Informatics, digital & computational pathology
Digital and whole slide imaging
Quality control
Authors: Akash Parvatikar, Ph.D., Julie Feldstein, M.D.
Editorial Board Members: Jerome Cheng, M.D., Lewis A. Hassell, M.D.
Last staff update: 3 December 2024 (update in progress)
Copyright: 2024, PathologyOutlines.com, Inc.
PubMed Search: Quality control for digital pathology
See Also: Digital imaging fundamentals & standards
Table of Contents
Definition / general | Essential features | Terminology | Applications | Implementation | Advantages | Limitations | Diagrams / tables | Microscopic (histologic) images | Additional references | Board review style question #1 | Board review style answer #1 | Board review style question #2 | Board review style answer #2Cite this page: Parvatikar A, Feldstein J. Quality control. PathologyOutlines.com website. https://www.pathologyoutlines.com/topic/informaticsqualitycontrolfordigpath.html. Accessed December 4th, 2024.
Definition / general
- Curation of pathologist annotated, high quality slide datasets is a serious bottleneck in digital pathology; these data fuel downstream analysis and the development of reliable artificial intelligence (AI) based solutions for various applications
- By streamlining the production of high quality histopathology images via fully automating tissue processing and digitization workflows, it expedites data interpretation and usage, thereby accelerating preclinical and clinical research
- Reference: Arch Pathol Lab Med 2019;143:222
Essential features
- Automated quality control (QC) systems: utilize AI models to identify common slide issues, such as blurriness, folds, tears and air bubbles, ensuring high quality and consistent digital slides
- Improved workflow efficiency: automated quality checks significantly reduce manual labor, speeding up the diagnostic process and allowing pathologists to focus on more complex tasks
- Enhanced diagnostic accuracy: high quality slides lead to more accurate diagnoses, minimizing errors and improving patient outcomes
- Integration challenges: addressing the complexities of integrating AI tools with existing laboratory information systems and workflows
- Regulatory and compliance: ensuring that AI based QC processes meet regulatory standards as well as maintain data security and patient privacy
- Reference: Diagnostics (Basel) 2024;14:990
Terminology
- Computational pathology: the application of advanced computational techniques, including machine learning and data analysis, to interpret pathology data and improve diagnostic accuracy, research and patient care
- Whole slide imaging (WSI): a digital imaging technology that captures high resolution scans of entire histopathological slides, enabling virtual microscopy and facilitating remote analysis and AI integration
- Quality control (QC): a systematic process in digital pathology to ensure the accuracy, consistency and reliability of slide preparation, scanning and digitization for improved diagnostic outcomes
- Artifact detection: the identification and classification of anomalies or imperfections (e.g., folds, tears, blurriness) in digital slide images that may compromise diagnostic quality or accuracy
- Automated QC systems: AI powered tools that detect and flag quality issues in histopathology slides, such as blurriness or artifacts, streamlining workflows and ensuring consistent high quality outputs
Applications
- Enhanced slide quality and consistency: AI models detect blurriness and physical artifacts like folds, tears and air bubbles, ensuring clear and precise histopathology images
- Increased efficiency and time savings: automated QC significantly reduces the time required for manual checks, allowing pathologists to focus on diagnostic tasks
- Improved diagnostic accuracy: ensuring high quality slides leads to more accurate diagnoses and reliable data for other AI diagnostic tools
- Streamlined workflow integration: AI based QC tools integrated into digital pathology platforms enable real time quality checks and continuous slide quality improvement
- Educational benefits: high quality digital slides with accurate annotations provide valuable resources for training and continuing education for pathologists and researchers
- Regulatory and compliance: automated QC processes help meet regulatory standards and maintain reliable records for audits and quality assurance
- Reference: Histopathology 2024;85:207
Implementation
- Assess current workflow
- Evaluate existing slide preparation, digitization and review processes
- Identify common quality issues, such as blurriness, folds and air bubbles
- Integrate automated QC tools
- Embed AI based QC tools into the digital pathology workflow
- Implement real time QC checks during slide scanning and digitization
- Validate and benchmark
- Compare AI model performance against manual QC by pathologists
- Use metrics such as AUC (area under the curve), accuracy and error rates for evaluation
- Deploy and monitor
- Roll out the automated QC system in the pathology lab
- Continuously monitor performance and gather feedback from pathologists
- Provide training and support
- Train lab personnel to use the new QC tools
- Offer ongoing support and updates to ensure the system remains effective
- Document and standardize
- Document the QC process and create standard operating procedures (SOPs)
- Ensure compliance with regulatory standards and maintain records for audits
- Review and improve
- Periodically review the QC system's effectiveness
- Incorporate new data and improve AI models to enhance accuracy and reliability
- Reference: Diagnostics (Basel) 2021;11:2167
Advantages
- Improved workflow efficiency
- Automated QC reduces the time spent on manual slide reviews, allowing pathologists to focus on diagnostic tasks
- Enhanced diagnostic accuracy
- Consistent high quality slides lead to more accurate diagnoses and reduce the risk of errors caused by poor image quality
- Increased lab productivity
- Faster QC processes streamline the overall workflow, increasing the number of cases that can be processed
- Cost savings
- Reducing the need for manual QC labor and reprocessing of poor quality slides leads to significant cost savings
- Standardized quality assurance
- Automated QC ensures consistent application of quality standards across all slides, enhancing reliability
- Better resource allocation
- Pathologists can allocate more time to complex cases and research, improving the quality of patient care
- Enhanced research capabilities
- Access to high quality, annotated datasets supports advanced research and development of new diagnostic tools
- Regulatory compliance
- Documented and automated QC processes help meet regulatory standards and facilitate audits
- Training and education
- High quality slides provide valuable resources for training new pathologists and for ongoing education
- Improved collaboration
- Standardized and high quality digital slides enhance collaboration between pathologists, researchers and other healthcare professionals
- Reference: Histopathology 2024;85:207
Limitations
- High initial costs
- Implementing AI technology requires significant investment in hardware, software and training
- Data privacy concerns
- Ensuring the confidentiality and security of patient data during digitization and storage is crucial
- Integration with existing systems
- Integrating AI tools with current lab information systems and workflows can be complex and time consuming
- Technical expertise
- Pathology labs need skilled personnel to manage and operate AI systems, which may not always be available
- Algorithm bias and reliability
- AI models may exhibit biases based on training data, leading to inconsistent results across diverse patient populations
- Regulatory and compliance issues
- Navigating regulatory approvals and ensuring compliance with healthcare standards can be challenging
- Resistance to change
- Pathologists and lab personnel may be resistant to adopting new technologies due to unfamiliarity or skepticism
- Quality and quantity of training data
- High quality annotated datasets are essential for training accurate AI models but obtaining such data can be difficult
- Maintenance and updates
- Continuous maintenance and periodic updates of AI systems are required to ensure optimal performance and address emerging quality issues
- Validation and benchmarking
- Ongoing validation and benchmarking of AI tools against manual QC processes are necessary to establish trust and reliability
- Reference: J Histotechnol 2024;47:39
Diagrams / tables
Contributed by Akash Parvatikar, Ph.D.
Schematic overview of QC workflow
Microscopic (histologic) images
Contributed by Akash Parvatikar, Ph.D.
No QC issue
Tissue blur
Tissue coverslip
Tissue dirt
Tissue folds
Tissue microvibrations
Tissue cracking
Tissue separation
Additional references
Board review style question #1
Which of the following is a digital scanning quality control (QC) issue that commonly occurs on whole slide imaging (WSI)?
- Blurriness
- Reagent lot number
- Scanner serial number
- Tissue depth
- Tissue size
Board review style answer #1
A. Blurriness is a common quality control issue on whole slide imaging that affects the clarity and detail resolution necessary for accurate pathology analysis. Answer E is incorrect because tissue size is not directly a QC issue but rather a specimen preparation concern. Answer D is incorrect because tissue depth pertains to the sectioning technique and is not evaluated by digital scanning QC. Answer C is incorrect because the scanner serial number is related to equipment tracking, not image quality. Answer B is incorrect because reagent lot number influences reagent quality but is not a digital scanning QC issue.
Comment Here
Reference: Quality control
Comment Here
Reference: Quality control
Board review style question #2
What quality control (QC) issue is demonstrated in the example shown above?
- Antibody issue
- Blurriness
- Coverslip issue
- Tissue fold
- Tissue separation
Board review style answer #2
B. Blurriness. The image demonstrates a lack of sharpness and clarity, indicating a blurriness issue, which impacts the ability to accurately analyze tissue details in digital pathology. Answer E is incorrect because tissue separation involves the detachment of tissue components, which would be seen as gaps or breaks within the tissue and are not present in the image. Answer D is incorrect because a tissue fold would appear as overlapped or pinched tissue, which alters the tissue architecture and is also not observed here. Answer A is incorrect because an antibody issue would affect the staining quality and specificity, leading to issues in immunohistochemistry results not image clarity. Answer C is incorrect because coverslip issues typically manifest as air bubbles, dust or scratches visible on the slide, which affect the optical quality but do not cause blurriness of the entire image.
Comment Here
Reference: Quality control
Comment Here
Reference: Quality control
