Second Malignancy After Treatment for Non-Hodgkin Lymphoma: a Systematic Review and a Meta-Analysis of Population-Based and Cohort Studies.

The risk of second malignancy in non-Hodgkin lymphoma (NHL) survivors have been described in several studies, but the available evidence have yielded conflicting results. Thus, we performed a systematic review and a meta-analysis on population-based and cohort studies to provide a quantitative assessment of the available evidence on the risk of secondary occurrence of cancer after treatment for NHL. Primary aims of our research were to evaluate the pooled Relative Risk (RR) of second cancer for overall malignancies and for every cancer.

A Medline search from 1985 to 2008 was conducted for identification of relevant observational studies that provide estimates of RR, as measured by standardized incidence ratios (SIR) that is the observed-expected ratio of second malignancy appearing during follow up of NHL. The reference lists of identified articles were inspected to identify additional papers. Criteria for including studies in the meta-analysis were: a) studies on naïve patients with any stage of NHL, b) studies reporting measure of SIR or data allowing such outcome to be derived and c) English language. The article included, had to have been published in peer-reviewed literature. We did not exclude papers on the base of therapeutic regimens. The Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines were followed. Pooled RR and 95% confidence interval (CI) were calculated using random effect models. Tests on heterogeneity and sensitivity analysis was conducted. Also, the publication bias was evaluated.

Eleven papers meet the inclusion criteria reporting RRs for all malignancies. These studies included 223,593 patients affected by NHL of which 14,952 presented a second cancer. RRs ranged from 0.93-1.90 and the meta-RR for second malignancy was 1.24 (95%CI: 1.11-1.39). The analysis on solid tumours, excluding haematological malignancies, based on seven studies did not show a significantly higher risk for secondary cancer: the meta-RR was 1.06 (95%CI: 0.84-1.34). However, some kind of cancer showed a statistically significant excess of risk, as lung cancer (meta-RR on nine studies: 1.46 95%CI: 1.33-1.59) and bladder cancer (meta-RR on eight studies: 1.42 95%CI: 1.33-1.51). Prostate (meta-RR on ten studies: 1.08 95%CI: 0.93-1.24) and breast cancer (meta-RR on eleven studies: 0.99 95%CI: 0.83-1.19) has demonstrated no evidence of association with NHL therapy. Moreover, we found a higher risk of developing Hodgkin lymphoma and myeloid leukemia (respectively meta-RR on seven studies: 6.44 95%CI: 5.59-11.55 and meta-RR on five studies: 7.81 95%CI: 2.59-24.46). Regarding leukemia, however we have to consider that they include either acute or chronic leukemia and sometime could also include higher risk myelodisplastic syndrome . No evidence of publication bias was observed.

Although there exist a number of paper on this topic, until now there are not been attempts to perform a meta-analysis on RR of second malignancies after treatment for NHL. Indeed comparative analysis on the incidence of second cancer presents several issues, including the heterogeneity of NHL, the source of data, the time during which the study was performed, the different schedule of chemotherapy, the dosage of radiotherapy used in the different period of time and the length of follow-up. Although these problems could reduce the accuracy of the meta-analysis, our results indicate that NHL treatment is associated with a significantly higher risk of second malignancy, in particular for some specific cancer like lung and bladder and same haematological malignancies as Hodgkin lymphoma and myeloid leukemia.

Finally, we think that it is important to determine by meta-analysis the incidence of each second cancer in survivor of NHL as for some malignancies screening test could be performed and early diagnosis could be made.

No relevant conflicts of interest to declare.