TO THE EDITOR: In their report on patients with relapsed or refractory follicular lymphoma, van Oers et al come to the conclusion that the presence of tumor cells in blood or bone marrow detected at the end of induction therapy by BCL2/IgH major break point region (MBR) polymerase chain reaction (PCR) is not predictive of progression-free survival. This conclusion is at variance with previous publications cited by the authors. The results may be a result of an inappropriate selection of patients rather than a genuine inability of BCL2/IgH PCR to predict outcome. Follicular lymphoma is a heterogeneous disorder. In Western countries, only approximately 70% to 90% of patients harbor the t(14;18) translocation juxtaposing BCL2 and IgH. In patients carrying the translocation, up to 50% have a break point outside the MBR of the BCL2 gene. Van Oers et al included all patients with follicular lymphoma regardless of t(14;18) status in their investigation, and they limited the PCR analysis to the MBR. Applying the distribution of genetic abnormalities to the trial cohort, the lymphoma cells of half the patients may not have been detectable by the approach chosen. This may explain why the proportion of patients with proven blood or bone marrow involvement was lower than expected and molecular evidence of marrow involvement did not correlate with histologic evidence. Although the presence or absence of lymphoma cells lacking the BCL2/IgH MBR fusion could not be evaluated, these patients were included in the group of patients without evidence of blood and marrow involvement. Thus, although the BCL2/IgH MBR–positive group was homogeneous with regard to the presence of tumor cells, the BCL2/IgH MBR–negative group was heterogeneous because, in addition to patients with an MBR fusion and no evidence of blood and marrow involvement, it also included an unknown number of patients whose status escaped detectability. It seems reasonable to assume that a substantial proportion of these latter patients also had residual tumor cells in blood and bone marrow. The conclusions drawn by van Oers et al would have been appropriate had they limited their analysis to lymphomas with proven BCL2/IgH MBR fusion. Instead, they analyzed all lymphomas irrespective of their molecular makeup and regarded failure to obtain a PCR product as evidence of absence of tumor cells. This conclusion, however, is illegitimate unless the tumor cells are known to harbor the genetic marker targeted by the PCR primers. Therefore, the investigations by van Oers et al are not suited to challenge the existing literature on the prognostic significance of molecular remission in follicular lymphoma. Whether maintenance therapy with rituximab is equally beneficial in patients with or without molecular remission remains unresolved.