Statistical Difference In Progression-free Survival Research Articles

Analysis of outcomes in patients with supra-diaphragmatic vs infra-diaphragmatic diffuse large B cell lymphoma treated with R-CHOP therapy

The prognostic implications of infra-diaphragmatic (InD) versus supra-diaphragmatic (SpD) primary lesions in limited-stage diffuse large B-cell lymphoma (DLBCL) remains unknown. This retrospective study aimed to assess the prognostic impact of spD and InD lesions as well as presence of gastrointestinal (GI) involvements in adults with limited-stage DLBCL. We analyzed data from 178 patients with limited-stage DLBCL who were treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone therapy at 7 institutions of the Yokohama City University Hematology Group between 2003 and 2009. The median age was 63 years (range, 18–80 years). The primary sites were SpD in 109 patients, and InD in 69. No statistical differences in progression-free survival (PFS) or overall survival (OS) were observed between patients with SpD lesions and those with InD lesions. However, when patients with SpD lesions, InD lesions with (n=35), and without (n=34) GI involvement were compared, the presence of GI lesions was associated with favorable PFS. The multivariate analysis revealed that SpD or InD localization had no independent effect on PFS or OS, whereas the presence of GI lesions was correlated with favorable PFS (P=0.024, HR 0.09).

No Outcome Differences Were Observed Between Supra-Diaphragmatic and Infra-Diaphragmatic Lesions in Limited-Stage Diffuse Large B Cell Lymphoma Treated with R-CHOP Therapy, Whereas the Presence of Gastro-Intestinal Lesions Was Associated with a Favorable Prognosis

Background: Some studies evaluating differences in clinical features and outcome between supra-diaphragmatic (SpD) and infra-diaphragmatic (InD) primary lesions in Hodgkin lymphoma (HL) have been reported (Cancer 1991;68:1476-1481: Haematologica 2006;91:32-39). In regard to non-Hodgkin’s lymphoma, there exist some previous studies that report on outcomes of patients with gastrointestinal (GI) involvement (J Clin Oncol 2005;23:2797-2804; Cancer 2003;97:2462-73). However, no studies comparing outcome between SpD and InD primary lesions have been conducted for patients with diffuse large B cell lymphoma (DLBCL). Thus, we retrospectively evaluated outcome differences between SpD and InD lesions, and the prognostic impact of GI involvement in limited-stage DLBCL.Patients and Methods: We analyzed data from 178 patients with limited-stage DLBCL who were treated with rituximab plus cyclophosphamide (CPA), doxorubicin (DXR), vincristine (VCR), and prednisone (PSL; R-CHOP) therapy at 7 institutions of the Yokohama City University Hematology Group between 2003 and 2009. Patients who needed a dose reduction of more than 20% per cycle of R-CHOP therapy were excluded from the study. We classified the patients into SpD lesions group or InD lesions group according to the location of lesions. The patients in the InD lesions group were subsequently classified into InD with GI involvement group or InD without GI involvement group according to the presence of GI lesions. The impact of primary site location on patient outcome was evaluated using univariate and multivariate analyses.Results: The study cohort included 104 men and 74 women with a median age of 63 years (range, 18-80 years). All patients were categorized as Ann Arbor stage I (n = 66) or II (n = 112). The primary sites were SpD in 109 patients and InD in 69 patients. There were no significant differences in distribution between the SpD lesions group and the InD lesions group with respect to age, sex, Ann Arbor stage, the IPI score, ECOG performance status, and the presence of a bulky mass. Significantly more patients in the InD group presented with B symptoms (P = 0.003). Comparing patients in the InD lesions group presenting with (n = 35) or without (n = 34) GI involvement, resulted in similar findings with respect to clinical characteristics. The group with GI involvement consisted of 20 patients with stomach lesions (57%), 7 with small intestine lesions (20%), 7 with colon lesions (20%), and 1 with both colon and rectum lesions (3%). The median observation period for all surviving patients was 52 months (range, 3 – 94 months). The 4-year progression-free survival (PFS) and overall survival (OS) rates of all 178 patients were 84% and 91%, respectively. Twenty-six patients presented with disease progression after R-CHOP therapy. In total, 16 patients died: 14 of recurrent lymphoma, 1 of secondary malignancy (acute lymphoblastic leukemia), and 1 as a result of an accident.No statistical differences in PFS or OS were observed between patients with SpD lesions and those with InD lesions (4-year PFS rate: 87% and 81%, respectively; P = 0.31; Fig 1; 4-year OS rates: 92% and 86%, respectively; P = 0.31). When patients were classified and assessed according to primary site location, and their GI involvement status, which were SpD group, InD with GI involvement group, or InD without GI involvement group, the PFS rate was higher in SpD group (4-year PFS rates 86%), and InD with GI involvement group (4-year PFS rates 97%) than InD without GI involvement group (4-year PFS rates 67%; P = 0.036, and P = 0.003, respectively; Fig 2). However, no significant differences in regard to OS were observed among 3 groups (4-year OS rates: 93%, 97%, and 78%, respectively; P = 0.09).The multivariate analysis revealed that SpD or InD localization had no independent effect on PFS or OS. In addition, primary sites according to the presence of GI lesions were incorporated into the multivariate analysis, whereas SpD, and InD were excluded as overlapping variable. The multivariate analysis revealed that the presence of GI involvement in patients with InD lesions was an independent prognostic factor in predicting favorable PFS (P = 0.024, HR 0.09).Conclusion: No outcome differences were observed between SpD and InD lesions in limited-stage DLBCL. However, the presence of GI lesions was associated with a favorable prognosis. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Secondary Monoclonal Gammopathy of Undetermined Significance Is Frequently Associated with High Response Rate and Superior Survival in Patients with Plasma Cell Dyscrasias

Secondary monoclonal gammopathy of undetermined significance (MGUS) is a special phenomenon that occurs during the treatment of multiple myeloma (MM). The incidence, biological characteristics, and prognostic value of secondary MGUS in patients with MM remain undefined. We proceed with a retrospective systematic review of serum immunofixation electrophoresis studies performed in 438 cases of patients with plasma cell dyscrasias, including 409 cases of newly diagnosed MM and 29 cases of primary plasma cell leukemia. Secondary MGUS was more common in patients with myeloma who had undergone stem cell transplantation than in those who had not (17 [29.8%] of 57 versus 5 [1.4%] of 352, P < .001). The clinical parameters and cytogenetic characteristics in patients with or without secondary MGUS were comparable. The complete response rates in patients with or without secondary MGUS were 81.8% and 21.8% respectively (P < .01). For the cohort as a whole, secondary MGUS was associated with significantly prolonged progression-free survival (median, 52.0 months versus 22.5 months; P = .002) and overall survival (median, not reached versus 35.0 months; P < .001). The presence of secondary MGUS retained independent prognostic value with a moderate impact on overall survival (hazard ratio .128 [95% confidence interval .018 to .922]; P = .041) in the multivariate Cox regression model. However, when analysis was restricted to patients undergoing stem cell transplantation, no statistical differences in progression-free survival and overall survival were found. In conclusion, we observe that secondary MGUS was frequently observed in MM patients after transplantation and conferred a survival prolongation. The favorable survival in patients with secondary MGUS may be explained by beneficial effect from myeloablative therapy.

Efficacy of gefitinib or erlotinib in patients with squamous cell lung cancer.

IntroductionThe aim of this study was to evaluate the feasibility of epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) and EGFR mutation frequency in advanced squamous cell lung cancer based on the data from our hospital.Material and methodsThe clinical data of 102 patients with advanced squamous cell lung cancer who were admitted to Zhejiang Cancer Hospital from January 2007 to December 2011 were retrospectively analyzed. Survival analysis was evaluated by the Kaplan-Meier method. The EGFR mutations were assessed in some of the patients using the pyrosequencing assay.ResultsNine patients (8.8%) showed a partial response (PR), 28 (27.5%) achieved stable disease (SD), and 65 (63.7%) had progressive disease (PD). The disease control rate was 36.3% and the median progression-free survival (PFS) was 1.93 months (95% CI: 1.57–2.29). The PFS in patients who obtained disease control in the prior TKI was 8.63 months and 1.37 months in the disease progression cases (p < 0.001). No statistical differences in PFS were observed between gefitinib and erlotinib (2.0 months vs. 1.87 months, p = 0.76). The toxicities associated with EGFR-TKI were generally acceptable. In 74 patients with adequate specimens for molecular analysis, 4 (5.4%) had EGFR mutations (2 with deletions within exon 19 and 2 with L858R mutation in exon 21).ConclusionsThe EGFR-tyrosine kinase inhibitor seems to be a potential therapeutic option for treatment of advanced squamous cell lung cancer patients. Erlotinib and gefitinib had a similar efficacy in advanced squamous cell lung cancer. The frequency of EGFR mutation was about 5.4% in our single hospital data.

Pretreatment serum biomarkers and clinical outcome in the phase II trial of zibotentan (endothelin A receptor antagonist) in metastatic CRPC.

97 Background: Three angiogenesis-related serum biomarkers [TIMP-1 (tissue inhibitor of metalloproteinase-1), CA IX (carbonic anhydrase 9), and VEGF] and PSA were correlated with outcome in the phase II zibotentan [endothelin A (ETA) receptor inhibitor] trial in metastatic CRPC. Methods: Pretreatment serum samples were available from 206 of 312 patients enrolled in the multicenter phase II trial evaluating zibotentan in patients with metastatic CRPC and bone metastases who were pain free or mildly symptomatic for pain, with a 1:1:1 randomization to zibotentan 10mg, 15mg, or placebo. The CA IX, TIMP-1, and VEGF ELISAs were from WILEX Inc./ Oncogene Science, Cambridge, MA. Progression-free survival (PFS) and overall survival (OS) were analyzed using the Kaplan-Meier method and Cox modeling. Results: In the final efficacy report, there were no statistical differences in PFS, but a trend for OS improvement was observed in the zibotentan arms compared with placebo: 15 mg (HR 0.76; 80% CI 0.61–0.94; p= 0.103) and 10 mg (HR 0.83; 80% CI 0.67–1.02; p= 0.254) (James et al, BJUI 106L966-973, 2010). In the retrospective serum biomarker analysis (206 patients), there were no apparent differences for PFS with any of the 3 novel biomarkers in the whole population, or within treatment arms. For OS in the whole population, patients with higher serum TIMP (&gt; median) had reduced OS (17.5 mos), compared with patients who had lower serum TIMP-1 (&lt; median) (28.4 mos). Within treatment arms, patients with higher serum CA IX (&gt; median) treated with zibotentan had increased OS in both the 10 mg arm (median 28.4 mos), and the 15 mg arm (median 25.7 mos) compared with those in the placebo arm (median 15.4 mos). In a Cox model, including serum biomarkers and bisphosphonate use, for PFS only PSA was significant (p&lt; 0.0001); for OS only PSA (p&lt; 0.0001) and TIMP-1 (p&lt; 0.002) were significant. Conclusions: Higher pretreatment serum CA IX (a marker of hypoxia) may identify a cohort of metastatic CRPC patients more sensitive to zibotentan treatment, and higher serum TIMP-1 may identify patients at risk for reduced overall survival. These serum biomarkers deserve further study in larger ETA receptor inhibitor trials.

Adjuvant Chemotherapy for Stage I Clear Cell Carcinoma of the Ovary

Although postoperative adjuvant chemotherapy is generally recommended for early-stage ovarian cancer, it remains unclear whether adjuvant chemotherapy is also effective for clear cell carcinoma (CCC). Seventy-three patients with stage I CCC of the ovary who had undergone complete surgical staging formed the study population (stage IA, 20 patients; stage IC, 53 patients). Survival and multivariate analyses were retrospectively performed to determine the effectiveness of postoperative chemotherapy in these patients. Of the total (73 patients), 30 patients received adjuvant chemotherapy (stage I C-positive), whereas 43 patients did not (stage I C-negative). The 5-year progression-free survival (PFS) and 5-year overall survival (OS) rates for the stage I C-positive group were 80.1% and 87.4% compared with 73.9% and 81.7% for the stage I C-negative group. The differences in survival between these groups were not significant (PFS: P = 0.610; OS: P = 0.557). Four of the patients with stage IA CCC underwent chemotherapy, whereas the remaining 16 patients received no additional therapy. No recurrence was observed in either group. Of the patients with stage IC CCC, 26 patients underwent chemotherapy (stage IC C-positive) and 27 received no additional therapy (stage IC C-negative). There was no statistical difference in PFS and OS between the stage IC C-positive and stage IC C-negative groups. Of the patients with stage IC without artificial rupture, the 5-year PFS rates of the C-positive and C-negative patients were 69.6% and 34.6%, respectively, but the 5-year OS rates were 75.0% and 70.0%, respectively (not significant). Multivariate analyses confirmed that the presence or absence of adjuvant chemotherapy was not a prognostic indicator. The current study was performed only in fully staged patients, suggesting that postoperative adjuvant chemotherapy is not necessary for stage IA CCC patients. For patients with stage IC CCC patients, adjuvant chemotherapy suppressed recurrence, but the effectiveness was insufficient in our limited study. Further studies are required to clarify this.

Pretreatment serum biomarkers and clinical outcome in the phase II trial of zibotentan (endothelin A receptor antagonist) in metastatic CRPC.

154 Background: Three angiogenesis-related serum biomarkers [TIMP-1 (tissue inhibitor of metalloproteinase-1), CA IX (carbonic anhydrase 9), and VEGF] and PSA were correlated with outcome in the phase II zibotentan [endothelin A (ETA) receptor inhibitor] trial in metastatic CRPC. Methods: Pretreatment serum samples were available from 206 of 312 patients enrolled in the multicenter phase II trial evaluating zibotentan in patients with metastatic CRPC and bone metastases who were pain free or mildly symptomatic for pain, with a 1:1:1 randomization to zibotentan 10mg, 15mg, or placebo. The CA IX, TIMP-1, and VEGF ELISAs were from WILEX Inc./Oncogene Science, Cambridge, MA. Progression-free survival (PFS) and overall survival (OS) were analyzed using the Kaplan-Meier method and Cox modeling. Results: In the final efficacy report, there were no statistical differences in PFS, but a trend for OS improvement was observed in the zibotentan arms compared with placebo: 15 mg (HR 0.76; 80% CI 0.61–0.94; p= 0.103) and 10 mg (HR 0.83; 80% CI 0.67–1.02; p= 0.254) (James et al, BJUI 106L966-973, 2010). In the retrospective serum biomarker analysis (206 patients), there were no apparent differences for PFS with any of the 3 novel biomarkers in the whole population, or within treatment arms. For OS in the whole population, patients with higher serum TIMP (&gt; median) had reduced OS (17.5 mos), compared with patients who had lower serum TIMP-1 (&lt; median) (28.4 mos). Within treatment arms, patients with higher serum CA IX (&gt; median) treated with zibotentan had increased OS in both the 10 mg arm (median 28.4 mos), and the 15 mg arm (median 25.7 mos) compared with those in the placebo arm (median 15.4 mos). In a Cox model, including serum biomarkers and bisphosphonate use, for PFS only PSA was significant (p&lt; 0.0001); for OS only PSA (p&lt; 0.0001) and TIMP-1 (p&lt; 0.002) were significant. Conclusions: Higher pretreatment serum CA IX (a marker of hypoxia) may identify a cohort of metastatic CRPC patients more sensitive to zibotentan treatment, and higher serum TIMP-1 may identify patients at risk for reduced overall survival. These serum biomarkers deserve further study in larger ETA receptor inhibitor trials.

Evaluation of paclitaxel/carboplatin in a dose dense or weekly regimen in 66 patients with recurrent or primary metastatic cervical cancer

Background and aimsTo evaluate paclitaxel/carboplatin in a dose dense (TCdd) and weekly (TCw) regimen in recurrent or primary metastatic cervical cancer. MethodsSix courses of paclitaxel (90mg/m2) and carboplatin (area under the curve (AUC) 4) were administered on d1, d8 q3 wks in TCdd. Eighteen courses of paclitaxel (60mg/m2) and carboplatin (AUC 2.7) were administered weekly in TCw. Response rates were determined using Response Evaluation Criteria in Solid Tumours (RECIST) 1.1 criteria. Toxicity was evaluated according to the National Cancer Institute-Common Toxicity Criteria (NCI-CTC) Criteria. ResultsSixty-six patients were included (44 TCdd and 22 TCw). TCdd and TCw were administered as first-line chemotherapy in 48% and 41%, second-line in 43% and 18%, and third/fourth-line in 9% and 41%, respectively. Response (confirmed or unconfirmed) was observed in 58% and 36% for TCdd and TCw, respectively. As first-line, the response rates for TCdd and TCw were the same (55%). As second or more line, the response rates for TCdd and TCw were 61% and 29%, respectively. In patients, receiving TCdd as first-line systemic treatment, median overall survival (OS) and progression-free survival (PFS) was 10 and 5months. As first-line, the median OS for TCw also was 10months (median PFS not reached). There was no statistical difference in PFS or OS between patients treated with TCw or TCdd. Grade 3–4 toxicity was mostly bone-marrow related and was more common with TCdd. Febrile neutropenia was observed in 0% and 12% of the patients treated with TCw and TCdd, respectively. ConclusionsCombination of paclitaxel and carboplatin in a dose dense regimen or weekly regimen resulted in favourable response rate and toxicity profile compared with cisplatin-based combination regimens. TCdd appears to be more toxic than TCw, but resulted in higher response rates than TCw in patients with recurrent metastatic cervical cancer who received prior chemotherapy.

Inclusion of Total BODY Computed Tomography (TB-CT) SCANS In the INITIAL WORK-up of Binet STAGE A CHRONIC LYMPHOCYTIC LEUKEMIA (CLL) PATIENTS On CLINICAL Grounds: PRELIMINARY RESULTS of the Prospective, MULTICENTER O-CLL1- GISL STUDY

Abstract 2837Rai and Binet staging systems are not devoid of some limitations, including the lack of evaluation of thoracic and abdominal lymphadenopathies. The widely-used IWCLL guidelines do not incorporate use of TB-CT scan in the diagnostic algorithm. In the present study, we investigated whether TB-CT scans could up-stage Binet stage A CLL patients included in the prospective multicenter O-CLL01 GISL study (clinicaltrial.gov ID: NCT00917549), and whether this subgroup presented differences in prognostic markers and in progression-free survival (PFS). To date, 454 patients have been enrolled and TB-CT scans were available in 238 patients. The median age was 60 years (range, 33–71) and 136 (57%) were male. According to Rai, 180 patients were at low risk (stage 0) and 58 at intermediate risk (stages I-II). b2-microglobulin was elevated in 35.5% of cases. Seventy-eight patients (32%) were IgVH unmutated, 108 patients (45%) had a high ZAP-70 expression, 45 patients (19%) were CD38 positive (>30%). FISH data were available in 226/238 cases; the most frequent abnormality was del(13)(q14) (105 pts, 46.5%), followed by trisomy 12 (24 pts, 10.6%), del(11q22.3) (13 pts 5.5%), del(17p13) (4 pts 1.8%) while 80 cases (35.4%) cytogenetics were normal. Cytogenetic abnormalities were clustered in 3 risk groups [i.e. low (del(13q14) and normal), intermediate (trisomy 12) and high risk (del(11q22) and del(17p13)]. Two hundred six out of 238 patients had a minimum follow-up of 6 months and were evaluable for PFS. Considering TB-CT scan, 54 out of 238 analyzed (22.7%) patients converted into Binet stage B. Notably, 63% were male, b2-microglobulin was elevated in 50% of cases, 42.6% were IgVH unmutated, 48.1% had a high ZAP-70 expression, 27.8% were CD38 positive, and 17.6% showed a high-risk FISH. Binet B patients showed a statistically higher rate of cases with high risk cytogenetic abnormalities than Binet A patients (17.6% vs 4.6%; p=0.032). While, no statistically different distribution of gender, age, B2-microglobulin, IgVH mutational status, CD38 or ZAP-70 expression were observed between the two subgroups. After a median follow-up of 24 months 46/206 (22%) evaluable cases showed disease progression. Binet B patients showed a PFS significantly shorter than those with a normal TB-CT (2-years PFS probability, 85.6% vs 68.5%; p<0.0001). According to the Rai classification 102/180 (56.7%) low risk patients were re-defined as intermediate risk with the integration of TB-CT scan. This subset of patients showed a statistically higher rate of cases with elevated ZAP-70 (51.5% vs 35.9%; p=0.049) and CD38 (22.5% vs 10.3%; p=0.045) than patients at low risk. After a median follow-up of 25 months, 23/154 (15%) of evaluable cases showed disease progression. Patients with an intermediate risk Rai stage showed a PFS significantly shorter than those with a low risk (2-years PFS probability, 82% vs 96%; p=0.002). In this setting 70 cases met the diagnostic criteria of monoclonal B-lymphocytosis (<5 × 109/L B- lymphocytes in the blood). With the integration of TB-CT scan 30/70 (42.9%) monoclonal B-lymphocytosis patients were re-defined as intermediate risk according the Rai classification. No statistically different distribution of clinical and biological parameters were observed between cases who remained in the low risk stage and those who became at intermediate risk. After a median follow-up of 28 months 4/57 cases evaluable for PFS showed a disease progression (2 cases for each subgroup). Considering low risk Rai stage, no statistical difference in PFS was observed among nonCT-upstaged MBL, CT-upstaged MBL, nonCT-upstaged Rai 0, while CT-upstaged Rai 0 cases showed a statistically shorter PFS (p<.0001) than the other groups (Figure 1). Finally, TB-CT scan allowed the early identification of a second neoplasia in 2 cases (lung cancer 1 pt, renal cell carcinoma 1 pt). Our preliminary data indicate that the integration of TB-CT scans in the clinical staging allows for an effective clinical discrimination of Binet A CLL cases in approximately 23% of cases at more advanced stages, predicting a worse clinical outcome. However, the use of TB-CT scanning for upstaging is not beneficial for predicting PFS in MBL cases. A longer follow-up will demonstrate whether the inclusion of TB-CT scan in the initial work-up of patients with early-stage CLL will provide clinically relevant prognostic information. [Display omitted] Disclosures:Di Raimondo:celgene: Honoraria. Foà:Bristol-Myers Squibb: Consultancy, Membership on an entity’s Board of Directors or advisory committees.

Impact of androgen receptor cytosine-adenine-guanine polymorphisms on clinical outcome in BRCA mutation-associated epithelial ovarian cancers

Purpose Androgen signaling may function in the pathobiology of epithelial ovarian cancers associated with mutations in the BRCA1/ 2 genes. Androgen receptor (AR) activity correlates inversely with length of a polymorphic cytosine-adenine-guanine (CAG) repeat in exon 1. We hypothesized that AR CAG allele length is a modifier of clinical outcome in BRCA1/ 2 mutation positive women with ovarian cancer. Experimental design We identified BRCA1/ 2 ovarian cancer patients with banked serum from which we PCR amplified the CAG repeat region. We abstracted clinical and survival data, and examined CAG repeat length < 19 as a short AR allelotype. We calculated a sample size of 60 patients to determine a 24-month difference in survival. Result In 62 patients, 43 (69%) had BRCA1 mutations and 19 (31%) had BRCA2 mutations. Fifteen (24%) were found to have a short AR allelotype. Patients with a short AR did not demonstrate statistical differences in progression-free survival (43 months vs. 28 months for long AR) or overall survival (78 months vs. 142 months for long AR). In patients with BRCA2 mutations alone, a short AR correlated with decreased overall survival (31 months) compared to 126 months for those with a long AR ( p = 0.01). Conclusions AR allelotype length did not correlate with survival in this statistically representative cohort of patients with BRCA1/2 mutations. Potential associations between short AR and outcome in BRCA2-associated ovarian cancers remain to be determined.

Long-Term Follow-Up of High-Dose Chemotherapy Followed by Autologous Stem Cell Transplantation in Peripheral T-Cell Lymphomas at Diagnosis.

Background: Limited experience is available on the feasibility and long-term efficacy of high-dose (HD) chemotherapy followed by autologous stem cell transplantation (ASCT) in patients with peripheral T-cell lymphoma (PTCL) at diagnosis.Patients and methods: Between January 1993 and December 2003, 54 patients (median age 44 years, range 20-61; 66% male) with PTCL with high-risk features were candidate to HD chemotherapy and ASCT as a first-line treatment. Histologic PTCL subtypes included: PTCL not otherwise specified (n=25), ALCL alk-positive (n=17), angioimmunoblastic (n=9), others subtypes (n=4). 44 patients (81%) presented with stage III or IV, 58% had elevated s-LDH and 73% presented with two or more adverse risk factors according to the International Prognostic Index (IPI). Treatment program consisted of: 1) induction and intensification phase with two APO and two DHAP courses than either the sequential administration at 15-20 days intervals of HD-cyclophosphamide, HD-cytarabine and HD-etoposide (n=26) or 8 weeks of MACOP-B followed by HD-cytarabine and mitoxantrone (MAD) (n=29); 2) consolidation phase with HD-mitoxantrone and HD-melphalan (n=12) or BEAM (n=27) followed by ASCT; 3) radiotherapy on bulky sites.Results: Thirty-nine patients (71%) attained a clinical response (n=32 CR; n=7 PR) and were autografted while 16 patients (29%) did not for the following reasons: progressive disease (n=10), toxicity (n=2), toxic death concomitant with refractory disease (n=2), refusal (n=2). At a median follow-up of 66 months (range, 10-127 months), 27 patients (49%) are alive (25 in CR) and 28 died (51%): 21 for disease progression, 5 for treatment-related-toxicity and 2 for other reasons (n=1 other disease, n=1 car accident). The estimated 2-year TRM was 7 %. Using an intention-to-treat-analysis, the estimated 5 year overall survival (OS) and progression-free survival (PFS) projections were 52% (95% CI, 38% to 66%) and 50% (95% CI, 36% to 64%) respectively. Patients autografted with ALCL alk-pos subtype had a better prognosis compared to other subtypes with 5 year OS of 84% versus 52% (P<0.03) and 5 year PFS of 84% versus 36% (P<0.01), respectively. No statistical difference in PFS or OS was observed between patients treated with either sequential HD therapy or MACOP-B/MAD. Univariate analysis revealed no significant differences in OS and PFS for patients with IPI score <2 versus ≥ 2.Conclusions: Our study indicates that up-front HD and ASCT: 1) induces a high rate of long term remission in patients with ALCLalk-pos subtype; 2) is of limited benefit in patients with non-ALCLalk-pos subtype whose expected long-term progression free survival was lower than 40%. In this subgroup other treatment options including consolidation with allogeneic transplantation from HLA identical donor, should be evaluated.