• T-cell non-Hodgkin lymphoma;
  • chemotherapy;
  • prognosis factor


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. References

The important prognostic factors were evaluated for T-cell non-Hodgkin lymphoma (NHL) patients in a prospective study using the CEOP-B protocol [a modified cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP)-like regimen that uses epirubicin instead of doxorubicin with the addition of bleomycin]. Fifty-two patients were enrolled in the study. The overall response rate was 63·5%. The median progression-free survival (PFS) and median overall survival (OS) was 18·0 and 39·5 months respectively. The most common toxicity was neutropenia. The factors related to poor outcome were a high International Prognostic Index (IPI) and a high ‘B’ score (bone marrow involvement, B symptoms, bulky disease). We developed a new prognostic model, namely the Prognostic Group for T cell NHL (PGT) that included four groups: PGT1 (low IPI/low B score), PGT2 (low IPI/high B score), PGT3 (high IPI/Low B score) and PGT4 (high IPI/Low B score). OS and PFS (not reached, 48 months) in the PGT1 group were significantly longer than those (11·5 and 4·8 months) in PGT2. The same result was observed in the PGT3 and PGT4 groups. The CEOP-B regimen was moderately active and tolerable for T-cell NHL patients, and the PGT system might be useful for the prediction of long-term survival of T-cell NHL patients.

T-cell non-Hodgkin lymphoma (NHL) is included in a group of heterogeneous lymphoid malignancies derived from T cells (Harris et al, 1994) that usually develops in middle-aged or elderly patients. At diagnosis it usually shows a higher incidence of disseminated disease with systemic symptoms, bone marrow (BM) invasion and extranodal involvement compared with B-cell NHL (Ascani et al, 1997). Therefore, the clinical features associated with T-cell NHL are generally aggressive with poor treatment response and prognosis (Coiffier et al, 1990). Most recently, the cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) chemotherapy regimen has been the standard treatment modality for T-cell NHL; however, treatment outcome after CHOP for T-cell NHL is worse than that for B-cell NHL. Although second or third generation chemotherapy, or high-dose chemotherapy with autologous stem cell transplantation, has been available, the prognosis for patients with T-cell NHL remains poor (Gisselbrecht et al, 1998; Rodriguez et al, 2001). Therefore, the optimal treatment approach for patients with T-cell NHL has not been found. This may be because of the rarity of the disease, the variable clinical course and the paucity of clinical trials (Haioun et al, 1997). In addition, guidelines for evaluating the prognosis and therapeutic decision-making for patients with T-NHL have not been completely determined. Some reports have shown that the International Prognostic Index (IPI) system, which is well established and widely used for the prediction of treatment outcomes in B-cell NHL, is a valuable prognostic tool for patients with T-cell NHL (Case, 1983; Sampi et al, 1983; Gottlieb et al, 1990). However, the role of the IPI system in T-cell NHL has not been clearly defined and only a few reports have evaluated the prognostic factors associated with T-cell NHL (Reiser et al, 2002; Rudiger et al, 2002; Escalon et al, 2005; Sonnen et al, 2005). Moreover, the available reports have all been retrospective studies on T-cell NHL patients treated with diverse chemotherapeutic regimens, and evaluated by limited prognostic factors. A prospective, well-designed study using a uniform chemotherapeutic regimen is needed for analysis of the factors suspected to influence the prognosis of T-cell NHL. The evaluation of treatment outcome and prognostic factors in such studies, may help improve survival in patients with T-cell NHL. Therefore, we performed a prospective study of patients treated with the CEOP-B protocol, a modified CHOP-like regimen using epirubicin instead of doxorubicin and with the addition of bleomycin to improve treatment outcomes for the patients with T-cell NHL; in addition we evaluated important prognostic factors predicting long-term survival.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. References

Patient eligibility

T-cell NHL patients who were initially diagnosed according to the histological revised European-American classification of lymphoid neoplasms (REAL classification), with at least one measurable lesion, were enrolled in this study between January 1999 and April 2005 at the Korea University Medical Centre (KUMC). Other eligibility criteria were: performance status (PS) of <2 on the Eastern Cooperative Oncology Group (ECOG) scale, life expectancy of >2 months, serum creatinine <176·8 μmol/l, serum total bilirubin <51·3 μmol/l, adequate BM function with an absolute neutrophil count >1·0 × 109/l and a platelet count >60 × 109/l. Contraindications to enrolment were: an active infectious process, active heart disease, central nervous system involvement or any concomitant second primary cancer. Written informed consent was obtained from all patients enrolled in this study and the institutional review board at KUMC approved the study protocol.

Patients were staged according to the Ann Arbor staging system, which included a thorough history, physical examination, routine blood and urine analysis, standard chest X-rays, computerized tomography (CT) of the abdomen and pelvis, and bilateral BM aspiration and biopsy. The clinical data, including: age, sex, complete blood count, erythrocyte sedimentation rate (ESR), lactate dehydrogenase (LDH) level and serum β2-microglobulin level, was reviewed. The following assessments were recorded: Ann Arbor staging, IPI, number and sites of extranodal disease, BM involvement, systemic (B) symptoms, bulky disease, PS, date of diagnosis, response to therapy, date of assessment of response, date of relapse, date of last follow up, status (alive or deceased) and if deceased, date and cause of death. Bulky disease was defined as a mass with the largest diameter at 10 cm. Systemic symptoms were defined, according to Ann Arbor criteria, as recurrent fever (>38°C), night sweats or the loss of more than 10% of body weight.


On day 1, cyclophosphamide 750 mg/m2, epirubicin 60 mg/m2, vincristine 1·4 mg/m2 and bleomycin 5 mg/m2 were intravenously infused, and prednisolone 100 mg was orally administered for 5 d. CEOP-B chemotherapy was repeated every 21 d for up to six cycles. Extended or involved field radiotherapy was applied to any residual mass; this was carried out in 14 patients starting 4 weeks after completion of chemotherapy. Custom cerrobend blocks were used to shield the areas that were not considered at risk of containing disease, and to block critical structures to avoid exceeding tissue tolerance (e.g. spinal cord). Radiation therapy (RT) was performed with a linear accelerator using 4 MV photons (CLINAC 1800, Varian, Palo Alto, USA). The doses used for RT varied from 30 to 40 Gy.

A variety of different salvage chemotherapy regimens were used for patients who relapsed and progressed after CEOP-B treatment: etoposide, solumedrol, cytarabine (ara-c), cisplatin (ESHAP), ifosfamide, etoposide, ara-c, methotrexate (IVAM) and BCNU, etoposide, ara-c, melphalan (BEAM) were used for further treatment.

Toxicity evaluation and dosage modification

Toxicity was evaluated before each treatment cycle according to the National Cancer Institute Common Toxicity Criteria (NCI-CTC) version 3·0. In cases of grade 3 or more of haematological toxicity, cyclophosphamide and epirubicin were initiated at a reduced dose (reduced by 25%). In cases with grade 4 neutropenia, granulocyte colony-stimulating factor was administered intravenously until haematological values were recovered. When severe toxicity was diagnosed at the time of scheduled treatment, the treatment was postponed by 1 week until the toxicity resolved. If drug administration was delayed for more than 2 weeks, the patient was withdrawn from the study.

Response evaluation

Response to treatment was assessed 6 weeks after the end of induction chemotherapy by CT scans. The definition for treatment response was based on those from the International Union Against Cancer (UICC; Schmits & Trumper, 2001). Complete remission (CR) was defined as the disappearance of all clinical evidence of disease, determined by two observations not <4 weeks apart. Partial remission (PR) was defined as a more than 50% reduction, for at least 1 month, at the largest dimension of each measurable lesion. Progressive disease/no response (NR) was defined as a 25% increase in at least one lesion, or the appearance of a new lesion. Time to progression was measured as the time from the first day of therapy until death from disease, appearance of new lesions, or a >25% increase in the diameter of the tumour lesions. Time to progression and survival time were summarised using Kaplan–Meier product limit curves.

Statistical analysis

All statistical analyses were performed using the Statistical Package for the Social Sciences (spss) software, version 10·0 (SPSS Inc., Chicago, IL, USA). Overall survival (OS) was measured from the date of start of treatment until death by any cause or the last follow-up date. Progression-free survival (PFS) for the patients who achieved CR or PR was calculated from the date of remission to the date of progression or death. The survival probabilities were calculated according to the product-limit method of Kaplan–Meier. A univariate association, between individual clinical features and OS or PFS, was determined by the log-rank test. Prognostic factors including: histological subtype, age, gender, PS, the presence of B symptoms, serum LDH level, serum β2-microglobulin level, number of extranodal involvement sites, presence of bulky disease and IPI were analysed. Factors independently associated with OS and PFS were identified by a multivariate analysis using a Cox proportional hazards regression model. Statistical significance was defined as P < 0·05.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. References

Characteristics of patients

The characteristics of the 52 patients are summarised in Table I. The median age of our patient population was 47 years (range: 17–78), with a male to female ratio of 2:1. Half of the patients had advanced stage diseases at presentation: 31% had stage I disease, 19% had stage II disease, 21% had stage III disease and 29% had stage IV disease. The most common histopathological type was peripheral T-cell lymphoma, unspecified (PTCL-u) followed by angioimmunoblastic T-cell lymphoma, anaplastic large cell lymphoma, subcutaneous panniculitis-like T-cell lymphoma and lymphoblastic lymphoma, in decreasing order of frequency (Table I). B symptoms were noted in 21 patients (40%). Commonly involved extranodal sites were BM, skin and nasal/paranasal sinuses (Table II).

Table I.   Characteristics of the patients.
CharacteristicNumber of patients (%)
  1. * The upper limit of normal for LDH at our institution was 450.

  2. LDH, lactate dehydrogenase; PTCL-u, peripheral T-cell lymphoma-unspecified; IPI, International Prognostic Index.

Male/female gender34 (65)/18 (35)
Age[years; median (range)]47 (17–78)
 >6013 (25)
 <6039 (75)
Performance status
 0/139 (75)
 2/313 (25)
B symptoms21 (40)
 I16 (30· 8)
 II9 (17· 3)
 III12 (23· 1)
 IV15 (28· 8)
 PTCL-u41 (79)
 Angioimmunoblasticlymphoma5 (10)
 Anaplasticlarge cell lymphoma4 (8)
 Subcutaneouspanniculitis-like T-cell lymphoma1 (2)
 Lymphoblasticlymphoma1 (2)
Extranodal involvement
 121 (40)
 ≥25 (10)
 Bonemarrow involvement10 (19)
Bulky disease
 Present11 (21)
 Absent41 (79)
 LDH > normal*11 (55)
IPI (Number of risk factors)
 Low (0–1)20 (38)
 Lowintermediate (2)14 (27)
 Highintermediate (3)13 (25)
 High (4–5)5 (10)
Table II.   Extranodal involvement sites.
Extranodal siteNumber of patients (%)
Bone marrow10 (19· 2)
Skin/subcutaneous7 (13· 4)
Paranasal sinus4 (7· 7)
Gastrointestinal3 (5· 8)
Lung3 (5· 8)
Pleura1 (1· 9)
Peritoneum1 (1· 9)
Conjuctiva1 (1· 9)
Kidney1 (1· 9)
Liver1 (1· 9)
Parotid gland1 (1· 9)
Bone1 (1· 9)

Response to treatment

Forty-six of 52 patients were evaluated for their response to treatment. One patient died due to severe sepsis after the first cycle of chemotherapy. Five patients were lost to follow up and excluded from the evaluation.

Complete remission was achieved in 17·3% and PR in 46·2% of patients based on an intention-to-treat analysis. The overall response rate (RR) was 63·5%. In patients with PTCL-u, the overall RR was 55% (complete response, 17%; Table III).

Table III.   Response to CEOP-B by histological diagnosis
CharacteristicsNumber of Patients CR (%) PR (%)CR  + PR (%)
  1. CR, complete remission; PR, partial remission.

Peripheral T-cell lymphoma417 (17· 1)18 (43· 9)26 (61)
Angioimmunoblastic lymphoma52 (40)2 (40)4 (80)
Anaplastic large cell lymphoma40 (0)3 (75)3 (75)
Subcutaneous panniculitis-like T-cell lymphoma10 (0)0 (0)0 (0)
Lymphoblastic lymphoma10 (0)1 (100) 
Total529 (17· 3)24 (46· 2)33 (63· 5)

The univariate analysis showed that BM involvement was related to a poor response to therapy (P < 0·05). Other factors, such as the histology subtype, LDH, PS and systemic (B) symptoms were not significantly related to response to therapy (Table IV). Recurrence or progression of disease developed in 17 of 32 patients who initially responded to chemotherapy.

Table IV.   Response and survival by clinical parameters of patients treated by CEOP-B.
Characteristic CR, PR (%) P-value5 years PFS (%) P-value5years OS (%) P-value
  1. LDH, lactate dehydrogenase; β2-MG, β2-microglobulin;IPI, International Prognostic Index, CR, complete remission; PR, partial remission; PFS, progression free survival; OS, overall survival.

Age (years)
 <6064 (25/39)0· 40320· 1510· 66
 >6061 (8/13) 0 0 
Performance status
 0–167 (26/39)0· 4340· 11560·1
 >154 (7/13) 0 0 
B symptom
 Absent67 (20/30)0· 48370· 03590· 03
 Present57 (12/21) 16 35 
 I/II70 (17/27)0· 28340· 04610· 04
 III/IV56 (14/25) 0 <37 
BM involvement
 Absent71 (30/42)0· 01320· 001550· 002
 Present30 (3/10) 0 <25 
Bulky disease
 Absent63 (26/41)0· 98360· 009520· 01
 Present64 (7/11) 0 39 
Extranodal involve
 0–162 (29/47)0· 49320· 12550· 01
 280 (4/5) 0 0 
LDH (IU/l)
 <45069 (22/32)0· 3420· 01530· 03
 >45055 (11/20) 9 45 
β2-MG (mg/l)
 <2· 472 (21/29)0· 49370· 94520· 45
 >2· 463 (12/19) 29 44 
 0–175 (15/20)0· 3452<0· 05770· 05
 257 (8/14) <7  
 361 (8/13) 0 0 
 4–540 (2/5) 0 0 


Of the 254 cycles of treatment (median six cycles, range: 1–6), 172 cycles could be evaluated for toxicity. There were two treatment-related deaths. One patient died from acute pancreatitis on day 14 of chemotherapy. Another patient died subsequently after an intracranial haemorrhage, which developed with severe thrombocytopenia. The treatment caused significant haematological toxicity, and grades 3 and 4 neutropenia and thrombocytopenia occurred in 54% and 11·5% of patients respectively. Febrile neutropenia occurred in seven patients (12%). Theses patients required hospitalisation for the administration of intravenous antibiotics. The most common non-haematological toxicity was nausea and vomiting, which was grades 3 and 4 in four patients (7%). No significant cardiac or pulmonary toxicity was observed. Table V shows the toxicities in 52 patients who could be assessed for adverse events after >1 course of treatment.

Table V.   Toxicity of CEOP-B.
Toxicity (Grades 3 and 4)Number of patients (%)
Granulocytopenia28 (54)
Febrile neutropenia7 (13· 5)
Thrombocytopenia6 (11· 5)
Nausea/vomiting4 (7· 7)
Renal insufficiency2 (3· 8)
Acute pancreatitis1 (1· 9)

Survival and prognostic factors

The median overall and PFS were 39·5 months (range: 0–77+) and 18·0 months respectively. The estimated 5-year survival rate was 49% with median of 13 months follow up (Fig 1). Univariate analysis showed that the presence of B symptoms, BM involvement, two or more sites of extranodal involvement, bulky disease, elevated LDH, advanced disease stage (III or IV), response to treatment and high intermediate/high-risk group by IPI, were all related to OS and PFS (Table IV). No significant differences were found according to the histological subtype. The multivariate analysis showed that high intermediate/high risk by IPI, bulky disease, B symptoms and BM involvement were related to poor OS (Table VI).


Figure 1.  Overall and progression-free survival (PFS) in patients with T-cell non-Hodgkin lymphoma (NHL) treated by CEOP-B.

Download figure to PowerPoint

Table VI.   Clinical parameters influencing survival in multivariate analysis.
ParametersRelative risk 95% CI P-value
  1. IPI, International Prognostic Index; BM, bone marrow.

IPI (high/high intermediate)6· 641· 646–26· 830· 008
Bulky disease1· 3251· 315–1· 8740· 03
B symptom1· 5311· 293–2· 1430· 021
BM involvement1· 6821· 321–2· 3420· 035

In this cohort of 52 patients, the IPI predictive model, developed for aggressive B-cell NHL, identified four categories of patients, each with a different prognosis: low in 20 patients, low intermediate in 14 patients, high intermediate in 13 patients and high in 5 patients. The median time for OS and median PFS has not yet been reached; for the low-risk group they were 19 months and 7·5 months, respectively, in the low intermediate risk group, 11·5 months and 7·0 months in the high intermediate risk group 2·0 months and 2·0 months in the high risk group (P < 0·0001, log-rank test; Fig 2). When the patients in this study were divided into a simplified 2-class subgroup, high or low, according to the IPI score (0–2/3–5), a more favourable OS and PFS were found in the low-risk group (Fig 3). Because the relative risk associated with each of the three B factors (BM involvement, B symptoms and bulky disease) was comparable, we constructed a new prognostic model by combining these prognostic variables as follows: low IPI group, IPI score 0–2; high IPI group, IPI score 3–5; low B score, 0–1 B factor; high B score, 2–3 B factors. This new model for prognosis of T-cell NHL identifies four groups of patients: Prognostic Group for T-cell NHL (PGT)1, low IPI/low B; PGT2, low IPI/high B; PGT3, high IPI/low B; PGT4, high IPI/high B, resulting in different outcomes (Fig 4, Table VII). For the 25 patients in PGT1, the median OS has not yet been reached and the PFS was 44 months; for the nine patients in PGT2, OS and PFS was 18·5 months and 5·6 months, respectively, for the 10 patients in PGT3, OS and PFS was 15·5 months and 6·0 months, respectively, and for the eight patients in PGT4, 11·5 months and 4·8 months. The differences between OS and PFS among the PGT groups were statistically significant (P = 0·0007, 0·0001 respectively).


Figure 2.  Overall and progression-free survival (PFS) according to the International Prognostic Index (IPI).

Download figure to PowerPoint


Figure 3.  Overall survival and progression-free survival (PFS) according to the simplified International Prognostic Index (IPI); low- and high-risk group.

Download figure to PowerPoint


Figure 4.  Overall and progression-free survival (PFS) according to the Prognostic Group for T-cell non-Hodgkin lymphoma (NHL) (PGT).

Download figure to PowerPoint

Table VII.   New prognostic model for T- cell NHL
PGT group1234 
  1. low-risk IPI 0–2, high-risk IPI 3–4 ; PGT, Prognostic Group for T-cell NHL;

  2. B score, scoring according to the presence of bone marrow involvement, Bulky disease and B.

  3. symptom, OS, overall survival; PFS, progression-free survival; M, months; NHL, non-Hodgkin lymphoma; IPI, International Prognostic Index.

IPI (n)Low and low intermediate (34)High and high intermediate (18)P-value
B score (n)B 0,1 (25)B 2,3 (9)B 0,1 (10)B 2,3 (8) 
Median OS (range)Not reached18· 5 M (10· 9-26· 0)15· 5 M (5· 7-14· 6)11· 5 M (4· 23-11· 3)0 · 0007
Median PFS (range)44 M (21–66)5· 6 M (3· 5–7· 6)6· 0 M (0· 0–13· 1)4· 8 M (0· 94–8· 6)0 · 0001


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. References

The CEOP-B combination chemotherapy comprises myelosuppressive (cyclophosphamide, epirubicine) and non-myelosuppressive (vincristine and bleomycin) drugs with a proven effectiveness in patients with NHL (Case, 1983; Sampi et al, 1983; Gottlieb et al, 1990). The potential antitumour synergy of these drugs with non-overlapping toxicity profiles, as well as the absence of cross-resistance, makes this combination an attractive frontline regimen for the treatment of patients with NHL. Moreover, epirubicin is an anthracycline known to have less severe cardiotoxicity than doxorubicin; therefore, addition of this drug can reduce cardiotoxicity without reducing regimen efficacy. In this prospective study, the CEOP-B regimen achieved an encouraging overall RR of 63·5%; however, the CR rate of 17·3% was disappointing. Our results showed that the combination of these five drugs, which comprise the CEOP-B regimen, provided a moderately active and tolerable regimen for patients with T-cell NHL, and compared favourably with the outcomes reported from other studies. Huang et al (2004) reported on the clinical outcome of 106 patients with peripheral T-cell lymphoma treated with the standard CHOP regimen. The initial RR to chemotherapy was 69·5%, and the CR rate was 44·1% (Huang et al, 2004). Peng et al (2004) treated 21 patients who had PTCL with the EPOCH regimen (doxorubicin/epirubicin, vincristine, etoposide, cyclophosphamide, prednisone); the investigators reported a RR of 85% (17 of 20), and a CR rate of 50% (10 of 20; Peng et al, 2004). Compared with the studies of Huang et al (2004) and Peng et al (2004), the CR rate in our study was low. This might be due to the difference in study design. However, the present study was prospective whereas Huang et al (2004) and Peng et al (2004) reported retrospectively. Another possible cause might be the high frequency of neutropenia, more than grade 3 (54%), resulting in the dose reduction of chemotherapeutic agents in our study. Therefore, further trials to minimise and overcome severe neutropenia using our protocol are needed. Despite of the relatively low incidence of CR in our study, the 5-year OS of the patients enrolled (49%) was similar to prior reports on T-cell NHL (Case, 1983; Sampi et al, 1983; Gottlieb et al, 1990). Differences in study design as well as patient characteristics, such as comparatively young median age (47 years) and good PS (PS 0–1, 64%), may explain these findings. At the time of this report, 20 of the 52 patients had died due to disease progression or relapse. The risk of death was higher in the first 2 years after diagnosis and decreased over time until a plateau was reached. The obtained survival curves suggest that the therapy used in this study changed the course of disease when the optimal treatment strategy could be used.

Another aim of this study was to determine the value of pre-existing prognostic models, and to propose a new improved system to evaluate the prognosis of patients with T-cell NHL. Clinical prognostic factors for T-cell NHL have been studied by several investigators, and the IPI system, a universal model for B-cell lymphoma patients, has been found to have significant value for the prediction of the prognosis of patients with T-cell NHL (Sampi et al, 1983; Gottlieb et al, 1990; Ansell et al, 1997; Arrowsmith et al, 2003; Escalon et al, 2005; Sonnen et al, 2005). In our study, the IPI system was also valuable for predicting the prognosis of patients with T-cell NHL; the survival of patients with a low IPI score (0–1) was significantly longer than for patients with a high IPI score (2–5). However, the IPI system did not accurately reflect the prognosis of T-cell NHL as it does in patients with B-cell NHL. Therefore, there have been trials to find an improved system for predicting prognosis that can supplement the IPI system in patients with T-cell NHL. Lopez-Guillermo et al (1998) reported that the presence of B symptoms, histological subgroup (ACLC versus other PTCL), and the IPI (low versus high) were independent predictors of prognosis for T-cell NHL using a multivariate analysis (Lopez-Guillermo et al, 1998). Gallamini et al (2004) analysed 385 PTCL-u patients and reported that some of the IPI components (age, PS, LDH level) and BM involvement were valuable for the prediction of prognosis. Sonnen et al (2005) analysed 125 T-cell lymphoma patients and reported that the presence of B symptoms was independently related to a poor prognosis in patients with T-cell NHL. Escalon et al (2005) analysed 135 T-cell lymphoma patients and reported that the presence of bulky disease was independently related to a poor prognosis. However, these reports were all retrospective studies, with diverse chemotherapeutic regimens; in addition, the evaluated prognostic factors were somewhat different in each study. For example, Gallamini et al (2004) and Escalon et al (2005) did not analyse B symptoms, and Sonnen et al (2005) did not include analysis of bulky disease. We performed a prospective study to avoid the problems associated with retrospective studies, such as differences in the histological subtype accrued and chemotherapeutic regimens, aiming to evaluate all possible factors suspected to influence the prognosis of T-cell NHL with a multivariate analysis. Therefore, our study might have some advantages over prior reports associated with the prospective study design, such as consistent pathological diagnosis, standardised clinical evaluation with a single chemotherapeutic regimen, even though the cohort was relatively small compared with those of prior retrospective studies. In the present study, multivariate analysis showed that B symptoms, bulky disease and BM involvement were independent risk factors for survival, and that the IPI system also had independent predictive value. Therefore, we designed a new B score system, composed of B symptoms, bulky disease and BM involvement. On the basis of these findings we propose a new prognostic model for patients with T-cell NHL, namely the PGT, which combines the IPI and B score systems. The PGT model demonstrated four groups, PGT1 (low IPI/low B score), PGT2 (low IPI/high B score), PGT3 (high IPI/low B score), PGT4 (high IPI/low B score) each with different outcomes. OS and PFS in PGT1 were significantly longer than in PGT 2, regardless of a low IPI score. The same finding was observed for PGT3 and PGT4. These observations suggest that the B score system has an important added prognostic value even with the same IPI score. The treatment outcomes of the patients in our study with CEOP-B in PGT1 were acceptable in terms of OS and PFS; however, for patients in PGT2, PGT3 and PGT4 the OS and PFS were disappointing. Therefore, additional protocols must be studied, with a combination of newer drugs or stem cell transplantation in patients with a low IPI and high B score as well as those with high IPI score. Therefore, trials to improve treatment outcomes for patients with T-cell lymphoma must be continued.

In conclusion, the CEOP-B combination chemotherapy with cyclophosphamide, epirubicin, vincristine, prednisone and bleomycin, was moderately active and tolerated in patients with T-cell NHL. The PGT system developed in this study, on the basis of IPI and B (B symptom, BM involvement, bulky disease) scores might be a useful tool for the prediction of long-term survival in patients with T-cell NHL. Therefore, a prospective, multicentre study with the recruitment of sufficient numbers of patients with T-cell NHL, and long-term follow up, should be performed to confirm our results.


  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. References
  • Ansell, S.M., Habermann, T.M., Kurtin, P.J., Witzig, T.E., Chen, M.G., Li, C.Y., Inwards, D.J. & Colgan, J.P. (1997) Predictive capacity of the International Prognostic Factor Index in patients with peripheral T-cell lymphoma. Journal of Clinical Oncology, 15, 22962301.
  • Arrowsmith, E.R., Macon, W.R., Kinney, M.C., Stein, R.S., Goodman, S.A., Morgan, D.S., Flexner, J.M., Cousar, J.B., Jagasia, M.H., Mccurley, T.L. & Greer, J.P. (2003) Peripheral T-cell lymphomas: clinical features and prognostic factors of 92 cases defined by revised European American Lymphoma classification. Leukemia & Lymphoma, 44, 241249.
  • Ascani, S., Zinzani, P.L., Gherlinzoni, F., Sabattini, E., Briskomatis, A., De Vivo, A., Piccioli, M., Fraternali Orcioni, G., Pieri, F., Goldoni, A., Piccaluga, P.P., Zallocco, D., Burnelli, R., Leoncini, L., Falini, B., Tura, S. & Pileri, S.A. (1997) Peripheral T-cell lymphomas. Clinico-pathologic study of 168 cases diagnosed according to the R.E.A.L classification. Annals of Oncology, 8, 583592.
  • Case, D.C. Jr (1983) Long-term results of patients with advanced diffuse, non-Hodgkin's lymphoma treated with cyclophosphamide, doxorubicin, vincristine, prednisone and bleomycin (CHOP-Bleo). Oncology, 40, 186191.
  • Coiffier, B., Brousse, N., Peuchmaur, M., Berger, F., Gisselbrecht, C., Bryon, P.A. & Diebold, J. (1990) Peripheral T-cell lymphomas have a worse prognosis than B-cell lymphomas: a prospective study of 361 immunophenotyped patients treated with the LNH-84 regimen. The GELA (Groupe d'Etude des Lymphomes Agressives). Annals of Oncology, 1, 4550.
  • Escalon, M.P., Liu, N.S., Yang, Y., Hess, M., Walker, P.L., Smith, T.L. & Dang, N.H. (2005) Prognostic factors and treatment of patients with T-cell non-Hodgkin lymphoma: the M. D. Anderson Cancer Center experience. Cancer, 103, 20912098.
  • Gallamini, A., Stelitano, C., Calvi, R., Bellei, M., Mattei, D., Vitolo, U., Morabito, F., Martelli, M., Brusamolino, E., Iannitto, E., Zaja, F., Cortelazzo, S., Rigacci, L., Devizzi, L., Todeschini, G., Santini, G., Brugiatelli, M. & Federico, M. (2004) Peripheral T-cell lymphoma unspecified (PTCL-U): a new prognostic model from a retrospective multicentric clinical study. Blood, 103, 24742479.
  • Gisselbrecht, C., Gaulard, P., Lepage, E., Coiffier, B., Briere, J., Haioun, C., Cazals-Hatem, D., Bosly, A., Xerri, L., Tilly, H., Berger, F., Bouhabdallah, R. & Diebold, J. (1998) Prognostic significance of T-cell phenotype in aggressive non-Hodgkin's lymphomas. Groupe d'Etudes des Lymphomes de l'Adulte (GELA). Blood, 92, 7682.
  • Gottlieb, A.J., Anderson, J.R., Ginsberg, S.J., Bloomfield, C.D., Norton, L., Barcos, M., Peterson, B.A., Nissen, N., Henderson, E.S. & Holland, J.F. (1990) A randomized comparison of methotrexate dose and the addition of bleomycin to CHOP therapy for diffuse large cell lymphoma and other non-Hodgkin's lymphomas. Cancer and Leukemia Group B study 7851. Cancer, 66, 18881896.
  • Haioun, C., Lepage, E., Gisselbrecht, C., Bastion, Y., Coiffier, B., Brice, P., Bosly, A., Dupriez, B., Nouvel, C., Tilly, H., Lederlin, P., Biron, P., Briere, J., Gaulard, P. & Reyes, F. (1997) Benefit of autologous bone marrow transplantation over sequential chemotherapy in poor-risk aggressive non-Hodgkin's lymphoma: updated results of the prospective study LNH87-2. Groupe d'Etude des Lymphomes de l'Adulte. Journal of Clinical Oncology, 15, 11311137.
  • Harris, N.L., Jaffe, E.S., Stein, H., Banks, P.M., Chan, J.K., Cleary, M.L., Delsol, G., De Wolf-Peeters, C., Falini, B., Gatter, K.C., Grogan, T.M., Isaacson, P.G., Knowles, D.M., Mason, D.Y., Muller-Hermelink, H.K., Pileri, S.A., Piris, M.A., Ralfkiaer, E. & Warnke, R.A. (1994) A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood, 84, 13611392.
  • Huang, H.Q., Peng, Y.L., Lin, X.B., Sun, X.F., Lin, T.Y., Xia, Z.J., Li, Y.H., Cai, Q.Q., He, Y.J., Jiang, W.Q. & Guan, Z.Z. (2004) Clinical outcomes of 106 patients with peripheral T-cell lymphoma treated by standard CHOP regimen. Chinese Journal of Cancer, 23, 14431447.
  • Lopez-Guillermo, A., Cid, J., Salar, A., Lopez, A., Montalban, C., Castrillo, J.M., Gonzalez, M., Ribera, J.M., Brunet, S., Garcia-Conde, J., Fernandez de Sevilla, A., Bosch, F. & Montserrat, E. (1998) Peripheral T-cell lymphomas: initial features, natural history, and prognostic factors in a series of 174 patients diagnosed according to the R.E.A.L. classification. Annals of Oncology, 9, 849855.
  • Peng, Y.L., Huang, H.Q., Lin, X.B., Xia, Z.J., Li, Y.H., Wang, W., He, Y.J., Pan, Z.H., Jiang, W.Q. & Guan, Z.Z. (2004) Clinical outcomes of patients with peripheral T-cell lymphoma (PTCL) treated by EPOCH regimen. Chinese Journal of Cancer, 23, 943946.
  • Reiser, M., Josting, A., Soltani, M., Staib, P., Salzberger, B., Diehl, V. & Engert, A. (2002) T-cell non-Hodgkin's lymphoma in adults: clinicopathological characteristics, response to treatment and prognostic factors. Leukemia & Lymphoma, 43, 805811.
  • Rodriguez, J., Munsell, M., Yazji, S., Hagemeister, F.B., Younes, A., Andersson, B., Giralt, S., Gajewski, J., De Lima, M., Couriel, D., Romaguera, J., Cabanillas, F.F., Champlin, R.E. & Khouri, I.F. (2001) Impact of high-dose chemotherapy on peripheral T-cell lymphomas. Journal of Clinical Oncology, 19, 37663770.
  • Rudiger, T., Weisenburger, D.D., Anderson, J.R., Armitage, J.O., Diebold, J., MacLennan, K.A., Nathwani, B.N., Ullrich, F. & Muller-Hermelink, H.K. (2002) Peripheral T-cell lymphoma (excluding anaplastic large-cell lymphoma): results from the Non-Hodgkin's Lymphoma Classification Project. Annals of Oncology, 13, 140149.
  • Sampi, K., Honda, T., Hattori, M. & Takayama, S. (1983) Cyclophosphamide, adriamycin, vincristine, bleomycin and prednisolone (CHOP-Bleo) combination chemotherapy for advanced non-Hodgkin's lymphoma. Cancer & Chemotherapy, 10, 12661271.
  • Schmits, R. & Trumper, L. (2001) Report on workshop: UICC workshop ‘Therapy of NHL in early stages’. Part 2: Aggressive lymphomas. Annals of Hematology, 80(s), 1619.
  • Sonnen, R., Schmidt, W.P., Muller-Hermelink, H.K. & Schmitz, N. (2005) The International Prognostic Index determines the outcome of patients with nodal mature T-cell lymphomas. British Journal of Haematology, 129, 366372.