Clinicopathological study on peripheral T-cell non-Hodgkin lymphoma with bone marrow involvement: a retrospective analysis from China
- First Online:
- Cite this article as:
- Tong, H., Ren, Y., Qian, W. et al. Int J Hematol (2009) 90: 303. doi:10.1007/s12185-009-0390-6
- 161 Views
We reviewed 173 patients with an initial diagnosis of peripheral T-cell non-Hodgkin lymphoma (PTCL) and compared the patients with bone marrow involvement (BMI) to those without to have a better understanding of the clinical characteristics, treatments, survival and prognosis of PTCLs with BMI. We found that 40% (70/173) of the patients had BMI, and its frequency was 64% in angioimmunoblastic T-cell lymphoma (TCL), 46% in PTCL unspecified, 29% in anaplastic large T-cell lymphoma, 23% in extranodal NK/T-cell lymphoma and 13% in enteropathy-type TCL. In the BMI group, 36% of patients had lymphoma-associated hemophagocytic syndrome (LAHS), compared with 8% of the patients without BMI (8/103, P < 0.001). The estimated 1-year overall survival (OS) rates of patients with LAHS in the BMI and non-BMI groups were 5 and 49%, respectively. The increased levels of lactate dehydrogenase, fasting triglycerides and β2-microglobulin between the BMI and non-BMI groups were not significantly different, but ferritin increased significantly and liver dysfunction-related diseases were seen more in the BMI group. As much as 51% of patients of the BMI group had anemia, compared with 27% of the patients without BMI (P = 0.001). The estimated 2-year OS rates in the two groups were 10 and 34%. The estimated 2-year OS rate of the 67 patients with BMI, who did not lose to follow-up, was 22%, compared with 38% in the non-BMI group. The median survival times of the 2 groups were 120 and 356 days. The estimated 2-year OS rate of patients treated by CHOP regimen was 9%, compared with 51% of those with intensive chemotherapy, with a significant difference (log rank P = 0.0008). The median survival time of the 14 patients subjected to chemotherapy combined with l-asparaginase was 365 days and that of the 7 patients undergoing hemopoietic stem cell transplantation (HSCT) was 575 days. A total of 3 patients in a critical condition underwent plasmapheresis as initial therapy and achieved stable condition. We conclude that patients with PTCLs with BMI on initial diagnosis usually have hemaphagocytic syndrome and poor prognosis. BMI without lymphadenopathy is a patent clinical feature in most PTCLs. Patients with anemia on initial diagnosis in the BMI group usually have poor prognosis than those without. Intense chemotherapy, addition of l-asparaginase in chemotherapy and HSCT are comparatively efficient treatments of PTCLs. For patients in critical conditions, plasmapheresis before chemotherapy would lower the risk and improve the tolerance to chemotherapy.
KeywordsT-cell lymphomaBone marrow involvementLymphoma-associated hemophagocytic syndromeOverall survivalChemotherapy
T-cell non-Hodgkin lymphomas (NHLs) are malignancies that represent approximately 12% of all lymphomas. The relative frequency of mature T-cell lymphoma appears to vary considerably according to geography, ranging from 1.5% in Vancouver, British Columbia and Canada to 18.3% in Hong Kong, 25% in Thailand and almost 30% in mainland China .
NHLs are diagnosed with bone marrow biopsy to rule out other pathology, assess the extent of disease and monitor treatment response. In B-cell NHLs, the frequency and histology of bone marrow involvement (BMI) on initial diagnosis vary greatly among different clinicopathological entities recognized by the WHO classification, ranging from as low as 3% in primary mediastinal large B-cell lymphoma to as high as 73% in B-cell small lymphocytic lymphoma/chronic lymphocytic leukemia and lymphoplasmacytoid lymphoma . In T-cell NHLs, the frequency of BMI on initial diagnosis was 20–40% , and reached as high as 90.7% (42/54) in China, of which 46.3% progressed to lymphosarcoma leukemia . The different frequencies of BMI may be associated with the different constituent ratio of subtypes in different research centers. It has been reported that T-cell NHLs had higher frequency of BMI than aggressive diffuse large B-cell lymphoma (DLBCL) .
T-cell lymphomas that arise from postthymic cells (mature T cells) are collectively called PTCLs. Since peripheral T-cell lymphomas (PTCLs) with a high frequency of BMI are common in Asia, we reviewed 173 patients in our center with initial primary PTCLs and compared the patients with BMI to those without to better understand the clinical characteristics, treatments, survival and prognosis of PTCL with BMI.
2 Materials and methods
From January 2000 to September 2007, 173 patients (68 male/35 female; median age: 44 years; range: 12–81 years) were diagnosed with PTCL in the First Affiliated Hospital, College of Medicine, Zhejiang University. Their clinical and laboratory records were reviewed. All patients underwent at least one site of bone marrow biopsy for lymphoma staging and 70 of these patients (45 male/25 female; median age: 42 years; range: 15–79 years) were confirmed with BMI at the time of diagnosis.
2.2 Histopathologic assessment and immunohistochemical staining
Specimens from bone marrow biopsies were assessed as well as those from the initial sites of involvement. The specimens were fixed in buffered formalin or Bouin fluid. Paraffin-embedded tissue sections were stained with hematoxylin-eosin, periodic acid-Schiff (PAS) or Wright-Giemsa. Preliminary diagnoses were made based on morphology and basic clinical information.
The immunophenotype was analyzed using mouse monoclonal antibodies of variable specificity to detect cellular antigens in the frozen or paraffin-embedded tissue sections. Sections of all patients were stained for B-cell markers (CD20 and CD79a) and T-cell markers (CD2, CD3 and CD45), and the staining results showed one or more positive T-cell markers and negative B-cell markers. Immunostains of CD30 and anaplastic leukemia kinase-1 (ALK-1) were further used in the differential diagnosis of systemic anaplastic large cell lymphoma. CD56, CD57, TIA-1 and granzyme B were used in the diagnosis of NK/T-cell lymphoma.
2.3 Diagnosis criteria
PTCLs consist of a variety of subtypes, some of which show typical clinical presentations according to the WHO classification (2001), including systemic anaplastic large T-cell lymphoma (ALTCL), angioimmunoblastic T-cell lymphomas (TCL), cutaneous TCL (mycosis fungoides and Sezary syndrome), extranodal NK/T-cell lymphoma, enteropathy-type TCL, hepatosplenicγδTCL, subcutaneous panniculitis-like TCL, and peripheral T-cell lymphoma, unspecified (PTCL-U). Considering that the biological characteristic of lymphoblastic TCL is similar to acute lymphocytic leukemia and usually the chemotherapy regimens employed are basically the same, we excluded this type from the analysis.
The diagnostic criteria for lymphoma-associated hemophagocytic syndrome (LAHS) are as follows: (1) fever; (2) splenomegaly; (3) cytopenia for at least 2 cell lines with hemoglobin <90 g/L (or below 120 g/L for more than 4 weeks), platelet <10.0 × 109/L and neutrophils <1.0 × 109/L; (4) hypertriglyceridemia (fasting triglycerides >3 mmol/L) and/or hypofibrinogenemia (fibrinogen <1.5 g/L); (5) ferritin >500 mg/L; (6) sCD25 >2400 U/mL; (7) decreased or absent NK-cell activity; (8) hemophagocytosis in bone marrow, cerebrospinal fluid (CSF) or lymph nodes. At least five of the eight criteria mentioned above are required to make a diagnosis .
As in many instances, the histological features of lymphoma are subtle and overlap with a variety of reactive T-cell proliferations and other malignant processes; the diagnosis often requires critical integration of the information obtained from clinical features, peripheral blood, aspirate findings and histopathological examination of the bone marrow biopsy . In our study, lymphoma cells were found in marrow aspirate or/and marrow biopsy that established the diagnosis of BMI, which needs hematopathology review of all slides with at least one paraffin block representative of the tumor and adequate immunophenotyping for the diagnosis.
2.4 Clinical staging
Disease grade was evaluated according to the Ann Arbor Staging System. Once BMI was detected, the disease was classified to the IV stage. Performance status was based on the Eastern Cooperative Oncology Group scale (0–4). International Prognostic Index (IPI) scores were determined in patients and used in the survival analysis.
Chemotherapy regimens including CHOP or CHOP-like regimens (idamycin, mitoxantrone, liposomal doxorubicin substituting for epirubicin) were employed. Intensive chemotherapies included ECHOP, CHOP with Ara-c (cytosine arabinoside 200 mg d1–7), MAVP (mitoxantrone 8 mg/m2 d1–3, cytosine arabinoside 200 mg d1–7, vincristine 1.4 mg/m2 d1, d8, and prednisone 60 mg d1–14), VDCP, MINE, DHAP and hyperCVAD. Besides, 14 patients were given l-asparaginase (l-ASP, 10 000 IU d1–10) in combination with the regimens mentioned above. FLAG, BFM-90 and proMACE-cytoBOM regimens were tried on three patients, respectively. Gemcitabine was tried on 2 patients. Velcade, monoclonal antibody to CD52 and arsenic trioxide were tried on three patients, respectively. Seven patients received local radiotherapy. Six patients underwent autologous hemopoietic stem cell transplantation (HSCT) and one allogene HSCT.
Three patients in a critical condition received plasmapheresis as initial treatment before initial chemotherapy by a Baxter CS-3000 Plus blood cell separator. Plasma exchange was done once a day for 2–3 days and 2000 mL of plasma was replaced each time.
2.6 Outcome analysis and statistical methods
Overall survival (OS) was measured from the date of diagnosis to the date of death owing to any cause or to the last follow-up visit and was estimated by the Kaplan–Meier method. Comparisons among those variables of interest were performed by the log-rank test. Chi-square test was used to compare the clinical and laboratory data between the BMI and the non-BMI groups. A P value less than 0.05 was considered to be statistically significant. SPSS Version 11.5 was used for statistical analyses.
3.1 Clinical features
Histological examination confirmed that 70 of the 173 patients at the first diagnosis were with BMI. Of the 70 patients, 54% (38) were diagnosed only by bone marrow specimens. In the non-BMI group, the diagnostic tissues were obtained from lymph node in 61 patients, oral and pharynx nasalis in 17, enteron in 11, cutaneous and subcutaneous tissues in 9, lung in 2, adrenal gland in 1, spleen in 1 and ascitic fluid in 1.
In the BMI group, 49 patients were with PTCL-U, 5 with ALTCL, 6 with extranodal NK/T-cell lymphoma, 7 with angioimmunoblastic lymphoma, 1 with enteropathy-type TCL, 1 with subcutaneous panniculitis-like TCL and 1 with mycosis fungoides. In the non-BMI group, 58 patients were with PTCL-U, 20 with extranodal NK/T-cell lymphoma, 12 with ALTCL, 4 with angioimmunoblastic lymphoma, 7 with enteropathy-type TCL, 1 with subcutaneous panniculitis-like lymphoma and 1 with mycosis fungoides. The frequency of BMI in every subtype was listed as follows: 64% (7/11) in the angioimmunoblastic TCL group, 50% (1/2) in the subcutaneous panniculitis-like TCL, 50% (1/2) in the mycosis fungoides group, 46% (49/107) in the PTCL-U group, 29% (5/17) in the ALTCL group, 23% (6/26) in the NK/T-cell lymphoma group and 13% (1/8) in the enteropathy-type TCL group.
Demographics and clinical characteristics of PTCL: BMI group versus non-BMI group
BMI group (n = 70)
Non-BMI group (n = 103)
Median age (years)
Age range (years)
Low risk (0–1)
Intermediate risk (2–3)
High risk (4–5)
Cytopenia for at least 2 cell linesa
Developed to lymphosarcoma leukemia
Laboratory findings of patients with and without BMI
Fasting triglycerides >3 mmol/L
LDH >1000 U/L
Β2-microglobulin >2300 μg/L
CA125 >35 U/mL
Ferritin >500 μg/L
3.2 Therapeutic results and survival analysis
The frequency of BMI varies greatly among different clinicopathological entities of PTCLs. We found in the literature that the highest frequency of BMI was more than 95% in hepatosplenicγδTCL, followed by 32–70% in angioimmunoblastic TCL, 20–40% in PTCL-U, 15–25% in ALCL, 5–8% in enteropathy-type TCL and 5–7% in extranodal NK/T-cell lymphoma [6–10]. Our data revealed that 40% of the 173 patients with PTCLs had BMI at initial diagnosis and that the frequency of BMI was up to 64% in angioimmunoblastic TCL, 46% in PTCL-U, 29% in ALTCL, 23% in extranodal NK/T-cell lymphoma, and 13% in enteropathy-type TCL. Our findings were corresponding to the previous reports, except that the frequency of BMI in enteropathy-type TCL and extranodal NK/T-cell lymphoma was higher than that reported in literature.
A notable characteristic of many cases of PTCL is that the source of tissue available for assessment can only be obtained by bone marrow biopsy. In our study, 60% of patients in the BMI group had no lymphadenopathy or other extranodal presentation of PTCLs, and its definite diagnosis was made only by repeated bone marrow biopsies. This result is consistent with those of other researches [11, 12] and indicates that BMI on initial diagnosis in the absence of lymphadenopathy is a clinical feature of most PTCLs.
We also found that patients with BMI more frequently had LAHS than those without (36 vs. 8%, P < 0.001), with an early mortality as high as 78%. We had reported the characteristics of T-cell lymphoma with LAHS in our previous study . The clinical manifestations, including B symptoms, hepatosplenomegaly and cytopenia in at least two lines, especially anemia, were more obvious in the BMI group than those in the non-BMI group, as were ferritin and diseases associated with liver dysfunction. But, increased levels of LDH, β2-microglobulin, CA125 and fasting triglycerides were irrelevant to whether bone marrow was involved or not. We suppose that the characteristics mentioned above were associated with the high rate of LAHS.
Since 1998, hemoglobin concentration has been confirmed as a dependent prognostic factor of Hodgkin lymphoma in the later stage. Patients with lower hemoglobin concentration respond poorly to treatment. This may be associated with a low oxygen concentration and an increase of hypoxic cells, which are not sensitive to radiotherapy and chemotherapy . In patients with PTCLs associated with BMI, our results showed that the OS of patients with anemia was lower than those without, and their estimated 2-year survival rates were 10 and 34%, respectively. We suggest that anemia would affect the prognosis of the disease.
It was reported that BMI was related to the poor response to chemotherapy and short survival time in DLBCL . A similar conclusion was drawn from our study, that is, a higher mortality of 45% for PTCLs in the early stage was observed in patients with BMI, compared with 26% in patients without BMI (P < 0.05). The OS of patients with BMI was dramatically lower than that of those without, and the estimated 1-, 2- and 5-year survival rates of 33, 22 and 10% were compared to those of 56, 38 and 19% between the 2 groups, respectively. A multiple-center retrospective research in Italy showed that BMI was a dependent prognostic factor of PTCL-U, and that based on this finding researchers from this group incorporated BMI into their scoring system for evaluating the prognosis of PTCL-U in which other parameters including age, serum LDH and physical state were also taken into account.  However, reports from China  and eastern countries expressed different opinions. Tsimberidou et al.  reported that the risk factors of PTCL were IPI score, the number of extranodal diseases and whether able to achieve complete remission after treatment, but were not related to BMI. An explanation for the difference in conclusions is not available. We suppose that such difference may exist because of the different constituent ratios of PTCL subtypes from different regions or regional locations of PTCLs. Further researches, especially the study on every subtype of PTCL, need to be done.
The prognosis of PTCLs with BMI was worse than that of DLBCL with BMI, and the efficacy of conventional chemotherapy remained poor . The 5-year survival rate of patients subjected to CHOP regimen was 25–45% [18–23], and the difference in the survival rate may be attributed to different research time and standard for histopathological classification. In our study, the estimated 2- and 5-year OS rates were 51 and 15% in the intense chemotherapy group, respectively, and the latter was even poorer than the lower limit of that reported. Improving the survival rate of patients of PTCLs with BMI remains a challenge. Our data clearly showed that the survival time for patients without any treatment (including critically ill patients and those who gave up treatment) was the shortest, and the survival time and OS rate were greater in the intense chemotherapy group than that in the CHOP regimen group.
We explored some treatments to improve the prognosis of PTCLs and achieved some better results. First, the median survival time of patients who received autologous HSCT was markedly prolonged, compared with patients who received chemotherapy. Rezania et al.  reported that high-dose chemotherapy followed by autologous or allogeneic HSCT has emerged as a potential curative treatment in patients with high-risk or relapsed and refractory PTCL. Second, chemotherapy with the aid of l-ASP, a major component of therapy for acute lymphoblastic leukemia, successfully prolonged the median survival time of patients. As in Yong et al’s report , the 2-year OS rate was 64.6% and there was a notable disparity compared with the group without l-ASP (45.5%), which supported the effectiveness of the addition of l-ASP in our study. Third, 3 patients with high fever, jaundice, hepatosplenomegaly, severe liver dysfunction and pancytopenia underwent plasmapheresis as initial therapy and achieved a stable condition. We therefore consider that plasmapheresis before chemotherapy may reduce the risk and improve the outcome of chemotherapy, although its application under this circumstance is far from being sufficient.
In conclusion, patients of PTCLs with BMI on initial diagnosis usually have hemaphagocytic syndrome, and BMI without lymphadenopathy is a patent clinical feature in most PTCL. BMI is less observed in extranodal NK/T-cell lymphoma and enteropathy-type TCL. The prognosis of patients in the BMI group with LAHS and anemia on initial diagnosis is poor, and the total survival rate of patients is lower than that without BMI. Intense chemotherapy, addition of l-ASP to chemotherapy and HSCT are comparatively efficient treatments for PTCL. For patients in critical a condition, plasmapheresis before chemotherapy would lower the risk and improve their tolerance to chemotherapy.
This work was supported by the Medical Science Research Fund of Zhejiang Province (No 2007B079) and also by the help of the Case Management Center of the First Affiliated Hospital of Medical College, Zhejiang University. We thank our interns for collecting and conducting the follow-up.