This session explores how digital pathology and computational pathology (AI) are transforming anatomic pathology. Dr. Marilyn Bui (Moffitt Cancer Center; Chair, CAP Digital & Computational Pathology Committee) outlines the “third revolution” of pathology, moving beyond manual “eyeballing” to augmented intelligence that improves detection, classification, quantification, prognosis, and prediction.
We cover practical workflows, measurable benefits (error reduction, shorter TAT, less unnecessary IHC), adoption models, change management, and what it takes to build an AI-ready lab in 2025 and beyond. NovoPath closes with a complimentary Workflow Assessment for labs seeking a readiness evaluation.
Key Takeaways
- AI augments, not replaces: Think augmented intelligence: pathologist + AI > either alone.
- Digital is the foundation: Whole slide imaging enables scalable consults and powers computational pathology.
- 5 core AI applications: Detection, Classification, Quantification (e.g., HER2, PD-L1, Ki-67), Prognosis, Prediction.
- Measured impact: Lower error rates (combining AI + pathologist), reduced unnecessary IHC usage, faster turnaround time (TAT), standardized quality across locations.
- Start small: You don’t need a fully digital lab to get value—begin with targeted QA/QC tools (e.g., IHC quality monitoring) and scale.
- Leadership & adoption: Education + clear “why” + early wins overcome resistance; late adopters often become champions after hands-on use.
- Roadmap: Digitize slides, pilot high-value use cases, integrate into LIS workflows, monitor outcomes, expand across biomarkers and prediction models.
- NovoPath offer: Complimentary Workflow Assessment, on-site review of current processes, gaps, and ROI opportunities for AI-readiness.
Who Should Watch
Lab directors, pathology leadership, QA managers, enterprise lab networks, operations leaders, IT/LIS owners, and teams evaluating digital pathology, AI apps, and LIS modernization.
Problems Addressed
- Manual/fragmented workflows that cap throughput and consistency
- Rising case volumes + pathologist shortages
- Inefficient IHC utilization and elongated TAT
- Unclear path to AI-ready operations and ROI
Solutions & Frameworks
- “Third Revolution” of Pathology From microscope → IHC → molecular → digital + AI (computational pathology).
- Three Pillars for AI Value
- Precision Pathology: diagnosis & biomarker interpretation
- Workflow & Quality Improvement: efficiency, QC/QA, fewer handoffs
- Impactful Reporting: moving toward multimodal, interactive reports
Five AI Application Zones
- Detection (e.g., sentinel lymph node metastasis)
- Classification (e.g., lung/bladder/sarcoma subtyping)
- Quantification (HER2, PD-L1, Ki-67; multiplex IF/IHC)
- Prognosis (disease course, risk)
- Prediction (therapy benefit; FDA-cleared multimodal example in prostate)
Use Cases Mentioned
- Metastasis detection in lymph nodes: combined AI + specialist pathologist reduces error and increases speed.
- Reduce unnecessary IHC: AI triage can cut waste and improve TAT.
- Prostate cancer (multimodal): HE + biomarkers + outcomes to predict therapy benefit and risk.
- QC without full digital: Single-slide scanners with cloud QC can tighten IHC quality immediately.
Implementation Roadmap for an AI-Ready Lab
- Baseline: Map current workflow, pain points, and KPIs (TAT, IHC rates, error trends).
- Digitize Core: Prioritize scanners + WSI coverage for high-value specimen types.
- Select Use Cases: Start with quantification (clear ROI, measurable outcomes).
- Integrate: Connect viewers, LIS, and AI apps; ensure human-in-the-loop sign-out.
- Measure: Track TAT, IHC utilization, concordance, re-cut rates, and cost/time saved.
- Scale: Expand to detection, classification, prognosis/prediction; standardize across sites.
Common Objections & Responses
- “AI slows us down.” Initial setup requires change; once stabilized, users report faster sign-out and higher confidence.
- “We’re not fully digital yet.” You can start with targeted QC tools (IHC quality monitoring) and incremental scanning.
- “What about reimbursement for stains auto-ordered by AI?” Use human-in-the-loop policies or generate local validation data to support payer discussions.
- “Pathologists resist change.” Show the “why,” deliver training, pilot with champions, and highlight flexibility (e.g., remote sign-out).
FAQ
AI-driven analysis of whole slide images + clinical/omic data to support diagnosis, prognosis, and prediction.
By reducing error rates, standardizing biomarker scoring, lowering unnecessary IHC, and improving TAT.
Yes. Begin with limited scanning and QA/QC tools, then expand to full digital and AI apps.
No. Augmented intelligence improves accuracy and efficiency while keeping experts in the loop.
Biomarker quantification (HER2, PD-L1, Ki-67), metastasis detection triage, and IHC quality monitoring.
TAT, IHC utilization, error/re-cut rates, concordance, productivity (cases per FTE), and cost/time saved.
Route via human-in-the-loop ordering or run a validation study demonstrating reliability and avoided waste.
Here's a Quick Breakdown, AI in Action in Your Lab
- Definition: Augmented intelligence in pathology pairs human expertise with AI to improve detection, biomarker quantification, prognosis, and therapy prediction.
- Value: AI + digital pathology reduce errors, shorten TAT, and standardize quality across sites.
- Start Here: Pilot quantification (HER2/PD-L1/Ki-67) and IHC QC; measure outcomes; scale.
- Readiness: An AI-ready lab has digitized slides, integrated viewers/LIS, human-in-the-loop checks, and tracked KPIs.
- Change: Education + early wins convert skeptics; many late adopters become champions after real use.
- Next Step: Book a NovoPath Workflow Assessment to map gaps and ROI.
Speaker & Host
- Speaker: Dr. Marilyn Bui — Senior Member & Professor, Moffitt Cancer Center; Chair, CAP Digital & Computational Pathology Committee; 200+ peer-reviewed papers; books & patents.
- Host: NovoPath Expert Speaker Series — Practical perspectives on AI-ready pathology operations.
Resources
- CAP AI Studio (directory of pathology AI apps)
- HER2 documentary/training resource (trailer shown)
- Request the Complimentary Workflow Assessment
Compliance / Disclaimer
Views expressed reflect the speaker’s professional opinions and do not represent official positions of Moffitt Cancer Center or CAP.
This session explores how digital pathology and computational pathology (AI) are transforming anatomic pathology. Dr. Marilyn Bui (Moffitt Cancer Center; Chair, CAP Digital & Computational Pathology Committee) outlines the “third revolution” of pathology, moving beyond manual “eyeballing” to augmented intelligence that improves detection, classification, quantification, prognosis, and prediction.
We cover practical workflows, measurable benefits (error reduction, shorter TAT, less unnecessary IHC), adoption models, change management, and what it takes to build an AI-ready lab in 2025 and beyond. NovoPath closes with a complimentary Workflow Assessment for labs seeking a readiness evaluation.
Key Takeaways
AI augments, not replaces. Think augmented intelligence: pathologist + AI > either alone.
Digital is the foundation. Whole slide imaging enables scalable consults and powers computational pathology.
5 core AI applications: Detection, Classification, Quantification (e.g., HER2, PD-L1, Ki-67), Prognosis, Prediction.
Measured impact: Lower error rates (combining AI + pathologist), reduced unnecessary IHC usage, faster turnaround time (TAT), standardized quality across locations.
Start small: You don’t need a fully digital lab to get value—begin with targeted QA/QC tools (e.g., IHC quality monitoring) and scale.
Leadership & adoption: Education + clear “why” + early wins overcome resistance; late adopters often become champions after hands-on use.
Roadmap: Digitize slides, pilot high-value use cases, integrate into LIS workflows, monitor outcomes, expand across biomarkers and prediction models.
NovoPath offer: Complimentary Workflow Assessment, on-site review of current processes, gaps, and ROI opportunities for AI-readiness.
Who Should Watch
Lab directors, pathology leadership, QA managers, enterprise lab networks, operations leaders, IT/LIS owners, and teams evaluating digital pathology, AI apps, and LIS modernization.
Problems Addressed
Manual/fragmented workflows that cap throughput and consistency
Rising case volumes + pathologist shortages
Inefficient IHC utilization and elongated TAT
Unclear path to AI-ready operations and ROI
Solutions & Frameworks
1) “Third Revolution” of Pathology
From microscope → IHC → molecular → digital + AI (computational pathology).
2) Three Pillars for AI Value
Precision Pathology: diagnosis & biomarker interpretation
Workflow & Quality Improvement: efficiency, QC/QA, fewer handoffs
Impactful Reporting: moving toward multimodal, interactive reports
3) Five AI Application Zones
Detection (e.g., sentinel lymph node metastasis)
Classification (e.g., lung/bladder/sarcoma subtyping)
Quantification (HER2, PD-L1, Ki-67; multiplex IF/IHC)
Prognosis (disease course, risk)
Prediction (therapy benefit; FDA-cleared multimodal example in prostate)
Use Cases Mentioned
Metastasis detection in lymph nodes: combined AI + specialist pathologist reduces error and increases speed.
Reduce unnecessary IHC: AI triage can cut waste and improve TAT.
Prostate cancer (multimodal): HE + biomarkers + outcomes to predict therapy benefit and risk.
QC without full digital: Single-slide scanners with cloud QC can tighten IHC quality immediately.
Implementation Roadmap (AI-Ready Lab)
Baseline: Map current workflow, pain points, and KPIs (TAT, IHC rates, error trends).
Digitize Core: Prioritize scanners + WSI coverage for high-value specimen types.
Select Use Cases: Start with quantification (clear ROI, measurable outcomes).
Integrate: Connect viewers, LIS, and AI apps; ensure human-in-the-loop sign-out.
Measure: Track TAT, IHC utilization, concordance, re-cut rates, and cost/time saved.
Scale: Expand to detection, classification, prognosis/prediction; standardize across sites.
Common Objections & Responses
“AI slows us down.” Initial setup requires change; once stabilized, users report faster sign-out and higher confidence.
“We’re not fully digital yet.” You can start with targeted QC tools (IHC quality monitoring) and incremental scanning.
“What about reimbursement for stains auto-ordered by AI?” Use human-in-the-loop policies or generate local validation data to support payer discussions.
“Pathologists resist change.” Show the “why,” deliver training, pilot with champions, and highlight flexibility (e.g., remote sign-out).
FAQ
Q1: What is computational pathology?
AI-driven analysis of whole slide images + clinical/omic data to support diagnosis, prognosis, and prediction.
Q2: How does AI improve quality?
By reducing error rates, standardizing biomarker scoring, lowering unnecessary IHC, and improving TAT.
Q3: Can I start without a fully digital lab?
Yes. Begin with limited scanning and QA/QC tools, then expand to full digital and AI apps.
Q4: Does AI replace pathologists?
No. Augmented intelligence improves accuracy and efficiency while keeping experts in the loop.
Q5: What are the fastest-ROI use cases?
Biomarker quantification (HER2, PD-L1, Ki-67), metastasis detection triage, and IHC quality monitoring.
Q6: What KPIs should we track?
TAT, IHC utilization, error/re-cut rates, concordance, productivity (cases per FTE), and cost/time saved.
Q7: How do we manage payer concerns on special stains?
Route via human-in-the-loop ordering or run a validation study demonstrating reliability and avoided waste.
Short Answer Snippet Pack
Definition: Augmented intelligence in pathology pairs human expertise with AI to improve detection, biomarker quantification, prognosis, and therapy prediction.
Value: AI + digital pathology reduce errors, shorten TAT, and standardize quality across sites.
Start Here: Pilot quantification (HER2/PD-L1/Ki-67) and IHC QC; measure outcomes; scale.
Readiness: An AI-ready lab has digitized slides, integrated viewers/LIS, human-in-the-loop checks, and tracked KPIs.
Change: Education + early wins convert skeptics; many late adopters become champions after real use.
Next Step: Book a NovoPath Workflow Assessment to map gaps and ROI.
Speaker & Host
Speaker: Dr. Marilyn Bui — Senior Member & Professor, Moffitt Cancer Center; Chair, CAP Digital & Computational Pathology Committee; 200+ peer-reviewed papers; books & patents.
Host: NovoPath Expert Speaker Series — Practical perspectives on AI-ready pathology operations.
Resources
CAP AI Studio (directory of pathology AI apps)
HER2 documentary/training resource (trailer shown)
Request the Complimentary Workflow Assessment
Compliance / Disclaimer
Views expressed reflect the speaker’s professional opinions and do not represent official positions of Moffitt Cancer Center or CAP.
