Impact of concomitant carcinoma in situ distribution on non‐muscle‐invasive bladder cancer progression risk

Abstract

Objective

To assess whether the distribution of concomitant carcinoma in situ (CIS; unifocal or multifocal) with papillary non‐muscle‐invasive bladder cancer (NMIBC) impacts the risk of progression, as concomitant CIS is an established risk factor for progression in papillary NMIBC and commonly used calculators do not make this distinction.

Patients and Methods

In this multi‐institutional retrospective cohort study from both academic and community hospitals, clinicopathological data were collected from patients with pTa/pT1 NMIBC treated from 2005 to 2022. Unifocal concomitant CIS was defined as CIS present in only one specimen (i.e., papillary disease with CIS at the tumour base or isolated CIS in one specimen). Multifocal concomitant CIS was characterised by CIS in multiple specimens. Progression was defined as the development of muscle‐invasive or metastatic disease. Fine‐Gray regression was performed to identify progression‐associated factors, using all‐cause mortality as a competing risk.

Results

Among 2923 patients, 383 (13%) progressed over a median (interquartile range) follow‐up of 5.1 (3.0–8.5) years. Concomitant CIS was found in 327 patients (11%), with 233 and 94 harbouring unifocal and multifocal CIS, respectively. Recurrent tumours, T1 stage, high‐grade disease, multifocal CIS, and multiple tumours were independently associated with increased progression risk (all P < 0.05). Among patients with concomitant CIS, multifocal CIS remained a significant prognosticator (sub‐distribution hazard ratio 1.90, 95% confidence interval 1.18–3.05; P = 0.008) adjusting for age, sex, tumour history, stage, grade, number of tumours, tumour diameter, and Bacillus Calmette–Guérin treatment.

Conclusions

Papillary NMIBC progression risk varies with concomitant CIS distribution. Only multifocal concomitant CIS is a risk factor for progression in patients with T1 NMIBC. If validated in further studies, risk calculators should consider including this CIS distinction. Submitting separate specimens at the time of transurethral resection, as recommended by guidelines, should be encouraged.

Abbreviations

CIS

carcinoma in situ

CUA

Canadian Urological Association

EAU

European Association of Urology

HG

high grade

IQR

interquartile range

(N)MIBC

(non‐)muscle‐invasive bladder cancer

MUC1

mucin 1

SHR

sub‐distribution hazard ratio

TURBT

transurethral resection of the bladder tumour

Introduction

Carcinoma in situ (CIS) is an aggressive form of non‐muscle‐invasive bladder cancer (NMIBC) characterised by a flat, high‐grade (HG) tumour confined to the bladder urothelium that can appear as an erythematous, velvet‐like area on the bladder. CIS is associated with an increased risk of cancer progression and worse cancer‐specific outcomes [1, 2, 3, 4]. When left untreated, the natural history of CIS poses a risk of progression to muscle‐invasive bladder cancer (MIBC) in up to 50% of cases at 5 years [4]. Due to its aggressive nature, guidelines classify the presence of any CIS in NMIBC as high‐risk disease [5, 6, 7, 8]. The current standard of care for high‐risk NMIBC includes transurethral resection of the bladder tumour (TURBT), followed by intravesical administration of BCG to reduce the risk of recurrence and progression [5, 6, 7]. BCG induction involves an initial 6‐week course, followed by maintenance instillations completed over 3 years for high‐risk NMIBC [9].

Carcinoma in situ can be further classified into: (i) primary CIS, which presents with no previous or concomitant papillary tumour; (ii) secondary CIS, which is detected during follow‐up of patients with prior NMIBC; and (iii) concomitant CIS, which is diagnosed concurrently with papillary tumour [10]. Concomitant CIS is present in ~10% of patients with NMIBC and has consistently been shown to be an independent adverse prognosticator for tumour progression to MIBC [11, 12]. As such, it has been incorporated into all guideline‐endorsed NMIBC prognostic risk calculators, including the most recent European Association of Urology (EAU) NMIBC risk calculator, Spanish Urological Club for Oncological Treatment (CUETO) scoring model, and the previous European Organisation for Research and Treatment of Cancer (EORTC) risk tables among others [13, 14, 15].

Carcinoma in situ lesions can also be stratified by distribution pattern into either unifocal or multifocal [1]. However, current risk calculators do not distinguish between these two entities, only differentiating between the presence or absence of CIS concomitantly with papillary tumours [13, 14, 15]. The prognostic implications of multifocal compared to unifocal concomitant CIS remain unclear. Therefore, this study sought to assess whether concomitant CIS distribution (unifocal or multifocal) impacts progression risk when concurrently present in patients with papillary Ta/T1 NMIBC.

Methods

Study Population

Following institutional approval, a retrospective cohort study was conducted on patients with NMIBC treated between 1 January 2005 and 30 June 2022 at four Canadian academic or community hospitals, including Princess Margaret Cancer Centre, University Health Network (Toronto, Ontario), Mount Sinai Hospital, Sinai Health System (Toronto, Ontario), Credit Valley Hospital (Mississauga, Ontario), and Mississauga Hospital (Mississauga, Ontario). Ethics approval was granted prior to the study data collection (University Health Network and Sinai Health System #22‐5280; Credit Valley Hospital and Mississauga Hospital #1140). Patients were included regardless of their prior NMIBC history and received BCG or other intravesical agents as per standard of care [6]. As per guidelines, re‐staging TURBT was performed in patients with T1 NMIBC, when a complete resection was not achieved with the first TURBT, when no detrusor muscle was present in the specimen, or in select cases of HG Ta tumours (e.g., large and/or multiple tumours) [5]. CIS was identified based on histological confirmation of macroscopically suspicious areas during TURBT. Systematic bladder mapping with biopsies of visually normal urothelium was not routinely performed [16]. In cases where CIS was not present on the initial TURBT but found on re‐TURBT, the pathology was incorporated into the primary pathology at initial TRUBT, as it would have been extremely unlikely that CIS would have appeared in the short period of time between the index and repeat TURBT. Adequate BCG administration was defined as the administration of at least five out of six doses of the initial induction course and two out of three doses of maintenance therapy or two out of six doses of a second induction course over a minimum period of 6 months. Exclusion criteria included any NMIBC treatment prior to 1 January 2005 or missing pathological information such as grade and/or stage. The follow‐up strategy, including cystoscopy and upper tract surveillance, was based on the Canadian Urological Association (CUA) guidelines [5].

Outcome Definition

The primary outcome was tumour progression, defined as the time from the date of index TURBT to first development of muscle‐invasive (Stage ≥T2 either at subsequent TURBT or radical cystectomy), nodal, or metastatic disease. The analysis did not include recurrence‐free survival or other survival endpoints, as disease progression was the primary focus of our study, as per the recent EAU risk calculator for progression [13]. Patients who did not experience the event were censored at the date of last follow‐up cystoscopy.

Statistical Analyses

The cumulative incidence of tumour progression after the index TURBT was calculated using the Aalen–Johansen estimator, with all‐cause mortality considered as a competing event. Cumulative incidence of progression was visualised and stratified by stage and distribution of CIS. Differences in cumulative incidence across strata were assessed using the log‐rank test. Fine and Gray competing risk regression was used to estimate the sub‐distribution hazard ratios (SHRs) for progression, accounting for all‐cause mortality as a competing event. Candidate predictors included in uni‐ and multivariable analysis were age (per 10 years), sex, tumour history (primary vs recurrent), pathological stage (Ta vs T1), grade (WHO 2004/2022 classification, low grade vs HG), distribution of CIS (unifocal vs multifocal), number of tumours (single vs multiple), tumour diameter (<3 vs ≥3 cm), and BCG treatment based on findings of index TURBT. Multiple tumours were defined as the presence of more than one papillary and/or CIS lesion. Unifocal CIS was defined as CIS in only one specimen (i.e., papillary disease with CIS at the tumour base or isolated CIS in one specimen). Multifocal CIS was characterised by CIS in multiple specimens.

Statistical analyses were performed using R version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria) and the ‘cmprsk’ version 2.2‐11 package. Plots were generated using Python version 3.9 and the ‘lifelines’ version 0.28.0 package.

Results

Of the 3324 patients treated across the four institutions during the study period, 401 were excluded due to missing information. The baseline characteristics of the final cohort (n = 2923), as well as those stratified by unifocal or multifocal CIS, are summarised in Table 1. In the final cohort, the median (interquartile range [IQR]) age was 71 (62–79) years, with 697 (24%) females. Among the cohort, 942 patients (32%) had T1 disease, while 1487 (51%) had HG disease. Concomitant CIS was found in 327 patients (11%), of which 233 (8%) and 94 (3%) presented with unifocal and multifocal CIS, respectively. In terms of intravesical therapy, 1471 patients (50%) were treated with BCG, while 822 (28%) received adequate BCG. Among patients with concomitant CIS, 281 (86%) and 190 (58%) received any BCG and adequate BCG, respectively. During a median (IQR) follow‐up of 5.1 (3.0–8.5) years, 383 patients developed tumour progression.

Table 1.

Baseline characteristics of the study cohort (n = 2923).

Characteristics	Overall	No CIS	Unifocal CIS	Multifocal CIS
Number of patients	2923	2596	233	94
Age, years, median (IQR)	71 (62–79)	71 (62–79)	72 (63–79)	71 (65–78)
Female, n (%)	697 (24)	644 (25)	42 (18)	11 (12)
Prior history of NMIBC, n (%)	287 (10)	247 (10)	28 (12)	12 (13)
Stage, n (%)
Ta	1981 (68)	1873 (72)	73 (31)	35 (37)
T1	942 (32)	723 (28)	160 (69)	59 (63)
HG, n (%)	1487 (51)	1172 (45)	222 (95)	93 (99)
Concomitant CIS, n (%)	327 (11)	0 (0)	233 (100)	94 (100)
Distribution of CIS, n (%)
Unifocal	233 (8)	0 (0)	233 (100)	0 (0)
Multifocal	94 (3)	0 (0)	0 (0)	94 (100)
Multiple tumours	1171 (40)	955 (37)	131 (56)	85 (90)
Tumour diameter ≥3 cm, n (%)	958 (33)	840 (32)	87 (37)	31 (33)
BCG, n/N (%)
Any BCG	1471 (50)	1189 (46)	204 (88)	78 (83)
Adequate BCG †	822 (28)	635 (25)	133 (57)	54 (57)
Progression	383/2923 (13)	304/2596 (12)	47/233 (20)	32/94 (34)
Follow‐up, years, median (IQR)	5.1 (3.0–8.5)	5.2 (3.0–8.6)	4.5 (2.3–8.7)	3.8 (1.9–6.3)

^† At least five of the six doses of initial induction course AND either at least two of the three doses of maintenance therapy or two of the sex doses of a second induction course.

Progression risk varied when stratified by stage and distribution of CIS (Fig. 1). The 5‐year cumulative incidence of tumour progression was 5% (95% CI 4–7%) for Ta‐only disease, 12% (95% CI 5–22%) for Ta with unifocal concomitant CIS, 12% (95% CI 4–24%) for Ta with multifocal CIS, 27% (95% CI 24–31%) for T1‐only disease, 26% (95% CI 19–33%) for T1 with unifocal concomitant CIS, and 49% (95% CI 34–63%) for T1 with multifocal CIS. Among patients who received adequate BCG, the 5‐year cumulative incidence of progression was 14% (95% CI 8–21%) for unifocal CIS, and 33% (95% CI 19–48%) for multifocal concomitant CIS (log‐rank test P = 0.02, Fig. S1). On subset analysis of stratified patients with T1 NMIBC who received adequate BCG, stratified by CIS distribution, only multifocal, and not unifocal, CIS was associated with increased progression (Fig. S2). On an additional subset analysis restricted to patients with T1 NMIBC who underwent repeat TURBT, the 5‐year progression rates remained unchanged (Fig. S3). On univariable analysis, all predictors except for sex were associated with tumour progression (Table 2). On multivariable analysis, recurrent tumours (SHR 1.83, 95% CI 1.49–2.25; P < 0.001), T1 stage (SHR 2.66, 95% CI 2.05–3.45; P < 0.001), HG disease (SHR 4.39, 95% CI 3.05–6.31; P < 0.001), multifocal CIS (SHR 1.60, 95% CI 1.09–2.33; P = 0.02), and multiple tumours (SHR 1.43, 95% CI 1.15–1.78; P = 0.001) were all significantly associated with tumour progression. In the subset analysis of patients with concomitant CIS, T1 stage (SHR 2.35, 95% CI 1.31–4.21; P = 0.004), multifocal CIS (SHR 1.90, 95% CI 1.18–3.05; P = 0.008), and tumour diameter ≥3 cm (SHR 1.75, 95% CI 1.13–2.73; P = 0.01) remained significant predictors of tumour progression, after adjusting for age, sex, recurrent tumour, grade, multiple tumours, and BCG treatment after index TURBT (Table 3). Multifocal CIS remained an independent prognosticator of progression among patients with concomitant CIS who received adequate BCG (Table S1).

Fig. 1.

Cumulative incidence of tumour progression following index TURBT, stratified by stage and distribution of CIS. All‐cause mortality was considered as a competing risk. The cumulative incidence of tumour progression at 5 years is indicated in parentheses.

Table 2.

Fine and Gray competing risk regression of the full cohort (n = 2923) to estimate SHRs of tumour progression, accounting for all‐cause mortality as a competing risk.

Predictor	Univariable SHR (95% CI)	P	Multivariable SHR (95% CI)	P
Age (per 10 years)	1.16 (1.07–1.25)	<0.001	1.04 (0.96–1.13)	0.4
Sex
Male	Ref	–	Ref	–
Female	0.83 (0.65–1.06)	0.14	1.02 (0.78–1.32)	0.9
Tumour history
Primary	Ref	–	Ref	–
Recurrent	1.63 (1.34–1.98)	<0.001	1.83 (1.49–2.25)	<0.001
Stage
Ta	Ref	–	Ref	–
T1	5.28 (4.26–6.53)	<0.001	2.66 (2.05–3.45)	<0.001
Grade
Low grade	Ref	–	Ref	–
HG	7.05 (5.28–9.42)	<0.001	4.39 (3.05–6.31)	<0.001
Concomitant CIS
No	Ref	–	Excluded, collinear with distribution of CIS
Yes	2.33 (1.82–2.98)	<0.001
Distribution of CIS
None	Ref	–	Ref	–
Unifocal	1.90 (1.39–2.60)	<0.001	0.91 (0.65–1.28)	0.6
Multifocal	3.47 (2.43–4.96)	<0.001	1.60 (1.09–2.33)	0.02
Number of tumours
Single	Ref	–	Ref	–
Multiple	1.84 (1.51–2.25)	<0.001	1.43 (1.15–1.78)	0.001
Tumour diameter
<3 cm	Ref	–	Ref	–
≥3 cm	1.61 (1.31–1.97)	<0.001	1.19 (0.95–1.47)	0.13
Treated with BCG
No	Ref	–	Ref	–
Yes	1.34 (1.09–1.64)	0.005	0.58 (0.47–0.72)	<0.001

Table 3.

Fine and Gray competing risk regression for patients with concomitant CIS (n = 327) to estimate SHRs of tumour progression, accounting for all‐cause mortality as a competing risk.

Predictor	Univariable SHR (95% CI)	P	Multivariable SHR (95% CI)	P
Age (per 10 years)	1.07 (0.87–1.32)	0.5	1.13 (0.92–1.38)	0.2
Sex
Male	Ref	–	Ref	–
Female	0.66 (0.33–1.33)	0.2	0.79 (0.37–1.66)	0.5
Tumour history
Primary	Ref	–	Ref	–
Recurrent	1.11 (0.67–1.84)	0.7	1.24 (0.78–1.98)	0.4
Stage
Ta	Ref	–	Ref	–
T1	2.65 (1.50–4.69)	<0.001	2.35 (1.31–4.21)	0.004
Grade
Low grade	Ref	–	Ref	–
HG	3.18 (0.44–23.1)	0.2	1.73 (0.22–13.8)	0.6
Distribution of CIS
Unifocal	Ref	–	Ref	–
Multifocal	1.81 (1.16–2.83)	0.009	1.90 (1.18–3.05)	0.008
Number of tumours
Single	Ref	–	Ref	–
Multiple	1.20 (0.74–1.95)	0.5	1.05 (0.61–1.79)	0.9
Tumour diameter
<3 cm	Ref	–	Ref	–
≥3 cm	1.97 (1.27–3.05)	0.003	1.75 (1.13–2.73)	0.01
Treated with BCG
No	Ref	–	Ref	–
Yes	0.49 (0.31–0.76)	0.002	0.54 (0.34–0.84)	0.006

For the distribution pattern of CIS, unifocal CIS was used as the reference.

Discussion

Non‐muscle‐invasive bladder cancer is a very heterogenous disease with significant variation in individual risk of progression to MIBC. Risk stratification by the AUA and EAU guidelines or by using nomograms/risk calculators developed from randomised trials can help stratify NMIBC but remain suboptimal [6, 7, 13, 14, 15]. They all include concomitant CIS as an independent risk factor for progression without distinguishing between CIS distribution, whether unifocal or multifocal. Progression risk in NMIBC can vary from 1% to >40% depending on the NMIBC risk category, whether low, intermediate, high or the recently described very high risk [13]. All tumours with concomitant CIS are now re‐classified as high risk. T1 HG NMIBC associated with CIS and at least one additional risk factor is considered in the recent new EAU risk calculator as very high risk, with 5‐year progression rates of 44% (95% CI 30–61%), reaching 59% (95% CI 39–79%) at 10 years [13].

In this large, multi‐institutional study involving both academic and community institutions, only multifocal concomitant CIS was an independent adverse prognosticator of NMIBC progression at 5‐years for T1 NMIBC. Additionally, no differences in progression rates were observed for patients with T1 disease when concomitant CIS was unifocal only. In patients who received adequate BCG (Fig. S1), the cumulative incidence of tumour progression following index TURBT was strikingly different between unifocal and multifocal concomitant CIS (14% vs 33%), although this was a subset analysis.

The observed progression rates for patients with Ta/T1 NMIBC were in keeping with the recent literature (Table S2) and those reported in the new EAU risk calculator [13]. In our study, concomitant CIS was found in 11% of patients with NMIBC, which is in line with previous reports as well [12]. In a subset analysis of patients with concomitant CIS, as expected, T1 disease, multifocal CIS only, and tumour diameter ≥3 cm were independently associated with an increased risk of progression. These findings highlight the potential clinical relevance of accurately characterising the distribution of concomitant CIS and distinguishing between unifocal and multifocal CIS in NMIBC to better inform management. The high progression rates observed among patients with T1 disease and multifocal CIS may, in part, reflect understaging, particularly in those who did not undergo re‐staging TURBT. On univariable analysis, no difference in progression rates was observed between Ta with concomitant multifocal or unifocal CIS; however, multifocal CIS was an independent risk factor on multivariable analysis in the entire cohort. Prior studies have also observed that progression rates did not significantly differ between Ta Grade 3 with or without concomitant CIS [12].

Carcinoma in situ lesions have distinct characteristics compared to papillary tumours potentially contributing to their aggressive phenotype [4]. Most CIS are Class 2a/2b (UROMOL), associated with the least favourable prognosis due to late cell‐cycle activation with TP53‐, ERBB2‐, and APOBEC‐related pathways [17]. Explanations underlying why multifocal concomitant CIS may be an independent prognosticator are unclear and speculative at this stage, but likely due to multiple contributing factors [17]. A comprehensive multi‐omics landscape analysis of papillary NMIBC and CIS has revealed distinct profiles, each exhibiting unique clinical, pathological, and molecular features. Patients with CIS‐derived NMIBC have worse outcomes and a higher incidence of metastasis [18]. However, most studies have not distinguished between unifocal and multifocal concomitant CIS, likely due to challenges of analysing a minute amount of tissue like unifocal CIS by RNAseq or multi‐omics technologies. Multifocal CIS could be a manifestation of widespread field cancerisation and a surrogate for increased genomic instability [17]. Clinically, multifocal CIS may be observed in patients who experienced a delay in diagnosis or treatment, resulting in the unmonitored progression of unifocal to multifocal CIS and subsequently an increased rate of aggressive disease [2]. In addition, the superficially diffusive, discohesive, and flat nature of CIS can make it challenging to manage [1, 19]. CIS is also often under‐detected by white‐light endoscopic evaluation of the bladder [20]. In recent years, this has encouraged the use of blue‐light cystoscopy and narrow‐band imaging, which have been shown to improve the detection and management of CIS [20]. Multifocal CIS may also have an inherently more aggressive biological profile compared to unifocal CIS due to differences in cell adhesion protein expression [21]. For example, while both unifocal and multifocal CIS express similar levels of the protein E‐cadherin, a calcium‐regulated epithelial adhesion molecule (the loss of which can result in dedifferentiation, cancer development and progression), multifocal CIS cells have been observed to express the oncoprotein mucin 1 (MUC1) at higher levels compared to unifocal CIS cells [21, 22]. Specifically, MUC1 is believed to interfere with E‐cadherin‐mediated cell–cell adhesion, fuelling cell detachment, subsequent invasion, and metastasis [21, 22].

Missense ERBB2 mutations have also been found in close to half of CIS [23]. Together with epithelial‐to‐mesenchymal transition, cellular absence of cohesion, cell motility involved in tumour progression, and cell proliferation, oncogenic ERBB2/ERBB3 signalling are important for bladder tumour progression from CIS to invasive disease [24]. Multifocal CIS of the bladder may possess a higher propensity for seeding and progression through this phenomenon compared to unifocal CIS, providing a biological basis for its use as an independent prognosticator in the future [21, 22, 23, 24]. However at this stage, these observations are hypothesis generating only and not conclusive.

Our findings also emphasise the importance of comprehensive tumour mapping and submitting separate tumour specimens during TURBT to identify the presence of multifocal concomitant CIS (Fig. 2), which aligns with current guideline recommendations [6, 25]. To further optimise CIS detection and minimise the risk of under sampling or under reporting, resection can be initiated a few millimetres away from the main papillary tumour. This approach ensures that adjacent mucosa is included in the specimen, improving the chance of diagnosing CIS near the tumour base. Enhanced detection of CIS, either through the use of blue‐light cystoscopy or random bladder biopsies, in suspected HG NMIBC should also be considered [5, 6, 7, 26, 27]. Previous reports have also outlined that this practice may inform management, including BCG [27, 28]. While CIS was identified based on histological confirmation of suspicious areas during TURBT, systematic bladder mapping was not consistently performed in our cohort in accordance with current guidelines and existing literature [6, 17]. Nevertheless, its potential role warrants further investigation. Routine mapping of visually normal areas may enhance CIS detection but must be balanced against procedural risks and resource use. Future research should aim to identify specific patient subgroups, such as those with HG, multifocal, or recurrent disease, who may benefit most from routine mapping strategies. Additionally, emerging urinary biomarkers hold promise in complementing cystoscopic detection of CIS and may facilitate earlier or more accurate diagnosis in both diagnostic and surveillance settings [29].

Fig. 2.

Example of submitting separate tumour specimens at the time of TURBT to facilitate tumour mapping; the resection starts a few millimetres from the main tumour site to improve detection of adjacent CIS and potentially reduce under sampling. Picture reproduced with patient's permission.

Our findings definitely warrant further investigations at this stage. But if validated in large studies, concomitant CIS distribution (multifocal vs unifocal) could be incorporated into NMIBC risk calculators in the future, which currently only consider the presence of concomitant CIS [13, 14, 15]. Given the substantial progression rates observed among patients with T1 disease with multifocal CIS, our study provides further support to categorise these patients into the ‘very high risk’ group.

Our study is not without limitations. First, the retrospective nature of this study may be susceptible to selection bias and the inability to control for all confounding variables at the time of data collection. Second, while our study included patients from multiple academic and community institutions, these were all within the province of Ontario, Canada, which may limit generalisability due to variations in patient demographics, treatment practices, and outcomes between institutions and geographic regions. Third, most, but not all, patients with high‐risk disease received BCG and even fewer received adequate BCG. This is not unexpected though, given the ongoing issues with BCG shortage and is in line with real‐world outcomes [4]. In a study that included 412 patients with high‐risk NMIBC within Department of Veterans Affairs Centers across the United States from 1 January 2000 to 31 December 2015, only 37% of patients received adequate BCG therapy [3]. Other reports have highlighted that BCG is far from being administered to all eligible patients despite guidelines recommendations [30]. Fourth, intravesical chemotherapy was not frequently used in this high‐risk cohort, accounting for only 5% of the overall population, precluding any meaningful analysis of its impact on outcomes. Fifth, blue‐light cystoscopy, which may improve tumour detection compared to conventional white‐light cystoscopy, although its impact on NMIBC progression is unclear, was not routinely used in our cohort [20]. Additionally, urine cytology, advanced imaging, and data on prostatic urethral CIS were not consistently available during TURBT in our cohort, which may impact the identification of CIS and overall risk stratification. Systematic random biopsies during TURBT were not performed, although previous studies have questioned their usefulness [16]. However, CUA guidelines were followed, namely the complete resection of all visible tumours suspicious areas. Analyses were not stratified by hospital, as it would render the findings less generalisable. Finally, our study did not account for practice variations among urologists. Identification of multifocal CIS is dependent on submitting multiple TURBT specimens separately to pathology as recommended by guidelines [5]. There was no consistent strategy although this is representative of a real‐world setting. Thus, multifocal concomitant CIS may be under reported among patients diagnosed with unifocal concomitant CIS or papillary‐only disease if TURBT deviated from this recommendation. Additionally, defining concomitant CIS using only pathology specimens likely underestimates its true incidence. Satellite lesions may have been under sampled in this real‐world setting, particularly among high‐risk patients (e.g., those with advanced age, significant comorbidities, anticoagulation use, or poor bladder quality). To improve detection accuracy, future studies should incorporate a thorough review of operative notes and compare cystoscopy findings at the time of TURBT with corresponding pathology results.

Conclusions

This large multi‐institutional study combining both academic and community hospitals demonstrates the adverse prognostic impact of multifocal concomitant CIS on NMIBC progression compared to unifocal or peritumoral CIS. It highlights that not all CIS are created equal, reinforcing the potential clinical value of submitting separate tumour specimens to pathology for precise tumour mapping rather than collecting tumours in bulk. Recognising the distinct prognostic value of multifocal concomitant CIS and incorporating this finding into future NMIBC risk calculators, if validated by future prospective studies, may enhance the accuracy of risk stratification tools and optimise treatment approaches.

Disclosure of Interests

Alexandre R. Zlotta reports participation on a data safety monitoring board or advisory board for enGene, CG Oncology, CAMEVI, Janssen, Verity Pharmaceuticals, Ferring, mIR Scientific, Tolmar, and Theralase; and consulting fees from Janssen, Verity Pharmaceuticals, Ferring, mIR Scientific, Tolmar, enGene and Theralase. GSK reports advisory, consultant, or trial work with Merck, BMS, EMD Serono, Pfizer, Janssen, Ferring, Theralase, Verity, TerSera, Knight Therapeutics, Abbvie, Tolmar, PhotoCure, Astra Zeneca, enGene. All other authors declare no competing interests.

Supporting information

Fig. S1. Cumulative incidence of tumour progression following index TURBT among patients who received adequate BCG, stratified by distribution of CIS (log‐rank P = 0.02). All‐cause mortality was considered as a competing risk. The cumulative incidence of tumour progression at five‐years is indicated in parentheses.

Fig. S2. Cumulative incidence of tumour progression among patients with T1 NMIBC who received adequate BCG, stratified by CIS distribution (log‐rank P < 0.001). All‐cause mortality was considered as a competing risk. The cumulative incidence of tumour progression at 5‐years is indicated in parentheses.

Fig. S3. Cumulative incidence of tumour progression stratified by stage and CIS distribution (log‐rank P < 0.001), limited to patients with T1 NMIBC who underwent repeat TURBT.

Table S1. Fine and Gray competing risk regression for patients with concomitant CIS who received adequate BCG (n = 187) to estimate SHRs of tumour progression, accounting for all‐cause mortality as a competing risk. For the distribution pattern of CIS, unifocal CIS was used as the reference.

Table S2. The 5‐year cumulative incidence rates of NMIBC progression reported in recent literature stratified by Ta, T1, Ta + CIS, and T1 + CIS stages.