Divergent Differentiation and Histologic Subtypes in Upper Tract Urothelial Carcinoma Demonstrate Distinct Patterns of Extraurothelial Recurrence

Take Home Message

Divergent differentiation and histologic subtypes in upper tract urothelial carcinoma is associated with significantly higher extraurothelial recurrence rates despite similar urothelial recurrence rates. These findings support risk-adapted surveillance strategies and early multidisciplinary management in this high-risk subgroup.

Abstract

Background and objective

Upper tract urothelial carcinoma (UTUC) shows lower urinary tract recurrence after nephroureterectomy, with rates of 22–47%. We aimed to evaluate recurrence patterns and survival outcomes in UTUC patients with divergent differentiation and histologic subtypes (DDHSs) compared with pure urothelial carcinoma (UC), with a focus on differentiating urothelial versus extraurothelial recurrences.

Methods

A retrospective review was conducted of 398 patients who underwent nephroureterectomy for UTUC at a single center (2014–2024). Clinical, pathologic, and treatment data were collected. Patients were stratified by histology (pure UC vs DDHSs). The primary outcomes include urothelial (U-RFS) and extraurothelial (EU-RFS) recurrence-free survival. Multivariable Cox proportional hazards regression was used to evaluate the predictors of recurrence-free survival. A Kaplan-Meier analysis was used to estimate recurrence-free and overall survival.

Key findings and limitations

Among 398 patients, 14% had DDHSs. Patients with DDHSs received adjuvant therapy more frequently (33% vs 13%) and presented with more advanced pathologic stage and nodal involvement at nephroureterectomy than those with pure UC. Overall recurrence rates were similar between groups. The Kaplan-Meier analysis demonstrated significantly worse EU-RFS in patients with DDHSs (p < 0.001), while U-RFS did not differ. On a multivariable Cox analysis, DDHSs (hazard ratio [HR] 3.18, p = 0.01) and advanced pathologic T stage (HR 4.64, p = 0.047) were independently associated with extraurothelial recurrence. Overall survival was similar between groups.

Conclusions and clinical implications

Patients with DDHS UTUC exhibit higher rates of extraurothelial recurrence despite similar rates of urothelial recurrence. These findings underscore the need for intensified surveillance and early systemic therapy consideration in DDHS UTUC patients.

Patient summary

We studied patients with upper tract urothelial cancer who had different tumor types. We found that patients with these specific rare tumor types were more likely to have cancer spread outside the urinary tract. These patients may benefit from closer follow-up and additional treatment after surgery.

1. Introduction

Upper tract urothelial carcinoma (UTUC) is a rare malignancy, comprising only 5–10% of all urothelial cancers, with an incidence rate of approximately two per 100 000 in Western countries [1]. Despite its rarity, UTUC imposes a significant health care burden, with 5-yr post-nephroureterectomy costs in the USA ranging from $41 474 to $385 464 per patient [2].

Recurrence, particularly in the lower urinary tract, is common, occurring in 22–47% of patients after nephroureterectomy.

Up to 25% of UTUC patients exhibit divergent differentiation and histologic subtypes (DDHSs), including squamous, glandular, neuroendocrine, or micropapillary and sarcomatoid, similar to patterns seen in bladder cancer [3]. DDHSs are associated with high-grade disease and worse recurrence-free survival (RFS) and cancer-specific survival compared with pure urothelial carcinoma [4], [5]. Prior studies suggest higher recurrence rates in patients with DDHSs following nephroureterectomy, which is the standard treatment for high-risk UTUC [6]. However, many focus solely on surgical bed, nodal, or distant recurrences, omitting the involvement of the lower urinary tract [7], [8].

To address this gap, we aimed to characterize recurrence patterns in DDHS UTUC patients at our quaternary academic center, assess the impact of tumor features on recurrence, and examine histologic concordance between primary and recurrent tumors. We hypothesized that DDHSs would be associated with higher recurrence rates and worse oncologic outcomes.

2. Patients and methods

2.1. Cohort selection

We conducted a single-institution retrospective review of all patients with UTUC who underwent nephroureterectomy. Institutional review board approval was obtained (IRB #16-823). Patients were identified through a query of our institutional Laboratory Information System for surgical specimens labeled as “kidney,” “ureter,” or “carcinoma” between 2014 and 2024, yielding an initial cohort of 2107 patients (Fig. 1). After screening, 616 patients were confirmed to have undergone radical nephroureterectomy. Of these patients, 218 were excluded for the following reasons: missing key clinical data (n = 7), primary bladder cancer with extension to the ureter (n = 3), metastatic disease at the time of surgery (n = 2), nonurothelial histology (n = 1), or a personal history of bladder cancer (n = 205). The final study cohort included 398 patients with either pure UTUC or urothelial carcinoma with DDHSs (previously referred to in the literature as variant histology).

Fig. 1.

Cohort selection flowchart.

2.2. Study variables and outcomes of interest

Baseline demographic, clinical, and pathologic variables were collected for all the patients. Patient characteristics included age, sex, race, smoking history, body mass index, personal or family history of bladder cancer or UTUC, known Lynch syndrome, renal function at diagnosis, and Eastern Cooperative Oncology Group performance status, when available. Diagnostic data included index tumor location, histologic grade, presence of DDHSs (as determined by biopsy or cytology), and tumor multifocality. Treatment variables included operative details and receipt of neoadjuvant or adjuvant systemic therapy, and perioperative intravesical therapy.

All available slides for urothelial carcinoma with DDHSs were retrieved and reviewed by a dedicated genitourinary pathologist (R.A.) to confirm original diagnoses along with other pertinent pathologic information. Pathologic outcomes of interest included TNM staging, grade, presence, and percentage of DDHSs (Supplementary Table 1), carcinoma in situ, lymphovascular invasion, and surgical margin status. The collected oncologic outcomes included urothelial recurrences (within 2 yr of surgery), all extraurothelial recurrences, and survival status.

The primary outcomes were urothelial recurrence-free survival (U-RFS) and extraurothelial recurrence-free survival (EU-RFS), stratified by the presence of DDHSs. The secondary outcomes included overall survival (OS) and a histologic comparison between primary tumors and recurrence specimens.

Urothelial recurrence was defined as the first documented recurrence within the urinary tract, including the bladder, urethra, or contralateral upper tract. Extraurothelial recurrences were defined as the first recurrence outside the urinary tract, including the renal fossa or surgical bed, regional lymphadenopathy (retroperitoneal or pelvic nodes), and distant lymphadenopathy or other visceral or osseous metastatic disease. Recurrence was determined based on histologic confirmation when available or radiographic evidence consistent with metastatic disease.

2.3. Statistical analysis

The primary study questions were whether DDHSs are associated with U-RFS, EU-RFS, and OS following nephroureterectomy for UTUC.

To address these questions, time-to-event outcomes were analyzed using Cox proportional hazards regression, with histologic subtype (DDHSs vs pure urothelial carcinoma) as the primary predictor. Multivariable models were prespecified and adjusted for age, sex, receipt of adjuvant therapy, and pathologic T stage (≥2 vs <2). Given the limited number of extraurothelial recurrence events, the EU-RFS model was restricted to clinically relevant covariates to reduce overfitting.

Kaplan-Meier methods were used to estimate U-RFS, EU-RFS, and OS, with comparisons between histologic groups evaluated using the log-rank test. Patients without events were censored at the dates of the last disease assessment for recurrence outcomes and the last known follow-up for OS.

As a secondary outcome, histologic concordance between primary tumors and recurrence specimens was evaluated. Histologic transitions across primary tumors and recurrences were visualized using a Sankey plot depicting progression from the index nephroureterectomy specimen to subsequent recurrence specimens, for which data were available.

Descriptive statistics were summarized using medians and interquartile ranges for continuous variables and counts with percentages for categorical variables. All analyses were conducted using R version 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria; http: //www.r-project.org), and a two-sided p value of <0.05 was considered statistically significant.

3. Results

3.1. Cohort assembly and baseline characteristics

After a thorough review of our institutional database, 398 patients were found to have undergone nephroureterectomy for UTUC with complete data elements. In this group, 55 patients (14%) were identified as having DDHSs (Fig. 1). The median follow-up for all patients who remain alive and recurrence free within the entire cohort was 24 mo.

The median age at diagnosis was 73 yr, and the cohort included predominantly non-Hispanic White and male patients. Patients with DDHSs were more likely to receive adjuvant therapy (33% vs 13%) compared with those with pure urothelial carcinoma. They also presented with more locally advanced disease, including higher rates of pT3 (56% vs 18%) and pT4 (13% vs 4.7%) stage, and increased nodal involvement (N1 = 5.5% vs 1.2%; N2 = 7.3% vs 2.9%) at the time of nephroureterectomy (Table 1).

Table 1.

Baseline characteristics and pathologic outcomes for patients who underwent nephroureterectomy for upper tract urothelial carcinoma

	Pure urothelial	DDHSs
N (%)	343 (86)	55 (14)
Age, median (IQR)	73 (66–80)	73 (66–80)
Sex, n (%)
Female	128 (37)	24 (44)
Race/ethnicity, n (%)
Asian	3 (0.87)	2 (3.6)
Black	19 (5.5)	3 (5.5)
Hispanic	10 (2.9)	2 (3.6)
Non-Hispanic White	294 (86)	48 (87)
Other	3 (0.87)	0 (0.0)
Declined to answer	13 (3.8)	0 (0.0)
Smoking status, n (%)
Active smoker	47 (14)	8 (15)
Former smoker	196 (57)	26 (47)
Never smoker	100 (29)	21 (38)
Body mass index, median (IQR)	28 (24–32)	28 (23–33)
Neoadjuvant chemotherapy, n (%)	30 (8.7)	6 (11)
Adjuvant therapy, n (%)	45 (13)	18 (33)
Pathologic T stage, n (%)
pTa	113 (33)	4 (7.3)
pTis	17 (5.0)	0 (0.0)
pT1	84 (25)	9 (16)
pT2	49 (14)	4 (7.3)
pT3	63 (18)	31 (56)
pT4	16 (4.7)	7 (13)
Pathologic N stage, n (%)
pN0	74 (22)	15 (27)
pN1	4 (1.2)	3 (5.5)
pN2	10 (2.9)	4 (7.3)
pNX	252 (74)	33 (60)
Median follow-up a (mo)	26	14

DDHSs = divergent differentiation and histologic subtypes; IQR = interquartile range.

^aMedian follow-up duration refers to patients who are event free (without recurrences or death).

Among cases with DDHSs, squamous differentiation (58%) was most common, followed by micropapillary (11%) and sarcomatoid subtypes (9.1%; Supplementary Table 1).

3.2. Overall, urothelial, and extraurothelial recurrences

In this cohort of 398 postnephroureterectomy patients, recurrences were identified in 130 patients. There was no significant difference in overall recurrence rates between patients with DDHSs and those with pure urothelial carcinoma.

Of 130 patients who recurred, 27% (108/398) recurred within the urothelium, while only 7% (28/398) recurred outside of the urinary tract. Notably, urothelial recurrence appeared more common among patients with pure urothelial carcinoma (29% vs 15%), whereas patients with DDHSs were more likely to experience extraurothelial recurrence (22% vs 4.7%; Table 2).

Table 2.

Oncologic outcomes for patients who underwent nephroureterectomy for upper tract urothelial carcinoma

	Pure urothelial	DDHSs
Total number of patients	343	55
Any recurrence	113	17
Urothelial recurrences	100	8
Extraurothelial recurrences	16	12
Surgical bed	2	4
Regional nodes	2	1
Distant (lymph node or other)	14	8
Overall survival
Deceased	121	23
Median overall survival (mo)	74.3	58.5

DDHSs = divergent differentiation and histologic subtypes.

Regional nodal recurrences refer to retroperitoneal or pelvic adenopathy. Distant recurrences refer to nonregional lymph node or distant visceral or osseous metastatic disease.

On multivariable Cox regression, DDHSs were independently associated with a more than threefold higher rate of extraurothelial recurrence (hazard ratio [HR] 3.18, 95% confidence interval [CI] 1.29–7.82, p = 0.01), as did pT ≥2 (HR 4.64, 95% CI 1.02–21.06, p = 0.047; Table 3). These findings were supported by Kaplan-Meier analyses, which showed significantly worse EU-RFS in the DDHS group (p < 0.001; Fig. 2A and 2B).

Table 3.

Multivariable Cox regression analysis of clinical and pathologic predictors for urothelial and extraurothelial recurrence-free survival

	Urothelial recurrence			Extraurothelial recurrence a
	HR	95% CI	p value	HR	95% CI	p value
Age (per year)	1.02	1.00–1.04	0.1
Sex (male)	0.67	0.40–1.12	0.1
Adjuvant therapy	0.81	0.43–1.55	0.5	1.53	0.63–3.74	0.4
Pathologic T stage ≥2	1.01	0.62–1.64	0.9	4.64	1.02–21.06	0.047
DDHS	0.84	0.35–2.04	0.7	3.18	1.29–7.82	0.01

CI = confidence interval; DDHS = divergent differentiation and histologic subtypes; HR = hazard ratio.

^aThe extraurothelial recurrence model included fewer covariates due to limited event counts.

Fig. 2.

Kaplan-Meier and cumulative incidence curves for (A and B) extraurothelial and (C and D) urothelial recurrences, stratified by the presence of divergent differentiation and histologic subtypes. The blue line represents pure urothelial carcinoma, and the red line represents divergent differentiation or histologic subtypes, with 95% confidence intervals illustrated with the corresponding shaded areas. GU = genitourinary.

In contrast, DDHSs were not significantly associated with urothelial recurrence on a multivariable analysis (HR 0.84, 95% CI 0.35–2.04, p = 0.7). Kaplan-Meier estimates similarly showed no difference in U-RFS when stratified by histology (p = 0.2; Fig. 2C and 2D).

Among histologic subtypes, the median RFS (in months) was shortest for plasmacytoid (4.4 mo) and neuroendocrine (6.5 mo) subtypes, while glandular and squamous histology had relatively longer RFS of 14.7 and 12.7 mo, respectively (Supplementary Table 2).

3.3. Overall survival

A Kaplan-Meier analysis demonstrated no significant difference in OS by histologic subtype, with no association on Cox regression (HR 1.33, 95% CI 0.85–2.08, p = 0.2; Fig. 3).

Fig. 3.

Kaplan-Meier curves for overall survival, stratified by the presence of divergent differentiation and histologic subtypes. The blue line represents pure urothelial carcinoma, and the red line represents divergent differentiation or histologic subtypes, with 95% confidence intervals illustrated with the corresponding shaded areas.

3.4. Histologic progression across recurrences

Histologic changes across successive urothelial recurrences were also examined as part of our study for both urothelial and extraurothelial recurrences. Among patients whose pathology on index nephroureterectomy was pure urothelial carcinoma, most retained pure urothelial histology across multiple urothelial recurrences; however, one patient in this group was observed to transition to squamous differentiation (Fig. 4A).

Fig. 4.

Sankey plot for histology of (A) urothelial and (B) extraurothelial recurrences, as compared with primary histology.

When extraurothelial recurrences were examined, however, the group that was initially diagnosed with DDHSs showed much more histologic variation between recurrences, with some patients showing persistent divergent features, while others from the squamous, micropapillary, and other variant groups transitioned to pure urothelial histology on subsequent recurrences (Fig. 4B).

3.5. Patterns of extraurothelial recurrences

In light of these findings, sites of extraurothelial recurrences were also analyzed for the DDHS group. The most common sites of recurrence were the lung (n = 10), bone (n = 9), and liver (n = 7). Other metastatic sites included the renal fossa (n = 6), retroperitoneal/pelvic nodes (n = 3), distant nodes (n = 3), and miscellaneous other sites including small bowel, and peritoneal carcinomatosis (n = 3) was also observed (Fig. 5).

Fig. 5.

Schema demonstrating sites of extraurothelial recurrences for cases with divergent differentiation and histologic subtypes.

4. Discussion

In this study, we analyzed a cohort of 398 patients who underwent radical nephroureterectomy for UTUC, of whom 14% exhibited DDHSs. This proportion aligns with the findings of prior studies, which have reported DDHS rates ranging from 7% to 13% [4], [5], [9], [10]. Notably, a recent Taiwanese study involving 1239 patients reported a substantially higher rate of incidence of DDHSs (31%) [11]. The reasons for this discrepancy remain uncertain, although it may be attributable to the widespread exposure to aristolochic acid prior to 2003—a recognized nephrotoxin and carcinogen—potentially contributing to the elevated incidence of UTUC in that geographic region [12].

Our findings indicate that while U-RFS did not differ significantly between patients with pure urothelial carcinoma and those with DDHSs, EU-RFS was both clinically and statistically worse in the DDHS group. Although this disparity did not extend to OS within our cohort, the results underscore important considerations in the management of DDHS UTUC and support our hypothesis that UTUC with DDHSs represents a biologically more aggressive disease entity.

Poor oncologic outcomes in patients with DDHS UTUC are well documented in the literature. In an institutional series of 47 patients with DDHSs, Nogueira et al [5] reported significantly worse OS (HR 1.74), although no difference in RFS was observed. A separate meta-analysis identified DDHSs as a predictor associated with inferior outcomes across multiple endpoints, including RFS (HR 1.48), cancer-specific survival (HR 1.86), and OS (HR 1.74) [13]. Most of the existing data are derived from retrospective studies with small numbers of DDHS patients, limiting the ability to analyze the prognostic significance of individual histologic subtypes. Nonetheless, one multi-institutional study found micropapillary and sarcomatoid variants to be independently associated with worse outcomes [10]. In our cohort, plasmacytoid and neuroendocrine subtypes showed worse RFS, although low event counts precluded a multivariable analysis.

While our data lacked granularity to assess individual subtypes, we introduce the concept of extraurothelial recurrence, encompassing both locoregional and distant disease. This distinction—between urothelial and extraurothelial recurrences—is meaningful, as their management strategies differ substantially. Urothelial recurrences are often addressed with local therapies, such as surgery or intravesical instillation, whereas extraurothelial recurrences typically require systemic therapy and are rarely amenable to local intervention [6], [14]. Given the differing management and morbidity, we advocate for separate reporting of these recurrence types.

Over the past decade, two pivotal developments have significantly influenced the treatment landscape of UTUC. First, the publication of the POUT trial provided level I evidence supporting the use of adjuvant chemotherapy in select patients with UTUC. Second, the approval of PD-1 inhibitors as adjuvant therapy for locally advanced or metastatic urothelial carcinoma has expanded systemic treatment options, with many studies including patients with UTUC [15], [16], [17].

Although PD-1/PD-L1 inhibitors have demonstrated comparable efficacy between bladder and upper tract tumors in the metastatic setting, their benefit in the adjuvant setting for UTUC remains uncertain [16], [18], [19], [20]. Most studies reporting on adjuvant immunotherapy outcomes for UTUC rely on post hoc or subgroup analyses, which are not adequately powered to determine UTUC-specific efficacy. As such, these findings should be interpreted with caution. Further prospective, UTUC-focused studies are needed to clarify the role of adjuvant or even neoadjuvant immunotherapy in this distinct patient population.

Several studies have evaluated the impact of DDHSs on outcomes in patients with UTUC treated with adjuvant chemotherapy. In an early report by Xylinas et al [21], no significant difference in RFS or cancer-specific survival was observed between patients with DDHSs and those with pure urothelial carcinoma when treated in the adjuvant setting. More recently, Lo et al [22] reported that adjuvant chemotherapy was associated with improved cancer-specific and disease-free survival in patients with DDHS UTUC. In a separate multi-institutional study, Chung et al [23] found that DDHSs were associated with significantly worse RFS, cancer-specific survival, and OS. However, these differences became nonsignificant after adjusting for the receipt of adjuvant chemotherapy, highlighting its potential therapeutic impact in this population.

Taken together with our finding of earlier extraurothelial recurrences in patients with DDHSs, the data support a more intensive approach to postoperative surveillance in these patients. Specifically, we recommend (1) increased vigilance during surveillance, including more frequent cross-sectional imaging to facilitate early detection of extraurothelial recurrences, and (2) early referral to medical oncology to initiate multidisciplinary discussions regarding the role of adjuvant therapy in this high-risk cohort.

Interestingly, upon examination of the histology of both urothelial and extraurothelial recurrences, the majority were found to be pure urothelial carcinomas, despite the presence of DDHSs in the primary tumor. This observation may reflect the clonal evolution seen commonly in urothelial carcinoma, or the intratumoral heterogeneity often characteristic of tumors with DDHSs. Moreover, phenotypic plasticity has been described in bladder cancer, with some tumors transitioning from a luminal to a basal subtype over time, while others exhibit early expression of basal markers as a precursor of squamous differentiation [24], [25]. The precise mechanisms driving these histologic shifts in our cohort remain unclear. Whether these changes are driven by systemic therapy utilization, intrinsic tumor biology, or sampling limitations is unknown and warrants further investigation.

This study has several limitations. Its retrospective design introduces a potential selection bias, as treatment and surveillance were not standardized. The rarity of UTUC and limited sample sizes for individual subtypes constrain the power of subgroup analyses, particularly for recurrence patterns. Additionally, while individual subtypes (eg, squamous, micropapillary, and sarcomatoid) may exhibit distinct biological behaviors and prognoses, these were grouped together under a single DDHS category in our analysis due to limited sample sizes, which may have attenuated subtype-specific outcome differences. Furthermore, in alignment with the 2022 WHO classification, the nested variant has been subdivided into “large” and “small” nested subtypes, and neuroendocrine differentiation may further be classified into large-cell neuroendocrine and small-cell carcinoma; however, these were analyzed as single categories in our dataset, which may limit the granularity of our histologic assessment. Our cohort excluded patients managed with nephron-sparing approaches, including endoscopic ablation and distal ureterectomy, to ensure adequate tissue for a pathologic review and minimize bias in oncologic outcomes. Differences in systemic therapy use between groups may have introduced treatment-related confounding, as detailed treatment data were not uniformly available. Nonetheless, the relatively large, contemporary cohort and extended follow-up—longer than those reported in most Western series—offer valuable real-world insights into this uncommon disease.

5. Conclusions

In this study, we found that patients with DDHSs had similar rates of urothelial recurrence to those with pure UTUC, but significantly higher rates of extraurothelial recurrence. These findings highlight the need for intensified surveillance and early multidisciplinary engagement in this high-risk population to enable prompt detection and treatment of aggressive recurrences. Future studies investigating the molecular characteristics of divergent histologic subtypes may further clarify the mechanisms driving their aggressive behavior and help identify novel therapeutic targets.



Author contributions: Betty Wang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.



Study concept and design: Wang, Davis, Bukavina.

Acquisition of data: Wang, Nandwana, Siva, Maina, Alaghehbandan, Haywood, S. Campbell, Abouassaly, Olivares, Berglund, Stein, Kaouk.

Analysis and interpretation of data: Wang, Nandwana, Bukavina.

Drafting of the manuscript: Wang, Davis, Nandwana.

Critical revision of the manuscript for important intellectual content: Weight, R. Campbell, Alaghehbandan, Bukavina.

Statistical analysis: Sindhani, Bukavina.

Obtaining funding: None.

Administrative, technical, or material support: None.

Supervision: Alaghehbandan, Bukavina.

Other: None.



Financial disclosures: Betty Wang certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.



Funding/Support and role of the sponsor: None.

Associate Editor: Maria Carmen Mir

Appendix A. Supplementary data

The following are the Supplementary data to this article: