Limited data exist about the role of second-line chemotherapy response assessed by positron emission tomography (PET) as a prognostic factor in patients with aggressive non-Hodgkin Lymphoma (NHL) who undergo autologous stem cell transplantation (ASCT). The objective of this analysis was to investigate the main determinants of prognosis in patients with aggressive B-cell NHL who undergo ASCT, focusing on the impact of pretransplantation PET, secondary age-adjusted International Prognostic Index (sAA-IPI) score, histology, and previous response to first-line chemotherapy.
Seventy-five patients with diffuse, large B-cell lymphoma or grade 3 follicular lymphoma who were treated at the author' institution with second-line chemotherapy (combined ifosfamide, etoposide, and epirubicin [IEV]) followed by ASCT between September 2002 and September 2006 were included. All patients were evaluated by PET after 1 to 3 courses of IEV chemotherapy before ASCT, and all patients received a conditioning regimen of combined carmustine, etoposide, cytosine arabinoside, and melphalan. The prognostic impact of pretransplantation PET, sAA-IPI score, histology, and previous response to first-line chemotherapy was evaluated by univariate and multivariate analyses.
Seventy-two of 75 patients underwent ASCT. In a univariate analysis for progression-free survival (PFS) and overall survival (OS), a significant association was observed with pretransplantation PET (PFS, P < .00001; OS, P < .01) and previous first-line response (PFS, P = .02; OS, P = .04). In the multivariate framework, pretransplantation PET was identified as the only independent prognostic factor (PFS, P < .001; OS, P = .01).
Most patients with aggressive B-cell NHL initially respond to first-line combined cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP)-like chemotherapy regimens, but only 40% to 50% of all patients achieve sustained remission and are cured with this approach.1 The advent of rituximab significantly reduced the percentage of refractory or relapsing B-cell non-Hodgkin lymphoma (NHL), which currently is approximately 30% to 40%.2, 3 For these patients and for those who do not achieve complete remission after first-line chemotherapy, the best therapeutic strategy available is autologous stem cell transplantation (ASCT); with this approach, the 5-year event-free survival rate among chemosensitive patients is approximately 40% to 45%.4
Until recently, the only established prognostic factors for such patients who are about to receive high-dose chemotherapy (HDT) have been their response to second-line chemotherapy and the secondary age-adjusted International Prognostic Index (sAA-IPI) score,4–6 which is a clinical parameter that was derived from the IPI score for primary high-grade NHL and was validated by Hamlin and coworkers, who observed that increased lactate dehydrogenase level, advanced clinical stage, and impaired performance status were predictors of a poor prognosis after ASCT.6
The concept of chemosensitivity is related strictly to the imaging techniques and criteria used to assess the chemotherapy response. Conventional response evaluation by computed tomography (CT) scan may be inconclusive, because patients who receive chemotherapy often are left with residual masses of uncertain significance. The positron emission tomography (PET) scan can differentiate viable tumor from necrosis and fibrosis of these masses, reducing the number of unconfirmed complete responses.7 In the first-line setting, the positive predictive value of the PET scan is statistically higher than that of the CT scan.8
The predictive role of PET scans in the pretransplantation setting has been assessed in few studies, and those studies lumped together patients with Hodgkin lymphoma (HL) and T-cell or B-cell, high-grade and low-grade NHL, reporting an event-free survival (EFS) rate of 54% to 83% for PET-negative patients and 7% to 43% for PET-positive patients.9–13 Recently, Schot and coworkers validated a clinical prognostic model for predicting the outcome of ASCT based on a combination of the clinical risk score (sAA-IPI or recurring HL prognostic score) and pretransplantation PET scans.9
In the current article, we report on a single-center, prospective study of 75 consecutive patients with aggressive B-cell NHL who received second-line chemotherapy and underwent ASCT. The objective of this study was to assess the relative prognostic impact of pretransplantation PET, sAA-IPI, histology (diffuse large B-cell lymphoma [DLBCL] vs grade 3 follicular lymphoma [FL]), and previous first-line response.
MATERIALS AND METHODS
Seventy-five consecutive patients with histologically confirmed, aggressive B-cell NHL (53 patients with DLBCL and 22 patients with grade 3 FL) who received second-line and underwent ASCT at our center between September 2002 and September 2006 were included prospectively in the study. All patients provided informed consent. Patients characteristics are shown in Table 1.
Table 1. Characteristics of the 72 Patients Who Completed the Study Protocol
No. of Patients
DLBCL indicates diffuse large B-cell lymphoma; FLIII, grade 3 follicular lymphoma; LDH, lactate dehydrogenase; ECOG PS, Eastern Cooperative Oncology Group performance status; sAA-IPI, secondary age-adjusted International Prognostic Index.
No. of patients
Median age (range), y
Median follow-up (range), mo
No. of extranodal sites
Bone marrow involvement
Seventy-two patients completed the study protocol. Three patients (2 with DLBCL and 1 with grade 3 FL) did not undergo ASCT after second-line ifosfamide, etoposide, and epirubicin (IEV) chemotherapy because of an insufficient left ventricular ejection fraction (<45%).
Twenty-one patients underwent ASCT in complete response (CR) (negative PET scans and negative CT scans) after first-line chemotherapy (high-risk patients with an IPI score ≥2 and/or unfavorable gene profiling, activated B-cell [ABC]-like or natural cytotoxic [NC]-like DLBCL, enrolled in a clinical trial with front-line ASCT), 39 patients had a partial response (PR), and 12 patients were refractory either primarily or at first relapse.
All 72 patients were received 1 to 3 cycles (median 2) of the IEV regimen14 followed by peripheral blood stem cell mobilization and collection. The ASCT conditioning regimen was combined carmustine, etoposide, cytosine arabinoside, and melphalan (BEAM) in all patients. The median value of CD34-positive reinfused stem cells was 4.5 × 106/kg (range 2-15.1 × 106/kg).
All patients were evaluated by PET and CT scans according the algorithm shown in Figure 1. Figure 2 summarizes the conversion from a positive PET scan (PET+) to a negative PET scan (PET−) according to the sequential chemotherapeutic steps.
Among the 51 patients who had positive PET scans and positive CT scans before the start of second-line chemotherapy, evaluation after the IEV regimen revealed negative PET scans in 27 of 51 patients (53%; 24 patients had negative CT scans, and 3 patients had CT scans that revealed a partial response [PR]) and persistently positive PET scans in 24 patients (18 patients had positive PET scans and positive CT scans, and 6 patients had positive PET scans and negative CT scans). Globally, after second-line chemotherapy, 48 patients had negative PET scans, and 24 patients had positive PET scans.
After ASCT, 9 patients persistently had positive PET scans, whereas the remaining 63 patients (87.5%) had negative PET scans (55 patients had negative PET scans and negative CT scans, and 8 patients had negative PET scans and positive CT scans). Among the 8 patients with negative PET scans and positive CT scans, 3 patients underwent to biopsies, and histologic evaluation revealed lymphoma involvement. For the 9 patients who had positive PET scans after ASCT, 7 patients had positive PET scans and positive CT scans, and the other 2 patients had positive PET scans and negative CT scans. The median follow-up for all patients was 24.4 months (range, 6.5-47.8 months)
Positron Emission Tomography
PET scan evaluations were scored as negative or positive,15 in a dichotomous manner, for univariate and multivariate analyses. Scans were interpreted as negative when no pathologic tracer uptake was revealed by PET. Sites of known physiologic uptake, including the kidney, ureter, bladder, and musculoskeletal areas that showed symmetrical uptake, were not described in the report; and the scan was considered negative. Positive lesions were defined as areas with greater fluorine-18 (18F)-fluorodeoxyglucose (FDG) activity compared with the contralateral site or with background.
The intensity of FDG uptake was quantified by calculating the standardized uptake value (SUV). For the calculation of SUV, circular regions of interest (≥70 pixels) were drawn on transaxial images around the areas with increased FDG uptake.
We considered as background the mean FDG uptake in sites that are well known for not accumulating the tracer in fasting condition (muscle); the same criteria was used for lung lesions; whereas, for liver lesions, background was considered the mean FDG uptake in the rest of the liver parenchyma. In the presence of bilaterally increased uptake (eg, in both axillae), the absolute degree of FDG uptake (maximum SUV) was considered: When it exceeding the liver background (usually values >2.5), it was considered disease (thus, positive). Otherwise, it was regarded as inflammation and was scored as negative for the purposes of the current study.
PET (GE Discovery Tomograph; GE Medical Systems, Milwaukee, Wis) was carried out according to standard procedures (PET acquisition was performed in patients who fasted for 6 hours after intravenous injection of 5.3 megabecquerels of 18F-FDG; uptake time, 60-90 minutes). Areas of focal uptake were interpreted as unequivocally positive for lymphoma when they were localized at sites of previous disease (residual disease or relapse), within asymmetrical lymph nodes, or within lymph nodes that were unlikely to be affected by inflammation (mediastinal [except for hilar] and abdominal). All scans were reviewed by 2 experienced nuclear medicine physicians (S.F. and G.M.) who were blinded to clinical radiologic and follow-up data.
Response Criteria and Statistical Analysis
Responses were assessed according to the International Workshop Criteria.16 The objective of the current study was to evaluate the relative impact of PET, sAA-IPI score, previous first-line response, and histology (grade 3 FL vs DLBCL) on progression-free survival (PFS) and overall survival (OS) for all consecutive patients with aggressive B-cell NHL who received second-line chemotherapy and underwent ASCT between 2002 and 2006 at our institution.
The results of the study were analyzed statistically using the SPSS software program (version 8.0 for Windows; SPSS Inc, Chicago Ill). Univariate analyses of survival and of the time to progression were carried out either by using the Kaplan-Meier method17 to evaluate differences between groups or by using the log-rank test. Multivariate analysis was performed with the Cox regression model.18 A P value ≤.05 was considered statistically significant.
Univariate Analysis of Progression-Free Survival
Of 24 patients who had positive PET scans before they underwent stem cell transplantation, 15 patients obtained a CR (62.5%), 6 patients obtained a PR (25%), and 3 patients had stable disease or disease progression (12.5%). Overall, in this group, only 8 of 24 patients (33%; 7 patients with negative PET scans and negative CT scans and 1 patient with a negative PET and a positive CT scan) remained in complete remission, 15 patients had recurrent or progressive disease during follow-up, and the 1 remaining patient (a partial responder) had underwent a second ASCT and, thus, was excluded from the analysis of PFS.
Of 48 patients who had negative PET scans before they underwent stem cell transplantation, only 6 of 48 patients (12.5%) had recurrent or progressive disease during follow-up; 2 of those patients had with negative PET scans and negative CT scans, and 4 patients had negative PET scans but positive CT scans. One patient who had early disease progression before ASCT received radiation therapy and then underwent allogeneic bone marrow transplantation (he was excluded from the analysis for PFS).
The actuarial PFS rate was 87.2% for PET-negative patients compared with 34.7% for PET-positive patients (median PFS, 15.4 months; P < .00001; log-rank test, 19.74) (Fig. 3, Table 2).
Table 2. Univariate Analysis of Progression-free Survival
Responders to first-line chemotherapy (n = 60 patients) had a statistically higher PFS rate than patients who had recurrent/refractory disease (n = 12 patients; 75.8% vs 41.6%; P = .02; log-rank test, 6.03). Histology (DLBCL vs FL) and sAA-IPI scores (0-1 vs 2-3) were not significant prognostic factors according to univariate analysis (P = .45 and P = .86, respectively) (Table 2).
Multivariate Analysis of Progression-Free Survival
Multivariate analysis addressed the factors that were related significantly to PFS in univariate analysis. The multivariate setting disclosed a significant independent association with pretransplantation PET alone (P = .0003; hazard ratio, 5.91). First-line chemotherapy response lost significance in the multivariate setting (P = .12) (Table 3).
Table 3. Multivariate Analysis for Progression-free Survival
All patients were evaluable for OS. The actuarial OS rate for PET-negative patients was 93.7% (with 3 events) versus 66.6% (8 events) for the PET-positive group (P = .009; log-rank test, 8.76) (Fig. 4, Table 4).
Table 4. Univariate Analysis of Overall Survival
Actuarial OS, %
HR (95% CI)
OS indicates overall survival; HR, hazard ratio; CI, confidence interval; PET, positron emission tomography; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; sAA-IPI, secondary age-adjusted International Prognostic Index.
FL grade 3
The OS rate among first-line responders was 88.3% versus 66.6% in patients with recurrent/refractory disease (P = .05; log-rank test, 4.31). Again, histology and sAA-IPI score were not predictive (P = .23 and P = .27, respectively) (Table 4).
Multivariate Analysis of Overall Survival
Multivariate analysis addressed the factors that were related significantly to OS in univariate analysis. The multivariate setting disclosed a significant independent association with pretransplantation PET alone (P= .01; hazard ratio, 5.06). Once again, first-line chemotherapy response lost significance in the multivariate setting (P = .17) (Table 5).
Table 5. Multivariate Analysis of Overall Survival
Finally, patients with negative pretransplantation PET scans were categorized into 2 groups: The first group included 21 patients who underwent ASCT in CR after first-line chemotherapy (19 patients had negative CT scans, and 2 patients had a partial response [PR] on CT scans), and the second group included 27 patients who obtained a CR after second-line chemotherapy (24 patients had negative CT scans, and 3 patients had a PR on CT scans). We observed that PFS and OS did not differ statistically between the 2 groups (PFS, P = .66; OS, P = .82) (Fig. 5, Fig. 6).
In univariate analysis for PFS and OS, pretransplantation PET was identified as a significant prognostic factor, with PFS and OS rates of 87.2% and 93.7%, respectively, for PET-negative patients compared with 34.7% and 66.6%, respectively, for PET-positive patients. These results are in line with data from the literature on this issue. Among the other factors that we assessed, only the previous response to first-line chemotherapy significantly affected the prognosis in univariate analysis.
Multivariate analysis indicated that the only independent prognostic predictor in patients undergoing ASCT for aggressive B-cell NHL was pretransplantation PET. The prognostic value of pre-transplantation PET has been addressed by univariate analysis in several studies in which various types of lymphoma (HL, B- or T-cell NHL) were considered,9–13 but only the recent study from Shot et al9 assessed the predictive value of pretransplantation PET in a multivariate setting: In that analysis, in which 20 HL and 57 NHL of various histologies were included, the independent prognostic factors were the pretransplantation PET response and the clinical risk score (Table 6).
Table 6. Published Studies on the Role of Pretransplantation Positron Emission Tomography as a Prognostic Factor in Patients Undergoing Autologous Stem Cell Transplantation for Relapsed/Refractory Non-Hodgkin Lymphoma and Hodgkin Lymphoma
No. of Patients
EFS indicates event-free survival; NHL, non-Hodgkin lymphoma; HL, Hodgkin lymphoma; DHAP, dexamethasone, cytosine arabinoside, and cisplatinum; VIM, etoposide, ifosfamide, and methotrexate; BEAM, carmustine, etoposide, cytosine arabinoside, melphalan; PET, positron emission tomography; sAA-IPI, secondary age-adjusted International Prognostic Index; LDH, lactate dehydrogenase; +, positive; −, negative; TBI, total body irradiation; ICE, ifosfamide, carboplatin, and etoposide; ESHAP, etoposide, methylprednisolone, cytosine arabinoside, and cisplatin; BCV, carmustine, etoposide, and cyclophosphamide; EDX, endoxan; SD, stable disease; PD, progressive disease; CR, complete response; PR, partial response.
In the current study, the response to chemotherapy was assessed by PET after first-line chemotherapy, second-line chemotherapy, and ASCT in all patients. Our patients were consecutive and unselected; the relatively low number of patients with recurrent disease who underwent ASCT in the current study, compared with other case series, may have been because, in our clinical practice, we tend to offer ASCT after first-line chemotherapy for young patients who do not achieve CR instead of offering radiation therapy. Twenty-one high-risk patients who also were included in a phase 2 trial underwent ASCT in CR after first-line chemotherapy (patients with IPI scores ≥2 and ABC-like/NC-like DLBCL). The strengths of the current study are that the patients had uniform histology (B-cell aggressive NHL) and received the same second-line regimen (IEV chemotherapy) and the same ASCT conditioning regimen (BEAM).
The sensitivity of PET in indolent B-cell NHL or T-cell NHL may vary. Other published works on this issue have considered patients with various histologies at the same time, so that the relative predictive power of FDG-PET with respect to other prognostic factors (such as secondary clinical risk scores) may have been underestimated.
Our univariate analysis indicated that the sAA-IPI score (like histology) was not a predictive factor for either PFS or OS. That finding may be because the clinical risk score was validated in a series that included only patients with primary refractory and recurrent disease and not patients in PR or CR after first-line chemotherapy. The first-line response was a prognostic predictor in our univariate analysis, but it lost significance in the multivariate setting.
Chemosensitivity is a crucial prognostic factor for patients undergoing ASCT. Most clinical trials limit the enrollment of patients for ASCT to those who are proven chemosensitive to second-line chemotherapy. Considering the 48 patients in the current study who had negative pretransplantation PET scans, the PFS and OS for 27 patients who had positive PET scans after first-line chemotherapy (that became negative after second-line chemotherapy) did not differ statistically from the PFS and OS for the patients who underwent ASCT and had negative PET scans after first-line chemotherapy. Thus, a negative PET scan is an overwhelming prognostic factor regardless of when this negativity is achieved (after first- or second-line chemotherapy).
Chemosensitivity, as assessed by pretransplantation FDG-PET, clearly emerged as the main prognostic factor, overshadowing first-line response, histology, and sAA-IPI scores, for both PFS and OS. These results confirm data from the literature on the role of PET in the first-line setting for aggressive NHL: The main role of PET, compared with the other previously established prognostic factors (eg, the IPI score), was validated in 3 studies in which the status of an early PET scan was the only independent prognostic factor.19–21
The median PFS for our patients who had positive pretransplantation PET scans was approximately 15 months. For these patients, other treatment strategies should be considered; and clinical trials that aim to improve the response rate to second-line chemotherapy, or ASCT (eg, immunotherapy combined with salvage chemotherapy, radioimmunotherapy as part of the conditioning regimen), or allogeneic stem cell transplantation should be offered. The current results confirmed the leading role of pretransplantation PET as prognostic factor in a homogeneous population of patients with aggressive B-cell non-Hodgkin lymphoma.