Clinicogenetic Risk Models in Patients Randomized to Receive Consolidative Autologous Stem-Cell Transplantation after Frontline R-CHOP for Advanced Follicular Lymphoma: An Analysis from the GLSG2000 Trial

Introduction: Advanced follicular lymphoma (FL) is a clinically and molecularly heterogeneous disease. About 20% of patients have early progression of disease (POD) and short overall survival (OS). We have previously shown that integration of lymphoma-specific gene mutations and clinical factors improves pretreatment risk stratification (Pastore, 2015) and prediction of early POD (i.e., within 24 months, POD24; Jurinovic, 2016). Recently, we have shown that high-dose therapy (HDT) followed by autologous stem-cell transplantation (ASCT) is an effective treatment option for eligible patients with high-risk disease as defined by POD24 (Jurinovic, 2018). Here, we aimed to explore whether HDT/ASCT is an effective frontline therapy for patients identified to be high-risk by clinical (i.e., FLIPI) or clinicogenetic risk models (i.e., m7-FLIPI, POD24-PI).

Methods: We performed targeted DNA deep-sequencing of >150 genes in available diagnostic FL biopsies from 165 patients ≤60 years with advanced FL from the GLSG2000 trial who uniformly received R-CHOP as frontline treatment. Of these, 87 patients (53%) were randomized to receive consolidative HDT/ASCT, 78 (47%) were randomized to interferon maintenance. We performed intention-to-treat (ITT) survival and regression analyses to explore whether known clinical and clinicogenetic risk factors can be overcome by ASCT.

Results: The HDT/ASCT and no-HDT/ASCT cohorts were balanced regarding age (48 vs 50 years), sex (49% vs 64% male patients), high-risk FL International Prognostic Index (FLIPI; 25% vs 29%), Eastern Cooperative Oncology Group Performance Score >1 (6% vs 5%) and mutation status of EZH2 (23% vs 18%) and TP53 (3% vs 3%). The incidence of POD24 was not significantly lower in the HDT/ASCT cohort (8% vs 14%, p=0.32). After a median follow-up of 7.5 years, 5-year failure-free survival (FFS) rates in the HDT/ASCT and no-HDT/ASCT cohorts were 77% and 69% (HR 0.7, p=0.16), 5-year OS rates were 95% and 90% (HR 0.6, p=0.21), respectively.

The high-risk cohorts identified by FLIPI, m7-FLIPI, and POD24-PI comprised 27% (n=45), 18% (n=29) and 22% (n=37) of patients, respectively (Fig. A). The m7-FLIPI reclassified 10% (n=16) of patients from high-risk FLIPI to low-risk m7-FLIPI. The POD24-PI reclassified 5% (n=9) of patients from high-risk FLIPI to low-risk POD24-PI; one patient was reclassified from low-risk FLIPI to high-risk POD24-PI. Patients identified to be high-risk by all three indices had shorter FFS (FLIPI: HR 2.8, p=0.0002; m7-FLIPI: HR 3.0, p=0.0003; POD24-PI: HR 2.5, p=0.0013), but OS was not different (FLIPI: HR 1.4, p=0.47; m7-FLIPI: HR 1.5, p=0.45; POD24-PI: HR 1.5, p=0.47). The risk to develop POD24 was increased in high-risk patients (FLIPI: OR 4.4, p=0.007; m7-FLIPI: OR 4.8, p=0.005; POD24-PI: OR 4.3, p=0.008).

Consolidative HDT/ASCT did not prolong FFS in high-risk patients as defined by FLIPI (HR 1.2, p=0.67), m7-FLIPI (HR 1.2, p=0.70; Fig. B) and POD24-PI (HR 1.3, p=0.63; Fig. B). Similarly, OS was not significantly improved in all three high-risk cohorts (FLIPI: HR 0.2, p=0.13; m7-FLIPI: HR n/a, p>0.99; and POD24-PI: HR 0.3, p=0.22). In low-risk patients, HDT/ASCT was associated with a non-significant trend towards prolonged FFS (FLIPI: HR 0.5, p=0.061; m7-FLIPI: HR 0.6, p=0.16; POD24-PI: HR 0.5, p=0.068; Fig. B), but again OS was not significantly different (FLIPI: HR 0.8, p=0.69; m7-FLIPI: HR 0.8, p=0.66; and POD24-PI: HR 0.7, p=0.52).

Conclusions: Our ITT-analysis confirms that consolidative HDT/ASCT should not be offered to unselected cohorts of patients with previously untreated, advanced FL after R-CHOP. Also, our current clinicogenetic risk models are not optimized to select high-risk patients who may benefit from frontline HDT/ASCT. The fraction of patients identified to be high-risk by FLIPI, m7-FLIPI and POD24-PI is low when applied to younger, medically fit patients. Moreover, the fraction of patients being reclassified by integrating gene mutation data is low in this patient cohort. Therefore, we are developing specific stratification algorithms for younger, medically fit patients who are eligible for dose-intensified approaches.

Figure.

View largeDownload slide

Figure.

View largeDownload slide

Close modal