This retrospective, multi-institutional study was approved by the Nara Medical University Ethics Committee (protocol ID: 2891) and conducted in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from participants through posters and/or the institutional website using an opt-out method [18]. We retrospectively reviewed electronic medical records from 642 patients with la/mUC diagnosed between January 2008 and December 2025 at 12 collaborating hospitals. Baseline characteristics, including age, sex, smoking history, Eastern Cooperative Oncology Group Performance Status (ECOG-PS) score, Charlson Comorbidity Index (CCI), primary tumor site, radical surgery status, laboratory data, and metastatic and target lesions, were collected. Follow-up data, including tumor response and duration of response (DOR) based on radiographic examinations, were assessed from the start of 1L therapy to the last documented follow-up or data lock (March 2026).

Patients with treatment-naïve la/mUC were evaluated using the Enfortumab Vedotin Ineligible Criteria (EVITA), defined as follows: i) Hemoglobin A1c ≥ 8% (if this parameter was unavailable, it was replaced by baeline glucose > 150 mg/dL in two consecutive blood samples taken 1 week apart); ii) Grade ≥ 2 sensory or motor neuropathy; iii) Any corneal or retinal abnormality; iv) Estimated glomerular filtration rate (eGFR) < 45 mL/min/1.73 m²; v) ECOG-PS score ≥ 2.

In this study, an estimated GFR < 45 mL/min/1.73 m², calculated using the Japanese equation for estimated GFR from serum creatinine, was adopted as a substitute for the original criterion “creatinine clearance or GFR < 45 mL/min” for practical purposes [19–21].

Because of the retrospective nature of this study, the regimen, dose reduction, dose interruption, treatment duration, and number of cycles were determined by the attending physicians in consultation with the patients, based on factors such as adverse events, organ function, CCI, and ECOG-PS score. Intravenous EV was administered on days 1 and 8, and intravenous pembrolizumab at a dose of 200 mg (or 400 mg every 6 weeks) was given after the enfortumab vedotin infusion on day 1 of each 3-week cycle. For EV administration, a dose of 1.25 mg/kg was considered the standard induction regimen 5. Dose modifications for EV were implemented in stepwise 0.25 mg/kg reductions as follows: 1.25, 1.0, 0.75, and 0.5 mg/kg as the minimal dose. Interruptions of EV and/or pembrolizumab were carefully considered, primarily based on adverse events. Conventional platinum-based CT (Pl-CT) regimens—including gemcitabine plus cisplatin (GC), gemcitabine plus carboplatin (GCarbo), and dose-dense methotrexate, vinblastine, doxorubicin, and cisplatin (DD-MVAC)—were administered as previously reported [22, 23]. The analysis also included patients receiving 1L non-Pl-CT, such as taxane-based regimens, which were classified as “others.”

The overall objective response was categorized as CR, PR, stable disease (SD), or progressive disease (PD) according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 [24], based on computed tomography and/or magnetic resonance imaging, as previously described. In the analysis of 1L EVP therapy, the percentage change from baseline in the sum of the diameters of measurable target lesions at the time of the objective response for each patient was illustrated using waterfall plots, and longitudinal changes in these diameters for each patient were depicted using spider plots.

Organotropism-related differences in treatment response (organ-specific response) were evaluated separately for primary or locally recurrent lesions, LN mets, lung mets, and liver mets. In patients with metastases involving multiple sites or organs, only the responses of the lesions of interest were considered to assess organotropism-related differences in treatment response. The response (CR, PR, SD, or PD) was determined in the same manner as the overall objective response.

This study investigated the prognostic impact of several blood-based inflammation and nutrition markers in la/mUC patients treated with 1L EVP, including the neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), systemic inflammation response index (SIRI, defined as monocyte count × neutrophil count/lymphocyte count), systemic immune-inflammation index (SII, defined as platelet count × neutrophil count/lymphocyte count), and C-reactive protein (CRP)–albumin–lymphocyte (CALLY) index (calculated as serum albumin × total lymphocyte count/[CRP × 10⁴]) [25].

TRAEs were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 5.0, based on clinical laboratory tests, physical examinations, and medical interviews conducted by physicians or investigators. The dates of AE onset and recovery were monitored. Adverse Events of Special Interest (AESI) identified for this study included five categories: skin toxicity, peripheral neuropathy, dysgeusia, thyroid disorder, and interstitial lung disease (ILD). Events that changed in grade for individual patients were analyzed independently using Kaplan-Meier methods for each respective grade. Patients without a corresponding AE of a particular grade were censored in the time-to-event analysis for that grade at the earlier of the following: last dose date plus 1 month, data cutoff, or end-of-study date.

Statistical analyses were performed, and figures were generated using GraphPad Prism version 10 (GraphPad Software, San Diego, CA, USA). Statistical significance was set at p < 0.05. Clinicopathological characteristics were compared using the chi-squared test, Fisher’s exact test, Mann-Whitney U test, and Kruskal-Wallis test, as appropriate. Disease progression was defined as radiographic PD according to RECIST version 1.1 or death from any cause. Progression-free survival (PFS) was estimated from the start of 1L therapy using the Kaplan-Meier method and compared between groups using the log-rank test, with p values provided.

The use of 1L EVP for la/mUC was approved by the PMDA of Japan on September 27, 2024. A total of 65 patients with systemic therapy-naïve la/mUC were diagnosed between October 2024 and December 2025. Of these, 55 (85%) received 1L EVP therapy, while 10 (15%) received non-EVP therapy. Among the 10 patients in the non-EVP group, five received GCarbo, three received DD-MVACarbo [26], one received GC, and one did not receive any systemic therapy (best supportive care [BSC]). Table S1 lists details of patients diagnosed with la/mUC between October 2024 and December 2025 who did not receive 1L EVP therapy, including reasons for not selecting EVP. Baseline characteristics of the patients are compared between the two groups (Table 1), showing that younger age (74 vs. 79 years), absence of diabetes mellitus (13.0% vs. 40.0%), presence of liver mets 24.0% vs. 0.0%), larger tumor volume (median sum, 49 mm vs. 32 mm), better PS (ECOG-PS ≤ 1: 89% vs. 50%), and a lower number of EVITA were significantly or marginally associated with the selection of EVP.

A multicenter database included 642 patients with mUC diagnosed between January 2008 and December 2025; among these, 65 (10.1%) received BSC without any systemic therapy, whereas 55 (8.6%), 281 (43.8%), 13 (2.0%), 178 (27.7%), and 50 (7.8%) patients were treated with 1L EVP, GC, DD-MVAC, GCarbo, and other regimens, respectively. Baseline characteristics of the six groups are summarized in Table S2, demonstrating significant differences in several clinical characteristics. Notably, renal function, as assessed by eGFR, was higher in the GC and DD-MVAC groups (eGFR ≥ 45 mL/min/1.73 m²) compared with the EVP, GCarbo, and other groups (eGFR < 45 mL/min/1.73 m²).

The ORR, disease control rate (DCR), and PD rate for 1L EVP therapy were 66% (CR: 11% and PR: 55%), 83% (SD: 17%), and 17%, respectively, whereas those for conventional CT were 42% (CR: 6.2% and PR: 36%), 68% (SD: 25%), and 32%, respectively (Figure 1A). The change in tumor size from baseline was illustrated in waterfall plots, showing a median change of –40.6% (95% confidence interval [CI]: –30.8% to –55.8%) (Figure 1B). Although there was no substantial difference in tumor response among conventional CT regimens (ORRs: 39–43%), the PD rate in the “other” group was lower compared with Pl-CT (32% vs 46%). Organotropism-related differences in treatment response were evaluated separately for primary or locally recurrent lesions, LN mets, lung mets, and liver mets (Figures 1C–F). The corresponding ORRs were 56%, 71%, 54%, and 85%, respectively. Notably, the CR, PR, and DCR rates for liver metastases were 31%, 54%, and 100%, respectively, indicating that organotropism for 1L EVP was highest in liver metastatic lesions.

Next, overall ORRs were compared among 1L systemic therapies across different disease subsets, including the lower urinary tract UC subset, upper urinary tract UC subset, LN mets only subset, visceral mets-present subset, liver mets-absent subset, and liver mets-present subset (Figure 2). Overall, 1L EVP demonstrated sustained oncologic efficacy compared with conventional CT across all disease subsets. Notably, the response to 1L EVP was relatively higher in the liver mets-present subset compared with other subsets (ORR: 72% vs. 61–68%). The response to 1L EVP in the upper urinary tract UC subset was comparable with the lower urinary tract UC subset (ORR: 65% vs. 67%). The comparison of clinical background between the bladder UC and the upper urinary tract UC subsets demonstrated lower eGFR in the latter subset (Table S3).

Figure 3A presents PFS from the start of 1L systemic therapy in the overall population across selected 1L regimens. The median PFS (95%CI) for EVP, GC, GCarbo, and other regimens were 14.5 months (10.5–not determined), 10.5 months (8.4–13.3), 9.2 months (6.7–14.2), and 8.5 months (4.7–11.0), respectively. No significant difference was observed between 1L EVP and other CT groups, partly due to the limited follow-up duration of the EVP group (log-rank p = 0.14). DOR among responders to 1L therapy was compared using Kaplan-Meier estimation across the 1L regimens, showing a consistent durable response across different disease subsets (Figures 3B–F). The majority of patients who achieved an objective response to 1L EVP therapy demonstrated a durable response. A spider plot was generated to visualize longitudinal changes in tumor size among patients receiving 1L EVP (Figure 3G). A swimmer plot of patients receiving 1L EVP illustrated the time to first radiologic response, treatment discontinuation of EV and pembrolizumab, disease course, and DOR among patients achieving CR or PR (Figure 3H). In la/mUC patients treated with 1L EVP, there was no association of PFS and OS with several blood-based inflammation and nutrition markers including CRP, NLR, MLR, PLR, SIRI, SII, PNI, and CALLY index (Figure S1). Among 55 patients treated with 1L EVP, 5 (9.1%) received subsequent systemic therapy consisting of GC in two, GCarbo in one, DD-MVAC in one, and FGFR3 inhibitors in one patient.

In the safety analysis set of 40 patients, all-grade TRAEs were reported in 36 patients (90%), with grade 1–2 events occurring in 30 patients (75%) and grade 3–4 toxicities in 6 patients (15%). No grade 5 TRAEs were observed in this cohort. Details of EV-related and pembrolizumab-related adverse events are presented in Table S4 and Table S5. The most common grade 1–2 TRAEs were skin toxicity (55%), anorexia (43%), anemia (38%), and dysgeusia (33%) (Figure 4A). Grade 3–4 TRAEs included skin toxicity (7.5%), anorexia (5.0%), anemia (5.0%), gastrointestinal disorders (2.5%), renal dysfunction (2.5%), and ILD (2.5%). Most AESIs occurred within 3 months, whereas the onset or grade increase of peripheral neuropathy was more frequent after 3 months (Figure 4B). The incidence of the first AESI onset demonstrated a temporal pattern characterized by either an acute peak (e.g., skin toxicity) or a delayed peak (e.g., peripheral neuropathy) (Figure 4C). The median number of administered EV cycles was 4 (range, 1–13), 5 (2–13), and 10 (4–13) in the overall population, responders (CR + PR), and CR population, respectively (Figure 4D). A similar trend was observed for the number of pembrolizumab cycles. Dose modifications and interruptions of EV were common, with most patients experiencing at least one dose reduction within the first 2 months, whereas dose modifications after 4 months were uncommon (Figure 4E). However, pembrolizumab dose interruptions due to TRAEs occurred less frequently than those of EV (Figure 4F).

Additionally, to investigate whether certain host-related conditions rather than tumor status were associated with the response to 1L EVP, we performed a subgroup analysis comparing responders and non-responders in 55 patients treated with 1L EVP. Among the evaluated parameters, age > 70 years, adequate renal function (eGFR ≥ 45 mL/min/1.73 m²), and an EVITA score of 0 were associated with a higher ORR compared with their respective counterparts (Figure 5A). In the safety analysis set of 40 patients, age (cutoff: 70 years) and CCI (0 or ≥ 1) did not significantly affect the incidence of TRAEs, whereas eGFR < 45 mL/min/1.73 m² and EVITA score ≥ 1 were associated with a higher incidence of skin toxicity (Figure 5B).