Insights and resources for AI in pathology

5 reasons to use AI in clinical diagnostics

Written by Aiforia | Jul 4, 2024 6:08:29 AM

Rising cancer rates increase sample numbers, creating a bigger burden on pathologists, an already underrepresented group of healthcare professionals. The challenge is exacerbated by the fact that traditional clinical tools and processes available to pathologists are manual and subjective.

The digitization of pathology has helped address some of these challenges by enhancing clinical lab workflows and enabling more efficient collaboration. However, to fully reap the benefits of digital pathology in clinical diagnostics, artificial intelligence (AI) must be considered. After all, effective treatment begins with a fast and precise diagnosis – and clinical AI can help to achieve this. 

Adding AI for image analysis to the clinical workflow is vital to:

 

1. Boost productivity 

AI is substantially faster at image analysis and can automate manual, time-consuming tasks. Speeding up the case review increases pathology labs’ output, allowing more new patients to be admitted. Additionally, with the time saved, pathologists can focus more on complex and rare cases.

A study on intraoperative brain tumor diagnosis found that an expert pathologist’s diagnosis during surgery, normally about a 40-minute process, can take under 3 minutes with the assistance of an AI model in the operating room.1 

In an interview with Dr. Kevin Sandeman, Clinical Pathologist and Head of department at Region Skåne pathology lab in Sweden, he explained: “In terms of time savings from case prioritization, for example, with prostate cancer diagnosis, when I work with prostate specimens, and I receive 16 slides and somewhere within these slides is a tumor that can take up less than 2% of space on all the slides combined. It takes vigilance to find it. Therefore, if the AI system can take that 2% area and present it to me to review first, the diagnosis is much faster. This benefits my productivity.”

"We were also positively surprised with the speed of execution of the Aiforia® Platform, which was able to process regions of interest for all cores in just 2 minutes, saving us a lot of time and effort.” – Rania Gaspo, Global Therapy Area Lead, Cerba Research

Read the full case study → 

 

2. Increase diagnostic performance

Pathologists are highly specialized healthcare professionals. However, the tools at their disposal for case review and clinical diagnostics are fallible and time-consuming. Artificial intelligence systems improve the accuracy of analysis, reduce bias, and standardize sample review. 

Several studies show how algorithms based on computational neural networks (CNN) demonstrate expert-level performance in pathology tasks prone to inter-observer variability.2 “This work shows that recent AI-based image analysis tools, such as the Aiforia® Platform, provide valuable assistance in the field of image analysis and allowed us to drastically reduce inter-pathologist variability in the Ki-67 scoring of solid tumors,” Rania Gaspo, Global Therapy Area Lead from Cerba Research, commented on the results from a case study where the AI-assisted results of Ki-67 scoring were compared against three independent pathologists on various solid tumors.

Jenni Säilä, a researcher at the HUSLAB Department of Pathology, working together with her supervisor Tiina Vesterinen, developed a deep learning AI model to quantify Ki-67 positive and negative tumor cells in rare pulmonary carcinoid (PC) tumors. The AI-assisted results were referenced against a pathologist completing the analysis manually and compared with those produced by non-AI-based image analysis software. "The AI modeI was better at analyzing the samples. You didn’t have to exclude any areas because Aiforia was able to detect everything. Compared to other image analysis methods in which if there are any issues with the samples like staining errors or broken tissue, the other programs are not able to count in these areas,” Jenni describes. Read more about the research here

The most beneficial use of AI is the combination of both a pathologist’s knowledge and AI’s accuracy and efficiency. AI is set not to replace pathologists but to supercharge them.

3. Reduce costs

As described above, AI assistance improves diagnostic accuracy, eliminating bias and subjectivity. AI systems analyze cases with 100% consistency. Reducing diagnostic error and misdiagnosis while improving treatment accuracy with more detailed results will result in direct cost savings due to greater precision.

Not only does misdiagnosis cost lives, but it is also a significant financial burden on both patients and hospitals. A 25-year study on malpractice3 in the US found that the average cost per claim in cases involving diagnostic error was $386,849. 

Improved productivity also reduces costs, as AI solutions enable pathologists to review more cases in less time.4

 

4. Enhance staff satisfaction

Better workload distribution can be achieved as pathologists spend less time on manual, repetitive tasks and more on assessing rare or complex cases requiring higher expertise and skills. Faster review times also alleviate the overall burden of rising caseloads.

Implementing AI into the clinical workflow can be an exciting learning experience for medical professionals. Furthermore, the digitalization of clinical workflows and the increased number of samples collected in web-based libraries will represent an incredible teaching opportunity for residents, students, and young pathologists to widen their knowledge.5

“It has been a very fun collaboration. I have gotten to learn brand new things about AI, and I have been able to provide my expertise in pulmonary pathology to hopefully make any algorithms we design as good as possible. The Aiforia team has been great about providing technical suggestions, advice, and support." – Dr. Jennifer M. Boland, Mayo Clinic

Read the full case study → 

 

5. Improve patient outcomes

Improved diagnostic accuracy and consistency of analysis with AI assistance benefit not only the hospital and pathologists but, more importantly, patients as well.

The potential benefits include:

  • Enhanced treatment efficacy
  • More personalized therapies are enabled
  • Reduction of the number of unnecessary interventions or surgeries
  • Enhanced quality of service as patients receive diagnoses faster

In Aiforia’s webinar, "Using prognostic AI models in pathology: case colorectal cancer," Dr. Rish Pai, MD, PhD, Pathologist at the Mayo Clinic, discussed how AI can aid oncologists in deciding which patients should get adjuvant chemotherapy and for how long. He explained that the decision whether to give chemotherapy in stage II and III colorectal cancer is based on pathologic features that are diagnostically challenging, features that pathologists often disagree with one another. His goal in developing the prognostic AI model, QuantCRC, was to help address this challenge. The AI model can make it easier to decide between three and six months of chemotherapy – reducing the total number of patients getting extensive chemotherapy but ensuring it is recommended for high-risk patients who can most benefit from it.

“To the healthcare system in general, integrating multiple data modalities will become more and more important in diagnostics and treatment decisions, especially if we think about personalized care. In this, AI could perhaps solve questions and suggest decisions that human experts can utilize in their educated decision-making,” explains Dr. Tuomas Mirtti, Consulting Clinical Pathologist and Chief Physician at the Helsinki University Hospital Diagnostic Center in Finland.

 

 

References

  1. Hollon et al. (2020, January). Near real-time intraoperative brain tumor diagnosis using stimulated Raman histology and deep neural networks. Nat Med., 26(1), 52-58. https://doi.org/10.1038/s41591-019-0715-9 

  2. Verghese et al. (2023, August 14). Computational pathology in cancer diagnosis, prognosis, and prediction – present day and prospects. The Journal of Pathology. https://doi.org/10.1002/path.6163 

  3. Tehrani et al. (2013, August). 25-Year summary of US malpractice claims for diagnostic errors 1986-2010: an analysis from the National Practitioner Data Bank. BMJ Qual Saf., 22(8), 672-680. https://doi.org/10.1136/bmjqs-2012-001550

  4. Homeyer et al. (2021, January-December). Artificial Intelligence in Pathology: From Prototype to Product. Journal of Pathology Informatics, 12(1). https://doi.org/10.4103/jpi.jpi_84_20 

  5. Eccher et al. (2024, May 14). Digital pathology structure and deployment in Veneto: a proof-of-concept study. Virchows Arch. https://doi.org/10.1007/s00428-024-03823-7