- Email: [email protected]
THE BIOMARKER-BASED DIAGNOSIS OF ALZHEIMER’S DISEASE: LESSONS FROM ONCOLOGY
Poster Presentations: Sunday, July 24, 2016 Table 1 Results of the precision study. Target range (pg/mL) Automated assay Reproducibility Rr-CV% Dd-CV% Total CV% Manual assay (5 days) Reproducibility Rr-CV% Dd-CV% Total CV%
930-1360
630-715
520-525
320-345
100-200
5.18 1.62 6.53 8.49
1.96 3.52 0.00 4.03
1.76 3.45 3.21 5.03
1.54 4.20 5.07 6.76
2.86 4.38 4.85 7.13
4.76 0.00 3.12 5.69
4.58 0.00 3.99 6.07
5.99 2.01 3.45 7.19
4.50 3.33 2.20 6.01
7.19 0.00 4.11 8.28
Background: The integration of CSF biomarker analysis in clinical routine or its global use in clinical trials is hampered in-part by the presence of different analyte concentrations in function of the assay design (e.g., matrix interference), the applied technology, a too high inter-center variability (reproducibility), and the absence of international reference standard materials. Part of the above mentioned problems were already solved recently by the integration in the market of a new series of immuno-assays or modified assay formats of existing assays for the most important CSF proteins. New designs focused, besides other aspects, on the removal of matrix interference and the integration of temperature-stable, ready-to-use calibrators. However, the manual test procedures for these new assays still do not solve completely the high inter-center variability. The goal: To develop an approach to transfer and qualify a manual test protocol for the CSF proteins (Ab(1-42), Ab(1-40), total tau) to fully automated test procedures. Methods: The manual test procedure was optimized for automated sample testing on two different instruments: the BepIII (Siemens, Erlangen, Germany) and the Janus (Perkin-Elmer, Waltham, MA, USA). Precision (reproducibility, between-run (rr) CV%, between-day (dd) CV%) after transfer was evaluated according to CLSI EP-5A-2 (10 days procedures) using neat CSF samples. Results: The results of the precision study are shown in Table 1. Precision data are much lower than published as part of the Alzheimer’s Association Quality Control Program. Conclusions: We have been able to transfer the manual test procedure as provided by the vendor to a fully-automated sample analysis, resulting in an improved test reproducibility for CSF and improved total CV% at the clinical decision level (520-715 pg/mL). The latter is considered as a cost-efficient (low investment) manner (automated colorimetric ELISA) for biomarker integration into clinical routine when compared with an IVD closed system.
P1-201
QUALITY INDICATORS OF PRE-ANALYTICAL VARIATION IN CEREBROSPINAL FLUID DETECTED WITH APTAMER SCREENING
Eline A. J. Willemse1,2, Kees WJ. van Uffelen1,2, Lev Yener1,2, Mark A. van de Wiel2, Charlotte E. Teunissen1,2, 1Neuroscience Campus Amsterdam, Amsterdam, Netherlands; 2VU University Medical Center, Amsterdam, Netherlands. Contact e-mail: [email protected] Background: During different steps of cerebrospinal fluid (CSF)
processing, pre-analytical variation factors could artificially influence proteins, resulting in changed biomarker levels not related
P481
to the patient biology. Pre-analytical variation, due to e.g. lot-tolot variation, accounted for up to 25% of incorrect AD classifications based on the CSF profile (Vos et al., 2014, PlosOne). To avoid these confounding effects, we aimed to identify quality markers for CSF and, complementary, study the stability of CSF proteins under pre-analytical conditions in general. In the current study we aimed to detect changes in protein levels during specific pre-analytical conditions. Methods: We experimentally exposed CSF to extreme pre-analytical conditions: delayed processing at room temperature and 4 C for 1, 2, and 24 hours; delayed storage at room temperature and 4 C for 1 hour, 1 day, and 1 week; and 1, 4, and 8 freeze-andthaw cycles. For each pre-analytical condition 3 pools were prepared. A reference sample per pool was kept at -80⁰C from the start. The CSF samples were measured with Somascan, to analyze the levels of 1129 proteins using aptamers. Effects of the treatments were statistically analyzed with ANOVA, corrected for pool number and for multiple testing (Bonferroni correction). Results: We detected 8 markers with significantly changed levels after a delay of 24 hours between collection and centrifugation at room temperature. For the delay between centrifugation and storage, we found 8 different markers to be significantly changed after 1 week at room temperature. No changes in protein levels were observed during time delays at 4⁰C. None of the levels of 1129 markers were changed after up to 8 freeze-and-thaw cycles. Conclusions: Most proteins in CSF are stable upon extreme pre-analytical conditions. For both delayed processing and delayed storage 8 different markers were detected which were significantly altered in the most extreme condition. These markers need to be validated for their potential use as sentinel molecules for CSF quality. CSF seems insensitive for multiple freeze-and-thaw steps for a large number of proteins. Our results provide support for studying CSF with more confidence regarding the effects of potential pre-analytical confounders.
P1-202
THE BIOMARKER-BASED DIAGNOSIS OF ALZHEIMER’S DISEASE: LESSONS FROM ONCOLOGY
Marina Boccardi1,2, Valentina Gallo3, Yutaka Yasui4,5, Paolo Vineis6, Alessandro Padovani7,8, Urs Mosimann9,10, Panteleimon Giannakopoulos11, Gabriel Gold12, Bruno Dubois13,14, Clifford R. Jack Jr.15, Bengt Winblad16, Giovanni B. Frisoni17,18, Emiliano Albanese19, the Geneva Task Force for the Roadmap of Alzheimer’s Biomarkers, 1IRCCS S Giovanni di Dio Fatebenefratelli, Brescia, Italy; 2LANVIE - Laboratory of Neuroimaging of Aging, University of Geneva, Geneva, Switzerland; 3Queen Mary, University of London, Barts and the London School of Medicine, Centre of Primary Care and Public Health, Blizard Institute, London, United Kingdom; 4St. Jude Children’s Research Hospital, Memphis, TN, USA; 5School of Public Health, University of Alberta, Edmonton, AB, Canada; 6School of Public Health, Imperial College of London, London, United Kingdom; 7Spedali Civili di Brescia, Brescia, Italy; 8University of Brescia, Brescia, Italy; 9 Gerontechnology and Rehabilitation Group, Bern, Switzerland; 10 University Hospital of Old Age Psychiatry and Psychotherapy, University of Bern, Bern, Switzerland; 11University Hospital of Geneva, Geneva, Switzerland; 12Division of Geriatrics, University Hospital of Geneva, Geneva, Switzerland; 13APHP, Paris, France; 14Sorbonne Universites, Universite Pierre et Marie Curie-Paris 6, Paris, France; 15Mayo Clinic, Rochester, MN, USA; 16Karolinska Institutet, Center for Alzheimer Research, Div. of Neurogeriatrics, Huddinge, Sweden; 17Memory Clinic and
P482
Poster Presentations: Sunday, July 24, 2016
LANVIE - Laboratory of Neuroimaging of Aging, University Hospitals and University of Geneva, Geneva, Switzerland; 18IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, Brescia, Italy; 19University of Geneva and University Hospitals of Geneva (HUG), Geneva, Switzerland. Contact e-mail: [email protected]
The use biomarkers for Alzheimer’s disease (AD) in clinical settings is regulated relatively loosely compared, for example, to requirements for introducing new drugs. In 2001, the framework used for drugs validation was adapted to design a strict and systematic validation procedure for oncology biomarkers. This work aims to adapt the oncology framework AD specific biomarkers. Methods: The 5phases framework by Pepe et al (Journal of the National Cancer Institute, 93(14), 2001) was adapted to meet the specificity of Background:
validation studies of biomarker for AD. Adaptations were made to: specific terms; types of studies design; context of use; target population. Limitations of these adaptations were thoroughly considered. Results: The adaptation led to Incidental and Substantial differences of the “AD”, compared to the “oncology”, framework. Incidental differences relate to target tissue (brain vs tumor), specific outcomes (disability, morbidity, institutionalization, quality of life, caregivers burden vs mortality), and study designs (prospective vs retrospective). Substantial differences relate to the target population and to the possible use of biomarkers within the two frameworks. As to target population, this validation framework is restricted to the MCI population, due to the need of early detection of clinical disease, to the fact that clinical criteria do not recommend
Phases for the development of biomarkers as adapted from the oncology framework (Pepe et al., J Natl Cancer Inst 2001) to the case of the pre-dementia diagnosis of Alzheimer’s disease. PHASES Phase 1 Pilot Studies Phase 2 Clinical Assay Development for Clinical Disease
AIMS
description
Primary Aims
To identify leads for potentially useful biomarkers and prioritize identified leads.
Primary Aim
To estimate the true and false positive rate or ROC curve and assess its ability to distinguish subjects with and without the disease. To optimize procedures for performing the assay and to assess the reproducibility of the assay within and between laboratories. To determine the relationship between biomarker tissue measurements made on tissue (phase 1) and the biomarker measurements made on the noninvasive clinical specimen (phase 2). To assess factors (e.g. sex, age, etc.), associated with biomarker status or level in control subjects. If such factors affect the biomarker. thresholds for test positivity may need to be defined separately for target subpopulations. To assess factors associated with biomarker status or level in diseased subjects—in particular, disease characteristics.
Secondary Aim 1 Secondary Aim 2 Secondary Aim 3
Secondary Aim 4 Phase 3 Prospective Longitudina 1 Repository Studies
Primary Aim 1
To evaluate the capacity of the biomarker to detect the earliest disease stages.
Primary Aim 2 Secondary Aim 1
To define criteria for a biomarker positive test in preparation for phase 4. To explore the impact of covariates on the discriminatory abilities of the biomarker before clinical diagnosis. To compare markers with a view to selecting those that are most promising. To develop algorithms for positivity based on combinations of markers. To determine a biomarker testing interval for phase 4 if repeated testing is of interest.
Secondary Aim 2 Secondary Aim 3 Secondary Aim 4 Phase 4 Prospective Diagnostic Studies
Primary Aim
Secondary Aim 1 Secondary Aim 2 Secondary Aim 3 Secondary Aim 4 Phase 5 Disease Control Studies
Primary Aim Secondary Aim 1 Secondary Aim 2 Secondary Aim 3
To determine the operating characteristics of the biomarker-based test in a relevant population by determining the detection rate and the false referral rate. Studies at this stage involve testing people and lead to diagnosis and treatment. To describe the characteristics of disease detected by the biomarker test–in particular, with regard to the potential benefit incurred by early detection. To assess the practical feasibility of implementing the diagnostic program and compliance of test-positive subjects with work-up and treatment recommendations. To make preliminary assessments of the effects of biomarker testing on costs and mortality associated with the disease. To monitor disease occurring clinically but not detected by the biomarker testing protocol. To estimate the reductions in disease-associated mortality, morbidity. and disability afforded by biomarker testing. To obtain information about the costs of biomarker testing and treatment and the cost per life saved or per quality-adjusted life year. To evaluate compliance with testing and work-up in a diverse range of settings. To compare different biomarker testing protocols and/or to compare different approaches to treating test positive subjects in regard to effects on mortality and costs.
Poster Presentations: Sunday, July 24, 2016
preclinical diagnosis for ethical reasons, and to the possibility of using “conversion to dementia” as a gold standard for diagnosis (in the lack of pathology data). The resulting 5 sequential phases were: 1) pilot studies, 2) clinical assay development for clinical disease, 3) prospective longitudinal repository studies, 4) prospective diagnostic studies, and 5) disease control studies (Table). Because of the required adaptations, biomarkers for AD can be used for biomarker-based diagnoses and not yet for screening purposes. Conclusions: The adaptation of the oncology framework to AD aims to systematize the validation of AD biomarkers. The important limitations restrict the generalizability of results to the general population and the use of such biomarkers for screening purposes. This initiative should be considered as a first, although necessary, step to the definition of a systematic validation of biomarkers for AD.
P1-203
VERIFICATION OF THE ANALYTICAL PERFORMANCE OF THE FULLY AUTOMATED LUMIPULSE G b-AMYLOID 1-42 AND LUMIPULSE G TOTAL TAU ASSAYS
Annelies Vandersteen1, Sandra Pereson1, Kathleen Gorteman1, Wim Vandezande1, Filip Dekeyser1, Tinne Dumont1, Roger Moonen2, Els Huyck1, Manu Vandijck2, Martine Dauwe2, Vesna Kostanjevecki2, Geert Jannes2, 1Fujirebio Europe N.V., Gent, Belgium; 2Fujirebio Europe, Gent, Belgium. Contact e-mail: [email protected] Background: Today b-amyloid(1-42) peptide (Ab1-42) and total tau
protein levels in cerebrospinal fluid (CSF) are well-accepted biomarkers representing Alzheimer’s disease (AD) progression from the earliest stages on. Widespread use of these biomarkers in AD diagnosis requires reliable, highly precise and accurate measurements. The current study evaluates the novel LumipulseÒ G b-Amyloid 1-42 and LumipulseÒ G Total Tau assays on these analytical aspects. Methods: The LUMIPULSEÒ G instrument series use single, ready-to-use immunoreaction cartridges with a throughput of 60 and 120 tests/hour for the G600II and the G1200 instrument, respectively. Each cartridge generates quantitative results for one biomarker within approximately 30 minutes and multiple assays can be easily combined in the system enabling full characterization of samples during one run. The Lumipulse G b-Amyloid 1-42 and Lumipulse G Total Tau assays were developed using established monoclonal antibodies. Analytical assay performance was characterized according to CLSI guidelines. Results: Precision evaluation using a panel of CSF sample preparations resulted in total CV 8% for both Lumipulse G b-Amyloid 1-42 and Total Tau assays. The LoD and LoQ for the Lumipulse G b-Amyloid 1-42 and Total Tau assays were shown to be <10 pg/mL and <15 pg/mL, and <40 pg/mL and <60 pg/mL, respectively. No high dose hook effect was observed for samples containing up to 200.000 pg/mL of Ab142 and total tau protein, respectively. Linearity was shown across the clinical application range for both assays. In a method comparison study with the routinely used INNOTEST b-AMYLOID(1-42) and INNOTEST hTAU Ag assays, good correlation (r>0.90) was demonstrated. In addition, a good correlation to the recently approved reference measurement procedure (RMP) for Ab1-42was shown. Conclusions: Automation, the mono test cartridge principle, short reaction times and instrument flexibility are key attributes of
P483
the LUMIPULSE G instrument series making it the ideal platform to fulfill today’s needs for rapid and accurate quantification of CSF biomarkers in both low and high throughput clinical laboratories. The novel Ab1-42 and total tau assays on the LUMIPULSE G instruments show good sensitivity and precision and correlate well with the established INNOTEST assays and the RMP for Ab1-42.
P1-204
NEUROGRANIN: A SPECIFIC SYNAPTIC BIOMARKER FOR AMNESTIC ALZHEIMER’S DISEASE?
Henrietta Wellington1, Ross W. Paterson1, Ulrika T€ornqvist2, Erik Portelius3, Nick C. Fox4, Kaj Blennow3, Henrik Zetterberg3,5, Jon M. Schott4, 1University College London, London, United Kingdom; 2 University of Gothenberg, Gothenbergy, Sweden; 3Institute of Neuroscience and Physiology, The Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden; 4UCL Institute of Neurology, London, United Kingdom; 5University College London, Institute of Neurology, London, United Kingdom. Contact e-mail: [email protected] Background: Synaptic dysfunction is an early event in Alzheimer’s
disease (AD), occurring in the prodromal and even pre-symptomatic stage of the disease, when damage is potentially reversible and treatments may be most effective. Several studies have confirmed that neurogranin (Ng), a neuron specific post-synaptic protein, is significantly higher in cerebrospinal (CSF) from AD patients compared to controls; and we recently reported that Ng elevation is specific for AD as opposed to other primary dementias and parkinsonian conditions. This specificity could reflect AD pathology per se, or that AD pathology is predominantly seen in regions (e.g. the hippocampi), where Ng is preferentially expressed. To investigate this, we aimed to compare Ng levels in typical amnestic AD with those with posterior cortical atrophy (PCA), an atypical variant of AD where the pathology is more focussed on parietal-occipital regions, with relative hippocampal sparing. Methods: We measured Ab42, total-tau (T-tau) and phospho-tau (P-tau), and a novel ELISA assay to measure Ng in the CSF of 162 subjects, comprising controls (n¼27), and patients with amnestic AD (n¼104) and fulfilling clinical criteria for PCA (n¼31). Ng levels between groups were compared using Kruskal Wallis test and Dunns multiple comparisons test. Spearman’s correlation was performed to explore the association of Ng with Ab42, T-tau and P-tau. Statistical significance was set at P<0.05. Results: Ng levels were significantly higher in AD patients (447 [median], 275-635 [interquartile range] pg/ml) compared to controls (207, 137-284 pg/ml, P<0.0001). CSF Ng levels in PCA patients were significantly lower (291, 193-396 pg/ml) compared to patients with amnestic AD (494, 311-739 pg/ml, P<0.0001). We found a negative correlation between Ng and Ab42 in the whole dataset (rs¼ -0.21, P<0.01) but a stronger, positive, correlation with T-tau and P-tau (rs¼0.79 and 0.76 respectively, both P<0.0001). Conclusions: We confirm an increase in CSF Ng in AD patients as previously reported. We show that Ng is significantly lower in patients with the atypical AD variant PCA compared to amnestic AD suggesting that Ng elevation in AD may relate to the presence and degree of hippocampal/limbic AD pathology. Ng may be a specific biomarker for typical, amnestic AD.
930-1360
630-715
520-525
320-345
100-200
5.18 1.62 6.53 8.49
1.96 3.52 0.00 4.03
1.76 3.45 3.21 5.03
1.54 4.20 5.07 6.76
2.86 4.38 4.85 7.13
4.76 0.00 3.12 5.69
4.58 0.00 3.99 6.07
5.99 2.01 3.45 7.19
4.50 3.33 2.20 6.01
7.19 0.00 4.11 8.28
Background: The integration of CSF biomarker analysis in clinical routine or its global use in clinical trials is hampered in-part by the presence of different analyte concentrations in function of the assay design (e.g., matrix interference), the applied technology, a too high inter-center variability (reproducibility), and the absence of international reference standard materials. Part of the above mentioned problems were already solved recently by the integration in the market of a new series of immuno-assays or modified assay formats of existing assays for the most important CSF proteins. New designs focused, besides other aspects, on the removal of matrix interference and the integration of temperature-stable, ready-to-use calibrators. However, the manual test procedures for these new assays still do not solve completely the high inter-center variability. The goal: To develop an approach to transfer and qualify a manual test protocol for the CSF proteins (Ab(1-42), Ab(1-40), total tau) to fully automated test procedures. Methods: The manual test procedure was optimized for automated sample testing on two different instruments: the BepIII (Siemens, Erlangen, Germany) and the Janus (Perkin-Elmer, Waltham, MA, USA). Precision (reproducibility, between-run (rr) CV%, between-day (dd) CV%) after transfer was evaluated according to CLSI EP-5A-2 (10 days procedures) using neat CSF samples. Results: The results of the precision study are shown in Table 1. Precision data are much lower than published as part of the Alzheimer’s Association Quality Control Program. Conclusions: We have been able to transfer the manual test procedure as provided by the vendor to a fully-automated sample analysis, resulting in an improved test reproducibility for CSF and improved total CV% at the clinical decision level (520-715 pg/mL). The latter is considered as a cost-efficient (low investment) manner (automated colorimetric ELISA) for biomarker integration into clinical routine when compared with an IVD closed system.
P1-201
QUALITY INDICATORS OF PRE-ANALYTICAL VARIATION IN CEREBROSPINAL FLUID DETECTED WITH APTAMER SCREENING
Eline A. J. Willemse1,2, Kees WJ. van Uffelen1,2, Lev Yener1,2, Mark A. van de Wiel2, Charlotte E. Teunissen1,2, 1Neuroscience Campus Amsterdam, Amsterdam, Netherlands; 2VU University Medical Center, Amsterdam, Netherlands. Contact e-mail: [email protected] Background: During different steps of cerebrospinal fluid (CSF)
processing, pre-analytical variation factors could artificially influence proteins, resulting in changed biomarker levels not related
P481
to the patient biology. Pre-analytical variation, due to e.g. lot-tolot variation, accounted for up to 25% of incorrect AD classifications based on the CSF profile (Vos et al., 2014, PlosOne). To avoid these confounding effects, we aimed to identify quality markers for CSF and, complementary, study the stability of CSF proteins under pre-analytical conditions in general. In the current study we aimed to detect changes in protein levels during specific pre-analytical conditions. Methods: We experimentally exposed CSF to extreme pre-analytical conditions: delayed processing at room temperature and 4 C for 1, 2, and 24 hours; delayed storage at room temperature and 4 C for 1 hour, 1 day, and 1 week; and 1, 4, and 8 freeze-andthaw cycles. For each pre-analytical condition 3 pools were prepared. A reference sample per pool was kept at -80⁰C from the start. The CSF samples were measured with Somascan, to analyze the levels of 1129 proteins using aptamers. Effects of the treatments were statistically analyzed with ANOVA, corrected for pool number and for multiple testing (Bonferroni correction). Results: We detected 8 markers with significantly changed levels after a delay of 24 hours between collection and centrifugation at room temperature. For the delay between centrifugation and storage, we found 8 different markers to be significantly changed after 1 week at room temperature. No changes in protein levels were observed during time delays at 4⁰C. None of the levels of 1129 markers were changed after up to 8 freeze-and-thaw cycles. Conclusions: Most proteins in CSF are stable upon extreme pre-analytical conditions. For both delayed processing and delayed storage 8 different markers were detected which were significantly altered in the most extreme condition. These markers need to be validated for their potential use as sentinel molecules for CSF quality. CSF seems insensitive for multiple freeze-and-thaw steps for a large number of proteins. Our results provide support for studying CSF with more confidence regarding the effects of potential pre-analytical confounders.
P1-202
THE BIOMARKER-BASED DIAGNOSIS OF ALZHEIMER’S DISEASE: LESSONS FROM ONCOLOGY
Marina Boccardi1,2, Valentina Gallo3, Yutaka Yasui4,5, Paolo Vineis6, Alessandro Padovani7,8, Urs Mosimann9,10, Panteleimon Giannakopoulos11, Gabriel Gold12, Bruno Dubois13,14, Clifford R. Jack Jr.15, Bengt Winblad16, Giovanni B. Frisoni17,18, Emiliano Albanese19, the Geneva Task Force for the Roadmap of Alzheimer’s Biomarkers, 1IRCCS S Giovanni di Dio Fatebenefratelli, Brescia, Italy; 2LANVIE - Laboratory of Neuroimaging of Aging, University of Geneva, Geneva, Switzerland; 3Queen Mary, University of London, Barts and the London School of Medicine, Centre of Primary Care and Public Health, Blizard Institute, London, United Kingdom; 4St. Jude Children’s Research Hospital, Memphis, TN, USA; 5School of Public Health, University of Alberta, Edmonton, AB, Canada; 6School of Public Health, Imperial College of London, London, United Kingdom; 7Spedali Civili di Brescia, Brescia, Italy; 8University of Brescia, Brescia, Italy; 9 Gerontechnology and Rehabilitation Group, Bern, Switzerland; 10 University Hospital of Old Age Psychiatry and Psychotherapy, University of Bern, Bern, Switzerland; 11University Hospital of Geneva, Geneva, Switzerland; 12Division of Geriatrics, University Hospital of Geneva, Geneva, Switzerland; 13APHP, Paris, France; 14Sorbonne Universites, Universite Pierre et Marie Curie-Paris 6, Paris, France; 15Mayo Clinic, Rochester, MN, USA; 16Karolinska Institutet, Center for Alzheimer Research, Div. of Neurogeriatrics, Huddinge, Sweden; 17Memory Clinic and
P482
Poster Presentations: Sunday, July 24, 2016
LANVIE - Laboratory of Neuroimaging of Aging, University Hospitals and University of Geneva, Geneva, Switzerland; 18IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, Brescia, Italy; 19University of Geneva and University Hospitals of Geneva (HUG), Geneva, Switzerland. Contact e-mail: [email protected]
The use biomarkers for Alzheimer’s disease (AD) in clinical settings is regulated relatively loosely compared, for example, to requirements for introducing new drugs. In 2001, the framework used for drugs validation was adapted to design a strict and systematic validation procedure for oncology biomarkers. This work aims to adapt the oncology framework AD specific biomarkers. Methods: The 5phases framework by Pepe et al (Journal of the National Cancer Institute, 93(14), 2001) was adapted to meet the specificity of Background:
validation studies of biomarker for AD. Adaptations were made to: specific terms; types of studies design; context of use; target population. Limitations of these adaptations were thoroughly considered. Results: The adaptation led to Incidental and Substantial differences of the “AD”, compared to the “oncology”, framework. Incidental differences relate to target tissue (brain vs tumor), specific outcomes (disability, morbidity, institutionalization, quality of life, caregivers burden vs mortality), and study designs (prospective vs retrospective). Substantial differences relate to the target population and to the possible use of biomarkers within the two frameworks. As to target population, this validation framework is restricted to the MCI population, due to the need of early detection of clinical disease, to the fact that clinical criteria do not recommend
Phases for the development of biomarkers as adapted from the oncology framework (Pepe et al., J Natl Cancer Inst 2001) to the case of the pre-dementia diagnosis of Alzheimer’s disease. PHASES Phase 1 Pilot Studies Phase 2 Clinical Assay Development for Clinical Disease
AIMS
description
Primary Aims
To identify leads for potentially useful biomarkers and prioritize identified leads.
Primary Aim
To estimate the true and false positive rate or ROC curve and assess its ability to distinguish subjects with and without the disease. To optimize procedures for performing the assay and to assess the reproducibility of the assay within and between laboratories. To determine the relationship between biomarker tissue measurements made on tissue (phase 1) and the biomarker measurements made on the noninvasive clinical specimen (phase 2). To assess factors (e.g. sex, age, etc.), associated with biomarker status or level in control subjects. If such factors affect the biomarker. thresholds for test positivity may need to be defined separately for target subpopulations. To assess factors associated with biomarker status or level in diseased subjects—in particular, disease characteristics.
Secondary Aim 1 Secondary Aim 2 Secondary Aim 3
Secondary Aim 4 Phase 3 Prospective Longitudina 1 Repository Studies
Primary Aim 1
To evaluate the capacity of the biomarker to detect the earliest disease stages.
Primary Aim 2 Secondary Aim 1
To define criteria for a biomarker positive test in preparation for phase 4. To explore the impact of covariates on the discriminatory abilities of the biomarker before clinical diagnosis. To compare markers with a view to selecting those that are most promising. To develop algorithms for positivity based on combinations of markers. To determine a biomarker testing interval for phase 4 if repeated testing is of interest.
Secondary Aim 2 Secondary Aim 3 Secondary Aim 4 Phase 4 Prospective Diagnostic Studies
Primary Aim
Secondary Aim 1 Secondary Aim 2 Secondary Aim 3 Secondary Aim 4 Phase 5 Disease Control Studies
Primary Aim Secondary Aim 1 Secondary Aim 2 Secondary Aim 3
To determine the operating characteristics of the biomarker-based test in a relevant population by determining the detection rate and the false referral rate. Studies at this stage involve testing people and lead to diagnosis and treatment. To describe the characteristics of disease detected by the biomarker test–in particular, with regard to the potential benefit incurred by early detection. To assess the practical feasibility of implementing the diagnostic program and compliance of test-positive subjects with work-up and treatment recommendations. To make preliminary assessments of the effects of biomarker testing on costs and mortality associated with the disease. To monitor disease occurring clinically but not detected by the biomarker testing protocol. To estimate the reductions in disease-associated mortality, morbidity. and disability afforded by biomarker testing. To obtain information about the costs of biomarker testing and treatment and the cost per life saved or per quality-adjusted life year. To evaluate compliance with testing and work-up in a diverse range of settings. To compare different biomarker testing protocols and/or to compare different approaches to treating test positive subjects in regard to effects on mortality and costs.
Poster Presentations: Sunday, July 24, 2016
preclinical diagnosis for ethical reasons, and to the possibility of using “conversion to dementia” as a gold standard for diagnosis (in the lack of pathology data). The resulting 5 sequential phases were: 1) pilot studies, 2) clinical assay development for clinical disease, 3) prospective longitudinal repository studies, 4) prospective diagnostic studies, and 5) disease control studies (Table). Because of the required adaptations, biomarkers for AD can be used for biomarker-based diagnoses and not yet for screening purposes. Conclusions: The adaptation of the oncology framework to AD aims to systematize the validation of AD biomarkers. The important limitations restrict the generalizability of results to the general population and the use of such biomarkers for screening purposes. This initiative should be considered as a first, although necessary, step to the definition of a systematic validation of biomarkers for AD.
P1-203
VERIFICATION OF THE ANALYTICAL PERFORMANCE OF THE FULLY AUTOMATED LUMIPULSE G b-AMYLOID 1-42 AND LUMIPULSE G TOTAL TAU ASSAYS
Annelies Vandersteen1, Sandra Pereson1, Kathleen Gorteman1, Wim Vandezande1, Filip Dekeyser1, Tinne Dumont1, Roger Moonen2, Els Huyck1, Manu Vandijck2, Martine Dauwe2, Vesna Kostanjevecki2, Geert Jannes2, 1Fujirebio Europe N.V., Gent, Belgium; 2Fujirebio Europe, Gent, Belgium. Contact e-mail: [email protected] Background: Today b-amyloid(1-42) peptide (Ab1-42) and total tau
protein levels in cerebrospinal fluid (CSF) are well-accepted biomarkers representing Alzheimer’s disease (AD) progression from the earliest stages on. Widespread use of these biomarkers in AD diagnosis requires reliable, highly precise and accurate measurements. The current study evaluates the novel LumipulseÒ G b-Amyloid 1-42 and LumipulseÒ G Total Tau assays on these analytical aspects. Methods: The LUMIPULSEÒ G instrument series use single, ready-to-use immunoreaction cartridges with a throughput of 60 and 120 tests/hour for the G600II and the G1200 instrument, respectively. Each cartridge generates quantitative results for one biomarker within approximately 30 minutes and multiple assays can be easily combined in the system enabling full characterization of samples during one run. The Lumipulse G b-Amyloid 1-42 and Lumipulse G Total Tau assays were developed using established monoclonal antibodies. Analytical assay performance was characterized according to CLSI guidelines. Results: Precision evaluation using a panel of CSF sample preparations resulted in total CV 8% for both Lumipulse G b-Amyloid 1-42 and Total Tau assays. The LoD and LoQ for the Lumipulse G b-Amyloid 1-42 and Total Tau assays were shown to be <10 pg/mL and <15 pg/mL, and <40 pg/mL and <60 pg/mL, respectively. No high dose hook effect was observed for samples containing up to 200.000 pg/mL of Ab142 and total tau protein, respectively. Linearity was shown across the clinical application range for both assays. In a method comparison study with the routinely used INNOTEST b-AMYLOID(1-42) and INNOTEST hTAU Ag assays, good correlation (r>0.90) was demonstrated. In addition, a good correlation to the recently approved reference measurement procedure (RMP) for Ab1-42was shown. Conclusions: Automation, the mono test cartridge principle, short reaction times and instrument flexibility are key attributes of
P483
the LUMIPULSE G instrument series making it the ideal platform to fulfill today’s needs for rapid and accurate quantification of CSF biomarkers in both low and high throughput clinical laboratories. The novel Ab1-42 and total tau assays on the LUMIPULSE G instruments show good sensitivity and precision and correlate well with the established INNOTEST assays and the RMP for Ab1-42.
P1-204
NEUROGRANIN: A SPECIFIC SYNAPTIC BIOMARKER FOR AMNESTIC ALZHEIMER’S DISEASE?
Henrietta Wellington1, Ross W. Paterson1, Ulrika T€ornqvist2, Erik Portelius3, Nick C. Fox4, Kaj Blennow3, Henrik Zetterberg3,5, Jon M. Schott4, 1University College London, London, United Kingdom; 2 University of Gothenberg, Gothenbergy, Sweden; 3Institute of Neuroscience and Physiology, The Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden; 4UCL Institute of Neurology, London, United Kingdom; 5University College London, Institute of Neurology, London, United Kingdom. Contact e-mail: [email protected] Background: Synaptic dysfunction is an early event in Alzheimer’s
disease (AD), occurring in the prodromal and even pre-symptomatic stage of the disease, when damage is potentially reversible and treatments may be most effective. Several studies have confirmed that neurogranin (Ng), a neuron specific post-synaptic protein, is significantly higher in cerebrospinal (CSF) from AD patients compared to controls; and we recently reported that Ng elevation is specific for AD as opposed to other primary dementias and parkinsonian conditions. This specificity could reflect AD pathology per se, or that AD pathology is predominantly seen in regions (e.g. the hippocampi), where Ng is preferentially expressed. To investigate this, we aimed to compare Ng levels in typical amnestic AD with those with posterior cortical atrophy (PCA), an atypical variant of AD where the pathology is more focussed on parietal-occipital regions, with relative hippocampal sparing. Methods: We measured Ab42, total-tau (T-tau) and phospho-tau (P-tau), and a novel ELISA assay to measure Ng in the CSF of 162 subjects, comprising controls (n¼27), and patients with amnestic AD (n¼104) and fulfilling clinical criteria for PCA (n¼31). Ng levels between groups were compared using Kruskal Wallis test and Dunns multiple comparisons test. Spearman’s correlation was performed to explore the association of Ng with Ab42, T-tau and P-tau. Statistical significance was set at P<0.05. Results: Ng levels were significantly higher in AD patients (447 [median], 275-635 [interquartile range] pg/ml) compared to controls (207, 137-284 pg/ml, P<0.0001). CSF Ng levels in PCA patients were significantly lower (291, 193-396 pg/ml) compared to patients with amnestic AD (494, 311-739 pg/ml, P<0.0001). We found a negative correlation between Ng and Ab42 in the whole dataset (rs¼ -0.21, P<0.01) but a stronger, positive, correlation with T-tau and P-tau (rs¼0.79 and 0.76 respectively, both P<0.0001). Conclusions: We confirm an increase in CSF Ng in AD patients as previously reported. We show that Ng is significantly lower in patients with the atypical AD variant PCA compared to amnestic AD suggesting that Ng elevation in AD may relate to the presence and degree of hippocampal/limbic AD pathology. Ng may be a specific biomarker for typical, amnestic AD.