Advanced Characterization of Silicone Oil Droplets in Protein Therapeutics Using Artificial Intelligence Analysis of Imaging Flow Cytometry Data

Abstract

Monitoring protein particles is increasingly emphasized in the development of biopharmaceuticals due to potential immunogenicity. Accurate quantitation of protein particles is complicated by silicone oil droplets, a common pharmaceutical component in pre-filled syringes. Though silicone oil is typically regarded as harmless, numerous reports have indicated protein adsorption may render these particles with immunostimulatory properties. Imaging flow cytometry (IFC) is an emerging pharmaceutical method capable of capturing high-resolution brightfield and fluorescence imagery from samples in suspension. In this study, we created a data analysis strategy using artificial intelligence (AI) software to classify brightfield images collected with IFC as protein or silicone oil. The AI software performs image classification using deep learning with a convolutional neural network architecture, for identification of subtle morphological phenotypes. The AI model yielded robust classification of particles >2 mm across various sources of protein and silicone oil particles and over the instrument life cycle. Next, the AI model was combined with IFC fluorescence images to differentiate potentially immunogenic protein-adsorbed silicone and innocuous native silicone. The methods reported herein provide added analytical capability for characterization of particulate matter in therapeutic formulations, and may be applied for optimization of protein formulations and evaluation of product consistency.

The in vitro micronucleus assay using imaging flow cytometry and deep learning

“…In particular, CNNs which are multi-layered neural networks that use kernel-based processing, are particularly effective at extracting information at multiple levels to comprehensively characterize image data, thereby facilitating enhanced discrimination of populations or identification of image subtleties 43 . As such, CNNs have been used in a number of fields to enhance accuracy and improve workflow including pathology 43 , cell biology 44 , and pharmaceuticals 45 . One disadvantage of CNNs has been the requirement for computer scientists to perform optimization and validation, making them relatively inaccessible to researchers lacking expertise in those areas.…”

“…The development of CNN-based classifiers using AAI is further supported by unique clustering and prediction functionalities which permit rapid ground truth model class assignment of morphologically similar images, greatly streamlining the construction of large training datasets. In a recent publication, we demonstrated that a model created and trained with the AAI software using only Brightfield imagery was able to robustly identify and differentiate silicone oil droplets and protein aggregates, a distinction of particular importance in the development of therapeutic protein formulations 45 .…”

Particles in Biopharmaceutical Formulations, Part 2: An Update on Analytical Techniques and Applications for Therapeutic Proteins, Viruses, Vaccines and Cells

“…135 Interestingly, artificial intelligence has lately also been applied as a tool for particle classification in brightfield channel images from IFC. 138 The authors used a CNN and fluorescence staining for verification to differentiate SO droplets, protein adsorbed silicone as well as protein aggregates. Importantly, image-based particle classification still remains challenging in the low micrometer size range because of limitations of the optical resolution of the applied imaging system.…”

Review: tomographic imaging flow cytometry

“…With this approach, the processing of data is massively parallelized and thus accelerated. [83][84][85] Another approach is the use of field-programmable gate arrays (FPGAs) [86][87][88] or vision system on chip (VSoC) sensors. 89,90 Here, a large part of the pre-processing of the image data is done close to the sensor and only the important data is transmitted to the controlling computer.…”