Anaplastic large cell lymphoma (ALCL) represents approximately 15% of all pediatric non-Hodgkin lymphomas (NHL). It has distinct clinical features, including frequent involvement of extranodal sites and rare localization to the central nervous system (CNS). Despite varying treatment approaches the outcome of patients with ALCL has not significantly improved during the last two decades.
From October 1997 to beginning of 2000, newly diagnosed ALCL patients were enrolled into AIEOP LNH-97 protocol for ALCL. Thereafter and until 2007, only CNS positive patients were included. AIEOP LNH-97 was based on the BFM-95 schema for ALCL and included six high-dose chemotherapy courses. CNS prophylaxis was obtained with one intrathecal injection of chemotherapy in each course, whereas treatment of CNS involvement included three intrathecal injections without irradiation.
Thirty-two patients were eligible for the study. Lymph-node disease was the most frequent localization (69% of the cases), followed by mediastinal (25%), CNS (22%), bone marrow (16%), and skin (13%) involvement. Probabilities of overall survival (OS) and of event-free survival (EFS) at 5 years for the whole population were 87% (SE 6%) and 68% (SE 8%), respectively.
Anaplastic large cell lymphoma (ALCL) represents approximately 15% of all pediatric non-Hodgkin lymphomas (NHL) 1. First described by Stein et al. 2, ALCL is characterized by proliferation of anaplastic cells of T or null phenotype with CD30 antigen expression. ALCL exhibits a broad morphological spectrum, including common, lymphohistiocytic, and small cell variants 3. The t(2;5)(p23;q35) translocation, resulting in the fusion of the nucleophosmin gene (NPM) at 5q35 and the tyrosine kinase gene ALK at 2p23, is identified in more than 90% of pediatric ALCL 4. Expression of the fusion transcript NPM-ALK can be detected by reverse transcriptase polymerase chain reaction (RT-PCR) and represents a tumor-specific marker that can be exploited to evaluate minimal disseminated disease (MDD) in the bone marrow (BM). MDD at diagnosis is frequently detected in pediatric ALCL and can identify patients at risk of relapse 5. The most common clinical features of ALCL in the pediatric population include systemic symptoms, especially high fever, generalized lymph-adenopathy and involvement of extranodal sites such as skin, bone, soft tissue, lung, and liver 6–13. The central nervous system (CNS) is rarely involved 14. In the literature, few cases of primary CNS ALCL have been described, but in those cases the tumor-associated mortality seemed to be lower in patients with ALK-positive ALCL and young age 15. The secondary spread of primary nodal or systemic ALCL to the CNS is a rare event 7–12, 16–18.
Optimal treatment for ALCL is not yet ultimately defined. Most European pediatric oncology groups have used short-pulse chemotherapy regimens, including high-dose methotrexate (MTX), cyclophosphamide, vincristine, doxorubicin, and corticosteroids with a total duration of 4–6 months 7–9, 16–18. In the early nineties, like in North America, our group treated patients with ALCL with prolonged, repeated-pulse chemotherapy, based on a modified LSA-L2 protocol 10–12. The results of the randomized clinical trial ALCL99 showed that a NHL-B-like therapy, based on six cycles of chemotherapy for 5 days each with MTX 3 g/m2 in 4-hour infusion, showed the same efficacy when MTX was given at 1 g/m2 in 24-hour infusion 19. This study also showed that intrathecal prophylaxis is not necessary to prevent disease recurrence to the CNS. Adding vinblastine to the same therapy backbone significantly delayed the occurrence of relapse but did not reduce the risk of failure 20. The event-free survival (EFS) rate ranges from 60% to 75% for most of the studies and recurrence is observed in up to 35% of patients 7–12, 16–19. Patients who suffer a relapse can be successfully treated with second line therapy, ranging from weekly vinblastine alone 21 to intensified therapy with hematopoietic stem cell transplantation (HSCT) 17, 22–25, resulting in high overall survival (OS) rate. Here, we report the characteristics and long-term outcome of ALCL patients treated in the Italian Association of Pediatric Hematology and Oncology (AIEOP) centers using AIEOP LNH-97 protocol.
PATIENTS AND METHODS
From October 1997 to beginning of 2000, newly diagnosed pediatric patients with ALCL were registered onto AIEOP LNH-97 protocol for ALCL. From the year 2000, only patients with CNS positivity not eligible for the newly established international randomized ALCL-99 trial, were enrolled. Patients were treated in 17 Italian centers. Local ethics committee approval and informed consent were obtained prior to patient enrolment. Exclusion criteria were: age >18 years, previous cytostatic treatment, BM involvement >25%, inclusion in any other study, NHL as a secondary malignancy, pre-existing severe combined immunodeficiency or acquired immunodeficiency syndrome and any other condition prohibiting protocol chemotherapy.
Diagnosis and Classification
Diagnosis was established by histological and immunohistochemical analyses of lymph-node or tumor mass biopsy and/or by cytological and immunologic examinations of malignant effusions. Cases were classified according to the updated Kiel classification and the revised European-American lymphoma classification 26, 27. Central histopathological review of diagnosis was performed by two independent pathologists in all cases with available material. Formalin-fixed paraffin-embedded tumor biopsies from all ALCL cases were analyzed by immunohistochemistry using a wide panel of antibodies including T- and B-lineage markers (CD2, CD3, CD4, CD5, CD7, CD8, CD20, CD43, CD45RO, CD79a), NK markers (CD56, CD57), CD30, ALK1, EMA, clusterin, and cytotoxic protein (Tia1, Granzyme B, and Perforin). Centralized investigation of ALK gene rearrangements was performed in all cases with fresh tissue available.
Staging and Stratification of Treatment
Disease stage was defined according to the Murphy modified system 28. All patients underwent physical examination, full blood count and biochemical profile, including serum lactate dehydrogenase (LDH), chest X-ray, BM aspirate and cerebrospinal fluid (CSF) cytospin examination, ultrasonography, computed tomography (CT) scan, and/or magnetic resonance imaging (MRI), skeletal scintigraphy and skin biopsy of cutaneous lesions. CNS involvement was defined as the presence of any number of morphologically identifiable lymphoma cell in the CSF or/and demonstration of a cerebral mass by cranial CT scan or MRI, or presence of cranial nerve palsy not caused by an extracranial mass. BM infiltration was considered positive when tumor cells, independently of the percentage, were morphologically detectable. Patients were stratified according to clinical stage and histological subtype into three therapeutic risk groups (R1, R2, and R3) as detailed in Figure 1.
Treatment and Response Criteria
Therapy plan and drug combinations are detailed in Figure 1 and Supplemental Table I. The AIEOP LNH-97 protocol was based on the BFM-95 study 17. All patients received a 5-day prephase followed by three to six 5-day chemotherapy courses, depending on patient risk group. CNS prophylaxis consisted of triple intrathecal (TIT) injections of MTX/cytarabine/prednisolone, one in the pre-phase and one in each course. In case of CNS involvement, additional intrathecal therapy was administered for a total of 19 doses. No cranial irradiation was given. Hematologic requirements for starting each course were platelets >50,000/mm3 and neutrophils >500/mm3. The first course started just after the prephase, whereas the minimum interval between two consecutive courses was 9 days. Tumor response was evaluated after each course of therapy. Complete remission (CR) was defined as absence of any residual mass and no blast in the BM and in the CSF. Second-look surgery at the end of therapy was performed only in case of a doubtful residual mass.
OS was calculated from diagnosis to death from any cause or to the last follow-up examination. EFS was calculated from the date of diagnosis to the first event (relapse, tumor progression after an incompletely resolved tumor, death from any cause, second malignancy) or to the most recent follow-up examination. Patients lost to follow up were censored at the time of their withdrawal. Survival probabilities were determined by using the Kaplan–Meier method with differences compared by the log-rank test 29. Cox model analysis was planned for multivariate analysis considering variables with a log rank P-value < 0.1 30. Statistical analysis was carried out by using the SAS statistical program (SAS-PC, version 9.1, SAS Institute, Inc., Cary, NC).
Thirty-seven patients were enrolled in the ALCL LNH-97 protocol. Five patients were excluded from analysis due to pre-treatment (n = 2), incorrect diagnosis at histological revision (n = 2), and >25% blasts in the BM (n = 1). Of the 32 evaluable patients, six with CNS involvement were enrolled after the year 2000.
Local histologic diagnosis was centrally reviewed in 88% of the cases. The distribution of subtypes according to the histological classification was: common type (n = 16), lymphohistiocytic (n = 4), mixed cell (n = 6), small cell type (n = 2), ALCL not further specified (n = 4). All patients were positive for CD30; 30 were studied for ALK expression (28 positive, 2 negative) and 22 for EMA (19 positive, 3 negative). The t(2;5)(p23;q35) translocation on tissue biopsy was determined by RT-PCR in 22/32 patients and was positive in 19. The majority of the cases showed a T-cell immunophenotype (n = 24), whereas eight patients were Null type.
Among the 32 eligible patients, 20 (63%) were male and 12 (37%) were female (M/F ratio 1.7:1). The median age at diagnosis was 10 years (range, 0.6–17.6). Eight patients had stage II, 14 stage III, and 10 stage IV disease (three of them with BM, five with CNS, and two with combined BM and CNS involvement). Fifteen children (47%) were included in the R2 group and 17 (53%) in the R3 group. The median serum LDH value was 500 IU/L (range, 237–5,400 IU/L). Single organ involvement was found in 14 patients: lymph-node in seven cases, mediastinum in four cases, abdomen in two cases and bone in one case. The remaining 18 patients had tumor presentation at multiple sites: 15 lymph-node combined with CNS (n = 4), BM (n = 3), mediastinum (n = 3), skin (n = 3), or lung (n = 2); 3 CNS combined with skin (n = 1) or BM (n = 1) or mediastinum and BM (n = 1). Lymph-node involvement was the most frequent localization, accounting for 69% of the cases, whereas mediastinal involvement was confirmed in 25% of the cases, followed by CNS (22%), BM (16%) and skin (13%) involvement. Among the seven CNS positive patients, four had blast positive CSF (one case with concomitant intra-cerebral mass), three had CNS mass; none presented with isolated CNS disease.
As summarized in Table I, CR was achieved in 30 of 32 patients (94%), after a median time from diagnosis of 2.2 months (range, 0.24–6 months). Two patients did not achieve CR: one because of early death (i.e., at the end of the prephase); one because of disease progression during the second therapy course, which caused death 52 days from diagnosis. Eight patients (25%) relapsed at a median time of 6 months from CR (range, 1.6–17.6 months). Three had an isolated lymph-node relapse, two an isolated CNS relapse (both in CNS positive patients at diagnosis), while three had combined relapses: mediastinum/abdomen (n = 1), mediastinum/skin/lymph-nodes (n = 1), and lung/lymph-nodes (n = 1).
Two of eight relapsed patients died of disease progression after second or subsequent line treatments, including autologous HSCT in one. At the time of last follow-up, the six remaining relapsed patients were alive in CR after second line chemotherapy alone (n = 4), or combined with allogeneic (n = 1) or autologous HSCT (n = 1). One patient had a secondary acute myeloid leukemia at 5.4 years from diagnosis and died after unrelated HSCT. Details of events are listed in Table II.
Table II. Events in Patients With ALCL
Tumor site at diagnosis
Type of event
Timing of event (days from diagnosis)
Site of relapse
Outcome/causes of death
Lymp, lymph-nodes; n.a., not applicable; CNS, central nervous system; BM, bone marrow; PD, progression of disease; CR, complete remission; HSCT, hematopoietic stem cell transplantation; Med, mediastinum.
Patients who did not achieved CR.
Patients who underwent allogeneic-HSCT as second line therapy.
Patient who underwent autologous-HSCT with additional cranial radiotherapy as second line therapy.
One patient was taken off-study after the second chemotherapy course while in CR due to parental decision. With a median follow-up of 10 years (range, 0.01–13 years), 26 out of the 32 evaluable patients (81%) are alive, including the off-study case. The probability of OS and of EFS at 5 years for the whole population were 87% [standard error (SE) 6%] and 68% (SE 8%), respectively (Fig. 2). Univariate analysis for the different parameters including gender, age, stage, risk group, immunology, histological subtype, tumor sites, and LDH value at diagnosis demonstrated no statistically significant impact on EFS. None of the two parameters (age and LDH) entered in the multivariate analysis demonstrated a significant impact on EFS (Table III).
Table III. Univariate and Multivariate Analysis for Different Risk Factors
5-year EFS % (SE)
Data regarding WHO grade 3–4 toxicity were available for all of the patients. The number of hematologic toxicity episodes was 33 in 15 R2 patients and 38 in 17 R3 patients, being neutropenia the most frequently reported toxicity. Overall, 21 episodes of mucositis and 13 episodes of hepatic toxicity were reported. One toxic death due to Escherichia coli sepsis occurred during prephase in a high risk patient, whereas other 12 not life-threatening sepsis occurred in nine patients, caused by Staphylococcus aureus (n = 4), Klebsiella pneumonia (n = 2), Escherichia coli (n = 2), Pseudomonas aeruginosa (n = 2), Capnocytophaga spp (n = 1), and Ochrocaterium anthropi (n = 1).
Patients with ALCL diagnosed in Italy in the early 1990s were treated with the AIEOP LNH-92 protocol, based on a modified LSA2-L2 strategy. Since 1997 children with ALCL, including CNS positive cases, were treated with the shorter BFM based protocol AIEOP LNH-97 11, 16, until the international trial ALCL99 was launched. Thereafter, only patients with positive CNS disease and who were not included in the randomized ALCL99 trial, were enrolled in the AIEOP LNH-97 protocol, thus explaining the high prevalence of CNS involvement in our cohort of patients.
The change of therapeutic strategy from a leukemia-like treatment, such as the AIEOP LNH-92 protocol 11, to the B-cell NHL based protocol AIEOP LNH-97, did not result in a better prognosis for children with ALCL (5-year EFS: 65% vs. 68% in LNH-92 and LNH-97, respectively). However, when comparing results of these two protocols, we should consider the higher percentage of CNS positive patients (22% vs. 0%) and of the lymphohistiocytic variant (13% vs. 3%) in the LNH-97 protocol compared to the LNH-92 protocol, both characteristics being correlated to unfavorable prognosis 15, 31. The AIEOP LNH-97 protocol used higher doses of chemotherapy compared to the ALCL99 protocol, but achieved comparable outcome (5-year EFS 68% vs. 2-year EFS 71%). This confirms the results of the randomized use of two different doses and schedules of high-dose MTX and of the addition of vinblastine, evaluated prospectively in the ALCL-99 trial, which failed to demonstrate any significant difference in terms of efficacy 19, 20. These observations raise the question of whether dose intensity, or dose of specific agents, such as MTX, may play a role in ALCL therapy or whether other characteristics in the treatment strategy, such as continuous low-dose chemotherapy, might be most important for the overall outcome. At this regard it would be challenging to pursue further the analysis of risk factors in ALCL, including histological subtypes, minimal disease and anti-ALK antibody titer, as we and others have demonstrated that they can identify distinct risk categories among pediatric ALK-positive ALCL 32, 33. Because the cumulative dose of chemotherapy and the total duration of treatment were lower in the LNH-97 protocol compared with the LNH-92 protocol, our results indicate that short pulse-based chemotherapy is not inferior to long leukemia-like therapies.
Univariate and multivariate analyses showed a trend for a worse outcome in patients with elevated LDH levels. The negative prognostic impact of high LDH levels was previously suggested by the French study LMB-89 7, 13. Our findings would support that observation but, possibly due to the relatively small cohort of patients, none of the common prognostic parameters analyzed showed an impact on prognosis, including CNS involvement. In this regard, it is worth noting that two of the three patients who died due to progressive disease had advanced stage ALCL with CNS involvement. This would suggest that CNS involvement might impact negatively on survival, although it does not significantly influence the EFS of pediatric patients with ALCL.
Results of the AIEOP LNH-97 protocol indicate that short pulse chemotherapy can achieve similar outcome compared with longer therapies, such as LSA2-L2 based regimens, in patients with ALCL, including patients with CNS involvement and with unfavorable histology. However, because there is no definitive demonstration that high-dose intensity based therapies be advantageous in ALCL compared with less intensive treatments, development of alternative therapeutic strategies is warranted. Refinement and optimization of therapy for ALCL may originate from a combination of clinical and biological prospective studies, as those in the pipeline of current international collaboration. Moreover, the early results of clinical use of targeted therapy, including conjugated anti-CD30 antibody and ALK inhibitors 34, 35, give new perspectives for a more effective treatment of high-risk ALCL patients.
We are indebted to Dr Elisa Carraro for her careful editing of the manuscript.