The WorldSymposiumTM sessions on Tuesday 9 February included a satellite symposium on immunogenicity, diagnostic testing data, and the potential use of artificial intelligence in lysosomal diseases.
Lysosomal diseases: Bench to Bedside and Beyond: understanding immunogenicity, diagnostic testing data, and the potential use of artificial intelligence
Satellite symposium supported by Sanofi Genzyme
Dr Susan Richards (Translational Medicine and Early Development, Sanofi Genzyme, Boston, MA, USA), who discussed the fundamentals of immunogenicity, delivered the first presentation of this symposium. She highlighted that the generation of an immune response to therapeutic proteins can cause the development of anti-drug antibodies, which may affect the efficacy and safety of some treatments.1 Dr Richards noted that both the Food and Drug Administration and the European Medicines Agency use a tiered approach strategy to test for immunogenicity and anti-drug antibodies.2,3 In some instances, anti-drug antibodies may correlate with a patient’s clinical response to treatment. For example, anti-drug antibodies may affect the pharmacodynamic and pharmacokinetic profile and exposure of a therapeutic protein, or its efficacy and safety.2,3 In line with the guidance developed by the European Medicines Agency, Dr Richards noted that the age and genetics of a patient and/or disease-related factors may also influence immunogenicity.3 The results of one publication by Lenders et al. (J Am Soc Nephrol 2018) in the context of immunogenicity and lysosomal storage diseases were highlighted. In this study, the effects of excess anti-drug antibodies on clinical outcomes in male patients with Fabry disease receiving enzyme replacement therapy were investigated. Patients with excess anti-drug antibodies during therapeutic infusion exhibited progressive loss of extracellular glomerular filtration rate and ongoing cardiac hypertrophy.4 These effects were not observed in patients with Fabry disease who had achieved therapeutic protein/antibody equilibrium following infusion with enzyme replacement therapy. Consequently, the results of this study emphasise the need for more personalised treatment of patients with Fabry disease as the development of anti-drug antibodies may affect clinical response to therapeutic interventions.4
The second presentation was delivered by Dr Stefaan Sansen (Rare Diseases, Sanofi Genzyme, Brussels, Belgium) who focused on strategies for diagnosis of rare diseases. In Dr Sansen’s own words, he described the challenges that can complicate rare disease diagnoses. These included low disease awareness, non-specific disease symptoms, lack of infrastructure for diagnostic testing, and variable access and differences in testing strategies. Recommendations of a diagnostic test should be based on evidence from randomised clinical trials, where the strategies are compared using patient-relevant outcomes.5 However, in the field of rare diseases, Dr Sansen highlighted that, in his opinion, cross-sectional and case-control studies are typically used when assessing the accuracy of diagnostic tests. However, use of case-control studies, in which patients with an established disease diagnosis are compared with a healthy control group, may lead to inflated estimates of accuracy when testing a new diagnostic strategy.6 Instead, cross-sectional studies include individuals who have a suspected diagnosis. In these studies, the new diagnostic strategy is tested initially, followed by the index test, which would typically be used to make a diagnosis.6 Dr Sansen then described a cross-sectional study that investigated the clinical utility of the biomarker globotriaosylsphingosine (lyso-Gb3) in diagnosing female patients with Fabry disease. Measurement of alpha-galactosidase A (α-Gal A) enzyme levels in addition to levels of lyso-Gb3 in dried blood spots was shown to enhance the diagnostic detection of Fabry disease in the 11,948 females who were included in the study. Abnormal levels of both α-Gal A and lyso-Gb3 were described as ‘highly suspicious’ indicators of Fabry disease (97% positive predictive value), whereas in cases where only one biochemical value was abnormal, levels of lyso-Gb3 were more indicative of Fabry disease diagnosis compared with α-Gal A (positive predictive values: 39% and 6%, respectively). Normal levels of both α-Gal A and lyso-Gb3 were considered unlikely to be associated with Fabry disease diagnosis.7 Dr Sansen indicated that the results of this cross-sectional study were externally validated in a separate publication. In this study, the positive predictive value of peripheral blood samples of lyso-Gb3 in confirming a diagnosis of Fabry disease was 100% in both males and females, compared with 84% and 58% in males and females, respectively, using assessments of α-Gal A enzyme activity.8 Dr Sansen concluded his presentation by highlighting that, in his opinion, diagnostic strategy accuracy should be taken into account during decision-making in clinical practice, and that this information may be derived from large cross-sectional studies.
The final presentation was given by Dr Amanda Wilson (Real World Evidence, Sanofi Genzyme, Cambridge, MA, USA), who firstly described the difficulties associated with diagnosing rare diseases and the potential of artificial intelligence in aiding diagnoses. In her opinion, the low prevalence, disease heterogeneity and the lack of knowledge of disease natural history can lead to misdiagnosis and under-diagnosis of lysosomal storage diseases, which could ultimately affect patient outcomes. Dr Wilson noted that, in her experience, real-world data have been used to help elucidate the pathway to diagnosis of patients with lysosomal storage diseases. Examples of real-world data, as outlined by Dr Wilson, may include international patient registries, observational studies and case series, which are all typically based on disease-specific datasets.9 From Dr Wilson’s viewpoint, the current real-world data on lysosomal storage diseases are limited as they are based on disease-specific databases and do not include, for example, information from administrative claims databases or electronic health records. However, some examples do exist.10-12 Dr Wilson then noted that, from her perspective, the historical limitations associated with use of current real-world data in lysosomal storage diseases may be overcome by using information from electronic health records and artificial intelligence. Using this large volume of patient information from a large number of patients across a variety of clinics may allow greater integration of patient data. Development of algorithms, or artificial intelligence, and analysis of patient data from electronic health records may lead to the identification of a greater number of disease features, as described by Dr Wilson. These diagnostic algorithms can then be tested on other datasets and used in other clinics or countries. In her opinion, the use of artificial intelligence is yet to be impactful in the field of lysosomal storage diseases; however, using data modelling and algorithms based on electronic health records may lead to:
- Greater identification of patients at risk of a lysosomal storage disease
- Recognition of the most common disease characteristics experience by patients with lysosomal storage diseases and a comparison of diagnosed versus undiagnosed patients
- Earlier detection of patients with an undiagnosed lysosomal storage disease.
Disclaimer: The views expressed here are the views of the presenting physicians. The content presented in this report is not reviewed, approved, or endorsed by WORLDSymposiumTM, or any of its employees, agents, or contractors. No speakers or staff were interviewed directly or involved in the development of this report. Satellite Symposia are not part of the official WORLDSymposiumTM programme and WORLDSymposiumTM does not approve or endorse any commercial products or services discussed during the Satellite Symposia or offered for sale by any corporate supporter of the Satellite Symposia. Unofficial content. Official content is available only to registered attendees of WORLDSymposiumTM 2021. All trademarks are the property of their respective owners.
C-ANPROM/INT/GAUD/0025; Date of preparation: February 2021
- Rosenberg AS, Sauna ZE. Immunogenicity assessment during the development of protein therapeutics. J Pharm Pharmacol 2018; 70: 584-594.
- U. S. Food and Drug Administration. Immunogenicity Testing of Therapeutic Protein Products — Developing and Validating Assays for Anti-Drug Antibody Detection: Guidance for Industry. Available at: https://www.fda.gov/media/119788/download. Accessed February 2021.
- European Medicines Agency. Guideline on Immunogenicity Assessment of Therapeutic Proteins. Available at: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-immunogenicity-assessment-therapeutic-proteins-revision-1_en.pdf. Accessed February 2021.
- Lenders M, Neußer LP, Rudnicki M, et al. Dose-dependent effect of enzyme replacement therapy on neutralizing antidrug antibody titers and clinical outcome in patients with Fabry disease. J Am Soc Nephrol 2018; 29: 2879-2889.
- Buehler AM, Ascef BO, Oliveira Júnior HA, et al. Rational use of diagnostic tests for clinical decision making. Rev Assoc Med Bras (1992) 2019; 65: 452-459.
- Rutjes AWS, Reitsma JB, Vandenbroucke JP, et al. Case-control and two-gate designs in diagnostic accuracy studies. Clin Chem 2005; 51: 1335-1341.
- Balendran S, Oliva P, Sansen S, et al. Diagnostic strategy for females suspected of Fabry disease. Clin Genet 2020; 97: 655-660.
- Stiles AR, Zhang H, Dai J, et al. A comprehensive testing algorithm for the diagnosis of Fabry disease in males and females. Mol Genet Metab 2020; 130: 209-214.
- Gliklich RE, Dreyer NA. Patient registries. In: Leavy MB, ed. Registries for Evaluating Patient Outcomes, 3rd edition: A User's Guide. Rockville, MD: Agency for Healthcare Research and Quality, 2014.
- Reynolds TM, Mewies C, Hamilton J, et al. Identification of rare diseases by screening a population selected on the basis of routine pathology results—the PATHFINDER project: lysosomal acid lipase/cholesteryl ester storage disease substudy. J Clin Pathol 2018; 71: 608-613.
- Ehsani-Moghaddam B, Queenan JA, MacKenzie J, et al. Mucopolysaccharidosis type II detection by Naïve Bayes Classifier: an example of patient classification for a rare disease using electronic medical records from the Canadian Primary Care Sentinel Surveillance Network. PLoS One 2018; 13: e0209018.
- Darbà J, Marsà A. Current status and use of resources of lysosomal storage diseases: analysis of a Spanish claims database. Endocr Metab Immune Disord Drug Targets 2020; 20: 263-270.