Disease stage-dependent expression of ADC-related biomarkers in urothelial carcinoma Authors: Steffen Rausch, Alicia Bombyk, Lena Müller, Jörg Hennenlotter, Moritz Maas, Simon Walz, Igor Tsaur, Veronika Bahlinger, Viktoria Stühler Published: 2026 ABSTRACT AbstractBackgroundBiomarker-driven therapies have expanded treatment options for advanced urothelial carcinoma (UC), making reliable assessment of therapeutic targets increasingly important for patient selection and treatment sequencing. The aim of this study was to evaluate the expression of clinically relevant biomarkers [Trop2, HER2/neu, Nectin-4, and programmed death-ligand 1 (PD-L1)] across normal urothelium (NT), primary tumors (PT), and lymph node metastases (LNM) in UC, and to explore their potential implications for biomarker-guided therapy.MethodsTissue samples from PT (n=157), NT (n=69), and LNM (n=51) were analyzed by immunohistochemistry using established scoring systems [combined positive score (CPS), H-score, immunohistochemical (IHC) score]. Statistical analyses included paired comparisons, correlation analyses, and survival analyses with Kaplan-Meier, and Cox regression.ResultsPD-L1 expression was overall low but significantly higher in PT and LNM compared to NT (PT: P=0.001; LNM: P<0.001). Nectin-4 showed low to moderate membranous expression without significant differences across tissue types (P=0.68) and no prognostic impact. Trop2 expression was significantly higher in NT compared to PT and LNM (P<0.001) and was not associated with survival. HER2/neu expression was significantly increased in LNM compared to PT and NT (P<0.01), and was associated with advanced nodal and metastatic stage. No significant co-expression patterns were observed among the four biomarkers.ConclusionsPD-L1, Nectin-4, Trop2, and HER2/neu show heterogeneous and largely independent expression across disease stages in UC, supporting individualized biomarker assessment and highlighting the potential relevance of metastatic tissue sampling for therapy guidance.Keywords: Antibody-drug conjugates (ADCs), HER2/neu, Nectin-4, programmed death-ligand 1 (PD-L1), Trop2 Highlight box Key findings • Programmed death-ligand 1 (PD-L1) and HER2/neu expression were significantly increased in urothelial carcinoma compared with normal urothelium, with HER2/neu showing the highest expression in lymph node metastases. • Trop2 expression was significantly reduced in primary tumors and lymph node metastases, whereas Nectin-4 remained relatively stable across tissue compartments. • No relevant co-expression patterns were identified, suggesting largely independent biomarker expression profiles. What is known and what is new? • PD-L1, Nectin-4, Trop2, and HER2/neu are clinically relevant therapeutic biomarkers in urothelial carcinoma and are serve as targets for immunotherapy and antibody-drug conjugate-based treatments. Biomarker expression in urothelial carcinoma is heterogeneous and may vary between primary tumors and metastatic lesions. Reliable biomarker assessment is becoming increasingly important for personalized treatment strategies in advanced urothelial carcinoma. • This study compared biomarker expression across matched normal urothelium, primary tumors, and lymph node metastases in a large urothelial carcinoma cohort. HER2/neu was enriched in lymph node metastases and associated with advanced nodal and metastatic disease, while Trop2 showed higher expression in normal urothelium than in tumor tissue. Independent expression patterns emphasize the need for individual biomarker assessment. What is the implication, and what should change now? • Biomarker testing should consider metastatic tissue whenever available, as expression patterns may differ between primary and metastatic sites. • Comprehensive molecular and immunohistochemical profiling may improve personalized treatment strategies in urothelial carcinoma. Introduction Advanced urothelial carcinoma (UC) remains a therapeutic challenge due to its biological heterogeneity and variable response to systemic treatments across disease stages. Biomarker-driven therapies, including immune checkpoint inhibitors and antibody-drug conjugates (ADCs), have expanded treatment options and improved outcomes in advanced and metastatic UC (1,2). ADCs such as enfortumab vedotin and sacituzumab govitecan target surface antigens including Nectin-4 and Trop2, while HER2-directed therapies are under active clinical investigation. As these strategies rely on target expression, reliable biomarker assessment is essential for patient selection and treatment sequencing (2). However, the spatial and biological heterogeneity of these biomarkers remains incompletely understood, particularly regarding concordance between primary tumors (PT) and metastases. Several molecular targets have gained clinical relevance in UC. Programmed death-ligand 1 (PD-L1) serves as a predictive marker for response to immune checkpoint blockade (3). Nectin-4 and Trop2 are key targets for approved ADCs, while HER2/neu represents a promising target for emerging therapeutic strategies (4,5). HER2/neu, a member of the EGFR family, and target for trastuzumab deruxtecan and disitamab vedotin, regulates cell proliferation and differentiation and shows variable expression in UC (6,7). Given limited data across tissue types and disease stages, this study examines the expression of these biomarkers in normal urothelium (NT), PT, and lymph node metastases (LNM), and explores their implications for biomarker-guided therapy in UC. We present this article in accordance with the REMARK reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/rc). Methods Patient cohort and tissue samples This retrospective study included 157 patients who underwent radical cystectomy for invasive UC (pT1–4) between 2004 and 2016. Follow-up was completed on October 25, 2023. Tissue samples were obtained from PT (n=157), corresponding NT (n=69), and synchronous LNM (n=51) for immunohistochemical (IHC) analysis of PD-L1, Nectin-4, Trop2, and HER2/neu. The study was approved by the Ethics Committee of the Eberhard-Karls-University Tuebingen (project No. 123/2023BO2), and all patients provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Immunohistochemistry TMAs were constructed from PT, LNM, and NT specimens. Immunohistochemistry followed standardized protocols. PD-L1 and HER2/neu were stained automatically on a BenchMark XT and the BenchMark ULTRA system (Ventana) using the mouse monoclonal antibody 22C3 (Dako) according to FDA-approved procedures (dilution 1:50) (8). Nectin-4 and Trop2 were stained manually using for Nectin-4 a polyclonal rabbit antibody (PA5-50463, Thermo Fisher; dilution 1:1,500), and for Trop2 the primary monoclonal rabbit antibody anti-Trop2 (Abcam Clone ERP20043, dilution 1:3,000). Slides were digitized (NanoZoomer 2.0 HT, Hamamatsu) and manually semiquantitatively evaluated using established UC scoring systems: PD-L1 by combined positive score (CPS) (9), Nectin-4 and Trop2 by H-score (0–300) (10), HER2/neu according to IHC scoring criteria of the ASCO/CAP breast cancer guidelines (scores 0–3+) (11). Statistical analysis Data were analyzed using SPSS. Marker correlations across tissue types were assessed by Spearman’s rank correlation with bootstrap resampling. Group comparisons were performed using paired t-tests or nonparametric tests, and associations with clinicopathological parameters were evaluated accordingly. Cancer-specific survival (CSS) was calculated from surgery and analyzed by dichotomized marker expression using Log-Rank test and Cox regression. For survival analyses, biomarker expression was dichotomized using the median value as the cut-off, classifying cases into low- and high-expression groups. Multiple marker correlations were examined within each tissue type. Correlation heatmaps were generated using JASP (University of Amsterdam). Statistical significance was set at P<0.05 with Bonferroni correction. Results Clinicopathological parameters The study cohort included 157 patients, the majority of whom (91.1%) had muscle invasive UC (≥ T2a). Clinical and pathological characteristics are summarized in Table 1. The median follow-up was 15 months (range 0–213 months). During follow-up, 79 patients (50.3%) died due to UC. Median CSS was 38 months. Kaplan-Meier analysis demonstrated significantly worse CSS in patients with advanced tumor stage (≥ T2), positive lymph node status, higher tumor grade (G2/3), synchronous metastases, or distant metastases compared to their respective reference groups (Table 2). Table 1Overview of the study cohort with detailed clinicopathologic characteristicsVariableValue (n=157)Gender Male 110 (70.1) Female 47 (29.9)Age (years), median (IQR)68 (40–92)Primary tumor size (cm)3.2 (0.8–12)Histology Urothelial141 (89.81) Squamous9 (5.73) Sarcomatoid3 (1.91) Adenocarcinomatous3 (1.91) Neuroendocrine1 (0.64)pT stage CIS4 (2.55) pT110 (6.37) pT2a10 (6.37) pT2b22 (14.01) pT3a36 (22.93) pT3b42 (26.75) pT4a25 (15.92) pT4b8 (5.10)Concomitant CIS Yes46 (29.30) No110 (70.06) NE1 (0.64)Nodal stage at first diagnosis N053 (33.76) N134 (21.66) N249 (31.21) N311 (7.01) NE10 (6.37)Metastasis at first diagnosis M0139 (88.54) M118 (11.46)Grading G12 (1.27) G233 (21.02) G3116 (73.89) NE6 (3.82)Lymphatic invasion primary tumor L063 (40.13) L189 (56.69) NE5 (3.18)Vascular invasion primary tumor V0115 (73.25) V135 (22.29) NE7 (4.46)Resection status primary tumor R0119 (75.80) R131 (19.75) R23 (1.91) NE4 (2.55)Neoadjuvant chemotherapy Yes9 (5.73) No148 (94.27)Metastasis Synchronous12 (7.64) Metachronous99 (63.06) None46 (29.30)Overall survival Alive59 (37.58) Deceased98 (62.42)Cancer-specific survival Alive/ non-cancer-related death78 (49.68) Cancer-related death79 (50.32)Time from primary tumor to last follow-up or death (months)15 (0–213)Time from primary tumor to metastasis (months)28.5 (4–103)Data are presented as n (%) or median (range) unless otherwise specified. CIS, carcinoma in situ; G, grading; IQR, interquartile range; L, lymphatic invasion; M, metastasis; N, nodal stage; NE, not evaluable; R, resection status; T, tumor stage; V, vascular invasion. Overview of the study cohort with detailed clinicopathologic characteristics Data are presented as n (%) or median (range) unless otherwise specified. CIS, carcinoma in situ; G, grading; IQR, interquartile range; L, lymphatic invasion; M, metastasis; N, nodal stage; NE, not evaluable; R, resection status; T, tumor stage; V, vascular invasion. Table 2Overview of the study cohort and CSS according to clinicopathological characteristicsVariableMedian CSS (months)HR (95% CI)P valueT stage < T2 vs. ≥ T2NR vs. 315.78 (1.42–23.61)0.005*T stage < T3 vs. ≥ T3NR vs. 252.75 (1.54–4.92)<0.001*pN0 vs. ≥ pN1NR vs. 222.47 (1.45–4.23)<0.001*G1/2 vs. G3NR vs. 311.97 (1.04–3.73)0.03*M0 vs. M154 vs. 73.27 (1.81–5.91)<0.001*Synchronous vs. metachronous21 vs. 1040.42 (0.18–0.99)0.04**, P<0.05. CI, confidence interval; CSS, cancer specific survival; G, grading; HR, hazard ratio; M, metastasis; N, nodal stage; NR, not reached; T, tumor stage. Overview of the study cohort and CSS according to clinicopathological characteristics *, P<0.05. CI, confidence interval; CSS, cancer specific survival; G, grading; HR, hazard ratio; M, metastasis; N, nodal stage; NR, not reached; T, tumor stage. PD-L1 expression PD-L1 expression was predominantly membranous, with occasional cytoplasmic staining, and was detectable in both benign and malignant tissue, showing enhanced expression in the immune cell-rich peritumoral stroma. Overall, PD-L1 expression was low but comparable between PT and LNM and significantly higher than in NT, with no significant difference between PT and LNM. A CPS ≥10 was observed in 10 of 139 cases (7.2%). Paired analyses between PT and LNM (n=38) showed no significant difference (P=0.16), although half of the cases exhibited higher expression in LNM. Compared with NT, LNM displayed significantly higher PD-L1 levels (P<0.001), see Figure 1A,1B, Table 3. Patient-matched pairwise comparisons between tissue compartments are visualized in a scatter plot (n=38; Figure 1B). Expression differed across histological subtypes, with higher median values in squamous variants compared with UC, while adenoid subtypes lacked PD-L1 expression (Table S1). No significant associations were observed between PD-L1 expression and sex, grading, T or M stage. PD-L1 levels did not significantly impact CSS in either PT (median 24 vs. 36 months for high vs. low expression; P=0.55) or LNM (median 108 vs. 36 months; P=0.92). Protein expression patterns in PT; LNM, and NT and their association with CSS. (A) Representative immunohistochemical staining (magnification: 40×) for PD-L1: left, weak expression in LNM (CPS 10.83); middle, moderate expression in PT (CPS 27.35); right, strong expression in PT (CPS 87.18); Nectin-4 in PT: left, no expression (H-score 0); middle, moderate expression in PT (H-score 20); right, strong expression (H-score 200); Trop2: left, strong expression in NT (score 300); middle, moderate expression in PT (score 160); right, moderate expression in LNM (score 200); HER2/neu: left, weak expression in NT (score 1+); middle, intense expression in PT (score 3+); right, intense expression in LNM (score 3+). (B) Comparison of PD-L1 expression (CPS, %) in PT (n=50), corresponding LNM (n=44), and NT (n=46): left, boxplots of PD-L1 expression in PT, LNM, and NT (n=140); right, paired dot plot of PD-L1 expression in PT, LNM, and NT. Comparison of Nectin-4 expression (H-score) in PT, corresponding LNM, and NT: left, boxplots of Nectin-4 expression in PT, LNM, and NT (n=208); right, paired dot plot of Nectin-4 expression in PT, LNM, and NT. Comparison of Trop2 expression in PT, corresponding LNM, and NT: left, boxplots of Trop2 expression in PT, LNM, and NT; right, paired dot plot of Trop2 expression in PT, LNM, and NT. Comparison of HER2/neu expression in PT, corresponding LNM, and NT: left, boxplots of HER2/neu expression in PT, LNM, and NT; right, paired dot plot of Trop2 expression in PT, LNM, and NT. (C) Kaplan-Meier analysis of CSS according to protein expression in PT and LNM. *, P<0.05. CI, confidence interval; CPS, combined positive score; CSS, cancer specific survival; HR, hazard ratio; IHC, immunohistochemical; LNM, lymph node metastasis; NT, normal tissue grading; PD-L1, programmed death-ligand 1; PT, primary tumor. Table 3Data on different protein expression levels in the PT compared with LNM and NTTissueNumber (n)Median (range)Mean± SDP valuePD-L1 PT: PT vs. NT50 vs. 400.57 (0–87.18)5.27±14.540.001* LNM: PT vs. LNM44 vs. 380.77 (0–42.15)3.79±8.080.16 NT: LNM vs. NT46 vs. 360.00 (0–10.14)0.49±1.93<0.001*Nectin-4 PT: PT vs. NT120 vs. 3615.00 (0–140)25.21±31.320.91 LNM: PT vs. LNM39 vs. 2620.00 (0–160)41.73±44.270.43 NT: LNM vs. NT49 vs. 2410.00 (0–210)32.31±44.730.31Trop2 PT: PT vs. NT133200 (0–300)163.54±106.37<0.001* LNM: PT vs. LNM51210 (0–300)204.80±88.200.03* NT: LNM vs. NT52267.50 (100–300)249.13±54.900.007*HER2/neu PT: PT vs. NT1320 (0–3)0.72±1.050.31 LNM: PT vs. LNM511.00 (0–3)1.15±1.140.008* NT: LNM vs. NT580 (0–2)0.41±0.60<0.001**, Fisher’s exact test, P<0.05. LNM, lymph node metastasis; NT, normal tissue grading; PD-L1, programmed death-ligand 1; PT, primary tumor; SD, standard deviation. Data on different protein expression levels in the PT compared with LNM and NT *, Fisher’s exact test, P<0.05. LNM, lymph node metastasis; NT, normal tissue grading; PD-L1, programmed death-ligand 1; PT, primary tumor; SD, standard deviation. Nectin-4 expression IHC analysis demonstrated predominantly membranous Nectin-4 expression. Neither overall comparisons nor patient-matched pairwise analyses revealed significant differences between PT, LNM, and NT. Expression distributions are illustrated by boxplots (Figure 1B). The highest Nectin-4 expression was observed in UC, while squamous cell carcinomas showed variable expression, whereas rare histological subtypes exhibited absent or minimal expression (Table S1). Nectin-4 expression was not associated with CSS in either PT (median 24 vs. 38 months for high vs. low expression; P=0.68) or LNM (36 vs. 108 months; P=0.604), nor with N or M stage, grading, sex. Only advanced tumor stage (≥ T3) was associated with significantly lower Nectin-4 expression (P=0.02, Table 4). Table 4Univariate analysis of protein expression in PT and LNM in relation to clinical parametersVariablePTLNMnMedianMean ± SDP valuenMedianMean ± SDP valuePD-L1 < T2a 10.880.88NENANANANA ≥ T2a490.575.36±14.68440.773.79±8.08 < T3 120.543.91±10.960.7492.032.37±2.630.57 ≥ T3 380.695.70±15.61350.584.16±8.96 G1/2 110.570.84±0.970.74110.53.97±8.610.78 G3360.555.90±16.27301.053.51±7.98 Men330.584.09±9.940.69260.873.68±6.860.94 Woman 170.517.57±21.02180.773.96±9.79 M0 420.815.74±15.650.81390.583.89±8.530.54 M1 80.312.84±6.1451.552.97±3.19Nectin-4 < T2a103033.33±37.530.29NANANANA ≥ T2a 1101524.47±30.79392041.72±44.26 < T33428.7536.35±37.810.02*858.7549.84±39.490.44 ≥ T3861020.81±27.38312039.63±45.79 G1/2 2822.532.49±39.400.211056.1549.61±36.020.45 G3 8811.2522.51±28.542717.541.16±48.02 Men8414.1523.76±32.450.22241534.37±39.920.18 Woman 362528.61±28.68154553.50±49.60 M01031525.01±31.760.83352041.00±46.110.33 M1 172026.45±29.41449.948.08±25.91Trop2 < T2a9230172.22±140.070.531NANA0.82 ≥ T2a 124197.5162.91±104.2250210204.90±89.09 < T336182.5148.61±105.240.2612205192.50±105.410.67 ≥ T397200169.08±106.8039200208.59±83.41 G1/2 27230206.11±96.910.02*11210202.73±93.870.87 G3101185156.20±104.8137210209.73±82.73 Men94202.5161.79±110.640.8033210211.36±97.760.17 Woman 39180167.77±96.5418197,5192.78±80.30 M0116190157.61±110.400.2244210205.80±87.320.82 M1 17210204.00±61,187205198.57±100.61 N0 43200151.28±113.840.59 N181190165.44±101.46HER2/neu < T2a 900.74±1.300.850NANANE ≥ T2a 12300.72±1.035111.15±1.14 < T33600.71±1.140.731100.50±0.920.02* ≥ T39600.72±1.024011.33±1.14 G1/2 2600.56±0.850.541100.77±1.080.20 G310100.77±1.093711.24±1.14 Men9300.80±1.080.143211.30±1.190.26 Woman3900.53±0.96190,50.89±1.04 M0 11500.61±0.960.004*4411.10±1.130.51 M11721.50±1.30711.43±1.27 N0 4300.47±0.950.03* N1 8100.88±1.09*, Fisher’s exact test, P<0.05. G, grading; LNM, lymph node metastasis; M, metastasis; N, nodal stage; NA, not available; NE, not evaluable; NR, not reached; PD-L1, programmed death-ligand 1; PT, primary tumor; SD, standard deviation; T, tumor stage. Univariate analysis of protein expression in PT and LNM in relation to clinical parameters *, Fisher’s exact test, P<0.05. G, grading; LNM, lymph node metastasis; M, metastasis; N, nodal stage; NA, not available; NE, not evaluable; NR, not reached; PD-L1, programmed death-ligand 1; PT, primary tumor; SD, standard deviation; T, tumor stage. Trop2 expression Trop2 immunostaining showed pronounced intra- and interindividual heterogeneity in staining intensity and distribution. Only membranous staining was evaluated, while cytoplasmic signals were excluded. The highest mean Trop2 expression was observed in NT, followed by LNM and PT. Significant differences in Trop2 expression were found between all tissue compartments (PT vs. NT: P<0.001; PT vs. LNM: P=0.03; LNM vs. NT: P=0.007; Table 3, Figure 1B). Trop2 expression levels differed across histological subtypes, with highest median values observed in UC (Table S1). Trop2 expression in PT was not associated with sex or TNM stage, whereas low-grade tumors (G1/2) showed significantly higher expression (P=0.023). In a stratified marker analysis, there was a significant higher Trop2 expression in PT in patients with metastatic stage (M1, Table 5). CSS did not differ significantly according to Trop2 expression in either PT or LNM; although patients with low Trop2 expression showed numerically longer median CSS in PT (38 vs. 30 months for low vs. high expression, P=0.56) and LNM (44 vs. 31 months, P=0.61), neither log-rank testing nor univariate Cox regression demonstrated a significant prognostic impact, see Tables 4-6, Figure 1C. Table 5Stratified analysis using clinically or semiquantitatively relevant cut-offs of the markers. Differences in biomarker positivity rates between disease stages were analyzed using clinically or semiquantitatively relevant cut-offs with Pearson’s-Chi² test or Fisher’s exact test. For PD-L1 positivity was defined with CPS ≥10, for Nectin-4 with H-score ≥100, for HER2/neu with IHC ≥2+, and for Trop-2 with H-score >0VariablePTLNMn/N (%)P value*n/N (%)P value*PD-L1 positive (CPS ≥10) < T2a0/1>0.990>0.99 ≥ T2a5/49 (10.20)4/44 (9.09) < T3a1/12 (8.33)>0.990/90.57 ≥ T3a4/38 (10.52)4/35 (11.42) N00NA–NA N+5/50 (10.00)4/44 (9.09) M04/42 (9.52)>0.994/39 (10.25)>0.99 M11/8 (12.50)0/5 < G30/110.561/11 (9.09)>0.99 ≥ G34/36 (11.11)2/30 (6.66) All 5/50 (10.00)–4/44 (9.09)–Nectin-4 (H-score ≥100) < T2a1/10 (10.00)0.360– ≥ T2a4/110 (3.63)5/39 (12.82) < T3a3/34 (8.82)0.141/8 (12.50)>0.99 ≥ T3a2/86 (2.32)4/31 (12.90) N02/41 (4.87)>0.99NA**NA N+3/70 (4.28)5/39 (12.82) M05/103 (4.85)>0.995/35 (14.28)>0.99 M10/170/4 < G33/28 (10.71)0.091/10 (10.00)>0.99 ≥ G32/88 (2.27)4/27 (14.81) All 5/120 (4.16)–5/39 (12.82)–Trop2 (H-score >0) < T2a6/9 (66.67)0.411/1 (100.00)>0.99 ≥ T2a98/124 (79.03)45/50 (90.00) < T3a28/36 (77.78)0.9410/12 (83.33)0.58 ≥ T3a76/97 (78.35)36/39 (92.31) N030/43 (69.77)0.14-NA N+66/81 (81.48)46/51 (90.20) M087/116 (75.00)0.02*40/44 (90.91)0.54* M117/17 (100.00)6/7 (85.71) < G324/27 (88.89)0.1810/11 (90.91)>0.99 ≥ G378/101 (77.23)34/37 (91.89) All 104/133 (78.20)46/51 (90.20)HER2/neu (IHC-score ≥2+) < T2a2/9 (22.22)>0.99-NA ≥ T2a27/123 (22.00)18/51 (35.29) < T3a6/36 (16.67)0.372/11 (18.18)0.29 ≥ T3a23/96 (23.96)16/40 (40.00) N06/43 (13.95)0.12-NA N+21/81 (25.93)18/51 (35.29) M020/115 (17.40)0.003*15/44 (34.09)0.69 M19/17 (52.94)3/7 (43.86) < G33/26 (11.54)0.153/11 (27.27)0.72 ≥ G325/101 (24.75)14/37 (37.84) All 29/132 (21.97)18/51 (35.29)*, Fisher’s exact test, P<0.05. CPS, combined positive score; G, grading; IHC, immunohistochemical; LNM, lymph node metastasis; M, metastasis; N, nodal stage; NA, not available; PD-L1, programmed death-ligand 1; PT, primary tumor; T, tumor stage. Stratified analysis using clinically or semiquantitatively relevant cut-offs of the markers. Differences in biomarker positivity rates between disease stages were analyzed using clinically or semiquantitatively relevant cut-offs with Pearson’s-Chi² test or Fisher’s exact test. For PD-L1 positivity was defined with CPS ≥10, for Nectin-4 with H-score ≥100, for HER2/neu with IHC ≥2+, and for Trop-2 with H-score >0 *, Fisher’s exact test, P<0.05. CPS, combined positive score; G, grading; IHC, immunohistochemical; LNM, lymph node metastasis; M, metastasis; N, nodal stage; NA, not available; PD-L1, programmed death-ligand 1; PT, primary tumor; T, tumor stage. Table 6CSS defined as a function of protein expression in PT and LNMVariableProtein expressionCSS (months), median (95% CI)HR (95% CI)P valuePD-L1 expression (CPS) PTHigh >0.57 (n=23) 24 (NR)1.09 (0.46–2.55)0.84Low ≤0.57 (n=26)36 (16.8–55.2) LNMHigh >0.77 (n=22) 108 (NR)0.79 (0.31–1.99)0.61Low ≤0.77 (n=22)36 (6.4–65.6)Nectin-4 expression (H-score) PTHigh >15.00 (n=60) 241.05 (0.63–1.75)0.85Low ≤15.00 (n=60)38 LNMHigh >20.00 (n=19) 360.85 (0.32–2.21)0.75Low ≤20.00 (n=20)108Trop2 expression (H-score) PTLow ≤200 (n=73) 380.87 (0.54–1.39)0.56High >200 (n=60)30 LNMLow ≤210 (n=28) 440.81 (0.35–1.87)0.61High >10 (n=23)31HER2/neu expression (IHC-score) PTLow ≤0 (n=78) 551.376 (0.85–2.24)0.20High >0 (n=53)24 LNMLow ≤1 (n=32) 311.015 (0.43–2.42)0.97High >1 (n=19)36CI, confidence interval; CPS, combined positive score; CSS, cancer specific survival; HR, hazard ratio; IHC, immunohistochemical; LNM, lymph node metastasis; NR, not reached; PD-L1, programmed death-ligand 1; PT, primary tumor. CSS defined as a function of protein expression in PT and LNM CI, confidence interval; CPS, combined positive score; CSS, cancer specific survival; HR, hazard ratio; IHC, immunohistochemical; LNM, lymph node metastasis; NR, not reached; PD-L1, programmed death-ligand 1; PT, primary tumor. Her2/neu expression Due to its transmembranous localization, HER2/neu immunostaining was predominantly membranous with largely homogeneous intensity. The lowest HER2/neu expression was detectable in NT and the highest in LNM. No significant difference was observed between PT and NT (P=0.31), whereas expression was significantly higher in LNM compared with PT (P=0.008) and NT (P<0.001), see Table 3. No differences in HER2/neu expression were observed across histological subtypes in PT (Table S1). HER2/neu expression in PT was not associated with sex, T stage, or tumor grade. However, higher expression levels were significantly associated with advanced N and M stage (P=0.03 and P=0.004, respectively). CSS did not differ significantly according to HER2/neu expression in either PT or LNM. Although a numerical trend toward shorter CSS was observed in PT with high HER2/neu expression (24 vs. 55 months, P=0.20), no prognostic impact, and no CSS differences were observed in LNM (36 vs. 31 months, P=0.97), see Tables 4-6, Figure 1C. When analyzing all four biomarkers jointly, no significant correlations between PD-L1, Nectin-4, Trop2, and HER2/neu were observed in PT (n=41). A non-significant positive trend was noted between Trop2 and HER2/neu expression (ρ=0.290, P=0.07) and between Nectin-4 and Trop2 (ρ=0.263, P=0.10). Similarly, no significant correlations were detected in LNM (n=33), with overall weak and heterogeneous associations (e.g., PD-L1 vs. HER2/neu: ρ =−0.308, P=0.08). In NT (n=29), no significant correlations were identified, although weak to moderate positive trends were observed between HER2/neu and both Trop2 and Nectin-4 (each ρ=0.320, P=0.09). Overall, biomarker co-expression patterns were inconsistent across tissue compartments, see Figure 2. Correlation analysis of PD-L1, Nectin-4, HER2/neu, and Trop-2 expression in PT, LNM, and NT. (A) Left: correlations among PD-L1, Nectin-4, HER2/neu, and Trop-2 in PT (n=41). Right: heatmap of pairwise Spearman’s rank correlation coefficients (ρ) among tumor biomarker expression levels in PT tissue (PDL1Tumor, Nectin4Tumor, Trop2Tumor, and HER2/neuTumor). Color intensity indicates the direction and magnitude of the correlations (blue/violet = positive; red = negative), with numeric values representing the corresponding Spearman’s ρ for each biomarker pair. (B) Left: correlations between Nectin-4, Trop-2, HER2/neu, and PD-L1 in LNM (n=33). Right: heatmap of pairwise Spearman’s rank correlation coefficients (ρ) among LNM biomarker expression levels in LNM. (C) Left: correlations among Nectin-4, Trop-2, HER2/neu, and PD-L1 in NT (n=29). Right: heatmap of pairwise Spearman’s rank correlation coefficients (ρ) among NT biomarker expression levels in normal urothelium. (D) Cross-tabulation analysis of the different expression levels of the investigated marker. *, P<0.05. CI, confidence interval; LNM, lymph node metastasis; NT, normal tissue grading; PD-L1, programmed death-ligand 1; PT, primary tumor. Discussion Despite recent advances in systemic therapy, including immune checkpoint inhibitors and ADCs, optimal patient selection in UC remains challenging due to the lack of robust predictive biomarkers. Current EAU guidelines do not recommend routine biomarker-driven treatment decisions, although several markers are under active investigation (12). In this context, understanding the expression patterns of therapeutically relevant targets across disease stages is of particular importance, as treatment decisions increasingly rely on target availability. This study systematically evaluates the expression of PD-L1, Nectin-4, Trop2, and HER2/neu across NT, PT, and LNM in UC. PD-L1 expression was generally low but significantly higher in both PT and LNM compared to NT with no significant difference between tumor sites. This suggests relative stability of PD-L1 expression during lymphogenic progression, consistent with prior reports describing focal and heterogeneous but largely conserved expression patterns (13,14). This observation aligns with previous studies that described a frequently focal, heterogeneous PD-L1 expression (15,16) without a consistent association to tumor stage or survival (13,17). However, conflicting data regarding PD-L1 dynamics in response to systemic therapy or across metastatic settings highlight the strong influence of cohort characteristics and methodological variability, including varying histological subtypes, therapy history, and methodological factors such as antibody selection and cut-off definitions (18-21). In line with previous studies, PD-L1 expression was not associated with survival in our cohort, supporting its limited prognostic value in UC. Nevertheless, its role as a predictive biomarker for immune checkpoint inhibition remains clinically relevant. Nectin-4 showed predominantly membranous and relatively homogeneous expression across all tissue types, without significant differences between PT, LNM, and NT. This finding supports previous reports describing Nectin-4 as a broadly expressed and relatively stable surface antigen in UC, reinforcing its relevance as a therapeutic target for ADCs such as enfortumab vedotin (22-24). Reports of reduced expression in advanced disease may reflect increasing biological heterogeneity or methodological differences (25). In the present cohort, no significant association was found between Nectin-4 expression and CSS, in line with the inconsistent prognostic data reported in the literature (26). I.e., Kobayashi et al. reported an association between high expression and favourable prognosis in early stages, while Tomiyama et al. described a poor prognosis with high Nectin-4 expression (26,27). Despite varying cut-off values, patient populations, and expression heterogeneity for Nectin-4 scores have been reported, these findings support its role as a therapeutic rather than prognostic biomarker. Trop2 showed a distinct expression pattern, with highest levels in NT and lower expression in PT and LNM, although slightly increased levels in LNM versus PT suggest dynamic regulation during metastasis. As a cell adhesion molecule, Trop2 is implicated in tumor progression and is expressed in both NT and tumors (5,28-30). While partly consistent with previous reports, conflicting findings across studies highlight the influence of tumor microenvironment and disease context (31), despite its proposed role in proliferation and survival pathways (32). In our cohort, Trop2 expression was not associated with survival or most clinicopathological parameters, supporting the lack of a robust prognostic role in UC (30) which contrasts with other cancers where high Trop2 expression is associated with poorer prognosis (33-35). Nevertheless, its relevance as a therapeutic target, particularly for ADCs such as sacituzumab govitecan, remains unaffected by these findings (28-30,36). HER2/neu expression was highest in LNM and in PT and significantly associated with nodal and metastatic stage, suggesting a link to more aggressive tumor biology. While previous studies have reported inconsistent associations with lymph node involvement and survival (37-43), no significant correlation with CSS was observed. Overall, HER2/neu appears to reflect aggressive tumor biology but lacks consistent prognostic value in UC (6). However, its increased expression in metastatic tissue highlights its potential relevance as a therapeutic target and supports reassessment in advanced disease settings. This finding is of clinical relevance, as many patients with UC develop metachronous metastatic disease following resection of the PT in curative intention. In current clinical practice, biomarker assessment is frequently performed on the PT and subsequently used to guide systemic treatment decisions throughout the disease course. Our data demonstrate that HER2/neu expression may substantially increase in metastatic lesions, suggesting that reliance on baseline biomarker assessment of the PT alone may underestimate the proportion of patients eligible for HER2-directed ADC therapies. Furthermore, increasing evidence indicates that resistance to ADCs may arise through dynamic alterations in target antigen expression during disease progression and treatment (44,45). Although treatment response and resistance mechanisms were beyond the scope of the present study, our findings support repeated biomarker assessment of metastatic tissue whenever feasible, particularly before initiating HER2-targeted ADC therapy. In an integrative analysis of PD-L1, Nectin-4, Trop2, and HER2/neu, these markers were found to be largely independently expressed, without stable co-expression across tissue compartments. A weak positive trend between Trop2 and HER2/neu in PT suggested potential functional overlaps, though this had no clinical relevance in the present cohort. Previous studies on the relationship between PD-L1 and Nectin-4 have shown inconsistent results, with Tomiyama et al. finding no significant correlation and low co-expression in UTUC (27), while Kobayashi et al. reported an inverse relationship between the two markers in a multivariate immune score (26). The observed lack of coordinated expression underscores the biological heterogeneity of UC and suggests that reliance on a single biomarker may be insufficient for treatment stratification. The heterogeneous and largely independent expression patterns of PD-L1, Nectin-4, HER2/neu and Trop-2 may have important implications for future biomarker-guided treatment strategies. The lack of a consistent correlation suggests that low expression of one biomarker should not be interpreted as a surrogate for low expression of another therapeutically relevant target, reinforcing the need for independent assessment of each biomarker. Rather than arguing against combined treatment approaches, the independent regulation of these biomarkers may provide a biological rationale for targeting complementary pathways simultaneously. This concept is particularly relevant in the context of the recently established combination of enfortumab vedotin and pembrolizumab, which exploits distinct mechanisms of action. Furthermore, patients lacking expression of one therapeutic target may retain expression of another, thereby preserving eligibility for alternative ADC-based therapies. Although these concepts require prospective clinical validation, our findings support comprehensive multimarker profiling rather than reliance on a single biomarker or baseline tissue sample for treatment selection. Furthermore, the observed variability between PT and metastatic sites supports the concept that biomarker assessment based solely on primary tumor tissue may not adequately reflect the therapeutic landscape in advanced disease. This study has several limitations, including its retrospective design, limited sample size for LNM and NT, and the use of immunohistochemistry without complementary molecular analyses. In addition, the lack of treatment and response data precludes conclusions regarding the predictive value of the investigated biomarkers. Prospective studies integrating molecular profiling and treatment outcomes are needed to further define the clinical relevance of biomarker dynamics in UC. Conclusions PD-L1, Nectin-4, Trop2, and HER2/neu exhibit heterogeneous and largely independent expression patterns across disease stages in UC. Taken together, these findings emphasize the potential value of reassessing biomarker expression in metastatic lesions to optimize patient selection for targeted therapies and to account for dynamic biological changes occurring during disease progression. Supplementary The article’s supplementary files as Acknowledgments The authors thank Olga Dobler and Karen Greif for excellent technical assistance. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was approved by the Ethics Committee of the Eberhard-Karls-University Tuebingen (project No. 123/2023BO2), and all patients provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Footnotes Reporting Checklist: The authors have completed the REMARK reporting checklist. Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/rc Funding: None. Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/coif). The authors have no conflicts of interest to declare. Data Sharing Statement Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/dss References Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/prf Supplementary Materials The article’s supplementary files as Data Availability Statement Available at https://tau.amegroups.com/article/view/10.21037/tau-2026-0483/dss