Should Adolescents with NHL Be Treated as Old Children or Young Adults?

Children with non-Hodgkin lymphoma (NHL) reportedly have an outcome superior to that of adults.^1,2 Within the pediatric population, there are reports of a poorer outcome for some adolescents than for younger children.^2,3 These differences among pediatric patients with NHL have been reported to vary according to gender and histologic subtype. The SEER (Surveillance, Epidemiology, and End Results) data covering the years 1975–1998 indicate that the probability of survival in patients with NHL declines with increasing age: when survival was analyzed according to age in 5-year increments from birth through 44 years, survival progressively declined as age increased (Figure 1; see Color Figures, page 517).² Many factors may contribute to these age-related differences in treatment outcome. Some factors are related to the host, whereas others pertain to tumor histology and biology.

Drug distribution and clearance are known to vary with age,⁴ and the maximum tolerated dose (MTD) of many anticancer agents is higher in children than in adults.⁴ Concomitant drug administration, which may also differ with age, may also influence drug clearance and metabolism. For example, certain anticonvulsants induce the hepatic cytochrome P450 enzymes, thereby enhancing the metabolism (i.e., clearance) of various chemotherapeutic agents (e.g., vincristine, etoposide). A study of children with ALL showed that those receiving anticonvulsants had a poorer treatment outcome.⁵ 

Performance status, including the underlying status of organ function, also influences treatment outcome. This factor is of concern in the elderly but is less likely to affect adolescents and adults in their twenties. Among adults, the international IPI score, which incorporates performance status, is used for screening and prognosis.⁶ This score is not generally used in the management of NHL in children and adolescents.

The patient’s compliance with the treatment plan is another critical factor in treatment outcome. In the pediatric population, adolescents may be less likely than others to remain compliant. Lack of compliance is often difficult to document, although certain studies, such as the measurement of red blood cell 6TGN in the case of oral 6-mercaptopurine therapy, may give some indication.⁷ Compliance may be improved if rigorous adherence to protocol guidelines is emphasized and optimal supportive care is provided in the clinic and at home.

Among children and adolescents, gender is reported to be associated with age-related outcome for certain histologic subtypes of NHL. For example, a large Berlin-Frankfurt-Münster (BFM) study showed that among patients with T-lymphoblastic lymphoma and DLBCL, the outcome of female patients was significantly inferior, particularly among adolescents.³ The reasons for these differences have yet to be elucidated, although the potential role of pubertal changes has been suggested.

Outcome is influenced by age-related differences in the distribution of NHL histologic subtypes and by the selection of appropriate histology-directed treatment.^2,8 Further complicating the picture, age-related biologic differences within specific histologic subtypes may also influence treatment results. It is therefore important to identify the relative proportion of histologic subtypes and associated biologic features within various age groups and to examine what treatment strategies have been reported to be the most effective for these groups. A common and somewhat controversial challenge centers on the selection of therapy for adolescents and young adults (ages 15 to 29 years) with NHL. The remainder of this review will focus on age-related differences in tumor histology, tumor biology, and treatment selection.

The annual incidence of NHL steadily increases throughout life, in contrast to the bimodal age distribution of Hodgkin lymphoma, which peaks at ages 20 to 25 years and has a smaller peak at ages 75 to 80 years.² In the year 2000, the estimated incidence of NHL per million in the United States (SEER) was 2.8 (66 patients) for ages < 5 years, 5.5 (110 patients) for ages 5–9 years, 8.8 (147 patients) for ages 10–14 years, 14.3 (290 patients) for ages 15–19, 21.8 (413 patients) for ages 20–24 years, and 39.3 (762 patients) for ages 25–29 years.² 

The spectrum of NHL subtypes in children has long been known to differ from that observed in adults (Figure 2; see Color Figures, page 517). The NHL subtypes in children are predominantly extranodal high-grade tumors, whereas those that occur in adults are more commonly low-and intermediate-grade tumors. In a large single-institution review of childhood NHL, the spectrum of subtypes included Burkitt lymphoma (38.8%), lymphoblastic lymphoma (28.1%), large-cell lymphoma (26.3%), and rare subtypes, including follicular lymphoma (6.8%).⁹ In striking contrast, adults in North America most commonly have follicular lymphoma and DLBCL; Burkitt lymphoma and lymphoblastic lymphoma occur relatively infrequently.

The SEER data provide insights about the specific age ranges in which these differences begin to emerge.^2,8 From ages 5 to 14 years, Burkitt lymphoma is the predominant histologic subtype, whereas diffuse large B-cell lymphomas are the most common subtype in the 15- to 19-year age group.^2,8 More specifically, the diffuse large B-cell subtype is the predominant histologic diagnosis in the 15–19, 20–24 and 25–29 year age groups.² The proportion of cases of follicular lymphoma is also greater in the 25- through 29-year age group than in younger age groups.² Thus, age 15 to 29 years appears to coincide with the transition from pediatric to adult subtypes of lymphoma (Figure 3; see Color Figures, page 517).² 

Stage- and histology-directed multiagent chemotherapy is the foundation of modern therapy for NHL. Involved-field radiation is not widely used for newly diagnosed disease, because it did not appear to improve treatment outcome if appropriately intensive chemotherapy was used in children with limited¹⁰ or advanced-stage¹¹ NHL and in adults with localized aggressive disease.¹² Debulking surgery is not indicated except in the case of a localized terminal ileocecal mass with mesenteric nodal involvement only, which can be completely resected. Complete resection alters the tumor stage and the indicated therapy.

The Children’s Cancer Group (CCG) trial, which compared the cyclophosphamide-based COMP regimen with the LSA₂L₂ regimen designed for ALL, demonstrated that the COMP regimen was superior for the treatment of advanced- stage Burkitt lymphoma, whereas the LSA₂L₂ regimen was superior for the management of advanced-stage lymphoblastic lymphoma.¹³ Subsequent refinement, including the addition of high-dose methotrexate and high-dose cytarabine to a cyclophosphamide-based backbone^14,–16 and the further intensification of therapy,^17,–19 have resulted in cure rates in excess of 80%. In the NCI protocol, results were improved by adding the IVAC regimen (ifosfamide, etoposide, high-dose cytarabine) to the previously studied CODOX-M backbone (cyclophosphamide, vincristine, high-dose methotrexate, and intrathecal therapy).¹⁹ Excellent results were also achieved with the BFM90 regimen, which featured three risk groups in which the intensity of therapy was tailored to the degree of surgical resection, serum LDH activity, and tumor stage.¹⁷ The 6-year event-free survival (EFS) for stages III and IV NHL and B-ALL were 86%, 73% and 74%, respectively. The French LMB-89 regimen also features a risk-based 3-arm regimen with intensification based on degree of resection and disease stage.¹⁸ The 5-year EFS for stages III and IV NHL and BALL were 91%, 87% and 87%, respectively. An international collaborative trial (LMB-96) confirmed these excellent results.^20,21 Unifying features of these highly effective pediatric regimens include risk-based treatment, brief but intensive multiagent cyclophosphamide-based chemotherapy, and aggressive central nervous system (CNS) prophylaxis and treatment.

Adults with Burkitt lymphoma can be successfully treated with regimens designed for children. For example, the NCI CODOX-M/IVAC regimen resulted in an equally excellent result for adults and children with Burkitt lymphoma.¹⁹ Similarly, French investigators have demonstrated that an excellent result can be achieved in adults with Burkitt lymphoma by using pediatric LMB-based regimens.²² In their study, the 3-year overall survival rate was 80% for patients with stage III NHL and 57% for those with stage IV NHL or B-ALL. Other effective regimens designed for adults share the features of pediatric regimens described above.^23,24 For example, the hyper-CVAD regimen features courses of cyclophosphamide, vincristine, doxorubicin, and prednisone alternating with courses of high-dose methotrexate and cytarabine given in a dose-intensive fashion.²³ The outcome of treatment with hyper-CVAD²³ appeared to be improved by the addition of rituximab, particularly in elderly patients.²⁴ 

Gene-expression profiling has recently advanced our understanding of the molecular features of Burkitt lymphoma.^25,–27 Studies that focus specifically on children are under way and may lead to a greater understanding of the potential biologic differences of the Burkitt lymphomas in children and in adults. A study comparing the karyotypic abnormalities in Burkitt lymphoma in children versus adults found no significant difference in the karyotype complexity or the nature of the chromosomal abnormalities; however, the prognostic significance of specific chromosomal abnormalities differed in the two groups.²⁸ In children, abnormalities of chromosome arms 22q and 13q were associated with a poorer outcome, whereas in adults, abnormalities of chromosome 17 appeared to be associated with risk.²⁸ 

Contemporary treatment for Burkitt lymphoma in children and adults is very similar. Thus, in the case of this histologic subtype, there is little controversy about using a dose-intensive therapeutic approach that emphasizes alkyators (e.g., cyclophosphamide, ifosfamide), high-dose methotrexate, and cytarabine. Continued study of both pediatric and adult Burkitt lymphoma is needed to identify potential prognostic markers, and more importantly, potential targets for novel therapeutic agents.

Strategies to improve the cure rate for children with lymphoblastic lymphoma have generally built on the results of the CCG trial demonstrating that an ALL-directed approach (induction, extended maintenance, and CNS prophylaxis components) is superior to the cyclophosphamide-based COMP regimen.¹³ The best result to date (a 5-year EFS of 90%) was achieved with the BFM90 regimen, which features an induction phase, a consolidation phase of 4 courses of high-dose methotrexate (5 g/m²), and a maintenance phase of 6-mercaptopurine and low-dose methotrexate.²⁹ Components of this and other successful regimens that are thought to contribute to improved treatment outcome include high-dose methotrexate, asparaginase, anthracyclines, a re-intensification (i.e., re-induction) phase, and a prolonged maintenance phase (approximately 2 years of therapy).³⁰ 

As in children, treatment outcome in adults with lymphoblastic lymphoma has been improved with regimens derived from ALL protocols.^1,31,32 Nevertheless, the outcomes in adults with lymphoblastic NHL appear to be inferior to those in children. In sequential GMALL studies that featured multiagent induction, consolidation, and re-induction phases delivered over 6 to 12 months, overall survival, continuous complete remission (CR), and disease-free survival estimates at 7 years were 51%, 65%, and 62%, respectively.³¹ The MD Anderson study of hyper-CVAD resulted in 3-year progression-free survival and overall survival estimates of 66% and 70%, respectively.³² A recent study in Vancouver suggested that outcome was improved by incorporation of an intensification phase featuring hematopoietic stem cell support for patients in partial remission (PR) or CR after induction therapy;¹ however, others have observed no benefit from addition of such an intensification phase.³³ The apparent worse outcome in adults than in children treated with ALL-derived regimens has yet to be explained but may reflect age-related differences in the biologic features of these tumors.

The majority of pediatric lymphoblastic lymphomas have a precursor T-cell immunophenotype; the precursor B-cell immunophenotype is less common. There is very little information available about the biologic features of pediatric lymphoblastic lymphoma. This lack probably reflects the limited tissue obtained from non-invasive biopsy samples of anterior mediastinal masses with associated airway compression. Recently, the cytogenetic analysis of 13 pediatric patients with advanced-stage lymphoblastic lymphoma revealed a high (85%) frequency of chromosomal abnormalities.³⁴ A larger cohort of children must be studied to determine the frequency and prognostic significance of genetic abnormalities in these tumors. Data from gene array studies may provide insight into the potential differences between lymphoblastic NHL in adults and children. These studies may also identify genes whose expression is associated with a poorer prognosis, as has been done in T-cell lymphoblastic leukemia. Specifically, HOX11 activation is reportedly associated with a superior outcome, and TAL1 or LYL1 expression with a poorer outcome, in patients with T-ALL.³⁵ The expression of these genes has yet to be studied in large numbers of patients with lymphoblastic lymphoma; however, preliminary array data from children have been reported,³⁶ and additional studies are under way in the current Children’s Oncology Group (COG) trial.

A multiagent regimen derived from successful strategies to treat T-cell ALL is indicated for children and young adults; the current COG trial for lymphoblastic lymphoma is open to patients up to 35 years of age. Moreover, the demonstration that young adults with ALL have a superior outcome when treated with more intensive pediatric regimens³⁷ provides a strong rationale for a similar approach in the treatment of young adults with lymphoblastic lymphoma. There is clearly a need for biologic studies of both pediatric and adult lymphoblastic NHL to identify patients with a poor prognosis for whom early intervention is indicated and to identify possible targets for novel therapeutic interventions.

Successful therapies for anaplastic large-cell lymphoma (ALCL) in children are more diverse than those for other histologic subtypes of childhood NHL. In Europe, an immunophenotype-directed approach has historically been used in the treatment of large cell lymphomas, including ALCL.^38,39 The BFM have reported an excellent result using a B-cell approach (i.e., a strategy like that used to treat Burkitt lymphoma), achieving a 5-year EFS of 76% for all stages combined.³⁸ In the United States, a histology-based approach has historically been used to treat children with large cell lymphoma. Many of these regimens are CHOP-based.⁴⁰ One of the most successful (a 70% EFS for advanced-stage disease) is the anthracycline-containing APO regimen, which features an induction phase followed by sequential maintenance phases delivered over one year.⁴¹ Current cooperative group trials in Europe and the United States are building on the BFM B-cell and APO strategies, respectively. Both studies are examining the benefit of adding vinblastine, which is active as a single agent against multiply relapsed ALCL.⁴² 

Adults with ALCL are also generally treated with a multiagent anthracycline-containing regimen.^43,–48 However, the treatment outcome is inferior to that achieved in children. Age-related biologic differences in the lymphomas may explain this disparity.

In a number of adult trials, patients with anaplastic lymphoma kinase (ALK)⁺ ALCL had an overall survival rate superior to that in ALK⁻ patients;^45,–47 however, some studies have not observed this survival advantage.⁴⁹ In those that do, the younger age of patients with ALK⁺ disease is suggested to be responsible for the advantage. The majority of pediatric patients with ALCL are ALK^+.43 Among adult patients with ALK⁺ ALCL, the 5-year overall survival rates have approached those achieved in children. It is important that ALK⁻ ALCL not be confused with lymphomatoid papulosis and primary cutaneous ALCL, which are also ALK⁻ and do not require intensive chemotherapy.

Other biologic markers have been thought to have prognostic significance in ALCL.⁵⁰ In a large study of 143 cases of adult and pediatric ALCL, CD56 expression was also identified as an independent predictor of outcome, regardless of ALK expression status.⁵⁰ Although children were included in this study, their number was not defined. A subsequent case report and review of the literature did not identify a uniformly aggressive clinical course for children with CD56⁺ ALCL but indicated the need to study a larger cohort of patients with this phenotype.⁵¹ 

A recent COG study of 40 pediatric patients with ALCL identified ALK expression and activation of STAT3 in most tumors; however, CD56 expression was infrequent, as was cytoplasmic localization of survivin and TIMP1.⁵² The authors suggested that although STAT3 is activated, the downstream anti-apoptotic effect of activated STAT3 is likely to be limited by other mechanisms, and that these effects may contribute to the relatively good prognosis of ALCL in children.

Although the EFS rates of 70% to 80% achieved with the pediatric regimens described above^38,41 suggest that they offer an advantage to adolescents and very young adults, they have not been proven in randomized trials to be superior to the most commonly used adult regimens with respect to either age group or ALK status. However, the outcome of the small number of ALK⁻ children treated with the BFM approach was not substantially inferior to that of ALK⁺ patients.³⁸ It is therefore reasonable to use the BFM or APO regimen for adolescents and very young adults with ALCL; however, outcomes may still worsen with increasing age because of the known and unknown biologic differences between ALCL in adults and children. Future advances may be achieved by including an immunotherapeutic agent such as anti-CD30.^53,54 Clinical trials of anti-CD30 in combination with conventional chemotherapy are currently under way in both children and adults. The success of allogeneic bone marrow transplantation (BMT) for refractory or high-risk pediatric ALCL also suggests the potential immunotherapeutic benefit of a graft-versus-lymphoma effect.⁵⁵ 

The outcome of children with DLBCL is excellent. The approach most commonly used is that described earlier for Burkitt lymphoma. For example, in the international LMB-96 trial, 4-year EFS for patients with DLBCL (excluding primary mediastinal disease) was 92%, compared with 93% for those with Burkitt lymphoma.²⁰ Those with primary mediastinal tumor had a slightly inferior 4-year EFS of 71%. The APO regimen has also been shown to have activity against DLBCL, but the 4-year EFS of 63.8% was inferior to that achieved with the Burkitt lymphoma regimen LMB-96.⁴¹ 

The prognosis for adults with DLBCL is inferior to that for children. CHOP has been the most widely used combination for more than 30 years for DLBCL in adults: the addition of other agents to the CHOP backbone (e.g., CHOP-BLEO, M-BACOD, MACOP-B, ProMACE-CytaBOM) failed to improve on the outcome achieved with CHOP.⁵⁶ Subsequent trials focused on intensification of therapy, but the addition of rituximab to the CHOP regimen has had the greatest effect on outcomes for adults with DLBCL. The treatment of adults with DLBCL is covered in detail by Dr. Habermann in this volume.

Gene expression profiling studies of adult DLBCL have identified two molecularly distinct subtypes: those with a germinal center (GC) B-like phenotype and those with an activated B-cell (ABC)–like phenotype.⁵⁷ Among adults, DLBCL with a GC phenotype has a better outcome than DLBCL with an ABC phenotype.⁵⁷ Most cases of DLBCL in children were recently found to have a GC phenotype.⁵⁸ The authors suggest that this finding may explain the superior treatment outcome observed in children. The high frequency of the GC phenotype in pediatric patients was confirmed in a subsequent international study.⁵⁹ That study also showed that pediatric patients had higher proliferation rates, greater c-MYC protein expression, and less BCL2 protein expression than is reported in adults.⁵⁹ The authors reported that pediatric DLBCL, whose phenotype is similar to Burkitt lymphoma, appear to be characterized by c-Myc–driven proliferation, whereas adult DLBCL is characterized by BCL2 protein expression and associated inhibition of apoptosis.⁵⁹ 

Gene-expression profiling of the mediastinal large B-cell lymphoma (MLBCL) differentiates it from other DLBCLs and suggests that it is related to Hodgkin lymphoma. Adults with MLBCL have a better outcome than adults with other DLBCLs.

The excellent outcome achieved in pediatric DLBCL using treatment designed for children with Burkitt lymphoma suggests that it is worthy of consideration for older adolescents and very young adults whose tumor is characterized by a pediatric phenotype. In contrast, for those whose tumor is characterized by an adult phenotype (i.e., BCL2⁺), a strategy that incorporates rituximab may be preferable. Clearly, additional clinical trials are needed to definitively answer these questions.

The follicular lymphomas occur very infrequently in children and adolescents;⁶⁰ however, when they do occur, questions often arise about the optimal treatment approach. Historically, many have been successfully treated with low-intensity CHOP-based regimens.⁶¹ Follicular lymphomas in children were found to differ biologically from those in adults. Those occurring in children are generally BCL2-negative; when BCL2-positive, they tend to be more advanced in stage and have a less successful treatment outcome.⁶¹ In some studies, children with completely resected limited-stage disease have been observed rather than being given systemic chemotherapy.⁶² 

The spectrum of treatment approaches for managing follicular lymphoma in adults is addressed in detail in the manuscript by Salles in this publication. Options include observation, involved-field irradiation, conventional-dose chemotherapy, high-dose chemotherapy with stem cell support, immunotherapy with rituximab, other new antibodies or radioimmunotherapy, alone and in combination with chemotherapy, depending on clinical stage and disease status.

Close observation only may be considered for adolescents and very young adults with completely resected, BCL2⁻, limited-stage disease, although mild systemic chemotherapy has often been given. For patients with BCL2⁻ but incompletely resected limited-stage disease or BCL2⁺ disease, low-dose systemic chemotherapy is usually given. Involved-field irradiation is generally avoided in children. The role of rituximab in this age group has not been determined, but the agent is worthy of consideration for bulky, BCL2⁺ disease.

The non-ALCL mature T-cell lymphomas occur very infrequently in children.⁶⁰ Therefore, there is very little information about their clinical and biologic features in various age groups. Historically, children and adolescents with these lymphomas have been treated on large-cell lymphoma regimens. Current treatment strategies are derived from those that have been successful in adults. Diagnoses that carry a particularly poor prognosis, such as hepatosplenic T-cell lymphoma, may require more intensive or novel frontline therapy. Additional clinical and biologic data on these tumors in both adults and children are needed before age-specific treatment recommendations can be made.

Age-related differences in the outcome of treatment for NHL reflect both host factors and tumor histology/biology factors. It is reasonable to consider pediatric strategies for some adolescents and very young adults with NHL. In fact, pediatric strategies are currently used to treat adults with certain subtypes of NHL (e.g., Burkitt lymphoma, lymphoblastic lymphoma). However, the use of pediatric strategies in the treatment of adults does not guarantee a comparable outcome, as illustrated above in the discusssion of lymphoblastic lymphoma. There is clearly a need for further biologic study of NHL in children, adolescents, and young adults. The tumor biology of ALCL and DLBCL has been shown to differ with age. Additional biologic data will not only help to improve prognosis and treatment stratification but, more importantly, will allow the identification of specific molecular targets for therapy.

Acknowledgments

Supported in part by a Grant from the National Cancer Institute (CA 21765) and by the American Lebanese Syrian Associated Charities (ALSAC).