Embracing AI-assisted flow cytometry to accelerate innovations in patient care
This guest editorial explores how emerging AI tools are significantly enhancing the utility of flow cytometry across life sciences and highlights how industry leaders must work together to overcome challenges
18 Mar 2026
This guest editorial was written by Zaida Vergara and Giulia Grazia, Global Product Managers for Flow Cytometry, at Beckman Coulter Life Sciences.
A 2025 report on artificial intelligence (AI) from the Massachusetts Institute of Technology presented a startling conclusion: Despite spending more than $30 billion to implement generative AI tools, 95% of businesses are seeing no return on that investment. In the healthcare and pharmaceutical industries, AI has been used to support rote tasks such as documentation and transcription, but the technology has not yet been widely embraced to improve clinical models, the authors concluded1.
The potential for AI to improve patient care is vast, and the technology is already making an impact in some areas of life sciences. For example, in drug development, pharmaceutical researchers are increasingly turning to artificial intelligence to help them make sense of the vast volumes of flow cytometry data they collect while screening thousands of experimental compounds. Thanks to advances in technology, there are now cloud-based platforms that integrate machine learning algorithms that enable scalable, collaborative data analysis.
AI also offers the potential to improve the use of flow cytometry in clinical trials and patient care. Flow cytometry is an essential tool for diagnosing several infectious diseases and cancers, as well as for classifying cell types and finding biomarkers that could predict treatment response. The technology has traditionally required specially trained lab technicians to set up flow cytometers and pathologists to interpret the data. Incorporating AI into these processes would not only accelerate workflows; it could also improve the quality of the data analysis, ultimately leading to better, faster insights that will positively impact patient care.
To get there, the life sciences industry will need to overcome several obstacles, not the least of which is a lack of familiarity and comfort with AI technology among regulators. As laboratories build AI into their flow cytometry workflows, they should be mindful of the need to demonstrate the reliability and reproducibility of the data they’re generating, and to ensure the AI is “explainable,” meaning laboratories using the technology can provide clear explanations of how AI tools generate their results. Here, too, there are a range of emerging technologies that laboratories can use to make AI explainable.
Optimizing flow cytometry data analysis in drug development
In recent years, AI-enabled data analysis with automatic gating has improved flow cytometry workflows in pharmaceutical development, making it easier for researchers to test thousands of compounds at different concentrations on multiple cell or tissue samples. Even with well-defined gating strategies to identify the cell populations of interest, pharmaceutical labs still need specially trained personnel to constantly examine the data and adjust the gates—no minor task when dealing with thousands of samples. AI-enabled tools now allow researchers to define their gating strategies and then use a small subset of samples to train models that can be applied to the remaining samples.
Recently, automatic gating algorithms have been optimized to include key performance indicators (KPI) metrics. KPIs can be used to assess false positives and negatives per population and generate a score to help researchers judge the quality of their models before batch processing. This methodology also helps ensure reproducibility2.
The task of analyzing flow cytometry data is becoming easier, thanks to cloud-based, AI-enabled analysis platforms.As flow cytometry improves, the dimensionality of datasets is increasing, as is the sample size per experiment, raising challenges in data management and visualization, collaboration and analysis. Laboratories can now use cloud-based platforms that integrate analysis algorithms and provide tools that facilitate collaborative clustering, dimensionality reduction and visualization. By using the cloud, labs can easily archive and share the data securely without bearing the expense of purchasing and maintaining their own quantum computers.
The future of flow cytometry in clinical settings
Applying AI to improve the use of flow cytometry in clinical settings holds tremendous promise, but it also raises a host of challenges that industry leaders will need to work together to solve. Among them is collecting the right data—and enough of it—to properly train AI tools for clinical diagnostic use. For example, it’s essential to ensure the data used for training is heterogeneous, so the conclusions generated by AI tools and the resulting recommendations are not biased toward any single demographic group.
Selecting the right data for training will also be essential for preventing diagnostic errors. The rise of generative AI tools such as ChatGPT has brought to light the risk of “hallucination,” where AI fabricates incorrect answers because the datasets it has been trained on are incomplete. In research settings, mistakes are common, but they could be disastrous in clinical settings, where patient lives are on the line.
In addition to building the right datasets for AI training, ensuring explainability will be essential in bringing AI-enabled flow cytometry tools to the clinic. In creating custom AI tools, clinical laboratories will need to collaborate closely with technology vendors to ensure the reliability of the datasets used to train models, and to establish processes for demonstrating how the tools generate the insights and conclusions that guide patient care.
The bottom line is that when it comes to using AI to improve the use of flow cytometry in patient care—from innovating new therapies to diagnosing and treating patients—this technology cannot be a black box. To develop the most effective AI tools, laboratories and their vendors must collaborate to anticipate potential pitfalls and design solutions to address them. AI tools must also be fully explainable, so regulators will gain the confidence they need to support this technology.
References
[1] https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf
[2] Beckman Coulter. How to establish and evaluate machine learning assisted automatic gating to improve reproducibility and reduce time spent on your flow cytometry data analysis.