Veterinary Pathology Online Prognostic Factors for Treated Canine Malignant Lymphoma

Transcription

Veterinary Pathology Online Prognostic Factors for Treated Canine Malignant Lymphoma
Veterinary Pathology Online
http://vet.sagepub.com/
Prognostic Factors for Treated Canine Malignant Lymphoma
M. Kiupel, E. Teske and D. Bostock
Vet Pathol 1999 36: 292
DOI: 10.1354/vp.36-4-292
The online version of this article can be found at:
http://vet.sagepub.com/content/36/4/292
Published by:
http://www.sagepublications.com
On behalf of:
American College of Veterinary Pathologists, European College of Veterinary Pathologists, & the Japanese College of Veterinary
Pathologists.
Additional services and information for Veterinary Pathology Online can be found at:
Email Alerts: http://vet.sagepub.com/cgi/alerts
Subscriptions: http://vet.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Jul 1, 1999
What is This?
Downloaded from vet.sagepub.com by guest on September 9, 2014
Vet Pathol 36:292–300 (1999)
Prognostic Factors for Treated Canine Malignant Lymphoma
M. KIUPEL, E. TESKE,
AND
D. BOSTOCK
Institut fu¨r Veterina¨r-Pathologie der Freien Universita¨t Berlin, Berlin, Germany (MK1);
Department of Clinical Sciences of Companion Animals, Utrecht University, Utrecht, The Netherlands (ET); and
Department of Clinical Veterinary Medicine, University of Cambridge, Cambridge, UK (DB)
Abstract. The aim of this study was to investigate the prognostic importance of different clinical, immunohistologic and tumorproliferation characteristics in dogs with malignant lymphoma treated with chemotherapy.
From 74 dogs with malignant lymphoma at least one enlarged peripheral lymph node was taken for biopsy
before chemotherapy following a standardized protocol (vincristine, cyclophosphamide, prednisolone, doxorubicin, and L-asparaginase). The variables evaluated as prognostic factors were age, sex, and tumor stage, as
well as histomorphologic grade (Kiel classification, Working Formulation), immunophenotype (using markers
for CD3 and CD79a), and cell proliferation (Ki-67, proliferation cell nuclear antigen, mitotic index, and argyrophil nucleolar organizer regions [AgNORs]) in extirpated lymph nodes. All markers were used on routinely
formalin-fixed, paraffin-embedded tissues. The AgNORs were assessed qualitatively, based on the AgNOR
pattern distribution, and quantitatively using image analysis and routine counting. In both univariate and multivariate survival analyses, AgNORs were a valuable prognostic marker for the treatment of canine malignant
lymphomas. Based on the results of the multivariate analysis longer survival time correlated with a B-cell type,
a larger mean AgNOR area, a larger total AgNOR area, a shorter distance between two AgNORs, and a smaller
AgNOR area to nucleus ratio. Longer disease-free survival time correlated with a smaller number of AgNORs
per nucleus, a larger mean AgNOR area, a larger maximal AgNOR area, and a larger total AgNOR area. This
study clearly demonstrates the additional benefit of the use of AgNORs in predicting treatment outcome in
dogs with malignant lymphoma.
Key words: Argyrophil nucleolar organizer region; canine malignant lymphoma, immunophenotype; Ki-67
monoclonal antibody; prognostic factors; proliferation cell nuclear antigen.
The behavior of canine malignant lymphomas is difficult to predict on the basis of histomorphology alone,
but the vast majority of canine malignant lymphomas
are high-grade malignant by most classifications.16,36,42
Histologic and immunologic classifications have been
shown to have some prognostic significance.6 Some
studies have applied the Kiel classification30 and the
Working Formulation to canine malignant lymphomas.7,19,28,36,41 In one study the Kiel classification was
found to be a prognostic factor for time to relapse in
treated dogs with malignant lymphoma, whereas the
Working Formulation was of prognostic significance
for the survival time.41 In a previous study the Kiel
classification was found to be of prognostic importance for overall survival time of treated and untreated
dogs with malignant lymphoma.28 Only a few studies
investigated the prognostic importance of the immunophenotype of canine malignant lymphomas1,41 and
found a poor prognosis for the survival time and the
disease-free survival time of dogs with T-cell lymphoma versus dogs with B-cell lymphoma.41 Nevertheless,
1 Present address: Animal Disease Diagnostic Laboratory,
Purdue University, West Lafayette, IN.
some differences occur within each of these groups
with respect to presentation, clinical causes, and response to treatment and prognosis. Therefore, additional parameters are needed to predict treatment outcome. The role of proliferation activity has received
special attention. The Ki-67 monoclonal antibody recognizes an antigen expressed in all phases of the cell
cycle except G0.18 Ki-67 has been reported as an important prognostic marker in human non-Hodgkin’s
lymphomasS24 however, to our knowledge no results of
Ki-67 in canine malignant lymphomas have been published. The proliferation cell nuclear antigen (PCNA)
is synthesized during the G1-phase, is maximal during
the S-phase, decreases during the G2-phase, and is
minimal during the M-phase and the G0-phase.43 Prognostic significance of the PCNA count has been shown
in canine mast cell tumors40 and canine mammary tumors,31 but not in canine malignant lymphomas.28,44
The disadvantages of these proliferation markers are
that they only give an indication of the percentage of
cells cycling, but no kinetics or dynamics can be derived from these proliferation marker labeling indices.4
The argyrophil nucleolar organizer regions (AgNORs)
may give additional information. AgNORs are proteins
292
Downloaded from vet.sagepub.com by guest on September 9, 2014
Vet Pathol 36:4, 1999
Prognostic Factors for Canine Malignant Lymphomas
associated with loops of DNA in the nucleoli in which
ribosomal RNA transcription takes place. The amount
of AgNORs not only relates to the percentage of cells
cycling but is also increased when the cell cycle is
faster. AgNORs are now widely used as a marker of
proliferation in tumor pathology in humans9,14 and
dogs2,40 and have been shown to be beneficial for the
grading of canine malignant lymphomas27,44 and human non-Hodgkin’s lymphomas.12,25,45
The purpose of this study was to investigate the additional benefit of the use of the proliferation markers
AgNOR, Ki-67, PCNA, and mitotic index, together
with the immunohistologic classification of the tumor
and the characteristics age, gender, and clinical tumor
stage to predict survival in dogs with malignant lymphomas treated with chemotherapy.
Materials and Methods
Animals
Specimens of lymph node biopsies were studied from a
series of 74 dogs with multicentric malignant lymphoma. All
dogs had enlarged lymph nodes and diagnosis of malignant
lymphoma was based on histopathologic evaluation of at
least one enlarged lymph node. Twenty-seven dogs were cases referred to the Department of Clinical Veterinary Medicine, University of Cambridge, over a period of 4 years.
Forty-seven lymphomas were collected during a period of 5
years in the Clinic for Companion Animals of the University
of Utrecht. No dog had received any form of cytostatic therapy at the time of the biopsy.
Treatment
Following clinical evaluation, staging, and biopsy all dogs
were treated with chemotherapy. The 27 dogs from Cambridge were treated with vincristine, cyclophosphamide, and
prednisolone8 and the 47 dogs from Utrecht received a more
advanced protocol, consisting of doxorubicin, L-asparaginase, vincristine, cyclophosphamide, and prednisolone.32,41
Survival data
Disappearance of all measurable tumor was considered to
be a complete response and only in these dogs the diseasefree survival was calculated as the time from the initial diagnosis until relapse or the date on which the dog was last
known to be free of disease, counting only relapses as
events. Overall survival was calculated as the time from initial diagnosis to death or to the date on which the dog was
last known to be alive, including all dogs, and counting only
deaths due to lymphoma as events.
Staging
A hematologic profile, bone marrow aspirates, and radiographs of the abdomen and thorax were taken from 54 dogs
for staging purposes. Staging was based on the modified
World Health Organization staging system for canine malignant lymphomas.35
293
Biopsy material
Tissues were fixed in 10% formol–saline and routinely
embedded in paraffin wax. Three-micrometer serial sections
were stained by hematoxylin and eosin, as well as Giemsa,
for histologic evaluation and estimation of the mitotic index.
Following the diagnosis of malignant lymphoma, histomorphologic grading of the tumor was performed in all 74 dogs
using both the Kiel classification29 and the Working Formulation (National Cancer Institut).34
Immunophenotype
For immunophenotyping CD3 was used for T-cell labeling
and CD79a was used for B-cell labeling. Sections of lymph
nodes from all 74 dogs were incubated in a 10-mM HCl/
citrate buffer (pH 6.0) and heated in a microwave oven (600
W) for 10 minutes following 20 minutes acclimatization at
room temperature. The slides were then transferred to 100%
alcohol and immersed in methanol containing 1.0% hydrogen peroxide for 30 minutes to block endogenous peroxidase
activity. After washing with distilled water and phosphatebuffered saline (PBS) with a pH of 7.6, nonspecific antibody
binding was blocked by means of normal goat serum for
CD3 or normal equine serum for CD79a for 15 minutes.
Immunostaining was performed by incubation in rabbit antiCD3 (DAKO, High Wycombe, UK) or mouse anti-CD79a
(DAKO) antibody at a dilution of 1 : 600 in 10% normal goat
or equine serum for 60 minutes. The antibody binding was
localized with the Vector ABC Elite kit (Vectastain PK6102,
Vector Laboratories, Peterborough, UK) and visualized with
3,3⬘-diaminobenzidine substrate (Vector SK4100). After a final washing in distilled, deionized water, the sections were
counterstained with Mayer’s hematoxylin, then dehydrated,
cleared, and mounted in DPX (BDH, UK). Canine lymphoid
tissue was used as a positive control. The immunophenotype
was assigned to an investigated canine malignant lymphoma
based on the majority and distribution of cells positive for
CD3 or CD79a.
PCNA
For PCNA immunostaining, sections of lymph nodes from
all 74 dogs were processed as previously described.28 Human
tonsillar tissue was used as a positive control. Evaluation of
the percentage of PCNA-positive cells was performed by
counting the number of cells with intensely positively
stained nuclei per 100 randomly selected cells as described
by Kiupel et al.28 No difference was found in the evaluation
of the percentage of PCNA-positive cells when counting
100, 200, or 300 cells.
Ki-67
For Ki-67 immunostaining, sections of lymph nodes from
all 74 dogs were incubated in a 10 mM HCl/citrate buffer
(pH 6.0) and cooked in a microwave oven (600 W) for 25
minutes following overnight acclimatization in room temperature. The slides were then transferred to 100% alcohol
and immersed in methanol containing 1.0% hydrogen peroxide for 30 minutes to block endogenous peroxidase activity. After washing with distilled water and PBS with a pH
of 7.6, nonspecific antibody binding was blocked by means
Downloaded from vet.sagepub.com by guest on September 9, 2014
294
Kiupel, Teske, and Bostock
of normal equine serum for 15 minutes. Immunostaining was
performed by incubation in mouse anti-MIB-1 antibody (Immunotech S.A., Marseille, France) at a dilution of 1 : 200 in
10% normal equine serum for 60 minutes. The antibody
binding was localized with the Vector ABC Elite kit (Vector
PK4000) and visualized with 3,3⬘-diaminobenzidine substrate (Vector SK4100). After a final washing in distilled,
deionized water, the sections were counterstained with Mayer’s hematoxylin, then dehydrated, cleared, and mounted in
DPX (BDH). Equine lymphoid tissue was used as a positive
control. The percentage of Ki-67-positive cells was determined by counting the number of Ki-67-positive cells per
100 randomly selected cells. Necrotic areas were excluded
as well as histiocytes, plasma cells, or other inflammatory
cells. No difference was found in the evaluation of the percentage of Ki-67-positive cells when counting 100, 200, or
300 cells.
AgNORs
Following the diagnosis of malignant lymphoma, sections
were stained for AgNORs with a one-step silver colloid
method, as described by Ploton et al.,37 using a previously
published protocol.27 Evaluation of AgNORs was performed
quantitatively and qualitatively. The mean AgNOR number
(NORNBC), the mean AgNOR area (MEANAR), the maximal AgNOR area (MAXNOR), the total AgNOR area
(S㛮AREA), the ratio between the total AgNOR area and the
nucleus (ARRAT), the ratio between the mean AgNOR area
and the nucleus (SIZERAT), the number of AgNORs located
centrally in the nucleus (CENTER), the number of AgNORs
located peripherally in the nucleus (BORDER), the position
of AgNORs between the point of balance and the nuclear
membrane (LOCAT), the mean distance between two
AgNORs (MDIST), and the maximal distance between two
AgNORs (MAXDIST) were determined on 100 neoplastic
cells using image analysis.22 AgNOR pattern distribution was
evaluated based on previous studies by Field et al.,17 Crocker
et al.,10 Hansen and Ostergard,21 Ru¨schoff et al.,39 and our
own investigations.27 We described five different types of
AgNOR pattern distribution. Depending on the majority of
cells expressing a similar pattern distribution, a grade was
assigned to each malignant lymphoma.27
Statistics
Statistical analysis of data was performed using the SPSS
and EGRET statistical packages. Survival curves were
drawn with the Kaplan–Meier method. Univariate tests for
comparison of groups of survival data were made with the
log-rank test and with a proportional hazard logistic regression model. Those parameters with P values less than 0.3 in
the univariate analysis were included in a multivariate Cox
proportional hazards model regression analysis with a forward step selection. In all analyses stratification on clinic
center was performed. A value of P ⬍ 0.05 was considered
significant.
Results
Of the 74 dogs treated with chemotherapy, 48 dogs
responded completely and only in these dogs disease-
Vet Pathol 36:4, 1999
free survival could be calculated. Clinical tumor stage
was only determined in 54 dogs and for one of these
dogs no immunophenotype could be determined. Because stage and immunophenotype were included in
the multivariate analyses for overall survival, these statistics were only performed in 53 dogs.
Age, weight, sex, and clinical stage had no prognostic significance for the overall survival time and
disease-free survival time of chemotherapeutically
treated dogs with malignant lymphoma. Histomorphology was also found to be an unreliable criterion
for the prognostic evaluation of canine malignant lymphomas. Only immunophenotype and proliferation
markers were significant for predicting overall survival
time and disease-free survival time.
In the univariate analysis, AgNOR area parameters
(mean AgNOR area: P ⫽ 0.001, maximal AgNOR
area: P ⫽ 0.002) and the mean AgNOR number (P ⫽
0.001) were of prognostic significance for the diseasefree survival time (Table 1). The AgNOR type was
only significant between groups of dogs with lymphomas of type 1 and 4 (P ⫽ 0.012) and the mitotic index
was only significant for groups of dogs with malignant
lymphomas with a value between 9 and 12 (P ⫽ 0.02).
For the overall survival time the following parameters
were of prognostic significance in the univariate analysis: PCNA (P ⫽ 0.009) and mitotic index (P ⫽
0.028), as well as AgNOR area parameters (mean
AgNOR area: P ⫽ 0.001, maximal AgNOR area: P ⫽
0.001), mean AgNOR number (P ⫽ 0.001), AgNOR
location (P ⫽ 0.019), and AgNOR type (P ⫽ 0.008).
Ki-67 was only significant between groups of dogs
with lymphomas with less than 20% Ki-67-positive
cells and with 40–60% Ki-67-positive cells. All other
investigated parameters were of no prognostic significance in the univariate analysis.
In the multivariate analysis, AgNOR area parameters (mean AgNOR area: P ⫽ 0.001, total AgNOR
area: P ⫽ 0.001, maximal AgNOR area: P ⫽ 0.01)
and the mean AgNOR number (P ⫽ 0.001) were of
prognostic significance for disease-free survival (Table
2). The immunophenotype (P ⫽ 0.027), AgNOR area
parameters (mean AgNOR area: P ⫽ 0.001, total
AgNOR area: P ⫽ 0.003), and AgNOR location parameters (ratio AgNOR area : nucleus: P ⫽ 0.008,
maximal AgNOR distance: P ⫽ 0.009) were of prognostic significance for overall survival time (Table 3).
All other investigated parameters were of no prognostic significance in the multivariate analysis. The disease-free survival time correlated strongly with the different groups of mean AgNOR numbers. This correlation is shown for all three groups of dogs with malignant lymphoma based on the mean AgNOR number
in Fig. 1. The overall survival time correlated strongly
with the different groups of mean AgNOR areas. This
Downloaded from vet.sagepub.com by guest on September 9, 2014
Prognostic Factors for Canine Malignant Lymphomas
Vet Pathol 36:4, 1999
correlation is shown for all three groups of dogs with
malignant lymphoma based on the mean AgNOR area
in Fig. 2.
Discussion
The use of antibodies for the assessment of cell proliferation and immunophenotyping has been limited,
because many are suitable on frozen sections only.
However, new markers have become available (MIB1,
CD79a) and antigen retrieval with the microwave oven
allows the application of these markers to routinely
processed paraffin sections. In this study we were able
to evaluate a whole panel of different prognostic markers in canine malignant lymphomas on routinely processed material.
AgNORs are a valuable prognostic marker for the
treatment of canine malignant lymphomas. Based on
the results of the multivariate analysis, longer diseasefree survival time correlated with a smaller number of
AgNORs per nucleus, a larger mean AgNOR area, a
larger maximal AgNOR area, and a larger total
AgNOR area. Longer overall survival time correlated
with a B-cell type, a larger mean AgNOR area, a larger
total AgNOR area, a shorter distance between two
AgNORs, and a smaller AgNOR area to nucleus ratio.
In the univariate analysis mitotic index, PCNA count,
and AgNOR type were of prognostic significance for
the overall survival. However, in the multivariate analysis this could not be confirmed.
AgNOR numbers in canine malignant lymphomas
have only been evaluated in two other studies.28,44 In
both studies the mean AgNOR number was a significant prognostic parameter for the overall survival time.
However, no multivariate survival analysis was performed in these studies. Studies of proliferation activity using multiple different markers, including AgNORs, are rare in dogs and were performed by Simoes
et al.40 in mast cell tumors, by Loehr et al.31 in canine
mammary tumors, and by Vail et al.44 and Kiupel et
al.28 in canine malignant lymphomas. For mast cell tumors and canine mammary tumors PCNA was suggested as a useful prognostic marker. In canine malignant lymphomas AgNOR frequency was superior to
PCNA in predicting outcome after therapy. AgNORs
have the advantage that they are only increased in actively dividing cells,5 whereas anti-PCNA simply labels proliferating cells in the S-phase of the cell cycle
and cells going through DNA repair.4 In addition, high
PCNA counts might be the result of false positive cells
due to the long half-life of PCNA.3 Ki-67, which labels
cycling cells, was of prognostic significance for canine
mammary tumors in one study.31 In our investigations
Ki-67 was found to be of no prognostic value. A high
mitotic index indicated higher malignancy and a poor
prognosis in this study. However, the number of dogs
295
with lymphomas with a high mitotic index is small and
therefore the practical application of the mitotic index
for prognostic purposes is limited.
Most proliferation markers only give an indication
of the proportion of cells cycling but not of the speed
of proliferation. In a tumor 80% of cells could be cycling, but the cell cycle could last 1 month. Compared
to a tumor with only 25% cells cycling, but with a cell
cycle length of 2 days, the later would of course be
more aggressive. The majority of canine malignant
lymphomas, as in this study, are of high grade and the
percentage of cycling cells is high in most canine malignant lymphomas. This may explain that the assessment of the growth fraction by Ki-67 was less useful
in identifying prognostic differences. Furthermore, the
ratio of cells undergoing apoptosis plays an important
role in assessing proliferation activity and no investigations to measure the ratio of apoptosis were conducted in this study. Mitotic index and PCNA represent the phase-index and their measurement is not only
difficult to standardize, but the phase-index would only
be useful to identify correlations between karyologic
abnormalities and survival. In contrast, an increased
speed of cell proliferation indicates a high metabolic
activity and probably correlates with the rate of tumor
cell infiltration and metastasis. AGNORs are the only
parameter available on formalin-fixed tissue to measure proliferation speed and should therefore be considered the most important proliferation marker in canine malignant lymphomas.
AgNORs have been characterized as the most powerful tool for the prognostic evaluation of human nonHodgkin’s lymphomas.12,20,23 The total number of
AgNORs per nucleus is a parameter of proliferation13,15,38 and the mean number of AgNORs per nucleus is correlated with tumor grade and survival times
in human non-Hodgkin’s lymphoma.11 Even though
different studies varied greatly in the estimated mean
AgNOR number, increasing malignancy was always
correlated to an increased mean AgNOR number per
nucleus.11,20,26,45
The major problems of using AgNOR counts are the
diversity of parameters and their interpretation and dependence of the staining protocol on external conditions (fixation, temperature, staining time), which
makes a standardized AgNOR protocol difficult to
achieve.33 The use of image analysis38 for the investigation of AgNOR parameters has to be considered as
state of art to avoid methodical errors. We are not
aware of complex image analysis studies of AgNORs
in canine tumors and specifically malignant lymphomas. Further investigations are being carried out on
the value of AgNOR cluster formation and pattern distribution in prognosis.39 These parameters are less susceptible to variation than individual AgNOR counts
Downloaded from vet.sagepub.com by guest on September 9, 2014
296
Kiupel, Teske, and Bostock
Vet Pathol 36:4, 1999
Table 1. Results of univariate analysis of factors predicting recurrence after complete remission (using disease-free
survival time) and death (using overall survival time) in canine malignant lymphoma, stratified to institution. Characteristics
are defined in the text.
Disease-Free Survival Time
(n ⫽ 48)
Patient’s Characteristic
P Value
Age
0.079
Weight (kg)
0–20
21–30
31–40
⬎40
0.733
Sex
Female
Male
Hazard Ratio
P Value
0.8872
0.165
Hazard Ratio
0.9249
0.549
0.522
0.791
0.911
0.6856
1.155
0.9375
0.212
0.801
0.554
0.5316
0.9015
0.7593
0.436
0.7269
0.116
0.6409
Stage
Stage II ⫹ III
Stave IV
Stage V
0.435
Kiel classificiation
Low malignant
High malignant
0.198
0.483
0.582
0.217
Working Formulation
Low grade
Medium grade
High grade
0.269
Immunophenotype
B-cell lymphomas
T-cell lymphomas
0.932
0.311
0.077
0.932
MITOSE
0–9
10–12
⬎12
0.137
PCNA
0–20
21–40
40–60
⬎60
0.226
Ki-67
0–20
21–40
41–60
⬎60
0.207
TYPE
1
2
3
4
5
0.118
NORNBC
0–3.00
3.01–6.00
⬎6.00
Overall Survival Time
(n ⫽ 74)
0.390
0.020
0.436
0.14
0.059
0.704
0.158
0.087
0.889
0.320
0.012
0.082
0.076
0.7235
1.266
0.6928
1.675
0.408
1.767
0.415
1.288
0.504
1.990
9.466
0.320
0.362
1.518
2.130
0.229
1.032
0.239
0.6987
0.028
1.597
17.36
0.287
0.005
1.670
6.380
0.009
1.470
2.124
2.875
0.457
0.033
0.003
1.368
2.506
3.890
0.078
1.240
2.134
2.670
0.439
0.024
0.131
1.405
2.676
2.107
0.008
1.098
1.762
28.28
3.063
⬍0.001
0.399
0.008
0.417
0.154
0.668
0.147
0.007
0.007
1.249
1.968
10.68
3.889
⬍0.001
1.772
6.647
Downloaded from vet.sagepub.com by guest on September 9, 2014
0.182
⬍0.001
1.877
9.340
Vet Pathol 36:4, 1999
Prognostic Factors for Canine Malignant Lymphomas
Table 1.
Continued.
Disease-Free Survival Time
(n ⫽ 48)
Patient’s Characteristic
P Value
S㛮AREA
0–2.0000
2.0001–3.0000
3.0001–4.0000
4.0001–5.0000
⬎5.0000
0.075
MEANAR
0–0.5000
0.5001–1.0000
⬎1.0000
⬍0.001
MAXNOR
0–1.0000
1.0001–1.5000
1.5001–2.0000
⬎2.0000
0.002
ARRAT
0–110
⬎110
0.138
SIZERAT
0–10
11–20
21–30
⬎30
0.959
CENTER
0–0.40
0.41–0.60
0.61–0.80
⬎0.80
0.490
BORDER
0–0.75
0.76–1.50
1.51–2.25
⬎2.25
0.217
LOCAT
0–0.475
0.476–0.550
⬎0.550
0.210
MAXDIST
0–0.425
0.426–0.500
0.501–0.575
⬎0.575
0.302
MDIST
0–0.325
0.326–0.375
0.376–0.425
⬎0.425
0.458
0.793
0.586
0.658
0.177
0.014
⬍0.001
0.574
0.004
0.003
0.153
0.731
0.733
0.913
0.551
0.917
0.452
0.069
0.222
0.103
0.708
0.126
0.155
0.093
0.149
0.207
0.373
0.147
297
Hazard Ratio
Overall Survival Time
(n ⫽ 74)
P Value
Hazard Ratio
0.074
1.240
1.564
1.438
0.175
0.189
0.349
0.471
0.321
1.987
1.632
1.583
0.4579
⬍0.001
0.2647
0.0760
0.022
⬍0.001
0.4060
0.0314
⬍0.001
0.7783
0.1797
0.1360
0.095
⬍0.001
⬍0.001
0.5625
0.1784
0.1161
0.103
5.849
0.096
2.415
0.271
1.152
0.8070
0.8650
0.244
0.49
0.322
1.455
0.6879
0.5164
0.437
0.6546
1.084
0.5163
0.273
0.812
0.271
0.5937
0.8822
0.5180
0.327
2.666
1.891
2.776
0.156
0.128
0.191
1.750
1.707
2.031
0.019
1.278
2.209
0.289
0.017
1.713
3.017
0.224
2.718
3.068
2.743
0.135
0.056
0.139
2.108
2.506
2.140
0.453
2.085
1.741
3.785
Downloaded from vet.sagepub.com by guest on September 9, 2014
0.194
0.297
0.164
1.886
1.690
2.730
298
Kiupel, Teske, and Bostock
Table 2. Results of multivariate analysis of factors predicting recurrence after achieving complete remission with
chemotherapy in dogs with malignant lymphoma. Characteristics are defined in the text.
Disease-Free Survival Time
(n ⫽ 48)
Patient’s Characteristic
NORNBC
0–3.00
3.01–6.00
⬎6.00
P Value
0.001
0.002
⬍0.001
S㛮AREA
0–2.0000
2.0001–3.0000
3.0001–4.0000
4.0001–5.0000
⬎5.0000
⬍0.001
MEANAR
0–0.5000
0.5001–1.0000
⬎1.0000
0.003
MAXNOR
0–1.0000
1.0001–1.5000
1.5001–2.0000
⬎2.0000
Hazard Ratio
0.002
0.070
0.521
0.001
⬍0.001
0.588
26.09
63.80
0.0363
0.1699
0.5246
0.0030
Table 3. Results of multivariate analysis of factors predicting death in dogs treated with chemotherapy for malignant lymphoma. Characteristics are defined in the text.
Survival Time (n ⫽ 53)
Patient’s Characteristic
P Value
Immunophenotype
B-cell lymphoma
T-cell lymphoma
0.027
S㛮AREA
0–2.0000
2.0001–3.0000
3.0001–4.0000
4.0001–5.0000
⬎5.0000
0.003
MEANAR
0–0.5000
0.5001–1.0000
⬎1.0000
⬍0.001
0.027
0.121
0.382
0.383
0.016
⬍0.001
⬍0.001
0.0725
0.5192
ARRAT
0–110
⬎110
0.008
0.009
2.456
0.0569
0.3414
MAXDIST
0–0.425
0.426–0.500
0.501–0.575
⬎0.575
0.010
0.156
0.027
0.397
Vet Pathol 36:4, 1999
* P ⬍ 0.001.
0.008
⬍0.001
0.005
0.039
Hazard Ratio
2.997
0.1994
0.4430
2.513
0.0414
0.1053
0.0058
37.45
13.16
10.64
6.278
* P ⫽ 0.026.
and should therefore be more reproducible. In a previous study27 the AgNOR type was suggested as an
important prognostic marker for untreated dogs with
malignant lymphoma. These results were confirmed
for treated dogs by the univariate analysis; however,
the results could not be confirmed in the multivariate
analysis. The smaller number of cases in the multivariate analysis and different response to chemotherapy may explain this difference.
Immunophenotype was an important prognostic
marker for the overall survival time of dogs with malignant lymphoma in this study. This result confirms
previous studies by Teske et al.41,42 However, in these
studies clinical stage and immunophenotype were
found to be prognostic factors for predicting the overall survival time and the disease-free survival time.41
Also, the Working Formulation had prognostic value
for predicting overall survival time in Teske et al. studies.41 Neither clinical stage nor histomorphology (Kiel
classification, Working Formulation) were found to be
of prognostic significance in the data presented here,
and the immunophenotype was also not prognostically
significant for the disease-free survival time in this
study. Different treatment protocols and individual differences in determining the clinical stage (results from
two different institutions were pooled) may be a major
factor to explain these contrary results.
In conclusion, this study clearly demonstrates the
benefit of AgNORs in predicting treatment outcome in
dogs with malignant lymphoma. In addition, it was
demonstrated that not only the number of AgNORs is
Fig. 1. Baseline disease-free survival curves for dogs
with malignant lymphoma with different mean AgNOR
numbers.
Downloaded from vet.sagepub.com by guest on September 9, 2014
Prognostic Factors for Canine Malignant Lymphomas
Vet Pathol 36:4, 1999
Fig. 2. Baseline overall survival curves for dogs with
malignant lymphoma with different mean AgNOR areas.
of importance, but also AgNOR area parameters such
as the mean AgNOR area. This underlines the importance of using image analysis for the evaluation of
AgNORs. Furthermore qualitative characteristics of
AgNORs such as AgNOR type may provide a simple
tool for the prognostic evaluation of canine malignant
lymphomas in routine laboratories.
Acknowledgements
Jane Dobson, Department of Clinical Veterinary Medicine, University of Cambridge, UK; Jane Oates and John
Crocker, Pathology Department, Birmingham Heartlands
Hospital, UK; and Dr. Peter Hufnagel, Institut fu¨r Pathologie, Charite´ Berlin, Germany, are gratefully acknowledged
for their assistance.
References
1 Appelbaum FR, Sale GE, Storb R, Charrier K, Deeg HJ,
Graham TC, Wulff JC: Phenotyping of canine lymphoma
with monoclonal antibodies directed at cell surface antigens: classification, morphology, clinical presentation
and response to chemotherapy. Haematol Oncol 2:151–
167, 1984
2 Bostock DE, Crocker J, Harris K, Smith P: Nucleolar
organizer regions as indicators of post-surgical prognosis
in canine spontaneous mast cell tumours. Br J Cancer
59:915–918, 1989
3 Bravo R, Frank R, Blundell PA, McDonald BH: Cycline/
PCNA is the auxiliary protein of DNA polymerase-delta.
Nature 326:515–517, 1987
4 Brugal G: Interpretation of proliferation markers. Comp
Cytol Histol Labor 234–239, 1994
5 Busch H, Lischwe MA, Michalik J, Chan PK, Busch R:
Nucleolar proteins of special interest: silver-staining proteins B23 and C23 and antigens of human tumor nucleoli. In: The Nucleolus, ed. Jordan EG, and Cullis CA,
pp. 43–72. Cambridge University Press, London, UK,
1984
6 Capurro C, Buracco P, Rossi L: Lymphoma in dogs. Veterinaria 4:15–27, 1990
299
7 Carter RF, Valli VEO, Lumsden JH: The cytology, histology and prevalence of cell types in canine lymphoma
classified according to the National Cancer Institute
working formulation. Can J Vet Res 50:154–164, 1986
8 Cotter SM: Treatment of lymphoma and leukemia with
cyclophosphamide, vincristine and prednisolone. I.
Treatment of dogs. J Am Anim Hosp Assoc 23:159–172,
1983
9 Crocker J: Nucleolar organizer regions. Curr Top Pathol
82:91–143, 1990
10 Crocker J, Boldy D, Egan MJ: How should we count
AgNORs? Proposals for a standardized approach. J
Pathol 158:185–188, 1989
11 Crocker J, Egan MJ: Correlation between NOR sizes and
numbers in non-Hodgkin’s lymphomas. J Pathol 156:
233–239, 1988
12 Crocker J, Nar P: Nucleolar organizer regions in lymphomas. J Pathol 151:111–118, 1987
13 Derenzini M, Pession A, Trere D: The quantity of nucleolar silver-stained proteins is related to proliferating
activity in cancer cells. Lab Invest 63:137–140, 1990
14 Derenzini M, Trere D: Importance of interphase nucleolar organizer regions in tumour pathology. Virchows
Arch B 61:1–8, 1991
15 Derenzini M, Trere D: AgNOR proteins as a parameter
of the rapidity of cell proliferation. Zentralbl Pathol 140:
7–10, 1994
16 Dorn CR, Taylor DON, Hibbard HH: Epizootiologic
characteristics of canine and feline leukemia and lymphoma. Am J Vet Res 28:993–1001, 1967
17 Field DH, Fitzgerald PH, Sin FY: Nucleolar silver-staining patterns related to cell cycle phase and cell generation of PHA-stimulated human lymphocytes. Cytobios
41:23–33, 1984
18 Gerdes J, Li L, Schlueter C, Duchrow M, Wohlenberg
C, Gerlach C, Stahmer I, Kloth S, Brandt E, Flad HD:
Immunobiochemical and molecular biologic characterization of the cell proliferation-associated nuclear antigen is defined by monoclonal antibody Ki-67. Am J
Pathol 138:867–873, 1991
19 Greenlee PG, Filippa DA, Quimby FW, Patnaik AK, Calvano SE, Matus RE, Kimmel M, Hurvitz AI, Liebermann
PH: Lymphomas in dogs. A morphologic, immunologic
and clinical study. Cancer 66:480–491, 1990
20 Hall PA, Richards MA, Gregory WM, d’Ardenne AJ,
Lister TA, Stansfeld AG: The prognostic value of Ki-67
immunostaining in non-Hodgkin’s lymphoma. J Pathol
154:223–231, 1988
21 Hansen AB, Ostergard B: Nucleolar organizer regions in
hyperplastic and neoplastic prostatic tissue. Virchows
Arch A 147:9–13, 1990
22 Hufnagel P, Guski H, Schulz HJ: Measuring of AgNORs
using image analysis. Zentralbl Pathol 140:31–35, 1994
23 Jakijc-Razumovijc J, Tentor D, Petroveicki M, Radman
I: Nucleolar organizer regions and survival in patients
with non-Hodgkin’s lymphomas classified by the Working Formulation. J Clin Pathol 46:943–947, 1993
24 Jan-Mohamed RM, Murray PG, Crocker J, Leyland MJ:
Sequential demonstration of nucleolar organizer regions
Downloaded from vet.sagepub.com by guest on September 9, 2014
300
25
26
27
28
29
30
31
32
33
34
Kiupel, Teske, and Bostock
and Ki-67 immunolabelling in non-Hodgkin’s lymphomas. Clin Lab Haematol 12:395–399, 1990
Kalir T, Chan KS, Liu Z, Strauchen J, Gil J: Semi-automatic quantitation of nucleolar organizer regions in
non-Hodgkin’s lymphomas. Pathol Res Pract 190:124–
128, 1994
Kim I, Park SH, Paik SY: Nucleolar organizer regions
of lymphomas in Korea. J Korean Med Sci 3:99–105,
1988
Kiupel M, Bostock DE, Bergmann V: A semiquantitative
method for the evaluation of AgNORs in canine malignant lymphomas. Eur J Vet Pathol 4:67–71, 1998
Kiupel M, Bostock DE, Bergmann V: The prognostic
significance of AgNOR- and PCNA-counts and histopathological grading of canine malignant lymphomas. J
Comp Pathol 119:407–418, 1998
Lennert K, Feller AC: Histopathologie der Non-Hodgkin-Lymphome, 2nd ed. Springer Verlag, Berlin, Germany, 1990
Lennert K, Mohri N: Malignant Lymphomas other than
Hodgkin’s Disease: Histopathology and Diagnosis of
Non-Hodgkin’s Lymphomas. Springer Verlag, New
York, NY, 1978
Loehr CV, Teifke JP, Failing K, Weiss E: Characterization of the proliferation state in canine mammary tumors
by the standardized AgNOR method with postfixation
and immunohistologic detection of Ki-67 and PCNA. Vet
Pathol 34:212–221, 1997
MacEwen EG, Rosenthal RC, Fox LE, Loar AS, Kurzman ID: Evaluation of L-asparaginase: polyethylene glycol conjugate versus native L-asparaginase combined
with chemotherapy. A randomized double blind study in
canine lymphoma. J Vet Intern Med 6:230–234, 1992
Munakata S, Hendricks JB: Morphometric analysis of
AgNORs in imprints and sections from non-Hodgkin’s
lymphomas. An approach to standardization. Anal Quant
Cytol Histol 15:329–334, 1993
National Cancer Institut: The non-Hodgkin’s lymphoma
pathologic classification project: National Cancer Institut
sponsored study of classifications of non-Hodgkin’s lymphomas: summary and description of a working formulation for clinical usage. Cancer 49:2112–2135, 1982
Vet Pathol 36:4, 1999
35 Owen LN, ed.: TNM Classification of Tumours in Domestic Animals, pp. 46–47. World Health Organization.
Geneva, Switzerland, 1980
36 Parodi AL, Dargent F, Crespeau F: Histological classification of canine malignant lymphomas. J Vet Med A 35:
178–192, 1988
37 Ploton D, Menager M, Jeannesson P, Himber G, Pigeon
F, Adnet JJ: Improvement in the staining and in the argyrophilic protein of the nucleolar organizer region at
the optical level. Histochem J 18:5–14, 1986
38 Ru¨schoff J: Nukleolus organisierende Regionen (NORs)
in der pathomorphologischen Tumordiagnostik. Fischer
Verlag, Stuttgart, Germany, 1992
39 Ru¨schoff J, Fauser G, Knu¨chel R, Hofsta¨dter F: AgNOR
quantification with special reference to staining patterns.
Zentralbl Pathol 140:23–30, 1994
40 Simoes JPC, Schoning P, Butine M: Prognosis of canine
mast cell tumors: a comparison of three methods. Vet
Pathol 31:637–647, 1994
41 Teske E, van Heerde P, Rutteman GR, Kurzman I, Moore
PF, MacEwen EG: Prognostic factors for treatment of
malignant lymphoma in dogs. J Am Vet Med Assoc 205:
1722–1728, 1994
42 Teske E, Wisman P, Moore PF, van Heerde P: Histological classification and immunophenotyping of canine
non-Hodgkin’s lymphoma. Unexpected high frequency
of T-cell lymphomas with B-cell morphology. Exp Hematol 22:1179–1187, 1994
43 Trere D: AgNOR quantification in tumor pathology:
what is actually evaluated? Anat Cell Pathol 5:191–201,
1993
44 Vail DM, Kisseberth WC, Obradivich JE, Moore FM,
London CA, MacEwen EG, Ritter MA: Assessment of
potential doubling time (Tpot), argyrophilic nucleolar organizer regions (AgNOR), and proliferating cell nuclear
antigen (PCNA) as predictors of therapy response in canine non-Hodgkin’s lymphoma. Exp Hematol 24:807–
815, 1996
45 Yekeler H, Ozercan MR, Yumbul AZ, Ajgan M, Ozercan
IH: Nucleolar organizer regions in lymphomas: a quantitative study. Pathologica 85:353–362, 1993
Request reprints from Dr. M. Kiupel, 1175 ADDL, Purdue University, West Lafayette, IN 47907-1175 (USA).
Downloaded from vet.sagepub.com by guest on September 9, 2014