Comparative study of Allergy Explorer (ALEX) versus ImmunoCAP platforms

Full Text

The ALEX® chip is a multiplex test that combines simultaneous measurement of total IgE (tlgE) and specific lgE (sIgE) to allergen extracts and allergen components in one assay. In total 282 analytes are measured, 156 extracts and 125 components, and total IgE. As compared to ImmunoCAP™ ISAC 112, 81 allergen components are overlapping, 44 are unique to the ALEX test, while 31 are only present on the ImmunoCAP ISAC test. The manufacturer of the ALEX test claims that the assay gives semi-quantitative results for tIgE up to 2500 kU/L, for sIgE to allergen extract and allergen components the test is said to give quantitative results up to 50 kUA/L [1]. ImmunoCAP single-analyte tests for total IgE, allergen extracts and allergen components all generate quantitative results, with traceability to the WHO standard [2, 3]. The ImmunoCAP ISAC carries 112 allergen components and generates semi-quantitative results (ISAC standardized units/ISU) [4]. Purpose To evaluate the technical performance of the ALEX test in measuring total IgE, sIgE to whole allergen extracts and allergen components in comparison with results from ImmunoCAP single tests (total IgE, sIgE) and the multiplex platform ImmunoCAP ISAC sIgE 112. For correspondence Статья поступила 04.12.2019 г Charlotte Nereliu, PhD, Thermo Fisher Scientific, Phadia Принята к печати 11.02.2020 г E-mail: charlotte.nerelius@thermofisher.com Рекомендована к публикации Е.С. Феденко 76 Российский Аллергологический Журнал 2020;17(1) I FOR PRACTITIONERS • Comparative study of Allergy Explorer (ALEX) versus ImmunoCAP platforms Study design and methods Eleven whole allergens and their corresponding allergen components representing different allergen groups (seasonal/perennial inhalant allergens, food allergens of animal and plant origin) were selected based on proven clinical relevance and for representation on both the ALEX and ImmunoCAP platforms. Samples from Phadia AB’s Diagnostic Sample Unit were selected for known sIgE positivity to the selected whole allergens, and for their respective components present on the microarrays. Samples with multiple extract positivity and with sIgE spanning the range from 0.35 to 50 kUA/L were chosen. In total, 64-66 serum samples were analyzed using the ALEX test, ImmunoCAP Specific IgE whole allergen tests, the ImmunoCAP Total IgE test and the ImmunoCAP ISAC sIgE 112 chip. 19-50 of the samples were also analyzed using ImmunoCAP Specific IgE allergen component tests (n=25), representing the selected whole allergens (Tab. 1). Results Total IgE ImmunoCAP Total IgE test results for serum samples (n=65) ranged between 7 and 7900 kU/L, while the ALEX test results ranged between 12 and 654 kU/L. The correlation between the ALEX test and ImmunoCAP test results was poor above 500 kU/L (Figure 1). According to the manufacturer, the measuring range of the ALEX test is 1-2500 kU/L, generating semiquantitative results [1]. The ImmunoCAP Total IgE test gives quantitative results [2] (Fig. 1). Allergen extracts: ALEX versus ImmunoCAP Whole Allergens Agreement in classes According to the manufacturers both the ALEX test and ImmunoCAP Specific IgE tests generate quantitative results [1, 3]. Nevertheless, to allow easy comparison over table 1. List of selected whole allergens and allergen components analyzed for each platform. Full ImmunoCAP product names are available at the end of the document Whole allergen extracts (n=11) Corresponding components common to the ALEX test and ImmunoCAP Allergens (n=25) Components common to the ALEX and ImmunoCAP™ ISAC sIgE 112 test 2 (n=26) House dust mite Der p 1, Der p 2, Der p 10, Der p 23 Der p 1, Der p 2, Der p 10, Der f 1, Der f 2 Wheat Gliadin - Peach Pru p 3 Pru p 3 Olive pollen Ole e 1 Ole e 1 Timothy Phl p 1, Phl p 2, Phl p 5, Phl p 6, Phl p 7, Phl p 12 Phl p 1, Phl p 2, Phl p 5, Phl p 6, Phl p 7, Phl p 12 Wall Pellitory Par j 2 Par j 2 Egg Gal d 1, Gal d 2, Gal d 3 Gal d 1, Gal d 2, Gal d 3, Gal d 5 Milk Bos d 4, Bos d 5, Bos d 8 Bos d 4, Bos d 5, Bos d 8 Hazelnut Cor a 1, Cor a 8, Cor a 9, Cor a 14 Cor a 1, Cor a 8, Cor a 9 Walnut Jug r 1 Jug r 1, Jug r 2 All assays were performed according to each manufacturers’ instructions by professional laboratory technicians. ALEX tests were run at Thermo Fisher Scientific’s Service laboratory, Allerod, Denmark, all other assays in the R&D department in Uppsala, Sweden. Statistical Analyses The agreement between the different test systems was evaluated for samples with at least six positive test results on both the ALEX and ImmunoCAP tests. A linear regression analysis was performed following the method comparison procedure of Passing-Bablok [5]. The analyses were done separately for each allergen and in the measuring range relevant for the ALEX test (0.3-50 kUA/L for sIgE; 1-2500 kU/l for total IgE) according to the manufacturer. Overall positive and negative agreement was calculated as well as class agreement based on classes defined for the ALEX test. ImmunoCAP (kll/L) Figure 1. Correlation of total IgE results for the ALEX test versus ImmunoCAP tests showing data points up to 2500 kU/L (n=58) with ImmunoCAP Seven data points giving results above 2500 kU/L with ImmunoCAP are omitted. The regression line is shown in red Russian Journal of Allergy 2020;17(1) 77 В ПОМОЩЬ ПРАКТИЧЕСКОМУ ВРАЧУ • Сравнительное исследование Allergy Explorer (ALEX) и платформ ImmunoCAP the entire measuring range, allergen extract results were transformed into classes as defined for ALEX (Tab. 2, Fig. 2). The agreement of the ALEX test data with ImmunoCAP test data was calculated per class. Correlations of all results are shown in Fig 3 а and 3 b. Table 2. Absolute agreement of the ALEX test results versus ImmunoCAP test results by classes defined for ALEX. the overall class agreement was 53% Class (ALEX definition) Agreement ALEX vs ImmunoCAP™ 0 (<0,3 kU/l) 99,7% 1 (0,3-1 kU/l) 4,0 % 2 (1-5 kU/l) 16,2% 3 (5-15 kU/l) 23,4 % 4 (>15 kU/l) 38,8 % Figure 2. Heat map of the ALEX test vs ImmunoCAP test results. The false negative rate, defined as <0.3 kUA/L in the ALEX test but >0.3 kUA/L in ImmunoCAP tests, was 55% (234/423 positive ImmunoCAP results) for the ALEX test. False negatives were found over the whole measuring range Quantitative comparisons Quantitative correlations of individual allergen extract test results were also performed. Fig. 4 and Fig. 5 show two examples each of allergen extracts with good and poor agreement, respectively, when comparing the ALEX test results to ImmunoCAP Specific IgE whole allergen test results. ImmunoCAP test results were considered positive above the threshold set for the ALEX test, i.e. >0.3 kUA/L. Note: The number of positive test results for ImmunoCAP tests would be higher if results down to 0.1 kUA/L (representing the LoQ of ImmunoCAP Specific IgE tests) were included). . ■ ■ ■ .. % я . ч m ■ "" ■ ■ Л ■ Л " , ■ ■ ■ ■ - ■ ■ 0 50 100 150 200 Figure 3 a. Correlation of all ALEX and ImmunoCAP Specific IgE whole allergen test results (n=709 data points). Samples with up to 203 kUA/L sIgE according to ImmunoCAP sIgE test results generate ALEX test results up to 36 kUA/L ImmunoCAP Specific IgE (kllA/L) 30 25 ' . _ 20 ... D . " ' if. 15 x . „«■ - ■ LU ' . ■ < 10 0 10 20 30 40 50 ImmunoCAP Specifc IgE (kUA/L) Figure 3 b. The shaded area of Fig. 3 а shown in enlargement Allergen components: ALEX versus ImmunoCAP Allergen Components The overall agreement of component sIgE results of the ALEX test versus ImmunoCAP tests was analyzed by transforming results into “ALEX classes”. Twenty-five components from the 11 chosen allergen extracts were analyzed using 19-50 serum samples each (Tab. 1). Agreement of the ALEX test data with Immuno CAP test data was calculated per class (Tab. 3, Fig. 6). As the manufacturer of the ALEX test states that the test gives quantitative results [1], correlations between results in kUA/L using the ALEX and ImmunoCAP Specific IgE tests for individual allergen components were also performed. For four components, namely Cor a 14 and Cor a 1 from hazelnut, Par j 2 from wall pellitory and Phl p 7 from timothy pollen, it was not possible to perform regression analyses since too few positive responses were obtained with the ALEX test (less than six positive sera). 45 1 18 31 1 2 21 21 25 4 8 8 5 285 94 100 30 10 0 12 3 4 ImmunoCAP Specific IgE results 78 Российский Аллергологический Журнал 2020;17(1) FOR PRACTITIONERS • Comparative study of Allergy Explorer (ALEX) versus ImmunoCAP platforms Figure 4 a. Correlation of the ALEX test vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for D. pteronyssinus extract. The complete set of data points is depicted in logarithmic scatter plot Figure 4 c. Correlation of the ALEX test vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for timothy grass extract. The complete set of data points is depicted in logarithmic scatter plot 50- 40- 30- 20- • 10- . • 0- • ( ) 1C I I I 20 30 40 I 50 Figure 4 b. Correlation of the ALEX test vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for D. pteronyssinus extract. Correlations of datapoints positive in both systems, and within the ALEX test measuring range of 0.3-50 kUA/L (regression line in red). For D.pteronyssinus (house dust mite, d1) 39 samples out of totally 65 samples were positive on ImmunoCAP Specific IgE tests (>0.3 kUA/L). The corresponding number for the ALEX test was 31 ImmunoCAP (kl)A/L) Figure 4 d. Correlation of the ALEX test vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for timothy grass extract. Correlations of datapoints positive in both systems, and within the ALEX test measuring range of 0.3-50 kUA/L (regression line in red). For timothy, 47 out of totally 65 samples were positive on ImmunoCAP Specific IgE tests (>0.3 kUA/L). The corresponding number for the ALEX test was 25 For Cor a 1 and Phl p 7 the number of sera positive in both systems were too few to draw any meaningful conclusions, while for Par j 2 and Cor a 14, 23/41 samples and 13/40 samples, respectively, were positive in ImmunoCAP tests (>0.3 kUA/L). The corresponding numbers of positive samples on the ALEX test were three for both Par j 2 and Cor a 14, resulting in the poorest positive agreement in this study, 13% and 23%, respectively. Out of the remaining twenty-one components analyzed, two examples of components with good and two with poor agreement when comparing the ALEX to ImmunoCAP Specific IgE Allergen Component test results are shown (Fig. 7 and Fig. 8). For ImmunoCAP tests, results were considered positive above the threshold set for the ALEX test, i.e. >0.3 kUA/L. Again, the number of positive test results for ImmunoCAP tests would be higher if results down to 0.1 kUA/L (representing the LoQ of the ImmunoCAP Specific IgE tests) were included). Analysis of negative extract results in the ALEX test vs. component results in the ALEX and ImmunoCAP tests: Hazelnut example Out of 64 samples tested, 25 generated positive results with ImmunoCAP hazelnut whole allergen test but were Russian Journal of Allergy 2020;17(1) 79 В ПОМОЩЬ ПРАКТИЧЕСКОМУ ВРАЧУ • Сравнительное исследование Allergy Explorer (ALEX) и платформ ImmunoCAP Figure 5 a. Correlation of the ALEX vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for Hazelnut extract. The complete set of data points is depicted in logarithmic scatter plot Figure 5 c. Correlation of the ALEX vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for Olive pollen extract. The complete set of data points is depicted in logarithmic scatter plot Figure 5 b. Correlation of the ALEX vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for Hazelnut extract. Correlations of datapoints positive in both systems, and within the ALEX test measuring range of 0.3-50 kUA/L (regression line in red). For hazelnut, 42 samples out of 64 were positive using ImmunoCAP Specific IgE tests (>0.3 kU./L). For the ALEX test results, the corresponding number was 17 Figure 5 d. Correlation of the ALEX vs ImmunoCAP Specific IgE whole allergen test results in kUA/L for Olive pollen extract. Correlations of datapoints positive in both systems, and within the ALEX test measuring range of 0.3-50 kUA/L (regression line in red). 41 samples out of 64 were positive using ImmunoCAP Specific IgE tests (>0.3 kU./L). For the ALEX test results, the corresponding number was 13 table 3. Agreement of the ALEX test component results versus ImmunoCAP Allergen Component test results in “ALEX classes”. the overall class agreement was 70% Class (ALEX definition) Agreement ALEX vs ImmunoCA) 0 (<0,3 kU/l) 98,9% 1 (0,3-1 kU/l) 14,9% 2 (1-5 kU/l) 36,3% 3 (5-15 kU/l) 43,9% 4 (>15 kU/l) 66,0% LC P X [J.I 15 68 26 36 27 1 11 45 18 8 4 17 10 5 461 86 43 8 3 0 12 3 4 ImmunoCAP slgE results Figure 6. Heat map of the ALEX test vs ImmunoCAP test results. The false negative rate for components IgE (defined as <0.3 kUA/L in the ALEX test but >0.3 kUA/L in ImmunoCAP tests), was 33% (140/426) for the ALEX test. False negatives were found over the whole measuring range 80 Российский Аллергологический Журнал 2020;17(1) FOR PRACTITIONERS • Comparative study of Allergy Explorer (ALEX) versus ImmunoCAP platforms Figure 7 a. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen component test results in kUA/L for Phl p 5 from timothy grass. All data points are depicted in logarithmic scatterplot ImmunoCAP (кид/L) Figure 7 b. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen component test results in kUA/L for Phl p 5 from timothy grass. Correlations of datapoints positive in both systems, and within the measuring range of 0.3-50 kUA/L defined for the ALEX test (regression line in red). 20 out of 44 samples for sIgE to Phl p 5 were positive on ImmunoCAPs Specific IgE tests (>0.3 kUA/L). The corresponding number on the ALEX test was 19 Figure 7 c. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen component test results in kUA/L for Der p 1 from D. pter-onyssinus. All data points are depicted in logarithmic scatterplot ImmunoCAP (кид/L) Figura 7 d. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen component test results in kUA/L for Der p 1 from D. pteronyssinus. Correlations of datapoints positive in both systems, and within the measuring range of 0.3-50 kUA/L defined for the ALEX test (regression line in red). 18 out of 38 samples for sIgE to Der p 1 were positive on ImmunoCAPs Specific IgE tests (>0.3 kUA/L). The corresponding number on the ALEX test was 14 negative using the ALEX test. The results for component IgE analyses for 24 of these samples are shown in Tab. 4. Twenty-three sera were positive for any hazelnut component on ImmunoCAP tests, while ALEX generated positive hazelnut component results in only nine. The results underline that negative ALEX extract test results are not necessarily truly negative. In addition, they demonstrate that false negative extract results are not counter balanced by component sIgE positivity, but that sensitizations to certain allergen sources can be missed entirely. CCD It is important to note, that the false negative rates (see. Tab. 5) did not change significantly when only analyzing sera lacking sIgE to cross-reactive carbohydrate determinants as determined by ImmunoCAP Specific IgE MUXF3 (o214, CCD) (n=54-56). It is possible that the lower sensitivity of the ALEX test, at least partly, is a consequence of the necessary 5x dilution of sera in the protocol for the ALEX assay. The 5x dilution of samples in the ALEX assay is the step that allows for CCD inhibition. According to the manufacturer, 85% reduction of CCD-reactivity is achieved by this incubation, which is said to increase to 95% if incubation is prolonged by 30 minutes [6]. When running the ImmunoCAP ISAC, no serum dilution is required, and to date only six CCD-con-taining allergen components are among the 112 present on the ImmunoCAP ISAC sIgE 112 chip. When the new ImmunoCAP ISACE112i test is launched only four CCD-containing allergen components will remain. Russian Journal of Allergy 2020;17(1) 81 В помощь ПРАКТИЧЕСКОМУ ВРАЧУ Сравнительное исследование Allergy Explorer (ALEX) и платформ ImmunoCAP n=9 Figure 8 a. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen Component results in kUA/L for Jug r 1 from walnut. All data points are depicted in logarithmic scatterplot n=5 n=19 £0.1 1 10 ImmunoCAP (kUA/L) 2100 Figure 8 c. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen Component results in kUA/L for Gal d 2 from hen's egg white. All data points are depicted in logarithmic scatterplot 50- 40- 30- 20- 10n=9 20 30 ImmunoCAP (kUA/L) 40 50 Figure 8 b. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen Component results in kUA/L for Jug r 1 from walnut. Correlations of data points positive in both systems, and within the measuring range of 0.3-50 kUA/L defined for the ALEX test (regression line in red). 19/42 samples were positive for sIgE to Jug r ImmunoCAP Specific IgE (>0.3 kUA/L). The corresponding positive number for the component on ALEX was 10 50- 40- 30- 20- 10- n=9 и! • •_- 0 1C i I ) 20 30 1 1 40 i 50 ImmunoCAP (kUA/L) Figure 8 d. Correlation of the ALEX and ImmunoCAP Specific IgE Allergen Component results in kUA/L for Gal d 2 from hen's egg white. Correlations of data points positive in both systems, and within the measuring range of 0.3-50 kUA/L defined for the ALEX test (regression line in red). 15/19 samples were positive for sIgE to Gal d 2 ImmunoCAP Specific IgE (>0.3 kUA/L). The corresponding positive number for the components on ALEX was 10 table 4. Comparison of component sIgE results in the two platforms for sera negative to hazelnut extract on the ALEX test, but positive using ImmunoCAP whole allergen tests. Both 0.1 and 0.3 kUA/L cut-offs for ImmunoCAP Specific IgE allergen component test results are shown Number of samples positive for components Method sIgE to extract sIgE sIgE to any comp Cor a 1 Cor a 8 Cor a 9 Cor a 14 Cut-off >0,3 >0,3 >0,1 >0,3 >0,1 >0,3 >0,1 >0,3 >0,1 >0,3 ImmunoCAP 24 23/24 8 2 18 17 12 10 13 5 ALEX components 0 9/24 НД 0 НД 5 НД 5 НД 0 82 Российский Аллергологический Журнал 2020;17(1) for practitioners • comparative study of allergy explorer (Alex) versus ImmunocAP platforms table 5. Summary of the overall sensitivity of the ALEX test Whole extracts ALEX vs ImmunoCAP sIgE extracts 55% false negatives Allergen components ALEX vs ImmunoCAP sIgE components 33% false negatives Allergen components ALEX vs ImmunoCAP ISAC 112 16% false negatives ImmunoCAP ISAC 112 vs ALEX 9,5% false negatives table 6. Agreement of the ALEX test component results versus ImmunoCAP ISAC 112 Allergen Components test results and vice versa (according to “ALEX classes”) Class (ALEX definition) 0 (<0,3 kU/l) 1 (0,3-1 kU/l) 2 (1-5 kU/l) 3 (5-15 kU/l) 4 (> 15 kU/l) Agreement ALEX vs ImmunoCAP ISAC 97,7% 19,3% 42,7% 51,0% 78,0% Agreement ImmunoCAP ISAC vs ALEX 96,0% 33,3% 47,0% 49,0% 65,2% 4 3 1 5 26 60 2 4 36 53 13 8 28 47 14 3 23 16 8 1 1402 34 14 10 1 0 12 3 4 ImmunoCAP ISAC Figure 9. Heat map of the ALEX test vs ImmunoCAP ISAC 112 Components test results. The false negative rate for the ALEX test vs the ImmunoCAP ISAC 112 test was 15.8% (59/374) while the false negative rate for the ImmunoCAP ISAC 112 test vs the ALEX test was 9.5% (33/348) Figure 10. The ALEX test results for component sIgE level off and do not reach values above 39 kUVL among 407 positive data points, while the ImmunoCAP ISACa112 test detects Specific IgE up above 100 ISU-E. Thirteen positive data points are lacking in the plot (truncated) ALEX vs ImmunoCAP ISAC sIgE 112 component test results For the comparison between ALEX and ImmunoCAP ISAC sIgE 112, class agreement was calculated in two ways, assuming that either system gives the “true” results. False negative results for the ALEX vs the ImmunoCAP ISAC sIgE 112 test was 15.8% (59/374), while the false negative rate for the ImmunoCAP ISAC sIgE 112 vs the ALEX test was 9.5% (33/348) (Tab. 6, Fig. 9). Overall, the result agreement is fair between the two systems. The overall class agreement ranges from 61-97% for the studied components, the vast majority being above 80%. The lower agreement in classes 1-3 (between 19 and 51%) is most likely due to a lower sensitivity of the ALEX test, results did not reach values above 39 kUA/L among 407 positive data points, while the ImmunoCAP ISAC 112 test detects Specific IgE up above 100 ISU-E (Fig. 10). The agreement within class 4, i.e. slgE levels above 15 kUA/L, is 78%. Overall sensitivity of the ALEX test Throughout the comparisons, the ALEX test was found to be less sensitive than ImmunoCAP tests as indicated by the rate of false negative results for the ALEX test (Tab. 5). Discussion According to the manufacturer, the ALEX multiplex test is an alternative to ImmunoCAP and ImmunoCAP ISAC tests, widely used on the market for measurement of total and specific IgE levels. A direct comparison of the technical performance of ALEX with ImmunoCAP tests revealed a lack of sensitivity of the ALEX test for both whole allergens (the false negative rate 55%, 234/423 positive ImmunoCAP results) and allergen components (the false negative rate 33%, positive Im-munoCAP results). In addition, the dynamic range of ALEX is significantly lower: no results reached values above 36 kUA/L for whole allergens and 38 kUA/L for allergen components. Levels in ImmunoCAP specific IgE tests for the same samples were >100 kUA/L and up Russian Journal of Allergy 2020;17(1) 83 В помощь практическому Врачу • сравнительное исследование Allergy Explorer (ALEX) и платформ ImmunoCAP to 150 kUA/L, respectively. At the same time, the ALEX test showed lower results in the entire dynamic range for both whole allergens and specific allergen components. In particular, when comparing ALEX and ImmunoCAP tests for whole allergens, overall class agreement was 53%, ranging from 4 - 40% for positive classes and positive quantitative agreement ranged from 23% to 80% for the 11 extracts studied. A slightly better picture was observed for allergen components: overall class agreement was 70% ranging from 15-66% for positive classes. False negative results obtained by analyzing the response to hazelnut extract provide additional confirmations for the conclusions stated above. At the same time, comparison of negative results demonstrated a high agreement between the ALEX and ImmunoCAP Specific IgE tests showing 99.7% agreement for extracts and 98.9% agreement for components for negative results, respectively. These data testify no apparent background problems in the ALEX test. Thus, the ALEX® tIgE, slgE to whole allergens and allergen components tests show low dynamic range and poor agreement with quantitative results comparing to ImmunoCAP singleplex tests. In comparison with the ImmunoCAP ISAC sIgE 112 chip, the agreement is better, but the sensitivity and dynamic range are still lower.

About the authors

C Nerelius

Thermo Fisher Scientific

Author for correspondence.
Email: charlotte.nerelius@thermofisher.com

M Andersson

Thermo Fisher Scientific

Email: mats.andersson@thermofisher.com

L Sogaard

Thermo Fisher Scientific; Thermo Fisher Diagnostics ApS

Email: lone.sogaard@thermofisher.com

M Schwanbeck

Thermo Fisher Scientific

Email: maria.schwanbeck@thermofisher.com

S Kofler

Thermo Fisher Scientific; Thermo Fisher Diagnostics Austria GmbH

Email: stefan.kofler@thermofisher.com

M Berthold

Cyxone AB

Email: malin.berthold@cyxone.com

References

Supplementary files