Presence, Subtypes, and Prognostic Significance of Tertiary Lymphoid Structures in Urothelial Carcinoma of the Bladder

Abstract Objective To evaluate the presence and subtypes of tertiary lymphatic structures (TLSs) in urothelial carcinoma of the bladder (UCB) and to analyze their associated clinicopathological characteristics and prognostic significance. Methods The study enrolled 580 patients with surgically treated UCB, including 313 non-muscle invasive bladder cancer (NMIBC) and 267 muscle-invasive bladder cancer (MIBC). The presence and subtypes of TLSs were identified by immunohistochemistry (CD20, CD3, Bcl-6, and CD21). TLSs were classified into non-GC (nGC) TLS and GC TLS subtypes based on germinal center (GC) formation. Disease-free survival (DFS) was used as an endpoint outcome to evaluate the prognostic significance of TLS and its subtypes in UCB. Results TLSs were more common in MIBC than in NMIBC (67.8% vs 48.2%, P < .001), and the tumor-infiltrating lymphocyte (TIL) mean density was significantly higher in MIBC than in NMIBC (24.0% vs 17.5%, P < .001). Moreover, a positive correlation was found between TLS presence and GC structure formation and TIL infiltration in UCB. Endpoint events occurred in 191 patients. Compared to patients with endpoint events, patients without disease progression exhibited higher TIL density and more TLSs (P < .05). Kaplan–Meier curves showed that TLS was associated with better DFS in NMIBC (P = .041) and MIBC (P = .049). However, the Cox multivariate analysis did not demonstrate the prognostic significance of TLS. Conclusions TLS is heterogeneous in UCB, and that TLS and GC structures are related to TIL density and prognostic events. However, TLS as a prognostic indicator remains unclear, warranting further investigation.


Introduction
Urothelial carcinoma of the bladder (UCB) is a common malignant tumor in the urinary system.It is classified into non-muscle invasive bladder cancer (NMIBC) and muscleinvasive bladder cancer (MIBC).The 5-year progression rate of high-risk NMIBC is up to 40% and transurethral resection of bladder tumor (TURBT) combined with intravesical instillation, depending on the risk classification, is its standard treatment approach. 1,2On the other hand, radical cystectomy (RCT) is the recommended treatment option for MIBC, which has a poor prognosis with a 60%-70% 5-year overall survival (OS) rate. 3In recent years, immunotherapy has emerged as a promising treatment alternative for patients with metastatic UCB.However, only a few patients have a long-term, durable response to immunotherapy. 3 Consequently, the central focus of UCB research has shifted to identifying novel biomarkers The Oncologist, 2024, Vol. 29, No. 2 e249 that can accurately stratify prognosis, manage treatment, and predict the response to or efficacy of immunotherapy.
TLSs are ectopic lymphocyte aggregates formed in nonlymphocytic tissues that can develop at sites or target organs of inflammation, autoimmune diseases, and tumors.They present all the features of lymph node structures associated with the generation of an adaptive immune response, including a T-cell zone with mature dendritic cells (DCs), a germinal center (GC) with follicular dendritic cells (FDCs) and proliferating B cells, and high endothelial venules (HEVs). 4Recent studies reported that TLSs correlate with improved clinical outcomes and immunotherapy responses in most tumor types, 5 with a few exceptions. 68][9] Consequently, the presence and characteristics of TLSs, as well as their clinical evaluation and prognostic significance, remain unclear.
Herein, TLSs were divided into the nGC TLS and GC TLS subtypes based on GC formation (a critical TLS maturation characteristic) for the practicality of this clinicopathological investigation.Additionally, we evaluated the TILs, which are significant constituents of the immune microenvironment, encompassing T cells, B cells, natural killer cells, and macrophages.Few studies have associated TILs 9 and their important cell subset CD8 + T cells 10,11 with better survival outcomes in UCB.We aimed to evaluate the presence and subtypes of TLSs in 580 patients with UCB, as well as their clinicopathological features and prognostic value.

Collection of Clinical Samples and Data
This retrospective study was ethically approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (approval No.: QYFY WZLL 27571), China.It included 580 patients with UCB recruited between January 2017 and December 2019.All patients underwent surgical treatment.The exclusion criteria were as follows: (1) Patients with a concurrent presence of other malignant tumors; (2) Patients who have undergone preoperative immunotherapy, targeted therapy, or neoadjuvant therapy; and (3) Insufficient surgical specimens for subsequent immunohistochemical (IHC) testing.Patients' clinicopathological data (gender, age, tumor size, tumor multiplicity, histological subtype, T stage, lymph node metastasis, nerve invasion, lymphovascular invasion (LVI), Ki67 expression, treatment method, and surgical resection status) were obtained from a medical record system and a pathological database.Two uropathologists systematically reviewed and verified all cases based on the eighth edition International Union against Cancer (UICC) TNM Classification of Malignant Tumors 12 and the 2016 World Health Organization Urogenital Tumor Classification. 13isease-free survival (DFS), the period between surgery and the occurrence of the first disease progression event (recurrence or metastasis) or death was used as the study endpoint.Patient follow-up was conducted through the clinical review of records or by telephone, and the last follow-up was scheduled for December 25, 2022.

Assessment of TLSs and TILs
Our analysis was focused on determining the presence and identifying the subtypes of TLSs.We first observed round or round-like lymphocyte aggregates in the hematoxylin and eosin (H&E) sections within the tumor stroma, and peritumoral and intratumoral tissues.Second, we used IHC to establish the presence of TLSs by combining the expression of CD20 + B cells and CD3 + T cells.The presence of TLS structures was not considered when only B cells or T cells aggregated without forming lymphoid aggregates.Finally, TLSs were subdivided into nGC TLS or GC TLS subtypes based on GC formation.Typical well-developed GC structures were observed on the H&E sections, and using the Bcl-6 and CD21 markers for identifying ambiguous GC structures proved beneficial.While the nGC TLSs (Fig. 1E, 1G and 1H) were characterized by CD20 + B cell structures surrounded by CD3 + T cells without GCs, the GC TLSs (Fig. 1F, 1I-1L) were defined by the presence of GC structures within lymphoid aggregates characterized by Bcl-6 + GC B cells and CD21 + FDCs.
Here, stromal TILs in UCB were manually evaluated using H&E sections based on the standardized methodology proposed by the International Immuno-Oncology Biomarkers Working Group for assessing TILs in solid tumors 14 which is detailed as follows.First, when assessing TILs in UCB, the TIL assessment area should be determined within the borders of the invasive tumor, which includes the invasive edge and the stroma pertaining to the fibrovascular cores of invasive papillary structures.At the same time, necrosis, coagulavon artefacts, and the previous biopsy area should be excluded.Subsequently, the stromal TILs (located within the tumor stroma) in the aforementioned areas should be evaluated, excluding intratumoral TILs (found within the tumor epithelial cancer nests).Notably, TILs encompass all mononuclear leukocytes, such as lymphocytes and plasma cells, and exclude polymorphonuclear leukocytes and eosinophils.Moreover, TIL density is defined as the percentage of stromal TIL area in the tumor stromal region.The illustrative tutorial detailing the standard approach for assessing TILs in urothelial carcinoma can be found in the "Supplementary_ Data_File_6" of the article published by the International Immuno-Oncology Biomarkers Working Group (https:// www.ncbi.nlm.nih.gov/pmc/articles/PMC5638696/). 14The UCB images with different TIL densities are depicted in Fig. 1A-1D.Herein, two pathologists completed the assessment of TLSs and TILs and reached a consensus regarding their findings.

Statistical Analysis
The chi-square test or Fisher's exact test was used to analyze categorical variables, while continuous variables were analyzed using the t-test or one-way analysis of variance (ANOVA).Survival curves were plotted and compared using the Kaplan-Meier method and the log-rank test, respectively.A multivariate Cox regression model was employed to evaluate the prognostic significance of TLSs and TILs, controlling for all relevant clinicopathological covariates.All statistical tests were 2-sided, and results with P < .05were considered statistically significant.Data processing and plotting were performed using IBM SPSS V.26.0 or GraphPad Prism V.8.1.

Characteristics of TLSs and TILs in the NMIBC and MIBC Patient Groups
TLSs often develop in the peritumoral area as round or round-like lymphocyte aggregates, easily noticeable in H&E sections, and are rarely found in the intratumoral region (within tumor epithelial cancer nests).Only 3 MIBC cases had intratumoral TLSs in H&E (not confirmed by IHC due to insufficient tissue), and all 3 cases exhibited clear GC TLS at the peritumoral location.

Comparison of the Clinicopathological Characteristics of Groups Classified Based on TLS Presence and Subtype
We further compared the correlation between clinicopathological features in both the NMIBC (Table 2) and MIBC (Table 3) patient groups using TLS presence (TLSs + vs TLSs-) and subtypes (nGC TLS vs GC TLS+) as the grouping variables.Interestingly, the TIL density was higher in the TLSs+ group than in the TLSs-group in both the NMIBC and MIBC patient groups (NMIBC: P < .001;MIBC: P < .001),with fewer endpoint events in the TLSs+ group (NMIBC: P = .032;MIBC: P = .043).Furthermore, the TIL density was higher in the GC TLS group than in the nGC TLS group (NMIBC: P < .001;MIBC: P < .001). Figure 2G and 2H depicts the relationship between TLSs and TIL density in the NMIBC and MIBC groups, showing a positive correlation between TLS formation and GC structures with TIL infiltration.Additionally, the results indicate that the TLSs+ group had a higher LVI than the TLSs-group (P = .020)in the NMIBC patient group, while

Prognostic Significance of TILs and TLSs
Finally, we obtained the prognostic information of 570 patients and analyzed the relationship between endpoint events, TLS presence, and TIL density in the NMIBC and MIBC patient groups.According to the results, patients without disease progression had a higher TIL density than those with endpoint events in both the NMIBC (P = .004,Fig. 2I) and MIBC (P = .012,Fig. 2J) patient groups.Furthermore, compared to patients with TLSs, patients with TLSs+ had significantly fewer endpoint events (NMIBC: P = 0.032; MIBC: P = .043),and patients with GC TLS had the lowest rate of endpoint events (NMIBC: 7/38; MIBC: 7/34; Fig. 2K-2L).

Discussion
Under pathological conditions, the presence of persistent antigenic stimulation leads to lymphocyte accumulation in nonlymphoid tissues, where they initiate lymphoid neogenesis.The lymphoid neogenesis sometimes leads to TLS formation, which can be found in infections, 15 autoimmune diseases, 16 and malignancies 17 and is generally considered the hallmark of an active immune response.The presence of TLSs in tumors varies with different differentiation stages, reflecting the various stages of lymphoid neogenesis.Specifically, the TLS development spectrum ranges from simple lymphocyte aggregate clusters in the early stages of the disease to fully mature, potent TLSs with a GC structure in the late stages.Moreover, the degree of TLS maturation may be related to antitumor immunity. 18In this regard, lower recurrence rates have been reported in colorectal and hepatocellular carcinoma patients with mature TLSs. 19,20ere, the presence of TLSs was defined based on the aggregation of CD3 + T cells and CD20 + B cells for the practicality of this clinicopathological investigation, and GC structure formation was used to determine TLS maturity stages.This classification approach is highly reproducible and convenient for clinical practice.The results revealed that TLSs were heterogeneous in UCB, with the highly invasive MIBC having more TLSs than NMIBC.These findings are consistent with those of Koti et al., who studied TLSs in 28 UCB specimens resected by TRUBT in 2017. 7It has been implied that TLSs change constantly during the dynamic process of UCB progression.Similar observations have been made in other solid tumors.TLSs have been observed in different tumor progression stages or specimens resected after various treatments, showing a high degree of heterogeneity. 4 Furthermore, our results showed that in both the NMIBC and MIBC patient groups, TLS presence and the formation of GC structures were accompanied by an increased TIL infiltration, implying the possibility of a more robust antitumor immune response.This finding is consistent with previous observations that TLSs in high-grade plasmacytoid ovarian cancer recruit various lymphocytes while enhancing B cell and CD8 + T-cell infiltration, resulting in an antitumor immune response. 21dditionally, we discovered that in both the NMIBC and MIBC groups, patients without disease progression had more TLSs and a higher TIL density than those with endpoint events.However, based on the multifactorial Cox analysis results, TLSs and TILs could not be used as independent

e254
The Oncologist, 2024, Vol. 29, No. prognostic indicators for UCB.The multifactorial nature of TLSs could explain this outcome, as the simple typing approach may ignore the composition and role of immune cells in TLSs.However, this does not negate the value of TLSs and TILs in UCB.Various immune cells, including T cells, B cells, DCs, FDCs, and HEVs, are found in TLSs, and their infiltration may mediate antitumor activity.For example, follicular B-cell density in TLSs is positively correlated with more prolonged patient survival in non-small cell lung cancer. 22Additionally, a high density of mature DCs in lymphoid aggregates in the tumor stroma was correlated with the intensity of T-cell activation and longer patient survival in primary melanoma. 23In breast cancer, HEVs in lymphoid aggregates were associated with longer metastasis-free, disease-free, and overall survival of patients. 24Furthermore, the prognostic impact of TLSs may differ depending on their spatial location.In gastric cancer, TLS density in the center of the tumor better reflects the prognosis of patients than that in the infiltrating margins. 25In intrahepatic cholangiocarcinoma, a high intratumoral TLS density indicates a good prognosis, whereas a high peritumoral TLS density correlates with a poor prognosis.This difference could be attributed to the heterogeneity of immune cell subpopulations between the TLSs of different locations. 26On the other hand, a high TIL infiltration in UCB could be used to identify patients with a favorable prognosis.This is consistent with our univariate Cox analysis results, which revealed that TILs influence the prognosis of patients with NMIBC and MIBC.However, there were differences in TIL infiltration in various molecular subtypes of UCB, with basal tumors having the highest degree of TIL infiltration. 9This implies that the use of TILs in clinical prognostic stratification should be supplemented with the molecular characteristics of UCB, which shows a limitation to our study.The high tumor mutational burden in UCB makes it sensitive to immunotherapy, and particularly, the use of checkpoint inhibitors, such as monoclonal antibodies against programmed cell death 1 (PD-1) and its ligand (PD-L1),

e256
The Oncologist, 2024, Vol. 29, No. 2 has improved OS in metastatic UCB. 27Previous research has demonstrated a significant correlation between TLS presence and positive PD-L1 expression in UCB. 28A TLS marker, CXCL13, was associated with improved survival in patients with advanced UCB who received immunotherapy. 29Additionally, a compositional analysis of the tumor microenvironment in a large cohort of MIBC gene expression datasets showed that TLS signatures were associated with improved survival and response to immunotherapy in MIBC.However, they are yet to be validated in preclinical models. 30Overall, while TLSs have great potential as biomarkers of response to immunotherapy in MIBC, their mechanism needs to be supported by further preclinical studies.Although our study did not involve immunotherapy for MIBC, we defined the presence and typing of TLSs in UCB from a clinicopathological standpoint, which is simple, reproducible, and easy to implement clinically, and may provide a foundation for future studies on the role of TLSs in MIBC immunotherapy from a practical clinical perspective.

Conclusion
Herein, we demonstrated the heterogeneity of TLSs in UCB and found that presence was closely associated with clinical outcomes in patients with UCB.However, the clinical application of TLSs as prognostic markers requires further exploration.Characteristics of TLS presentation and GC formation alone are insufficient indicators of the overall status of TLSs.As a result, using TLS to accurately stratify the clinical prognosis of patients with UCB may be challenging.However, TLSs showed great promise as prognostic stratification markers, and their immune cell composition may be essential in their application as prognostic markers.At the same time, there is a lack of uniform evaluation criteria and specific markers for defining and characterizing TLSs, another area that requires attention for their practical clinical application.

Table 1 .
Comparison of TILs density and TLSs in NMIBC and MIBC.

Table 2 .
Clinicopathological characteristics comparison according to the presence and subtypes of TLSs in NMIBC.

Table 3 .
Clinicopathological characteristics comparison according to the presence and subtypes of TLSs in MIBC.