SGH fungi expert trains AI for infection diagnosis

​SGH Pathology collaborates with Synapxe to develop artificial intelligence that can help identify common fungi species that cause infections in patients.

Did you know that there are over 100,000 species of fungi in the world? From the mushrooms we eat, to those that cannot be seen with the naked eye. With such a wide variety, identifying them is no easy feat. Yet it is crucial to identify the species accurately given that there are over 4,000 clinically important species that may cause life-threatening diseases.

And did you know that Singapore General Hospital has a staff dedicated to the study of fungi, in a field known as Mycology? Meet Tan Mei Gie, Senior Medical Laboratory Scientist in SGH's Diagnostic Bacteriology laboratory.

"As a mycologist, I identify and characterise the different species of fungi, moulds and yeast that cause infections, contributing to diagnosis and treatment of diseases. I also do research to understand the behaviour of these fungi."

"We are seeing increasingly complicated cases that are difficult to identify. With a shortage of trained staff, it can be challenging to provide lab results and diagnosis quickly."

To leverage the capability of Artificial Intelligence (AI), SGH's Department of Microbiology and Synapxe's Data Analytics and AI department came together to develop a computer vision AI model, Fungal Species Detection using Artificial Intelligence (FungAI), to medically identify common fungal species from patients' specimens.

Currently in pilot phase, FungAI aims to digitalise mycology identification processes and enhance diagnostic capabilities through AI technology.

Says Dr Tan Yen Ee, SGH's Senior Consultant in Microbiology, "With the rise in antifungal resistance, it is important to identify fungal species at a faster rate for earlier appropriate intervention. The FungAI technology aims to augment laboratories with limited resources by reducing the reliance on skilled workers which may take many years of training and experience to be proficient in the specialty. This is especially important given the difficulties of getting trained staff in both the current and future job market."

Integration of human expertise 

The project team from Synapxe and SGH Pathology bring together their expert knowledge in IT, bioinformatics and pathology.

The project team first tested the concept with a particular fungal strain which takes about four days to be identified. They were able to design a solution that could halve the time required, allowing for earlier diagnosis and potential productivity gains.

Data collection was a major challenge, says Mei Gie. "Training AI requires high quality dataset. There is limited accurate dataset that can be used to train the AI. So we had to collect whole new sets of data images and document them, which took up most of our time. We also have to get multiple types of data that accurately models the behaviour of the fungi species."

In October 2019, SGH and Synapxe expanded the scope of their project to five species of commonly encountered fungal species. They trained the AI with over 8,500 images collected over two years. Through the experiment, the AI model demonstrated that the turnaround time for identification could be reduced to just two days, with an accuracy of 80 to 90%.

The team endeavours to scale the experiment into a proof-of-concept, focusing on a broader range of fungal species, an efficient data collection process and a simple user interface.

If successful, FungAI may be deployed in the SGH mycology laboratory to address the shortage of skilled expertise in this area by using the AI model to augment junior laboratory staff to identify commonly encountered fungal species. This frees trained staff to work on complicated fungal cases.

Says Mei Gie, "This project required expertise from multiple fields, including computer science, bioinformatics and statistics, to come together. Each challenge highlights the complexity of combining fungal biology with AI, but overcoming them could lead to significant advancements in hospital settings to improve care for our patients."

"I am a curious person and tend to explore beyond what is required for clinical reporting. Being able to solve a complicated case or identify a novel strain bring satisfaction to me and this keeps me going in what I do."

The FungAI project is supported by the Artificial Intelligence Automation (AIA) unit under the Department of Future Health System (FHS) at Singapore General Hospital (SGH). FHS was established to drive care delivery transformation through redesigning care, leveraging automation technology, and developing a future ready workforce. AIA aims to enable adoption of Artificial Intelligence (AI) in the hospital through providing project management support and facilitating strategic collaborations with external partners.

 If you are seeking guidance on how to navigate your SGH innovation journey and bring your ideas to life, please get in touch with FHS by submitting your completed Innovation Idea Form on Infopedia. (infopedia access required)

Edited with some excerpts from Synapxe HealthTech Connect, Aug 2024 issue. Read the full story here https://www.synapxe.sg/blog/artificial-intelligence/leveraging-artificial-intelligence-to-identify-fungi-strains

We love mail! Drop us a note at [email protected] to tell us what you like or didn't like about this story, and what you would like to see more of in LighterNotes.