Muscle-invasive bladder cancer (MIBC) is treated with radical cystectomy and possible neoadjuvant chemotherapy (NAC) or adjuvant chemotherapy (AC). NAC provides a 5%–8% absolute increase in overall survival (OS) at 5 years compared to surgery alone, with a number-needed-to-treat of ˜10–20.¹ NAC aims to reduce the primary tumor burden while treating subclinical micrometastases, though it delays definitive extirpative therapy for nonresponders. Even though ˜30%–50% of patients with MIBC are downstaged following NAC on final surgical pathology to ≤pT1N0 and ˜20%–40% to pT0, this translates to a mere ˜5% OS benefit (fig. 1).^2-4 Surgical treatment is postponed in these patients, and they are subjected to the toxicities of chemotherapy. Thus, there is the risk of overtreatment with NAC. Moreover, for logistical and socioeconomic reasons, only about 15%–30% of patients with MIBC who undergo radical cystectomy receive NAC.⁵ Nevertheless, NAC is standard of care and is recommended by major guidelines to eligible patients, perhaps owing to the initial higher level of evidence supporting its use compared to AC.⁶

Figure 1. Sankey diagrams showing the proportional flow of MIBC patients treated with neoadjuvant gemcitabine and cisplatin (GC, left), and those treated with neoadjuvant methotrexate, vinblastine, doxorubicin and cisplatin (MVAC, right). The 3 leftmost blue bands represent patients’ proportional preoperative clinical T-stage, and the rightmost 5 nodes represent the proportions of final surgical pathology following radical cystectomy. The flow between these 2 represents the number of patients experiencing downstaging (green), upstaging (red) and consistent staging (beige). The leftmost percentages describe the proportion of downstaging and upstaging within each clinical T-stage cohort. This figure is adapted from table 2 in a report by Zargar et al,² with the exclusion of any Tx and Nx patients (8).

In contrast, the adjuvant approach avoids delay to surgical therapy and provides pathological information that can inform the initiation of systemic treatment if high risk pathological features are present (pT3, pT4a or pN+), ergo providing chemotherapy in those patients who are most likely to benefit from it and curbing overtreatment. AC is used following up-front radical cystectomy and was shown in a recent meta-analysis to provide an absolute OS benefit of 6% at 5 years compared to surgery alone, and a 9% absolute OS benefit when adjusted for covariates, resulting in a number-needed-to-treat of ˜10–20 (fig. 2).⁷ It is inviting, though inappropriate, to equate the OS benefit across settings as the patient populations receiving NAC or AC are inherently different.

There is a selection bias in the studies from which these data are based. The majority of patients included in the meta-analyses investigating cisplatin-based NAC had a performance status of 0 or 1, had a creatinine clearance >50 ml per minute and were less than 70 years old.¹ Given that the average age of bladder cancer diagnosis is 73, the benefit seen in NAC trials may not extend across the entire bladder cancer patient population. The use of AC minimizes overtreatment by reserving therapy for those with high risk features on surgical pathology, and importantly only treats patients who survive surgery and are postoperatively fit, resulting in a survivorship bias. Therefore, OS comparison between NAC and AC must be interpreted cautiously.

In the adjuvant setting, immunotherapy in the form of nivolumab has been shown to increase disease-free survival (DFS) in patients who were at high risk for recurrence compared to placebo.⁸ Importantly, though, and in contrast to AC, adjuvant immunotherapy has not yet been shown to be a reasonable substitute for NAC. Bajorin et al noted in a subgroup analysis that DFS was only statistically significantly improved in patients who received neoadjuvant therapy in addition to adjuvant nivolumab.⁸ Also, DFS was higher among patients with tumor programmed death ligand 1 (PD-L1) expression level of 1% or more. Again, we see the theme of precision medicine: tailoring treatment to patient and tumor characteristics.

Figure 2. Forest plot (A) and Kaplan-Meier (B) curves (nonstratified) of the effect of adjuvant (Adj) chemotherapy (CT) on overall survival. In part A of figure, each trial is represented by a square, the center of which denotes the hazard (haz) ratio for that trial (comparison), with the horizontal lines showing the 95% and 99% confidence intervals. The size of the square is directly proportional to the amount of information contributed by the trial. The black diamond gives the pooled hazard ratio from the fixed-effect model; the center of this diamond denotes the hazard ratio and the extremities of the 95% CI. This figure and caption are from a study by the Advanced Bladder Cancer Meta-analysis Collaborators Group.⁷

Precisely selecting patients most likely to benefit from treatment will be the cornerstone of chemotherapy utilization in the future management of MIBC. Circulating tumor DNA (ctDNA) has shown early promise as a serum biomarker that has prognostic and predictive power at several points in the disease course and is suggestive of early metastatic disease–and thus may serve as a trigger for the initiation of NAC.⁹ In the adjuvant setting, patients with high-risk final pathology who are positive for ctDNA have been shown to derive survival benefit from atezolizumab versus observation, while this benefit did not exist for those patients who were ctDNA negative.¹⁰ ctDNA as a marker for molecular residual disease has the potential to dictate the initiation of chemotherapy and immunotherapy in the treatment of MIBC.

Meta-analysis data now underscore a choice between NAC and AC that depends on both clinical context and clinician and patient preference. Though comparison between options in terms of OS is limited due to bias, there is clear benefit in both cases, albeit at a cost of notable side effects. The burden is on the clinician to decide which patients merit which medicines as we await more advanced and accessible treatment stratification; until then, we remember the words of William Osler: “Medicine is a science of uncertainty and an art of probability.”