The 'Unsexy' Side of Pathology Informatics: Interoperability and Standardization

When people think of pathology informatics, they often imagine cutting-edge artificial intelligence, whole-slide imaging, or complex algorithms identifying patterns in tumor cells. But the real backbone of pathology informatics—the work that ensures patient data moves seamlessly between systems, institutions, and even countries—is much less glamorous.

Interoperability and standardization of data aren’t flashy topics, but they’re critical to modern healthcare. Without them, even the most advanced AI tools or digital pathology systems become isolated silos of information, limiting their clinical utility.

Why Does Interoperability Matter in Pathology?

Pathology is inherently a data-heavy field. Every biopsy, blood test, or molecular assay generates structured (numerical values, test results) and unstructured (pathology reports, images) data. This data must be:

1. Accessible across different electronic health record (EHR) systems

2. Consistently formatted so it can be analyzed and compared

3. Integrated with clinical decision-making tools

   
   

Yet, pathology informatics is plagued by a lack of true interoperability. Many laboratories use proprietary information systems that don’t communicate well with others. As a result, sending a patient’s pathology report from Hospital A to Hospital B can still involve fax machines, PDFs, and manual data entry—archaic processes that increase the risk of errors and delays.

Case in Point: The Molecular Pathology Report Bottleneck

Consider a patient with non-small cell lung cancer (NSCLC) undergoing molecular testing for targeted therapy selection. A next-generation sequencing (NGS) test performed at one institution might identify an actionable EGFR mutation, but if the patient transfers care to another hospital, that genetic data might not integrate into their new oncologist’s system.

Why? Because molecular pathology reports are often embedded in PDFs rather than structured formats, making it nearly impossible for EHR systems to extract and analyze key mutations automatically. This means oncologists may need to manually review and re-enter data, risking transcription errors and delays in treatment.

The Challenge of Standardization

Even when data is shared, inconsistency in formatting can be a nightmare. Different labs may use different naming conventions, reference ranges, or units for the same test. A simple example:

• One lab reports serum creatinine as 1.2 mg/dL, while another expresses it as 106 µmol/L—same result, different units.

• A molecular lab may report BRAF V600E, while another might write c.1799T>A—same mutation, different notation.

  
  

These discrepancies create unnecessary hurdles for automated clinical decision support (CDS) tools, which rely on standardized inputs to provide actionable recommendations.

HL7, FHIR, and Other Efforts—Not a Magic Fix

Efforts like HL7 (Health Level Seven) and FHIR (Fast Healthcare Interoperability Resources) aim to create universal standards for exchanging healthcare data. While they’ve improved interoperability in some areas, pathology data remains particularly challenging due to its complexity.

For instance, the CAP Cancer Protocols provide standardized pathology report templates, but implementation is inconsistent. LOINC (Logical Observation Identifiers Names and Codes) standardizes lab test names, yet many labs don’t map tests correctly.

Similarly, SNOMED CT (Systematized Nomenclature of Medicine—Clinical Terms) provides standardized codes for diagnoses and laboratory findings, ensuring consistency across systems. However, its adoption remains uneven, limiting its full potential for interoperability.

Real-World Consequences

Poor interoperability and data standardization aren’t just IT headaches—they have real-world implications:

• Delayed Diagnoses – A patient’s lab results might be trapped in a system that doesn’t communicate with a specialist’s platform.

• Increased Costs – Redundant testing happens because prior results are inaccessible.

• Compromised Research – Large-scale studies depend on harmonized datasets, but inconsistent formats make data aggregation difficult.

  
  

Where Do We Go from Here?

1. Push for Structured Data – Pathology reports should be machine-readable, not just PDFs. Standardized synoptic reporting should become the norm.

2. Expand Use of Interoperability Standards – Labs must actively adopt HL7 FHIR, LOINC, and SNOMED CT coding in a meaningful way.

3. Advocate for Vendor Collaboration – Laboratory information system (LIS) vendors should prioritize interoperability instead of locking customers into proprietary ecosystems.

   
   

The work of making pathology data flow seamlessly across systems may not be as exciting as AI-driven diagnostics, but without it, the future of precision medicine remains stuck in the past.