Cytoplasmic and nuclear programmed death ligand 1 expression in peritumoral stromal cells in breast cancer: Prognostic and predictive value

Abstract

BACKGROUND

Breast cancer (BC) continues to occupy a leading position in terms of morbidity and mortality from malignant neoplasms among the female population. One of the promising markers associated with BC progression is programmed death ligand 1 (PD-L1). Previously, we investigated PD-L1 expression in BC via a new antibody against programmed cell death protein 1 ligand 1 (PDCD1 LG1) and reported that high PDCD1 LG1 expression in tumor cells is an independent factor for a high risk of regional metastasis in patients with BC. However, the prognostic significance of PDCD1 LG1 expression in BC stromal cells has not been adequately studied.

AIM

To study the features of PDCD1 LG1 expression in BC stromal cells and its relationship with BC clinicopathological characteristics.

METHODS

In a prospective single-center observational study, tumor samples from 148 patients with newly diagnosed BC were examined. The tumor sections were immunohistochemically stained with antibodies against PDCD1 LG1. In the tumor samples, the PDCD1 LG1-positive lymphocyte (PDCD1 LG1+ LF) score, presence of nuclear PDCD1 LG1 expression in the LFs, PDCD1 LG1 expression in polymorphic cell infiltrates (PDCD1 LG1+ polymorphic cell infiltrates [PCIs]), and cells of the fibroblastic stroma and endothelial cells of the tumor microvessels were assessed. Statistical analyses were performed using Statistica 10.0 software.

RESULTS

A PDCD1 LG1+ LF score ≥ 3 was detected more often at stages N0 and N3 than at N1 and N2 (P = 0.03). Moderate and pronounced PDCD1 LG1+ PCIs and the presence of PDCD1 LG1+ fibroblastic stroma were associated with negative estrogen receptor status (P = 0.0008 and P = 0.03, respectively), human epidermal growth factor receptor 2-positive (HER2+) BC (P < 0.00001 and P = 0.0005), and luminal B HER2+, non-luminal HER2+ and triple-negative BC (P < 0.00001 and P = 0.004). The risk of metastasis to regional lymph nodes (RLNs) depend on lymphovascular invasion (LVI) and the PDCD1 LG1+ LF score. In the absence of LVI and a PDCD1 LG1+ LF score < 3 or ≥ 3, metastases in RLNs were absent in 66.6% and 93.9% of patients with BC, respectively. In the presence of LVI and a PDCD1 LG1+ LF score < 3 or ≥ 3, metastases in RLNs were detected in 82.6% and 92.7% of patients with BC, respectively.

CONCLUSION

The results indicated that the combined assessment of the PDCD1 LG1+ LF score and LVI can improve the accuracy of predicting the risk of metastasis to RLNs in patients with BC.

Key Words: Breast cancer; Programmed death-ligand 1; Regional metastasis; Tumor stroma

Core Tip: Programmed death ligand 1 expression was studied in breast cancer (BC) stromal cells using a new antibody against programmed cell death protein 1 ligand 1 (PDCD1 LG1). This study revealed that moderate and pronounced PDCD1 LG1 expression in polymorphic cell infiltrates and the presence of PDCD1 LG1 expression in the fibroblastic stroma are associated with unfavorable BC prognosis. Additionally, PDCD1 LG1+ lymphocyte score ≥ 3 in the absence of lymphovascular invasion (LVI) is associated with a reduced risk of BC regional metastasis; by contrast, in the presence of LVI, it is associated with an increased risk of regional BC metastasis.

INTRODUCTION

Breast cancer (BC) occupies a leading position in most countries of the world in the structure of morbidity and mortality from malignant neoplasms[1,2]. Between 1990 and 2021, the number of BC cases increased from 875657.23 to 2121564.32. The highest age-standardized incidence rate of BC was observed in countries with a high sociodemographic index (SDI), and the lowest was observed in countries with a low SDI[3]. Despite the increase in the incidence of BC, the global age-standardized death rate from this pathology has decreased. However, in different regions, both a decrease and an increase in mortality from BC have been noted, and this indicator does not depend on the SDI[3].

Reducing mortality from BC depends on both early diagnosis of this pathology and correctly choosing treatment tactics. According to the 18^th St. Gallen International Breast Cancer Consensus Conference recommendations, mastectomy or organ-preserving surgery with sentinel lymph node (SLN) biopsy may be performed at the first stage of treatment for early BC. Patients with locally advanced BC, as well as in early BC and unfavorable prognosis factors, including high tumor grade (G3), triple-negative BC (TNBC), overexpression of human epidermal growth factor receptor 2-positive (HER2+) and the presence of lymphovascular invasion (LVI), are indicated for neoadjuvant chemotherapy[4]. In the presence of a complete clinical response or the absence of metastases in the axillary LNs, according to the primary clinical and instrumental examinations, patients may also undergo SLN biopsy.

The results of BC treatment are determined by early diagnosis and the effectiveness of the treatment. The effectiveness of BC treatment depends on an accurate assessment of the disease prognosis and the sensitivity of the tumor to therapy. The most valuable prognostic factors in BC are the T and N stages, tumor grade (G) and BC molecular biological subtype[5-8]. In the study by Pan et al[9], in patients with stage T1 BC, disease recurrence occurred in 13%, 20%, and 34% of cases, respectively, at stages N0, N1, and N2, whereas in patients with stage T2 BC, disease recurrence was detected in 19%, 26%, and 41% of cases, respectively. The risk of disease recurrence is greater in patients with a negative estrogen receptor (ER-) status than in those with a positive estrogen receptor (ER+) status and increases with the number of regional LNs (RLNs) with metastases[8].

Thus, the preoperative assessment of the status of RLNs is highly important. It is carried out on the basis of clinical and instrumental studies (palpation, ultrasound, and magnetic resonance imaging) and cytological examination of LN punctures suspected of being metastatic lesions. Research has shown that the risk of regional metastasis increases with a tumor node size of more than 2 cm[10,11], a multifocal and multicentric type of tumor growth[12,13], a high degree of tumor malignancy[10,14], a close distance from the tumor to the nipple[15], and the presence of LVI[10,14,16,17]. The final conclusion is based on the histological examination of the removed LNs, which can be performed on stationary or frozen sections, as well as cytological and immunocytochemical examination methods. However, these methods have serious drawbacks associated with low sensitivity, a wide range of false-negative results, and the need for expensive specialized equipment[18-20]. In addition, these methods do not allow the assessment of the individual risk of regional metastasis. Moreover, the presence or absence of metastases in the RLNs determines the indications for the volume of axillary LN dissection and neoadjuvant chemotherapy. Thus, the search for new markers associated with a high risk of regional metastasis of BC has not lost its relevance.

In a recent study, we investigated programmed death ligand 1 (PD-L1) expression in BC using a new polyclonal antibody against the protein programmed cell death 1 ligand 1 (PDCD1 LG1). We found that a high expression coefficient of PDCD1 LG1 in tumor cells is an independent factor associated with a high risk of BC regional metastasis[21]. We hypothesized that PDCD1 LG1 expression in stromal cells may also be associated with the risk of BC regional metastasis.

Given the results of a previous study and that the intracellular distribution of PD-L1 is associated with tumor progression factors and the response to chemotherapy and immunotherapy[22,23], the aim of this study was to investigate the features of cytoplasmic and nuclear expression of PDCD1 LG1 in various elements of the tumor stroma and its associations with clinical and pathological BC characteristics.

MATERIALS AND METHODS

Patient characteristics

A total of 148 patients with newly diagnosed BC were included in this prospective single-center observational cohort study. Patients were treated at the Orenburg Regional Cancer Clinic between January 15, 2023, and June 30, 2023. The inclusion criteria were stage Tis-2N0-1M0, radical surgical treatment (R0) at stage 1, and voluntary informed consent to participate in the clinical trial. A total of 107 patients underwent lumpectomy; and 41 patients underwent radical mastectomy. LN dissection at levels 1-2 was performed in all patients. The exclusion criteria were as follows: Receiving chemotherapy, targeted, hormone therapy or radiation therapy before surgery and taking corticosteroids or nonsteroidal anti-inflammatory drugs. Before treatment, all patients underwent a standard clinical and instrumental examination, which included palpation of the mammary glands and RLNs; bilateral mammography; ultrasound examination of the mammary glands and RLNs; trephine biopsy of mammary gland formation with morphological and immunohistochemical examination of the obtained material; fine-needle aspiration biopsy of RLNs suspected of being metastatic lesions, chest X-ray in two projections; ultrasound examination of the liver, retroperitoneal LNs, and pelvic organs; osteoscintigraphy; and examination by a general practitioner and gynecologist.

The study was conducted in accordance with the Helsinki Declaration and internationally recognized guidelines. Ethical approval was obtained from the Ethics Committee of Orenburg State Medical University (Protocol No. 311 dated January 13, 2023). The BC stage was determined according to the 8^th TNM classification of malignant tumors[24]. Clinical data on patients, treatment, histological and immunohistochemical findings, and the number of removed LNs and LNs with metastases were obtained from outpatient records. The characteristics of the patients included in the study are presented in Table 1. Most patients with BC had stages T1c and T2 (60.1% and 29.7%, respectively), stages N0 and N1 (54.7% and 31.8%, respectively), ductal BC (90.5%), G2 and G3 tumors (60.1% and 29.1%, respectively), ER+ (89.2%) and positive status of progesterone receptor (PR+) (72.3%), HER2- BC (85.5%), and luminal A and luminal B HER2- BC (37.8% and 44.6%, respectively). LVI was detected in 56.8% of the tumor samples. In 20 patients, after surgery, according to morphological examination data, metastases were detected in 4 or more LNs, due to which the stage was changed upward.

Table 1 Baseline clinical and pathological information of patients with breast cancer.

Breast cancer characteristics	Patients
	n	%
T stage
Т in situ	4	2.7
Т1в	11	7.5
Т1с	89	60.1
Т2	44	29.7
N stage
N0	81	54.7
N1	47	31.8
N2	16	10.8
N3	4	2.7
Histology
Ductal	134	90.5
Lobular	2	1.4
Papillary	4	2.7
Mucinous	4	2.7
Intraductal	4	2.7
Grade (G)
G1	16	10.8
G2	91	61.5
G3	41	27.7
Lymphovascular invasion
Yes	84	56.8
No	64	43.2
Perineural invasion
Yes	115	77.7
No	33	22.3
Molecular biological subtype
Luminal A	56	37.8
Luminal B HER2-negative	66	44.6
Luminal B HER2-positive	10	6.8
Non-luminal HER2-positive	10	6.8
TNBC	6	4.0
ER status
Negative	16	10.8
Positive	132	89.2
PR status
Negative	41	27.7
Positive	107	72.3
HER2 status
Negative	128	85.5
Positive	20	14.5

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; TNBC: Triple-negative breast cancer.

Immunohistochemistry

The tumor sections were processed routinely and subjected to immunohistochemistry with antibodies against PDCD1 LG1 at a 1:100 dilution (Cloud-Clone Corp., Shanghai, China) and VENTANA PD-L1 (SP142), which was diluted prediluted (Ventana Medical System, Oro Valley, AZ, United States). The staining procedure was performed according to the manufacturer's protocol using a fully automated BOND-MAX staining system (Leica Biosystems Melboume Pty Ltd., Mount Waverley, Australia) for PDCD1 LG1 and BenchMark XT IHC/ISH (Ventana Medical Systems) for SP142. The visualization system included DAB with hematoxylin. For the negative control sections, the primary antibodies were replaced with phosphate-buffered saline, and the samples were processed in the same way. Benign tonsil tissue was used as a positive control. Histological preparations were examined via light microscopy (Levenhuk D740T Digital Microscope connected to a 5.1 MP camera, Levenhuk, Russia). All samples were examined by two researchers (Zubareva EY and Saidler NV) who were blinded to the clinical and pathological data of the patients.

PDCD1 LG1 IHC scoring

The following characteristics of PDCD1 LG1 expression were assessed in the peritumoral stroma: (1) The PDCD1 LG1-positive lymphocyte (PDCD1 LG1+ LF) score was determined in five fields of view at 800 × as follows: The relative density of LFs was determined as follows: Up to 5 cells, 1; 6 to 15 cells, 2; and 16 or more cells, 3. The percentages of PDCD1 LG1+ LFs were estimated as follows: 0%, 0; 1%-25%, 1; 26%-50%, 2; 51%-75%, 3; 76%-100%, 4. The PDCD1 LG1+ LF score was calculated by multiplying the relative density of LFs by the percentage of PDCD1 LG1+ LFs. (range 0–12). Examples of the calculation of the PDCD1 LG1+ LF score are shown in Figure 1. For analysis, samples were divided into two groups according to the 35th percentile: Those with a PDCD1 LG1+ LF score < 3 and those with a PDCD1 LG1+ LF score ≥ 3; (2) The presence or absence of nuclear PDCD1 LG1 expression in the LFs of the peritumoral stroma was assessed in three fields of view at a magnification of 800 × (Figure 2); (3) The severity of PDCD1 LG1 expression in polymorphic cellular infiltrates (PDCD1 LG1+ polymorphic cell infiltrates [PCIs]) was determined in 3 fields of view at 200 × as follows: Absence of PDCD1 LG1+ PCIs, 0; mild PDCD1 LG1+ PCIs (up to 30% of the tumor stroma area), 1; moderate PDCD1 LG1+ PCIs (from 30%-50% of the tumor stroma area), 2; and pronounced PDCD1 LG1+ PCIs (more than 50%), 3. Examples of the severity of PDCD1 LG1+ PCIs are shown in Figure 3; (4) The presence or absence of PDCD1 LG1+ fibroblastic stroma was assessed in three fields of view at 200 × magnification (Figure 4); and (5) The presence or absence of tumor microvessels with PDCD1 LG1 expression in endothelial cells was assessed in five fields of view at 800 × magnification (Figure 5).

Figure 1 Assessment of the programmed cell death protein 1 ligand 1-positive LF score. A: Relative density of immune cells (ICs) – 2; percentage of programmed cell death protein 1 ligand 1-positive (PDCD1 LG1+) ICs - 1. PDCD1 LG1+ LF score = 2 (1 × 2); B: Relative density of ICs - 2; percentage of PDCD1 LG1+ ICs - 1. PDCD1 LG1+ LF score = 2 (2 × 1); C: Relative density of ICs - 3; percentage of PDCD1 LG1+ ICs - 2. The PDCD1 LG1+ LF score = 6 (3 × 2); D: Relative density of ICs - 3; percentage of PDCD1 LG1+ ICs - 4. PDCD1 LG1+ LF score = 12 (3 × 4). Immunohistochemistry staining with antibodies against PDCD1 LG1, 800 ×.

Figure 2 Nuclear expression of programmed cell death protein 1 ligand 1 in lymphocytes of the peritumoral stroma. A: Absence of nuclear programmed cell death protein 1 ligand 1 (PDCD1 LG1) expression in lymphocytes; B: Presence of nuclear PDCD1 LG1 expression in lymphocytes. Immunohistochemistry staining with antibodies against PDCD1 LG1, 800 ×.

Figure 3 The severity of programmed cell death protein 1 ligand 1 expression in polymorphic cell infiltration. A: No polymorphic cell infiltration; B: Weak programmed cell death protein 1 ligand 1-positive (PDCD1 LG1+) polymorphic cell infiltration; C: Moderate PDCD1 LG1+ polymorphic cell infiltration; D: Pronounced expressed PDCD1 LG1+ polymorphic cell infiltration. Immunohistochemistry staining with antibodies against PDCD1 LG1, 200 ×.

Figure 4 Presence of programmed cell death protein 1 ligand 1-positive fibroblastic stroma. A and B: Absence of programmed cell death protein 1 ligand 1-positive (PDCD1 LG1+) fibroblastic stroma; C and D: Presence of PDCD1 LG1+ fibroblastic stroma. Immunohistochemistry staining with antibodies against PDCD1 LG1. Magnification: A and C: 200 ×; B and D: 800 ×.

Figure 5 Presence of programmed cell death protein 1 ligand 1 expression in tumor microvessels. A: Vessel without programmed cell death protein 1 ligand 1 (PDCD1 LG1) expression; B: Vessels with PDCD1 LG1 expression, immunohistochemistry staining with antibodies against PDCD1 LG1, 800 ×.

SP 142 IHC scoring

The assessment of tumor-infiltrating SP142+ immune cells (ICs) was performed according to the Ventana interpretation guide for TNBC. The SP142 + IC score was calculated as the proportion of the tumor area occupied by SP142 + ICs of any intensity. A tumor sample was considered SP142 positive if the SP142 + IC score was ≥ 1% and negative if it was < 1% (Figure 6).

Figure 6 Assessment of the SP142+ immune cell score. А: SP142+ IC0 (< 1%); В: SP142+ IC1 (≥ 1%), immunohistochemistry staining with antibodies against SP142, 200 ×.

Statistical analyses

Statistical processing of the results was performed using Statistica 12.0 software. Correlations between different indicators were estimated by nonparametric gamma correlation. The 35^th percentile was used to distribute patients into groups according to a PDCD1 LG1+ LF scores. χ² tests were performed to analyze differences in distribution between categorized data. Univariate and multivariate analyses were performed to establish independent predictors associated with the risk of regional metastasis in patients with BC. To determine the concordance between the results of PD-L1 expression assessment using PDCD1 LG1 and SP142 antibodies, negative and positive percentage agreements were calculated. P < 0.05 was considered statistically significant.

RESULTS

PDCD1 LG1 expression in peritumoral lymphocytes

The analysis revealed that the PDCD1 LG1+ LF score in the peritumoral stroma was 4.88 ± 3.17 (median 4.50). A PDCD1 LG1+ LF score < 3 was detected in 41 (27.7%) tumor samples, whereas a PDCD1 LG1+ LF score ≥ 3 was detected in 107 (72.3%) tumor samples. The distributions of patients with BC with PDCD1 LG1 + LF scores < 3 and PDCD1 LG1+ LF scores ≥ 3 according to their clinical and pathological characteristics are presented in Table 2.

Table 2 Distribution of patients with programmed cell death protein 1 ligand 1-positive lymphocyte score < 3 and ≥ 3, n (%).

Breast cancer characteristics	PDCD1 LG1+ LF score		P value χ² tests
	< 3	≥ 3
T stage
Т in situ	0 (0)	4 (100)	0.50
Т1в	2 (18.2)	9 (81.8)
Т1с	27 (30.3)	62 (69.7)
Т2	12 (27.3)	32 (72.7)
N stage
N0	16 (19.7)	65 (80.3)	0.03
N1	19 (40.4)	28 (59.6)
N2	6 (37.5)	10 (62.5)
N3	0 (0)	4 (100)
Tumor grade (G)
G1	6 (37.5)	10 (62.5)	0.47
G2	26 (28.6)	65 (71.4)
G3	9 (21.9)	32 (78.1)
Lymphovascular invasion
No	18 (21.4)	66 (78.6)	0.05
Yes	23 (35.9)	41 (64.1)
Perineural invasion
No	31 (27.0)	84 (73.0)	0.70
Yes	10 (30.3)	23 (69.7)
ER status
Negative	3 (18.7)	13 (81.3)	0.39
Positive	38 (28.8)	94 (71.2)
PR status
Negative	15 (36.6)	2 (63.4)	0.13
Positive	26 (24.3)	81 (75.7)
HER2 status
HER2-	39 (30.5)	89 (69.5)	0.05
HER2+	2 (10.0)	18 (90.0)
Molecular biological subtype
Luminal A	18 (32.1)	38 (67.9)	0.26
Luminal B HER2-	20 (30.3)	46 (69.7)
Luminal B HER2+	0 (0)	10 (100)
Non-luminal HER2+	2 (20.0)	8 (80.0)
TNBC	1 (16.7)	5 (83.3)

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; TNBC: Triple negative breast cancer; PDCD1 LG1: Programmed cell death protein 1 ligand 1.

According to the obtained data, a PDCD1 LG1+ LF score ≥ 3 was significantly more common in N0 and N3 than in N1 and N2 (80.3%, 59.6%, 62.5% and 100% of cases, respectively, for N0, N1, N2 and N3, P = 0.03), in the absence of LVI than in its presence (78.6% and 64.1% of cases, P = 0.05), and in HER2+ BC than in HER2- BC (90% and 69.5% of cases, respectively, P = 0.05).

Nuclear expression in peritumoral stromal lymphocytes

LFs with nuclear expression of PDCD1 LG1 were detected in 130 (87.8%) BC samples. The distribution of patients with BC with and without nuclear expression in peritumoral LFs according to their clinical and pathological characteristics is presented in Table 3.

Table 3 Distribution of patients with and without nuclear expression of programmed cell death protein 1 ligand 1 in peritumoral lymphocytes, n (%).

Breast cancer characteristics	Nuclear expression of PDCD1 LG1 in peritumoral lymphocytes		P value χ² tests
	Absent	Present
T stage
Т in situ	0 (0)	4 (100)	0.69
Т1в	2 (18.2)	9 (8.18)
Т1с	12 (13.5)	77 (86.5)
Т2	4 (9.1)	40 (90.9)
N stage
N0	8 (9.9)	73 (90.1)	0.57
N1	8 (17.0)	39 (83.0)
N2	2 (12.5)	14 (87.5)
N3	0 (0)	4 (100)
Tumor grade (G)
G1	4 (25.0)	12 (75.0)	0.15
G2	8 (8.8)	83 (91.2)
G3	6 (14.6)	35 (85.4)
Lymphovascular invasion
No	6 (7.1)	78 (92.9)	0.03
Yes	12 (18.8)	52 (81.2)
Perineural invasion
No	12 (10.4)	103 (89.6)	0.23
Yes	6 (18.2)	27 (81.8)
ER status
Negative	0 (0)	16 (100)	0.11
Positive	18 (13.6)	114 (86.4)
PR status
Negative	10 (24.4)	31 (75.6)	0.004
Positive	8 (7.5)	99 (92.5)
HER2 status
HER2-	18 (14.1)	110 (85.9)	0.07
HER2+	0 (0)	20 (100)
Molecular biological subtype
Luminal A	6 (10.7)	50 (89.3)	0.20
Luminal B HER2-	12 (18.2)	54 (81.8)
Luminal B HER2+	0 (0)	10 (100)
Nonluminal HER2+	0 (0)	10 (100)
TNBC	0 (0)	6 (100)

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; PDCD1 LG1: Programmed cell death protein 1 ligand 1; TNBC: Triple-negative breast cancer.

Nuclear expression in peritumoral LFs was observed significantly more often in the absence of LVI than in its presence (92.9% and 81.2% of cases, respectively, P = 0.03) and with PR+ than with PR- (in 92.5% and 75.6% of cases, respectively, P = 0.004). These data indirectly indicate that peritumoral LFs with nuclear expression of PDCD1 LG1 may have antitumor properties.

Severity of PDCD1 LG1 expression in polymorphocellular infiltrates of the tumor stroma

PDCD1 LG1 expression in PCIs was detected in 54% of cases and represented a cluster of heterogeneous stromal cell populations that morphologically corresponded to LFs, macrophages, granulocytes, fibroblast-like cells, and isolated tumor cells. Mild, moderate, and pronounced PDCD1 LG1+ PCIs were detected in 22%, 19%, and 12% of the tumor samples, respectively. In 68 (46%) tumor samples, PDCD1 LG1+ PCIs were absent. The distributions of patients with BC with different severities of PDCD1 LG1+ PCIs according to their clinical and pathological characteristics are presented in Table 4.

Table 4 Distribution of patients with different degrees of programmed cell death protein 1 ligand 1-positive polymorphic cell infiltration, n (%).

Breast cancer characteristics	Degrees of PDCD1 LG1+ polymorphic cell infiltrations		P value χ² tests
	No, weak	Moderate, pronounced
T stage
Тis	2 (50.0)	2 (50.0)	0.02
Т1в	4 (36.4)	7 (63.6)
Т1с	59 (66.3)	30 (33.7)
Т2	36 (81.8)	8 (18.2)
N stage
N0	58 (71.6)	23 (28.4)	0.70
N1	31 (66.0)	16 (34.0)
N2	10 (62.5)	6 (37.5)
N3	2 (50.0)	2 (50.0)
Tumor grade (G)
G1	10 (62.5)	6 (37.5)	0.09
G2	68 (74.7)	23 (25.3)
G3	23 (56.1)	18 (43.9)
Lymphovascular invasion
No	60 (71.4)	24 (28.6)	0.34
Yes	41 (64.1)	23 (35.9)
Perineural invasion
No	79 (68.7)	36 (31.3)	0.82
Yes	22 (66.7)	11 (33.3)
ER status
Negative	5 (31.3)	11 (68.7)	0.0008
Positive	96 (72.7)	36 (27.3)
PR status
Negative	25 (60.9)	16 (39.0)	0.23
Positive	76 (71.0)	31 (29.0)
HER2 status
HER2 -	99 (77.3)	29 (22.7)	< 0.00001
HER2+	2 (10.0)	18 (90.0)
Molecular biological subtype
Luminal A	42 (75.0)	14 (25.0)	< 0.00001
Luminal B HER2-	54 (81.8)	12 (18.2)
Luminal B HER2+	0 (0)	10 (100)
Nonluminal HER2+	2 (20.0)	8 (80.0)
TNBC	3 (50.0)	3 (50.0)

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; TNBC: Triple-negative breast cancer; PDCD1 LG1: Programmed cell death protein 1 ligand 1.

According to the obtained results, moderate and pronounced PDCD1 LG1+ PCIs were significantly more often observed in Tis and T1b than in T1c and T2 stages (50%, 63.6%, 33.7% and 18.2%, respectively, for Tis, T1b, T1c, and T2, P = 0.02); ER- than in ER+ (68.7% and 27.3% of cases, respectively, P = 0.0008); HER2+ BC than in HER2- BC (90.0% and 22.7% of cases, respectively, P = 0.000001); and non-luminal HER2+ BC, luminal B Her + BC and TNBC than in luminal A BC and luminal B HER2- BC (25.0%, 18.2%, 100%, 80.0% and 50% of cases, respectively, in luminal A, luminal B HER2-, luminal B HER2+, non-luminal HER2+ and TNBC, P = 0.000001). The obtained data indicate that moderate and pronounced PDCD1 LG1+ PCIs are associated with more aggressive BC subtypes.

PDCD1 LG1+ fibroblastic stroma

PDCD1 LG1+ fibroblastic stroma was detected in 41 (28%) tumor samples and was represented by PDCD1 LG1 expression in clusters of fibroblast-like cells of an elongated shape. The distribution of patients with BC with and without PDCD1 LG1+ fibroblastic stroma according to their clinical and pathological characteristics is presented in Table 5.

Table 5 Distribution of patients with and without programmed cell death protein 1 ligand 1-positive fibroblastic stroma, n (%).

Breast cancer characteristics	PDCD1 LG1+ fibroblastic stroma		P value χ² tests
	Absent	Present
T stage
Т in situ	4 (100)	0 (0)	0.026
Т1в	6 (54.5)	5 (45.5)
Т1с	59 (66.3)	30 (33.7)
Т2	38 (86.4)	6 (13.6)
N stage
N0	60 (74.1)	21 (25.9)	0.73
N1	33 (70.2)	14 (29.8)
N2	12 (75.0)	4 (25.0)
N3	2 (50.0)	2 (50.0)
Tumor grade (G)
G1	10 (62.5)	6 (37.5)	0.61
G2	66 (72.5)	25 (27.5)
G3	31 (75.6)	10 (24.4)
Lymphovascular invasion
No	63 (75.0)	21 (25.0)	0.39
Yes	44 (68.8)	20 (31.2)
Perineural invasion
No	80 (69.6)	35 (30.4)	0.16
Yes	27 (81.8)	6 (18.2)
ER status
Negative	8 (50.0)	8 (50.0)	0.03
Positive	99 (75.0)	33 (25.0)
PR status
Negative	31 (75.6)	10 (24.4)	0.57
Positive	76 (71.0)	31 (29.0)
HER2 status
HER2 -	99 (77.3)	29 (22.6)	0.0005
HER2+	8 (40.0)	12 (60.0)
Molecular biological subtype
Luminal A	44 (78.6)	12 (21.4)	0.004
Luminal B HER2-	52 (78.8)	14 (21.2)
Luminal B HER2+	3 (30.0)	7 (70.0)
Non-luminal HER2+	5 (50.0)	5 (50.0)
TNBC	3 (50.0)	3 (50.0)

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; TNBC: Triple-negative breast cancer; PDCD1 LG1: Programmed cell death protein 1 ligand 1.

Thus, PDCD1 LG1+ fibroblastic stroma was significantly more common in T1b and T1c than in Tis and T2 stages (0%, 45.5%, 33.7% and 13.6% of cases, respectively, for Tis, T1b, T1c and T2, P = 0.026), in ER- than in ER+ (50% and 25% of cases, respectively, for ER- and ER+, P = 0.03), and in HER2+ BC than in HER2- BC (60.0% and 22.6% of cases, respectively, for HER2+ and HER2, P = 0.0005). Fibroblastic stroma was significantly more frequently observed in non-luminal HER2+ BC than in other BC subtypes (21.4%, 21.2%, 70%, 50% and 50% of cases, respectively, in luminal A BC, luminal B HER2- BC, luminal B HER2+ BC, non-luminal HER2+ BC and TNBC, P = 0.004). These data indicate that the presence of PDCD1 LG1+ fibroblastic stroma is also associated with more aggressive BC subtypes.

PDCD1 LG1 expression in the endothelial cells of tumor micro-vessels

Tumor microvessels with PDCD1 LG1 expression in endothelial cells were detected in 51 (34%) tumor samples. The distribution of patients with BC with and without tumor microvessels with PDCD1 LG1 expression in endothelial cells according to their clinical and pathological characteristics is presented in Table 6.

Table 6 Distribution of patients with and without programmed cell death protein 1 ligand 1 expression in the endothelial cells of tumor microvessels, n (%).

Breast cancer characteristics	PDCD1 LG1 expression in endothelial cells of tumor microvessels		P value χ² tests
	Absent	Present
T stage
Т in situ	2 (50.0)	5 (50.0)	0.06
Т1в	4 (36.4)	7 (63.6)
Т1с	57 (64.0)	32 (36.0)
Т2	2 (50.0)	2 (50.0)
N stage
N0	58 (71.6)	23 (28.4)	0.36
N1	27 (57.5)	20 (42.5)
N2	10 (62.5)	6 (37.5)
N3	2 (50.0)	2 (50.0)
Tumor grade
G1	12 (75.0)	4 (25.0)	0.41
G2	56 (61.5)	35 (38.5)
G3	29 (70.7)	12 (29.3)
Lymphovascular invasion
No	59 (70.2)	25 (29.8)	0.16
Yes	38 (59.4)	26 (40.6)
Perineural invasion
No	74 (64.3)	41 (35.7)	0.56
Yes	23 (69.7)	10 (30.3)
ER status
Negative	10 (62.5)	6 (37.5)	0.78
Positive	87 (65.9)	45 (34.1)
PR status
Negative	29 (70.7)	12 (29.3)	0.41
Positive	68 (63.5)	39 (36.5)
HER2 status
HER2 -	87 (68.0)	41 (32.0)	0.11
HER2+	10 (50.0)	10 (50.0)
Molecular biological subtype
Luminal A	38 (67.9)	18 (32.1)	0.13
Luminal B HER2-	46 (69.7)	20 (30.3)
Luminal B HER2+	3 (30.0)	7 (70.0)
Non-luminal HER2+	7 (70.0)	3 (30.0)
TNBC	3 (50.0)	3 (50.0)

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; PR: Progesterone receptor; TNBC: Triple-negative breast cancer.

The results obtained indicate that significant correlations between the presence or absence of PDCD1 LG1 expression in the endothelial cells of tumor microvessels and the clinical and pathological characteristics of BC have not been established.

Concordance between assessing PD-L1 expression via PDCD1 LG1 and SP142 antibodies

The analysis revealed positive correlations between PCI severity and the SP142+ IC score (gamma = 0.504; P < 0.00001). Among the 106 BC samples with negative SP142+ IC scores (< 1%), 82 had no or weak PDCD1 LG1+ PCIs. The negative percentage agreement for the markers was 77.4%. Moreover, of the 42 BC samples with positive SP142+ IC scores (≥ 1%), 23 had moderate or severe PDCD1 LG1+ PCI. The positive percentage agreement for the markers was 54.8%.

Predictive value of stromal PDCD1 LG1 expression for assessing the risk of regional BC metastasis

To determine the markers to be included in the univariate and multifactorial analyses, a correlation analysis was first performed. The results are presented in Table 7.

Table 7 Results of the correlation analysis.

Breast cancer characteristics	Gamma correlation	P value
Т stage	-0.240	0.022
Tumor grade	-0.160	0.133
Lymphovascular invasion	0.967	< 0.00001
Perineural invasion	0.380	0.003
ER status	0.606	0.00008
PR status	0.158	0.068
HER2 status	-0.359	0.029
Ki67 index	-0.114	0.147
Subtype	0.104	0.28
PDCD1 LG1+ LF score	-0.414	0.0004
Lymphocytes with nuclear expression of PDCD1 LG1	0.295	0.12
PDCD1 LG1+ polymorphic cellular infiltrations	0.151	0.105
PDCD1 LG1+ fibroblastic stroma	-0.029	0.796
PDCD1 LG1 expression in endothelial cells of tumor microvessels	0.289	0.011
SP142+ ICs	-0.283	0.033

ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; ICs: Immune cells; LFs: Lymphocytes; PDCD1 LG1: Programmed cell death protein 1 ligand 1; PR: Progesterone receptor; TNBC: Triple-negative breast cancer.

According to the data obtained, LVI, perineural invasion (PNI), ER status, HER2 status, the PDCD1 LG1+ LF score and PDCD1 LG1 expression in the endothelial cells of tumor micro-vessels were correlated with the presence of regional BC metastases. These indicators were included in the univariate and multivariate analyses. Considering that a PDCD1 LG1+ LF score ≥ 3 was observed significantly more often both in the absence of metastases and at stage N3, we grouped the presence and absence of LVI with a PDCD1 LG1+ LF score < 3 or a PDCD1 LG1+ LF score ≥ 3. As a result, we obtained 4 combinations of LVI with a PDCD1 LG1+ LF score, which were also included in the univariate and multivariate analyses. The results of the univariate and multivariate analyses are presented in Table 8.

Table 8 Results of univariate and multivariate analyses.

Breast cancer characteristics	Univariate analysis OR (95%CI)	P value	Multivariate analysis OR (95%CI)	P value
T stage
Tis-T1b	1	> 0.05
T1c	1.08 (0.43-5.81)
T2	0.54 (0.11-3.96)
Tumor grade (G)
G1	1	0.05
G2	1.78 (0.59-5.31)
G3	0.86 (0.26-2.87)
Lymphovascular invasion
No	1	< 0.00001	1	< 0.00001
Yes	60.26 (21.60-168.1)		82.73 (24.02-284.89)
Perineural invasion
No	1	0.047	1	> 0.05
Yes	2.23 (1.01-4.91)		0.28 (0.07-1.19)
ER status
Negative	1	0.03	1	> 0.05
Positive	4.08 (1.11-14.98)		7.35 (0.91-59.45)
PR status
Negative	1	> 0.05
Positive	1.08 (0.25-2.23)
HER2 status
HER2 -	1	> 0.05
HER2+	0.47 (0.17-1.30)
Molecular biological subtype
Luminal A	1	> 0.05
Luminal B HER2-	1.08 (0.53-2.21)
Luminal B HER2+	1.73 (0.44-6.81)
Нелюминальный HER2+	0.57 (0.16-1.21)
TNBC	1.15 (0.21-6.22)
Vessels with PDCD1 LG1 expression
absent	1	0.08
present	1.81 (0.91-3.59)
PDCD1 LG1+ LF score
< 3	1	0.018	1	> 0.05
≥ 3	0.41 (0.20-0.86)		0.46 (0.14-1.48)
Combination of LVI and PDCD1 LG1+ LF score
LVI- and PDCD1 LG1+ LF score < 3	1	0.004	1
LVI- and PDCD1 LG1+ LF score ≥ 3	0.13 (0.03-0.53)		0.11 (0.02-0.46)	0.003
LVI+ and PDCD1 LG1+ LF score < 3	9.50 (2.21-40.79)	0.002	11.09 (2.31-53.15)	0.003
LVI+ and PDCD1 LG1+ LF score ≥ 3	25.33 (5.48-117.1)	< 0.00001	44.35 (7.35-267.5)	< 0.00001

CI: Confidence interval; ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; LFs: Lymphocytes; LVI: Lymphovascular invasion; OR: Odds ratio; PDCD1 LG1: Programmed cell death protein 1 ligand 1; PR: Progesterone receptor; TNBC: Triple-negative breast cancer.

According to the univariate analysis data, the following indicators are associated with the risk of regional BC metastasis: The presence of LVI and PNI, the ER status, a combination of a PDCD1 LG1+ LF score < 3 with LVI, and a combination of a PDCD1 LG1+ LF score ≥ 3 with LVI. In the multivariate analysis, the independent predictors of a high risk of regional metastasis were only the presence of LVI, a combination of a PDCD1 LG1+ LF score < 3 with LVI and a combination of PDCD1 LG1+ LF score ≥ 3 with LVI. Moreover, a PDCD1 LG1+ LF score ≥ 3 in the absence of LVI was an independent predictor of a low risk of regional BC metastasis. We analyzed the presence and absence of regional BC metastases according to the combination of the PDCD1 LG1+ LF score with the presence of LVI. The results are presented in Table 9. Thus, in the presence of LVI and PDCD1 LG1+ LF scores < 3 or PDCD1 LG1+ LF scores ≥ 3, metastases in RLNs were detected in 82.6% and 92.7% of patients with BC, respectively, whereas in the absence of LVI and PDCD1 LG1+ LF scores < 3 or PDCD1 LG1+ LF score ≥ 3, metastases in RLNs were absent in 66.6% and 93.9% of patients with BC, respectively.

Table 9 Frequency of regional lymph nodes metastases according to the combinations of programmed cell death protein 1 ligand 1-positive lymphocytes score and lymphovascular invasion, n (%).

Combinations of PDCD1 LG1+ LF score and LVI	N stage		P value
	N0	N1-3
LVI- and PDCD1 LG1+ LF score < 3	12 (66.6)	6 (33.7)	0.00001
LVI- and PDCD1 LG1+ LF score ≥ 3	62 (93.9)	4 (6.1)
LVI+ and PDCD1 LG1+ LF score < 3	4 (17.4)	19 (82.6)
LVI+ and PDCD1 LG1+ LF score ≥ 3	3 (7.3)	38 (92.7)

LFs: Lymphocytes; LVI: Lymphovascular invasion; PDCD1 LG1: Programmed cell death protein 1 ligand 1.

DISCUSSION

BC remains the most common type of cancer in women and one of the leading causes of cancer death worldwide[25]. Determining the RLN status is of key importance when choosing treatment strategies for BC[26-29]. Its assessment is based on clinical and instrumental diagnostic methods and is confirmed by the results of cytological examination of fine-needle biopsy of LNs. The accuracy of such assessments is not high enough and is determined by the experience of the physician, the availability of modern equipment and other factors[30]. However, the RLN status determines the indications for neoadjuvant drug therapy and the volume of axillary LN dissection performed. Thus, in the early stages of BC, SLN biopsy is recommended. If three or more SLNs are affected by metastases, patients undergo axillary LN dissection. However, in the case of metastatic lesions of one or two SLNs, this mutilating operation can be avoided[31,32].

Notably, despite the widespread introduction of SLN biopsy into clinical practice, this invasive procedure has several disadvantages associated with a high frequency of false-negative results, the lack of a unified method for detecting tumor cells in SLNs, and the uncertain diagnostic value of neoadjuvant strategies[19]. Thus, the search for new noninvasive methods for assessing the state of SLNs and new markers associated with a high risk of BC metastasis remains relevant.

According to the literature, one promising marker associated with BC metastasis is PD-L1 expression. PD-1 and its ligands PD-L1 and PD-L2 are representative of the "immunological checkpoint" system and play important roles in the regulation and modulation of the immune response. By suppressing the activity of ICs, this receptor and its ligands help reduce damage to organs and tissues during inflammatory processes and prevent the triggering of autoimmune reactions. In malignant neoplasms, the overexpression of these markers leads to tumor evasion from immune surveillance through direct inhibition of PD-1+ T-cell cytotoxicity and inhibition of apoptosis in tumor cells, thereby promoting tumor progression[33]. Experimental studies have shown that an increase in PD-L1 expression is associated with the activation of epithelial–mesenchymal transition and an increase in the migration of tumor cells[34,35]. However, the results of clinical studies of PD-L1 expression in BC are extremely contradictory. Thus, some authors have shown that high expression of PD-L1 in tumor cells is associated with BC metastasis to RLNs[36-39]. By contrast, other researchers have noted the association of high PD-L1 expression with the absence of metastases to RLNs[40]. There are also studies that have not reported a relationship between the severity of PD-L1 expression and regional BC metastasis[41-43]. Research results assessing the associations between PD-L1 expression and BC clinicopathological characteristics are presented in Table 10.

Table 10 Research results assessing programmed death ligand 1 expression in breast cancer.

Ref.	Number of patients/breast cancer subtype	Anti- programmed death-ligand 1 clone	Association of programmed death-ligand 1 expression with clinical and morphological characteristics of breast cancer	Association of programmed death-ligand 1expression with metastasis in regional lymph nodes
Li et al[36], 2018	112/Ductal BC/subtypes not specified	Polyclonal (Abcam)	Higher in G2-G3 (P = 0.03)	Associated with the presence of metastases
Cirqueira et al[40], 2021	6468/All subtypes	SP263, E1 L3N, 22C3, CAL 10, SP142, B7-H1, Polyclonal	Higher in high Ki67 (P = 0.005), HER2- BC (P = 0.001)	Associated with the absence of metastases
Zhou et al[41], 2018	136 Ductal BC/All subtypes	Monoclonal (ab213524)	Higher in luminal B HER2+, HER2+ non-luminal and TNBC (P = 0.001)	Not associated with the presence of metastases
Catacchio et al[42], 2019	180/All subtypes	Not specified	More often observed in lobular BC (P = 0.0287)	Not associated with the presence of metastases
Evangelou et al[43], 2020	45/All subtypes	Monoclonal E1 L3N	More often observed with a tumor size of more than 2 cm (P = 0.023), G3 (P = 0.017), ER- status (P = 0.001), PR- status (P = 0.002) and Ki67 more than 20% (P = 0.030)	Not associated with the presence of metastases
Zhang et al[37], 2020	4336/All subtypes	SP263, E1 L3N, 22C3, CAL 10, SP142, B7-H1, Polyclonal, NAT105, 28–8, ab58810, Abcam	More often observed in stage III (P = 0.015), ER- status (P = 0.000), HER2+ BC (P = 0.03), and higher levels of Ki67 (P = 0.000)	Associated with the presence of metastases
Kim et al[38], 2017	4578/All subtypes	Abcam	Higher in G3 (P = 0.001)	Associated with the presence of metastases
Zhang et al[39], 2017	2546, All subtypes	ab58810, NA, М1Н1	Higher in G3 (P = 0.001), ER- status (P = 0.008), TNBC (P = 0.001)	Associated with the presence of metastases

BC: Breast cancer; ER: Estrogen receptor; HER2: Human epidermal growth factor receptor 2; LVI: Lymphovascular invasion; PR: Progesterone receptor; TNBC: Triple-negative breast cancer.

The data in the table confirm the inconsistency of the results of studies regarding the relationship of PD-L1 expression with BC clinicopathological characteristics. One of the reasons for these contradictions may be the lack of a unified methodology for assessing PD-L1 expression because the authors used different reagents and different criteria for assessing its expression, such as the types of cells assessed (tumor and stromal), the nature of cell staining (cytoplasmic, membrane) and different threshold values of PD-L1 expression[44,45]. In addition, most authors consider the presence of PD-L1 expression to be a factor that is only positively or negatively associated with the presence of metastases. In our study, we found that a high PDCD1 LG1+ LF score (> 3) can be associated with both the presence and absence of metastases in RLNs, depending on the presence of LVI.

Previously, we investigated PD-L1 expression in BC via a new antibody against PDCD1 LG1[21]. This study revealed that, along with the T stage, the presence of LVI, the BC molecular biological subtype and cytoplasmic PDCD1 LG1 expression in tumor cells are independent predictors of a high risk of BC regional metastasis. In addition, PDCD1 LG1 expression was observed in different cells of the tumor stroma; therefore, we investigated the features of marker expression in stromal cells and their correlation with clinical and pathological characteristics and the risk of BC regional metastasis.

In this study, PDCD1 LG1 expression was detected in LFs of the peritumoral stroma, PCIs, fibroblastic stroma and endothelial cells of tumor microvessels. The analysis revealed that the presence and severity of PDCD1 LG1 expression in the tumor stroma may be associated with both favorable prognostic factors for BC and more aggressive characteristics of BC. For example, moderate and pronounced PDCD1 LG1+ PCIs and PDCD1 LG1+ fibroblastic stroma are associated with unfavorable prognostic factors for BC, including ER- status, HER2+ BC, non-luminal HER2+ BC, luminal B HER+ BC and TNBC, whereas a medium and high relative density of PDCD1 LG1+ LFs in the peritumoral region is observed both in the absence of regional metastases and at stage N3 BC.

According to the results of numerous studies, large tumor size, the presence of LVI, ER+ status and HER2+ BC are associated with the risk of regional metastasis of BC. In a retrospective study by Elleson et al[10] including 738 primary patients with stages I-III BC, the independent predictors of regional metastasis risk were a tumor size greater than 26 mm (P < 0.001), age younger than 40 years (P < 0.001), distance from the tumor to the nipple less than 45 mm (P < 0.001) and the presence of LVI (P < 0.001). Gallagher et al[11] presented data on 618 patients with BC, of whom 296 had metastases in the RLNs. All patients had an ER+ status, and more than 90% of patients had HER2- tumors. A larger tumor size and shorter distance from the tumor to the nipple were associated with the presence of metastases in the RLNs. The study by Malter et al[14] included 1360 patients with primary invasive BC (cT1-cT3). The authors showed that in a univariate analysis, the presence of metastases in the RLNs was associated with a multicentric tumor growth pattern (P = 0.027), G3 (P = 0.046), a tumor location in the retroareolar zone (P = 0.019), the presence of LVI (P < 0.001), and T2 and T3 tumors (P < 0.001). However, in the multivariate analysis, only LVI (P = 0.0001) and tumor size (P = 0.0001) were independent predictors of metastases in the RLNs. Chen et al[16] reported that in the presence and absence of LVI, stage N2 was observed in 19.1% and 9.4%, respectively (P < 0.001), and stage N3 was observed in 26.8% and 7.0%, respectively (P < 0.001). The presence of LVI was associated with a larger tumor size (P = 0.004), G3 tumors (P = 0.001), HER2+ BC (P = 0.016) and a Ki67 index greater than 20% (P = 0.019). In the study of Abdulla et al[17], the independent predictors of metastases in SLNs were G2-G3 (P = 0.011), ER+ (P = 0.009) and the presence of LVI (P = 0.003). A total of 160 patients with newly diagnosed invasive BC who had no metastases in the RLNs according to clinical and instrumental data were included in the study. All patients underwent SLN biopsy, and one to five LNs were examined.

In our study, LVI, PNI, ER status, and HER2 status correlated with the risk of BC regional metastasis. In addition to these characteristics, the PDCD1 LG1+ LF score and the presence of PDCD1 LG1 expression in the endothelial cells of tumor micro-vessels are also associated with the risk of BC regional metastasis. Considering that a PDCD1 LG1+ LF score ≥ 3 was observed significantly more often both in the absence of metastases and at stage N3 of BC, we grouped the presence or absence of LVI with a PDCD1 LG1+ LF score < 3 and a PDCD1 LG1+ LF score ≥ 3. As a result, we obtained 4 combinations of LVI with a PDCD1 LG1+ LF score, which, along with the clinicopathological characteristics of BC, were included in the univariate and multivariate analyses. According to the results of multivariate analysis, the presence of LVI, a combination of a PDCD1 LG1+ LF score < 3 with the presence of LVI, and a combination of a PDCD1 LG1+ LF score ≥ 3 with the presence of LVI were independent predictors of a high risk of regional BC metastasis. However, in the presence of LVI and a PDCD1 LG1+ LF score ≥ 3, metastases in the RLNs were detected significantly more often than in the presence of LVI and a PDCD1 LG1+ LF score < 3 (92.7% and 82.5% BC cases, respectively). Moreover, a combination of a PDCD1 LG1+ LF score ≥ 3 in the absence of LVI was an independent predictor of a low risk of BC regional metastasis. In the absence of LVI and a PDCD1 LG1+ LF score < 3, metastases in RLNs were absent in 66.6% of patients, whereas in the absence of LVI and a PDCD1 LG1+ LF score ≥ 3, metastases in RLNs were absent in 93.9% of patients with BC. Thus, a combined assessment of LVI and the PDCD1 LG1+ LF score allowed us to clarify the risk of regional metastasis of BC.

Importantly, according to the literature, the importance of lymphoid infiltration in BC is quite ambiguous. On the one hand, ICs play an important protective role by recognizing and destroying tumor cells. On the other hand, a tumor is able to change its microenvironment and avoid the immune response; in this case, lymphoid infiltration can contribute to the spread of tumor cells and disease progression, particularly metastasis to RLNs[46]. We believe that in the absence of LVI, the medium and high relative density of PDCD1 LG1+ ICs is associated with favorable prognostic factors, reducing the risk of metastasis to RLNs by 27.6% (from 33.7% to 6.1%), whereas in the presence of LVI, it is associated with unfavorable prognostic characteristics of BC, increasing the risk of regional metastasis by 10.1% (from 82.6% to 92.7%).

Notably, moderate and pronounced PDCD1 LG1+ PCIs, on the one hand, are associated with aggressive characteristics of BC (ER- status of BC, HER2+ BC, non-luminal HER2+ BC, luminal B Her+ BC and TNBC). However, moderate and pronounced PDCD1 LG1+ PCIs were also significantly more often observed in the Tis and T1b stages than in the T1c and T2 stages. We believe that these data may indirectly confirm the results of studies indicating the dual role of PD-L1 expression in malignant neoplasms. Some researchers have shown that PD-L1 expression not only contributes to tumor evasion from the immune response and tumor progression but also, under certain conditions, can activate antitumor immunity[41-44,47]. Thus, Wang et al[47] reported that PD-L1+ macrophages obtained from human breast tumors are more mature and, like M1-type macrophages, have antitumor properties, whereas PD-L1- macrophages are less mature and, like M2-type macrophages, contribute to tumor progression.

An interesting feature of the studied marker (PDCD1 LG1) is the presence of not only cytoplasmic but also nuclear expression in IC. According to the results of recent studies, distinguishing between membrane PD-L1, cytoplasmic PD-L1, nuclear PD-L1 and serum PD-L1 is customary[48-50]. The intracellular distribution of PD-L1 is believed to be essential both for the prognosis of the disease and for the effectiveness of drugs and immunotherapy. Current treatment strategies focus mainly on PD-L1 expressed on the cell membrane. However, these strategies may be ineffective against cytoplasmic and nuclear PD-L1, the role of which has not been sufficiently studied[50]. Thus, the obtained results and literature data indicate the importance and promise of further research aimed at studying the role of the intracellular distribution of PD-L1 in the progression of BC and the development of drug resistance.

CONCLUSION

Thus, the results of this study indicate that assessing the DCD1 LG1+ LF score can help increase the accuracy of predicting the risk of regional metastases in patients with BC. This study had limitations associated with the small sample of patients, the single-center nature of the study, the predominance of patients with early BC and luminal subtypes of BC, and an insufficient number of patients with TNBC and HER2+ BC. In the future, it is necessary to confirm the results obtained in multicenter studies and in larger cohorts of patients, which would increase the generalizability of the results. It is also necessary to validate the method for determining the PDCD1 LG1+ LF score for use in clinical practice. Given that our data indirectly confirm the dual role of PD-L1 expression in malignancies, identifying PDCD1 LG1+ ICs may improve our understanding of their role in BC progression. Thus, we believe that further studies of the role of stromal PDCD1 LG1 expression in metastasis and the development of chemoresistance in BC are promising.