Pediatric Hodgkin Lymphoma

Staging

Physical examination and diagnostic imaging evaluations are used in pediatric patients to designate a clinical stage according to the Ann Arbor staging system, while in adult practice this has been modified by the Lugano staging classification of lymphomas. This historical staging system is based on anatomical groups of regional lymph nodes as delineated at the 1970 Ann Arbor symposium. It was subsequently revised at the Cotswolds meeting in 1989, although not all recommendations are in current use. This staging system persists to date with stage I identifying disease confined to one nodal region. Stage II signifies disease confined to two or more regions, but on one side of the diaphragm. Stage III indicates nodal disease that is on both sides of the diaphragm, while stage IV is diffuse or disseminated disease that includes extranodal sites. A special subset of stages I and II disease may be designated as extralymphatic or extranodal with the “E” suffix, which historically had been interpreted to be a disease distribution that could be encompassed within a reasonable radiation treatment portal. This operational definition is currently problematic, as the role of RT is being curtailed, but still is needed for prognostic purposes to designate early stage. Largely ignored from the Cotswold recommendations in current practice is the suffix “X” to designate bulky disease (> 10 cm maximum dimension) and a category of response to therapy, unconfirmed/uncertain complete remission (CR[u]), introduced to accommodate the difficulty of persistent radiological abnormalities of uncertain significance prior to the current use of metabolic imaging. PET imaging has become important for initial staging as well as for response assessment to therapy.

Lymphoma staging continues to be unique with the inclusion of prognostically significant constitutional symptoms. The absence of symptoms is labeled with the suffix “A” while “B” symptoms included in the staging assignment are unexplained fever with temperatures taken orally that are higher than 38°C, unexplained weight loss of 10% within 6 months preceding diagnosis, and drenching night sweats. At one point, pruritus had been included as a B symptom but was dropped many decades ago.

In the past, pathological staging, based on the findings of a staging laparotomy, including splenectomy, was routinely used to assess infradiaphragmatic disease. The increasing use of systemic therapy in children and the development of more accurate diagnostic imaging modalities eventually led to the routine use of clinical staging and abandonment of surgical staging except to assess any equivocal findings. Currently, surgical staging—most typically, nodal sampling without splenectomy—is pursued only if the anticipated findings will significantly alter the treatment plan.

Clinical Manifestations

Pediatric patients most commonly present with painless cervical or supraclavicular lymphadenopathy. Mediastinal lymphadenopathy occurs in up to two-thirds of patients and may be associated with a nonproductive cough or other symptoms of tracheal or bronchial compression. Axillary or inguinal lymphadenopathy is less frequently seen as the first presenting sign. Primary infradiaphragmatic disease is rare in pediatric patients, occurring in fewer than 5% of cases. Splenic involvement occurs in 30% to 40% of pediatric patients with HL, whereas hepatic involvement is exceedingly rare. The pulmonary parenchyma, chest wall, pleura, and pericardium are the most commonly involved extranodal sites of disease. Bone marrow involvement at the time of initial presentation of HL is also uncommon in children. Approximately 65% of children have stages I and II disease and 35% have stages III and IV disease ( Table 85.1 ).

Nonspecific systemic symptoms are often associated with lymphadenopathy and may include fatigue, anorexia, mild weight loss, and pruritus. The prognostically significant B symptoms have already been defined earlier. B symptoms occur in approximately 33% of patients (see Table 85.1 ). Laboratory changes observed at presentation are nonspecific but may provide clues about the extent of disease. Hematological abnormalities may include anemia, neutrophilic leukocytosis, lymphopenia, eosinophilia, and monocytosis. Anemia may be associated with the presence of advanced disease and may result from impaired mobilization of iron stores or, less commonly from, hemolysis. Several autoimmune disorders have been reported in patients with HL, including nephrotic syndrome, autoimmune hemolytic anemia, autoimmune neutropenia, immune thrombocytopenia, and autoimmune lymphoproliferative disorder. These conditions typically remit as the lymphoma responds to therapy. Several acute-phase reactants—including erythrocyte sedimentation rate, serum copper, ferritin, and C-reactive protein levels—may be elevated at diagnosis and useful in follow-up evaluations.

Patient Evaluation

Posteroanterior and lateral thoracic radiographs should be performed as soon as HL becomes part of the differential diagnosis to assess mediastinal involvement, airway patency, and other intrathoracic structures. This is particularly important if sedation is planned for diagnostic procedures. Either an excisional lymph node biopsy or core needle biopsy is the preferred diagnostic procedure (avoiding cytological fine needle aspirates), because it permits evaluation of the malignant Hodgkin Reed-Sternberg cells within the background of characteristic architectural changes of the specific histological subtypes. All nodal regions, including Waldeyer's ring, should be assessed by careful physical examination. Historically, an upright chest radiograph was also important to assess mediastinal bulk that is defined by mediastinal lymphadenopathy measuring 33% or more of the maximum intrathoracic cavity. Whether this definition for bulk mediastinal adenopathy may be replaced by CT-based mass dimensions or volumes is currently under debate. CT is most frequently used to evaluate the nodal regions in the neck, axilla, thoracic and abdominal cavities, and pelvis. Administration of both oral and intravenous contrast agents is required for CT to accurately distinguish lymphadenopathy from other infradiaphragmatic structures. Organ size is an unreliable indicator of lymphomatous involvement in the liver or spleen because tumor deposits may be less than 1 cm in diameter and not visualized by diagnostic imaging modalities. Increased size of either organ can, of course, be caused by nonmalignant conditions. The presence of hypodense lesions by CT and/or abnormal functional avidity PET scanning provides stronger evidence of tumor infiltration in these organs.

Functional nuclear-imaging studies are appropriately used in patients with HL as a diagnostic and monitoring modality. PET scanning uses uptake of the radioactive glucose analogue, FDG, as a correlate of tumor activity. PET scanning is now widely available and has become a standard part of the staging workup and assessment of response to therapy. Fused PET/CT offers the advantage of integrating functional and anatomic tumor characteristics. Residual or persistent FDG avidity appears to be useful in predicting prognosis and the need for additional therapy in posttreatment evaluation. Moreover, PET may be useful in evaluating abnormalities that become clinically manifest or appear on imaging in order to assess recurrence. PET use for follow-up is not recommended, as other reports suggest low rates of diagnosing relapsed disease and a high degree of false-positive findings. Because extranodal disease involving the bones and bone marrow is relatively uncommon in children, these staging evaluations can be omitted in patients presenting with localized and asymptomatic disease. Bone pain should be evaluated with plain radiographs; Magnetic resonance imaging (MRI) may also be necessary in this situation for questions about focal bone involvement. PET scanning seems to be supplanting the need for technetium-99 bone scans for skeletal evaluation, which historically have also been performed when there was an elevated serum alkaline-phosphatase concentration beyond that expected for age or extranodal disease identified by other staging evaluations. Although bone marrow involvement is relatively rare in pediatric HL, bilateral bone marrow biopsy had historically been recommended in any patient with clinical stage III or IV disease or B symptoms. Yet, a meta-analysis of nine clinical studies, including both pediatric and adult patients, showed that PET/CT achieved high sensitivity (96.9%) and high specificity (99.7%) in detecting bone marrow involvement in newly diagnosed patients with HL. Thus, this standard recommendation has been revised by European investigators when PET imaging shows no bone or bone marrow involvement. Because the pattern of infiltration in the bone marrow may be diffuse or focal and is often accompanied by reversible marrow fibrosis, a bone marrow aspirate alone is inadequate to assess the marrow for disease.

Prognostic Factors

The identification of prognostic factors has taken on importance as a determinant of risk- and response-adapted treatment algorithms. Many prognostic factors have been identified from adult trials. Regardless, this effort is complicated by several concepts. First, prognostic factors are in flux as more effective or higher-intensity therapy may negate adverse risk factors previously demonstrated. Second, various risk stratification schemes have been used by different institutions and cooperative groups to allocate patients to various treatment regimens, making comparisons of patient populations challenging. Most data are based on reports that primarily include adults. While there is overlap in the biology and response to therapy of pediatric and adult forms of HL, there is divergence in prognostic groupings just as there are differences in treatment regimens. Even among pediatric protocols and cooperative groups, there is a lack of standardization in prognostic factors ( Table 85.2 ). Third, there is an evolving understanding that early response to systemic therapy may be an important prognostic factor that may be used to guide further treatment decisions. While this concept is under investigation in adults with HL, one may acknowledge that response-based therapy has recently developed firm roots in pediatric HL treatment paradigms. Thus, there is much variability in classifications of Hodgkin patients into favorable, intermediate, and unfavorable/advanced strata.

Dating from the time that early-stage HL was treated with subtotal or total nodal irradiation that included splenic irradiation—often termed extended-field RT—numerous adverse prognostic factors in stage I–II disease have identified those patients who benefit from combined-modality therapy. In pediatric management, extended-field RT was abandoned in favor of primary chemotherapy and reduced-dose involved-field RT in the 1970s and 1980s in order to reduce late effects of full-dose extended-field RT. This switch in adult management of early-stage disease did not occur until the mid-1990s. Regardless, the concept of risk-adapted therapy was first developed in which the presence of poor prognostic factors drives more intensive therapy. At the same time, favorable factors identify a population appropriately treated with less intensive therapy designed to maintain high cure rates with fewer acute and late side effects. Prognostic factors identified in these analyses include the number of involved lymphoid regions, size of individual nodes, extent of mediastinal disease, patient gender and age, presence of B symptoms or pruritus, histology, ESR, and overall tumor burden as measured by number of sites and disease bulk.

Even for pediatric management, as there are efforts to develop new protocols for AYA patients, it is relevant to note that the adult cooperative groups have refined favorable versus unfavorable groupings. For instance, the European Organization for Research and Treatment of Cancer (EORTC) and Groupe d'Etudes des Lymphomes de l'Adulte (GELA) specify the following as unfavorable factors: age greater than 50 years, ESR greater than or equal to 50 in the absence of B symptoms, ESR greater than or equal to 30 with B symptoms, and greater than or equal to 4 sites of involvement or bulky mediastinal involvement. For the German Hodgkin Lymphoma Study Group (GHSG), the following are considered unfavorable factors: ESR greater than or equal to 50 in the absence of B symptoms; ESR greater than or equal to 30 with B symptoms; and greater than or equal to 3 sites of involvement, extranodal involvement, or bulky mediastinal mass.

For pediatric practice, adverse prognostic features have included advanced stage, B symptoms, extranodal extension, peripheral or mediastinal bulky disease, hilar adenopathy, and 3 or more nodal regions. Based on the most recent intermediate-risk COG protocol AHOD0031, a preliminary prognostic scoring system has been proposed, with the acronym CHIPS for Childhood Hodgkin International Prognostic Score. Based on a multivariate analysis, four predictors—stage IV, large mediastinal adenopathy, albumin less than 3.5, and fever—were identified as predictive of adverse event-free survival (EFS). This simple scoring system needs further validation, particularly as treatment algorithms change and prognostic factors may change and become more difficult to identify. For the present, several prognostic factors continue to influence the success and choice of therapy for pediatric practice:

• 1.

  Stage of disease: Stage persists as the most important prognostic variable. Patients with advanced-stage disease, especially stage IV, have a poorer outcome than patients with early-stage disease.

• 2.

  B symptoms: These constitutional symptoms likely result from cytokine secretion and correlate with biological aggressiveness. Thus, unexplained fevers, drenching night sweats, or significant weight loss with definitions noted earlier in the Ann Arbor staging classification system continue to have prognostic importance and, in turn, influence management decisions. The presence of B symptoms has correlated with a higher likelihood of systemic disease, including occult subdiaphragmatic disease when staging laparotomies were once performed. Evidence suggests that fevers and weight loss have more prognostic significance than night sweats alone. The statistical analysis that generated CHIPS for intermediate-risk Hodgkin lymphoma, however, suggested that only fevers had prognostic significance.

• 3.

  Bulk : The bulk of disease combines the number of disease sites and the volume of involvement at each site. Patients with several sites of involvement, defined variably as either three or four or more sites of disease, fare less well. This prognostic factor may also influence treatment selection as to the reduction or elimination of RT after primary chemotherapy for patients with early-stage HL. Moreover, the presence of large mediastinal adenopathy or bulky disease in nonmediastinal sites has been another consistent risk factor. A variety of definitions of large mediastinal adenopathy have been reported in the literature. The most commonly used definition is based on a measurement of the maximum width of the mediastinal mass on standing posteroanterior (PA) chest radiograph compared with the maximum intrathoracic diameter. A ratio greater than one-third is defined as “bulky.” Other reports have used a ratio with the intrathoracic width at T5–6 as the denominator, while still others use absolute measurements, surface area calculations, or volume measurements. In the absence of a chest radiograph, CT-based measurements of mediastinal masses have not reached a consensus recommendation, but some commentators have suggested that a 10-cm maximal diameter may be reasonable. Bulky disease in nonmediastinal sites has similarly been classified with varying definitions. Some protocols define bulky as greater than or equal to 10 cm, while others use greater than or equal to 5 cm, greater than or equal to 6 cm, and even greater than or equal to 7 cm. In Europe, the EuroNet Pediatric Hodgkin Lymphoma Group uses a volumetric definition of bulk. This was initially assessed by a definition of 50 mL, but subsequent experience has found the 200-mL volume definition to be more prognostically discriminatory and is now used in their EuroNet PHL C2 trial to stratify patients. Moreover, volume of a given site of disease is calculated by a simple formula estimating an ellipsoid: V = (xyz)/2, where x, y, and z are the diameters of the mass in three dimensions. This becomes much more complex when measuring mediastinal masses for which adding multiple ellipsoids of disease to avoid counting normal structures is the procedure. Moreover, FDG-PET at diagnosis allows for novel three-dimensional (3D) measures of disease burden with parameters of metabolic tumor burden and total glycolysis uptake that are proving to have prognostic importance beyond traditional anatomic definitions of bulk.

• 4.

  Laboratory studies , including the ESR, hemoglobin level, and serum albumin, have been reported to predict worse outcomes. This could reflect disease biology or bulk. A recent low-risk pediatric HL trial demonstrated that the ESR has prognostic importance. As already noted for adult cooperative group trials, the ESR has been an important stratification factor in treatment protocols. Albumin is a key component of CHIPS. Moreover, both low hemoglobin and albumin levels have been found to be important factors in an International Prognostic Score, which has been validated only for patients older than 15 years.

• 5.

  Histological subtype has importance. As already noted, patients with NLPHL are biologically different and generally have improved DFS and overall survival (OS) relative to classical HL. Separate approaches with minimal therapy have been demonstrated for early-stage NLPHL patients. Patients with lymphocyte depleted Hodgkin lymphoma (LDHL) fare poorly. Mixed reports suggest better or poorer outcome of other histologies that may be related to other prognostic factors as well. Several reports suggest that mixed-cellularity subtype of classical HL may have better prognosis than nodular sclerosis subtype in the pediatric age group.

• 6.

  Age is a significant factor, with survival rates for children with HL approaching 85% to 95% and higher than adults stage for stage. In a report from Stanford, the 5- and 10-year survival rates for children with HL less than or equal to 10 years of age was 94% and 92%, respectively, compared with 93% and 86% for adolescents (aged 11 to 16 years old) and 84% and 73% for adults, respectively. Differences in treatment approaches may play a role in these differential outcomes, as higher PFS rates have been observed in patients with intermediate-risk HL treated on pediatric versus adult protocols.

• 7.

  Rapidity of response to initial chemotherapy is an important prognostic variable. Early response to therapy was initially observed in advanced-stage HL patients treated in the Pediatric Oncology Group (POG) 8725 trial, where 93% of patients who attained a CR after 3 cycles of chemotherapy remained disease free. This was also confirmed for lower-stage patients and afterwards incorporated into the latest (COG) frontline trials. Early CR to therapy has also been successfully incorporated into the German trials with low-risk patients who achieve CR after 2 cycles of OEPA not requiring further RT. Response-based therapy is currently the paradigm on which modern pediatric trials are based in North America and Europe, tailoring treatment intensity to interim response based on PET/CT.

Risk-Adapted Treatment Approach

Contemporary treatment for children and adolescents with HL involves a risk-adapted—and now, a response-adapted—approach based on the patient's presenting features at diagnosis and reevaluation after one or two cycles of chemotherapy. Factors included in the risk assessment may vary across studies, but they most often include the presence of B symptoms, mediastinal and peripheral lymph node bulk, extranodal extension of disease to contiguous structures, number of involved nodal regions, Ann Arbor stage, and gender. A favorable clinical presentation is typically characterized as localized (stage I/II) nodal involvement in the absence of B symptoms and bulky disease. While the historical definition of mediastinal bulk has been based on the ratio greater than one-third between the transverse dimension of the mediastinal mass to the intrathoracic cavity on an upright chest radiograph, some trials have moved to using a simple size criterion on cross-sectional imaging as employed for peripheral lymph node bulk. That definition, however, is highly variable across studies, as already noted, ranging from a 4- to 10-cm diameter as a minimal threshold while European trials use a more rigorous volume classification. Moreover, there is additional subjectivity in the definition of bulk when there are multiple matted or adjacent nodes, contributing to confusion on such risk stratification in practice. Fewer than three or four involved nodal regions are considered favorable. In some risk-adapted treatment protocols, patients with localized disease presenting with unfavorable features are designated intermediate in risk and treated similarly to those with advanced-stage disease; in others, a therapy intermediate in intensity is prescribed. The criteria for unfavorable clinical presentations has also differed among studies, but most often it is comprised of the presence of B symptoms, bulky lymphadenopathy, hilar lymphadenopathy, more than 3 to 4 involved nodal regions, extranodal extension to contiguous structures, or advanced-stage (IIIB-IV) disease. The results of contemporary trials indicate that children and adolescents with early-stage or favorable presentations of HL are excellent candidates for reduced therapy. Recent trials evaluated whether intensification of therapy improves outcomes in patients with intermediate- and high-risk presentations.

Although not widely used to guide therapy assignment in pediatric trials, other factors such as gender, age at diagnosis, and histology are most definitely considered in individual patients. The trials organized by the German-Austrian Pediatric Oncology Group (GPOH) and one trial by the Children's Cancer Group (now integrated into the COG) have been unique in their aims to prospectively evaluate gender-based therapy. Long-term follow-up of the GPOH 90 and 95 studies demonstrate that the substitution of etoposide for procarbazine in the vincristine, prednisone, procarbazine, and doxorubicin (OPPA) regimen does not compromise DFS and provides less potential risk for gonadal toxicity. Although age at presentation has not been used as a criterion to assign therapy in prospective trials, reports describing outcomes after treatment with chemotherapy alone stress the benefits of this approach in younger children at higher risk of radiation-related toxicity.

Prognostic groupings now dictate the intensity of therapy with the goal of matching up just enough cytotoxic therapy to yield a high chance for cure while minimizing those exposures responsible for late complications. Within a recent cycle of COG trials from 2002 to 2011, patients were allocated into a favorable group if they had stage IA or IIA disease without bulk. Unfavorable patients were those with stage IIIB or IVB disease. The heterogenous group of patients in between were deemed intermediate risk (see Table 85.2 ). The current COG trials beginning in 2014 categorize patients into early favorable (stage I/II, no bulky disease), early unfavorable (I/IIA with bulk, IB and IIB no bulk) and high-risk strata (IIB with bulk, IIIB, IVA, and IVB). In Europe, the GPOH has merged into a geographically broader EuroNet cooperative group in which patients have been categorized by early, intermediate, or advanced treatment groups (TGs) as follows :

• 1.

  TG-1: Patients of stages IA/B and IIA without bulk (< 200 mL) and without ESR (< 30 mm/h)

• 2.

  TG-2: Patients of stages IEA/B, IIEA, IIB or IIIA and patients of stages IA/B and IIA with bulk (≥ 200 mL) and/or ESR (≥ 30 mm/h)

• 3.

  TG-3: Patients of stages IIEB, IIIEA/B, IIIB or IVA/B

Finally, a collaborative group of investigators from St. Jude, Stanford, and Harvard have defined a favorable early-stage cohort that may be treated with minimal therapy with stage I-IIA disease without mediastinal bulk, no extranodal disease, and less than 3 sites of disease.

A summary of recent trials and results in children with early- and intermediate/advanced-stage HL is provided in Tables 85.3 and 85.4 .