4. DISCUSSION
We enrolled a homogenous group of 374 children diagnosed with the four main NHL subtypes and who were treated with modified BFM 95 regimens over the past 6 years in south China. Compared to many studies, our sample number was not small for survival analysis. The sex ratio was approximately 3:1, with more males than females. BL and LBL were observed more frequently than DLBCL and ALCL. At diagnosis, the dominant majority of patients had advanced-stage disease and were stratified into intermediate- or high-risk group. The incidence rate of BM involvement was about twice that of CNS involvement. Generally, others have reported similar clinical characteristics of pediatric NHL in both white and Asian population.1, 27-31 Overall, the patients responded well to the chemotherapy regimens at our center. The rates of CR (86.1%), 5-year EFS (74.2%) and OS (85.7%) rate in the whole group were comparable to that of different working groups.1, 32, 33 Although less than half of the patients in the present study were genotyped for the C677T/A1298C polymorphism, the subgroup of 158 genotyped patients could be considered as a representative sample of the whole group of pediatric NHL, as no significant difference was found in any clinicopathological characteristics between the two subgroups with or without MTHFR genotyping (TABLE 1). In this analysis, both C677T and A1298C alleles were in H-W equilibrium. The mutation rate of C677T was slightly greater than that of A1298C (47.5% vs. 37.7%). C677T genotype frequencies were similar to that of other reports in Asian pediatric patients with NHL.34, 35 Data on A1298C are limited in Asia. However, some studies on North American and European children with NHL reported a higher mutation rate of C677T and A1298C.27, 36 Therefore, we confirm the high regional and geo-graphic variability of MTHFR polymorphism prevalence in previous publications. 8 The distribution of A1298C genotypes was significantly different among the NHL subtypes (TABLE 2,P = 0.045); patients with DLBCL had the highest mutation rate (58.3%), while patients with LBL had the lowest rate (28.9%).
The influence of MTHFR C677T/A1298C on the survival of patients with cancer (non-hematologic and hematologic malignancy) has attracted great attention from researchers. Regarding adult solid cancers, available evidence on CRC is conflicting; different studies have reported that both C677T and A1298C could be a positive or negative prognostic factor or have no effects on survival of patients with CRC.20-24 No significant differences in the OS were found among the wild-type 677CC and variant CT/TT genotypes in patients with glioma.37 Martha et al. reported that both the C677T and A1298C polymorphisms were not significantly associated with risk of death, but that the 677T variant allele may adversely affect long-term survival in advanced breast cancer.38Lambrecht et al. found that the homozygous 677TT genotype was associated with a higher relapse rate but was not predictive for OS in pediatric osteosarcoma.39 Similarly, Xie et al. suggested that genetic polymorphism of MTHFR C677T in the MTX metabolic pathway had no obvious effect on progression-free survival and tumor necrosis rate in sarcoma (age range, 5-52 years). 40 For hematological malignancy, early studies on childhood ALL have reported controversial results.16-18 However, a recent review concluded that there was no association or opposite effects in the majority of published data, and that these two polymorphisms are not suitable predictors of outcome in pediatric ALL. 5However, the MTHFR 677TT homozygous genotype has been associated with shorter survival in patients with myelodysplastic syndromes41.
Only a few studies have investigated the influence of the C677T and A1298C polymorphisms on the survival of patients with NHL, with controversial results, and there is a particular lack of evidence on A1298C. In an Italian retrospective study with a total of 110 adult patients with NHL treated according to CHOP (cyclophosphamide, doxorubicin hydrochloride, vincristine sulfate, prednisone) or MACOP-B (methotrexate, doxorubicin, cyclophosphamide, vincristine, prednisone, bleomycin) protocol, survival analysis (log-ranks) implied that 677T carriers had a significant reduced probability of 5-year EFS in the whole group and in the MACOP-B–treated group (n = 68). The difference was more obvious when 1298CC cases were excluded from the reference group. However, no significant correlations were identified in the CHOP- treated group(n = 42).7 D’Angelo et al. also found that 677T carriers could have increased risk of relapse and reduced disease-free survival compared with wild-type patients in Caucasian and European children or adolescent NHL treated with LNH-97 and EURO LB-02 multiagent chemotherapy protocols developed by the AIEOP (Associazione Italiana di Ematologia e Oncologia Pediatrica).27 The multicenter NHL-BFM95 trial, involving centers in Austria, Germany, and Switzerland (n = 484), determined that the MTHFR 677 C > T polymorphism was not associated with clinical characteristics (sex, age, tumor stage) and did not appear to influence EFS in pediatric patients with NHL.36
However, in our study, the C677T polymorphism was associated with all-events incidence. Patients carrying the wild-type were more likely to experience events. Moreover, K-M survival analysis revealed that 677T allele could significantly increase the probability of 5-year EFS and was associated with a non-significant increased probability of 5-year OS. Therefore, the 677T variant allele might have played a protective role in the survival of childhood NHL in the present analysis. The influence of T variant allele on survival was more obvious when only BL/DLBCL were included in K-M analysis. As no significant difference was found when only LBL or ALCL were included, we assumed that the significant differences in the whole group of NHL resulted from the large proportion of BL/DLBCL cases (n = 81, 51.3%). Furthermore, we speculated that cancer type, age, treatment regimens, and particularly ethnicity might account in part for the discrepant results. A large meta-analysis concluded that the C677T polymorphism showed a statistically significantly increased risk for whites, but a decreased risk for Asians6. Therefore, it is also possible that the C677T polymorphism has an opposite influence on the survival of European and Chinese pediatric NHL even when patients are treated using the same chemotherapy protocol. The reduced MTHFR activity due to the 677T variant allele might have positive effects on survival of pediatric NHL in the Chinese population.
Interestingly, the A1298C polymorphism has been associated with increased NHL risk for in Asians.6 In the present study, except for NHL subtype, no significant associations of A1298C in clinical characteristics (age, sex, stage, risk, BM/CNS involvement) and outcome (EFS and OS) could be detected. However, a fact that should not be neglected is that MTHFR enzymatic activity is affected by the A1298C polymorphism to a lesser extent than by C677T, and the remaining activity in the 1298CC homozygous genotype still represents 60% of the normal status despite of the substitution of glutamate for alanine in the amino acid sequence.5, 9, 10 Furthermore, due to partial linkage disequilibrium, it is possible, but indeed very rare, for 677T and 1298C variant alleles in cis to coexist, supporting the hypothesis that triple mutations (677TT/1298AC or 677CT/1298CC) or double homozygous conditions (677TT/1298CC) are probably de novo recombinant events.42 In the present study, virtually all patients with 677TT (n = 14) had the wild-type 1298AA genotype. For these reasons, it was difficult to observe a clear allele–dosage effect for the variant 1298C being better accounted for in homozygous conditions. This could partly explain the lack of significant associations between the A1298C polymorphism and survival in our study, base on the difficulty in determining the effects of 1298 variant itself on survival.
Our results indicated that patients with ALCL had a significantly higher probability of experiencing events than patients with the other three subtypes, while patients with DLBCL had a relatively superior survival. Overall, patients with ALCL have inferior EFS compared to the other three types.1, 30, 43 Besides, we found that age, sex, and BM involvement did not affect survival, but CNS involvement, advanced stage, and intermediate/high risk predicted inferior survival in pediatric NHL. Similarly, a report from BFM Group study NHL-BFM95 (n = 505; age range, 1.4–19.7 years) revealed that CNS-positive patients had remarkably lower 3-year EFS compared to the whole group (69% vs. 89%), and that stage and risk were adversely associated with 3-year EFS.31 Besides, a large multicenter retrospective study of 749 pediatric patients with NHL in east Asia suggested that BM involvement does not affect treatment outcome, while CNS involvement and advanced stage (≥ stage III) were associated with poor outcomes (P < 0.05).30 More recently, Choeyprasert et al. identified advanced disease stage as a significant poor prognostic factor for 5-year EFS of pediatric NHL in Thailand.29
We are aware of the limitations of our study, given the retrospective nature and relatively small sample size enrolled for analyzing the influence of MTHFR on the survival of pediatric NHL. Moreover, not all newly diagnosed pediatric NHL at our center during the study period were included for analysis, as a small group of them did not receive modified BFM 95 regimens, but were alternatively treated with an adult protocol such as the R(rituximab)-CHOP regimen.
In conclusion, we found that the MTHFR C677T polymorphism and some clinicopathological characteristics including, NHL subtype, CNS involvement, stage, and risk at diagnosis, are significantly related to the survival of pediatric patients with NHL treated with modified BFM 95 regimens. The 677T allele could increase the probability of 5-year EFS and 5-year OS. No evidence showed that the MTHFR A1298C polymorphism influenced the EFS or OS rate in pediatric patients with NHL. Further larger studies are encouraged to verify these results properly.