A registry showed that 27 patients with DLBCL and residual disease had been treated with ASCT at the Asan Medical Center between March 1996 and May 2005. Residual disease of at least 1×1 cm was documented by CT scan, performed just prior to high-dose therapy. Each residual EN involvement was documented by pathologic examination. Disease status was assessed by CT of the neck, chest, abdomen, and pelvis, performed just prior to high-dose therapy. Bone marrow biopsies were repeated to evaluate the lymphoma status. All patients were staged, according to the Ann Arbor system. Response was assessed using the International Working Group criteria22). Disease status of eligible patients was classified as initial partial response, chemotherapy-sensitive relapse, and primary refractory disease. Primary refractory disease was defined as stable or progressive disease documented at restaging, immediately after the completion of induction therapy23). To be eligible for ASCT, patients had to be 15-65 years old and have biopsy-proven DLBCL; adequate hepatic, renal, pulmonary, and cardiac function; negative serology for human immunodeficiency virus; and no history of other malignancies or central nervous system involvement.

All histological specimens were reviewed by one pathologist (JH). Before 1998, histopathological results had been classified according to the International Working Formulation24), but we retrospectively reclassified such specimens according to the World Health Organization/Revised European-American Lymphoma (WHO/REAL) system25). Since 1998, all histologic diagnoses have been performed using the WHO/REAL classification.

The ASCT protocol was approved by the institutional review board of Asan Medical Center. Written informed consent was obtained from each patient.

In patients assessed as candidates for ASCT, peripheral blood progenitor cell (PBPC) mobilization and collection were performed as described previously26). PBPCs were mobilized with chemotherapy, followed by lenograstim (Neutrogin, Choongwae Ltd., Seoul, Korea), both to reduce the tumor burden and to facilitate PBPC harvesting. The chemotherapy regimens for mobilization were 4 g/m² cyclophosphamide or ESHAP (etoposide, methylprednisolone, cytarabine, cisplatin), with or without rituximab. All patients received 10 µg/kg/day lenograstim subcutaneously, starting on the day after the completion of mobilization chemotherapy and continuing until the last leukapheresis procedure. PBPCs were collected with a continuous-flow blood cell separator (Fenwal CS3000 plus, Baxter Healthcare, Deerfield, IL, USA). The total CD34⁺ cell count was monitored daily following each collection, with the target amount being 5×10⁶ CD34⁺ cells per kilogram patient weight.

The IPI consists of five risk factors: age older than 60 years, LDH greater than the upper limit of normal, ECOG performance status greater than 1, stage III or IV disease, and more than one EN site. The age-adjusted IPI (AAIPI) consists of three of the IPI risk factors: LDH, stage, and ECOG performance status13).

The BEAM15) or BEAC3) regimen was used as a high-dose conditioning regimen, and lenograstim was used to facilitate engraftment. Patients were cared for in a single room, with reverse isolation strictly maintained to prevent infectious complications. All patients were administered prophylactic antimicrobials, consisting of ciprofloxacin, fluconazole, and acyclovir. Patients received transfusions of red blood cells and platelets as clinically indicated. Generally, platelets were single donor transfusions, administered to keep the platelet count above 20,000/mm³ or to relieve clinical bleeding.

All continuous variables were analyzed using the Mann-Whitney test. Proportions were compared using the χ² test or Fisher's exact test, as appropriate. Overall survival (OS) was defined as the time from the date of transplantation until death or the last follow-up. Event-free survival (EFS) was measured from the date of ASCT until the time of a relapse, evidence of disease progression, or death, regardless of cause. The date of the first event was used in calculating EFS. Survival curves were plotted using the product-limit method, according to Kaplan and Meier, and were compared using the log-rank test. The Cox proportional hazards regression model was used for multivariate analysis. Statistical analysis was performed with SPSS (version 12.0 for Windows, SPSS Inc., Chicago, IL, USA) and the significance level was set at 0.05.

Twenty-seven patients with DLBCL were eligible for this analysis. Of them, 13 (48%) had no residual EN involvement and 14 (52%) had one or more residual EN disease at ASCT. Characteristics of the groups of patients at ASCT are listed in Table 1. Age and gender were not different between the groups. The most frequent first-line chemotherapy regimens were CHOP and rituximab plus CHOP. Dose-intense regimens, such as ESHAP, DHAP, and ICE, were administered to one patient in the no-EN group and four patients in the residual-EN group, who were found to have a significant tumor burden and poor prognostic features by their treating physicians. First-line chemotherapy regimens showed no statistically significant difference between the groups. For salvage regimens, ESHAP, DHAP, rituximab plus ESHAP, or rituximab plus DHAP were administered to the majority of patients. Three patients from the no-EN group and five patients from the residual-EN group were given aggressive regimens, such as CODOX-M/IVAC and GMALL, chosen by the treating physician. Two patients from the no-EN group and four patients from the residual-EN group were not administered second-line salvage chemotherapy because of their initial partial response (PR) status, and their PBPCs were mobilized by high-dose cyclophosphamide. Salvage chemotherapy regimens showed no difference between the groups. Radiotherapy history and disease status at ASCT were not different between the groups. Among the IPI factors, age and performance status were not significantly different between the groups. Patients showing LDH above the normal range were more frequent in the residual-EN group (86%) than in the no-EN group (39%; p=0.018). Advanced stage III/IV disease was observed in 13 patients (93%) in the residual-EN group (93%) and in one patient (15%) in the no-EN group (p<0.001). EN involvement score, categorized by the IPI system, was significantly worse in the residual-EN group than in the no-EN group (p=0.007). The IPI score was high-intermediate or high in seven patients (50%) in the residual-EN group and in one patient (8%) in the no-EN group (p=0.033). Ten patients (71%) in the residual-EN group had high-intermediate or high AAIPI scores and 12 patients (92%) in the no EN group showed low or low-intermediate risk (p=0.001).

Details of the distribution of EN involvement sites at ASCT are presented in Table 2. Eight patients had involvement of one EN site, five had involvement of two EN sites, and one had involvement of three EN sites. Of the 21 total EN sites in these 14 patients with residual EN disease at ASCT, the most frequently observed were the lung and pleura (6 sites), followed by the stomach (3 sites), head and neck (3 sites), intestine (2 sites), and liver (2 sites).

Because no ASCT-related mortality occurred in the 27 patients, all were evaluable for response. Twelve patients (44%) achieved a complete response (CR) post-ASCT, seven patients (26%) a PR, and eight patients (30%) failed. Of 19 patients transplanted in chemosensitive status, 11 (58×) achieved CR with the ASCT. However, only one patient (13×) of the eight transplanted in chemoresistant status achieved a CR (p=0.03).

The median follow-up duration for survivors was 24 months (range, 7-36). Eighteen patients (67%) have died and nine patients are alive. Twelve patients died of disease progression and six patients died of other illnesses. No occurrence of secondary neoplasm was observed. The Kaplan-Meier estimate of the proportion of patients remaining alive at 2 years following ASCT was 37%, with an estimated EFS at 2 years of 36% (Figure 1).

The univariate analysis for OS and EFS is shown in Table 3. A better outcome was associated with chemosensitive disease status at ASCT, normal LDH level, good performance, and no EN involvement at ASCT. However, based on a multivariate analysis, only normal LDH level [hazard ratio (HR) 0.05; 95& confidence interval (CI) 0.01-0.39; p=0.004 for OS; and HR 0.04; 95% CI 0.01-0.35; p=0.003 for EFS] was associated with superior survival.

Note that when categorized according to the original IPI, the 2-year OS was 37% for patients with 0-1 EN sites and 33% for those with more than 1 EN involvement at ASCT (p=0.874; Figure 2A). When we applied a modified categorization of EN involvement, grouped as negative or positive, the 2-year OS was 64% for patients with no EN site at ASCT, significantly better than the 14% 2-year OS in patients with EN involvement at ASCT (p=0.021; Figure 2B). The 2-year EFS was 36% for patients with 0-1 EN site and 33% with more than 1 EN involvement at ASCT (p=0.893; Figure 3A). With the modified categorization of EN involvement, the 2-year EFS was 62% for patients with no residual EN site at ASCT, significantly better than the 14% 2-year EFS in patients with residual EN involvement at ASCT (p=0.02; Figure 3B).