Hematology Automation: Challenges Beyond Basic Sample Processing

Automation has transformed routine hematology workflows across diagnostics. Yet the complexity of morphological interpretation means removing expert judgment entirely remains out of reach for most laboratory settings. This article examines where hematology automation stands today, what still limits it, and how Full-Field digital imaging platforms are helping close the gap between processing speed and clinical insight.

What Is Hematology Automation in Modern Laboratories?

Hematology automation refers to the use of integrated systems and artificial intelligence to streamline laboratory workflows, which includes sample processing through to diagnostic interpretation.

Today, automation spans multiple stages of the laboratory process:

• Pre-analytical handling: Sample collection, labeling, and preparation before analysis.
• Instrument-based analysis: Automated cell analyzers deliver reproducible counts and flags across high-volume specimen runs.
• Data management: Laboratory information management systems (LIMS) connect instruments across departments, enabling real-time data sharing and reducing transcription errors. At the analytical level, systems apply consistent parameters across samples, improving reproducibility in high-volume laboratory environments.
• Initial morphological review: AI-supported systems pre-classify cells and flag abnormalities before human review.

Artificial intelligence has further expanded these capabilities. The International Council for Standardisation in Haematology (ICSH) formally endorsed digital imaging and AI-based pre-classification in 2019.⁸

These advances have significantly improved throughput in hematology laboratories. However, sustaining performance still depends on:

• Ongoing staff training
• System interoperability
• Adherence to clinical and quality standards

Why Sample Processing Is No Longer the Primary Automation Barrier

The logistical challenge of moving samples efficiently through the laboratory has largely been addressed. The automated cell analyzer delivers reliable and consistent results for routine specimens, while total laboratory automation (TLA) systems integrate pre-analytical, analytical, and post-analytical workflows.³ When supported by appropriate informatics, these systems improve both efficiency and cost-effectiveness.¹

However, automation has shifted — rather than eliminated — the bottleneck. When abnormal results are flagged, further analysis depends on morphological interpretation. This requires recognising subtle cellular features, rare abnormalities and context-specific variations which may not be fully addressed by machine learning algorithms trained on specific datasets.⁶

Unlike sample processing, these interpretive steps cannot be fully standardised.

In practice, routine samples move efficiently through automated pipelines, while complex or flagged cases still require expert evaluation. The primary limitation therefore now lies in interpretation of slide morphology and result, not the processing of the sample.

Key Automation Limitations in Complex Hematology Diagnostics

Workflows in hematology diagnostics are inherently complex and data-intensive. They typically begin with complete blood counts and extend to detailed morphological evaluation of peripheral blood and bone marrow, with each step informing the next.⁴

Several key limitations continue to limit automation:

1. Domain shift

Artificial Intelligence (AI) models trained in one laboratory environment often underperform in another. Variations in staining, imaging equipment, and patient populations can significantly affect accuracy.⁶ These discrepancies are not always immediately visible, increasing the risk of diagnostic error.

2. Limited interpretability

Deep learning systems often function as “black boxes,” producing outputs without transparent reasoning. In cytomorphology — where diagnosis depends on subtle nuclear and cytoplasmic features — this lack of explainability is a major concern, particularly as diagnoses often involve malignant or extremely serious medical conditions, which are time sensitive.⁷

3. Incomplete sampling

Selective scanning approaches may miss clinically important areas of a slide. Rare or abnormal cells located outside sampled regions can go undetected, particularly in partial-field imaging systems that capture snapshots of selected cell fields, missing the broader slide context required for reliable rare cell detection. Comprehensive coverage of the clinically relevant slide area — including regions where abnormal cells concentrate — is essential for reducing the risk of false-negative findings in complex or high-stakes cases.

4. Adoption barriers

Despite decades of availability, digital morphology is not universally adopted.⁵ Common barriers include:

• Limited training opportunities
• Lack of standardisation
• Data security concerns
• High implementation costs
• Resistance to workflow change

The Role of Expert Review in Automated Hematology Systems

Automation has not replaced morphological expertise, but has redefined where that expertise is deployed and applied.

Manual microscopy remains the reference standard for evaluating peripheral blood and bone marrow smears. Fully digital systems are being used to evaluate the cell morphology in these samples, but only once review cases are flagged by automated systems, obviating the need to assess every sample. This is particularly important for white blood cell classification as well as nucleated red blood cell identification.

These assessments are essential for diagnosing conditions such as leukaemia and lymphoma, yet are prone to error in high-volume, high-fatigue environments.

AI systems support this process by:

• Pre-classifying cells
• Highlighting abnormalities
• Prioritising cases

However, they do not replace clinical judgment. Expert review remains essential to ensure diagnostic accuracy in these time sensitive and potentially life threatening conditions.⁹

There is also a growing concern that reduced exposure to manual microscopy may erode diagnostic confidence in complex cases, however this is largely being displaced by the growing confidence in digital systems which enable accurate and sensitive cell classification and allow more expert review time on regions of interest and complex cases.

Advancing Hematology Automation Beyond Rule-Based Analysis

The next phase of automation in hematology is not about faster processing and rather about generating deeper clinical insight.

Modern hematology produces large and complex datasets, including:

• Digital blood smear images
• Flow cytometry data
• Genomic and multi-omics data

These datasets often exceed what clinicians can interpret within routine time constraints.

AI is beginning to demonstrate value in this space:

• Risk assessment tools for thrombosis and venous thromboembolism
• Treatment planning systems integrating clinical and genomic data
• Prognostic models predicting outcomes and treatment response

Full-Field digital microscopy represents a significant step forward. By capturing entire blood smears at high resolution, it provides a more complete and context-rich view than isolated cell imaging.²

AI-enabled decision support systems can:

• Assess specimen quality
• Identify regions of interest
• Pre-classify thousands of cells
• Generate draft reports for expert review

Remote access further enhances these capabilities by enabling:

• Faster expert consultation
• Reduced diagnostic bottlenecks
• Improved turnaround times

For an automated differential blood test that requires specialist adjudication, remote access can mean the difference between same-day reporting and multi-day delays.

Scopio Labs has developed technology aligned with this approach. The Scopio X100 Full-Field system for peripheral blood smear* analysis uses AI-driven decision support to classify WBCs into multiple subtypes, identify nucleated RBCs, smudge cells, and artefacts, and automate platelet estimation across the entire clinically relevant slide area. Clinical validation demonstrated up to a 60% improvement in workflow efficiency for peripheral blood smear review — FDA-cleared and CE-marked. For laboratories aiming to advance hematology automation, the platform demonstrates how AI decision support can be integrated into real-world diagnostic workflows without removing expert oversight. For bone marrow analysis*, Scopio’s Full-Field platform incorporates whole-slide imaging at 100x, enabling remote expert review* and AI-assisted 500-cell differential without microscope dependency.

Disclaimers:

*Scopio Labs’ Full-Field Peripheral Blood Smear application is CE-marked and FDA-cleared, and it’s commercially available across the U.S., UK and Europe and other territories.
*Scopio’s Full-Field Bone Marrow Aspirate Application is CE-marked cleared for sales in additional regions. Not commercially available in the US for in vitro diagnostic procedures.
*Scopio’s Full-Field remote capabilities are available through the secure hospital network.

References:

1. Alhammad LA, Ainosah TK, Ahmad AM, Samarkandi MS, Jawi NH, Alharthi MA, et al. The impact of laboratory automation on efficiency and accuracy in healthcare settings. Int J Community Med Public Health [Internet]. 2023;11(1):459-463.
2. Fan BE, Yong BSJ, Li R, Wang SSY, Aw MYN, Chia MF, et al. From microscope to micropixels: A rapid review of artificial intelligence for the peripheral blood film. Blood Rev. 2024;64:101144. doi: 10.1016/j.blre.2023.101144
3. Yu HE, Lanzoni H, Steffen T, Derr W, Cannon K, Contreras J, et al. Improving Laboratory Processes with Total Laboratory Automation. Lab Med. 2019;50(1):96-102. doi: 10.1093/labmed/lmy031
4. Daves M, Roccaforte V, Lombardi F, Panella R, Pastori S, Spreafico M, et al. Modern hematology analyzers: beyond the simple blood cells count (with focus on the red blood cells). J Lab Precis Med. 2024;9:4. doi: 10.21037/jlpm-23-32
5. Pozdnyakova O. Hematopathology Practice in the Digital Era: What has Changed? Int J Lab Hematol. 2025 Aug 14. doi: 10.1111/ijlh.14515
6. Chossegros M, Delhommeau F, Stockholm D, Tannier X. Improving the generalizability of white blood cell classification with few-shot domain adaptation. J Pathol Inform. 2024;15:100405. doi: 10.1016/j.jpi.2024.100405
7. Zini G. Hematological cytomorphology: Where we are. Int J Lab Hematol. 2024;46(5):789-794. doi: 10.1111/ijlh.14330
8. Kratz A, Lee SH, Zini G, Riedl JA, Hur M, Machin S; International Council for Standardization in Haematology. Digital morphology analyzers in hematology: ICSH review and recommendations. Int J Lab Hematol. 2019;41(4):437-447. doi: 10.1111/ijlh.13042
9. Mandal PK. Applications of artificial intelligence in hematology: Present and the future. J Hematol Allied Sci. 2026;6:1-3. doi: 10.25259/JHAS_18_2026