Digital Pathology Image Formats: A Technical Comparison for Workflow Optimization

Digital pathology whole slide images (WSIs) are large data files whose storage, transmission, and analysis depend heavily on the image format chosen at the time of scanning. The format determines compression efficiency, pyramidal structure, metadata retention, and compatibility with analysis tools. This article compares common WSI formats including SVS, MRXS, NDPI, TIFF, and DICOM, and provides practical guidance for selecting formats based on workflow requirements including storage cost, analysis compatibility, and interoperability.

The Role of Image Format in Digital Pathology Workflows

Whole slide imaging converts glass slides into digital files that can be viewed, analyzed, and archived. The file format is the structural container that organizes image data, compression, and metadata. Format choice affects every downstream step including file size, viewing speed, analysis tool compatibility, and long-term archival cost.

Most commercially available slide scanners produce proprietary formats tied to specific vendors. A 2006 evaluation of 31 digital slide systems found that file format, compression techniques, and visualization solutions varied widely across systems, with most systems capable of high-resolution digitization of a whole slide within about one hour using a 40x objective. The image quality of these virtual microscopy systems was considered suitable for clinical, educational, and research purposes. This early survey established that format diversity was already a defining characteristic of the digital pathology landscape.

The practical consequence of format diversity is that laboratories must manage multiple file types, maintain vendor-specific viewing software, and navigate compatibility issues when moving images between systems. A 2026 review of DICOM in digital pathology noted that at least 10 whole slide scanners have been cleared or approved by the FDA in the United States, yet widespread adoption has been hampered by the continued proliferation of proprietary, vendor-specific whole slide image formats that limit interoperability between systems. Formats such as SVS, iSyntax, BIF, MRXS, and NDPI have proven adequate for siloed deployments but are less well-suited for integration into enterprise imaging solutions.

Core Technical Concepts in WSI Formats

Compression Algorithms

Compression reduces file size by removing redundant data. Lossy compression discards some image information to achieve smaller files, while lossless compression preserves all original data. Most WSI formats use lossy compression based on JPEG or JPEG 2000 algorithms.

A 2018 study on optimized JPEG 2000 compression for histopathological WSIs found that currently used WSI formats are based on lossy image compression algorithms, either using JPEG or its more efficient successor JPEG 2000. The study defined an optimized parametrization designated JP2-WSI that applies very high compression to the background portion of the WSI while using conventional compression on tissue-containing regions. This approach produced the smallest file sizes and highest overall compression ratios for all 17 slides tested when compared to fixed 35:1 JPEG 2000 compression and the default formats of Aperio, Hamamatsu, and 3DHISTECH scanners. The average file size by JP2-WSI amounted to 15, 9, and 16 percent respectively of the file sizes of the three commercial scanner vendor proprietary formats (3DHISTECH MRXS, Aperio SVS, and Hamamatsu NDPI). Image quality was equal to or better than the compared formats based on visual inspection and peak signal-to-noise ratio measurements.

This finding demonstrates that compression parameters matter as much as the format itself. Laboratories using the same scanner can achieve substantially different file sizes depending on how compression is configured.

Pyramidal Structure

WSI formats use a pyramidal structure to enable rapid viewing at multiple magnifications. The pyramid contains the full-resolution base layer plus progressively downsampled levels. Viewers request only the tiles needed for the current zoom level, which reduces memory requirements and speeds navigation.

The pyramidal structure is integral to how viewing software interacts with the file. Formats that lack proper pyramid support may require the viewer to load the entire image or to generate pyramid levels on the fly, which slows performance. When evaluating formats, verify that the pyramid levels are complete and correctly built, as corrupted or missing levels can cause viewing failures at specific magnifications.

Metadata and Annotation Storage

WSI formats vary in how they store metadata including patient identifiers, stain type, magnification, scanner settings, and annotations. Some formats embed metadata in standardized fields, while others use proprietary structures that only vendor software can read. The ability to preserve and export metadata affects data management, quality control, and compliance with data sharing policies.

The NIH Genomic Data Sharing Policy at https://sharing.nih.gov/genomic-data-sharing-policy establishes expectations for data sharing in NIH-funded research. While this policy focuses on genomic data, the principles of data documentation and sharing apply to digital pathology datasets used in research. The FAIR Guiding Principles at https://doi.org/10.1038/sdata.2016.18 emphasize that data should be findable, accessible, interoperable, and reusable. Format choices that preserve metadata and enable interoperability support these principles.

At a Glance: WSI Format Comparison

Format-Specific Analysis

SVS (Aperio)

SVS is the proprietary format produced by Aperio scanners, now part of Leica Biosystems. It uses JPEG or JPEG 2000 compression and includes a pyramidal structure. SVS files are widely supported by open-source tools including OpenSlide and QuPath, making them a common choice in research settings.

The 2018 JPEG 2000 optimization study used Aperio SVS as one of the commercial comparison formats. The optimized JP2-WSI format produced files averaging 9 percent of the size of SVS files while maintaining equal or better image quality. This suggests that SVS files, while functional, may not be optimally compressed for storage efficiency.

MRXS (3DHISTECH)

MRXS is the proprietary format from 3DHISTECH scanners. It uses JPEG compression and stores image data across multiple files in a folder structure instead of a single file. This multi-file structure can complicate file management and transfer.

The same 2018 study found that JP2-WSI files averaged 15 percent of the size of MRXS files. The multi-file structure of MRXS can also create challenges when moving or archiving images, as all component files must be kept together.

NDPI (Hamamatsu)

NDPI is the proprietary format from Hamamatsu scanners. It uses JPEG compression and includes a pyramidal structure. NDPI is supported by OpenSlide and QuPath, making it accessible to analysis workflows.

A 2021 study on automatic segmentation of digital pathology slides used 100 WSI files in NDPI format from breast, lung, and thyroid slides. The study demonstrated that NDPI files could be processed with unsupervised segmentation algorithms based on superpixel and fully convolutional neural network approaches. This confirms that NDPI is compatible with computational analysis pipelines.

A 2026 protocol for digitizing natural history collections using the Hamamatsu NanoZoomer S20 scanner provides practical data points for NDPI file sizes. Scanning times ranged from 41 seconds for small invertebrates to 18 minutes for palynological samples, with final compressed file sizes of 0.15 to 28 GB. High-resolution images at 0.23 micrometers per pixel captured diagnostic morphological features across pollen, diatoms, radiolarians, plant and fungi tissues, and invertebrates. This demonstrates that NDPI format can accommodate a wide range of specimen types and file sizes.

TIFF and OME-TIFF

TIFF is a flexible container format that can store uncompressed or compressed image data. OME-TIFF extends TIFF with standardized metadata for microscopy images. The Open Microscopy Environment specification defines how to store image dimensions, channel information, and acquisition metadata within the TIFF structure.

OME-TIFF is particularly valuable for research workflows because it is vendor-neutral and supported by a wide range of tools including ImageJ, napari, and QuPath. A 2026 open-source tool called MCD Stitcher converts proprietary Imaging Mass Cytometry .mcd files into OME-TIFF images with automated whole-slide stitching. The generated OME-TIFF outputs preserve spatial, channel, and acquisition metadata for downstream analysis in tools such as QuPath, napari, and ImageJ/Fiji. This demonstrates that OME-TIFF serves as an effective interoperability format for converting proprietary data into analyzable form.

DICOM

DICOM (Digital Imaging and Communications in Medicine) is the standard that revolutionized radiology and is now being applied to digital pathology. The DICOM standard offers a vendor-neutral whole slide image format that enables seamless exchange of images and metadata across systems. A 2026 review noted that DICOM supports efficient workflows, accurate diagnosis, and secondary use such as artificial intelligence research.

The same review identified that proprietary formats such as SVS, iSyntax, BIF, MRXS, and NDPI are less well-suited for integration into enterprise imaging solutions. DICOM adoption represents a critical step toward an interoperable, vendor-neutral imaging ecosystem as digital pathology matures and computational pathology applications proliferate.

DICOM pathology support includes standardized storage of whole slide images with pyramidal structure and metadata. The standard defines how to encode image tiles, compression, and clinical metadata in a consistent way that any DICOM-compliant system can read.

Emerging Formats

The Iris File Extension (IFE) is a low overhead performance-oriented WSI file format designed to improve image rendering experience and simplify image management. A 2025 paper described the Iris RESTful Server, an HTTP server with a RESTful API compatible with the DICOMweb WADO-RS API. Testing showed that a single Raspberry Pi equivalent system could handle a peak of 5061 requests per second with a median latency of 21 milliseconds on a private network. This demonstrates that format design can significantly affect server performance and viewer responsiveness.

Practical Workflow Considerations

Storage Cost and Capacity Planning

File size directly determines storage cost. The 2018 JPEG 2000 study found that optimized compression could reduce file sizes to 15, 9, and 16 percent of MRXS, SVS, and NDPI file sizes respectively. For a laboratory scanning thousands of slides per month, these differences translate into substantial storage savings.

A 2026 study on whole-slide scanner tissue detection algorithms examined how scan area affects file size. The study scanned 260 routine diagnostic glass slides on 8 whole-slide scanners representing 4 commercial manufacturers. Scanner-specific tissue detection algorithms altered the scanned area, scan time, and file size. Variations in scan area directly influenced scan time and file size, which are major expenses in storage and data management. This means that scanner settings and tissue detection algorithms are as important as format choice in determining storage requirements.

When planning storage capacity, consider the following factors:

• Average file size per slide based on tissue type and scanner settings
• Monthly slide volume and projected growth
• Retention period required by institutional policy or regulations
• Backup and redundancy requirements
• Compression settings and whether they can be optimized

Analysis Tool Compatibility

The choice of analysis tools constrains format selection. Open-source tools including QuPath, ImageJ, and napari support multiple formats through libraries such as OpenSlide. However, not all formats are equally well supported, and some proprietary formats may require vendor-specific plugins or conversion steps.

A 2022 study on weakly supervised learning for endometrial whole slide image diagnosis used iSyntax format files. The study achieved over 85 percent validation accuracy and over 87 percent test accuracy using a weakly supervised learning algorithm that required only slide-level labels during training. This demonstrates that iSyntax files can be used in deep learning pipelines, but the study also noted that fully supervised learning is problematic due to the requirement for costly manual annotation.

The METRICS quality scoring tool for radiomics research, described in a 2024 paper, includes categories for imaging data, image processing and feature extraction, and open science. The tool emphasizes that methodological quality depends on documenting how images were acquired, processed, and analyzed. Format choices that preserve metadata and support reproducible processing contribute to higher quality scores.

Conversion and Interoperability

Converting between formats is sometimes necessary when moving images between systems or tools. Conversion can be lossy if compression settings change, and metadata may be lost if the target format does not support all fields.

A 2024 paper on data format conversion and compatibility assurance described a technology middleware approach for integrating multi-source data. The method achieved 98 percent and 99 percent accuracy in handling JSON and SQL database data respectively. While this study focused on general data formats instead of pathology images, it illustrates the broader challenge of format conversion and the importance of maintaining data integrity during conversion.

The MCD Stitcher tool demonstrates a practical approach to format conversion for Imaging Mass Cytometry data. The tool converts proprietary .mcd files into OME-TIFF with automated whole-slide stitching, supporting rectangular and polygonal regions of interest, accommodating variable pixel sizes, and using memory-aware chunked reading to process large datasets on standard workstations. This provides a reproducible workflow for converting raw data into interoperable image formats without reliance on vendor-specific software.

Quality Control

Quality control is essential in digital pathology workflows. A 2026 study described the deployment of an AI-based automated quality control platform in a large tertiary cancer center. The study evaluated 94,995 WSIs generated over one month across 10 subspecialty services and multiple stain types. Artifact detection demonstrated high reproducibility across scanner platforms, with tissue fold detection showing over 97 percent concordance and air bubble detection exceeding 98 percent concordance. Scanning-related artifacts including missing tissue and blurred regions varied by scanner model but were consistently identified. The mean analysis time was 17 seconds per image with no analysis failures during load testing.

The study found that mean total artifact burden per image was 1.91 percent, with most images meeting predefined quality control thresholds. This suggests that automated quality control can be integrated into high-throughput workflows, but the variation in artifact types across scanner models means that quality control parameters may need to be calibrated for each scanner.

Records and Measurements

File Size Tracking

Maintain records of file sizes by scanner, format, and tissue type. This data supports storage planning and identifies anomalies that may indicate scanning problems. Track the following metrics:

• Average file size per slide by tissue type
• File size range and outliers
• Compression ratio achieved
• Scan time per slide
• Scan area as a percentage of total slide area

A 2026 study on tissue detection algorithms found that scan area directly influences scan time and file size. Tracking these metrics together helps identify when scanner settings drift or when tissue detection algorithms perform poorly.

Compression Performance

Document compression settings and measure their effect on file size and image quality. The 2018 JPEG 2000 study demonstrated that optimized compression can reduce file sizes dramatically while maintaining diagnostic quality. Record the following for each scanning configuration:

• Compression algorithm and quality setting
• File size before and after compression
• Visual quality assessment or quantitative metrics such as peak signal-to-noise ratio
• Any clinical or research use cases where image quality was insufficient

Metadata Completeness

Verify that required metadata is captured in each WSI file. Missing metadata can cause problems in downstream analysis, data sharing, and regulatory compliance. Check the following fields:

• Patient or specimen identifier
• Stain type and protocol
• Scanner model and settings
• Magnification and pixel size
• Date and time of scanning
• Any annotations or regions of interest

Common Failure Patterns

Corrupted Pyramid Levels

Pyramid levels can become corrupted during scanning or file transfer. Symptoms include blank areas at specific zoom levels, slow loading, or viewer crashes. Verify pyramid integrity by opening files in multiple viewers and checking all zoom levels.

Metadata Loss During Conversion

Converting between formats can result in lost metadata, particularly for proprietary fields that have no equivalent in the target format. Document what metadata is preserved during conversion and verify that critical fields survive the process.

Incompatible Compression Settings

Some analysis tools may not support certain compression types or quality settings. For example, a tool that expects JPEG compression may fail on JPEG 2000 files. Test analysis pipelines with representative files before committing to a format.

Multi-File Format Management

Formats such as MRXS store image data across multiple files. If these files are separated during transfer or archival, the image becomes unreadable. Implement procedures to keep multi-file formats intact and verify file integrity after transfer.

Scanner-Specific Artifacts

Different scanner models produce different artifact patterns. The 2026 quality control study found that scanning-related artifacts including missing tissue and blurred regions varied by scanner model. Document known artifact patterns for each scanner and incorporate this knowledge into quality control procedures.

Limitations and Constraints

Proprietary Format Lock-In

Proprietary formats can create vendor lock-in, making it difficult to switch scanners or software without converting existing archives. The 2026 DICOM review noted that proprietary formats limit interoperability between systems. Consider the long-term implications of format choice when selecting scanners.

Compression and Diagnostic Accuracy

Lossy compression can affect diagnostic accuracy in some contexts. A 2010 study compared digital image formats to film for diabetic macular edema evaluation. The study found that agreement between evaluations of monoscopic digital formats and film was similar to that between stereo digital formats and film, and the performance of uncompressed images versus film was similar to that of compressed images versus film. However, the readers detected clinically significant macular edema and retinal thickening at the center of the macula in film images more often than in digital formats. This suggests that digital formats, including compressed formats, may be less sensitive for detecting certain subtle findings.

File Size Variability

File sizes vary substantially based on tissue type, stain, and scanning parameters. The natural history collection protocol reported file sizes ranging from 0.15 to 28 GB for different specimen types. This variability complicates storage planning and requires capacity estimates based on actual data instead of averages.

Analysis Tool Limitations

Not all analysis tools support all formats. Some tools may require conversion before analysis, which adds processing time and may introduce artifacts. Test analysis workflows with the formats you plan to use before committing to a format.

Safety and Regulatory Context

Data Sharing and Privacy

Digital pathology images contain patient health information and are subject to privacy regulations. The NIH Genomic Data Sharing Policy at https://sharing.nih.gov/genomic-data-sharing-policy establishes expectations for data sharing in NIH-funded research. While this policy focuses on genomic data, the principles of responsible data sharing apply to pathology images used in research.

The FAIR Guiding Principles at https://doi.org/10.1038/sdata.2016.18 emphasize that data should be findable, accessible, interoperable, and reusable. Format choices that support these principles facilitate data sharing and reuse.

Regulatory Approval of Scanners

The 2026 DICOM review noted that at least 10 whole slide scanners have been cleared or approved by the FDA in the United States. This regulatory oversight provides some assurance of scanner performance, but format choices remain the responsibility of the laboratory.

Professional Escalation Criteria

Escalate to senior staff or information technology support when:

• Files cannot be opened in any available viewer
• Metadata is consistently missing or incorrect
• File sizes deviate substantially from expected ranges
• Analysis tools fail on specific formats
• Quality control identifies artifacts that may affect diagnosis

Frequently Asked Questions

What is the difference between lossy and lossless compression in WSI formats?

Lossy compression reduces file size by discarding some image information, while lossless compression preserves all original data. Most WSI formats use lossy compression based on JPEG or JPEG 2000 algorithms. A 2018 study found that optimized JPEG 2000 compression could reduce file sizes to 15, 9, and 16 percent of MRXS, SVS, and NDPI file sizes respectively while maintaining equal or better image quality. Lossless compression produces larger files but preserves all image data, which may be required for certain research or regulatory purposes.

Which WSI format is best for research workflows?

OME-TIFF is often the best choice for research workflows because it is vendor-neutral and supported by a wide range of tools including ImageJ, napari, and QuPath. The MCD Stitcher tool demonstrates that OME-TIFF can serve as an effective interoperability format for converting proprietary data into analyzable form. However, the best format depends on the specific analysis tools and workflows being used. Test your analysis pipeline with representative files before committing to a format.

How does DICOM compare to proprietary WSI formats?

DICOM offers a vendor-neutral whole slide image format that enables seamless exchange of images and metadata across systems. A 2026 review noted that proprietary formats such as SVS, iSyntax, BIF, MRXS, and NDPI are less well-suited for integration into enterprise imaging solutions. DICOM adoption represents a critical step toward an interoperable, vendor-neutral imaging ecosystem. However, DICOM support in analysis tools may be less mature than support for established proprietary formats.

Can I convert between WSI formats without losing image quality?

Conversion can be lossy if compression settings change, and metadata may be lost if the target format does not support all fields. The MCD Stitcher tool demonstrates a reproducible workflow for converting proprietary Imaging Mass Cytometry data into OME-TIFF while preserving spatial, channel, and acquisition metadata. Document what metadata is preserved during conversion and verify that critical fields survive the process.

How does compression affect diagnostic accuracy?

Lossy compression can affect diagnostic accuracy in some contexts. A 2010 study found that digital photography was not as sensitive as film in detecting clinically significant macular edema and retinal thickening at the center of the macula. However, agreement between evaluations of monoscopic digital formats and film was similar to that between stereo digital formats and film. The impact of compression depends on the specific diagnostic task and the compression level used.

What file sizes should I expect for whole slide images?

File sizes vary substantially based on tissue type, stain, and scanning parameters. A 2026 protocol for digitizing natural history collections reported final compressed file sizes of 0.15 to 28 GB for different specimen types. A 2018 study found that optimized JPEG 2000 compression could reduce file sizes dramatically compared to standard formats. Track file sizes for your specific scanner and tissue types to establish expected ranges.

How do scanner settings affect file size?

Scanner-specific tissue detection algorithms alter the scanned area, scan time, and file size. A 2026 study found that variations in scan area directly influenced scan time and file size, which are major expenses in storage and data management. Optimizing scanner settings to minimize scan area while capturing all tissue can reduce file sizes and storage costs.

What should I do if analysis tools cannot read my WSI files?

First, verify that the file is not corrupted by opening it in the vendor viewer. If the file opens in the vendor viewer but not in analysis tools, check whether the analysis tool supports the format. Some tools may require conversion to a supported format. The MCD Stitcher tool provides an example of converting proprietary formats to OME-TIFF for analysis in standard tools. If conversion is not possible, consider whether the analysis tool has plugins or extensions that add format support.

Related Bioinformatics Guides

• How To Use Alphafold To Predict Structure: Structural Analysis and Computational Methodologies in Bioinformatics
• AI Protein Binder Design Tools: RFdiffusion, ProteinMPNN, BindCraft-Style Filtering, and Target-Specific Discovery Workflows
• Pan-Cancer Analysis of Whole Genomes (PCAWG)
• FASTA File Format: Structure, Specifications, and Parser Implementations
• FASTQ File Format: Decoding Phred Quality Scores and Quality Control Workflows

References and Further Reading

• EMBL-EBI Training. European Bioinformatics Institute.
• NCBI Data Resources. National Center for Biotechnology Information.
• Genomic Data Sharing Policy. National Institutes of Health.
• The FAIR Guiding Principles. Scientific Data.
• Comparison of multiple stereoscopic and monoscopic digital image formats to film for diabetic macular edema evaluation.. Investigative ophthalmology & visual science, 2010.
• Critical comparison of 31 commercially available digital slide systems in pathology.. International journal of surgical pathology, 2006.
• The Lucerne Toolbox 3: digital health and artificial intelligence to optimise the patient journey in early breast cancer-a multidisciplinary consensus.. The Lancet. Oncology, 2025.
• METhodological RadiomICs Score (METRICS): a quality scoring tool for radiomics research endorsed by EuSoMII.. Insights into imaging, 2024.
• Weakly supervised learning and interpretability for endometrial whole slide image diagnosis.. Experimental biology and medicine (Maywood, N.J.), 2022.
• Optimized JPEG 2000 Compression for Efficient Storage of Histopathological Whole-Slide Images.. Journal of pathology informatics, 2018.
• [Automatic Segmentation of Digital Pathology Slides Based on Unsupervised Learning].. Sichuan da xue xue bao. Yi xue ban = Journal of Sichuan University. Medical science edition, 2021.
• Comparative analysis of tissue reconstruction algorithms for 3D histology.. Bioinformatics (Oxford, England), 2018.
• The evolving role of DICOM in digital pathology.. 2026.
• MCD Stitcher: An open-source tool for whole-slide stitching and conversion of Imaging Mass Cytometry data. 2026.
• Comparative analysis of whole-slide scanner tissue detection algorithms: Implications for scan area, scan time, and file size in high-volume digital pathology workflows.. 2026.
• Deployment of AI-driven automated quality control of whole-slide images in a large tertiary cancer center.. 2026.
• Digitizing microscope slide-based natural history collections: A protocol using slide scanner technology.. 2026.
• MHub.ai: A Standardized Platform for Reproducible AI Research in Medical Imaging. 2026.
• Real-World Insights in Designing SteatoStat: An End-to-End Deep Learning Pipeline for Hepatic Steatosis Quantification. 2026.
• Protocol for AI-powered classification of IDH-mutant astrocytoma using GUIDE.. 2026.
• Iris RESTful Server and IrisTileSource: An Iris implementation for existing OpenSeaDragon viewers. Journal of Pathology Informatics, 2025.
• Algorithm for Data Format Conversion and Compatibility Guarantee in Technical Platforms for Cross Platform Application Integration. 2024 IEEE 4th International Conference on Data Science and Computer Application (ICDSCA), 2024.
• JPEG Image Encryption With Improved Format Compatibility and File Size Preservation. IEEE transactions on multimedia, 2018.
• Exploring the Impact of the Output Format on the Evaluation of Large Language Models for Code Translation. 2024 IEEE/ACM First International Conference on AI Foundation Models and Software Engineering (Forge) Conference Acronym:, 2024.
• Determining format compatibility across a data processing device and another data processing device prior to transfer of a multimedia file therebetween. 2014.
• A compatibility analysis framework for freshwater quality standards of pesticides across multiple countries: Human health, regulatory compliance, and water treatment perspectives.. Science of the Total Environment, 2025.
• CrystalShift: A Versatile Command-Line Tool for Crystallographic Structural Data Analysis, Modification, and Format Conversion Prior to Solid-State DFT Calculations of Organic Crystals. De Computis, 2025.
• Aird-MSI: a high compression rate and decompression speed format for mass spectrometry imaging data. bioRxiv, 2025.
• Overview of modern digital diagnostic image markup tools. Kazan Medical Journal, 2023.
• Mammographic density assessed on paired raw and processed digital images and on paired screen-film and digital images across three mammography systems. Breast Cancer Research, 2016.
• Scalable HEVC for histological whole-slide image compression. Informatik Aktuell, 2020.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.