Weight of the resected specimen after transurethral resection as a new predictive variable for recurrence of non-muscle-invasive bladder tumour


Virginia Hernández Cañas, C/Budapest 1, 28922 Alcorcon (Madrid), Spain. e-mail: hernandezcv@gmail.com


What's known on the subject? and What does the study add?

Size is one of the currently used variables for predicting recurrence and progression of non-muscle-invasive bladder tumour (NMIBT), but size is subjective and not easy to measure.

We evaluate a new and objective variable to replace size, not previously reported. We assess the weight of the transurethral resection specimen as a predictor of recurrence and progression and conclude that this outperforms size as a predictor of recurrence of NMIBT.


  • • To evaluate the role of the weight of the resected specimen after transurethral resection as a predictive factor for recurrence and progression of non-muscle-invasive bladder tumour (NMIBT).


  • • The weight of the resected tumour was measured consecutively in 144 subjects who underwent transurethral resection of bladder tumours at our institution. The median (interquartile range [IQR]) follow-up was 58 (61.3) months.
  • • The probability of recurrence and progression at 1 and 5 years were calculated using the currently accepted variables.
  • • Thresholds for the specimen weight were determined according to percentiles and receiver–operating characteristic curves.


  • • The median (IQR) weight of the specimen was 6 (16) g.
  • • Multivariate analysis showed that the weight of the resected specimen was an independent predictive risk factor for recurrence at a threshold value of 6 g with a hazard ratio of 1.7 (95% confidence interval: 1.048–2.761) P= 0.03.
  • • Progression was not associated with the weight of the resected specimen.


  • • The weight of the resected specimen is a new variable for predicting the risk of recurrence of NMIBT.
  • • Tumours weighing >6 g, according to the present data, have a 1.7-fold higher likelihood of recurrence than those tumours that weigh less.

interquartile range


non-muscle-invasive bladder tumour


carcinoma in situ


transurethral resection of bladder tumour


European Organisation for the Research and Treatment of Cancer


transurethral resection


mitomycin C


receiver–operating characteristic


hazard ratio


area under the curve.


The objective assessment of the recurrence and progression of non-muscle-invasive bladder tumour (NMIBT) has been carried out by using a number of clinical and pathological tumour features [1]. These variables were given a score value and entered into a web-based calculator to help the clinician in the prediction of the clinical outcome [2]. The weighted score was based on six variables (number of tumours, tumour size, previous recurrence rate, T category, carcinoma in situ[CIS] and grade). The European Association of Urology subsequently adopted this system into its guidelines and, based on these scores [3], patients were stratified into low-, intermediate- and high-risk for recurrence [4]. This clinical tool has been validated in our practice [5], while other authors claim that underestimation of the results occurs in some subsets of patients [6].

In addition, there is an ongoing debate on which predictive variables should be used for recurrence and progression of NMIBT. In a study by Fernández-Gomez et al. [6], multiplicity, previous tumours, female gender and CIS were significant predictors of recurrence in multivariate analysis; however, according to other authors, grade and tumour stage were not predictive of recurrence and only tumour size and the use of intravesical instillations, in addition to the rest of those previously reported, were accurate predictors [1,7].

To evaluate size as a prognostic variable, Millán et al. [1] defined tumour size as the largest tumour diameter measured with the resection loop that is ordinarily 1 cm long. Accordingly, size was categorized as <1.5 cm, 1.5–3 cm, and >3 cm. A threshold of 3 cm was set for the prognosis of NMIBT [1].

There are various clinical scenarios in which the measurement of tumour size is hampered by the tumour features, e.g. multiple small tumours and small tumours adjacent to a larger tumour. In addition, in many cases, tumour size does not represent the overall tumour burden, e.g. in solid tumours mixed with papillary lesions or in bulky tumours with a small pedicle. Furthermore, endoscopic assessment of tumour size is highly subjective and not reproducible, being highly operator-dependent. With this in mind, we aimed to evaluate the role of tumour specimen weight after transurethral resection of bladder tumour (TURBT) as a new and objectively measurable variable in the prediction of recurrence and progression of NMIBT.



From January 1999 to December 2009, 520 patients underwent TURBT in our institution for primary NMIBT. These patients' pathological and clinical data were prospectively entered into our bladder cancer database, and the transurethral resection (TUR) specimen of a subset of 144 consecutive patients was weighed. Additional clinical data were entered in the database by an independent reviewer. The study was approved by our institutional review board and was conducted in accordance with the provisions of the Declaration of Helsinki. No additional tests were carried out on the patients apart from those included in routine standard clinical care.

Demographic variables were collected, as well as those included in the European Organisation for the Research and Treatment of Cancer (EORTC) risk tables by Sylvester et al. [2], in addition to the specimen weight. The study endpoints were recurrence and progression rates. The latter was defined as upstaging to muscle-invasive or metastatic disease. High-risk tumours Ta-T1 G3, with or without CIS, were not excluded from the study. Multiple random biopsies were not routinely performed. All re-TURs were excluded from the study. Histopathology results were classified according to the 2009 TNM system [8] and WHO (1973) grading system [9].


Complete TURBT was always attempted and a biopsy of the deep muscle at the tumour base was obtained using the resection loop until healthy muscle or perivesical fat were seen. This base sample was submitted separately to the laboratory. Operations were performed by eight urologists with >10 years of experience or by residents supervised by these same urologists. Tumour size was defined as the largest dimension, as assessed by endoscopic viewing during resection. A second resection in high-risk tumours was considered when the initial resection was incomplete, when the tumour was large or there were multiple tumours, or when the pathologist reported that the specimen contained no muscle tissue. Patients who were upgraded/upstaged (2.8%) were assigned to the most unfavourable category. One single immediate, postoperative intravesical instillation of chemotherapy with mitomycin C (MMC) was administered in all cases whenever: (i) gross active haematuria was absent; (ii) the ureteric orifice had not been resected; or (iii) the operating urologist did not have a well-founded suspicion of bladder perforation. In short, intermediate-risk patients, with highly recurrent disease, were offered a 6-week course of MMC and high-risk patients were offered BCG induction, plus maintenance if a response was obtained. Follow-up was carried out every 3 months for the first year with cystoscopy and cytology. Cytology plus cystoscopy or ultrasonography were used every 4 months during the second year and every 6 months thereafter.

The TUR specimen was placed in formaldehyde for its fixation before being sent to the pathology laboratory. The laboratory technicians measured, weighed and processed the specimens following the same protocol and using the same precision scale (GF-200R, A&D Co. Ltd, Tokyo, Japan), with a lower detection value of 0.001 g. All the TUR specimens obtained at the time of TURBT were included in the analysis. Tumour weight was included in the pathology report along with the standard evaluation.


Measurable variables are expressed as mean (sd) if normally distributed or median (interquartile range [IQR]) if not. Quantitative variables were compared using Student's t-test after an evaluation of normal distribution (Kolmogorov–Smirnov) test and equality of variance. Categorical variables are expressed as proportions and were compared using Pearson's chi-squared test with continuity correction or Fisher's exact test, as appropriate. Correlation between the weight and tumour size was performed using Spearman's correlation coefficient.

Recurrence- and progression-free survival curves were estimated using the Kaplan–Meier method, and differences between curves were evaluated using the log-rank test. Patients known to have died from causes unrelated to bladder cancer comprised <5% of the study population and they were censored in the recurrence and progression analysis.

The following methods were used to estimate the weight thresholds:

  1. Percentiles: the sample was divided into deciles and the weight threshold was taken as the decile in which a change in risk was detected.

  2. Receiver–operating characteristic (ROC) curves: the optimum weight threshold was estimated by the point that maximizes the sum of specificity and sensitivity as expressed by the area under the ROC curve.

We used multivariate Cox proportional hazard regression to assess the association of weight as an independent variable with recurrence. All variables significantly associated with the outcome in the univariate analyses were considered in the maximum multivariate model as well as corresponding interaction terms. A manual backward modelling strategy was used to eliminate variables from the maximum model to obtain the most parsimonious model to assess the effect of the independent variables on outcome. We tested the proportional hazard assumption by examining log survival plots for different categories against time.

All tests were two-tailed and a P value <0.05 was considered to indicate statistical significance. Statistical analyses were performed using the SPSS 15.0 statistical package for Windows (SPSS Inc, Chicago, IL, USA).


Table 1 lists the clinical and pathological characteristics of the series. The median (IQR) follow-up time was 58 (61.3) months. Seven patients (4.8%) died from causes unrelated to bladder cancer and eight patients (5.5%) were lost to follow-up. The 3-month recurrence-free survival was 97.9%

Table 1. Patients' clinical and pathological characteristics
 Patients with weighed tumours, n= 144
Mean (sd) age68.08 (12.9)
Male gender, n (%)125 (86.8)
Single tumour, n (%)109 (75.7)
Grade, n (%) 
 G174 (51.4)
 G249 (34)
 G318 (12.5)
 CIS, n (%)4 (2.8)
Stage, n (%) 
 Tx9 (6.3)
 Ta74 (51.4)
 T161 (42.3)
Size <3 cm, n (%)85 (59)
Postoperative single dose MMC, n (%)86 (59.7)
MMC induction, n (%)23 (16)
BCG induction, n (%)25 (17.4)
BCG maintenance, n (%)15 (10.4)

Median (IQR) resected weight was 6 (16) g, with a minimum and maximum weight of 0.03 and 115 g, respectively. The weight in g showed a distribution skewed to lower values. Correlation between weight and size of the tumour specimen was notably low (r= 0.32), with a great variability in size for each weight percentile range (Fig. 1).

Figure 1.

Correlation between weight and size of the tumour specimen.


The recurrence-free survival rates at 1, 3 and 5 years were 66.7, 49.5, 43.2% respectively. Table 2 shows the association of tumour weight with recurrence. We found that for a threshold value of 6 g the risk of recurrence greatly increased. After controlling for the remaining variables, a tumour weight >6 g entailed an increased risk of recurrence estimated at a hazard ratio (HR) of 2.47 (95% CI: 1.19–5.1, P= 0.01 [Fig. 2]). Above this weight there was no linear increase in the recurrence risk, which reached a HR of 2.64 (95% CI: 1.048–2.761, P= 0.009) for tumours weighing >18 g. The ROC curve for this threshold of 6 g showed an area under the curve (AUC) of 0.65 (95% CI: 0.57–0.73), with a sensitivity of 60% (CI 95%: 46.8–70.3) and specificity of 62% (CI 95%: 49.7–73.2). In the present study, the ROC curve for a tumour size of 3 cm, which is an accepted threshold in the literature, showed an AUC of 0.57 (95% CI: 0.47–0.66) with a sensitivity of 32.9% and a specificity of 78.9% (Fig. 3).

Table 2. Analysis of the weight thresholds by percentiles
Weight, g (percentiles)HRCI (95%) P–2.040.39
Figure 2.

Recurrence-free and progression-free survival stratified according to tumour weight.

Figure 3.

ROC curve for tumour weight and size.

In the univariate analysis, multiplicity, stage, grade and weight (≥6 g) were associated with recurrence (Table 3). Those variables that showed significance in univariate analysis and those that could be clinically relevant were included in the maximum multivariate regression model. After adjusting for size, grade, multiplicity and postoperative single-dose MMC, weight >6 g was a predictor of the risk of recurrence with a HR 1.7 (CI 95%: 1.048–2.761), P= 0.03

Table 3. Univariate analysis
VariablesHRCI (95%) P
Size <3 cm1.4370.907–2.2770.123
Postop MMC0.7130.438–1.1620.175
Weight ≥6 g1.7751.112–2.8320.016


Progression-free survival rates at 1, 3 and 5 years were 96.5, 95 and 95%, respectively.

Weight was not an independent predictor of progression in multivariate analysis, nor was it associated with a higher risk of progression (Fig. 2). Owing to the limited number of the sample, strong conclusions cannot be drawn at this point.


The assessment of the prognostic factors recurrence and progression of NMIBT is of great help in selecting the type of adjuvant treatment after TURBT and in categorizing patients into risk groups that enable us to compare different treatment regimens [10]. This categorization should eventually provide clinically relevant information about the length of adjuvant treatment [11] or for the change in the therapeutic strategy towards radical surgery for those patients harbouring non-responding high-risk tumours [12].

In addition to the EORTC risk calculator [2] and the CUETO tables [13], a nomogram has been proposed to predict the recurrence risk based on age, gender, cytology and urinary nuclear matrix protein 22 in patients with NMIBT [14].

In the landmark reports by Millan et al[1,7], and the later meta-analysis by Sylvester et al. [2], the tumour burden, expressed either by the size of the tumour or by the number of tumours, was the most important prognostic factor.

In the report by Sylvester et al. [2], size provides a 17% (3/17) relative value of the total score. As an example of the importance of size in the risk assessment, a single primary pT1G2 without CIS of 2.5 cm would have a 1- and 5-year recurrence risk of 24% and 46%, respectively. If this same tumour were 5 mm larger, i.e. 3 cm, the 1- and 5-year recurrence risk would be 38% and 63%, respectively. This means that a 58% and 36% risk increase at 1 and 5 years is calculated by only a 5-mm increase in size which is measured in a non-objective way and subject to significant interobserver variability.

This observer-dependent measurement could be excluded by using an objective, measurable and reproducible variable such as the weight of the complete TUR specimen [15] following strict quality control criteria [16].

No information is available on how size was measured in any of the trials included in the meta-analysis by Sylvester et al. [17–22]. In the study by Millan et al. [1] size was categorized into three groups; however, significance of the recurrence risk (HR 1.6) was only reached for a size threshold of 3 cm. In the present study, we noted a relationship between weight and the risk of recurrence; tumours weighing >6 g had a higher risk of recurrence with a HR of 1.7.

Tumour volume and the pattern of growth are critical issues in the biology of the tumour. This growth rate has also been correlated with the potential for recurrence and the development of metastasis [23].

The weight of the resected specimen more accurately represents the real tumour volume than does tumour size. By adding the weight of multiple resected tumours, it is possible to obtain the total tumour volume of a commonly multifocal disease, as is the case in NMIBT. In addition, as opposed to visual observation, weight of the specimen informs us of the total tumour volume and not only of the exofitic part of the neoplasm. In the present study, a tumour with the same size had significant differences in weight owing to the configuration of the lesion. Solid tumours were denser and weighed more (median 16.5 g) than papillary lesions (median 6 g), P= 0.002.

Progression of NMIBT is also influenced by other factors: accuracy of the staging by the pathologist, use of random biopsies, definition of progression, quality of the TUR, use of adjuvant therapy and duration of follow-up as well as some molecular expression of the tumour [24]. These factors may have influenced the lack of significance of weight in predicting progression.

External validity assessment was done by testing the homogeneity of the study group with the overall series of patients operated on in our hospital with NMIBT. This group showed no significant difference from the global population in any of the variables, which makes both groups clinically similar. No significant differences were found either between the recurrence-free survival of the overall series and the study group at 1, 3 and 5 years with a HR of 1.20 (95% CI: 0.93–1.57) P= 0.15, or progression-free survival at 1, 3 and 5 years with a HR of 1.13 (95% CI: 0.48–2.63) P= 0.70.

One limitation of this study is related to sample size and therefore to the accuracy of our data. The power of our study for this sample size was estimated at 95%. Moreover, the number of relapsed patients in the series (73 patients) did not represent a limitation for inclusion of variables in the logistic regression model. There are other limitations to the present study that should be acknowledged, including the matter of resection completeness. All procedures were performed by eight experienced urologists or supervised residents with the goal of achieving a complete TUR of all visible tumour in all cases. The resection of some healthy tissue is performed in all resections so we must concede that the weight values were overestimates; however, this applied to all cases and should not have led to bias. Nevertheless, the lack of randomization makes it necessary to confirm these results in future prospective studies.

In conclusion, the weight of the resected specimen after TUR of NMIBT is a prognostic variable of recurrence that has not previously been described. Standardized complete TUR allows us to obtain a measurable predictor unrelated to observer subjectivity that can eventually be used in the current risk calculator should our findings be confirmed. In the present series, those tumours that weigh >6 g had a 1.7-fold risk of recurrence.


None declared.