So what does the patient with high grade, non-muscle invasive bladder cancer who has disease that is unresponsive to (or has quickly recurred after) adequate intravesical therapy with Bacillus Calmette Guerin (BCG) do, particularly if he (she) does not want to lose his (her) bladder? Two agents that received priority, fast track priority review by the Food and Drug Administration (FDA) for such patients, Vicinuim  and Nadofaragene (rAd-If Na/Syn 3) , which have efficacy in this circumstance have experienced production problems and are not available [3, 4]. In a non-randomized (single arm) phase 3 study, Vicinium, an antibody to the transmembrane cell adhesion molecule, EpCam, conjugated to Pseudomonas exotoxin A (which stops protein translation and induces apoptosis) there was a complete response (CR) rate of 39% at three months and 17% at 12 months , while Nadofaragene, an adenovirus mediated “gene therapy” producing interferon alpha  had 53% CRs at 3 months and 45% at 12 months in another single arm phase 3 study.
This leaves intravesical chemotherapy with one or two agents (usually gemcitabine and/or docetaxel)  or other drugs (e.g. mitomycin C or valrubicin), or intravenous infusions of Pembrolizumab (an anti-PD-1 antibody)  for these patients.
Pembrolizumab, in the Keynote 057 single armed trial, had a three month CR rate of 38% and a two-year recurrence free survival rate of 15%  and also is FDA approved for this clinical scenario. While there are numerous agents in testing, including intravesical instillations of Coxsackievirus A21  and a gemcitabine impregnated device placed in the bladder (TAR 200) , none are available currently and are unlikely to be until ongoing trials are completed.
Thus it is particularly disappointing that a recent publication indicates that even with its modest efficacy, Pembrolizumab (200 mg IV every three weeks until recurrence up to two years), is not cost-effective compared to intravesical chemotherapy (which has limited efficacy) or even the radical cystectomy (RC) that patients are trying to avoid .
In an elegant cost analysis study using a decision analysis Markov model, Wymer and colleagues authors found that at it’s current cost (over $100,000 per year) , Pembrolizumab was far less cost-effective than either RC or sequential intravesical chemotherapy with gemcitabine-docetaxel (gem-doce) consisting of six weekly instillations and then monthly maintenance for two years [5, 9]. While the authors found RC most cost effective at the five year time horizon, the full five years was needed (this is a problem since assumptions about outcomes beyond two years had to be made for both gem-doce and Pembrolizumab). Moreover, in the analysis, the expected rates of adverse events (especially for RC) were somewhat optimistic. For example, five year incidences of chronic bowel toxicity including bowel obstruction requiring hospitalization or surgery, of parastomal or incisional hernia warranting revision, and of stricture revision of the ureter were assumed to be 1% each (although the range for each tested in sensitivity analyses was 0–13%). Moreover if the rate of metastases at two years estimated for gem-doce dropped minimally, from 6.1% to 5.9% , then intravesical chemotherapy would be most cost effective.
That said, based purely upon one report of outcomes and complications, which appeared in abstract form, the cost of Pembrolizumab was so great, that even assuming no metastatic events at two years, a cost reduction of more than 90% was needed for Pembrolizumab to become cost effective with RC or gem-doce. While sensitivity analyses varying the incidences of adverse events was used for all the assumptions, Pembrolizumab was still not cost effective without drastic reductions in its price.
There are numerous limitations in this analysis, particularly a reliance on very few studies to estimate adverse events and outcomes (which the authors readily acknowledge), and that long term outcomes and adverse events were based on assumptions. However, the basic conclusion was that in treating BCG unresponsive or recurrent disease, RC or intravesical gem-doce were the best available options, and individual circumstances, including predicted five-year survival other than from bladder cancer, already present and likely future urinary symptoms, the perceived inconvenience and morbidity from further intravesical treatments, and, of course, the ability to withstand and willingness to undergo RC, must all be considered before the “best” approach is chosen.
CONFLICTS OF INTEREST
The author has no conflicts of interest to report.
 Shore N , O’Donnell M , Keane T , et al. Phase 3 results of Vicinium in BCG-unresponsive non-muscle invasive bladder cancer (PD03-02) J Urol. 203:(Suppl 4) doi: 10.1097/JU0000000000000823.02
 Boorjian SA , Alemozaffar M , Konety BR , et al. Intravesical nadofaragene firadenovec gene therapy for BCG unresponsive non-muscles invasive bladder cancer, a single arm, open table repeat dose clinical trial. Lancet Oncol. 2021;22:107.
 sessen bio, August 13, 2021
 FerGene, May 17, 2020
 Steinberg RL , Thomas J , Brooks N , et al. Multi-institutional evaluation of sequential gemitavine and docetaxel as rescue therepy for non muscle invasive bladder cancer. J Urol. 2020;203:902–9.
 Balar AV , Kulkarni GS , Uchio EM , et al. Phase II trial of Pembolizumab for patients with high risk non-muscle invasive bladder cancer unresponsive to bacillus Calmette-Guerin. J Clin Oncol. 2019;37:350.
 Messing E . Intravesical Coxsackievirus virus A21 Therapy for non-muscle invasive bladder cancer. Bladder Cancer. 2020;6:97–8.
 Tyson MD , Morris D , Palou J , et al. Safety, tolerability and efficacy of TAR-200 in patients with muscle-invasive bladder cancer who were unfit for curative intent therapy: A phase 1 trial to be presented at the Society of Urologic Oncology (SUO) meeting, Dec 1-3, 2021.
 Wymer KM , Sharma V , Sargal CS , et al. Cost effectiveness analysis of Pembrolizumab for bacillus-Calmette-Guerin- unresponsive carcinoma in site of the bladder. J Urol. 2021;205:1326–36.