| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Article |
From the Institutes of Pathology,*
GSF-National Research Center for Environment and Health,
Neuherberg, Germany; and Technical University,
Munich, Germany; and the Instituto Nacional de la
Nutricion,
and the Instituto Nacional de
Cancerologia,
Mexico City, Mexico
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunophenotypic and genotypic studies have shown that nasal NK/T-cell lymphomas, in addition to the almost universal association with Epstein-Barr virus (EBV),10,11 express the NK cell marker CD56,12 and are usually negative for surface CD3 (Leu4), although cytoplasmic CD3 can be detected in paraffin sections. This phenotype, together with the lack of clonal T-cell receptor gene rearrangements, strongly supports the notion that these tumors are derived from NK cells, and not from conventional T cells.13
Although our understanding of nasal NK/T-cell lymphoma has increased in the last years, very little is known about the cytogenetic and molecular changes of this disease.14 This might be because, in part, of the small samples available with frequently extensive necrotic changes, and to the rarity of this entity. In a recent immunohistochemical study, we showed that p53 was overexpressed in a high percentage of cases of nasal NK/T-cell lymphoma from Peru.15 However, the mutational status of the p53 gene was not determined.
In the present study of a series of nasal NK/T-cell lymphoma from Mexico, we aimed to investigate the frequency of p53 gene alterations [mutations and loss of heterozygosity (LOH)], and to correlate them with the expression of p53 protein, and its downstream target, the cyclin-dependent kinase inhibitor p21, as a surrogate marker of p53 functionality. In addition, we examined the potential prognostic implication of p53 alterations for clinical outcome.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Thirty-six paraffin-embedded blocks of nasal lymphoma biopsy
material from the Instituto Nacional de la Nutricion and the Instituto
Nacional de Cancerologia, Mexico City, Mexico were studied. Eight cases
immunophenotypically corresponded to B-cell non-Hodgkin’s lymphomas
(NHLs) and three cases showed extensive necrosis with lack of viable
tissue and, therefore, were excluded from the series. The remaining 25
cases form the basis of this study. Clinical information was obtained
from the patients’ medical records. Hematoxylin and eosin-stained
slides and immunoperoxidase studies were reviewed in all cases by two
of the authors (LQ-M and FF). In addition, cytological features and
immunophenotype were recorded. The term NK/T-cell lymphoma is used in
this study as defined in the new World Health Organization
classification16
and corresponds to angiocentric lymphoma
in the Revised European American Lymphoma classification
(REAL).17
Cases of lymphoma presenting in nasal or
paranasal tissues were included in this study if they showed expression
of cytoplasmic CD3
, cytotoxic granule associated proteins (TIA-1),
EBERs, and in most cases CD56.
Immunohistochemistry
Immunohistochemistry was performed on an automated immunostainer
(Ventana Medical System, Inc., Tucson, AZ) according to the company’s
protocols, with slight modifications. After deparaffinization and
rehydration, the slides were placed in a microwave pressure cooker in
0.001 mol/L citrate buffer, pH 6.0, containing 0.1% Tween 20, and
heated in a microwave oven at maximum power for 30 minutes. The
antibody panel used included L26 (CD20; DAKO, Glostrup, Denmark)
polyclonal-CD3 (DAKO), CD56 (Novocastra, Newcastle, UK), TIA-1
(Coulter, Miami, FL), the monoclonal antibody DO7 against the
wild-type/mutant p53 (DAKO), and p21 (BD Transduction Laboratories, San
Diego, CA). Tissues were scored as p53-positive if 
10% of the tumor
cells had nuclear staining. Because p21 expression is found only in
isolated lymphocytes and in the epithelium of normal tonsils, any
staining 5% was considered positive. A grid ocular objective was
used to count 300 cells over three high-power fields (x40) and the
percentage of positive cells was reported as 0 to 100%.
In Situ Hybridization
In situ hybridization for EBV early RNA (EBER) was performed on fixed paraffin-embedded sections as previously described.15 The in situ hybridization procedure was performed under RNase-free conditions. The fluorescein-conjugated EBER peptide nucleic acid probe was obtained from DAKO. A well-characterized EBV+ Hodgkin’s lymphoma case was used as positive control.
DNA Extraction and Polymerase Chain Reaction (PCR) Amplification of p53 Gene
Genomic DNA was extracted from serial, dewaxed paraffin sections mounted on glass slides. In the majority of cases, depending on the amount and distribution of the tumor infiltrate, neoplastic cells were enriched either by manual microdissection or by laser capture microdissection of hematoxylin-stained slides as previously described.18 Adjacent sections stained for CD56 or p53 were used to identify the areas with the highest tumor cell content. The tissue fragments were digested with proteinase K overnight at 56°C, and the crude extract was used for PCR.
The highly conserved exons 5 to 8 of the p53 gene, including the
intron/exon boundaries, were amplified by PCR. The primers used are
listed in Table 1
. The integrity of the
DNA was assessed by amplification of a 268-bp fragment of the
ß-globin gene. The quality and correct size of the PCR products were
checked on agarose gels.
|
Mutational screening was performed according to the method
described by Oefner and Underhill19
on an automated DHPLC
analysis system (Transgenomic, Omaha, NE). The PCR products were
denatured for 4 minutes at 94°C and cooled to room temperature at a
rate of 1°C/minute. Three to 15 µl of PCR product were applied to a
preheated reverse-phase column (DNA-Sep, Transgenomic). Elution of DNA
was performed in a linear acetonitrile gradient. The temperature for
optimal resolution of heteroduplex and homoduplex DNA detection was
determined by analyzing the melting behavior of each PCR fragment while
the temperature was increased at 1°C steps beginning at 50 to 55°C
until the fragment was completely melted. The optimal temperatures for
the identification of mutations was established for each of the four
examined exons using a set of test DNA samples with known mutations.
The overall sensitivity of this method for the detection of p53
mutations in exons 5 to 8 is in the range of 95%. A detailed
discussion of this technique is given in a separate
report.20
A mutation can be detected reliably, if
10 to 20% of the PCR product shows the mutation. In cases with a
high tumor cell content, DHPLC was performed both with and without
admixture of wild-type PCR product of the corresponding exon to ensure
detection of mutations, because absence of wild-type DNA as a result of
LOH at the p53 locus, could potentially lead to a false-negative
result.
Direct Sequencing
Automated fluorescent sequencing was performed with the BigDye Terminator Cycle Sequencing kit (PE Applied Biosystems, Foster City, CA) and ABI Prism 377 automated sequencer (PE Applied Biosystems). All chromatograms were evaluated independently by two observers (GK and FF). Comparison with the wild-type sequences was performed with DNASIS 2.6 software (Hitachi Software Engineering, Yokohama, Japan).
LOH Analysis
Normal and neoplastic tissues were obtained by laser capture microdissection of hematoxylin-stained slides as described above. The primer sequences of the microsatellite marker TP53 used in this study target a highly informative dinucleotide repeat polymorphism.21 PCR amplification was performed using a TC1 thermal cycler (Perkin Elmer) under standard conditions in a final volume of 15 µl containing 2 µl of template DNA, 1.5 µl of 10x PCR buffer (15 mmol/L MgCl2, 0.5 mol/L KCl, 0.1 mol/L Tris-HCl, pH 9.0), 1.25 mmol/L dNTPs, 20 pmol of each primer oligonucleotide (forward primer is fluorescently labeled), and 1.5 U of Taq polymerase (Amersham Pharmacia). PCR cycles of denaturation, primer annealing, and extension were performed at 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 90 seconds for 35 cycles. Reaction products were separated and detected using an automated sequencer (ABI 377) and analyzed with the ABI Prism GeneScan software. LOH was defined as a reduction of one allele in tumor tissue by at least 50% when compared to the corresponding allele in normal tissue. Homozygous cases were considered to be not informative.
Statistical Analysis
The significance of the association of p53 mutations with morphology and clinical stage at presentation and the significance of overall survival and clinical stage was assessed using Fisher’s exact test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Clinical data and histology are summarized in Table 2
. Of the 25 patients, 17 were male and 8
female (M: F ratio, 2.1:1), with a median age of 43 years (range, 21 to
93 years). The most common presenting symptom was nasal obstruction. In
16 cases the tumor sample was from the nasal cavity, in 5 cases from
the nasopharynx, and in 4 cases from the palate. Patients were treated
with combined radiation and chemotherapy modalities in 15 cases,
chemotherapy alone in 4 cases, and radiation alone in 1 case. Eleven
patients presented with early stage disease (stage I to II), and 14
patients with advanced disease (stage III to IV). Thirteen patients
died with disease; 1 patient declined treatment because of advanced
disease, 2 patients died before treatment could be started, and 10
patients died despite treatment. Two of these patients initially had a
complete response (cases 7 and 16) and died with recurrence of the
disease at 16 and 8 months after the initial diagnosis, respectively.
Nine patients are alive; seven with no evidence of disease, and two
patients with disease. Three patients were lost to follow-up, one
during the treatment period, and two patients (cases 2 and 11), because
the treatment was declined.
|
. Cytological atypia characterized by
nuclear irregularities and clumped chromatin were often identified.
Inflammatory cells including neutrophils, eosinophils, plasma cells,
and histiocytes, frequently were intermingled with the tumor cells. In
the remaining four cases, the tumor was composed almost exclusively of
large transformed cells with irregular nuclei, clumped chromatin and
often prominent nucleoli (Figure 1B)
. The mitotic activity was high and
the inflammatory component was less prominent.
|
The results of the immunophenotypic studies and the EBV analysis
are summarized in Table 3
. All cases
expressed CD3. The neoplastic cells showed CD56 expression in 20 of
25 cases (80%). TIA-1, which is a well-characterized cytotoxic
molecule primarily restricted to cytotoxic T cells and NK cells, was
strongly positive in the cytoplasm of the tumor cells in all cases.
Overexpression of p53 was observed in >10% of the tumor cells in 15
of the 25 cases (60%). In 9 of these 15 cases the percentage of p53+
cells ranged between 14% and 45%, the intensity of the expression
varied from cell to cell in any given case. In the remaining six cases,
p53 was strongly positive in >50% of the tumor cells. Overexpression
of p21 was observed in 15 cases (60%). All cases showed labeling for
EBERs in a large proportion of viable, neoplastic cells.
|
The results of the mutational analysis are summarized in Table 3
.
Successful DNA amplification was obtained in 21 of the 25 cases. In 9
of 21 cases the mutational screening of exons 5 to 8 of the p53 gene
showed an abnormal (five cases) or questionable (four cases) DHPLC
pattern in at least one of the analyzed exons (Figure 2
; A, B, and C). All PCR products showing
an abnormal or questionable DHPLC pattern were subjected to direct
sequencing of both DNA strands. In all five cases with abnormal DHPLC
curve, missense mutations leading to amino acid substitutions were
detected (one each in exon 5 and 6, and three in exon 7) (Figure 2C)
.
In contrast, in none of the four questionable cases were p53 mutations
found (Figure 2B)
. In addition, all four exons from the cases with a
high level of p53 staining without abnormalities of the DHPLC pattern
(cases 8, 9, 11, and 12), and further randomly chosen PCR products were
sequenced. In none of these cases were p53 mutations identified.
|
Three different groups determined by the expression of p53 protein
and p53 gene status were identified (Figure 3
; A to D). Group 1 contained the five
cases with p53 missense mutations. All five cases revealed strong
nuclear overexpression of p53 in the majority of the tumor cells with a
mean of 86% (range, 53 to 90%) (Figure 3A)
. Unexpectedly, three of
these cases also showed overexpression of p21 (Figure 3B)
. Group 2
included the 10 cases that showed overexpression of p53, but retained
wild-type (wt) configuration of the p53 gene in the nine amplifiable
cases. The percentage of p53-positive cells was lower than in the
previous group with a mean of 23% (range, 14 to 52%). Only one case
in this group showed p53 positivity in >50% of the tumor cells.
Although this case revealed a normal DHPLC elution curve, direct
sequencing of exons 5 to 8 was performed to corroborate the presumable
wt status of the p53 gene. P21 was positive in 7 of the 10 cases
including the case with high p53 expression (Figure 3C)
. Group 3
included 10 cases with low or null expression of p53. All seven
amplifiable cases in this group showed a wt p53 gene. P21 was
overexpressed in 5 of the 10 cases. In two cases (cases 16 and 20) the
expression of p21 was relatively high in comparison with the other
tumors (Figure 3D)
.
|
The tumor morphology and clinical characteristics of the patients
were analyzed in the three different groups described above (Table 2)
.
Group 1 included five cases of mutated p53+/p53. Histologically, three
cases were composed of large transformed cells and two cases were
composed of a mixture of medium-sized cells and large cells. Four
patients in this group presented with advanced disease (Stage IVB) and
died shortly after diagnosis. The fifth patient presented with stage
IIB, declined treatment and was lost to follow-up. Group 2 included 10
cases of p53+/p53 wild type. Histologically, nine cases were composed
of a mixture of medium and large-sized cells, and one case was composed
predominantly of large cells. Five patients are alive12 to 120 months
later (mean, 24 months), three with no evidence of disease, and two
with disease. The three patients that are alive with no evidence of
disease presented clinically with stage IIA-B. Case 15 presented with
stage IA and recurred locally (stage IA) after 10 years of complete
remission. He is currently undergoing chemotherapy. Four patients died
with disease, three of them had presented with advanced disease. Case 7
presented with stage IB and initially had a complete remission;
however, 16 months later recurred and died with disease shortly
thereafter. Case 11 presented with stage IIB, the patient refused
treatment and was lost to follow-up. Group 3 included 10 cases of
p53-/p53 wild type. Histologically, all cases in this group were
composed of a mixed population of medium-sized and large cells. Four
patients are alive with no evidence of disease 9 to 60 months later
(mean, 10 months), three of these patients presented clinically with
stages IA-B and IIB. Five patients died with disease, four of them
presented with advanced disease. Case 16 presented with clinical stage
IIA and had initially a complete remission, however, recurred 8 months
later with disease in lung and liver and died with disease. Case 20
presented with clinical stage IVB and was lost to follow-up during the
treatment period.
No significant differences in the frequency of the main clinical parameters, ie, B symptoms (100% versus 80% versus 70%), advanced disease (stage III and IV; 80% versus 40% versus 60%), and overall survival (0% versus 56% versus 44%) was observed between groups 1, 2, and 3. Nevertheless, it is of clinical interest that four of five (80%) patients in group 1 presented with stage IVB and died with disease. Furthermore, morphologically, three of five cases in group 1 were composed of large transformed cells (60%), in contrast to only one case in groups 2 and 3 (5%). Comparison between morphology and the presence of p53 mutations showed that nasal NK/T-cell lymphomas with large cell morphology were more likely to be associated with p53 mutations (P = 0.0162).
LOH Analysis
Despite the use of laser capture microdissection, only 12 cases
rendered sufficient normal tissue for LOH analysis. Three cases
corresponded to group 1 (cases 1 to 3), six to group 2 (cases 6, 9, 10,
13 to 15), and three to group 3 (cases 16, 17, and 20). Analysis of
informative loci showed LOH in 2 of 12 cases (17%), both of which were
associated with p53 mutations of the remaining allele (cases 1 and 3)
(Figure 4)
. In one case with p53 mutation
in exon 5, no LOH was identified (case 2). One case in group 3 (case
17) was noninformative.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The p53 gene is the most frequently mutated gene in human cancer, and the prototype of checkpoint regulator.22 However, p53 gene mutations are relatively rare in NHL. In a large study of B- and T-cell NHL, only Burkitt’s lymphoma and progressed B-cell chronic lymphocytic leukemia (Richter’s transformation) were associated with p53 mutations.23 Among peripheral T-cell lymphomas, with the exception of adult T-cell leukemia/lymphoma,24 p53 mutations are found in <10% of the cases.25 In comparison, we found a relatively high incidence of p53 mutations in nasal NK/T-cell lymphoma from Mexico (24%). Accordingly, a high incidence of p53 mutations (47.6%) was also found in a recent study of nasal NK/T-cell lymphomas from Japan and China.26 The incidence of p53 mutations observed in that study varied depending on the geographic origin of the patients, suggesting that the development of p53 mutations is influenced by geographic, environmental, or ethnic differences. The mutational pattern observed in the five cases in our study is similar to previous findings in NHL.23,27 All mutations corresponded to missense mutations, with a predominance of transitions over transversions. However, none of our mutations involved a CpG dinucleotide or any of the predominant mutational hot spots. In contrast to our series, Li and colleagues26 observed a relatively high percentage of silent mutations, as well as an absence of p53 expression in some cases with missense mutations. The reasons for these differences are not clear.
As a screening method for the detection of p53 mutations we used DHPLC, a relatively new technique based on the different melting characteristics of hetero- and homoduplex forms between wild-type and mutated DNA fragments.19 Until now, the method has mainly been used to detect germline mutations in various inherited diseases for which a high degree of sensitivity ranging from 95 to 100% has been reported.28-31 Compared to currently used conventional screening techniques including single-strand conformation polymorphism and denaturing gradient gel electrophoresis, DHPLC is an easy and reliable screening method that significantly reduces the screening time for large numbers of tumors. Its sensitivity, in the range of 95%, compares favorably with these more labor-intensive techniques. All mutations in this study confirmed by sequencing had shown a clearly abnormal DHPLC pattern. None of the PCR products with questionable (four cases) or normal DHPLC elution curves were found to have a mutation. Although we cannot completely exclude that mutations outside of the examined exons 5 to 8 were responsible for p53 overexpression in some of the cases of group 2, we consider this unlikely. Hematological neoplasms, in general, show a very low rate of mutations outside the conserved domain, and when present, these are frequently nonsense mutations not leading to p53 protein accumulation.27
In an effort to see whether the overexpressed p53 protein in cases of
nasal NK/T-cell lymphoma with retention of wild-type p53 gene is
functional, we studied the level of expression of its downstream
target, the cyclin-dependent kinase inhibitor p21.22
Previous studies have shown that the expression of p21 is a very good
surrogate marker for the status of the p53 gene.32,33
In
cases with overexpression of wild-type p53, p21 protein is usually
detectable, whereas the absence of p21 correlates with the presence of
missense p53 mutations. To our surprise, even though most of the cases
in group 2 (p53+/p53 wild-type) showed expression of p21, as expected,
three of five cases with missense p53 mutations (group 1), as well as 5
of 10 cases without p53 expression also showed high expression of p21.
This finding indicates that p21 is up-regulated independent of p53 gene
status. An interesting, but still speculative possibility is that EBV
might be responsible for, or contributes to p21 nuclear accumulation.
Accordingly, in the study of Villuendas and colleagues32
a
tendency for EBV-positive cases to harbor high levels of both p53 and
p21 was found, indicating that EBV could be involved in the
nuclear accumulation of both proteins in NHL. In
nasopharyngeal carcinoma, another EBV-associated malignancy, p53 gene
mutations are rarely identified despite the frequent detection of p53
overexpression, again suggesting a role for EBV in the p53 accumulation
in EBV-associated neoplasias.34
A further support for the
notion that EBV itself, rather than the neoplastic transformation, is
responsible for p53 overexpression in tumors with wild-type p53 is the
up-regulation of p53 in EBV-infected, nonneoplastic B-cells in
infectious mononucleosis,35
and in angioimmunoblastic
T-cell NHL.36
The mechanism(s) through which EBV could
potentially achieve accumulation of p53 and p21 is still unresolved.
For p53, several hypotheses have been proposed, including p53 binding
to EBV nuclear antigen 5 (EBNA-5)37
or ZEBRA
protein,38,39
and up-regulation of p53 through induction
of nuclear factor kappa B (NF-
B) by EBNA-2 and LMP1.40
In accordance with the tumor suppressor gene model, mutation of one allele is often accompanied by deletion of the remaining allele, although this loss is not a requisite. In this study, of the 12 cases analyzed, only 2 cases (17%) showed LOH on chromosome 17p13, both of which showed p53 mutations of the remaining allele. A recent LOH study of different NK-cell neoplasias from China41 showed that LOH at chromosome 17p13 occurred only in nasal NK/T-cell lymphoma (four of nine, 44%), with a higher incidence at presentation (three of six) than at relapse (one of three). However, in addition to the small number of cases, the clinical stage of the patients at presentation was not mentioned, nor was a mutational analysis of p53 performed in that study.
Although the number of cases in our study is too small to allow any firm conclusion to be drawn, it is important to mention that four of the five patients whose tumors revealed p53 mutations, presented with advanced disease and died soon after diagnosis. Secondly, seven of the nine patients that presented with clinical stages I and II from whom follow-up was available, are alive, six with no evidence of disease. The seventh patient had recently a local recurrence after 10 years of complete remission. In contrast, 11 of 13 patients with stages III and IV are dead, some of them despite combined treatment. (overall survival, 78% versus 15%; P = 0.0059) The high percentage of cases with advanced stages (III and IV) in this study, most probably reflects the fact that many patients in Mexico seek medical help late in their disease allowing it to take its natural course. Nevertheless, our data suggest that similar to other NHL subtypes,23,42,43 p53 mutations in nasal NK/T-cell lymphoma are associated with tumor progression and late-stage disease. The outcome of the patients in this series is in agreement with other studies that have identified clinical stage as probably the most important prognostic parameter.44-51 Although patients in stages I and II seem to have a good response to combined treatment, the relapse rate in extranodal sites, especially in those patients treated only with radiotherapy, is still relatively high and confers a poor prognosis.46,52,53
Interestingly, three of the five cases with p53 mutations revealed large cell morphology, in contrast to only 1 of the 20 cases without p53 mutations (60% versus 5%, P = 0.0162). To our knowledge, there are no studies so far that have analyzed the possible relationship of morphology and molecular changes and/or prognosis in nasal NK/T-cell lymphoma. Our study indicates that large cell morphology is associated with the presence of p53 mutation in this neoplasm.
In summary, p53 mutations in nasal NK/T-cell lymphoma from Mexico are present in 24% of the cases and show a good correlation with high overexpression of p53. LOH at chromosome 17p13 was found in 2 of 12 (17%) cases investigated, and associated with mutations of the remaining allele. p53 mutations are associated with large cell morphology and advanced disease, and thus, might be involved in tumor progression. So far, clinical stage seems to be the most important prognostic indicator. The overexpression of p53 and p21 independent of p53 gene status, although as yet not clear, might be the result of EBV infection and an important mechanism of viral pathogenesis.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported in part by grants from the Hochschulsonderprogramm of the Technical University of Munich (to M. N.) and by the German Science Fund Wilhelm Sander Stiftung (to L. Q. M. and F. F.).
Accepted for publication August 17, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B. J Virol 1996, 70:4849-4853[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  F. Jardin, P. Ruminy, J.-P. Kerckaert, F. Parmentier, J.-M. Picquenot, S. Quief, C. Villenet, G. Buchonnet, M. Tosi, T. Frebourg, et al. Detection of somatic quantitative genetic alterations by multiplex polymerase chain reaction for the prediction of outcome in diffuse large B-cell lymphomas Haematologica, April 1, 2008; 93(4): 543 - 550. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Pagano, A. Gallamini, G. Trape, L. Fianchi, D. Mattei, G. Todeschini, A. Spadea, S. Cinieri, E. Iannitto, M. Martelli, et al. NK/T-cell lymphomas 'nasal type': an Italian multicentric retrospective survey Ann. Onc., May 1, 2006; 17(5): 794 - 800. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-X. Li, B. Yao, J. Jin, W.-H. Wang, Y.-P. Liu, Y.-W. Song, S.-L. Wang, X.-F. Liu, L.-Q. Zhou, X.-H. He, et al. Radiotherapy As Primary Treatment for Stage IE and IIE Nasal Natural Killer/T-Cell Lymphoma J. Clin. Oncol., January 1, 2006; 24(1): 181 - 189. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-S. Chim, S.-Y. Ma, W.-Y. Au, C. Choy, A. K. W. Lie, R. Liang, C.-C. Yau, and Y.-L. Kwong Primary nasal natural killer cell lymphoma: long-term treatment outcome and relationship with the International Prognostic Index Blood, January 1, 2004; 103(1): 216 - 221. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
