Richard K. Lee, Brady Urologic Health Center, Weill Cornell Medical College, New York Presbyterian Hospital, 525 East 68th Street, Box 94, Starr 900, New York, NY 10065, USA. e-mail: firstname.lastname@example.org
Study Type – Therapy (cost effectiveness analysis)
Level of Evidence 2a
What's known on the subject? and What does the study add?
Bladder cancer is one of the costliest malignancies to treat throughout the life of a patient. The most cost-effective management for low–risk non-muscle-invasive bladder cancer is not known.
The current study shows that employing cystoscopic office fulguration for low-risk appearing bladder cancer recurrences can materially impact the cost-effectiveness of therapy. In a follow-up protocol where office fulguration is routinely employed for low-risk bladder cancers, peri-operative intravesical chemotherapy may not provide any additional cost-effectiveness benefit.
• To examine the cost-effectiveness of fulguration vs transurethral resection of bladder tumour (TURBT) with and without perioperative intravesical chemotherapy (PIC) for managing low-risk carcinoma not invading bladder muscle (NMIBC). Low-risk NMIBC carries a low progression rate, lending support to the use of office-based fulguration for small recurrences rather than traditional TURBT.
MATERIALS AND METHODS
• A Markov state transition model was created to simulate treatment of NMIBC with vs without PIC, with recurrence treated by formal TURBT vs treatment with fulguration.
• Costing data were obtained from the Medicare Resource Based Relative Value Scale.
• Data regarding the success of PIC were obtained from the peer-reviewed literature, as were corresponding utilities for bladder cancer-related procedures.
• Sensitivity analyses were performed.
• At 5-year follow-up, a strategy of fulguration without PIC was the most cost-effective (mean cost-effectiveness = US $654.8/quality-adjusted life year), despite a lower recurrence rate with PIC.
• Both fulguration strategies dominated each TURBT strategy.
• Sensitivity analysis showed that fulguration without PIC dominated all other strategies when the recurrence rate after PIC was increased to ≥14.2% per year.
• Similarly, the cost-effectiveness of TURBT becomes more competitive with fulguration when the total cost of TURBT declines < US $1175.
• The present study shows that fulguration without PIC was the most cost-effective strategy for treating low-risk NMIBC.
• The effectiveness of PIC and the cost of TURBT can materially impact the cost-effectiveness of the different management strategies.
• These results should be considered in treatment decisions in the context of preserving oncological control.
Bladder urothelial carcinoma is one of the costliest malignancies to treat in the USA . Most cases of carcinoma not invading bladder muscle (NMIBC), which account for 70–80% of patients initially diagnosed with bladder urothelial carcinoma, show a low progression rate to more advanced disease but do display a high rate of recurrence . As such, patients with NMIBC will often incur a significant burden of treatment in the absence of disease progression as they are monitored for several years. This treatment burden encompasses both the office-based procedures used to monitor patients, as well as the surgical operating room-based procedures employed to treat those recurrences.
Level I evidence shows that a single dose of perioperative intravesical chemotherapy (PIC) reduces recurrences of NMIBC when used as an adjunct treatment to TURBT. Both European and US guidelines recommend PIC as adjuvant monotherapy for patients with NMIBC who are at low risk for progression and recurrence [3,4]. The most significant reduction in bladder cancer recurrence rates as a result of PIC use may manifest in these low-risk tumours (i.e. those that are small, solitary and presenting for the first time) . Despite this, many choose to forego PIC as a result of the inherent low risk of these tumours for which oncological control may be adequately achieved through straightforward fulguration alone .
Although NMIBC recurrences are traditionally treated with formal resection under regional or general anaesthesia in the operating room, it is also possible to manage these low-risk tumours through less expensive office fulguration . With these variables in mind, we aimed to model the cost-effectiveness of various approaches for managing low-risk NMIBC. Specifically, we looked at the use vs deferral of PIC at the time of initial resection, as well as office-based fulguration vs formal TURBT for the treatment of recurrences. We hypothesized that strategies minimizing the need for expensive operating room-based procedures would be the most cost-effective.
MATERIALS AND METHODS
A Markov state transition model was constructed to simulate different management strategies for a hypothetical cohort of patients with low-risk NMIBC after the initial TURBT (Fig. 1). In this model, patients exist in one of two disease states during each cycle: disease-free or disease-recurrence. We defined low-risk NMIBC with the features: primary, single tumour, <3 cm with the appearance of a low-grade, low-stage lesion. Additionally, high-grade preoperative urine cytology would exclude a patient from this hypothetical cohort. Because we are considering primary tumours, and PIC is typically given immediately after TURBT, the model assumes that PIC is given before obtaining histological diagnosis. After the initial TURBT, all patients were assumed to have been rendered disease-free. Next, four management strategies were considered: (i) single-dose PIC followed by the use of formal TURBT for any recurrences seen on flexible cystoscopy; (ii) single-dose PIC followed by the use of office-based fulguration for low-risk appearing tumour recurrences; (iii) no PIC followed by TURBT for any recurrences seen on flexible cystoscopy; and (iv) no PIC followed by office fulguration for low-risk appearing recurrences. Treatment and subsequent return to a disease-free state was assumed with each recurrence. Patients in the PIC treatment arms were assumed to only receive intravesical therapy at the time of initial TURBT and not at the time of recurrence. A cycle length of 3 months was chosen, which is consistent with common local practice for NMIBC surveillance. Disease progression was not modelled for two reasons: (i) rates of progression for small, solitary, initial TaG1 tumours are very low  and (ii) there is no evidence to suggest that the use vs deferral of a single dose of PIC would have any impact on disease progression .
Transition probabilities, which represent the probability of moving from a disease-free to a disease-recurrent state, were based on the meta-analysis reported by Sylvester et al.. In their study, the overall recurrence rate for patients treated with PIC was 36.7% with a median follow-up of 3.4 years. At a constant rate of recurrence, the 3-month recurrence rate (i.e. the cycle length of the present model) was calculated at 2.70%. The Markov model was analyzed for a total of 20 cycles, which is equivalent to 5 years of follow-up. Of note, subjects comprising the cohort reported by Sylvester et al. were treated with either mitomycin C, epirubicin, thiotepa or pirarubicin. Cost-modelling was based on mitomycin C for the purposes of the present study.
Direct procedural costs were derived from the Medicare Resource Based Relative Value Scale, which functions as a standard for other fee schedules in the USA. Utility data for bladder cancer were obtained from similar, previously published analyses (Table 1) [9,10]. The model was evaluated using Treeage Pro 2011 software (Treeage Software Inc., Williamstown, MA, USA).
Table 1. Costs, based on Medicare reimbursement schedules, of various aspects of treatment for carcinoma not invading bladder muscle
Cost (US $)
Utilities are quality-of-life weights attached to each health state in the model (three values listed are non-US $ amounts). CPT, current procedural terminology; PIC, perioperative intravesical chemotherapy; OR, operating room.
Cost-effectiveness was calculated for each cycle and multiplied by the proportion existing in that given state during the particular cycle. One-way sensitivity analysis was performed to test the importance of deviations from base case model assumptions, varying the efficacy of PIC in preventing bladder tumour recurrence, as well as the cost of PIC and formal TURBT. Two-way sensitivity analysis assessed the interactions: PIC efficacy vs TURBT cost and PIC cost vs TURBT cost.
BASE CASE ANALYSIS
At 5-year follow-up, a strategy of fulguration without PIC was the most cost-effective (mean cost-effectiveness = US $654.8/quality-adjusted life year [QALY]), followed by PIC + fulguration (mean cost-effectiveness = US $687.84/QALY), despite a lower recurrence rate when PIC was used (Table 2). Although the comparative effectiveness of PIC + fulguration was greater than fulguration alone (QALY 14.50 vs 14.36), the incremental cost increase associated with the former strategy outweighed the benefit of the incremental effectiveness in a US $/QALY analysis. Similarly, both fulguration strategies showed the superiority of cost-efficiency over both TURBT strategies, which is dependent largely upon the higher incremental costs of TURBT. The PIC + TURBT strategy showed a similar efficacy to PIC + fulguration (14.48 vs 14.5 QALY), although the high incremental cost (US $934.41) of PIC and formal TURBT made this strategy the least cost-effective.
Table 2. Cost-effectiveness (C-E) of various treatment strategies for carcinoma not invading bladder muscle: base case analysis
PIC, perioperative intravesical chemotherapy; QALY, quality-adjusted life year.
PIC + TURBT
PIC + fulguration
No PIC + TURBT
No PIC + fulguration
One-way sensitivity analysis modelled the effect of varying the efficacy of PIC on the recurrence rate of NMIBC with respect to the probability of a yearly recurrence in the range 4–20% (Fig. 2A). The two fulguration strategies showed superior cost-effectiveness (i.e. fewer US $/QALY across the full range of recurrence rates). PIC + fulguration and fulguration alone were co-dominant strategies until the yearly recurrence rate after PIC increased to ≥14.2% per year, when fulguration alone became singularly dominant. Additionally, with a yearly recurrence rate <10%, PIC + TURBT showed a cost-efficency beyond TURBT alone.
Varying the cost of PIC from US $50 to US $1000 showed the cost-efficiency of a PIC + TURBT strategy over TURBT alone for total intravesical therapy costs < US $263 (Fig. 2B). Additionally, only when PIC costs moved towards zero did the cost-effectiveness of PIC + fulguration approach that of fulguration alone.
Similarly, varying the cost of TURBT (Fig. 2C) showed that TURBT becomes competitive (from a cost-effectiveness standpoint) with fulguration strategies when the total cost of TURBT falls below US $1175. Note the intersection of plots between PIC + fulguration and PIC + TURBT, as well as between fulguration alone and TURBT alone, with both occurring at US $1175.
Two-way sensitivity analysis modelled the effects on cost-efficiency of varying both the efficacy of PIC and the cost of TURBT (Fig. 3A). Strategies involving TURBT were cost-efficient only at the lower range of TURBT cost (i.e. < US $1175). PIC + fulguration and fulguration alone were similarly cost-effective at higher TURBT costs. As an example, an equivalent cost-effectiveness was shown for those two strategies when the cost of TURBT and the rate of yearly recurrence after PIC was US $3725 and 7.2%, respectively, as well as US $4050 and 13.6%, respectively.
A second two-way sensitivity analysis was performed, varying the cost of PIC vs the cost of TURBT (Fig. 3B). For example, PIC + fulguration and fulguration alone showed similar cost-effectiveness at TURBT and PIC costs of US $3875 and US $335, respectively, as well as at costs of US $3650 and US $97.50. Notably, neither the efficacy of PIC, nor its cost has a significant impact on cost-efficiency compared to the cost of TURBT.
The most cost-effective approach for managing patients with low-risk NMIBC is unknown. Moreover, the dominant strategy may be different in the USA, Europe and Canada as a result of differences in reimbursement structure. Cost considerations may become less important as the aggressiveness of a malignancy increases but do become important when dealing with diseases such as low-risk NMIBC, which comprises an entity with a low progression rate but a recurrence rate that necessitates long-term surveillance and follow-up. Given these cost considerations, some studies have suggested that small, low-risk appearing recurrences may be treated with office-based fulguration [6,7] or even followed expectantly . In addition, although the use of PIC decreases recurrence rates in low-risk NMIBC [4,5,12], some studies challenge the true benefit of treating these ‘nuisance tumours’ and call into question the cost-efficiency of such an approach .
In our base case analysis, we found that deferring PIC at the time of initial TURBT and treating subsequent recurrences with TURBT was more cost efficient than a strategy of PIC + TURBT (Table 2). This is in contrast to Feifer et al. who reported that PIC + TURBT represented a dominant and cost-effective strategy over TURBT alone as early as 4 years after the initial TURBT. In agreement with the findings of Feifer et al., sensitivity analysis (Fig. 2A) in the present study showed that slight improvements in the efficacy of PIC would render PIC + TURBT a more cost-efficient strategy than TURBT alone. Given that the present analysis, as well as that of Feifer et al., used transition probabilities informed by the meta-analysis of Sylevester et al., the disparity in base case findings may reflect differences in costing assumptions inherent to a US vs Canadian model. Additionally, in contrast to the present study, these authors did not examine the economic effect of a fulguration-for-recurrence strategy .
We found that fulguration for recurrence strategies were more cost-efficient than the TURBT strategies (Table 2). Sensitivity analysis, varying the efficacy of PIC, showed that fulguration alone and PIC + fulguration were co-dominant strategies for a post-PIC yearly recurrence rate of up to 14.2% (Fig. 2A). Beyond this recurrence rate, fulguration alone became a singularly dominant strategy (i.e. the efficacy of PIC deteriorated to a point such that the cost could not be justified). Two-way sensitivity analysis (Fig. 3) showed the minimal importance of PIC cost and efficacy on net cost-efficiency compared to TURBT cost. Therefore, the cost of TURBT is the over-riding factor when deciding between strategies.
The findings obtained in the present study are supported by those of Rao and Jones , where an analysis of the economic benefit of routine PIC as described by the meta-analysis of Sylvester et al. showed that the treatment of bladder cancer recurrences in the outpatient setting offered an approach that obviated most (if not all) of the purported cost benefit of PIC, despite the known reduction in tumour recurrence with such therapy. Rao and Jones  calculated that the cost of the outpatient management of small recurrences with office fulguration was US $1167, which is a significant savings over the cost of US $2114 seen with TURB in an ambulatory surgery centre vs US $2666 in a hospital outpatient centre vs US $7025 for inpatient TURB. Rao and Jones  state that the magnitude of these savings was such that the cost benefit of intravesical therapy would be seriously eroded if recurrences were managed on an outpatient basis. Office-based management of low-risk NMIBC can therefore provide a compelling economic method for lowering the cost burden of managing disease, the magnitude of which could potentially obviate the benefit of PIC.
Herr et al. and Donat et al. have also shown excellent oncological control in a cohort of 215 patients with low-risk NMIBC managed with office-based fulguration for small-volume recurrences at a median duration of follow-up of 8 years. Tumours >0.5 cm in size or five in quantity were treated with TURBT under anaesthesia. In a cohort where 67% (143/215) of patients had at least one recurrence, patients required one fulguration every 2 years and only one TURBT every 3 years. In total, 17 (8%) patients progressed in grade or stage, all of whom had multiple tumours at the initial diagnosis; one (0.5%) died from disease. Fulguration for recurrence in low-risk bladder tumours may well provide a cost-effective and oncologically sound treatment paradigm.
There are several limitations to the present study. For example, there is no existing direct utility data for TURBT; the present study drew from utility data reported by others in the literature, which was itself extrapolated from data from similar endoscopic procedures [9,10]. We also assumed a constant rate of bladder tumour recurrence at every 3-month cycle, which may not be representative of the true biological behaviour of NMBIC. The present study also draws PIC data from high-volume, tertiary care practices whose patient composition may not correlate with that seen in a community practice. The cost parameters reflected in the present study are, in addition, applicable mainly to the US and may not apply to other countries, although this limitation was mitigated by the use of sensitivity analyses. Additionally, the model requires that PIC is given before histological diagnosis. We did not take into account the detriment upon QALY associated with receiving PIC, which histology would later prove was an unnecessary indication.
Finally, our conclusion that low-risk recurrences can be treated with office-based fulguration is predicated upon the reliability of visually diagnosing such lesions. Although experts have shown a high concordance rate between visual impression and pathological results, such accuracy is largely operator-dependent . However, improvements in endoscopic instrumentation, emerging imaging techniques and urinary biomarker development  (albeit at an increased cost) will hopefully continue to improve and assist in the accurate minimally-invasive diagnosis of bladder tumour recurrences in the future.
In conclusion, a single perioperative intravesical chemotherapy treatment will decrease the rate of recurrence of low-risk NMIBC, although the cost-efficiency of such adjuvant treatment in a TURBT for recurrence model is diminished if recurrence after PIC is greater than 10% per year. Should such occurrences arise, the more cost-effective strategy includes office-based fulguration of low-risk appearing lesions. The main benefit in cost-effectiveness arises from the use of office-based fulguration rather than TURBT as a result of the significant cost differences between the procedures. In comparison, the cost-effectiveness of using PIC in a fulguration-for-recurrence schema is only moderate in magnitude.
Supported in part by the Frederick J. and Theresa Dow Wallace Fund of the New York Community Trust.