OLYMPUS AFP - ALPHA-FETOPROTEIN

{0} 1 # 510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION DECISION SUMMARY A. 510(k) Number: k081709 B. Purpose for Submission: New Device C. Measurand: Alpha-Fetoprotein (AFP) D. Type of Test: Automated chemiluminescent immunoassay E. Applicant: Olympus America Inc. F. Proprietary and Established Names: Olympus AFP Test System G. Regulatory Information: 1. Regulation section: 866.6010 Tumor-associated antigen immunological test system 862.1150 Calibrator 862.1660 Quality control material (assayed and unassayed) 2. Classification: Class II 3. Product code: LOJ Kit, Test, Alpha-fetoprotein for testicular cancer JIT Calibrator, secondary JJX Single (specified) analyte controls (assayed and unassayed) 4. Panel: Immunology (82) H. Intended Use: 1. Intended use(s): The Olympus AFP assay is a paramagnetic particle (Dynabeads®), chemiluminescent immunoassay for the quantitative determination of alpha-fetoprotein levels in human serum/plasma using the Olympus AU3000i Immunoassay System. The Olympus AFP assay is intended for use as an aid in the management (monitoring) of patients with non-seminomatous germ cell tumors. The Olympus AFP calibrator is for calibrating the quantitative Olympus assay on the Olympus AU3000i Immunoassay. The Olympus AFP control is used for the quality control of the Olympus AFP assay on the Olympus AU3000i Immunoassay System. 2. Indication(s) for use: Same as above 3. Special conditions for use statement(s): Prescription use only 4. Special instrument requirements: Olympus AU3000i Immunoassay System (k062581) {1} 2 I. Device Description: The Olympus AFP Test System consists of two reagents (Reagent 1 and Reagent 2), calibrator and control material. Reagent 1 consists of paramagnetic particles coated with murine monoclonal anti-AFP antibody, Tris buffer, protein stabilizers and preservative. Reagent 2 consists of alkaline phosphatase labeled murine monoclonal anti-AFP antibody conjugate, MES buffer, protein stabilizers, detergent and preservatives. The calibrator is AFP prepared in bovine matrix with preservatives and the control is AFP prepared in human matrix with preservatives. J. Substantial Equivalence Information: 1. Predicate device name(s): Roche Elecsys AFP Assay Roche Elecsys PreciControl Tumor Marker Control Roche Elecsys AFP CalSet 2. Predicate K number(s): k981282 k050387 k043095 3. Comparison with predicate: AFP Test System | Similarities | | | | --- | --- | --- | | Item | New Device | Predicate | | Intended Use | The Olympus AFP assay is intended for use as an aid in the management (monitoring) of patients with non-seminomatous germ cell tumors. | Same | | Measurement | Quantitative | Same | | Assay Methodology | Chemiluminescent, two-site immunoassay | Same | | Capture Antibody | Murine monoclonal | Same | | Solid phase | Microparticle | Same | | Storage | 2-8°C | Same | | Differences | | | | --- | --- | --- | | Item | New Device | Predicate | | Instrument | Olympus AU3000i™ Test System | Roche Elecsys 1010/2010 and Modular analytics E170 immunoassay analyzers | | Matrix | Serum and plasma (Lithium heparin) | Serum and plasma (sodium heparin, EDTA, or sodium citrate) | | Measuring Range | 0.1 – 390 ng/mL | 0.6 – 1210 ng/mL | {2} | Differences | | | | --- | --- | --- | | Item | New Device | Predicate | | Conjugate | Alkaline phosphatase | Streptavidin | | Sample volume | 20μl | 10μl | | Stability | 2-8°C for 28 days | 2-8°C for 4 -12 weeks depending on instrument | Calibrator | Similarities | | | | --- | --- | --- | | Item | New Device | Predicate | | Intended Use | For calibration | Same | | Traceability | 1stIRP WHO Reference Standard 72/225 for human AFP | Same | | Storage | 2-8°C | Same | | Differences | | | | --- | --- | --- | | Item | New Device | Predicate | | Instruments | Olympus AU3000i™ Test System | Elecsys immunoassays systems | | Reagent Preparation | Liquid | Lyophilized | | Quantity | One | 4 x 1.0 mL | | Stability; Open | 2-8°C for 28 days | When reconstituted: 5 hours on board 6 weeks at 2-8°C 12 weeks at -20°C | | Composition | Bovine | Human | | Levels | One (~390 ng/mL) | Two (~6.0 and 60 ng/mL) | Control | Similarities | | | | --- | --- | --- | | Item | New Device | Predicate | | Intended Use | Quality control | Same | | Composition | Human serum | Same | | Stability; Open | 2-8°C for 28 days | When reconstituted: 5 hours on the analyzer Two weeks at 2-8°C 1 month at -20°C | | Storage | 2-8°C | Same | | Differences | | | | --- | --- | --- | | Item | New Device | Predicate | | Instruments | Olympus AU3000i™ Test System | Elecsys and cobas e immunoassay analyzers | {3} | Differences | | | | --- | --- | --- | | Item | New Device | Predicate | | Reagent Preparation | Ready-to-use | Lyophilized; 4 vials that are reconstituted into 3 mL distilled water each. | | Levels | One (~10.2 ng/mL) | Two | # K. Standard/Guidance Document Referenced (if applicable): CLSI EP5-A2 Evaluation of Precision Performance of Quantitative Measurement Methods CLSI EP9-A2 Method Comparison and Bias Estimation Using Patient Samples CLSI C28-A2 How to Define and Determine Reference Intervals in the Clinical Laboratory. # L. Test Principle: The Olympus AFP assay is a two-step paramagnetic particle enzyme immunoassay. Samples are incubated with a monoclonal anti-AFP antibody bound to paramagnetic particles. The AFP reacts with the paramagnetic particles is washed and incubated with a second monoclonal anti-AFP antibody conjugated with alkaline phosphatase to form a sandwich complex. A chemiluminescent substrate is added to react with the bound phosphatase. Light generated by the reaction is measured by the luminometer. The light emission is proportional to the quantity of AFP in the sample. Results are calculated from a predefined calibration curve. # M. Performance Characteristics (if/when applicable): # 1. Analytical performance: # a. Precision/Reproducibility: Four sites evaluated precision of 3 serum pools and 2 controls. Precision pools were prepared centrally and provided to the study sites. The low precision pool was prepared from patient serum. The medium and high precision pools were prepared from patient serum spiked with human AFP. Three different lots were evaluated at the 4 sites and testing was based on CLSI EP5-A2. Acceptance criteria $\% \mathrm{CV} \leq 5\%$ was met. | | | | Repeatability (Within Run) | | Between Run | | Between Day | | Within Laboratory (Total) | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | Pools | Site | Mean (ng/ml) | SD (ng/ml) | CV (%) | SD (ng/ml) | CV (%) | SD (ng/ml) | CV (%) | SD (ng/ml) | CV (%) | | Low | 1 | 1.222 | 0.018 | 1.4 | 0.014 | 1.2 | 0.005 | 0.4 | 0.023 | 1.9 | | | 2 | 1.231 | 0.016 | 1.3 | 0.017 | 1.3 | 0.012 | 1.0 | 0.026 | 2.1 | | | 3 | 1.279 | 0.014 | 1.1 | 0.022 | 1.7 | 0.000 | 0.0 | 0.025 | 2.0 | | | 4 | 1.250 | 0.014 | 1.1 | 0.020 | 1.6 | 0.016 | 1.2 | 0.029 | 2.3 | {4} 5 | Medium | 1 | 24.654 | 0.420 | 1.7 | 0.557 | 2.3 | 0.000 | 0.0 | 0.697 | 2.8 | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | 2 | 24.376 | 0.377 | 1.5 | 0.568 | 2.3 | 0.166 | 0.7 | 0.701 | 2.9 | | | 3 | 25.232 | 0.324 | 1.3 | 0.650 | 2.6 | 0.098 | 0.4 | 0.733 | 2.9 | | | 4 | 23.149 | 0.349 | 1.5 | 0.322 | 1.4 | 0.316 | 1.4 | 0.570 | 2.5 | | High | 1 | 140.795 | 2.459 | 1.7 | 2.557 | 1.8 | 0.788 | 0.6 | 3.634 | 2.6 | | | 2 | 147.815 | 2.688 | 1.8 | 3.614 | 2.4 | 2.516 | 1.7 | 5.160 | 3.5 | | | 3 | 152.006 | 1.864 | 1.2 | 4.169 | 2.7 | 0.000 | 0.0 | 4.566 | 3.0 | | | 4 | 141.883 | 2.606 | 1.8 | 3.630 | 2.6 | 2.920 | 2.1 | 5.338 | 3.8 | The data for each site was combined. The %CV for the low, medium and high samples were determined for within-run (1.3, 1.5, and 1.7), between run (1.5, 2.2, and 2.4), respectively. The control (~10.1 ng/mL) provided with the Olympus AFP was also evaluated and met the acceptance criteria. Lot-to-lot: To measure the variation in concentration between multiple lots of the Olympus AFP Test system, 3 serum samples of low (~1.2 ng/mL), medium (~25ng/mL) and high (~145 ng/mL) were tested in parallel on 3 different lots. The %CV ≤ 3.7%. b. Linearity/assay reportable range: Linearity: To assay linearity across the measuring range of the assay, human serum pool was spiked with AFP (from human amniotic fluid) to just above the measuring range (~408 ng/mL) and then diluted with Sample Diluent (SDIL1) to create 11 concentrations that spanned the measuring range. Samples were evaluated in triplicate. The acceptance criteria was met. | Range (ng/mL) | Slope (95% CI) | Intercept ng/mL (95% CI) | Correlation Coefficient | n | | --- | --- | --- | --- | --- | | 0.036 to 408 ng/mL | 0.9762 (0.96 – 0.99) | -0.632 (-3.06 – 1.80) | 0.999 | 11 | Dilution recovery: To demonstrate the linearity of the assay, 3 patient samples were prepared in 2-fold dilutions to a total of 4 levels. In addition a "neat" sample was run. Percent recovery is calculated by comparing the observed AFP results with the expected value. Samples concentrations prior to dilution: ~74 ng/mL, 151 ng/mL and 370 ng/mL. Mean % recoveries were between 92% and 100%. Acceptable recovery within ±10% was met. Spiked Recovery: Three serum pools across the assay linear range were spiked with three different amounts of AFP. Each spiked sample was then diluted into a low (~10 ng/mL), medium (~56 ng/mL) and high sample (~168 ng/mL). The neat spiked solution and corresponding dilutions were run in quadruplicate. Mean % recoveries ranged ~91% to 104%. Acceptance criteria ±10% were met. c. Traceability: The Olympus AFP is traceable to the 1st IRP WHO Reference Standard {5} 72/225 for human AFP. # d. Stability: Real-time, on-board, open vial stability were conducted for AFP test reagents, calibrator and controls. The results support the 28 day claim. # f. Detection limit: The limit of blank was determined by running 60 replicates of the blank sample and taking the $95^{\text{th}}$ percentile of the blank samples. The limit of detection was determined using 5 serum samples tested 12 times. The limit of detection was determined to be $(0.048\mathrm{ng / mL})$ The limit of quantification (LoQ) study was based on EP17A. The AFP LoQ pools were prepared using 3 different serums. LoQ pool levels 1-4 and level 6 were prepared by diluting a female human serum (base concentration $= 0.98$ ng/mL) with assay diluent. The 2 remaining levels (5 and 7) were stripped serum samples. A total of 40 replicated were measured incorporating 2 reagent lots, 2 instruments, across 4 days. The LoQ was determined to be $0.077\mathrm{ng / mL}$ which represents the lowest concentration of AFP that can be measured with a total imprecision of $\leq 17.2\%$ # g. Analytical specificity: Interference by Heterophilic antibodies and Rheumatoid Factor was determined using human serum samples spiked with AFP to the desired levels to create base pools with two levels of AFP ( $\sim 5.0 \, \mathrm{ng/mL}$ and $\sim 47 \, \mathrm{ng/mL}$ ). Expected analyte values were compared to observed values and $\%$ recovery calculated. No significant interference was seen with $2225 \, \mathrm{IU/mL}$ RF in samples at the designated AFP concentrations. Interference by HAMA (1825 ng/mL) was not significant; however a statement cautioning against HAMA interference is included in the package insert. Interference and cross-reactivity: To test the susceptibility of the AFP test to common interfering substances, the following substances were tested by spiking them into a human serum sample containing AFP $8.7\mathrm{ng / mL}$ . The interference was calculated by comparing the recovery of AFP in the samples containing interferents to the control sample containing no interferents. Bilirubin is unconjugated. Interference was $\leq 5\%$ at the concentrations indicated: | Bilirubin | ≤ 3% up to 40 mg/dl Bilirubin | | --- | --- | | Haemolysate | ≤ 5% up to 5 g/L Haemolysate | | Intralipid™ | ≤ 5 % up to 10 g/L Intralipid™ | The following interferent and cross-reacting substances were tested by adding the identified substances in known concentrations to a serum pool containing AFP at a concentration $\sim 5\mathrm{ng / mL}$ . The compounds did not show interference $>10\%$ at the specific levels indicated. {6} 7 | Drugs | Amount added | % Recovery (spiked control) | | --- | --- | --- | | Acetaminophen | 200 μg/mL | 101% | | Acetyl Cysteine | 164 μg/mL | 101% | | Aspirine | 1,032 mg/mL | 104% | | Ampicillin - Na | 1,064 mg/mL | 99% | | Ascorbic acid | 289,2 μg/mL | 104% | | Bleomycin | 10 mUI/mL | 104% | | Carboplatin | 1,036 mg/mL | 101% | | Cefoxitin | 2,44 mg/mL | 101% | | Cisplatin | 1,980 mg/mL | 95% | | Cyclophosphamide | 984 μg/mL | 100% | | Cyclosporine | 5 μg/mL | 103% | | D-actinomycin | 2,5 μg/mL | 101% | | Doxycycline | 50 μg/mL | 102% | | Etoposide | 100 μg/mL | 101% | | Ibuprofen | 1,13 mg/mL | 100% | | Ifosfamide | 4,5 mg/mL | 93% | | Levodopa | 27,7 μg/mL | 105% | | Methotrexate | 1,125 mg/mL | 94% | | Methyldopa + 1,5 | 21,6 μg/mL | 104% | | Metronidazole | 232 μg/mL | 102% | | Naprosyn sodium | 1,2 mg/mL | 103% | | Phenylbutazone | 420,8 μg/mL | 99% | | Rifampicin | 68,4 μg/mL | 101% | | Paclitaxel | 10 μg/mL | 100% | | Theophylline | 119,5 μg/mL | 101% | | Vinblastine | 109 μg/mL | 100% | | Vincristine | 100 μg/mL | 98% | | Human Proteins | Amount added | % Recovery (spiked control) | | --- | --- | --- | | human alpha globulin (≥2 Macrogl) | 2,5 mg/mL | 95% | | human alpha-1-acid glycoprotein | 10 mg/mL | 99% | | human alpha-1-antitrypsin | 20 mg/mL | 99% | | hCG | 1000 IU/mL | 98% | | human Gamma Globulin | 100 mg/mL | 92% | | human Placental Lactogen | 500 μg/mL | 105% | | human Serum Albumin | 25,1 mg/mL | 107% | | human Transferrin | 100 mg/mL | 106% | Hook Effect (Prozone): The presence of high dose effect was tested by analyzing a concentrated sample of purified AFP antigen both neat and on dilution with the measuring range of the AFP assays. No Hook Effect was demonstrated at the highest concentration evaluated (2,160,000 ng/mL). h. Assay cut-off: Refer to clinical studies below. 2. Comparison studies: a. Method comparison with predicate device: A total of 289 serum samples were analyzed using the Olympus AFP Test and compared to the predicate. Samples were run in singlicate and the data was analyzed using Deming regression. The correlation coefficient (r) was calculated using ordinary linear fit regression. The results are shown below: | Comparator | Sample range | Slope (95% CI) | Intercept ng/mL (95% CI) | r | n | | --- | --- | --- | --- | --- | --- | | Roche Elecsys AFP | 0.89 – 349.56 | 1.02 (0.99 – 1.05) | -0.19 (-0.56 – 0.18) | 0.996 | 289 | {7} b. Matrix comparison: To demonstrate the performance of Li-heparin plasma when compared to serum, AFP concentrations were measured in 50 matched samples were tested. The slope, intercept were calculated by Deming regression. The results of the study demonstrate that results obtained using Li-heparin plasma samples are consistent with those obtained using serum across the range. | Range (ng/mL) | Slope (95% CI) | Intercept ng/mL (95% CI) | Correlation Coefficient | n | | --- | --- | --- | --- | --- | | 1.044 to 359.1 ng/mL | 1.007 (0.988 – 1.027) | -1.065 (-1.72 – 1.51) | 0.997 | 50 | 3. Clinical studies: A clinical study was performed to assess the performance of the Olympus AFP assay to monitor patients with non-seminomatous testicular cancer. Seventy-three (73) retrospective serial serum sample sets (total of 308 evaluable samples) with clinical data from men diagnosed with testicular cancer were tested. Inclusions and exclusion criteria for the samples were provided. Samples were selected for age (range 1 to 56 years old), race/ethnicity (specimens from African America were not evaluated in this study), and stage of disease (stage I through IV). A longitudinal analysis of serial draws from 73 patients was performed. All patients were categorized as Active/Progressing, Responding, Stable or No Evidence of Disease (NED). Disease progression was determined by the patient's physician based on physical examination, radiographic findings, and surgical procedures. The Reference Change Value (RCV) was used to identify a significant change in AFP levels. For this calculation, the RCV was derived using the formula $\mathrm{RCV} = 2.33\left(\mathrm{Sw} + \mathrm{a}^{2}\right)^{1 / 2}$ where $\mathrm{Sw} + \mathrm{a}^{2}$ is the addition of the analytical variation based on the claimed imprecision $(5\%)$ and the biological variation $(12\%)$ squared. (Calculations taken from Trapé J. et al. Reference change value for alpha-fetoprotein and its application in early detection of hepatocellular carcinoma in patients with hepatic disease. Clin Chem 2003;49:1209-1211.) The RCV for the Olympus AFP test was calculated to be $30\%$ and $31.6\%$ for the predicate. Per Visit Analysis: Changes in AFP concentrations and in disease status were analyzed on a per visit basis. Patients were categorized as Active/Progressing, Responding, Stable or No evidence of Disease (NED) by the attending physician based on clinical information. The table below shows the distribution of results when compared to the disease status for the Olympus AFP Test: {8} 9 | Change in AFP | Change in Disease State | | | | | | --- | --- | --- | --- | --- | --- | | | Responding N (% T) | Stable N (% T) | No Evidence of Disease N (% T) | Progressing N (% T) | Total N (% T) | | > 30.0 % increase | 6 (2.6 %) | 6 (2.6 %) | 9 (3.8 %) | 24 (10.2 %) | 45 (19.2 %) | | No Significant Change | 5 (2.1 %) | 17 (7.2 %) | 70 (29.8 %) | 14 (6.0 %) | 107 (45.1 %) | | > 30.0 % decrease | 20 (8.5 %) | 22 (9.4 %) | 18 (7.7 %) | 24 (10.2 %) | 83 (35.8 %) | | Total | 31 (13.2 %) | 45 (19.2 %) | 97 (41.3 %) | 62 (26.4 %) | 235 (100 %) | The following two tables show per visit clinical performance results for the Olympus AFP test and predicate device when analyzed as “Progression” and “No Progression” with “No Progression” consisting of responding, stable and NED. Olympus AFP Value vs. Disease Progression | | Change in Disease State | | | | --- | --- | --- | --- | | Change in AFP | Progression | No Progression | Total | | > 30.0% increase | 24 | 21 | 45 | | ≤ 30.0% increase | 38 | 152 | 190 | | Total | 62 | 173 | 235 | | | Estimate | 95% Confidence Interval | | --- | --- | --- | | Sensitivity | 38.7% | (26.6% - 51.9%) | | Specificity | 87.9% | (82.0% to 92.3%) | The table below shows the distribution of results when compared to the disease status for the Predicate Test: | Change in AFP | Change in Disease State | | | | | | --- | --- | --- | --- | --- | --- | | | Responding N (% T) | Stable N (% T) | No Evidence of Disease N (% T) | Progressing N (% T) | Total N (% T) | | > 31.6 % increase | 6 (2.6 %) | 7 (3.0 %) | 11 (4.7 %) | 26 (11.1 %) | 50 (21.3 %) | | No Significant Change | 5 (2.1 %) | 15 (6.4 %) | 65 (27.7 %) | 10 (4.3 %) | 95 (40.4 %) | | > 31.6 % decrease | 20 (8.5 %) | 23 (9.8 %) | 21 (8.9 %) | 26 (11.1 %) | 90 (38.3 %) | | Total | 31 (13.2 %) | 45 (19.1 %) | 97 (41.3 %) | 62 (26.4 %) | 235 (100 %) | Predicate Device AFP Value vs. Disease Progression | | Change in Disease State | | | | --- | --- | --- | --- | | Change in AFP | Progression | No Progression | Total | | >31.6% increase | 26 | 24 | 50 | | ≤ 31.6% increase | 36 | 149 | 185 | | Total | 62 | 173 | 235 | | | Estimate | 95% Confidence Interval | | --- | --- | --- | | Sensitivity | 41.9% | (29.5% - 55.2%) | | Specificity | 86.1% | (80.1% - 90.9%) | {9} Olympus AU3000i AFP Concordance to Comparative Method (on a per visit basis) | | >31.6 % increase | ≤ 31.6 % increase | Total | | --- | --- | --- | --- | | >30.0 % increase | 42 | 3 | 45 | | ≤ 30.0 % increase | 8 | 182 | 190 | | Total | 50 | 185 | 235 | | | | 95% Confidence Interval | | --- | --- | --- | | % Overall agreement | 95.3% | (91.8% – 97.6%) | | % Positive agreement | 84.0% | (70.9% – 92.8%) | | % Negative agreement | 98.4% | (95.3% – 99.7%) | 4. Clinical cut-off: Not applicable. 5. Expected values/Reference range: The distribution of AFP values in normal individuals, patients with benign conditions and malignant conditions was established. In this study, 97.5% of healthy males had AFP levels less than 7.14 ng/mL. | | Distribution of AFP values n(%) | | | | | | | --- | --- | --- | --- | --- | --- | --- | | | N | 0 - 5 IU/mL [0 - 6 ng/mL] | 5 - 10 IU/mL [6 - 12 ng/mL] | 10 - 100 IU/mL [12 - 120 ng/mL] | 101 - 325 IU/mL [121 - 390 ng/mL] | >325 IU/mL [>390 ng/mL] | | Apparently Healthy | | | | | | | | Male | 206 | 197 (95.6%) | 9 (4.4%) | - | - | - | | Benign Conditions | | | | | | | | Prostate | 108 | 101 (93.5%) | 7 (6.5%) | - | - | - | | GI/Lung | 109 | 103 (94.5%) | 5 (4.6%) | 1 (0.9%) | - | - | | Diabetes | 106 | 97 (91.5%) | 8 (7.5%) | - | - | 1 (0.9%) | | Heart/Liver | 108 | 102 (94.4%) | 6 (5.6%) | - | - | - | | Malignant Conditions (treated) | | | | | | | | Liver | 18 | 6 (33.3%) | 3 (16.7%) | 3 (16.7%) | 1 (5.6%) | 5 (27.8%) | | Lung | 83 | 74 (89.2%) | 8 (9.6%) | 1 (1.2%) | - | - | | Upper GI | 43 | 40 (93.0%) | 2 (4.7%) | 1 (2.3%) | - | - | | Prostate/Testicular/Bladder | 228 | 212 (93.0%) | 9 (3.9%) | 5 (2.2%) | 1 (0.4%) | 1 (0.4%) | | Colorectal | 61 | 50 (82.0%) | 8 (13.1%) | 3 (4.9%) | - | - | N. Proposed Labeling: The labeling is sufficient and it satisfies the requirements of 21 CFR Part 809.10. O. Conclusion: The submitted information in this premarket notification is complete and supports a substantial equivalence decision.