Radiol Oncol 2024; 58(1): 99-109. doi: 10.2478/raon-2024-0010
99
research article
The prognostic significance of programmed cell 
death protein 1 and its ligand on lymphoma 
cells and tumor-immune cells in diffuse large 
B-cell lymphoma, not otherwise specified
Teja Cas Slak
1,2
, Simona Miceska
1,2
, Gorana Gasljevic
3,4
, Lucka Boltezar
2,5
,  
Veronika Kloboves-Prevodnik
1,4
1 
Department of Cytopathology, Institute of Oncology Ljubljana, Ljubljana, Slovenia
2
 Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia
3
 Department of Pathology, Institute of Oncology Ljubljana, Ljubljana, Slovenia
4
 Faculty of Medicine, University of Maribor, Maribor, Slovenia
5 
Department of Medical Oncology, Institute of Oncology Ljubljana, Ljubljana, Slovenia
Radiol Oncol 2024; 58(1): 99-109.
Received 7 November 2023 
Accepted 25 November 2023
Correspondence to: Assoc. Prof. Veronika Kloboves-Prevodnik, M.D., Ph.D., Institute of Oncology Ljubljana, Zaloška 2, SI-1000 Ljubljana, 
Slovenia. E-mail: vkloboves@onko-i.si 
Disclosure: No potential conflicts of interest were disclosed. 
This is an open access article distributed under the terms of the CC-BY license (https://creativecommons.org/licenses/by/4.0/).
Background. Diffuse large B-cell lymphoma, not otherwise specified (DLBCL, NOS) is the most common type non-
Hodgkin’s lymphoma, where the treatment of relapsed/refractory cases is the major challenge. Programmed cell 
death protein 1 (PD-1) and its ligand PD-L1 play a crucial role in the negative regulation of the immune response 
against the disease. The aim of the study was to analyze the expression of PD-1 and PD-L1 on lymphoma cells (LCs) 
and tumor-immune cells (TICs) and to investigate their correlation with outcome. 
Patients and methods. Samples from 283 patients diagnosed with DLBCL, NOS (both germinal center B cell like 
[GCB] and non-GCB subtypes) were included in the study. Expression of PD-1 and PD-L1 was determined using double 
immunohistochemical staining (D-IHC) for PD-1/PAX5 and PD-L1/ PAX5 on tissue microarrays. LCs were highlighted by 
D-IHC to obtain more accurate results. Clinical data and histologic diagnoses were obtained from electronic data 
records. We correlated clinical characteristics, and PD-1 and PD-L1 expression on LCs and TICs with progression-free 
survival (PFS) and overall survival (OS). 
Results. Expression of PD-1 on TICs was observed in 38.4% and on LCs in 8.8% of cases, while PD-L1 was expressed on 
TICs in 46.8% and on LCs in 6.5% of cases. PD-L1 expression on LCs was more frequent in non-GCB subtype (p = 0.047). 
In addition, patients with PD-L1 expression on LCs had significantly shorter PFS (p = 0.015), and the expression retained 
significant in the multivariate model (p = 0.034). 
Conclusions. PD-L1 was more frequently expressed in LCs of the non-GCB subtype. Additionally, PD-L1 in LCs may 
predict shorter PFS time. D-IHC staining for PD-L1/PAX5 is a feasible method to assess PD-L1 expression on LCs of DLBCL, 
NOS patients and can be used to identify patients who may benefit from targeted immunotherapy with checkpoint 
inhibitors.
Key words: diffuse large B-cell lymphoma; immunohistochemistry; PD-1; PD-L1; PAX5
Introduction
Diffuse large B-cell lymphoma, not otherwise 
specified (DLBCL, NOS) is the most prevalent type 
of non-Hodgkin’s lymphoma (NHL) arising from 
a complex interplay of genetic and molecular fac-
tors. This heterogeneity results in approximately 
35% of DLBCL, NOS cases not responding to 

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 100
standard treatment method that combines rituxi-
mab with anthracycline-based chemotherapy. As 
a result, patients who do not respond to standard 
therapy experience relapsed or refractory disease, 
which remains the leading cause of mortality.
1,2
 In 
recent years, a variety of novel therapies, including 
immunotherapies, have emerged that may provide 
effective treatment strategies for DLBCL, NOS pa-
tients, particularly for relapsed or refractory dis-
ease. Identifying additional biomarkers and care-
fully assessing them to precisely define the effec-
tiveness of immunotherapy are therefore critical 
for improved treatment outcomes, which are being 
investigated in several ongoing studies.
3-5
The programmed cell death protein 1 (PD-1)/
programmed cell death ligand 1 (PD-L1) immune 
checkpoint pathway, which is crucial for main-
taining self-tolerance and excessive immune re-
sponses, has emerged as a novel biomarker target 
for various malignant neoplasms to suppress the 
anti-tumor immune response and evade immune 
surveillance.
6
 Immunotherapy based on the PD-
1/PD-L1 signaling pathway has already been in-
cluded in standard treatment guidelines for vari-
ous carcinomas such as melanoma, non-small cell 
lung cancer, urothelial carcinoma, triple negative 
breast cancer and many others.
7
 Moreover, it has 
also attracted wide attention for the treatment of 
lymphomas, with the greatest success achieved in 
classical Hodgkin’s lymphoma and T-cell lympho-
ma.
8-13
 However, in DLBCL, NOS, immunotherapy 
based on the PD-1/PD-L1 pathway has not yet 
become part of the standard treatment approach. 
Expression of PD-1 and PD-L1 in DLBCL, NOS is 
often elevated, leading to impaired immune cell 
function and tumor growth. The expression of 
PD-1 has been reported mainly on immune cells, 
emphasizing tumor-immune cells (TICs), while 
PD-L1 expression on lymphoma cells (LCs) and 
TICs. Some studies have already shown that in-
creased PD-1 expression on TICs is associated with 
favorable overall survival (OS) of DLBCL, NOS pa-
tients.
14
 In contrast, increased PD-L1 expression 
on LCs has been associated with poorer prognosis 
and increased resistance to chemotherapy
8
, which 
also correlates with the DLBCL, NOS non-germi-
nal center B-cell like (non-GCB) DLBCL, NOS sub-
type defined by the Hans algorithm.
15
 These data 
confirm their pivotal role in the tumor microen-
vironment of DLBCL, NOS and also their contri-
bution to the poor clinical outcomes.
8,14
 However, 
to date, reported data have yielded conflicting 
results, particularly in relation to PD-1 expression 
on TICs, which has been by some authors associ-
ated with better progression-free survival (PFS) 
and OS, while some other studies have found no 
association with survival.
1,6,16-20
 On the other hand, 
some findings revealed an association of PD-L1 ex-
pression on LCs with worse OS
1,6,8,14,16,18-21
, although 
there are data disputing the prognostic signifi-
cance of PD-L1 or even showing a correlation 
with better rather than worse outcomes. It is also 
worth noting that most studies on the expression 
of PD-1 and PD-L1 in DLBCL, NOS have focused 
on the Asian population, where there is a higher 
prevalence of non-GCB DLBCL, NOS subtypes
22
, 
which has led to a lack of comprehensive research 
in the European population. Furthermore, almost 
all published studies have generally included only 
a very small number of DLBCL, NOS patients in 
their analyses.
1,18,19,23-25
Moreover, the majority of published research on 
DLBCL, NOS has merely focused on analyzing PD-
1 and PD-L1 expression on LCs or TICs, and when 
this has been the case, the analyses had been based 
on only a single immunohistochemical (IHC) 
staining assessment. To our knowledge, the use of 
markers to accurately identify LCs, such as PAX5, 
has rarely been reported.
14
 PAX5, a member of the 
paired box gene family of transcription factors, is 
a B cell-specific activator protein that plays an im-
portant role during B lymphopoiesis. It shows con-
sistent expression across various stages of B-cell 
maturation and can be identified in the majority 
of B-cell neoplasms, even in cases where mature 
B-cell markers are not expressed.
26
 Furthermore, 
no studies have investigated the simultaneous ex-
pression of PD-1 and PD-L1 on both LCs and TICs 
and their prognostic significance.
Therefore, our aim was to evaluate the simulta-
neous expression of PD-1 and PD-L1 on LCs and 
TICs in a Slovenian cohort of DLBCL, NOS patients 
using double IHC staining in combination with 
PAX5 and to investigate their association with 
prognosis.
Patients and methods
Patients
Patients who were diagnosed with de novo DLBCL, 
NOS at the Institute of Oncology Ljubljana (IOL), 
Slovenia, between February 2004 and May 2018, 
were included in the study. All patients were older 
than 18 years, tested negative for HIV and under-
went lymph node biopsy and histology assessment 
before receiving any specific oncologic treatment. 
In addition, all patients were treated with stand-

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 101
ard treatment (R-CHOP: rituximab-cyclophospha-
mide, vincristine, doxorubicin and prednisone 
or an R-CHOP-like protocol) and radiotherapy of 
residual disease if needed. Each patient has given 
written informed consent.
Study design
The study was conducted retrospectively, and rep-
resentative lymph node excision biopsy samples 
were utilized. All DLBCL, NOS patients were di-
agnosed and subtyped according to the Hans algo-
rithm
27
, as was previously described
28
 (classifying 
DLBCL, NOS by the cell-of-origin into germinal 
center B-cell [GCB] and non-germinal center B-cell 
like [non-GCB]). Double PD-1/PAX5 and double 
PD-L1/PAX5 IHC staining were performed on tis-
sue microarrays (TMAs). Results were evaluated 
by one experienced haemato-pathologist (GG), 
following an already published criteria.
14,16,17,25,29,30
 
Clinical data were obtained from the patients’ elec-
tronic medical record and were used to calculate 
the correlation with the survival outcomes, as well 
as with PD-1 and PD-L1 expression. International 
Prognostic Score (IPI) was calculated for each 
patient.
31
 Survival analysis was based on a mini-
mum of a 5-year patient follow-up. The study was 
conducted in accordance with the Declaration of 
Helsinki and was also approved by the Republic of 
Slovenia National Medical Ethics Committee (No. 
0120-151/2019/4).
Immunohistochemical staining
Double IHC staining was performed for PAX5 and 
PD-1 as well as for PAX5 and PD-L1 to determine 
PD-1 and PD-L1 expression on LCs and TICs, re-
spectively. PAX5 nuclear staining was visualized 
by the presence of red chromogen in B cells, in-
cluding DLBCL cells, while PD-1 and PD-L1 mem-
branous staining was visualized by the presence 
of brown (diaminobenzidine, DAB) chromogen. 
IHC staining was performed using Benchmark XT 
and Benchmark Ultra automated immunostainers 
TABLE 1. Description of the PAX5, PD-1 And PD-L1 antibodies and immunohistochemistry staining protocols
Primary 
Ab
Clone Vendor Reaction type
Antigen 
retrieval 
[100°C]
Ab
dilution
Ab
incubation 
time [min]
IHC detection kit
PAX5 SP34 Ventana Nuclear
CC1
56 min
RTU 32 (37°C)
UltraView
Universal Alkaline Phosphatase Red 
PD-1 NAT105 Dako
Cytoplasmic, 
Membranous
CC1
88 min
1:200 60 (37°C) OptiView DAB
PD-L1 SP263 Ventana
Cytoplasmic, 
Membranous
CC1
64 min 
RTU 16 (37°C) OptiView DAB
Ab = antibody; CC1 = cell conditioning solution 1; DAB = diaminobenzidine; IHC = immunocytochemistry; RTU = ready to use
FIGURE 1. Representable images of the double immunohistochemical staining for (A) PD-1/PAX5 and (B) PD-L1/PAX5. Red 
chromogen indicates PAX5 in DLBCL, NOS nuclei of LCs, with brown chromogen is labeled PD-1 (A) or PD-L1 (B), respectively 
(40x magnification).
DLBCL = diffuse large B-cell lymphoma; LCs = lymphoma cells; NOS = not otherwise specified; PD-1 = programmed cell death protein 1;  
PD-L1 = PD-1 ligand
A B

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 102
(Ventana Medical Systems, Inc., Tucson, AZ, USA). 
A detailed description of the used antibodies and 
IHC staining protocols are shown in Table 1.
Evaluation of PD-1 and PD-L1 expression
TICs and LCs were defined by re-evaluating he-
matoxylin and eosin H&E slides and IHC slides 
stained for Bcl-6, CD5, CD10, CD20, and MUM1. 
LCs were recognized according to PAX5 expres-
sion and morphology. The expression of PD-1 and 
PD-L1 was assessed semi-quantitatively for both 
LCs and TICs, using already published cut-off 
values. PD-1 and PD-L1 expression on TICs was 
assessed in three high-powered fields (HPF), and 
the score was categorized in four groups: score 0 
(no positive cells), score 1 (less than 10 cells), score 
2 (10-30 cells) and score 3 (more than 30 cells). 
According to the references
16,17,25
 we considered 
scores 0 and 1 as negative and 2 and 3 as positive. 
PD-1 and PD-L1 expression on LCs was catego-
rized into negative and positive group by using 
cut-off of 10% and 30%, respectively.
14,29,30
Statistical analysis
Descriptive statistics were used to describe the 
basic characteristic of the data. The median and 
range were calculated for the age of the patients, 
OS, PFS, and observation time. The Chi-square 
test or Fisher´s exact test were used to analyze if 
there is a difference between PD-1 and PD-L1 ex-
pression and clinicopathological characteristics of 
the patients. PFS was calculated as the time from 
diagnosis until disease progression or death from 
any cause, and OS was calculated as the time from 
diagnosis to death from any cause. The median 
survival of the patients was expressed in months. 
Kaplan Maier with log-rank test was used to com-
pare PFS and OS between two groups. Hazard 
ratio (HR) and 95% confidence interval (CI) were 
calculated for both univariate and multivariate 
analysis (Cox regression model). Parameters that 
proved to be significant in the univariate analysis 
were included in the multivariate analysis. p < 0.05 
was considered significant. IBM SPSS Statistics 
(version 28.0.1.0, IBM, Armonk, NY, USA) was used 
for the analysis.
Results 
Patients’ characteristics
The study included 283 Slovenian patients diag-
nosed with DLBCL, NOS. However, due to incom-
plete data on patients’ treatment, lost follow-up or 
inadequate biological material, 67 patients were 
excluded, resulting in a final cohort of 216 patients 
TABLE 2. Clinicopathological characteristics of the Slovenian 
patient cohort (N = 216) included in the analysis
Age at diagnosis (years)
Median 64
Range 27-89
≤60 84
>60 132
Sex, N (%)
Male 104 (48)
Female 112 (52)
Ann Arbor stage, N (%)
I 37 (17)
II 47 (22)
III 45 (21)
IV 87 (40)
Involvement of an extranodal organ, N (%)
Yes 72 (33)
No 103 (48)
No data 41 (19)
Involvement of spleen, N (%)
Yes 34 (16)
No 129 (60)
No data 53 (24)
B symptoms, N (%)
Yes 76 (35)
No 116 (54)
No data 24 (11)
IPI score, N (%)
0, 1 63 (29.2)
2 51 (23.6)
3 50 (23.1)
4, 5 54 (24.1)
Classification according to Hans Algorithm, N (%)
Non-GCB 92 (43)
GCB 124 (57)
Survival status of the patients, N (%)
Alive 102 (47)
Dead 114 (53)
GCB = germinal center B-cell diffuse large B-cell lymphoma (DLBCL) 
subtype; IPI = International Prognostic Index; N = number; non-GCB = 
non-germinal center B-cell like DLBCL subtype

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 103
for subsequent analyses. The median observation 
time for the analyzed patients was 162 months 
(range 60-234 months). Clinicopathological char-
acteristics of our patient’s cohort are presented in 
Table 2.
The expression of PD-1 and PD-L1
PD-1 was expressed on TICs in 38.4% of cases and 
on LCs in 8.8% of cases, while PD-L1 was expressed 
on TICs in 62.5% of cases and on LCs in 6.5% of 
cases (Figure 1). We also investigated whether 
there was a difference between the non-GCB and 
GCB subtypes regarding PD-1 and PD-L1 expres-
sion on both TICs and LCs. Our results showed no 
difference in the expression of PD-1 on TICs and 
LCs (p = 0.291 and p = 0.224, respectively), nor for 
PD-L1 on TICs (p = 0.393). Interestingly, we con-
firmed significantly increased PD-L1 expression 
on LCs within the non-GCB subtype compared to 
the GCB subtype (p = 0.047). The detailed results of 
PD-1 and PD-L1 expression analysis can be found 
in Table 3.
Regarding clinicopathological characteristics 
of the patients (Table 4), no significant differences 
were observed in PD-1 and PD-L1 expression on 
TICs or LCs when analyzed in relation with vari-
ables such as the age, sex, Ann Arbor stage, in-
volvement of an extranodal organ, involvement of 
spleen, presence of B symptoms or IPI score. 
Clinicopathological characteristics 
and correlation with progression free-
survival and overall survival
The results of the survival analysis, which was 
performed on the basis of the clinicopathological 
characteristics of the patients, are summarized 
in Figure 2 and Table 5. The median PFS was 77.4 
months (range 0.23-224.89) and the median OS 
TABLE 3. Clinicopathological characteristics of the Slovenian patient cohort (N = 216) included in the analysis
PD-1 on TICs
Expression
PD-1 on LCs 
Expression
PD-L1 on TICs 
Expression
PD-L1 on LCs 
Expression
(N, %) Positive Negative Positive Negative Positive Negative Positive Negative
All cases (N = 216) 83 133 19 197 135 81 14 202
Non-GCB subtype (N = 92) 31 (37.3) 61 (45.9) 11 (57.9) 81 (41.1) 58 (43.0) 34 (42.0) 10 (71.4) 82 (40.6)
GCB subtype (N = 142) 52 (62.7) 72 (54.1) 8 (42.1) 116 (58.9) 77 (57.0) 47 (58.0) 4 (28.6) 120 (59.4)
Non-GCB versus GCB 
subtype (p value)
0.258 0.224 0.887 0.047
GCB = germinal center B-cell DLBCL subtype; LCs = lymphoma cells; N = number; non-GCB = non-germinal center B-cell like DLBCL subtype; PD-1 = programmed cell 
death protein 1; PD-L1 = PD-1 ligand; TICS = tumor-immune cells
FIGURE 2. Kaplan-Meier curves for (A) progression-free survival and (B) 
overall survival, representing only significant differences among all analyzed 
clinicopathological characteristics of diffuse large B-cell lymphoma, not 
otherwise specified patients.
was 83.3 months (range 0.23-224.89). In the group 
of patients under 60 years of age, the OS was sig-
A
B

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 104
nificantly longer than in the group of patients over 
60 years of age (p < 0.001). At the same time, we 
did not find age to be statistically significant for 
PFS. Moreover, patients in Ann Arbor stages I or II 
showed a significant association with longer PFS 
(p = 0.044) and OS (p < 0.001) compared to patients 
in stages III and IV. In addition, a longer PFS (p = 
0.025) and OS (p = 0.004) were observed in patients 
without B-symptoms compared to patients with 
B-symptoms. A low IPI score (score between 0 and 
2) was associated with a longer PFS (p = 0.003) and 
a longer OS (p < 0.001). No correlation with PFS and 
OS was observed for the other clinicopathological 
characteristics such as Hans algorithm, gender, an 
extranodal organ and spleen involvement.
Correlation of PD-1 and PD-L1 
expression with progression free-
survival and overall survival
We conducted individual survival analyses for PD-
1 and PD-L1 on TICs. The results were obtained 
TABLE 4. PD-1 and PD-L1 expression in association with clinicopathological characteristics of patients with diffuse large B-cell lymphoma, not 
otherwise specified
PD-1 expression on TICs PD-1 expression on LCs PD-L1 expression on TICs PD-L1 expression on LCs
[N, (%)] Positive Negative p value Positive Negative p value Positive Negative p value Positive Negative p value
Total 83 (38.4) 133 (61.6) 19 (8.8) 197 (91.2) 135 (62.5) 81 (37.5) 14 (6.5) 202 (93.5)
Age 0.775 0.466 0.885 0.406
≤60 31 (14.4) 53 (24.5) 9 (4.2) 75 (34.7) 53 (24.5) 31 (14.4) 7 (3.2) 77 (35.6)
>60 52 (24.1) 8 0 (37.0) 10 (4.6) 122 (56.5) 82 (38.0) 50 (23.1) 7 (3.2) 125 (57.9)
Sex 0.889 0.811 0.265 1.000
Male 39 (18.1) 65 (30.1) 10 (4.6) 94 (43.5) 61 (28.8) 43 (19.9) 7 (3.2) 97 (44.9)
Female 44 (20.4) 68 (31.5) 9 (4.2) 103 (47.7) 74 (34.2) 38 (17.6) 7 (3.2) 1,5 (48.6)
Ann Arbor stage 1.000 1.000 0.116 0.134
I-II 32 (14.8) 52 (24.1) 7 (3.2) 77 (35.6) 47 (21.8) 37 (17.1) 3 (1.4) 81 (37.5)
III-IV 51 (23.6) 81 (37.5) 12 (5.6) 120 (55.6) 88 (40.7) 44 (20.4) 11 (5.1) 121 (56.0)
Involvement of an
extranodal organ
0.643 0.412 0.332 0.738
Yes 33 (18.9) 39 (22.3) 8 (4.6) 64 (36.6) 44 (25.1) 28 (16.0) 3 (1.7) 69 (39.4)
No 43 (24.6) 60 (34.3) 7 (4.0) 96 (54.9) 71 (40.6) 32 (18.3) 6 (3.4) 97 (55.4)
Involvement of spleen 0.847 1.000 0.540 1.000
Yes 15 (9.2) 19 (11.7) 3 (1.8) 31 (19.0) 25 (15.3) 9 (5.5) 1 (0.6) 33 (20.2)
No 61 (37.4) 68 (41.7) 11 (6.7) 118 (72.4) 87 (53.4) 42 (25.8) 7 (4.3) 122 (74.8)
B symptoms 0.366 0.598 0.536 0.085
Yes 27 (14.1) 49 (25.5) 5 (2.6) 71 (37.0) 52 (27.1) 24 (12.5) 9 (4.7) 67 (34.9)
No 50 (26.0) 66 (34.4) 11 (5.7) 105 (54.7) 73 (38.0) 43 (22.4) 5 (2.6) 111 (57.8)
IPI score 0.780 0.228 0.575 1.000
0-2 45 (20.8) 69 (31.9) 13 (6.0) 101 (46.8) 69 (31.9) 45 (20.8) 7 (3.2) 107 (49.5)
3-5 38 (17.6) 64 (29.6) 6 (2.8) 96 (44.4) 66 (30.6) 36 (16.7) 7 (3.2) 95 (44.0)
Hans Algorithm
classification
0.258 0.224 0.887 0.047
Non-GCB 31 (14.4) 61 (28.2) 11 (5.1) 88 (37.5) 58 (26.9) 34 (15.7) 10 (4.6) 82 (38.0)
GCB 52 (24.1) 72 (33.3) 8 (3.7) 116 (53.7) 77 (35.6) 47 (21.8)) 4 (1.9) 120 (55.6)
Patients’ outcome 0.124 0.639 0.779 0.788
Alive 45 (20.8) 57 (26.4) 10 (4.6) 92 (42.6) 65 (30.1) 37 (17.1) 6 (2.8) 96 (44.4)
Dead 38 (17.6) 76 (35.2) 9 (4.2) 105 (48.6) 70 (32.4) 44 (20.4) 8 (3.7) 106 (49.1)
GCB = germinal center B-cell DLBCL subtype; IPI = International Prognostic Index; LCs = lymphoma cells; N = number; non-GCB = non-germinal center B-cell like DLBCL 
subtype; PD-1 = programmed cell death protein 1: PD-L1 = PD-1 ligand; TICs = tumor-immune cells; % = percentage

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 105
while classifying the samples into four groups, as 
well as grouping them in two categories: negative 
(0 and 1) and positive (2 and 3). However, no statis-
tically significant differences were found in either 
data set (Figure 3).
Correlation analysis between PD-L1 expression 
on LCs and patient survival showed a significant 
correlation for patients who had no PD-L1 expres-
sion on LCs (i.e. less than 30% of LCs were PD-L1 
positive), with significantly longer progression-
free survival (PFS) (p = 0.015) compared to patients 
who had detectable PD-L1 expression (77 .7 months 
vs. 15.6 months). However, no correlation with OS 
was observed. Furthermore, correlation PD-1 ex-
pression on LCs cell showed no correlation with 
PFS and OS. Detailed results of the correlation 
analysis for PD-1 and PD-L1 expression with PFS 
and OS are summarized in Table 4 and Figure 4.
Multivariate analysis of the significant 
parameters in the univariate analysis
Only the significant clinicopathological features 
from the univariate analysis were included in the 
multivariate analysis: the Ann Arbor stage, pres-
ence of B symptoms, IPI score, and PD-L1 expres-
sion on LCs were used for the PFS analysis, and 
the age, Ann Arbor stage, presence of B symptoms, 
and IPI score were used for the OS analysis. Our 
results showed that the IPI score (p = 0.048, HR = 
1.945) and the presence of PD-L1 on the LCs (p = 
0.034, HR = 2.393) retained their significant prog-
nostic impact for PFS. As expected, patient age (p 
< 0.001, HR = 2.907) was found to be a significant 
prognostic factor for OS, while the other variables, 
including IPI score, remained non-significant. The 
results of the multivariate analysis are shown in 
Table 5.
Discussion
In the present study, we investigated the expres-
sion of PD-1 and PD-L1 on LCs and TICs in the 
tumor microenvironment of DLCBL, NOS patients 
in relation to non-GCB and GCB subtypes and pa-
tients’ survival.
The expression of PD-1 and PD-L1 has drawn 
great attention to the impact of lymphoma treat-
ment, particularly in aggressive lymphomas such 
as DLBCL, NOS, where they evade immune re-
sponse and drive aggressiveness.
21
 Identification 
of patients for PD-1/PD-L1 immunotherapy, pos-
sibly through IHC evaluation, holds promise for 
better patient outcomes and further research in 
checkpoint inhibitor treatment. Currently, there 
are few ongoing clinical trials investigating the 
use of anti-PD-1 and anti-PD-L1 treatments in pa-
tients with relapsed or refractory DLBCL, NOS. 
Preliminary results in small patient cohorts show 
promising results for prolonged disease-free inter-
vals.
29
 However, the selection criteria for patient 
enrollment are not dependent on PD-1 or PD-L1 ex-
pression on either TICs or LCs. One of the explana-
tions for this could be the lack of consensus in the 
FIGURE 3. Kaplan-Meier curves for (A) progression-free survival (PFS) and (B) overall 
survival (OS) for PD-1 and PD-L1 on tumor-immune cells. The cases were divided 
into four groups based on the cell count per high-power field. Furthermore, these 
cases were stratified into two classifications: negative (cell counts 0 and 1) and 
positive (cell counts 2 and 3). 
PD-1 = programmed cell death protein 1; PD-L1 = PD-1 ligand; TICs = tumor immune cells.
A
B

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 106
evaluation criteria, especially for PD-1 expression, 
as it is less frequently assessed and less defined. 
Different evaluation criteria are used to assess cell 
positivity, and we have used the most commonly 
used ones.
8,14,16,17,25,29,32
 We found only a few stud-
ies investigating whether there is a correlation 
between the expression of PD-1 or PD-L1 on TICs 
or LCs and patient outcomes. For the assessment 
of PD-1 and PD-L1 on LCs and TICs in our study, 
we performed double staining for PD-1/PAX5 and 
PD-L1/PAX5 to simplify the assessment of PD-1 
and PD-L1 expression and obtain more reliable re-
sults. Indeed, P AX5 was required for accurate iden-
tification of B cells, including LCs, because PAX5 
is expressed in mature B cells and LCs.
14,24,25,33
 The 
use of PAX5 increased the accuracy of LC identifi-
cation, which in combination with the simultane-
ous staining of PD-1 and PD-L1 is one of the major 
advantages of this study. This double staining was 
so far reported in Kiyasu’s study of 1091 patients 
with DLBCL
14
, NOS and Chen’s study of various 
lymphoma subtypes, including 66 patients with 
DLBCL, NOS
33
, but both studies used only PD-L1/
PAX5 staining. To the best of our knowledge, our 
study is the first to perform PD-L1 and PD-1 stain-
ing simultaneously with PAX5.
In this way, we confirmed a PD-1 expression of 
38.4% on TICs and 8.8% on LCs. Our results were 
similar to the already published data, where PD-
1 expression on TICs ranged from 22.2-60.0%. For 
PD-1 on LCs, we observed a slightly lower expres-
sion rate compared to the expression levels of PD-
1 reported in the literature (22.2-65.0%).
16,24,25,29,34
 
Using the same assessment method as in the other 
published studies
14,16,17,25,29,30
, we reported 62.4% 
PD-L1 expression on TICs and 6.5% on LCs. Our 
data again differed from already reported PD-L1 
expression levels, where expression ranged from 
15.3-37.0% on TICs and 8.9-61.1% on LCs.
14,17,24,29
 
In summary, we observed a lower expression of 
PD-L1 and PD-1 on LCs and a higher expression 
of PD-L1 on TICs. We speculate that the major rea-
son for the low expression on LCs is the additional 
staining with PAX5. With the double staining, we 
were able to recognize PD-1/PD-L1 positive LCs 
more precisely, so the numbers are probably more 
TABLE 5. Univariate and multivariate analysis of the patients’ survival based on their clinicopathological characteristics and PD-1 and PD-L1 
expressions on lymphoma cells and tumor-infiltrating immune cells in tissue samples of diffuse large B-cell lymphoma, not otherwise specified
Univariate analysis Multivariate analysis
PFS OS PFS OS
p value
Median 
when 
patients 
have relapse 
[months]
p value
Median 
when 
patients died 
[months]
p value HR (95% CI) p value HR (95% CI)
Age ≤ 60 vs. > 60 0.330 91.8 vs. 59.7 < 0.001 110.1 vs. 73.5 < 0.001
2.907
(1.710-4.940)
Sex
Male vs. 
Female
0.945 69.1 vs. 80.9 0.324 78.3 vs. 90.8
Ann Arbor stage I-II vs. III-IV 0.044 91.3 vs. 59.7 < 0.001 113.8 vs. 72.0 0.845
1.072
(0.532-2.130)
0.073
1.654
(0.955-2.865)
Involvement of an 
extranodal organ
(-) vs. (+) 0.886 77.4 vs. 74.1 0.451 81.8 vs. 82.3
Involvement of the 
spleen
(-) vs. (+) 0.915 69.9 vs. 81.8 0.844 80.1 vs. 81.8
B symptoms (-) vs. (+) 0.025 85.3 vs. 30.8 0.004 91.8 vs. 65.3 0.338
1.319
(0.748-2.326)
0.170
1.354
(0.879-2.087)
IPI score 0-2 vs. 3-5 0.006 88.8 vs. 29.3 < 0.001 101.2 vs. 62.7 0.048
1.945
(1.005-3.767)
0.494
1.205
(0.706-2.058)
Hans classification
Non-GCB vs. 
GCB
0.914 66.5 vs. 80.7 0.095 77.6 vs. 85.9
PD-1 on TICs (-) vs. (+) 0.797 81.6 vs. 76.0 0.478 85.9 vs. 80.7
PD-1 on LCs (-) vs. (+) 0.657 77. 8 vs. 76.0 0.882 84.9 vs. 76.3
PD-L1 on TICs (-) vs. (+) 0.955 85.9 vs. 76.2 0.623 111.0 vs. 79.2
PD-L1 on LCs (-) vs. (+) 0.015 77.7 vs. 15.6 0.373 85.1 vs. 22.3 0.034
2.393
(1.070-5.352)
GCB = germinal center B-cell DLBCL subtype; IPI = International Prognostic Index; LCs = lymphoma cells; N = number; non-GCB = non-germinal center B-cell like DLBCL 
subtype; OS = overall survival; PD-1 = programmed cell death protein 1; PD-L1 = PD-1 ligand; PFS = progression-free survival; TICs = tumor-immune cells

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 107
reliable than in other studies where other cells 
such as macrophages could be misinterpreted as 
LCs and resulted in higher expression of PD-L1 
and PD-1 on LCs. Two already published stud-
ies using PAX5 staining reported 8.9%
14
 and 11% 
PD-L1 positivity on LCs
33
, which is comparable to 
our results.
We also speculate that a possible reason for 
this discrepancy in PD-L1 and PD-1 expression 
rates could be a consequence of several other 
factors. One reason might be the different inclu-
sion criteria of the patients. Almost all published 
studies included patients with Epstein-Barr virus 
(EBV) positive DLBCL, NOS, which typically ex-
hibit PD-L1 positive status due to the association 
between PD-L1 expression and EBV infection. 
EBV positive patients were not a part of our pa-
tient cohort, so this could be one of the reasons 
for the discrepant results. In addition, different 
algorithms were used to define the DLBCL, NOS 
subtypes. The majority of the studies have used 
Hans’ algorithm
14,16,17,25,29,30
, but some have applied 
Choi’s algorithm.
35
 The choice of one of these al-
gorithms may contribute to different percentages 
of GCB and non-GCB subtypes, as well as differ-
ent percentages of PD-L1 positive LCs and TICs 
within each subtype.
35
 Furthermore, different an-
tibody clones were used in different studies, and 
in some cases the clones used were not clearly 
stated.
8,25
 Additionally, some studies lacked a clear 
description of the criteria used for the assessment 
of PD-L1 positivity.
24
 Since the major focus was on 
the Asian population, where non-GCB subtypes 
of DLBCL, NOS, were more common than GCB 
subtypes, a higher number of PD-L1 positive cases 
was expected.
22
The second aim of our study was to investigate 
the association of PD-1 and PD-L1 expression on 
TICs or LCs with clinicopathological characteris-
tics. We investigated the potential differences be-
tween various clinicopathological characteristics 
of patients such as age at diagnosis, gender, Ann 
Arbor stage, extranodal organ involvement, in-
volvement of the spleen, presence of B symptoms, 
and IPI score, but our results showed no difference 
in PD-1 and/or PD-L1 expression on LCs and/or 
TICs between these groups. Additionally, we ex-
amined the histological subtypes determined by 
the Hans algorithm and found that PD-L1 posi-
tive expression was more frequent in the non-GCB 
subtype, as already reported.
14,15,17,19
 However, in a 
univariate analysis, the Hans algorithm showed 
no significant impact on PFS or OS. Regarding pa-
tient survival, we found that patients with PD-L1 
expression on the LCs are correlated with shorter 
PFS, which also had a significant impact in the 
multivariate analysis including the IPI score. This 
indicates that although the number of patients 
with PD-L1 positive expression on the LCs is 
small, their survival is worse in comparison with 
the patients with no PD-L1 expression, regardless 
of the IPI score. Since the IPI score is still one of the 
strongest and most reliable prognostic markers in 
DLBCL, NOS, even in the era of novel therapies, 
we believe this is an interesting conclusion that re-
quires further research.
In the multivariate analysis for PFS, where the 
Ann Arbor stage, presence of B symptoms, IPI 
score, and PD-L1 expression on LCs for PFS were 
included as significant parameters from the uni-
variate analyses, PD-L1 expression on LCs was 
found to be an independent prognostic marker 
for PFS. As expected, the IPI score also proved 
to be an independent prognostic marker for PFS. 
In the multivariate analysis of OS, where the age, 
Ann Arbor stage, B symptoms and IPI score were 
included, we showed that age remained as an in-
dividual marker for OS. This result was consist-
ent with already published data based on 5-year 
follow-up analyses for PFS and OS of DLBCL, NOS 
patients.
8
Noteworthy, there are some limitations of our 
study that need to be considered. For example, due 
FIGURE 4. Kaplan-Meier curves for (A) progression-free survival and (B) overall 
survival representing the influence of PD-1 and PD-L1 expression on lymphoma 
cells (LCs). PD-1 expression on LCs was categorized as negative below 10%. PD-L1 
expression on LCs was categorized as negative below 30%. 
PD-1 = programmed cell death protein 1; PD-L1 = PD-1 ligand
A
B

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 108
to the retrospective nature of the study, the longer 
archiving time of FFPE tissue blocks may poten-
tially influence the staining results. In addition, it 
is still unclear whether the expression of PD-1 and 
PD-L1 on tumors and TICs is a key factor for the 
clinical prognosis of DLBCL, NOS patients treated 
with PD-1/PD-L1 blockade therapy. On the other 
hand, our study has certain advantages, such as 
centralized evaluation of specimens, homogene-
ous treatment and long follow-up time. To the best 
of our knowledge, we are the first to simultaneous-
ly assess the expression of PD-1 and PD-L1 on both 
TICs and LCs from the same cohort of patients us-
ing a double immunostaining approach and their 
impact on PFS and OS, as well as their association 
with other clinicopathological characteristics. Our 
results were consistent with individual studies on 
the European population as well as studies on the 
Asian population. 
In conclusion, we demonstrated that PD-L1 ex-
pression on LCs was associated with shorter PFS 
and was more frequently observed in the non-GCB 
subtype. Double IHC staining with PAX5 proved 
to be a feasible method to assess PD-1 and PD-L1 
expression in tissue samples. Further research and 
clinical studies are required to assess the impor-
tance of assessing PD-1 and PD-L1 in DLBCL, NOS 
patients as well as methods to determine their 
expression, particularly with regard to planning 
immunotherapy treatments. It is also important 
to understand the mechanisms of tumor immune 
evasion induced by PD-1/PD-L1 and to explore ap-
proaches to modulate the host immune response 
accordingly, which requires further research.
Acknowledgments 
We would like to thank to the Department of 
Pathology and the Department of Cytopathology 
at the Institute of Oncology Ljubljana for their sup-
port, contributions and collaboration.
The publication of this article was financed 
by Slovenian Research and Innovation Agency 
(ARIS), grant No. P3-0289. The founder had no role 
in study design, data collection and analysis, deci-
sion to publish, or preparation of the manuscript. 
References
1. Qiu L, Zheng H, Zhao X. The prognostic and clinicopathological significance 
of PD-L1 expression in patients with diffuse large B-cell lymphoma: a meta-
analysis. BMC Cancer 2019; 19: 273. doi: 10.1186/s12885-019-5466-y 
2. Minghan Q, Shan W, Xinrui C, Huaqing W. Update on diffuse large B-cell 
lymphoma: highlights from the 2022 ASCO Annual Meeting. Cancer Biol 
Med 2022; 19: 1117-20. doi: 10.20892/j.issn.2095-3941.2022.0403
3. Wang C, Shi F, Liu Y, Zhang Y, Dong L, Li X, et al. Anti-PD-1 antibodies as a 
salvage therapy for patients with diffuse large B cell lymphoma who pro-
gressed/relapsed after CART19/20 therapy. J Hematol Oncol 2021; 14: 106. 
doi: 10.1186/s13045-021-01120-3
4. Lu T, Zhang J, Xu-Monette ZY , Young KH. The progress of novel strategies on 
immune-based therapy in relapsed or refractory diffuse large B-cell lym-
phoma. Exp Hematol Oncol 2023; 12: 72. doi: 10.1186/s40164-023-00432-z
5. Guan J, Zhang J, Zhang X, Yuan Z, Cheng J, Chen B. Efficacy and safety of PD-
1/PD-L1 immune checkpoint inhibitors in treating non-Hodgkin lymphoma: 
A systematic review and meta-analysis of clinical trials. Medicine 2022; 101: 
e32333. doi: 10.1097/MD.0000000000032333
6. Xie M, Huang X, Ye X, Qian W . Prognostic and clinicopathological significance 
of PD-1/PD-L1 expression in the tumor microenvironment and neoplas-
tic cells for lymphoma. Int Immunopharmacol 2019; 77: 105999. doi: 
10.1016/j.intimp.2019.105999
7. Boydell E, Sandoval JL, Michielin O, Obeid M, Addeo A, Friedlaender A. 
Neoadjuvant immunotherapy: A promising new standard of care. Int J Mol 
Sci 2023; 24: 11849. doi: 10.3390/ijms241411849
8. Hu LY, Xu XL, Rao HL, Chen J, Lai RC, Huang HQ, et al. Expression and clini-
cal value of programmed cell death-ligand 1 (PD-L1) in diffuse large B cell 
lymphoma: a retrospective study. Chin J Cancer 2017; 36: 94. doi: 10.1186/
s40880-017-0262-z
9. Goodman A, Patel SP , Kurzrock R. PD-1-PD-L1 immune-checkpoint blockade 
in B-cell lymphomas. Nat Rev Clin Oncol 2017; 14: 203-20. doi: 10.1038/
nrclinonc.2016.168 
10. Jelinek T, Mihalyova J, Kascak M, Duras J, Hajek R. PD-1/PD-L1 inhibitors in 
haematological malignancies: update 2017. Immunology 2017;152: 357-71. 
doi: 10.1111/imm.12788
11. Ok CY, Young KH. Checkpoint inhibitors in hematological malignancies. J 
Hematol Oncol 2017; 10: 103. doi: 10.1186/s13045-017-0474-3
12. Wang Y, Wu L, Tian C, Zhang Y. PD-1-PD-L1 immune-checkpoint blockade 
in malignant lymphomas. Ann Hematol 2018; 97: 229-37. doi: 10.1007/
s00277-017-3176-6
13. Annibali O, Crescenzi A, Tomarchio V, Pagano A, Bianchi A, Grifoni A, et al. 
PD-1 /PD-L1 checkpoint in hematological malignancies. Leuk Res 2018; 67: 
45-55. doi: 10.1016/j.leukres.2018.01.014
14. Kiyasu J, Miyoshi H, Hirata A, Arakawa F, Ichikawa A, Niino D, et al. 
Expression of programmed cell death ligand 1 is associated with poor over-
all survival in patients with diffuse large B-cell lymphoma. Blood 2015; 126: 
2193-201. doi: 10.1182/blood-2015-02-629600 
15. Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, et al. 
WHO classification of tumours of haematopoietic and lymphoid tissues, 
WHO classification of tumours, Revised 4th edition, Volume 2. France: 
International Agency for Research on Cancer; 2017.
16. Fang X, Xiu B, Yang Z, Qiu W, Zhang L, Zhang S, et al. The expression 
and clinical relevance of PD-1, PD-L1, and TP63 in patients with dif-
fuse large B-cell lymphoma. Medicine 2017; 96: e6398. doi: 10.1097/
MD.0000000000006398
17. Kwon D, Kim S, Kim PJ, Go H, Nam SJ, Paik JH, et al. Clinicopathological analy-
sis of programmed cell death 1 and programmed cell death ligand 1 expres-
sion in the tumour microenvironments of diffuse large B cell lymphomas. 
Histopathology 2016; 68: 1079-89. doi: 10.1111/his.12882
18. Gravelle P, Burroni B, Péricart S, Rossi C, Bezombes C, Tosolini M, et 
al. Mechanisms of PD-1/PD-L1 expression and prognostic relevance in 
non-Hodgkin lymphoma: a summary of immunohistochemical studies. 
Oncotarget 2017; 8: 44960-75. doi: 10.18632/oncotarget.16680
19. Zeng Q, Liu Z, Liu T. Prognostic value and clinicopathological characteristics 
of PD-L1 overexpression in non-Hodgkin lymphoma: a meta-analysis. BMC 
Cancer 2020; 20: 59. doi: 10.1186/s12885-020-6550-z
20. Xie W, Medeiros LJ, Li S, Yin CC, Khoury JD, Xu J. PD-1/PD-L1 pathway and 
Its blockade in patients with classic Hodgkin lymphoma and Non-Hodgkin 
large-cell lymphomas. Curr Hematol Malig Rep 2020; 15: 372-81. doi: 
10.1007/s11899-020-00589-y

Radiol Oncol 2024; 58(1): 99-109.
Cas Slak T et al. / Prognostic significance of PD-1 and PD-L1 in B-cell lymphoma 109
21. Xing W, Dresser K, Zhang R, Evens AM, Yu H, Woda BA, Chen BJ. PD-L1 
expression in EBV-negative diffuse large B-cell lymphoma: clinicopathologic 
features and prognostic implications. Oncotarget 2016; 7: 59976-86. doi: 
10.18632/oncotarget.11045
22. Shiozawa E, Yamochi-Onizuka T, Takimoto M, Ota H: The GCB subtype of 
diffuse large B-cell lymphoma is less frequent in Asian countries. Leuk Res 
2007; 31: 1579-83. doi: 10.1016/j.leukres.2007.03.017
23. Muenst S, Hoeller S, Willi N, Dirnhofera S, Tzankov A. Diagnostic and prog-
nostic utility of PD-1 in B cell lymphomas. Dis Markers 2010; 29: 47-53. doi: 
10.3233/DMA-2010-0725
24. Laurent C, Charmpi K, Gravelle P, Tosolini M, Franchet C, Ysebaert L, 
et al. Several immune escape patterns in non-Hodgkin’s lymphomas. 
Oncoimmunology 2015; 4: e1026530. doi: 10.1080/2162402X.2015.1026530
25. Karakatsanis S, S Papadatos S, Syrigos N, Marinos L, Pouliou E, Papanikolaou 
A. Clinical significance of PD-1 and PD-L1 molecules in patients with diffuse 
large B-cell lymphoma, not otherwise specified: Correlation with clinical and 
pathological findings. J BUON 2021; 26: 569-79. 
26. Desouki MM, Post GR, Cherry D, Lazarchick J. PAX-5: a valuable immunohis-
to-chemical marker in the differential diagnosis of lymphoid neoplasms. Clin 
Med Res 2010; 8: 84-8. doi: 10.3121/cmr.2010.891
27. Hans CP, Weisenburger DD, Greiner TC, Gascoyne RD, Delabie J, Ott G, et 
al. Confirmation of the molecular classification of diffuse large B-cell lym-
phoma by immunohistochemistry using a tissue microarray. Blood 2004; 
103: 275-82. doi: 10.1182/blood-2003-05-1545 
28. Boltezar L, Kloboves-Prevodnik V, Pohar-Perme M, Gasljevic G, Jezersek-
Novakovic B. Comparison of the algorithms classifying the ABC and GCB 
subtypes in diffuse large B-cell lymphoma. Oncol Lett 2018; 15: 6903-12. 
doi: 10.3892/ol.2018.8243
29. Chen BJ, Dashnamoorthy R, Galera P, Makarenko V, Chang H, Ghosh S, et 
al. The immune checkpoint molecules PD-1, PD-L1, TIM-3 and LAG-3 in dif-
fuse large B-cell lymphoma. Oncotarget 2019; 10: 2030-40. doi: 10.18632/
oncotarget.26771
30. Xu-Monette ZY , Xiao M, Au Q, Padmanabhan R, Xu B, Hoe N, et al. Immune 
profiling and quantitative analysis decipher the clinical role of immune-
checkpoint expression in the tumor immune microenvironment of DLBCL. 
Cancer Immunol Res 2019; 7: 644-57. doi: 10.1158/2326-6066.CIR-18-0439
31. International Non-Hodgkin’s Lymphoma Prognostic Factors Project. A pre-
dictive model for aggressive non-Hodgkin’s lymphoma. N Engl J Med 1993; 
329: 987-94. doi: 10.1056/NEJM199309303291402
32. McCord R, Bolen CR, Koeppen H, Kadel EE 3rd, Oestergaard MZ, Nielsen T , et 
al. PD-L1 and tumor-associated macrophages in de novo DLBCL. Blood Adv 
2019; 3: 531-40. doi: 10.1182/bloodadvances.2018020602
33. Chen BJ, Chapuy B, Ouyang J, Sun HH, Roemer MG, Xu ML, et al. PD-L1 
expression is characteristic of a subset of aggressive B-cell lymphomas 
and virus-associated malignancies. Clin Cancer Res 2013; 19: 3462-73. doi: 
10.1158/1078-0432.CCR-13-0855
34. Zhang T, Xie J, Arai S, Wang L, Shi X, Shi N, et al. The efficacy and safety 
of anti-PD-1/PD-L1 antibodies for treatment of advanced or refractory 
cancers: a meta-analysis. Oncotarget 2016; 7: 73068-79. doi: 10.18632/
oncotarget.12230
35. Kwon HJ, Yang JM, Lee JO, Lee JS, Paik JH. Clinicopathologic implication 
of PD-L1 and phosphorylated STAT3 expression in diffuse large B cell lym-
phoma. J Transl Med 2018; 16: 320. doi: 10.1186/s12967-018-1689-y