Protein Therapeutics

At SentrySciences, we understand the analytical challenges of compliance with USP<787> and USP<788> requirements for sub-visible particulates. Light Obscuration, the approved compendial method for subvisible particle testing for injectable drug lot release, is significantly challenged by inherent protein particles. The light obscuration method has known technical limitations when applied to protein drugs that result in under-counting and under-sizing translucent, low-contrast inherent protein aggregates and subvisible particles due to refractive index-matching issues. Orthogonal methods such as flow imaging microscopy (FIM), microfluidic imaging (MFI), and backgrounded membrane imaging (BMI) do not face the same refractive-index-matching limitations. Each of these methods is identified as an orthogonal method to light obscuration for drug lot release. These imaging-based methods provide information on particle size and count, but they lack robust algorithms capable of in-depth analysis of the unique textural and morphological features of the protein aggregates and subvisible particles embedded within their images.

ParticleSentryᴬᴵ software combines deep machine learning and computational statistics to deliver actionable, quantitative data from these orthogonal methods. For the first time, protein formulation development and chemistry, manufacturing, and controls (CMC) scientists have a robust supervised learning software able to quantitatively determine if a collection of protein aggregate images match a known stressor condition or represent a new outlier condition - all with statistical certainty!