Bladder cancer risk stratification with the Oncuria 10-plex bead-based urinalysis assay using three different Luminex xMAP instrumentation platforms

Background: No single marker of bladder cancer (BC) exists in urine samples with sufficient accuracy for disease diagnosis and treatment monitoring. The multiplex Oncuria BC assay noninvasively quantifies the concentration of 10 protein analytes in voided urine samples to quickly generate a unique molecular profile with proven BC diagnostic and treatment-tracking utility. Test adoption by diagnostic and research laboratories mandates reliably reproducible assay performance across a variety of instrumentation platforms used in different laboratories. Methods: We compared the performance of the clinically validated Oncuria BC multiplex immunoassay when data output was generated on three different analyzer systems. Voided urine samples from 36 subjects (18 with BC and 18 Controls) were reacted with Oncuria test reagents in three 96-well microtiter plates on Day 1, and consecutively evaluated on the LED/image-based MagPix, and laser/flow based Luminex 200 and FlexMap 3D (all xMAP instruments from Luminex Corp., Austin, TX) on Day 2. The BC assay uses magnetic bead-based fluorescence technology (xMAP, Multi-analyte profiling; Luminex) to simultaneously quantify 10 protein analytes in urine specimens [i.e., angiogenin (ANG), apolipoprotein E (ApoE), carbonic anhydrase IX (CA9), CXCL8/interleukin-8 (IL-8), matrix metalloproteinase-9 (MMP-9), matrix metalloproteinase-10 (MMP-10), serpin A1/alpha-1 anti-trypsin (A1AT), serpin E1/plasminogen activator inhibitor-1 (PAI-1), CD138/syndecan-1 (SDC1), and vascular endothelial growth factor-A (VEGF-A)]. Results: All three platforms categorized all 10 analytes in identical samples at nearly identical concentrations, with variance across systems typically <5%. While the most contemporary instrument, the FlexMap 3D, output higher raw fluorescence values than the two comparator systems, standard curve slopes and analyte concentrations determined in urine samples were concordant across all three units. Forty-four percent of BC samples registered ≥1 analyte above the highest standard concentration, i.e., A1AT (n=7/18), IL-8 (n=5), and/or ANG (n=2). In Controls, A1AT was higher in one sample. Conclusion: Multiplex BC assays generate detailed molecular signatures useful for identifying BC, predicting treatment esponsiveness, and tracking disease progression and recurrence. The similar performance of the Oncuria assay across three different analyzer systems supports test adaptation by clinical and research laboratories using existing xMAP platforms. Trial Registration: This study was registered at ClinicalTrials.gov as NCT04564781, NCT03193528, NCT03193541, and NCT03193515.


INTRODUCTION
Bladder cancer (BC) is the second most common urogenital malignancy, and is the sixth most common cancer in men (5% of all cancers excluding non-melanoma skin cancer) and the 17th most common cancer in women (1.5% of cases). 1 Of the 85,000 annual BC diagnoses in the USA, 2 ≈75% will be nonmuscle-invasive disease (NMIBC) that require years-long monitoring for recurrence and progression after undergoing initial transurethral resection and/or Bacillus Calmette-Guerin (BCG) therapy.Cystoscopy and voided urine cytology remain the gold standards for evaluating BC status. 3,4Cystoscopy is uncomfortable and invasive, and carries signi cant costs and risks (e.g., infection, trauma).Voided urine cytology is noninvasive and economical and has high speci city for BC but also has suboptimal sensitivity, especially with low-grade and early-stage tumors. 5Biological marker evaluation in urine samples has evolved as a noninvasive means to more effectively identify BC, stratify patient risk, and monitor treatment progress. 67][8][9] Additionally, the levels of certain individual protein-based markers (e.g., nuclear matrix protein 22, NMP22, and bladder tumor antigen, BTA) are increased in urine in scenarios such as in ammation unrelated to BC, 10,11 which can lead to false-positive interpretations.Evaluating a single BC biomarker in urine samples may be a useful adjunctive test for con rming ndings by cystoscopy and histology, but remains insu cient for primary diagnosis and treatment planning. 3,4ltiplex assays that simultaneously evaluate diverse BC biomarkers in urine increases the likelihood of correctly identifying neoplasms of variable etiology and presentation, predicting treatment response, and accurately tracking therapy effectiveness. 5These noninvasive approaches generate comprehensive patient-speci c BC molecular pro les that can better inform diagnosis and personalized treatment planning, ultimately resulting in improved outcomes. 53][14][15] Biomarker levels are converted into composite risk scores using differently-weighted algorithms tailored for either assisting BC diagnosis, predicting response to BCG therapy in early-stage intermediate to high-risk disease, or tracking treatment progress.The assay is CE marked in Europe and was assigned FDA Breakthrough Device status for expedited review in the USA. 16The current study compared assay performance and output when urine samples were evaluated with the Oncuria assay using three different uorescenceanalyzing instruments commonly used in diagnostic laboratories worldwide.

MATERIALS AND METHODS
Subjects and Urine Samples.Subjects included 18 individuals bearing BC (17 de novo and 1 recurrent) and 18 nonBC controls (15 with voiding dysfunction/hematuria and 3 with a history of BC on surveillance).Data are reported according to PROBE criteria. 17Exclusion criteria were a history of renal insu ciency (i.e., glomerular ltration rate < 60 mL/min) and/or reduced urinary creatinine (< 40 mg/dL), because these conditions can cause proteinuria that can interfere with protein immunoassays.Midstream voided urine samples that had been collected for cytology were centrifuged at 1,000×g for 10 min, with supernatants frozen and undergoing only one freeze-thaw cycle before multiplex analysis.This study received approval and a waiver of consent to use previously banked de-identi ed urine samples from the Cedars-Sinai Medical Center Institutional Review Board, Los Angeles, CA (IRB #00001459).Study performance complied with the tenets of the Declaration of Helsinki.). 18Within a single sample, Oncuria simultaneously captures the 10 analytes using a pool of 10 distinct 6.5-µm magnetic bead + antibody sets, with each bead set differentiated by a unique internal uorescent label.Beads are recovered, identi ed, and their captured target antigens quanti ed by ow cytometry analysis.Oncuria is in clinical trials to support FDA approval as an in vitro diagnostic test for predicting BCG response in patients with BC (Oncuria-Predict), 19 for detecting de novo BC in patients with hematuria (Oncuria-Detect) 20,21 for detecting recurrent BC in patients with a history of BC (Oncuria-Monitor) 22 .In a recent clinical validation studies to detect de novo BC, the assay demonstrated an Area Under Receiver Operating Curve, AUROC, value of 0.95 (95% CI 0.90-1.00),with 93% speci city and 93% sensitivity, and PPV of 0.65 and NPV of 0.99 (Table 1). 12In a pilot study to predict responsiveness to intravesical BCG therapy for the treatment of NMIBC, the assay demonstrated an AUROC value of 0.89 (95% CI: 0.80-0.99),with a test sensitivity of 82% and a speci city of 85%. 13 xMAP Instrumentation.Experimental Overview.Voided urine samples were passively thawed at 4°C and centrifuged at 15,000 × g for 10 min at 4°C to remove potential particulates.Samples, standards, and controls (50 µL/well) were added to a 96-well plate in duplicate wells per condition.Standards comprised a pool of the 10 analytes, from which a sevenpoint 3-fold dilution series was created that covered the dynamic range (> 3-log) of every analyte.Bead + sample/standard incubation was performed on Day 1 using three 96-well plates (one for each instrument), and the reaction plates were covered with an adhesive aluminum foil seal and stored overnight in the dark at 4°C.On the morning of Day 2, the FlexMap-delegated plate was warmed to ambient temperature (18-22°C).Beads were immobilized by placing the plate on a magnetic separator for 2 min followed by wash buffer aspiration, and beads were then resuspended in 150 µL fresh wash buffer, shaken for 2 min to assure uniform distribution, and then assayed on the FlexMap 3D instrument.At midday and late afternoon of Day

RESULTS
Analyte Detection Ranges.The dynamic range of quanti cation (lowest to highest standard concentration) for the 10 analytes are shown in Fig. 1.The most sensitive of the 10 concurrently performed assays was for CA9, with a lower detection limit of 1.4 pg/mL.The greatest upper detection limit was for A1AT, at 185,250 pg/mL.
-INSERT FIGURE 1 HERE-Subject Characteristics.Urine samples were obtained from 18 subjects with a BC diagnosis and 18 control subjects who presented for a voiding condition (Table 2).Most participants were aged ≥ 65 years (53%) and were male (94%).Of the 18 individuals with BC, 50% had Stage T2-4 disease and 89% had high-grade neoplasms.3).This is due to differences in the optical platforms used in the different instruments, and does not impact analyte concentration determinations.Shown are raw data outputs from six representative urine samples, from three con rmed BC subjects ("Tumor") and three Control subjects.Values are averages of duplicate wells per analyte, per subject, rounded to the nearest whole number.
Biomarker Quanti cation by Instrument.The calculated concentration of all 10 analytes was very similar across all three instruments, in 100% (36/36) of urine samples (Table 4, Fig. 2, Supplemental Table S1).Although there were statistical discrepancies in mean protein concentrations determined across instruments for three biomarkers (Table 5).For example, MMP-9 concentrations determinations were mathematically different between the MagPix and 200 instruments, but the discrepancy was only ≈ 5%.The four other statistically signi cant mismatches had even lesser percentages differences between mean analyte concentrations, i.e., 2.2-4.4%.
While this report is intended to demonstrate assay reproducibility across different xMAP instruments and not for clinical validation, noteworthy elevations of analytes were not noted in BC versus Control urine samples (Table 4).Values Exceeding Dynamic Range.All three instruments captured and de ned all 10 biomarkers at or below their highest analyte-speci c standard curve concentration in nearly all urine samples (Supplemental Table S1).Of the BC samples, 44% (8/18) registered at least one analyte above the highest standard curve concentration, observed with A1AT (n = 7/18), IL-8 (n = 5) and/or ANG (n = 2); in Control samples, a single instance was observed of a biomarker (i.e., A1AT) exceeding the assay's dynamic range.In 11/15 instances, the dynamic range was exceeded with all 3 instruments (Supplemental Table S1).
The slopes of standard curves generated by all three instruments were essentially identical at all points in the assay range, for all analytes (not shown).

DISCUSSION
This study con rmed reproducible assay performance when voided urine samples were interrogated by the Oncuria multiplex BC assay, with very similar data output obtained from three different xMAP ow cytometry analysis instruments commonly used in diagnostic and research laboratories.3][14][15] The current demonstration that the concentrations of all 10 BC biomarkers were adjudicated nearly identically across three instrument platforms indicates that the Oncuria assay is highly amenable to standardization across laboratories that use different xMAP systems. 24ltiplex assays that evaluate a composite molecular signature in urine have greater utility in detecting and monitoring BC than efforts to identify a single BC biomarker. 6,7,10,11Advantages of multiplex immunoassays include increased e ciency and lower costs versus evaluating multiple analytes individually, and high-throughput capabilities that are further enhanced by using the automated features of modern instrumentation platforms. 25The practical utility of generating unique biomarker signatures is highlighted by the recent increase in FDA approvals of multiplex proteomic assays for clinical use, including cancer detection. 25,26The molecular pro le of the 10 biomarkers is converted into a BC risk score based on the relative contribution of individual analytes; ongoing research goals include adjusting and optimizing the Oncuria assay's algorithm based on patient demographics and medical history to provide more opportunities for personalized application. 27stoscopy and voided urine cytology remain the frontline methods for assessing BC status. 3,4Urine testing is a noninvasive approach without the safety risks of cystoscopy, which becomes particularly important in elderly and frail patients.Cystoscopy is sensitive for papillary lesions but tends to miss at lesions as carcinoma in situ (CIS), although newer imaging techniques provide improved contrast to differentiate tumor from normal tissue. 28While cystoscopy is often used in individuals with NMIBC and in MIBC patients who have undergone bladder-sparing treatments, no global consensus exists for endoscopic follow-up scheduling. 29There is growing evidence that cystoscopy may be overutilized, increasing both direct treatment costs and risks. 30In NMIBC patients, cystoscopy overuse has been linked to a 2-fold increase in transurethral resections performed, with an increased proportion of resection specimens not containing cancer, thus attesting to the di culty in visually identifying cancers. 31While urine cytology is noninvasive, it often produces indeterminate (atypical) calls and has suboptimal sensitivity for detecting early and low-grade tumors. 5e prospective study reported cytology sensitivities of 84% for high-grade but only 16% for low-grade NMIBC. 32By contrast to cystoscopy and cytology, the Oncuria assay, when adjusted for patient demographics, previously showed sensitivity values for high-grade BC, low-grade BC, MIBC and NMIBC of 94%, 89%, 97% and 93%, respectively (using the Model 200 ow analyzer). 12The 10-analyte panel also had a negativepredictive value of 99% for a BC call, which may preclude potentially super uous additional testing and procedures.Oncuria testing may be an important noninvasive adjunctive method for con rming and adding clinical value to the BC ndings of cystoscopy and cytology.
A primary study limitation was that the three xMAP instruments compared were produced by one manufacturer.
Fluorescent bead-based assays are easily standardized to ensure inter-lab reproducibility. 24,33,34Generalization of ndings is limited by the inclusion of primarily male urine samples, though prior and ongoing studies have included a larger number of samples from females.Additionally, updated algorithms used to analyze BC risk with the Oncuria assay take gender into account for clinical interpretation of assay output. 12While MIBC accounts for ≈ 25% of BC cases, our study employed a higher percentage (50%) due to our hospital being a tertiary institution that sees many advanced BC cases.
While this discrepancy has relevance in a clinical evaluation, it has less bearing on the intent or outcome of the current methodological investigation.
In conclusion, the Oncuria BC assay performed similarly well across three different ow analysis platforms for all 10 analytes simultaneously evaluated in urine samples.This agreement across instruments indicates that the test is amenable to standardized performance in laboratories using existing xMAP, without requiring costly outlays for new equipment.The multiplex Oncuria assay shows promise as a noninvasive and rapid-reporting adjunctive approach to cystoscopy and cytology in helping to identify BC, predict disease response to various therapies, track treatment progress, and monitor for recurrence.

Declarations
Ethics approval and consent to This study received approval and a waiver of consent to use previously banked de-identi ed urine samples from the Cedars-Sinai Medical Center Institutional Review Board, Los Angeles, CA (IRB #00001459).Study performance complied with the tenets of the Declaration of Helsinki.
Consent for publication: Not The assay was run on the LED/image-based MagPix, and laser/ owbased Luminex 200 and FlexMap 3D xMAP instruments operated with xPONENT Software V4.2 (MagPix and FlexMap 3D) and V4.3 (Luminex 200) (all from Luminex Corp.).23The classic 200 unit is designed for multiplex analysis up to 100 analytes in a single sample, and reads 96-well microtiter plates in ≈ 45 min.The MagPix instrument is more compact and portable than the 200 model to accommodate settings with space constraints or eldwork, and simultaneously measures 50 analytes in 96-well plates in ≈ 60 min.Both the 200 and MagPix models provide single-digit pictogram/mL sensitivity for protein targets and ≥ 3.5 logs of dynamic range.The newer FlexMap 3D allows evaluation of up to 500 analytes in a single sample.It has increased sensitivity (sub-picogram/mL) and dynamic range (≥ 4.5 logs) compared to earlier instruments, and accommodates high-throughput analysis and more advanced automation.The FlexMap 3D reads 96-well plates in ≈ 20 min and 384-well plates in ≈ 75 min.

Figures Figure 1
Figures

Table 1 Diagnostic performance of Oncuria assay in identifying high-grade/low-grade and highstage/low-stage BC.
* 1 case was missing a single analyte and thus excluded.N = 362 subjects presenting for bladder cancer evaluation.Instrumentation was Luminex 100/200 analyzer.
post-test for multiple comparisons.For protein calculations, analyte measurements above the highest standard curve value were replaced with that analyte's respective highest standard value, as is performed when calculating clinical risk scores.
Data Analysis.Data were analyzed using Prism v.9 graphing and statistical analysis software (GraphPad Software, Inc., San Diego, CA), and Excel v.16 (Microsoft Inc.Redmond).Analyte concentrations were determined by comparing sample readings to standard curves generated using a 5-parameter logistical curve t algorithm (xPONENT software from Luminex).Analyte concentrations are presented as pg/mL ± SD, range, median uorescence units (MFI, the instruments' raw data output) or number (% of samples), as appropriate.Mean values were compared by repeated measures ANOVA with Tukey

Table 2
Subject Characteristics.Signal Strength by Instrument.Raw uorescence signals output by the Model 200 and MagPix ow cytometry instruments were very similar for all 10 analytes, and both instruments' outputs were lower than signals from the FlexMap 3D instrument (Table

Table 3
Raw uorescence data outputs across three ow analyzers (median uorescence intensity, arbitrary units).

Table 4
Biomarker protein concentrations in urine samples compared across three ow analyzers (pg/mL).Shown is output from six representative urine samples, from three con rmed BC subjects ("Tumor") and three Control subjects.Values are averages of duplicate wells per analyte, per subject, rounded to the nearest whole number.Values that exceeded the highest standard curve concentration for any individual analyte were assigned that biomarker's highest standard value.The full dataset for all 36 subjects is provided in Supplementary TableS1.

Table 5
Details of instrument output discrepancies with statistical signi cance.Differences calculated as the absolute value of the ∆ divided by the lower of the two mean protein values × 100, with % differences rounded to nearest 0.1%.† P-values calculated by repeated ANOVA with Tukey post-test correction for multiple comparisons.All concentration values rounded to nearest 0.1 picogram.