Circulating tumor DNA based minimal residual disease detection and adjuvant treatment decision-making for muscle-invasive bladder cancer guided by modern clinical trials

Abstract

Up to 430,000 cases of bladder cancer are diagnosed each year worldwide. A proposed method for non-invasive monitoring has been to utilize a “liquid biopsy.” Liquid biopsy has been proposed as a non-invasive method of testing biomarkers in bodily fluids in order to detect and survey cancer. The liquid biopsy could be utilized to obtain information regarding circulating tumor cells, circulating cell-free tumor DNA, circulating cell-free tumor RNA, and more. It is currently being investigated to help guide adjuvant therapy and improve oncological outcomes. We highlight an array of exciting past and ongoing clinical trials regarding ctDNA and adjuvant therapy in regard to urothelial carcinoma which we believe to be amongst the leaders in the field.

Introduction

Bladder cancer remains a significant health problem. Up to 430,000 cases of bladder cancer are diagnosed each year worldwide [1]. The initial diagnosis and surveillance of bladder cancer usually requires a combination of cystoscopy, urine cytology, biopsy, and cross-sectional imaging. These methods can be costly and invasive, with varying specificity and sensitivity, as well as inter-observer variability [2]. There is a need for non-invasive methods of detection and surveillance. Increasing efforts have been made to further the knowledge and development of non-invasive methods to diagnose and monitor the recurrence or progression of cancer.

Bladder cancer often displays significant heterogeneity at the genomic, transcriptional, and cellular levels. Despite standard of care neoadjuvant chemotherapy, heterogeneity contributes to drug resistance and relapse after therapy, resulting in poor survival outcomes [3]. These poor outcomes are exemplified by a 5-year recurrence-free survival for muscle-invasive bladder cancer (MIBC) of only approximately 50%, despite aggressive treatment [4]. Therefore, it is imperative to identify these patients who are at high risk of relapse, as the addition of adjuvant therapy could ultimately prevent death from disease.

There has been interest in understanding urothelial carcinoma biology to thus guide future and novel therapies. By molecularly characterizing MIBC, essential pathways are decrypted and novel targets for therapy are identified to guide personalized treatment. However, bladder cancer targeted therapy is still in the early stages of clinical research. The cancer genome atlas (TCGA) analyzed DNA from 131 patients with MIBC and identified 32 mutant genes with high frequency, and showed these mutations to occur mostly in the cell cycle, chromatin regulation, and kinase signaling pathways [5]. A few examples include the FGFR3/2 mutation, as well as the PD-1 mutation. A notable recent target for advanced urothelial carcinoma is the FGFR3/2 mutation, with which clinical trials showed the overall response rate for the drug erdafitinib to be 49% in patients with FGFR3 mutations [[6], [7], [8]].

Urothelial carcinoma is known to have high tumor heterogeneity and high mutational burden, which can change cell differentiation with each cell division [9]. In fact, it has one of the highest mutation rates among cancers, which leads to a higher degrees of both intertumor, or between different patients, and intratumor, or different tumor sites within one patient, heterogeneity [10].

There have been studies that have shown consequences of intratumor heterogeneity (ITH) driving neoplastic progression and therapeutic resistance. A pan cancer analysis of 1165 tumors from 12 cancer subtypes, including MIBC, was demonstrated by Andor et al., which showed that 86% of all tumors had at least 2 clones, and the risk of mortality increased when more than 2 clonal populations existed, but decreased again with more than 4 clones, implying that a higher degree of genomic instability was deleterious for tumor growth, or advantageous for treatment response [11]. Another study by Lamy et al. analyzed 29 paired metachronous bladder tumors from patients who started with non-muscle invasive disease and found shared mutations between all matched pairs, suggesting that the tumors were clonal in origin [9]. ITH was shown to be greater in tumors that progressed, suggesting that the acquisition of additional genomic alterations contributed to progression.

However, while this heterogeneity can at times be an obstacle to treatment, it can also be advantageous for detection via ctDNA-based assays. Among all cancer types, bladder cancer has amongst the highest rates of non-silent mutations in cancer-associated consensus genes. Previous studies have demonstrated that >95% of bladder cancer samples harbor non-silent mutations in at least one consensus gene [12,13]. This unique characteristic opens the door to cost-effective and reliable detection through the use of ctDNA technologies. Technologies which are optimized for sensitive detection of disease via modification of ctDNA thresholds [14].

Following the administration of standard of care neoadjuvant cisplatin-based chemotherapy for MIBC, there is sometimes a pattern of tumor response followed by recurrence and progression. This chemotherapy regiment can select for pre-existing tumor subclones resistant to chemotherapy. This was further explored by Faltas et al. who examined ITH in MIBC with whole exome sequencing (WES) in 72 urothelial tumors from multiple sites and time points in 32 individual patients [15]. This study compared the genomes of primary and metastatic tumor sites before and after chemotherapy. They showed that there was no significant difference in the number of mutations between pre and post chemotherapy tumors, however, only 28% of the mutations were shared between matching pre and post chemotherapy samples. Furthermore, even mutations in key known driver genes were not shared between pre and post chemotherapy samples, and sometimes the same genes that were mutated prechemotherapy had new mutations post chemotherapy. This suggests that there could be a significant shift in the mutational genome of urothelial carcinoma along the timeline of treatment with chemotherapy. Specifically, they found that the combined frequency of copy number alterations and mutations in ATM, RB1, or FANCC in 73% of cases prechemotherapy and 38% post chemotherapy, and these chemo sensitive mutant clones were replaced by resistant subclones with wild type ATM, RB1, and FANCC.

There have been multiple techniques that have been utilized to evaluate the liquid biopsy in bladder cancer, including droplet digital polymerase chain reaction (ddPCR), beads, emulsion, amplification, and magnetics (BEAMing), tagged-amplicon deep sequencing (TAm-Seq), urine cancer personalized profiling by deep sequencing (uCAPP-Seq), whole genome bisulfite sequencing (WGBS-Seq), whole exome sequencing (WES), and shallow whole genome sequencing (sWGS) [[16], [17], [18], [19], [20], [21]].

Liquid biopsy has been proposed as a non-invasive method of testing biomarkers in bodily fluids in order to detect and survey cancer. This could potentially help reduce or avoid invasive procedures while providing relevant clinical information [[22], [23], [24]]. It could also be utilized to obtain information regarding circulating tumor cells, circulating cell-free tumor DNA (cfDNA), circulating cell-free tumor RNA, and more [25]. Bladder cancer is unique in the sense that cells and cfDNA shed from tumors can be readily accessed and trended temporally through the testing of easily acquirable urine samples.

Circulating tumor DNA (ctDNA) has been shown to be highly prognostic. Christensen et al. demonstrated that ctDNA testing in patients with bladder cancer who undergo chemotherapy and cystectomy to be highly sensitive and specific for early risk stratification of patients, prediction of treatment response, and early detection of metastatic relapse [26]. Performance of ctDNA testing varies tremendously based on the timepoint at which samples are collected. Analysis of samples collected after neoadjuvant chemotherapy (NAC), prior to cystectomy, detected residual disease with sensitivities and specificities of 50% and 96% respectively.  When pooled across all sample timepoints, ctDNA analysis correctly identified all patients with metastatic relapse during disease monitoring (100% sensitivity, 98% specificity) [26]. Similarly, Birkenkamp-Demtröder et al. showed that patients with metastatic relapse or disease progression had significantly higher ctDNA levels after cystectomy than those who remained disease free (p < 0.001) [27].

In addition to the prognostic utility of ctDNA, sequencing efforts can offer key insights into the tumor mutational burden (TMB) of a patient's cancer. Urothelial carcinoma patients with high TMB have been reported to benefit substantially from treatments such as immune checkpoint inhibitors such as pembrolizumab [28].

The field of ctDNA is fast moving. Here, we highlight an array of exciting past and ongoing clinical trials regarding ctDNA and adjuvant therapy in regard to urothelial carcinoma which we believe to be amongst the leaders in the field.

Completed trials

IMVigor010

Original study and analysis

IMVigor010 was a randomized phase 3 trial that investigated adjuvant atezolizumab versus observation alone in patients with urothelial cancer following radical surgery [29]. This trial included both bladder cancer patients as well as those with upper tract disease and primarily assessed disease-free survival. Atezolizumab is a monoclonal antibody that acts as an inhibitor of PD-L1. It has previously shown efficacy in treating several different cancer types [[30], [31], [32]].

Patients that were ypT2-T4a or pT3–4a were included in the study. Those that had received adjuvant radiotherapy or prior adjuvant chemotherapy were excluded. Patients (n = 809) were randomized (1:1) to receive 16 cycles of atezolizumab or observation. An intention-to-treat analysis was performed to assess their primary endpoint of disease-free survival.

There was no improvement found in disease-free survival as a result of atezolizumab therapy. Additionally, there were higher rates of adverse events leading to discontinuation than what has been reported in previous studies. Follow-up at the time was insufficient to adequately assess overall survival.

Post-hoc analysis

While IMvigor010 did not detect any significant benefit from adjuvant atezolizumab within their full unselected cohort, the authors were aware that there were likely subgroups that may have experienced tangible benefit from adjuvant immunotherapy. Powles et al. were originally interested in investigating ctDNA at several time points and whether it correlated with a patient's clinical response to atezolizumab [33].

Signatera™ (Natera, Inc; Austin, Texas, USA), an assay based on personalized amplicon-based deep NGS tracking of 16 tumor-informed targets, was used to determine ctDNA positivity. Subgroup analyses were performed on patients that were ctDNA positive, revealing significant improvements in both disease-free survival and overall survival as a result of atezolizumab therapy when compared to observation (DFS HR = 0.58; OS HR = 0.59).  Given the associations between ctDNA levels with MRD and disease recurrence, stratification based on ctDNA positivity selected for patients that were most likely to harbor residual disease post-operatively. Those with residual disease by ctDNA analysis, in turn, benefitted the most from adjuvant therapy. No survival differences were detected between atezolizumab and observation in the group of patients that were ctDNA negative.

The authors additionally examined rates of ctDNA clearance. This was defined as the proportion of patients that were ctDNA positive at cycle 1 day 1 that had subsequently converted to become ctDNA negative by the time of cycle 3 day 1. ctDNA clearance occurred in 18.2% of the atezolizumab arm compared to only 3.8% in the observation arm. Patients that were shown to have achieved ctDNA clearance were demonstrated to have superior survival compared to those that remained ctDNA positive by cycle 3 day 1.

In terms of transcriptional analysis, tumors from patients found to be ctDNA positive were more likely to be enriched in genes associated with the cell cycle and angiogenesis. These enrichments have been demonstrated to correspond with more aggressive cancer phenotypes [34,35]. This highlights yet another underpinning between ctDNA positivity and tumor biology.

Overall, this study serves to demonstrate a role for liquid biopsy and ctDNA analysis within the clinical context. Stratification based on ctDNA positivity revealed a subgroup of patients that directly benefited from atezolizumab therapy. This further underscores the utility of these methods for the purposes of patient stratification and treatment personalization.

Ongoing trials

IMvigor011

An ongoing clinical trial that is generating great enthusiasm is IMvigor011 (NCT04660344) which follows nicely from IMvigor010 and Powles’ analysis [36]. The study design of IMvigor011 was presented at the 2021 European Society for Medical Oncology congress. This trial is prospectively studying adjuvant atezolizumab vs placebo in patients with high-risk MIBC who are ctDNA+ post-cystectomy. The plan is to enroll 405 MIBC patients within 6–14 weeks after undergoing cystectomy. All patients must have undergone lymph node dissection, tumor sampling for PD-L1 testing (VENTANA SP142 IHC assay) and whole exome sequencing.

The inclusion criteria is ypT2–4a or ypN+, M0 tumors for patients after neoadjuvant chemotherapy, and pT3–4a or N+, M0 tumors for patients not given neoadjuvant chemotherapy who decline adjuvant chemotherapy or are cisplatin ineligible. Patients will enter a surveillance phase where they will undergo testing for plasma ctDNA positivity every 6 weeks up to 36 weeks and every 12 weeks up to 21 months. If they test positive for ctDNA (>= 2 mutations), the patients who are ECOG performance status <=2 with no radiographic recurrence will undergo randomization 2:1 to atezolizumab or placebo every 4 weeks for 12 cycles, or up to 1 year, or until recurrence, or toxicity. The primary endpoint is disease-free survival for patients who are ctDNA positive <= 20 weeks after cystectomy. Secondary endpoints include overall survival, and disease-free survival for patients with ctDNA positivity >= 20 weeks after cystectomy [37]. The study will also assess safety and exploratory predictive, prognostic, and pharmacodynamic biomarkers. Exclusion criteria include adjuvant chemotherapy, as well as postsurgical radiation.

TOMBOLA

Out of Denmark, an exciting trial is currently underway. The Treatment of Metastatic Bladder Cancer at the Time of Biochemical reLApse Following Radical Cystectomy (TOMBOLA) trial seeks to continue to explore the question of whether ctDNA positivity benefits from adjuvant immunotherapy. The TOMBOLA trial (NCT04138628) is a non-randomized, open-label, single-armed, phase II interventional clinical trial study. The hypothesis is that early treatment at the time of detection of ctDNA biochemical relapse with atezolizumab will improve overall survival. ctDNA positivity is determined by utilizing four different tumor & patient-specific digital droplet PCR assays targeting single nucleotide variants (SNVs) and indels that are absent in germline samples.

The plan is to enroll 282 patients with cT2–4a urothelial carcinoma (including subtypes) who received neoadjuvant chemotherapy followed by radical cystectomy. Patients will undergo surveillance plasma blood tests for ctDNA. The study drug, atezulizumab, will be initiated within 28 days of detection of plasma ctDNA. It will be given systemically every third week for 12 months, or until progression. Complete response, overall survival, cancer-specific survival, as well as other prognostic biomarkers will be measured throughout the study. The trial's endpoint is complete response, defined as negative findings on both ctDNA and imaging in the post-treatment setting.

BISCAY

Another interesting biomarker-driven clinical trial targeting invasive urothelial carcinoma is the BISCAY trial (NCT02546661). It is a phase 1b open label multicenter trial sponsored by AstraZeneca that investigates durvalumab treatment in patients with MIBC that have progressed on prior therapies. Patients are randomized to receive additional treatments based on genomic alterations detected in ctDNA. For example, those that are detected to harbor fibroblast growth factor receptor mutations are sorted into a AZD4547 treatment arm in addition to durvalumab, while those with homologous recombination repair pathway mutations receive olaparib in addition to durvalumab.

As a phase 1b trial, the primary endpoints are geared towards investigating the safety and side effect profiles of the drugs and drug combinations included. Secondarily, the trial is attempting to examine outcomes such as progression-free survival, disease-control rate, as well as objective response rate. The research team published preliminary results in 2021, reporting 135 patients had been allocated into six study arms. Response rates in study arms ranged from 9 to 36% and further development was halted due to failure to meet efficacy criteria [[38], [39]].

Although overall survival and progression-free survival did not significantly differ between the durvalumab monotherapy and any of the combinatorial arms (1-year OS rate 56% vs. 42%), the authors did find that changes in fibroblast growth factor receptor mutation status, detected by serial ctDNA samples, correlated with clinical outcomes. This suggests a potential oncogenomic role of non-invasive serial ctDNA sampling to further guide bladder cancer management.

Clinical vignette

Below we describe a vignette which describes our experience with incorporation into a real clinical scenario.

A 72-year-old patient presented with gross hematuria, and was found to have a large bladder tumor. Transurethral resection of the bladder tumor revealed at least muscle invasive urothelial carcinoma. She underwent neoadjuvant chemotherapy but was only able to tolerate one cycle out of the recommended four cycles. She then underwent anterior pelvic exenteration, with final pathology revealing ypT3bN0M0, urothelial carcinoma with sarcomatoid variant.

Due to significant postoperative weakness and deconditioning, she was unable to undergo adjuvant therapy. She underwent Signatera™ testing and standard of care imaging four months post op and this test was positive at 3.14 mean tumor molecules per ml, as well as a 1.5 cm lesion highly concerning for liver metastases. She was begun on immunotherapy. On three subsequent MRD tests, her mean tumor molecules decreased, as shown in Fig. 1. Her subsequent follow up imaging now shows her liver lesion has significantly decreased by more than half the size, correlating strongly with the decline in ctDNA observed on Signatera. For this patient, the down-trending ctDNA on Signatera in combination with a decline in tumor burden on imaging informed the decision to end targeted therapy and directly impacted future follow-up.

Fig. 1.

Real world clinical vignette.

This clinical example brings up exciting possibilities. First, the patient's Signatera™ testing was drawn using a mobile phlebotomy station, highlighting improved convenience. Previous studies have indicated that ctDNA changes often precede observable changes on imaging [40]. Given the positive ctDNA assay and positive imaging findings, a negative ctDNA assay could potentially lead to less frequent imaging and change the standard of care. Lastly, given the positive imaging findings and ctDNA assay, an invasive and costly confirmatory biopsy may not be necessary.

Future directions

We have described both previous and current clinical trials that are prioritizing the use of early detection of micro-metastatic disease relapse, which eludes traditional imaging (Table 1). Additional trials are needed to continue to build upon the data provided by ctDNA for the early detection of recurrence, which can prompt the initiation of personalized targeted therapies.

Table 1.

An overview of the studies described.

Study/Trial	Year	Type	ctDNA Assay	Intervention(s)	Outcome Measure
IMVigor010 (n = 809)	2021	RCT	N/A	Atezolizumab	DFS
Powles et al. (n = 581)	2021	Post-hoc	Signatera	Atezolizumab	DFS
IMVigor011 (n = 495)	2021	RCT	Signatera	Atezolizumab	DFS
TOMBOLA	Ongoing	NRS	–	Atezolizumab	Complete response, OS, CSS
BISCAY	Ongoing	NRS	–	AZD4547, Durvalumab	Adverse event rates, PFS, OS
RCT - randomized controlled trial; DFS - disease-free survival; NRS - non-randomized study; OS - overall survival; CSS - cancer specific survival; PFS - progression-free survival

The ultimate goal of treatment would be to identify patients who would benefit from adjuvant therapy and help improve their outcomes, and thus avoid overtreatment and unnecessary toxicities. Many patients receive toxic treatments with little therapeutic benefit in part due to the limitations of conventional imaging in detecting micro-metastatic disease. Improved development and implementation of commercial liquid biopsy platforms for ctDNA minimal residual disease detection can help improve the accuracy of adjuvant oncologic treatments.

The role of neoadjuvant chemotherapy has been well established, however, despite aggressive multimodal treatment, patients with MIBC still have a high rate of disease relapse. Additional adjuvant therapeutic options are needed as is further treatment precision. Early detection of micro-metastatic relapse using modern ctDNA technology is a promising route to trigger personalized intervention with adjuvant therapy.

In recent years, ctDNA-informed approaches have been increasingly utilized in the treatment of other malignancies such as colorectal cancer. Specifically, post-operative ctDNA results have been applied to the decision-making process for adjuvant therapy. DYNAMIC and CIRCULATE are two trials investigating these applications for stage II colorectal cancer [41,42]. While CIRCULATE is still ongoing, results have been published for the DYNAMIC trial. Patients were randomized into either ctDNA-guided vs. standard management for adjuvant chemotherapy decision making. The ctDNA-guided group had a lower percentage of patients that were treated with adjuvant chemotherapy with no decrease in 2-year recurrence-free survival. Therefore, by sparing ctDNA negative patients from receiving chemotherapy, the potential for overtreatment was mitigated without introducing significant risk to patients oncological outcomes.

In addition to improving patient outcomes, this technology has the potential to decrease patient burden in regards to follow up. The outpatient surveillance schedules for multiple cancers have been well documented to be rigorous and onerous [43,44], which can lead to decreased patient compliance, increased radiation exposure, and increased cost. These ctDNA blood tests have the potential to identify disease recurrence and decrease the amount of imaging tests, in-office visits, and potentially obviate the need for confirmatory biopsies of disease recurrence. In fact, there are already mobile phlebotomy programs implemented in communities to help address these issues.

Tri-modal therapy is a treatment modality for MIBC that involves maximal transurethral resection of bladder tumor, followed by radiation and concomitant chemotherapy. The ultimate goal of this type of treatment is bladder preservation without sacrificing disease control. There are challenges to monitor these patients after treatment, given that their bladder may still harbor micro-metastatic disease that may propagate. There is an exciting prospective biospecimen repository that is currently open for accrual (NCT05630131) that is collecting plasma ctDNA on these patients to predict tumor response. Incorporation of ctDNA into this type of treatment could result in more reliable disease monitoring, improved patient selection, and less operative morbidity.

Although there have been advances in sequencing technology for the detection of circulating tumor DNA, it is still an expensive test that is not readily available to most patients. In general, these tests and protocols are available only at high volume academic centers. Standardization by commercial and academic partners is going to be necessary to maintain a high level of consistency among assay results. This will lead to standardization of protocols for preanalytical variables (e.g. sample collection, storage, processing, and reporting). However, these tests still remain imperfect and there are always risks of false negative results. A positive test does not indicate a clear answer on the next step in therapy, or how the patient should be counseled.

In conclusion, ctDNA incorporation into treatment plans continues to rapidly evolve. As the field progresses, we believe liquid biopsy will become integral in dramatically improving personalization and ultimately patient outcomes (Fig. 2).

Fig. 2.

A figure showing the potential for personalized medicine.

CRediT authorship contribution statement

Cayce Nawaf: Conceptualization, Writing – original draft, Writing – review & editing. Alexander Shiang: Conceptualization, Writing – original draft, Writing – review & editing. Pradeep S. Chauhan: Conceptualization, Writing – original draft, Writing – review & editing. Aadel A. Chaudhuri: Conceptualization, Writing – original draft, Writing – review & editing, Project administration, Resources, Software, Supervision. Gautum Agarwal: Conceptualization, Writing – original draft, Writing – review & editing, Project administration, Resources, Software, Supervision. Zachary L. Smith: Conceptualization, Writing – original draft, Writing – review & editing, Project administration, Resources, Software, Supervision.

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: A.A.C. has patent filings related to cancer biomarkers, and has licensed technology to Droplet Biosciences, Tempus Labs and to Biocognitive Labs. A.A.C. has served as a consultant/advisor to Roche, Tempus, Geneoscopy, NuProbe, Illumina, Daiichi Sankyo, AstraZeneca, AlphaSights, DeciBio, and Guidepoint. A.A.C. has received honoraria from Roche, Foundation Medicine, and Dava Oncology. A.A.C. has stock options in Geneoscopy, research support from Roche, Illumina and Tempus Labs, and ownership interests in Droplet Biosciences and LiquidCell Dx.

G.A. is a consultant for Signatera