Long-term results with the adapted LMB 96 protocol in children with B-cell non Hodgkin lymphoma treated in Iraq: comparison in two subsequent cohorts of patients.

Since 2000, an adapted LMB 96 protocol was implemented at the Children-Welfare-Teaching-Hospital in Baghdad for the treatment of childhood B-cell non-Hodgkin lymphoma. The first experience (2000–2005) demonstrated efficacy and feasibility of this protocol in Iraq. In 2006, further adjustments were made in an attempt to reduce therapy-related toxicities. The outcome of the second cohort of 190 children (2006–2010) and the comparison with the previous study are hereby reported. Out of the 180 treated patients, 120 achieved a complete response; during treatment 51 died and 9 abandoned. The 60-month overall survival (OS) and event-free survival (EFS) were 64.7 and 56.3%, respectively. No differences were observed in the 24-month OS and EFS between the 2000–2005 and 2006–2010 cohorts (66.3% vs. 65.1%; p =.89 and 53.3% vs. 57.3%; p =.28, respectively). Therapeutic group-B in the second cohort showed better outcome, although not significant, compared to the first one (EFS 62.9% vs. 53.8%; p =.088). Therapy-related mortality remained high.

Keywords: B-cell non-Hodgkin lymphoma; children; low-income country

In high-income countries, pediatric treatment protocols for B-cell non-Hodgkin Lymphoma (NHL) lead to 5-year overall survival (OS) rates approaching 90% [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10]]. Although there is limited information on therapies and outcome from low-middle income countries, similar excellent results are not achieved and survival for children with NHL is in the range of 60–70%, falling to less than 30% in African countries where a public health service is lacking [[11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]]. The outcome differences between high and low-middle income countries can be ascribed to a combination of suboptimal therapy, lack of access to cancer care facilities, delay in diagnosis and in the initiation of treatment, therapy-related toxicity, inaccurate diagnosis together with inherent NHL biologic differences and treatment abandonment.

Within the framework of a Telemedicine Project, developed over the last 15 years between our Hematology Center in Rome and the Children Welfare Teaching Hospital (CWTH) in Baghdad, a reference pediatric center in Iraq, it was possible to review the biologic/histologic diagnostic work-up of patients, set up a prospective clinical registry, and design and conduct clinical trials for the treatment of pediatric onco-hematologic diseases tailored according to the local reality. A follow-up of patients was also carried out, in order to address the obstacles and help in the implementation of the programs.

In January 2000, an LMB 96-derived protocol, adapted to the local situation, was implemented at the CWTH for the treatment of children with B-cell NHL. The results of the first 5-year experience (2000–2005), which included 239 children, have been previously published [[11]] and demonstrated the feasibility and efficacy of this treatment in Iraq. At 24 months, the OS and event-free survival (EFS) were 66.3 and 53.8%, respectively. However, treatment-related mortality (TRM) and abandonment rate were very high (29 and 12%, respectively). In 2006, a prospective new protocol was designed with further adjustments in an attempt to reduce induction therapy-related toxicities and mortality. These mainly consisted in the introduction of a steroid pre-phase for patients with poor performance status (PS), fractionation of the cyclophosphamide daily dose for children with bulky tumor and/or poor PS, the addition of a second pre-phase with COP (cyclophosphamide, vincristine, and prednisone) after induction in case of a critical clinical status and the omission of doxorubicin in the first induction course, for patients who had normal LDH levels at diagnosis and achieved a complete remission (CR) after COP.

We report hereby the Iraqi experience treating children with B-cell NHL with the two consecutive LMB 96-derived protocols. In particular, the results in the second cohort of 190 Iraqi children diagnosed and treated between 2006 and 2010 are described and compared with those obtained in the previous study.

Children aged less than 14 years and diagnosed with mature B-cell NHL entered the study. All patients were treated according to the LMB 96 adapted Iraqi protocol [[11]] at the Pediatric Oncology Unit of CWTH in Baghdad. Data regarding age, gender, clinical presentation, diagnostic procedures, staging, treatment, complications, and outcome were prospectively collected. Data were analyzed at the GIMEMA (Gruppo Italiano Malattie EMatologiche dell'Adulto) Data Center in Rome.

Diagnosis was made by incisional/excisional biopsy, cytological examination of cerebrospinal fluid (CSF) or effusions and morphologic examination of bone marrow (BM) aspirate/biopsy. Histochemical stains, immunophenotype, and karyotype were not available. However, from 2005 the histologic diagnosis was centrally reviewed and confirmed at the Department of Experimental Medicine/Pathology of the Sapienza University of Rome. The histopathologic classification was carried out according to the International Working Formulation (NCI, 1982) [[24]]. The St. Jude staging system was employed [[25]]. Pretreatment evaluation included the patients' history, physical examination, complete blood count, serum electrolytes, liver and kidney profile, incisional/excisional tumor biopsy, BM aspirate/biopsy, CSF analysis, and radiologic investigations (chest X-ray, abdominal sonography, CT scan, or MRI in selected cases). Serum LDH levels were determined when available; gallium scan and bone scan were not performed.

The protocol received official institutional review board approval. Oral informed consent was obtained from the parents for all children under study. Patients were classified according to the LMB 96 protocol risk classification in the following groups: Group A (low risk): completely resected stage I and abdominal stage II; Group B (intermediate risk): unresected stages I and II, non-abdominal completely resected stages II and III; and Group C (high risk): BM and/or CNS involvement. The adapted protocol schedule was applied as follows: patients assigned to Group A received a pre-phase with cyclophosphamide, vincristine and prednisone (COP), and two induction courses of cyclophosphamide, vincristine, prednisone and doxorubicin (COPAD) without intrathecal therapy (IT). Patients in Group B received the COP pre-phase, two induction courses with cyclophosphamide, vincristine, prednisone, doxorubicin, methotrexate (COPADM), IT therapy and two consolidation courses with cytarabine and methotrexate (CYM). Therapy for Group C included the COP pre-phase, two induction COPADM courses, two consolidation courses with cytarabine and etoposide (mini–CYVE) and four maintenance cycles (M1, M2, M3, and M4), the details of which have been previously reported [[11]].

From January 2006, the protocol has been further adjusted by introducing a steroid pre-phase for patients with a poor PS, fractionated cyclophosphamide (50 mg/m2/dose every 12 hours instead of 300 mg/m2/single daily dose) for patients with bulky tumor and/or poor PS, and a second COP course in case of critical conditions in an attempt to reduce the induction toxicity. For patients who had normal LDH levels at diagnosis and achieved a CR after COP, doxorubicin was omitted in the first course of COPADM. Induction courses (COPAD or COPADM) were initiated at day 8 after the start of COP. The second induction and the consolidation courses started on day 21 and the second mini-CYVE consolidation cycle started 1 week after the methotrexate infusion if the absolute neutrophil count (ANC) was >1.0 × 109/l and the platelet count >100 × 109/l. Maintenance courses were administered at 28 days interval.

All patients received hydration (3000 ml/m2/day) and allopurinol (300 mg/m2/day), 24 hours before chemotherapy, during and up to 24 hours after cyclophosphamide infusion; hydration with alkalinization (3000 ml/m2/day with NaHCO3 50 mmol/l and KCl 10 ml/l) started 6 hours before the methotrexate infusion and continued for 48 hours. Folic acid was administered according to the protocol schedule. Urate-oxidase and methotrexate plasma level determination were not available.

No prophylactic antibiotics were given during neutropenia; G-CSF in combination with available broad-spectrum antibiotics was administered to patients who developed fever when the ANC count was <0.5 × 109/l and continued up to ANC recovery (ANC >1.0 × 109/l). Blood components were restricted to patients suffering from severe anemia (Hb <7g/dl), thrombocytopenia (Plts <20 × 109/l) and/or hemorrhages.

Treatment responses were defined as follows: CR, the complete disappearance of all measurable lesions, no blasts in the BM or CSF; partial response (PR), a greater than 20% reduction of the tumor mass and/or blasts in the BM or CSF; no response (NR), less than 20% reduction of the tumor mass and/or blasts in the BM or CSF. Disease relapse was defined as tumor re-growth after initial shrinkage.

Responses were evaluated after the COP course and after completion of each therapy phase; it consisted of clinical examination, chest X-ray, abdominal sonography, BM aspirate/biopsy, CSF examination, according to the initial presentation. Patients classified as group B with no response (less than 20%) after COP or those who still showed a residual mass after the second COPADM course were shifted to protocol C. Follow-up investigations were planned after therapy completion, every 2 months during the first year, every 3 months during the second year and every 4–6 months up to fifth year.

A descriptive analysis of the demographic and clinical characteristics of the patients was performed including median and range for continuous variables, and absolute and relative frequencies for categorical variables. Nonparametric tests were used to evaluate differences among groups (Fisher exact test and Wilcoxon test for categorical and continuous variables, respectively). OS was defined as the time from diagnosis to death or the date of the last follow-up. EFS was defined as the time from diagnosis to the date of failure (no CR, treatment abandonment, relapse, death) or the date of the last follow-up. Disease-free survival (DFS) was calculated from the time of achieving CR to relapse, death or the date of the last follow-up. Survival curves (OS, DFS, and EFS) were estimated according to the Kaplan-Meier product-limit method and were tested for significant differences using the log-rank test in univariate analysis and by means of the Cox regression model in multivariate analysis. In all analyses, 95% confidence intervals (CI) were reported for the main summary statistics and all statistical comparisons were based on two-tailed tests accepting p ≤ .05 as statistically significant. All analyses were performed using the SAS system software (version 9.4) and R (http://www.R-project.org) [[26], [27], [28], [29]].

Between January 2006 and December 2010, 190 B-cell NHL patients (age <14 years) were registered. Patients' characteristics are shown in Table 1 and compared with those of the 239 previously reported children, diagnosed and treated between 2000 and 2005 [[11]]. Briefly, there was a male preponderance (132 male; 58 female); median age was 4.5 years (range 1.34–13.55); 75 children (39.5%) lived in Baghdad, 115 (60.5%) outside Baghdad; the median duration of symptoms prior to diagnosis was 4 weeks (range 1–24); ≤6 weeks in 107 children (56.6%) and >6 weeks in 82 patients (43.4%). Seventy-four children presented with fever ≥38 °C; 35 (18.4%) were underweight, below the fifth percentile at diagnosis, and 61 (32.1%) were in very poor clinical status (WHO PS ≥3). Anemia (Hb ≤8 g/dl) was present in 15 children (8%). LDH, tested in 131 children, proved greater than 500 IU/L in 63 (33.2%).

Table 1. Diagnostic characteristics of the two cohorts of patients.

1 *In many patients more than one site was involved at diagnosis.

At diagnosis, the abdomen was the more involved site (161 children, 85%); maxillofacial masses (oropharynx, jaw, retro-orbital, and nasal) were identified in 31 children (16%). A peripheral node enlargement was present in 47 children (25%). BM was involved in 20 cases (10.5%); 22 children (11.6%) had CNS disease. In 27 (14.2%), the CSF evaluation at disease onset was not performed, but those with cranial nerve palsy were considered as CNS positive. One hundred and twenty-eight children (67.4%) had bulky disease (tumor mass diameter ≥10 cm).

The diagnosis was based on a tumor biopsy in 110 cases (58%), fine needle aspirate in 64 cases (34%), cytologic examination of ascitic fluid in 11 (6%) and BM aspirate/biopsy in 4 cases (2%). According to the International Working Formulation, 110 children (57.9%) were classified as small noncleaved cell lymphoma (SNCL), 10 (5.2%) as large cell lymphoma (LCL) and 68 (35.8%) as nonspecified high-grade lymphoma (HG NHL). Histology was not available in 2 cases (1%).

Most patients presented with advanced stage disease (stages III and IV; 126 and 37, 86%).

The two cohorts of patients showed similar clinical and biologic characteristics in terms of gender, stage, BM, and CNS involvement, histologic subtype and duration of symptoms prior to diagnosis; significant differences were found in median age that was lower in the second cohort of patients (p < .001) and nutritional status resulted below the fifth percentile in 36 and 18% (p < .0001) of children in the first and second cohort, respectively (Table 1). The abdomen was the more frequent involved site at disease presentation in both cohorts (85% and 89%; p = .145); jaw disease was present in 25% of children of the present study compared with 18% of the previous one (p = .0003).

Six patients (3.2%) died before starting treatment and 4 (2%) were not treated (refusal). According to the treatment protocol, 180 patients were assigned to the following risk groups: 10 (5.2%) group A, 116 (61.6%) group B, and 54 (28.4%) group C. Sixty-nine children (38%), with a poor PS, received the steroid pre-phase. Three Group A (completely resected) children did not undergo the COP pre-phase. In 25 cases, a second COP cycle was administered due to the critical clinical conditions. In 58 patients (32.4%), treatment modifications had to be made: omission or dose reduction of one or more scheduled drugs (methotrexate, vincristine, cyclophosphamide, doxorubicin) in seven patients, due to shortage of chemotherapy; omission or dose reduction in 24 children, due to excessive toxicity and, finally, doxorubicin was omitted in the first COPADM course in 24 patients, most of therapeutic Group B, who had normal LDH levels at diagnosis and had achieved a CR after COP.

Of the 180 treated patients, 177 received the COP pre-phase which resulted in a CR for 40 patients. The overall CR rate after induction (COP/COPAD/COPADM) was 68% (122 children); a PR was observed in 3 (2%) patients; 6 (3%) were NR and 2 were lost to follow-up, prior to response evaluation. Therapy-related deaths occurred during induction in 47 (26%) patients (21 with COP1/2 pre-phase). These results are comparable with those obtained in the previous study in which a CR rate of 56% was recorded; the difference is not statistically significant.

Nine patients abandoned therapy at a median time of 2 months from diagnosis (range 1–4 months; 5 after CYM, 3 after COPADM and 1 after COP) and 4 cases never responded to treatment. Four children, in CR, died (2 miniCYVE1 and 2 maintenance M1). Eleven children relapsed. Relapses occurred at a median time of 4 months (range 2–8); during maintenance in 4 and after therapy discontinuation in 7 patients. The primary site involved at relapses was the CNS (7/11 children), isolated in 5 and combined with an abdominal and jaw mass in the other 2.

In the previous cohort of patients, 69 died due to therapy-related complications and 7 resulted nonresponder. A higher number of patients (29) abandoned treatment; relapse occurred in a similar number of cases (10 patients), mostly after therapy completion.

In the 180 treated children, severe toxicities occurred mainly during/after the first cycles of the LMB96 protocol (COP/COPAD/COPADM) (Table 2(A)). Tumor lysis syndrome (TLS) was observed in 25 patients (fatal in 21) during the COP pre-phase, with hyperuricemia, electrolyte imbalance, and acute renal failure. Severe infections were recorded during or after COP in 21 children, COPADM1 in 103 (fatal in 20) and COPADM2 in 73 (fatal in 2). Severe mucositis and gastro-intestinal (GI) toxicity (WHO ≥ 3) occurred mainly during COPADM1/COPADM2 in 28 and 17 children, respectively. Neurologic toxicities (mainly seizures) were reported in 8 children after COP and in 2 during COPADM1, liver toxicity in 7 children during COPADM1 (fatal in 3) and hemorrhages in 2 after COP and in 14 during COPADM1. Toxicities reported during consolidation and maintenance cycles consisted mainly in infections (91 children; 4 fatal) and hemorrhages (18 children) (Table 2(A)). Three children died at home and the cause of death could not be specified.

Table 2. Toxicities and therapy-related deaths according to protocol phases in the two cohorts of patients.

Therapy-related complications observed in the first cohort of patients are shown in Table 2(B). Also in this group of patients, TLS and infections were the most frequent causes of deaths (13 and 49 children, respectively) with a strong impact in the first phases of treatment.

The 60-month OS, DFS and EFS for the 180 treated patients were: 64.7% (95%CI 58–72.2%) (Figure 1(A)), 86.8% (95% CI 80.8–93.3) and 56.3% (95%CI 49.4–64.1) (Figure 1(B)), respectively. The median follow-up is 67.9 months (range 0.5–118 months).

Graph: Figure 1. Overall Survival (A) and Event-free Survival (B).

The influence on outcome of the following diagnostic features was analyzed: age distribution (<5 years, 5–9 years, and 10–15 years), stage, BM, or CNS involvement, Hb levels, therapeutic group, performance status, body weight <5th percentile, fever, time to reach the hospital >3 hours, symptoms duration >6 weeks, steroid pre-phase, doxorubicin omission. In univariate analysis, a statistically unfavorable prognostic impact on OS was observed for: PS (WHO < 3 vs. WHO ≥ 3; p < .0001; hazard ratio [HR] 0.250, 95% CI 0.151–0.416) (Figure 2(A)), Hb (<8g/dl vs. ≥8 g/dl; p = .0061; HR 2.702, 95% CI 1.329–5.497), BM involvement (negative vs. positive; p = .002; HR 0.355, 95% CI 0.184–0.685), CNS disease (negative vs. positive; p = .0129; HR 0.407, 95% CI 0.200–0.827) and therapeutic group (B vs. C; p = .0452; HR 0.595, 95% CI 0.358–0.989) (Figure 2(C)) (Supplemental Table 1). The univariate analysis for EFS showed the same statistical significant prognostic effect of PS (WHO < 3 vs. WHO ≥3; p < .0001; HR 0.330, 95% CI 0.211–0.518) (Figure 2(B)), Hb (<8g/dl vs. ≥8 g/dl; p = .039; HR 2.077, 95% CI 1.034–4.171), BM involvement (negative vs. positive; p = .0018; HR 0.386, 95% CI 0.212–0.702), CNS disease (negative vs. positive; p = .0012; HR 0.361, 95% CI 0.195–0.670) and therapeutic group (B vs. C; p = .0023; HR 0.493, 95% CI 0.313–0.776) (Figure 2(D)) (Supplemental Table 1).

Graph: Figure 2. Overall Survival by WHO (A) and Event-free Survival by WHO (B). Overall Survival by Therapeutic Group (C) and Event-free Survival by Therapeutic Group (D).

Multivariate analysis confirmed PS (WHO ≥ 3) as the strongest prognostic factor for OS (p = .0005; HR 0.320, 95% CI 0.168–0.609) and EFS (p = .0176; HR 0.500, 95% CI 0.282–0.886); therapy group C was also confirmed as unfavorable prognostic factor for OS (B vs. C; p = .0433; HR 0.437, 95% CI 0.196–0.975) and EFS (B vs. C; p = .0272; HR 0.451, 95% CI 0.222–0.914) (Supplemental Table 2).

When the present results are compared with those obtained in the previously published cohort of patients (2000–2005 vs. 2006–2010), we did not observe statistically significant differences in terms of 24-month OS and EFS: 2000–2005 OS 66.3% (95% CI 60.2–72.9) vs. 2006–2010 OS 65.1% (95% CI 58.3–72.6), p = .89 and 2000–2005 EFS 53.3% (95% CI 47.3–60.1) vs. 2006–2010 57.3% (95% CI 50.4–65%), p = .28 (Figure 3(A,B)). However, in the last cohort of children, therapeutic group B showed a better, but no statistically significant, EFS (63.9% vs. 53.8%; p = .088) (Figure 4). The outcome of the two cohorts of patients was still similar in the adjusted analysis for known prognostic factors differing between the two groups.

Graph: Figure 3. Overall Survival in the two cohorts of patients (2000-2005 vs 2006-2010) (A) and Event-free Survival in the two cohorts of patients (2000-2005 vs 2006-2010) (B).

Graph: Figure 4. Event-free Survival by therapeutic group B in the two cohorts of patients (2000-2005 vs 2006-2010).

In the present study, we report the long-term results of a second cohort of children treated at the CWTH of Baghdad with the Iraqi LMB 96 protocol between 2006 and 2010 (median follow-up 67.9 months). The 60-month OS and EFS were 65.1 and 56.6%, confirming the efficacy of this treatment. In spite of the protocol adjustments aimed at decreasing the toxicity, no significant differences in outcome were observed comparing these results with those previously published; only a trend for a better EFS was recorded in the second study (53.3% vs. 57.3%; p = .28).

In developing countries where the greatest obstacle to the success of therapy is represented by toxic deaths, the identification of subsets of patients at a lower risk of treatment failure, could allow to reduce the intensity of therapy in these cases. A new risk stratification system was adopted in a group of 69 children with B-cell NHL, treated with a LMB-derived protocol in Saudi Arabia, from 2005 to 2011 [[20]]. Children were divided into five therapeutic groups according to the initial disease presentation. Risk-adapted treatment resulted in a lower overall toxicity with OS and EFS of 78.1 and 75.4%, respectively [[20]]. The early treatment response (pre-phase; first treatment course) could also allow to identify a group of patients with a favorable outcome who could benefit from a reduction in therapy intensity. It is interesting to note that in our series the omission of doxorubicin in patients responding to COP did not have a negative impact on the outcome. Most of these patients belonging to the therapeutic group B showed an improved outcome compared to the same counterpart of the previous study (EFS: 63%, 95% CI 54.7–72.4; vs. 54%; 95% CI 46.9–61.6; p = .088). However, in our experience, therapy-related deaths were still high (total 51 patients, 29%). As in the previous study, TLS (25/180 patients; fatal in 21) and infections (197 episodes; fatal in 22) remained the primary challenges affecting the outcome of children diagnosed with B-cell NHL in Iraq. The poor social conditions caused a delay in diagnosis and the poor PS of many children was the strongest poor prognostic factor; in addition, a high abandonment rate continued to occur. The lack of laboratory facilities, adequate supportive care for preventing or treating TLS and infections heavily influenced children survival [[30], [31], [32]]. These data further support the need to improve the availability in this country of key drugs, such as urate-oxidase that is effective in preventing TLS with renal failure and mortality. We, also, strongly recommend the availability of prophylactic antibiotics (quinolones) whose use in afebrile neutropenic episodes have significantly reduced the infection-related mortality in hematologic malignancies for both adults and children [[33]].

These simple measures would most likely have a substantial impact on the prognosis of childhood B-cell NHL in the main pediatric center in Iraq. The contribution of pharmaceutic companies in supporting protocols designed ad hoc for low-resourced countries could allow to use drugs, essential for pediatric neoplastic diseases that seem to have a higher incidence in these countries [[34]].

Furthermore, modern and new approaches, specifically effective for B-cell neoplasms with proven limited therapy-related toxicity are mandatory for the management of patients, children and adults, in this country. The anti-CD20 monoclonal antibody rituximab is the first target therapy used in B-cell malignancies. Rituximab has revolutionized treatment results without excess of toxicity [[35], [36], [37], [38], [39], [40]]. Ribrag et al. [[38]] have clearly demonstrated, in a phase III randomized adult trial, that the addition of rituximab to the LMB regimen significantly improves EFS and OS without increasing toxic effects. Furthermore, Goldman et al. analyzed the combinatorial approach of chemotherapy and rituximab in a pediatric cohort of patients with stage III/IV mature B-NHL and demonstrated the feasibility and the efficacy of chemo-immunotherapy also in the pediatric setting (EFS 95%) [[36]].

Recently, rituximab combined with chemotherapy has proven its superiority in the randomized high-risk Inter-B NHL protocol (LMB-derived) compared with chemotherapy alone [[40]]. Consequently, rituximab is now employed for the treatment of all pediatric patients with high-risk B-cells NHL.

The introduction of rituximab in the next adapted Iraqi protocol could allow a decrease in chemotherapy burden and the related-toxicity, and impact further on the outcome.

The authors wish to thank INTERSOS in supporting the transfer of pathology specimens from Baghdad to Rome and for sponsoring and supporting the Telemedicine program between the Children's Welfare Teaching Hospital in Baghdad, Iraq and the Hematology team at the Sapienza, University of Rome, Italy. The authors thank Prof. Luigi Ruco for the kind contribution to the discussion of the pathological diagnosis of the cases.

Disclosure forms provided by the authors are available with the full text of this article online at http://doi.org/10.1080/10428194.2018.1519810.