Diagnostic uncertainty of clinical T2 disease and its impact on treatment stratification in upper tract urothelial carcinoma: a multicenter retrospective study

Abstract

Accurate clinical staging of upper tract urothelial carcinoma (UTUC) remains challenging, and discrepancies between clinical and pathological T stages may result in missed opportunities for appropriate treatment. This study aimed to elucidate real-world discrepancies between clinical and pathological T staging in UTUC, emphasizing the diagnostic uncertainty associated with clinical T2 (cT2) disease. A retrospective analysis was conducted involving 739 patients with cTa–4N0M0 UTUC who underwent radical nephroureterectomy (RNU) at multiple institutions. Patients who received neoadjuvant therapy or had missing pathological T stage data were excluded. Discrepancies between clinical and pathological staging, treatment patterns, and survival outcomes were evaluated. Among 739 patients, substantial discrepancies between clinical and pathological T stages were observed, with an overall diagnostic concordance rate of 44.7%. Only 16% of cT2 cases were pathologically confirmed as pathological T2 (pT2), whereas most were either down-staged to ≤pT1 or up-staged to ≥pT3, indicating marked diagnostic uncertainty. Clinical heterogeneity within the cT2 group resulted in marked survival differences based on pathological stage, as pT3 patients had significantly worse outcomes than pT2 patients. Within the pT3 subgroup, lymphovascular invasion independently correlated with poor prognosis, whereas adjuvant therapy improved survival. However, only approximately one-third of patients with pT3 disease received adjuvant therapy, indicating a treatment gap in this high-risk population. This multicenter study identified real-world discrepancies between clinical and pathological staging in UTUC. cT2 disease represents a heterogeneous and diagnostically uncertain category, requiring careful interpretation for appropriate treatment planning.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-42876-3.

Introduction

Urothelial carcinoma (UC) of the upper urinary tract (UTUC) accounts for approximately 5%–10% of all UC cases¹. Radical nephroureterectomy (RNU) with bladder cuff excision remains the standard treatment for patients with non-metastatic UTUC (cTa-3N0M0). However, improvements in oncological outcomes over time have been limited, creating the need for new treatment approaches^2,3. Accurate staging is essential for optimizing treatment strategies in UTUC. Although computed tomography urography (CTU) represents the most commonly used imaging modality, preoperative determination of the clinical T stage remains difficult with current imaging techniques^4,5.

To address this issue, two nationwide studies were previously conducted using Japan’s Hospital-Based Cancer Registry (HBCR) to evaluate discrepancies between clinical and pathological T stages in UTUC. One study revealed that T-stage discrepancies occurred more frequently in UTUC than in bladder cancer, especially among patients with clinically organ-confined disease⁶. Another study demonstrated that patients whose disease was up-staged from clinical to pathological T stage exhibited significantly worse overall survival⁷. Although pathological T stage directly determines prognosis, these findings collectively raise concerns that clinical under-staging may compromise preoperative risk assessment and result in missed opportunities for appropriate perioperative treatment, including both adjuvant and potentially neoadjuvant treatment strategies.

However, these registry-based analyses lacked detailed clinical information, including imaging modality, tumor size, pathological characteristics, and treatment patterns such as lymph node dissection (LND) and perioperative chemotherapy. To address these limitations, a retrospective multi-institutional study was conducted using comprehensive clinical data from the Michinoku database. The objectives were to elucidate real-world discrepancies between clinical and pathological T staging in UTUC, particularly focusing on the diagnostic uncertainty associated with clinical T2 (cT2) disease and on prognostic determinants in pathological T3 (pT3) tumors to inform optimal treatment planning.

Patients and methods

Data sources

This study utilized data from the Michinoku database, which includes patients with UTUC who underwent RNU at 10 academic centers between January 1994 and April 2024, as previously described⁸. Among 1,124 patients with UTUC registered in the database, 1,034 were included in the present analysis after excluding those with concomitant bladder cancer. Patients with missing cT stage information (n = 27) or cN1–2 disease (n = 122) were then excluded, resulting in 885 cTa–4N0M0 cases. After further excluding those who received neoadjuvant therapy (NAC) (n = 118), did not undergo RNU (n = 9), or had missing pathological T stage data (n = 19), 739 patients remained in the final analysis (Fig. 1).

Fig. 1.

Patient eligibility criteria. UTUC, upper tract urothelial carcinoma; RNU, radical nephroureterectomy.

Statistical analyses

Variables between groups were compared using Pearson’s chi-squared test. Overall survival (OS) and metastasis-free survival (MFS) were analyzed using the Kaplan–Meier method and compared with the log-rank test. Cox proportional hazards regression analysis was applied to identify prognostic factors for OS and MFS. OS was defined as the time from surgery to death from any cause, whereas MFS was defined as the time from surgery to the occurrence of distant metastasis or death. Statistical significance was set at P < 0.05. All statistical analyses were conducted using R software, version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria).

Ethical considderations

The study protocol and data processing were approved by the Institutional Review Board of the University of Tsukuba Hospital (approval number: R05-121) based on the central approval granted by the Ethics Committee of the Hirosaki University School of Medicine (approval number: 2023–063–1), which served as the central ethics committee for this study. The requirement for informed consent was waived by the Ethics Committee of the Hirosaki University School of Medicine due to the retrospective nature of the study.

Results

A total of 739 patients with cTa–4N0M0 UTUC who underwent RNU were included from the Michinoku database after excluding those who received NAC or had missing pT stage data (Fig. 1). Baseline characteristics of the study cohort are presented in Table 1. The median age was 73 years, and 73% of patients were male. Hydronephrosis was observed in 43% of patients, and 15% had impaired renal function, defined as an estimated glomerular filtration rate (eGFR) ≤ 40 mL/min/1.73 m². The cT stage distribution was cTa–1 in 35.9%, cT2 in 43.7%, cT3 in 19.9%, and cT4 in 0.5%. Pathologically, 49.5% were classified as pTa–1, 14.7% as pT2, 33.7% as pT3, and 2.0% as pT4. Figure 2 A–B shows MFS and OS stratified by pathological T stage. Survival outcomes progressively worsened with increasing pathological T stage, and patients with pT3 or pT4 disease had significantly poorer MFS and OS compared to those with pTa–1 or pT2. The prognosis of patients with pT3 was clearly inferior to that of pT2, confirming the adverse impact of advanced pathological stage.

Table 1.

Baseline characteristics of 739 patients with cTa–4N0M0 UTUC treated with RNU.

Fig. 2.

Kaplan–Meier curves for (A) metastasis-free survival (MFS) and (B) overall survival (OS) according to pathological T stage in patients with upper tract urothelial carcinoma (UTUC).

Discrepancies between clinical and pathological T stages were next assessed (Fig. 3A and Supplementary Table 1). Among patients clinically staged as cTa–1, cT2, and cT3, accurate staging—defined as concordance between clinical and pathological T stage—was achieved in 71%, 16%, and 60% of cases, respectively. For cT3, 81% of cases were pathologically staged as pT2 or higher, suggesting that cT3 had relatively high sensitivity for detecting advanced disease. In contrast, only 16% of cT2 cases were pathologically confirmed as pT2, whereas 46% were down-staged to pTa–1 and 38% were up-staged to pT3 or higher. To account for potential temporal effects, era-specific analyses were performed, with patient distributions shown in Supplementary Fig. 1A. Because only one patient was included in Era 1 (1994–2004), analyses were limited to Era 2 (2005–2014) and Era 3 (2015–2024). The patterns of T-stage discrepancy were broadly similar between these eras (Supplementary Fig. 1B and C). These findings demonstrate that cT2 represents a diagnostically uncertain category encompassing tumors with widely divergent pathological stages and prognoses. Given this heterogeneity, the subsequent analysis focused on the cT2 subgroup to clarify its clinical and prognostic characteristics.

Fig. 3.

Proportion of discrepancies between clinical and pathological T stages in each clinical T stage

Baseline characteristics of patients with cT2 disease according to pathological stage (≤ pT2 vs. ≥pT3) are presented in Table 2. No significant differences were observed between the two groups regarding age, sex, presence of hydronephrosis, biopsy status, or LND. In contrast, impaired renal function, defined as an eGFR ≤ 40 mL/min/1.73 m², occurred more frequently in patients with ≥pT3 disease compared to those with ≤pT2 disease (22.3% vs. 10.7%, p = 0.008). Kaplan–Meier analysis demonstrated significant differences in MFS and OS among patients with ≤pT2 and ≥pT3 disease (Fig. 4 A and B). These results indicate that although cT2 disease appears to represent a single clinical category, it encompasses a heterogeneous population, with a substantial subset upstaged to ≥pT3 and exhibiting markedly poorer outcomes. Treatment and recurrence patterns are shown in Fig. 4C. The rate of LND was comparable between the two groups, whereas adjuvant therapy was more frequently administered in patients with ≥pT3 disease (40% vs. 4%), who also experienced a higher recurrence rate (37% vs. 11%). Collectively, these findings demonstrate that cT2 represents an unreliable clinical category that includes a significant proportion of patients with biologically aggressive disease, posing challenges for optimal treatment stratification. To identify prognostic determinants of these aggressive cases, the subsequent analysis focused on patients with pT3 disease.

Table 2.

Baseline characteristics of patients with cT2 disease according to pathological T stage (≤pT2 vs. ≥pT3).

Fig. 4.

Kaplan–Meier curves for (A) MFS and (B) OS in patients with cT2 disease, stratified by pathological T stage (≤pT2 vs. ≥pT3). (C) Treatment and recurrence patterns in cT2 disease according to pathological T stage. MFS, metastasis-free survival; OS, overall survival; LND, lymph node dissection.

To evaluate the prognostic value of clinical T stage in patients with pT3 disease, the cohort was stratified by preoperative cT classification (cTa–1, cT2, and cT3). As presented in Supplementary Table 2, baseline characteristics such as age, sex, renal function, and tumor histology were generally comparable among the three groups. Treatment and recurrence patterns are shown in Fig. 5A. The rates of LND, adjuvant therapy, and recurrence were similar across the three groups. Multivariate analysis of patients with pT3 disease identified lymphovascular invasion (LVI) as an independent adverse prognostic factor (hazard ratio [HR] 1.73, 95% confidence interval [CI] 1.05–2.85, p = 0.032), whereas adjuvant therapy was significantly associated with improved OS (HR 0.48, 95% CI 0.26–0.88, p = 0.018) (Fig. 5). Clinical T stage, LND, and other factors were not prognostic. Similar results were observed in the models for MFS, where LVI remained a significant adverse prognostic factor, and adjuvant therapy demonstrated a favorable association (Supplementary Fig. 1). Collectively, these findings indicate that prognosis in pT3 disease is primarily determined by pathological and treatment-related factors rather than preoperative clinical T stage. The results underscore the limited accuracy of clinical staging in UTUC, especially for cT2 disease, which exhibits substantial variability in pathological characteristics and survival outcomes.

Fig. 5.

(A) Treatment and recurrence patterns in patients with pT3 disease according to clinical T stage. (B) Multivariate Cox regression analysis for overall survival in patients with pT3 disease. Adjuvant therapy consisted exclusively of chemotherapy regimens (mainly GCis and GCarbo, with minor use of MVAC and others); no patients received adjuvant immune checkpoint inhibitors such as nivolumab. HR, hazard ratio; CI, confidence interval; LND, lymph node dissection.

Discussion

Despite the routine use of CTU for clinical staging of UTUC, no reliable imaging criteria exist to accurately distinguish ≤pT2 disease preoperatively⁹. Consistent with previous nationwide analyses using the HBCR, this study also showed that cT2 had the lowest diagnostic accuracy⁶. In contrast, cT3 disease showed relatively high concordance with pathological staging, as most cases were pathologically ≥pT2, aligning with previous reports¹⁰. These observations suggest that current imaging modalities are relatively sensitive for detecting advanced disease; however, their ability to accurately predict pathological T stage preoperatively remains inadequate.

Notably, although pT2 disease is difficult to predict accurately, cT2 was the most frequent clinical stage in this cohort. This finding indicates that, in the absence of definitive diagnostic criteria, intermediate-stage tumors that cannot be clearly classified as either early-stage or advanced disease are often categorized as cT2, reflecting the diagnostic uncertainty inherent in this classification. Consequently, the cT2 category includes a heterogeneous population encompassing both organ-confined (≤ pT2) and advanced (≥ pT3) tumors, potentially leading to inappropriate treatment selection and variable clinical outcomes.

The clinical value of LND in UTUC has primarily been evaluated in the context of high-risk disease¹¹. The European Association of Urology (EAU) guideline recommends template-based LND for high-risk UTUC—defined as high-grade tumors and/or radiographically visible invasive disease (cT2–T4a)—as this approach may provide both staging accuracy and potential therapeutic benefit with minimal additional morbidity⁹. LND was performed in approximately 40% of cases regardless of clinical T stage, and no clear association with survival outcomes was observed, including among patients with pT3 disease. However, this finding should be interpreted with caution, as the cohort was restricted to clinically node-negative (cN0) patients and detailed information regarding the extent of LND and the number of lymph nodes removed was unavailable. Consequently, the present analysis cannot distinguish between a true lack of therapeutic benefit and heterogeneity in surgical practice across centers. Collectively, these findings highlight the need to refine selection criteria for LND and establish evidence-based guidelines to optimize its use in managing UTUC.

NAC is recommended for ≥pT2 disease in muscle-invasive bladder cancer, and increasing evidence suggests similar benefits in high-risk UTUC, despite the lack of level 1 evidence^12–14. However, these studies defined eligibility based on risk stratification rather than clinical T staging, further emphasizing the limited diagnostic accuracy of UTUC. Although patients with high-risk UTUC—particularly those pathologically diagnosed as ≥pT3 and associated with poor prognosis—may benefit substantially from NAC, careful consideration is required when selecting NAC for cases clinically diagnosed as cT2, as this category often includes pathologically lower-stage tumors such as pT1.

Consistent with previous studies, pT3 disease was associated with poor prognosis, and due to the inaccuracy of preoperative staging, this group comprised a wide range of clinical stages^7,15. Adjuvant chemotherapy has been demonstrated to improve outcomes in patients with ≥pT2 or pN+ UTUC in randomized trials such as POUT, and the present data similarly indicated a survival benefit among patients with pT3 disease who received postoperative systemic therapy¹⁶. Nevertheless, real-world administration of adjuvant therapy remains limited due to patient-selected factors such as advanced age, impaired renal function, and frailty. In this cohort, only approximately 36% of patients with pT3 disease received adjuvant treatment. Therefore, accurate postoperative risk stratification is essential to identify individuals who would benefit most from adjuvant therapy. Recent prognostic models, including the JIKEI-YAYOI score, have emphasized the significance of pathological parameters such as pT stage, nodal involvement, tumor grade, and LVI for postoperative risk assessment¹⁷. The present findings similarly identified LVI as an independent adverse prognostic factor, highlighting the critical role of pathological evaluation in optimizing postoperative management and determining candidates for adjuvant therapy. Collectively, these findings underscore that limitations in preoperative staging—particularly the diagnostic uncertainty associated with cT2 disease—necessitate individualized treatment strategies for UTUC rather than uniform management. Future integration of advanced imaging, molecular profiling, and pathological biomarkers may enhance preoperative risk assessment and guide optimal perioperative treatment planning^18,19.

This study had several limitations. Firstly, although the multicenter design enabled a large sample size, the retrospective nature of the analysis may have introduced selection bias. Secondly, the absence of standardized imaging protocols across participating institutions could have affected the accuracy of clinical T stage assessment. In addition, clinical T staging was determined according to routine practice at each institution without centralized imaging review, which may have resulted in inter-institutional variability, particularly for cT2 disease. Furthermore, information regarding the time interval between perioperative imaging and RNU was unavailable, precluding assessment of potential disease progression during the waiting period. Such progression has been reported to be associated with pathological upstaging and aggressive disease features in UTUC²⁰. Third, the cohort was restricted to clinically cN0 patients, and detailed information regarding the extent of LND and lymph node yield was unavailable, which may have limited the ability to detect a prognostic benefit of LND. Finally, information distinguishing renal pelvic tumors from ureteral tumors was unavailable, which may have influenced staging accuracy and treatment selection. Despite these limitations, the study provides valuable real-world evidence regarding the discrepancies between clinical and pathological staging and their clinical implications in UTUC.

In conclusion, this study demonstrated significant real-world discrepancies between clinical and pathological staging in UTUC and their effects on treatment outcomes. cT2 disease represents a heterogeneous and diagnostically uncertain category, contributing to variability in treatment strategies and prognoses. Careful interpretation of clinical staging, particularly in cT2 disease, remains essential for accurate risk assessment and optimal management planning. Further studies combining imaging, pathological, and molecular data are needed to clarify the causes of staging inaccuracy and improve preoperative assessment.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Concept and design: Masanobu Shiga, Shuya Kandori, Shingo HatakeyamaData collection: Shuya Kandori, Shingo Hatakeyama, Naoki Fujita, Mizuki Kobayashi, Daiki Ikarashi, Hiroki Fukuhara, Takuma Sato, Shingo Myoen, Motohide Uemura, Takamitsu Inoue, Masaaki Oikawa, Yasuhiro Kaiho, Jun Miyazaki, Yoshiyuki Kojima, Hisanobu Adachi, Akihiro Ito, Norihiko Tsuchiya, Wataru Obara, Tomonori Habuchi, Chikara Ohyama, Hiroyuki NishiyamaData analysis and interpretation: Masanobu ShigaManuscript drafting: Masanobu Shiga, Shuya KandoriCritical revision and supervision: Shingo Hatakeyama, Hiroyuki NishiyamaAll authors reviewed and approved the final version of the manuscript.

Funding

This study was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI grants 25K20057 (M.S.) and 23K24469 (H.N.).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

Competing interestsShingo Hatakeyama received honoraria from Janssen Pharmaceutical K.K., Astellas Pharma Inc., AstraZeneca K.K., Ono Pharmaceutical Co., Ltd., Bayer AG, Pfizer Inc., Bristol-Myers Squibb, Merck Biopharma Co., Ltd., Kaneka Corporation, and Nipro Corporation. Hiroyuki Nishiyama received honoraria from Astellas, BMS, Janssen, MSD, Ono, and Pfizer, and a grant from Chugai and Astellas. The other authors have no conflicts of interest to declare. Tomonori Habuchi received honoraria from Janssen Pharmaceutical K.K., Nippon Kayaku Co., LTD., Takeda Pharmaceutical Company Ltd., Astellas Pharma Inc., Daiichi Sankyo Company, Ltd., AstraZeneca K.K., Sanofi S.A., Novartis Pharmaceuticals and Bayer AG. Tomonori Habuchi also received research funding supports from Mochida Pharmaceutical Co., Novartis Pharmaceuticals Co, LTD., Pharmaceuticals Co, LTD. and Sysmex Co. Wataru Obara received honoraria from Astellas Pharma Inc., AstraZeneca K.K., Ono Pharmaceutical Co., Ltd., Bristol-Myers Squibb, Merck Biopharma Co., Ltd., Takeda Pharmaceutical Company Ltd., Merck Biopharma. Norihiko Tsuchiya received honoraria from Janssen Pharmaceutical K.K., Pfizer Inc., Takeda Pharmaceutical Co. Ltd., Astellas Pharma Inc., Ono Pharmaceuticals Co., Ltd., MSD K.K., Merk & Co., Inc. Other authors have no conflicts of interest to declare.

Ethical approval

This retrospective, multi-institutional study was approved by the Institutional Review Boards or Ethics Committees of all participating centers, based on the central approval granted by the Ethics Committee of the Hirosaki University School of Medicine (approval number: 2023–063–1), which served as the central ethics committee for this study. The requirement for informed consent was waived by the Ethics Committee of the Hirosaki University School of Medicine due to the retrospective nature of the study, and information was disclosed to patients using an opt-out approach. All methods were performed in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki and institutional policies for human research.

Consent to participate

The need for informed consent to participate was waived by the Institutional Review Board of the University of Tsukuba Hospital and all participating hospitals. Written informed consent was not obtained from individual patients due to the use of an opt-out approach.

Consent to publish

All authors approved for the publication.