Lymphomas, the third most common cancer in pediatric patients (1, 2), are commonly divided into two categories: Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL). NHL includes a group of mature B lymphomas, comprising Burkitt Lymphomas (BL) and Diffuse Large B-Cell Lymphomas (DLBCL), responsible for approximately 60% of the total lymphomas in children and adolescents (2–5), with a greater prevalence in males (3). Furthermore, BL and DLBCL are characterized by strong CD20 expression (4, 6–9), and despite having aggressive behavior (6), the response to intensive short-term chemotherapy is very good, reaching a cure rate of more than 85% (1, 2, 9–12), even when patients are diagnosed in advanced stages of the disease.

This group also includes Primary Mediastinal B-cell Lymphomas (2%), which have similar aggressiveness but lower survival rates and higher frequencies in female patients (13, 14), and indolent lymphomas, which are rare in children and adolescents.

Despite the progressive advances in first-line protocols, patients with relapsed or refractory disease have an expressive reduction in overall survival (OS) (1, 4, 5, 9–11, 15), which is usually below 30% (7).

Rituximab is a chimeric monoclonal antibody that targets the CD20 lymphocyte surface antigen, which is expressed in more than 98% of high-grade mature B lymphomas (16), causing apoptosis and cytotoxicity (7, 14, 17). In clinical trials set in place thus far, rituximab in the pediatric population has been added to conventional chemotherapy and is reserved for high-grade mature B lymphomas and advanced-stage or high-risk populations, in addition to refractory/relapsed patients. Some studies already demonstrate its efficacy and safety, like those described in Cairo and Beishuizen's (18) review and Goldman et al. (19) trial. In this latter study (19), no serious adverse events were classified as probably or definitely caused by rituximab, and there was an increase in event-free survival (EFS) compared with a historic control not treated with rituximab (95% vs. 84%). Cairo (7) reported a 2-year EFS above 95% for intermediate-risk patients when rituximab was associated with chemotherapy treatment and a 2-year EFS of 90% for high-risk patients compared with the historic control. Moreover, the study of Minard-Colin et al. (20) prematurely ended because of the advantage of 13 percentage points in 1-year EFS for patients who received rituximab in contrast to treatment with chemotherapy alone.

Unfortunately, low-income countries have restrictions regarding the use of rituximab, the main reason being its substantial price. In Brazil, most of the population receives treatment through the public health system, and the allocated amount for the treatment of children with lymphomas is insufficient to add rituximab to conventional chemotherapy.

Therefore, the present study aims to evaluate the evidence that supports the use of rituximab in children/adolescents with high-grade mature B-NHL and map its application among Brazilian cancer treatment centers.

For this systematic review, our guiding question was “Does the use of rituximab in addition to standard chemotherapy improve overall and event-free survival rates in pediatric patients affected by high-grade, mature and advanced-stage B-cell non-Hodgkin lymphoma as first-line treatment or in the treatment of relapsed/refractory disease?”. Following the PICO strategy, searches were made in literature databases with selected terms joined by Boolean operators, with refinements when allowed by the database. The detailed information is available in the Supplementary Materials.

The inclusion criteria were as follows: studies with patients between 0 and 21 years of age who had mature and aggressive non-Hodgkin lymphomas of the B lineage, with CD20 expression, including BL or Burkitt-like, DLBCL or B-NHL, not otherwise specified (NOS) in advanced stages (III and IV) or refractory/relapsed, or primary mediastinal B lymphoma (PMBL) in any stage, that used rituximab combined with conventional chemotherapy, ideally (but not essentially) compared with conventional isolated chemotherapy, allowing additional treatment, such as bone marrow transplant, in both groups.

The exclusion criteria were as follows: trials that did not evaluate survival with the treatment used (with or without rituximab); trials that included patients with indolent lymphomas or immunocompromised patients; trials that used intrathecal rituximab or less than four doses applied parenterally (except, in this case, patients in second-line or later treatment); trials that included patients who had primary central nervous system lymphomas; studies without adequately described methodology; “case report” and “review” articles; studies in which biases that could compromise data extraction were detected; or studies that were available in their complete form in languages that were not English, Portuguese, Spanish or French.

Owing to the low availability of studies on this topic in the pediatric population, the present review included retro- and prospective studies, which were randomized or not randomized. Single-arm trials were included only when rituximab was used.

The project was submitted and approved by the PROSPERO platform (register number CRD42021292912).

All search, evaluation, selection, data extraction and risk of bias evaluation processes were independently conducted by two researchers. Discrepancies were evaluated by a third researcher.

Statistical analysis and forest plot creation were performed via R versions 4.0.2 and 4.3.0 (21). The meta-analysis was conducted with the “meta” package version 7.0-0 (22); odds ratio calculations with 95% confidence intervals (CI) were performed via a fixed effect model, and hazard ratio calculations with 95% CI were performed via common and random effects models.

The risk of bias was determined via the Cochrane network's, Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) (23) and Risk of Bias 2 (RoB 2.0) (24) tools. RoB 2.0 was used in randomized trials. In domain 2 of the evaluation, we considered the effect of attribution to the intervention. To evaluate non-randomized single-arm trials, the ROBINS-I tool was adapted without considering the third criterion (“bias in the classification of interventions”), with the remaining criteria normally examined.

Furthermore, an online questionnaire on the use of rituximab was offered to pediatric oncologists affiliated with the Brazilian Society of Pediatric Oncology (SOBOPE). The purpose of this research was to evaluate patients' access to the medication in question. The questions referred to current practices in the treatment center where the pediatrician was practicing, and all the data were analyzed on a treatment center basis. This research was approved by the IRB board of our institution (CAAE 48264421.3.0000.5580).

The 17 selected trials (Figure 1) were separated into three groups for adequate evaluation: patients receiving first-line therapy with BL, Burkitt-like, DLBCL and mature B-NHL, NOS; patients receiving first-line treatment with PMBL; and patients with refractory or relapsed disease receiving second-line or later treatment.

Our review and meta-analysis showed evidence of the benefits of using rituximab for treating children and adolescents with advanced-stage or relapsed/refractory high-grade mature B-NHL, consistent with the outcome of the only phase III study available for this purpose until now, in advanced-stage patients on first-line treatment [Minard-Colin et al. (12), included in this meta-analysis]. There is a great difficulty in conducting randomized prospective trials in patients with pediatric lymphomas, owing to the lower prevalence than in adult patients. Furthermore, there were several difficulties in paper selection, starting with the lack of standardization regarding the age of this population: there are trials that include, in pediatric studies, patients between 15 and 39 years, the AYA population. In this last circumstance, some of the articles that were found had to be excluded from the present review due to the impossibility of separating the data of the population that is our interest (patients up to 21 years old).

Another difficulty, which is a potential cause of bias, is the use of many different protocols for the treatment of the same pathology, several of which suffer from modifications from its original form that are not always described for the reader's assessment. Furthermore, the lack of stratification regarding the risk of patients evaluated in most trials caused us to alter our inclusion criteria for patients with advanced-stage disease, instead of those with high-risk disease, as was the original intention.

Additionally, some trials compared the use of rituximab with a historical control from the same research group that did not include this agent. Here, a potential bias can be identified, since protocols undergo modifications over time, and overall support for these patients tends to improve in this interval, which privileges the arm treated more recently.

The present review also aimed to evaluate whether the inclusion of rituximab in the treatment protocol increased the risk of serious infections or toxic death, but that was not possible because of the lack of these data in most of the included trials. Fu et al. (28) described early death in each group, or 2.2% in the group that used rituximab, in contrast with 4.3% in the group without rituximab. Minard-Colin et al. (12) reported a greater number of infections of Grade 4 or higher in the group treated with rituximab (18.5% vs. 11.1%), with a p-value of 0.07, but the rate of toxic deaths was 1.8%, with an equal distribution among both groups. Dourthe et al. (36) did not report any death by toxicity in either group in the trial. In the remaining double-armed clinical trials, when there was a description of these data, the description was made jointly.

Some questions are still lacking when we discuss the treatment of pediatric mature B lymphomas, especially in the group of refractory/relapsed patients. Overall survival in this group of patients improves when consolidated with bone marrow transplant after immunochemotherapy treatment [Anoop et al. (38); Jourdain et al. (26)], especially when a full response is reached. Jourdain et al. (26) reported an OS of 54.3% in patients who underwent transplantation in these circumstances vs. 0%, p = 0.016. However, there are still questions about the best type of transplant that can be performed: autologous (the most commonly used type) or allogeneic? Furthermore, Anoop et al. (38) reported better OS in relapsed or refractory patients who received at least 4 doses of rituximab than in those who received two or more, similar to patients in first-line treatment: more trials are needed to evaluate whether this difference is confirmed. Finally, would it be beneficial if rituximab was used after bone marrow transplantation as a maintenance treatment?

Despite the ethical conflicts that this type of research can cause, the trial of Samochatova et al. (27), which showed good response in the treatment of patients undergoing first-line treatment with reduced doses of methotrexate when rituximab was used, may pave the way for other studies that have this purpose, with the intent of reducing late adverse effects, which are very harmful for survivors of pediatric neoplasms. As another example, we can cite the study of Goldman et al. (42), in which the intervention group demonstrated superiority with the inclusion of rituximab and reduction of the dose of anthracyclines in patients stratified as FAB Group B with stage I/II and stage III with lactic acid dehydrogenase <2X upper normal limits. Finally, regarding the use of rituximab in the Brazilian pediatric oncology and hematology institutions interviewed, we can observe that most of the institutions rely on their own financing or on the financing provided by nongovernmental organizations for the use of the medication with their patients. Notably, many patients have to file lawsuits to obtain this therapeutic agent. These alternatives to offer the best treatment for patients are praiseworthy, but we know that these financing alternatives may encounter setbacks that generate delays or failures in their administration and, consequently, jeopardize treatment. Our meta-analysis revealed a clear improvement in patient survival with the addition of rituximab to chemotherapy regimens in the population studied, and we believe that financing for the acquisition of rituximab starts being carried out in the same way as that for conventional chemotherapy drugs.

In conclusion, the results of our systematic review demonstrate that, in the vast majority of the included trials, either on first-line therapy or later, survival (event-free or overall) is greater when rituximab is employed in association with chemotherapy. Nevertheless, many of the trials had nonsignificant p values for these comparisons, probably due to the small sample sizes included.

Our meta-analysis revealed that there is a tangible benefit in the use of rituximab in the evaluation of events and deaths occurring in patients receiving first-line treatment (with the exception of PMBL). Furthermore, this benefit remains when evaluating the OS of patients with relapsed and refractory disease who use this medication.