Introduction of artificial intelligence for high throughput culture-based MRSA screening

In the laboratory, AI-based systems that aid in the plate-reading process may increase the overall sample throughput and can also improve the workflow and the overall quality by reducing subjectivity and increasing precision of reads. The APAS Independence in combination with MRSA Analysis module was developed and evaluated using over 17,000 routine specimens over a 6-month period. MRSA screening was performed using chromID® MRSA (bioMérieux). Samples were processed using the AutoPlak® (Beckman Coulter), PREVI Isola, and evaluated after 24 and 48 hours of incubation. Results of APAS AI-based classification were compared to conventional plate reading by experienced medical technicians and microbiologists to determine sensitivity and specificity.

In a clinical setting, the APAS Independence performed with a sensitivity of 100%, and a specificity of 98.1% showing extremely positive results. APAS Independence reliably screens for MRSA and would significantly reduce time to report enabling reprioritisation of technician/microbiologist time.

Poster Presentation: by Labor Dr Wisplinghoff

Conference: ECCMID 2019, Amsterdam

Date: April 2019

Authors: Aurbach U, Wirth S, Gigilo S, Pohl B, Wisplinghoff H

Citation: Aurbach U, Wirth S, Gigilo S, Pohl B, Wisplinghoff H., ECCMID 2019. Introduction of artificial intelligence for high throughput culture-based MRSA screening.