Automated Gram Stain Assessment
STATUSProof of Concept
START DATEMay 2018
AutoGram is an extension of LBT’s proven intelligent imaging analysis platform technology, APAS.
The application uses AI combined with whole slide imaging for the automated interpretation of gram stain image analysis.
Challenges for Microbiology labs
Correct gram stain interpretation underpins microbiological interpretation of a sample. Manual microscopy of gram stains is time-consuming and highly specialised.
When scanning large volumes of stained slides, complacency resulting in errors in interpretation can arise.
There is a 9% vacancy rate in the US, and microbiologists’ average age is high (51 AUS / 42 US).
LBT’s proven APAS technology is a flexible image analysis system that can rapidly assess a range of images. The platform technology quickly discards areas of no interest, before performing a detailed analysis of the areas that remain.
LBT has applied this technology and approach to the analysis of digital gram stain images for the identification of gram positive and negative bacteria. This initial proof-of-concept study used LBT’s AI platform technology. It has proved effective at discriminating between:
- Cellular elements: (e.g. epithelial cells, erythrocytes and leucocytes);
- Gram-positive and Gram-negative stained bacteria; and
- Bacterial morphotypes (e.g. cocci and bacilli).
As the trend towards digital pathology grows, there is greater scope for advanced technologies like APAS Gram Stain to assist as a clinical diagnostic tool.