The follicular lymphoma (FL) is the second more frequent histotype of non-Hodgkin's lymphomas in the Western world, and represents one of the entities where the minimal residual disease (MRD) has been more frequently studied. Indeed, the history of MRD in FL is now 15 years old, starting with the demonstration that rituximab (R) significantly reduced the molecular tumor burden when administered after the CHOP (Cyclophosphamide, Doxorubicin, Vincristine, Prednisone) regimen: in that setting, patients achieving MRD negativity showed a significantly longer event-free survival (Rambaldi et al., 2002). Then, many other works focused on the MRD in the transplanted patients: cases MRD-negative after autologous transplantation remained relapse-free and alive in a higher percentage than those persistently MRD-positive (Galimberti et al., 2003; Melillo et al., 2005).

Subsequently, when yttrium-ibritumomab tiuxetan as consolidation was compared to the observation only, it was evident that radio-immunotherapy was able to rapidly eradicate MRD in the majority of patients (Goff et al., 2009): after this treatment 77% of patients became PCR-negative, with a significant favorable impact on progression-free survival (PFS) (Ibatici et al., 2014).

In the 2010, van Oers and colleagues demonstrated, in a series of relapsed patients, that the presence of the BCL2/JH at diagnosis had a negative impact on the outcome, whereas the molecular status before rituximab maintenance did not discriminate patients with worse outcome (van Oers et al., 2010).

More recently, two large studies conducted by the Fondazione Italiana Linfomi (FIL) showed that after R-CHOP, R-CVP (cyclophosphamide, vincristine, prednisone), R-FM (fludarabine and mitoxantrone) and R-FND (fludarabine, mitoxantrone, dexamethasone) the MRD status significantly conditioned the PFS; indeed, patients in partial response (PR) but MRD-negative showed longer PFS than those in complete remission (CR) but still MRD-positive (Ladetto et al., 2013; Galimberti et al., 2014).

More recently, the appearance in the scenario of the indolent lymphomas of the new anti-CD20 monoclonal antibody, obinutuzumab, prompted researchers to evaluate its value also in the context of the molecular response. In the Gallium trial, where CHOP, CVP or bendamustine were combined with rituximab or obinutuzumab, the new antibody resulted more powerful than rituximab, offering 92.5% of MRD-negativity when combined with bendamustine, 91.3% with CHOP, and 91.4% with CVP. Interestingly, the use of obinutuzumab in respect of rituximab added 13.5% of good molecular responses in the cohort receiving CHOP, 15.4% in those patients treated with CVP, but only 3% in the bendamustine arm, thus suggesting that molecular clearance offered by bendamustine was higher than that induced by the other two chemotherapeutic regimens (Pott et al., 2016).

Then, MRD is still today a hot topic and often more frequently it is introduced as aim in several phase-3 trials, as in the FIL FOLL12 study (trial.gov identifier number: NCT02063685), where the experimental arm is based on the PET and MRD results: a de-intensified treatment was reserved to MRD- and PET-negative cases, a consolidation with radio-immunotherapy to patients still PET-positive after induction or a pre-emptive therapy was adopted for PET-negative but MRD-positive cases.

Nevertheless, MRD is still not a decisional tool in the routine clinical practice: indeed, the ESMO guidelines recently edited, even if clearly recognizing the prognostic role of the MRD, state that the MRD “still cannot lead the therapeutic strategy” (Dreyling et al., 2017).

In this context, few data about R-bendamustine and MRD in FL are today available: Zohren and colleagues reported that 89.5% of cases receiving R-bendamustine achieved the MRD negativity, with a significant advantage in terms of PFS (Zohren et al., 2015).

More recently, in the context of Gadolin trial, obinutuzumab combined with bendamustine induced 82% of MRD negativity vs. 43% of the bendamustine alone, with a significant longer PFS for MRD-negative cases (Pott et al., 2015).

With these premises, we decided to assess the impact of the R-bendamustine on MRD in a cohort of 48 FL patients treated in 4 different Italian centers. The aim of this retrospective study was to measure both by qualitative and quantitative PCR the molecular tumor burden before and 2 months after the end of treatment with R-bendamustine. The presence and quantization of the BCL2/JH rearrangement at baseline and its behavior after induction was then compared with clinical features, response rates, and 2-years survivals.

Notwithstanding its long history, the MRD evaluation in FL and its significance in the clinical practice are still matter of debate. The recent European guidelines affirm that MRD has really got a predictive and prognostic significance, but that it is not still suitable for driving treatment of FL patients (Dreyling et al., 2017).

This is probably due to several considerations: (1) different studies proved that MRD-negative patients show a better prognosis, either when receiving chemo-immunotherapy or high-dose treatment. Nevertheless, these experiences have been conducted or in the context of large phase-III multicenter trials or, at the opposite, on small series of patients, and thus many physicians consider those results not fully exportable to their routine activity; (2) notwithstanding the standardization performed by some cooperative groups (van der Velden et al., 2003; Pott et al., 2013), the availability of a common, easily performable and cheap molecular method for investigating MRD is still lacking; (3) about MRD and R-bendamustine, few data are still available, and thus further information on this specific setting would be waited.

Our study is retrospective and involved a small number of patients, but it represents the picture of that occurs to FL patients in the real life. That statement is well supported by the observation that our clinical results are superimposable to those from other larger and pivotal trials: in particular, in our series CR rate resulted 79%, 2-years PFS 85%, and 2-years OS 94%. These values result comparable to those recently reported by Mondello et al. in 192 cases randomized to receive R-bendamustine or R-CHOP (Mondello et al., 2016). Those authors reported the superiority of R-bendamustine, with a significant prolongation of the median PFS (152 vs. 132 months); in that study, bendamustine offered 94% of ORR, and 63% of CR. Probably, the higher rate of overall response (100%) found in our series could be explained by the retrospective nature of this study (where enrolment was only based on the DNA availability).

Nevertheless, even with the limitation of the number of our cases, the present study could add some information about MRD in the setting of R-bendamustine to those previously reported by Zohren and colleagues in the contex of the StiL trial (Rummel et al., 2013).

Moreover, it is relevant to observe that our molecular data are perfectly in line with those previously reported by other authors: the BCL2/JH rearrangement at baseline has been detected in 62% of our patients; this is the same percentage reported in the Gadolin trial (Pott et al., 2015) and in the StiL study (Zohren et al., 2015), and slightly higher than that ourselves reported in previous FIL trials (52% and 51%) (Ladetto et al., 2013; Galimberti et al., 2014).

Also the number of cases with BM involvement documented by microscopic observation but without molecular marker at baseline was analogous to that previously reported (14% in the present study vs. 17% in the FOLL05 trial) (Galimberti et al., 2014): this phenomenon could be surely related to a possible poor BM sampling, but also to the emerging demonstration that at least 8–10% of the “PCR-negative” cases present some “rare” breakpoints in the BCL2 gene (these rearrangements have not be evaluated in the present study).

Thus, our study would suggest that R-bendamustine exerts a so frequent and deep clearance of MRD that it would be able to by-pass the negative prognostic impact usually played by the MRD in other contexts, such as when R-CHOP, R-CVP, R-FM, or R-FND were adopted as induction regimens.

Indeed, in the present work the baseline molecular tumor burden did not condition the quality of response, opposite to that we observed in the FOLL05 study, where CR was achieved by the half of the initially PCR-positive in respect of the PCR-negative cases (Galimberti et al., 2014).

In the present study, relapsed patients presented an initial molecular burden 15 times higher than that measured in not relapsed/progressed cases, but this difference did not translate in a significant advantage in terms of PFS. This could be probably due to the small number of patients, but we cannot exclude that it could be also the consequence of the more rapid and deeper MRD clearance exerted by bendamustine: in our series, more than 90% of cases achieved the MRD negativity, and quantitative data showed that the majority of cases achieved a very deep reduction of the molecular disease. The median reduction of 4 logs is encouraging in respect of the 2 logs reported after R-CHOP (Galimberti et al., 2014) and the 2.5 logs reported after yttrium-ibritumomab tiuxetan treatment (Ibatici et al., 2014).

Interestingly, the rapidity of the molecular clearance offered by bendamustine has been already well demonstrated in the Gallium (Pott et al., 2016) and Gadolin (Pott et al., 2015) trials, where bendamustine was combined with obinutuzumab: in both trials, the rapid efficacy of the R-bendamustine also in molecular terms has been clearly demonstrated. It is also relevant that in the arm with R-bendamustine of the Gallium trial, 90% of patients became MRD-negative, a percentage comparable to that observed in our study (93%) and that in that trial, as in the present study, the MRD status did not condition the PFS length.

In conclusion, our study suggests that the combination of rituximab and bendamustine, by exerting a rapid and deep clearance of the molecular disease, could represent a highly effective therapeutic approach to FL, also in terms of molecular disease. The larger ongoing studies will show if the MRD would be used as adjunctive tool for better selecting the more effective treatment for FL patients.

This study was carried out in accordance with the Declaration of Helsinki. All patients gave the consent to leave the leftover of the routine analyses for scientific purposes. this statement was approved by the legal office of the AOUP.

All authors participated to the conception of the study. EC, GE, SuGr, and FG performed PCR assays. SG, NC, MR, GC, FM, LI, FF, AB, LR, SK, DV, and MP enrolled patients, performed treatment and clinical follow-up. LM and EC performed the microscopic analyses. SG and MP performed statistical analyses. All authors approved the final version of the manuscript.