may explain the different results in these two studies.

High-dose chemotherapy with unpurged peripheral blood stem-cell (PBSC) support has also resulted in durable remissions in patients with relapsed indolent lymphomas [200,201]. An argument in favor of PBSC is that they may be less contaminated by neoplastic cells than is bone marrow.

Given the generally poor outcome once transformation occurs and given the encouraging preliminary results with BMT in relapsing patients, investigators are examining autologous BMT early in the course of the disease, ie, in first remission. Such trials are in their early phases, and longer follow-up will be required to show whether early BMT will have an impact on disease-free or overall survival [198,202].

At present, high-dose chemotherapy with autologous rescue remains investigational. Although early data from the Dana-Farber Cancer Institute (DFCI) [203] suggested that those who achieved CR with conventional chemotherapy prior to BMT had better disease-free survival than those in PR, a recent update no longer shows a significant difference between the two groups [204].

A recent study from Germany examined the assumption that PBSC may be less frequently contaminated with neoplastic cells than is the bone marrow and, thus, may serve as a better source for stem-cell support in conjunction with high-dose chemotherapy. The peripheral blood autografts were PCR-positive in 22 of 30 patients (as opposed to the expected 100% yield from an unpurged bone marrow following standard chemotherapy). Of those who received PCR-positive autografts, 6 of 22 converted to PCR-negative 6 to 16 months post-transplantation. This suggests that neoplastic cells may not be viable after peripheral blood stem cell collection and reinfusion. The median follow-up was too short to reach any conclusions beyond that [205].

Allogeneic BMT has not been commonly employed in follicular lymphomas, but encouraging data do exist. In a recent trial, 8 of 10 patients with refractory or recurrent disease achieved remission and remain relapse-free at a median follow-up time of 816 days [206]. Though preliminary, these data seem to be an improvement over results with autologous transplant, suggesting a role for the graft-vs-lymphoma (GVL) effect. However, this has to be weighed against the increased transplant-related morbidity and mortality from possible graft-vs-host disease (GVHD). In addition, despite promising data from early clinical trials, laboratory data demonstrate that follicular lymphoma cells are poor stimulants of allogeneic T-cells in mixed lymphocyte reactions [207]. This may be explained by the follicular lymphoma cells lack of expression of the B7 family of molecules that initiate a necessary costimulatory signal in T-cells via CD28. It remains to be seen, however, whether this will translate into lack of clinically evident GVL effect in transplanted patients.

PCR in Assessing Response and Monitoring for Relapse

Molecular techniques using PCR amplification of bcl-2 minor and major breakpoints detect clonal cells in the bone marrow in most patients at the time of diagnosis, after remission induction with conventional therapy, and at the time of relapse [208]. PCR has also been used to assess efficacy of ex vivo purging of bone marrow. In a DFCI trial, all 114 patients in the study had PCR-positive bone marrow after treatment with CHOP at the time of harvest. Following purging with MoAb, 50% of the bone marrow turned PCR negative and such patients had a markedly more favorable disease-free survival posttransplant [209,210]. The group from St. Bartholemew's Hospital could not replicate such a high percentage of negative conversion following purging [211], and three of the four patients whose bone marrow did turn negative with ex vivo purging relapsed. Such discrepancies may be explained by the fact that the British investigators used only one antibody to purge the bone marrow, whereas the DFCI investigators used three. This difference notwithstanding, overall survival of patients at the two centers is the same [196].

PCR has also been used to monitor response after high-dose chemotherapy and autologous bone marrow rescue [212] as well as following conventional-dose chemotherapy [151]. Preliminary results from the DFCI group were encouraging in that those who were persistently PCR-negative or became negative several months after BMT had no relapses (none of 77 patients at 6 years), while those who were persistently PCR positive ultimately relapsed (71% of 35 patients at 6 years). Similarly, McLaughlin et al [151] reported 19 patients with advanced-stage disease treated with three intensive sequential chemotherapy regimens. These 19 patients were originally positive in peripheral blood for bcl-2 rearrangement by PCR and had serial monitoring; 13 turned PCR negative after therapy, a superior result to what was previously observed with CHOP or other less intensive therapy [213-215]. Such “molecular remissions” seem to correlate with a lower likelihood of relapse (1 of 13 molecular remissions relapsed vs 2 of 6 patients with persistently positive PCR).

The above data indicate that some high-dose chemotherapy programs and some intensive standard-dose therapy are capable of converting patients into PCR-negative status. In the BMT patients, outcome also appears to depend on the successful purging of the autograft. The indication that PCR negativity correlates with disease-free survival may represent a breakthrough, since it identifies for the first time a surrogate molecular marker for long-term disease-free survival in patients who would otherwise require very long follow-up to show a favorable survival outcome.

A positive PCR for bcl-2 breakpoints may not be a highly specific predictor for relapse, however. The group from St. Bartholomew's Hospital reported six patients in remission for more than 10 years who are persistently PCR-positive [215]. Similar observations were made by investigators from M.D. Anderson [216]. In addition, some groups [27] have obtained positive PCR results in the peripheral blood of normal individuals. In fact, recent data suggest that when peripheral blood cells from normal individuals were sorted for B-cells, more than half of the individuals tested were found to harbor t(14;18) breakpoints, and sometimes, of several unrelated clones in the same individual as demonstrated by DNA sequencing [217]. Therefore, it is not surprising that patients can be clinically disease free for a long time with persistently amplifiable fragments.

Nevertheless, a negative PCR may indeed be a highly specific predictor of favorable prognosis; this seems most convincing in the context of BMT, but the preliminary data described above with intensive conventional chemotherapy is also promising. One recent study [218] challenged this concept and concluded that PCR has no positive or negative prognostic significance. However, this study included only eight patients treated with conventional CVP (with or without radiation), and the whole argument is based on a single patient in whom, despite clinically progressive disease, PCR was always negative; the patient's PCR status at the time of diagnosis was not mentioned.

The premise that peripheral blood may be less contaminated by neoplastic cells than the marrow will have an impact on which source is used for disease monitoring (and harvest). The DFCI data looking at marrow and peripheral blood at the time of harvest and relapse indicates that peripheral blood is indeed less sensitive for monitoring disease [208]. At diagnosis, however, the PCR yield for bcl-2 is high in both peripheral blood and marrow, and the results are highly concordant [126]. After therapy, reversion to PCR negativity of either peripheral blood or marrow is prognostically favorable, although most strikingly so with marrow monitoring [208]. It remains to be seen whether the simplicity and practicality of peripheral blood monitoring is outweighed by the increased sensitivity of BM monitoring. Conceivably, the development of reliable quantitative PCR assays (titers) will provide another useful perspective on this threshold of detection issue [219].

Conclusions

A number of factors can influence treatment decision for patients with newly diagnosed indolent lymphomas, including age, stage, histologic diagnosis, and the patient's general health and performance status. Prognostic factors are not as universally agreed upon as is the case in intermediate-grade lymphoma. The watch-and-wait policy is still an acceptable option in asymptomatic individuals without threatening disease. Older patients or those with poor performance status or other medical problems may be palliated with oral chlorambucil or cyclophosphamide. Oral purine analogs may also become available for this purpose [220].

Younger patients with advanced-stage disease may warrant consideration of early intensive chemotherapy. However, no single regimen is demonstrated to be superior to others. The importance of including doxorubicin is not as clear as in the more aggressive lymphomas, but there are data to support the use of early intensive therapy in some subsets of patients, eg, those with follicular mixed lymphomas. The role of fludarabine-containing regimens as front-line therapy is not yet well defined and is currently being investigated in clinical trials. Integration of alpha interferon (IFN-alfa) into front-line therapy appears to result in improved relapse-free survival. The younger patient in relapse may benefit from high-dose therapy with bone marrow or peripheral stem-cell support in the context of a clinical trial.

Histologic Progression and Clinical Transformation

In general, histologic progression corresponds with an accelerated clinical course. It usually evolves from follicular to diffuse histology with increasing numbers of large cells [221]. This histologic progression is associated with loss of the characteristic dendritic reticular cells constantly seen in follicular lymphomas. While approximately 30% of patients demonstrate progression on rebiopsy [221], an autopsy study suggests that about 70% of patients with an initial diagnosis of follicular lymphoma who died with the disease had only a diffuse pattern at the time of death, while 6% had preserved follicular pattern. Cytogenetic analysis often indicates the acquisition of additional chromosomal abnormalities upon transformation. In one series, these included +7, +3, del(13)q32, and +18 [69]. In addition, several series demonstrated p53 mutations in a sizable number of patients with follicular lymphoma who undergo histologic transformation (up to 30% in one series).

Whether this simply represents a secondary phenomenon or has a mechanistic role in progression is not clear at this time. The latter assumption is reasonable since loss of p53 leads to enhanced cycline dependant kinase (cdk) activity with subsequent increase in retinoblastoma protein phosphorylation; this leads to the release of transcription factor E2F, resulting in increased expression of several genes that contribute to S-phase entry, including c-myc, fos, and myb among others [222,223].

The incidence of clinical transformation is difficult to assess, but approaches 40% to 70% at 8 to 10 years of follow-up, and its incidence does not seem to be affected by previous therapy [224,225]. Patients have clearly transformed without having received treatment. Indeed, the data from Ig-bcl-2 transgenic mice suggests that transformation is an inherent feature. The time to transformation is highly variable, ranging between 8 months and 25 years, and, at least in the Stanford series, there does not seem to be a plateau in the rate of transformation in both initially treated and untreated patients. This extremely wide range of time to transformation is another demonstration of the high degree of heterogeneity in this disease.

In general, patients with histologic transformation have a poor prognosis, with survivals of less than 1 year often reported [226,227]. Prognostic factors at the time of transformation include disease bulk and response to chemotherapy [228]. In the Stanford data, the CR rate in transformed patients approached 40% with a median survival of 95 months for those complete responders [228]. The role of high-dose chemotherapy and autologous BMT in relapsed patients remains uncertain and is discussed elsewhere.

In summary, transformed patients should be treated with intensive chemotherapy regimens. As is the case for intermediate-grade lymphoma, there is no evidence that any regimen is superior to CHOP, but for those who have already received CHOP or other regimens during the indolent phase of the disease, it is probably advisable to select alternative non-cross-resistant regimens. While the definitive role of BMT in transformed lymphoma remains unsettled, patients demonstrating adequate response to conventional therapy should be good candidates for enrollment in clinical trials that incorporate BMT.

Future Directions

To date, most advanced-stage indolent lymphomas remain incurable, and new approaches to therapy are clearly needed. High-dose chemotherapy with autologous bone marrow or PBSC support is being increasingly used earlier in the disease course. Purging techniques and PCR monitoring are under investigation. New active drugs are being incorporated in combinations [229,230].

Novel approaches to therapy are also being explored. Monoclonal antibodies [231] have been extensively investigated. These include unconjugated antibodies, radioimmunoconjugates, and immunotoxins. The exact role and efficacy of these agents remain to be established. Individualized therapy by generating anti-idiotype monoclonal antibodies has resulted in good responses, and the concept is being developed into stimulating endogenous anti-idiotype MoAbs by utilizing vaccines [232-234]. New biological agents are also being examined, including IL-2, IL-4, and recombinant fusion toxins [235] (eg, DAB486IL-2).

Recently, intense research has focused on understanding the molecular mechanisms of apoptosis in B lymphocytes. The demonstration that murine and human “immature” B-cells may undergo apoptosis instead of proliferation [236] in response to cross-linking surface immunoglobulin receptors raises interesting questions about exploiting similar signal transduction pathways to induce apoptosis in human follicle center lymphomas. Bcl-2 antagonism and Bax upregulation through manipulating signalling remain elusive therapeutic goals that may prove synergistic with direct triggering of apoptosis by chemotherapy or radiation. The role of Fas-FasL interaction in inducing apoptosis in B-cells [7,237] is becoming better understood and raises questions about how the enforced expression of bcl-2 in follicular lymphomas may influence apoptotic signals through fas in neoplastic cells.

We still face many challenges in finding curative therapy for the indolent lymphomas. Future effective therapies will most likely emerge from a close interaction between basic scientists and clinical investigators, bringing knowledge acquired through fundamental research from the bench to the bedside.