MFI is a more sensitive orthogonal method to LO that detects particles (≥ 1 µm) using bright-field images captured as a solution passes through a flow cell. The detection method relies on refractive index changes rather than blockage of light, which makes it more sensitive to translucent particles, such as protein.

The images collected are used by the software to create an extensive database of information (size, shape, transparency) on the particles detected that can be used to classify and differentiate particles ≥ 5 µm. The ability of MFI to distinguish silicone oil from proteinaceous or extrinsic particles is a significant advantage of this technique.

The visual inspection of images can also provide important information on the types of protein particles being formed (density, fibrillar vs compact, complexes of silicone oil and protein, etc.). MFI has become a valuable technique routinely used to complement compendial methods for subvisible particle detection. Recent additions to such methods, such as USP<788>, include the addition of MFI-like techniques to routine analytical procedures.