Prognostic value of FDG-PET prior to autologous stem cell transplantation for relapsed and refractory diffuse large B-cell lymphoma

High-dose therapy (HDT) plus autologous stem cell transplantation (ASCT) is standard-of-care as consolidation for patients with relapsed and refractory (rel/ref) diffuse large B-cell lymphoma (DLBCL).¹  HDT-ASCT is indicated in other rel/ref aggressive B-cell non-Hodgkin lymphomas (B-NHL), including transformed histology.²  Two recent large prospective randomized phase III studies with interventions of conditioning modification³  and maintenance therapy⁴  failed to improve long-term disease-free survival rates above 50% to 60%.

Eligibility for HDT-ASCT is based on the presence of chemosensitive disease to salvage therapy (ST) as defined by complete or partial remission assessed by computed tomography (CT) imaging.⁵  The controversy surrounding the prognostic significance of interim positron emission tomography (PET) using ¹⁸F-deoxyglucose (FDG) restaging during induction therapy for DLBCL has prompted the generation of revised criteria for more accurate determination of chemosensitive response.⁶  The prognostic significance of FDG-PET pre-HDT-ASCT for DLBCL has been demonstrated previously^7-10 ; however, not with modern criteria. We aimed to determine the prognostic impact of FDG-PET-assessed response to ST before HDT-ASCT for rel/ref DLBCL/transformed B-NHL by modern criteria established at the first international workshop on FDG-PET reporting in lymphoma held in Deauville, France ¹¹  and recently recommended at the 12th International Conference of Malignant Lymphoma.¹² 

We retrospectively reviewed a database of 129 adult patients with rel/ref DLBCL and transformed B-NHL previously exposed to at least1 prior line of induction chemotherapy and who were chemosensitive to ST and proceeding to HDT-ASCT after a restaging FDG-PET scan at Memorial Sloan-Kettering Cancer Center (MSKCC) from 2002 to December 2012. Rel/ref disease was confirmed by biopsy in 123 of 129 (95%) of patients. Eligibility for HDT-ASCT was based on ST response per above, as well as adequate organ function, performance status, and CD34⁺ stem cells for re-infusion per institutional standard. A waiver of authorization to carry out this analysis was approved by the MSKCC Institutional Review Board.

Chemosensitivity was assessed per standard CT criteria for B-NHL.⁵  All PET/CT scans were obtained 60 minutes after radiotracer injection, extending from skull base to upper thigh, on Discovery STE or LS cameras (GE Medical Systems, Milwaukee, WI). A low-dose CT scan (120-140 kV, ∼80 mA) was followed by emission images (3 minutes per bed position). Deauville criteria¹¹  for FDG-PET response was applied for purposes of analysis on the 5-point scale: 1, no uptake; 2, uptake ≤mediastinal blood pool; 3, uptake > mediastinal blood pool but ≤ liver; 4, uptake > liver at any site; and 5, uptake > liver and/or new sites of disease. No patients had Deauville 5 response, because this represented progression of disease to ST, thus determining ineligibility to proceeding to HDT-ASCT consolidation. Two staff nuclear medicine radiologists (H.S. and G.A.U.) reviewed and scored all images. Traditional pre-ST risk factors⁴  were applied for purposes of analysis, including histologic phenotype at relapse (including cell of origin in DLBCL as determined by Hans criteria¹³ ), secondary age-adjusted International Prognostic Index (sAA-IPI), and previous remission duration to initial induction chemotherapy.

Overall survival (OS) and progression-free survival (PFS) were defined as the time from HDT-ASCT until death from any cause and time from HDT-ASCT to disease progression or death, respectively. Univariate probabilities and 95% confidence intervals (CIs) of OS and PFS were estimated using Kaplan-Meier methodology, and survival distributions were compared across patient and treatment characteristics using a log-rank test, with P values <.05 considered significant.

The characteristics of the 129 patients appear in Table 1. The majority of patients (66%) were de novo rel/ref DLBCL. All patients were previously exposed to rituximab and were chemosensitive by CT criteria.⁵  FDG-PET scans after ST occurred at a median of 34 days (range: 5-173 days) before HDT-ASCT and at a median of 14 days after ST. Of the 6 patients with pre-HDT-ASCT FDG-PET occurring >3 months from HDT-ASCT, all received pre-HDT-ASCT IFRT. The median follow-up among survivors was 43.6 months (range: 4.6-135.6 months). PFS and OS rates for the entire cohort at 3 years were 67% (95% CI: 58-75) and 74% (95% CI: 65-81), respectively (Figure 1A). No significant differences in PFS and OS were observed according to standard pre-ST risk factors including sAA-IPI, relapse <12 months or primary refractory disease vs relapse ≥12 months, and type of ST. In addition, transformed or de novo DLBCL and cell-of-origin factors lacked prognostic significance. The only pre-HDT-ASCT factor associated with PFS and OS was response to ST by FDG-PET Deauville criteria. At 3 years, patients with Deauville 1 to 3 experienced superior PFS (Figure 1A) and OS rates (Figure 1B) of 77% (95% CI: 65- 86) and 86% (95% CI: 75-92), respectively, compared with patients with Deauville 4 who had respective PFS and OS rates of 49% (95% CI: 34-62) and 54% (95% CI: 39-68) (P < .001).

This analysis represents the only series of chemosensitive rel/ref DLBCL and transformed B-NHL patients, predominately of de novo DLBCL histologic phenotype, proceeding to HDT-ASCT with responses to ST assessed by FDG-PET according to modern Deauville criteria on the 5-point scale. Our retrospective data demonstrate significant prognostic impact of FDG-PET response to ST (Deauville 1-3) with significant PFS and OS benefit (P < .001). Whereas other series’ have demonstrated prognostic significance with FDG-PET response before HDT-ASCT,^7-9,14  none has used the contemporary Deauville 5-point scale.¹¹  This 5-point scale of response criteria was recommended by consensus of the Imaging Working Group at the 12th International Conference of Malignant Lymphoma in Lugano, Switzerland.¹²  Additional key differences of previous publications assessing risk by FDG-PET pre-HDT-ASCT, as compared to our study, include patients undergoing HDT-ASCT in first remission,⁹  varying lymphoma histologies,^7,14  and smaller numbers of patients.^7-9  In the other larger, and most contemporary, study of 143 patients with rel/ref DLBCL or transformed lymphoma from the Dana-Farber Cancer Institute/Massachusetts General Hospital, a negative FDG-PET scan conferred a rate of OS benefit at 4 years of 75% vs 56% in FDG-PET-positive patients (P = .025), results notably similar to those obtained in our study.¹⁰  Response criteria by FDG-PET were not specified in the Dana-Farber Cancer Institute/Massachusetts General Hospital study.

Limitations of our study should be noted. First, this is not a prospective study and, thus, it is subject to the inherent limitations of retrospective data. Patient selection within this single-center study may account for the improved PFS and OS over historic controls of 2 recent prospective multicenter studies of HDT-ASCT for de novo rel/ref DLBCL.^3,4  Also, our study contained 15 patients (12% of the cohort) with transformed histology DLBCL and 8 patients (6%) who received 2 lines of ST, which was different from the recent prospective HDT-ASCT studies for de novo DLBCL. Lastly, 54 of the 129 patients with limited-stage disease received pre-HDT-ASCT IFRT after post-ST restaging as part of the HDT-ASCT^15,16  at the discretion of the treating physician. It is important to note that the PFS and OS rates between pre-HDT-ASCT IFRT and no pre-HDT-ASCT IFRT were similar (P = .20 and P = .73, respectively), even when restricted to the 48 patients who achieved Deauville 4 response to ST (PFS: P = .32; OS: P = .40).

In conclusion, our study represents the first report of the prognostic impact of pre-HDT-ASCT FDG-PET response to ST according to Deauville criteria in chemosensitive rel/ref DLBCL. Clearly, given the inferiority of patients achieving a Deauville 4 response to ST, this group of patients should be the subject of risk-adapted investigation. Such investigational interventions could include, but are not limited to, immunotherapeutic approaches to overcome relative chemotherapy insensitivity such as consolidation with chimeric antigen receptor–modified autologous T-cell therapy directed against CD19 post-HDT-ASCT (NCT01840566), immune-checkpoint blockade post-HDT-ASCT,¹⁷  or allogeneic hematopoietic cell transplantation. With regard to the latter treatment modality, we have recently demonstrated that the allogeneic graft-versus-lymphoma effect may overcome relative chemotherapy insensitivity, achieving similar long-term outcomes with allogeneic hematopoietic cell transplantation after nonmyeloablative conditioning in chemosensitive patients to ST according to CT criteria across all Deauville responses by FDG-PET.¹⁸ 

The authors thank the staff of the Lymphoma and Adult Bone Marrow Transplant Service at MSKCC for their enduring dedication to patient care.

Contribution: C.S.S. and C.H.M. designed the study; C.S.S., M.J.M., A.D.Z., and C.H.M. interpreted the data; C.S.S., J.M, J.C.M, S.M.D, P.H, G.A.U., and H.S. analyzed the data; all authors wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Craig S. Sauter, Memorial Sloan Kettering Cancer Center, Box 276, 1275 York Ave, New York, NY 10065; e-mail: sauterc@mskcc.org.