MCM5 urine expression (ADXBLADDER) is a reliable biomarker of high-risk non- muscle-invasive bladder cancer recurrence: A prospective matched case-control study

Abstract

BACKGROUND:

Mini Chromosome Maintenance 5 (MCM5) is considered as a urinary biomarker of bladder cancer. ADXBLADDER is a commercially available test to detect MCM5 antibodies.

OBJECTIVE:

External validation of ADXBLADDER test as a urinary biomarker of histopathologically confirmed non-muscle invasive bladder cancer (NMIBC) recurrence.

METHODS:

The study enrolled 119 consecutive patients with a history of NMIBC and 37 healthy volunteers matched as controls. Single, full-void urine samples were collected from patients before cystoscopy $\pm$ TUR. To measure MCM5 expression, Arquer Diagnostics ADXBLADDER test was used. The study protocol was registered within the clinical trials database (NCT03796299).

RESULTS:

Among patients with NMIBC history, recurrence was diagnosed in 83 patients (69.7%). ADXBLADDER demonstrated sensitivity of 73.5% (95% confidence interval (CI) 62.7%–82.6%), specificity of 33.3% (95% CI 18.6% to 51%), overall negative predictive value (NPV) of 35.3% (95% CI 23.3% to 49.5%) and overall positive predictive value of 71.8% (95% CI 66.1% to 76.8%) for detecting recurrence. In a control group, false positive ADXBLADDER results were noticed in 18 patients (48.6%). The sensitivity and NPV were the highest in invasive tumors (100% and 100%, respectively) and in high-grade recurrences (81.8% and 94.1%, respectively).

CONCLUSIONS:

ADXBLADDER has a moderate sensitivity and poor specificity in detecting NMIBC recurrence. However, it properly diagnoses patients with T1$+$ stage recurrence or high-grade tumors.

1. Introduction

Management of patients with bladder cancer remains a challenge for clinicians. At the time of diagnosis about 78% patients present with non-muscle invasive bladder cancer (NMIBC) [1]. After transurethral resection of the tumor (TUR) followed by intravesical therapy, these patients face disease recurrence in 31–78% of cases at 5 years depending on individually calculated risk [2]. For this reason, NMIBC patients need long-term follow-up to early identify the relapse. Cystoscopy and urine cytology remain standard of care in monitoring and detecting NMIBC recurrence. However, cytology is characterized by high specificity (86%) and sensitivity (84%) only in high-grade cancers [3], while very low sensitivity for low-grade cancers makes it an unperfect tool [4]. Cystoscopy is invasive, expensive and may be problematic in some conditions (chronic urinary infection, neurogenic bladder). Moreover, a substantial percentage of patients (even 18.8%) receives surveillance based on cystoscopy and cytology that significantly deviates from guidelines for appropriate follow-up care [5, 6]. No matter if the source of this discrepancy is on the patients’ or urologists’ site, it may lead to suboptimal follow-up after NMIBC treatment. Therefore reliable biomarkers of NMIBC recurrence are needed.

MCM5 is one of proteins of the minichromosome maintenance family and is regarded as a novel biomarker of NMIBC [7, 8]. However, until lately the laboratory methods were complicated and not suitable for clinical use. Recently, a novel commercially available enzyme-linked immunosorbent assay (ELISA) to detect MCM5 (ADXBLADDER; Arquer Diagnostics Ltd, Sunderland, UK) was introduced and studied as a bladder cancer biomarker among patients presenting with hematuria [9].

The aim of our study was to externally validate the clinical value of MCM5 as a urinary biomarker of NMIBC recurrence in a comparison to histopathological findings.

2. Material and methods

The study was designed as a prospective cohort study and was approved by the Ethical Board of Medical University of Warsaw. The experiments were undertaken with the understanding and written consent of each subject. The study conforms with The Code of Ethics of the World Medical Association (Declaration of Helsinki). The study protocol was registered within the Clinical Trials database (NCT03796299).

2.1. Participants

The study enrolled 119 consecutive patients with a history of NMIBC. There were also 37 healthy volunteers enrolled into the study, who were patients with other urological disorders (such as benign prostatic hyperplasia, hydrocele testis, stress or urge urinary incontinence) matched regarding age, sex and smoking status as controls. All the patients gave informed consent and went through a structured interview. Exclusion criteria were as follows: urolithiasis, prostatitis, other malignancies within the urinary tract, active urinary tract infection. A single, full-void urine sample of at least 10 ml was collected from patients on the day before cystoscopy $\pm$ TUR. Pathological evaluation of the surgical specimen was performed in the standard manner.

2.2. Urine analysis

To measure MCM5 expression, we used Arquer Diagnostics ADXBLADDER test, a double-monoclonal antibody sandwich enzyme-linked immunoassay (ELISA). Full-void urine samples were collected and urine sediment was achieved by centrifugation for 5 minutes at 1500 g. ADXBLADDER lysis buffer was added and kept for 30 minutes at room temperature to allow complete lysis and then stored at $-$20${}^{\circ }$C. The ELISA was performed according to manufacturer protocol and MCM5 result was calculated based on formula: “(studied sample absorbance $-$ negative control)/positive control $\times$ 100%”. The result of $>$ 0.985 (producer’s value) was defined as positive. Every sample was tested twice. Urinalysis was performed as a clinical sample.

2.3. Statistical analysis

To test the normality of variables the Shapiro-Wilk test was used. Chi-square test was used to examine the association between diagnosis nominal variables, Mann-Whitney U test was used to evaluate the difference between quantitative variables. The results are presented as means $\pm$ standard deviation (SD) or medians (1${}^{\text{st}}$, 3${}^{\text{rd}}$ quartiles). Statistical analysis was performed using Statistica 13.1 software, StatSoft, USA.

3. Results

The histopathological results of the primary tumor of patient with a history of NMIBC were as follows: pTaLG – 80 patients, pTaHG 14 patients; pT1LG – 2 patients; pT1HG – 23 patients. Among 119 NMIBC cases, 83 patients had disease recurrence and underwent TUR (group 1) and 36 patients had no recurrence and underwent rigid cystoscopy under general anesthesia (group 2). Stage of recurrent tumors in group 1 was as follows: Ta in 73 patients, Tis in 4 patients, T1 in 5 patients and T2 in one patient. One patient was diagnosed with carcinosarcoma and was excluded from further analysis. Patient characteristics are summarized in Table 1.

Table 1.

Patient characteristics

	Controls	Group 1 $+$ Group 2	$p$-value (controls vs. Group 1 $+$ Group 2)	Group 1 (NMIBC recurrence)	Group 2 (no recurrence)	$p$-value (difference between all available groups)
Number of patients	37	119		83	36
Age	68.27 ($\pm$ 8.23)	70 (64; 78)	0.07${}^{*}$	71 (64; 80)	68 (63; 73.5)	0.1${}^{***}$
Sex	13F/24M	45F/77M	0.77${}^{**}$	32F/51M	13F/23M	0.93${}^{**}$
% active smokers	24.3%	31.9%	0.38${}^{**}$	32.5%	30.6%	0.66${}^{**}$
% ever smokers	59.4%	77.3%	0.054${}^{**}$	78.3%	75.0%	0.09${}^{**}$
Prior recurrence rate per year	N/A			0.8 (0.5; 1.4)	0.7 (0.5; 1.1)	0.55${}^{*}$
Number of TURs in the past	N/A			3 (1; 4)	2 (1; 4)	0.33${}^{*}$
% previous intravesical BCG-therapy	N/A			8.4%	25.0%	0.03${}^{**}$
Time from last cystoscopy	N/A			6 (5; 12) weeks	6 (5; 12) weeks	0.83${}^{*}$
Median follow-up time				46 (18–117) months	41 (22–72) months	0.69${}^{*}$

${}^{*}$Mann-Whitney U test. ${}^{**}$chi-square test. ${}^{***}$Kruskal-Wallis test.

ADXBLADDER demonstrated moderate sensitivity of 73.5% (95% confidence interval (CI) 62.7%–82.6%), specificity of 33.3% (95% CI 18.6% to 51%), overall negative predictive value (NPV) of 35.3% (95% CI 23.3% to 49.5%) and overall positive predictive value (PPV) of 71.8% (95% CI 66.1% to 76.8%) for detecting bladder cancer recurrence (Table 2). Subgroup analyses are presented in Tables 3 and 4. Of note, sensitivity and NPV to detect high-grade recurrence were 81.8% and 96.2% respectively. All the patients with stage $⩾$ T1 bladder cancer ($n=$ 6/6) and half of patients with carcinoma in situ ($n=$ 2/4) had a positive ADXBLADDER result (Table 3). Noteworthy is also the fact that ADXBLADDER sensitivity achieved 80% in patients with erythrocyturia, leukocyturia, bacteriuria and proteinuria, as well as in never smokers and women (Table 4).

Table 2.

A 2 $\times$ 2 table of ADXBLADDER performance among patients with a history of NMIBC

	NMIBC recurrence ($+$)	NMIBC recurrence ($-$)
ADXBLADDER ($+$)	61	24
ADXBLADDER ($-$)	22	12

Sensitivity of 73.5% (95% CI 62.7%–82.6%), Specificity of 33.3% (95% CI 18.6% to 51%), NPV of 35.3% (95% CI 23.3% to 49.5%), PPV of 71.8% (95% CI 66.1% to 76.8%).

Table 3.

Sensitivity and NPV of ADXBLADDER according to tumor grade and stage

Stage	Numer of cases	Number of positive ADXBLADDER tests	Sensitivity (95% CI)	NPV (95% CI)
Low-grade tumors	72	52	72.2% (60.4%–82.1%)	41.2% (28.3%–55.5%)
High-grade tumors (including CIS)	11	9	81.8% (48.2%–97.7%)	94.1% (81.5%–98.3%)
Carcinoma in situ	4	2	50% (6.8%–93.2%)	94.1% (85.2%–97.8%)
Stage Ta	73	53	72.6% (60.9%–82.4%)	41.2% (28.3%–55.4%)
Stage T1–2	6	6	100% (54.1%–100%)	100%

Table 4.

ADXBLADDER performance in follow-up of patients with NMIBC – subgroup analysis

Criterion	$n$	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)
Erythrocyturia	45	94.6% (81.8%–99.3%)	25% (3.2% to 65.1%)	85.4% (79.5%–89.8%)	50% (14.1%–85.9%)
Leukocyturia	31	95.7% (78.1%–99.9%)	37.5% (8.5%–75.5%)	81.5% (71.9%–88.3%)	75.0% (26.6%–96.1%)
Bacteriuria	49	86.5% (71.2%–95.5%)	41.7% (15.2%–72.3%)	82.1% (73.6%–88.2%)	50% (25.8%–74.2%)
Leukocyturia or bacteriaria	60	88.6% (75.4%–96.2%)	43.8% (19.8%–70.1%)	81.3% (73.5%–87.1%)	58.3% (34.1%–79.1%)
Erythrocyturia or leukocyturia	75	87.5% (75.9%–94.8%)	36.9% (16.3%–61.6%)	80.3% (74.1%–85.4%)	50.0% (28.7%–71.3%)
or bacteriaria
Proteinuria	41	88.5% (69.9%–97.6%)	40% (16.3%–67.7%)	71.9% (62.3%–79.8%)	66.7% (36.9% to 87.3%)
Never smokers	27	83.3% (58.6%–96.4%)	55.6% (21.2%–86.3%)	79% (63.7%–88.9%)	62.5% (33.7%–84.5%)
Smokers	38	74.1% (53.7%–88.9%)	18.2% (2.3%–51.8%)	69% (60.9%–76.1%)	22.2% (6.5%–53.8%)
Males	74	68.6% (54.1%–80.9%)	34.8% (16.4%–57.3%)	70% (62.2%–76.8%)	33.3% (20%–50%)
Females	45	81.3% (63.6%–92.8%)	30.8% (9.1%–61.4%	74.3% (66%–81.2%)	40% (18.3%–66.4%)

n: number of cases meeting the criterion.

In a control group, 18 patients (48.6%) had false positive results. Half of these patients were suffering from BPH, most of them had erythrocyturia.

4. Discussion

We have performed a prospective study evaluating the clinical accuracy of ADXBLADDER, a new commercial test based on MCM5 protein expression, for diagnosing NMIBC recurrence. While it is characterized by overall moderate accuracy, it has high sensitivity and NPV in groups of high-grade and invasive NMIBC recurrence.

Due to high recurrence rates, intensive surveillance strategies and treatment costs, the management of bladder cancer patients contributes significantly to medical costs [10]. Extensive studies regarding biomarkers may lead to limiting unnecessary interventions in NMIBC and reduce costs [11]. The perfect biomarker for bladder cancer should be non-invasive, available and easy to perform. Nevertheless, the main goal of developing biomarker should be the identification of a test which can improve clinical decision-making. Until now six urine tests have been approved by the US Food and Drug Administration (FDA) in the follow-up of patients with NMIBC (NMP22 test kit, NMP22 BladderChek Test, BTA TRAK, BTA stat, UroVysion, and uCyt $+$) [12]. However, so far none of the urine-based biomarker tests have been widely accepted for clinical use [13]. Recently two novel tests have been presented as potential tools which may help in diagnosis among NMIBC patients – Bladder EpiCheck and ADXBLADDER [9, 14].

Previously it has been described that detection of MCM5 protein in urine sediment may be a promising biomarker of bladder cancer [7, 8], but the research methods were complicated and impractical in clinical practice due to low availability. Recently, Dudderidge et al. reported a novel commercially available ELISA test to detect bladder cancer – ADXBLADDER [9]. In this study we focused on ADXBLADDER test in detecting NMIBC recurrence.

On the one hand, our results show sensitivity to detect recurrence similar to previously noted in detecting primary bladder cancer among patients with hematuria. Similarly to the previous reports the best performance regarding sensitivity was noted in invasive tumors [7, 8, 9]. On the other hand, the specificity for detection of NMIBC recurrence is much lower than previously reported by Dudderidge et al. [9] (33.3% vs. 68.4%). Due to the nature and the concept of our study, the values of NPV are much lower. Patients with NMIBC recurrence accounted for 69.7% of all participants in our study, while in the study by Dudderidge et al. only 8.6% of participants were diagnosed with bladder tumor, what greatly affects NPV values.

Interesting is the fact that the rate of false-negative results is the lowest among patients with NMIBC recurrence and abnormalities in urinalysis. Although this fact still needs to be confirmed, it can show an additional pathway for the clinical use, while most biomarkers tend to achieve worse results in patients with abnormalities in urinalysis.

In contrary to previously reported data [8, 9] our study shows high percentage of false positive results. It is possible that substantial number of these patients will develop NMIBC relapse in the short future (the anticipatory effect [15]). Such phenomenon has also beenobserved by Brems-Eskildsen et al. In their study, patients with and without NMIBC recurrence were tested for MCM5 mRNA transcript and 56% of the false positives had the recurrence shortly after urine sampling, thus ended up as true positive [16].

The main limitation of our study is limited number of patients with high grade and invasive tumors, in whom the performance of ADXBLADDER was the best. In contrary to previously published studies, we enrolled a group of patients with suspected bladder cancer recurrence to reliably report sensitivity and specificity of the test. However, this was at the cost of possible selection bias. Moreover, we have studied MCM5 expression at the time of suspected recurrence. It could be worth to measure the MCM5 expression at the time of diagnosis of the primary tumor, as well as after complete resection or before recurrence. The results of such study could help determine the true value of ADXBLADDER in follow-up for NMIBC recurrence.

Novel noninvasive biomarkers or panels of biomarkers detecting bladder cancer recurrence are still being sought [17, 18, 19]. ADXBLADDER as a solitary tool does not seem to be an accurate biomarker of NMIBC recurrence. However, it may be a promising adjunct to other methods. Nevertheless, evaluation of this potential role needs further studies.

5. Conclusions

ADXBLADDER has a moderate sensitivity and poor specificity in detecting NMIBC recurrence. However, our study supports its good diagnostic accuracy among patients with invasive or high-grade tumors, however it must be noticed that the number of patients with invasive tumor in our study was not high ADXBLADDER may be a promising adjunct to other diagnostic tools, however, based on our results it cannot replace cystoscopy nor become a primary test for diagnosis of NMIBC recurrence.