The Value of Multiparametric Magnetic Resonance Imaging Sequences to Assist in the Decision Making of Muscle-Invasive Bladder Cancer

Over the last few years, the landscape of bladder cancer (BC) management has profoundly changed, thanks to increased knowledge of disease biology and the identification of novel therapeutic approaches and biomarkers.¹ No more than 5 years ago, the treatment-decision process for non muscle-invasive BC (NMIBC) or muscle-invasive BC (MIBC) was represented by radical surgery in most cases, with an opportunity for perioperative systemic therapy in a few cases. To date, the diagnostic and therapeutic armamentarium has been exceedingly enlarged for these patients.

For Bacillus Calmette-Guérin (BCG)-unresponsive NMIBC, the United States Food and Drug Administration (FDA) has recently approved the use of systemic therapies like pembrolizumab,² and is reviewing the novel intravesical compounds like nadofaragene firadenovec,³ a non-replicating adenovirus vector harboring the human interferon (IFN)-alpha-2b gene. In MIBC, the new data have also been exiting, although still not practice-changing and only in the context of clinical trials. Several studies have tested the activity of single-agent immune-checkpoint inhibitors^4,5 or their combination with chemotherapy,⁶ and many more are currently enrolling patients. Indeed, the availability of genomic and molecular biomarkers, such as the molecular subtype classification,^7,8 fibroblast growth factor receptor-3⁹ (FGFR3), programmed cell death-ligand-1 (PD-L1) expression and tumor mutational burden (TMB),¹⁰ have provided investigators with a unique opportunity to tailor neoadjuvant treatments and shift the standard management. Based on these opportunities for either NMIBC and MIBC, we have proposed a new and more comprehensive definition of the early disease stages, not limited to the identification of muscle layer invasion. The use of early bladder cancer (EBC) definition,¹¹ which included selected high-risk NMIBC and limited-stage MIBC, has been proposed for use within clinical trials testing the activity of novel compounds.

However, in the early stages, defining bladder cancer clinical states requires the ability to correctly evaluate the tumor extension within the bladder wall and stage the lymph nodes. Regarding the former goal, the use of multiparametric magnetic resonance imaging (mpMRI) in early disease may be quite useful for either NMIBC and MIBC patients. In fact, mpMRI was proven to be reliable to find signs of muscle infiltration in patients with NMIBC,¹² and this aspect is important especially for avoiding multiple diagnostic assessments, which can be detrimental, in patients receiving a long-term course of treatments (e.g. pembrolizumab or nadofaragene firadenovec). On the other hand, mpMRI was also proven to be useful in MIBC, when it was tested for identifying signs of residual disease after neoadjuvant treatment. In this regard, mpMRI was extensively evaluated in the context of neoadjuvant pembrolizumab within the PURE-01 study.¹³ Here, combining morphologic T2-weighted (T2W) assessment, functional diffusion-weighted imaging (DWI), and dynamic contrast enhancement (DCE) assessments, we were able to achieve an accurate prediction of pathological complete response at radical cystectomy (pCR, i.e. ypT0ypN0) within the initial 82 patients. In our recently updated publication, we have further analyzed our data on a larger patient sample which included 143 patients, of whom 123 had suitable paired imaging assessments before and after pembrolizumab (N = 246 mpMRI in total).¹⁴ These numbers are unprecedented in the context of novel imaging applied to innovative perioperative treatments in MIBC. The area-under-the-curve (AUC) of the combination of all mpMRI sequences to predict the pCR was 0.74. The AUC of the combination of all sequences to predict major tumor response or downstaging to NMIBC (ypT1/a/isypN0) was 0.87. Therefore, these results largely confirmed our initial findings. Additionally, we also found that the prediction of ypT0ypN0 and ypT1/a/is ypN0 responses were equally achieved regardless of the use of DCE sequences. In fact, when excluding the DCE assessment, the AUC was 0.74 (95% CI: 0.66–0.82) for prediction of ypT0ypN0 and 0.87 (95% CI: 0.81–0.93) for ypT1/a/is ypN0 responses. Therefore, the evidence supported the notion that we could simplify the mpMRI procedure by avoiding the use of intravenous contrast without compromising the accuracy.

Indeed, delivering mpMRI prior to radical cystectomy in all men with MIBC, to assess the response to treatment, may be challenging in the context of routine clinical practice, for both technical and budgetary reasons. Removal of the DCE sequence would reduce the costs by omitting the contrast, and also by reducing the duration of the exam and avoiding the need for medical practitioner attendance. Finally, biparametric MRI is less invasive, shorter, and a more comfortable scan for patients and could allow many more patients to experience the benefits of such a noninvasive assessment after neoadjuvant therapy. Pending validation with data from multicenter studies and recognizing the exploratory and hypothesis-generating nature of the present results, we support future clinical trials aimed to validate the use of biparametric MRI to stage organ-confined bladder tumors and evaluate response after neoadjuvant treatments.

Written by: Marco Bandini, and Andrea Necchi, Unit of Urology, Division of Experimental Oncology, Urological Research Institute (URI), IRCCS Ospedale San Raffaele, Milan, Italy and Department of Medical Oncology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Published Date: August 13th, 2020

Read the Abstract

Login