The global spread of antimicrobial resistance

Over the last century, the increase in antimicrobial resistance (AMR) has become a pressing global issue. This phenomenon occurs when bacteria become resistant to antimicrobials, rendering the medicine ineffective precisely when it is most needed.

It has been reported that drug-resistant bacteria have a mortality rate approximately three times that of normal bacteria,(*1) and it is predicted that the number of deaths caused each year by drug-resistant bacteria will exceed 10 million by 2050, surpassing the number of deaths due to cancer.(*2)

As the movement of people and goods continues both domestically and abroad, there is growing concern at the global level regarding this issue. In 2021, the WHO declared AMR to be one of the top 10 global public health threats currently facing humanity.(*3) In Japan, the Ministry of Health, Labour and Welfare has announced an “Action Plan for Drug-Resistant Bacteria”.(*4)

An issue impacted by the overwork of Microbiology Laboratory Technician

One of the causes of the rise of AMR is the inappropriate administration of antimicrobials commonly prescribed in hospitals and clinics. This state of affairs is driven by a severe shortage of microbiology laboratory technicians relative to the number of bacteriological tests performed.

When examining a bacteria sample, the first test that a laboratory technician performs to determine its species is the Gram stain, in which bacteria are stained and observed under a microscope to classify them into four types based on color and morphology.

Although Gram staining is useful as a preliminary step towards identifying bacterial species, microbiology laboratory technicians are usually handling a substantial number of examinations manually. The need to determine the color and morphology of numerous species under a microscope, combined with insufficient human resources, means that in Japan it currently takes anywhere from half a day to two days in the slowest cases for results to be reported. Doctors usually cannot afford to wait this long when dealing with sick patients, resulting in cases where ineffective or broad-spectrum antibiotics are administered by mistake and driving an increase in drug-resistant species of bacteria.

First-ever fully-automated AI Gram-staining-to- species-determination solution

Provided that they are used appropriately, antimicrobials are an indispensable part of modern medicine. For this reason, one of the most impactful approaches to addressing the problem of AMR is to identify pathogens as quickly, easily, and inexpensively as possible so that physicians can prescribe more appropriate antimicrobials to their patients.

This is what GramEye, a startup from Osaka University, seeks to achieve with the use of AI and robotics.

GramEye has developed a microbial and cell-staining analyzer that fully automates all processes from Gram staining to classification and identification of bacterial species using AI. The company plans to begin selling the analyzer as a medical device in 2024. While automated Gram-staining technology already exists, GramEye’s analyzer will be the world's first fully-automated device capable of handling every step from staining to bacterial species classification and identification.(*5)

Significantly reduces laboratory technician workload & shortens reporting timeframe

With GramEye’s analyzer, the entire testing process—from staining to bacterial classification—can be completed automatically within 10 minutes, simply by inserting a glass slide coated with a specimen.

One survey has shown that microbiology laboratory technicians spend 40% of their time on Gram staining.6 The introduction of GramEye’s system, which can operate 24 hours a day, would reduce technician workload and free up availability for performing other tasks. As a result, the use of GramEye’s analyzer will help shorten the half-a-day to two-day reporting window currently required by manual testing.

Left: The manual testing process and associated issues
Conventional Gram stain process: Delayed reporting of results, requiring physicians to select antimicrobials without the benefit of microbiology results

Right: Issues that can be addressed by by GramEye’s new analyzer
Automation shortens the time required for Gram staining and allows physicians to prescribe antimicrobials based on actual test results

Potential to serve as a Gram-staining digital image platform through global expansion

GramEye’s analyzer aims to efficiently collect standardized image data that does not depend on laboratory technician competence. As Gram staining is currently a manual process, on a global scale the digital standardization of test images taken from Gram staining has yet to be achieved.

With this being the case, the image data collection capability of GramEye’s device is likely to become a major advantage in the global market. Within the next few years, the company plans to expand its business to Europe and the United States, where laboratory automation is advancing, and it believes that the device has the potential to serve as a platform for collecting and standardizing Gram-stained image data from around the world.

While working towards the global dissemination of its product, GramEye is also considering horizontal development to enable identification of tuberculosis and other acid-fast bacteria, which are a concern for hospital-acquired infections. With a mission of creating a world where antimicrobials are prescribed appropriately, the university-launched startup and their attempts to help solve the global problem of drug-resistant bacteria have recently drawn attention both within Japan and abroad.

Same capabilities as an experienced laboratory technician by 2025

A prototype of GramEye’s analyzer has already been completed, and the company has gathered large-scale Gram-stain image data from dozens of participating medical institutions in preparation for the product’s launch as a medical device in 2024. Moving forward, GramEye will continue to train their AI on regularly-collected data from institutions where the product is being used, thereby delivering exponential improvements to the analyzer’s accuracy.

By 2025, GramEye plans to increase the possible Gram staining classifications from the current standard of four to more than twenty, and also expects to be able to identify bacterial species which are difficult to distinguish using Gram staining. At present, only a small number of skilled microbiologists are capable of performing classification and species identification at this level, and not every hospital or clinic has someone of this proficiency on staff. If GramEye's AI can allow for standardized and highly-sophisticated Gram staining, this could be expected to result in improved antimicrobial selection by physicians as well as greater efficiency in the post-Gram stain testing process.

• (*1) Nature, “The staggering death toll of drug-resistant bacteria” (Tosin Thompson)
• (*2) National Center for Global Health and Medicine, Antimicrobial Resistance Clinical Reference Center (Japanese)
• (*3) WHO,“Antimicrobial resistance”
• (*4) Ministry of Health, Labour and Welfare, “Antimicrobial Resistance Countermeasures” (Japanese; accessed September 10, 2023)
• (*5) GramEye internal research (as of September 2023)
• (*6) GramEye internal research (as of September 2023)