The possible role of minimal residual disease (MRD) in the management of patients with follicular lymphoma (FL) has been a recurrent topic in this field for at least two decades. In this issue of the journal, Pulsoni et al (2019) report a study using MRD, determined in bone marrow (BM) and/or peripheral blood (PB), in patients with early stage FL treated with involved-field radiotherapy (IF-RT). MRD status after IF-RT was used to guide the decision of whether or not to administer systemic rituximab; this strategy resulted in longer progression-free survival (PFS). The study raises again the interest of MRD in FL, particularly in the setting of localized disease in which only few data are available. Although many of the published series have limitations, including the retrospective nature, small sample size, mixed DNA sources (BM vs. PB), and lack of prospective planning for MRD time points, nowadays considerable evidences suggest that persistence of polymerase chain reaction (PCR)-detectable residual tumour cells in BM, and to a lesser extent in PB, is an independent predictor of relapse and shorter PFS in patients with FL (Rambaldi et al, 2002; Rambaldi et al, 2005; Hirt et al, 2008; Goff et al, 2009; Morschhauser et al, 2012; Ladetto et al, 2013). However, no significant differences in terms of overall survival (OS) have been detected according to the MRD status, specifically in the only prospective study performed with this exploratory aim (Ladetto et al, 2013). Furthermore, besides its prognostic significance, the main issue in the clinical setting is whether or not MRD could be useful for making treatment decisions, as in acute lymphoblastic leukaemia or promyelocytic leukaemia. Unfortunately, no solid data that can be applied in the clinic have been published to date. On the other hand, 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) scanning was shown to be a strong predictor of outcome (Trotman et al, 2014) and it is currently recommended for assessing response in FL (Cheson et al, 2014). The combination of MRD by PCR and FDG-PET has been suggested to be an excellent tool to define the prognosis of these patients (Luminari et al, 2016). Recently, the Fondazione Italiana Linfomi ran a phase 3 randomized trial that included MRD-based decision-making, with the objective of comparing standard maintenance with a tailored maintenance/consolidation programme based on MRD and FDG-PET in advanced stage FL. Unfortunately, preliminary data revealed no superiority of the tailored strategy (Federico et al, 2019). More recently, interest has been raised regarding the application of more modern techniques to detect MRD, including next generation sequencing (NGS), in the search for biomarkers that may help in adapting therapy to the disease risk. The sensitivity of NGS is substantially higher and, in addition, could be applied to a considerably higher number of patients with FL. Thus, the new tool might further improve MRD studies in FL. Follicular lymphoma usually presents in advanced stage, although up to 20% of newly diagnosed cases show localized disease. IF-RT with curative intent is limited to non-bulky stages I–II FL (Dreyling et al, 2016). Thus, patients with localized FL treated with IF-RT have an excellent outcome in terms of OS, although at least half of these individuals will eventually relapse. The detection of patients with higher risk of relapse who might be candidates for further systemic therapy, such as rituximab, as used by Pulsoni et al (2019), is an interesting challenge. With its limitations (heterogeneous series with a relatively small number of patients), this study is the first, to our knowledge, to address the impact of MRD in localized FL and adapting therapy to disease risk. Although the administration of rituximab to patients with MRD positivity after IF-RT could seem reasonable to many clinicians, the excellent prognosis of the whole series, irrespective the therapeutic approach, prevents any strong conclusions. In summary, the study by Pulsoni et al (2019) is interesting because it again centres the debate regarding the usefulness of MRD in FL and its probable role in strategies adapted to the risk of the disease. There is no doubt on the prognostic importance of MRD in FL; however, with the data currently available, it is still too early to integrate MRD into clinical practice. Future research could be useful to (i) refine our dynamic capacity for risk stratification; (ii) enable early termination or intensification of treatment based on MRD status and (iii) enable the discontinuation continuous therapy with targeted agents in those patients who become MRD-negative during treatment. The authors declare no conflict of interest regarding current manuscript. Both authors reviewed the bibliography and wrote the manuscript.