Revolutionizing Digital Pathology With the Power of Generative Artificial Intelligence and Foundation Models

Waqas, Asim; Bui, Marilyn M.; Glassy, Eric F.; El Naqa, Issam; Borkowski, Piotr; Borkowski, Andrew A.; Rasool, Ghulam

doi:10.1016/j.labinv.2023.100255

Cited by 27 publications

(6 citation statements)

References 96 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These AI models are designed to assist pathologists with routine, time-consuming tasks, such as cell counting and screening large numbers of biopsies, as well as those with limited reproducibility, such as tumor grading and immunohistochemistry scoring. 3,4,6 As such, AI represents a potential solution to address the shortage of pathologists by streamlining the diagnostic workflow. 7,8 Artificial Neural Networks (ANNs) are softwares that mimic the connections in the human brain; put in series, they can be trained to specific tasks, such as recognizing a dog on a picture -or a melanoma on a slide.…”

Section: Introduction -Background and Significancementioning

confidence: 99%

Section: Introduction - Background and Significancementioning

confidence: 99%

“…9 These models are underpinned by multi-layered ANN and rely upon transformer architecture -a peculiar neural network structure introduced by Google researchers in 2017. 6,[10][11][12] In particular, LLMs are built by training with massive amounts of textual data (billions of words derived from the internet, including websites, academic books, scientific articles and websites), allowing them to recognize the patterns and structures of language by identifying relationships between words and phrases and, consequently, generate (by predicting linguistic patterns) natural human-like language text outputs that are contextually coherent to the input. [11][12][13][14][15][16] The multi-stage training process involves varying degrees of human intervention and includes an unsupervised learning step (pre-training step) for language structure acquisition, followed by a supervised learning step (fine-tuning step) for task-specific optimization.…”

Section: Introduction -Background and Significancementioning

confidence: 99%

See 2 more Smart Citations

Evaluation of ChatGPT’s Usefulness and Accuracy in Diagnostic Surgical Pathology

Guastafierro,

Corbitt,

Bressan

et al. 2024

Preprint

View full text Add to dashboard Cite

ChatGPT is an artificial intelligence capable of processing and generating human-like language. ChatGPT's role within clinical patient care and medical education has been explored; however, assessment of its potential in supporting histopathological diagnosis is lacking. In this study, we assessed ChatGPT's reliability in addressing pathology-related diagnostic questions across 10 subspecialties, as well as its ability to provide scientific references. We created five clinico-pathological scenarios for each subspecialty, posed to ChatGPT as open-ended or multiple-choice questions. Each question either asked for scientific references or not. Outputs were assessed by six pathologists according to: 1) usefulness in supporting the diagnosis and 2) absolute number of errors. All references were manually verified. We used directed acyclic graphs and structural causal models to determine the effect of each scenario type, field, question modality and pathologist evaluation. Overall, we yielded 894 evaluations. ChatGPT provided useful answers in 62.2% of cases. 32.1% of outputs contained no errors, while the remaining contained at least one error (maximum 18). ChatGPT provided 214 bibliographic references: 70.1% were correct, 12.1% were inaccurate and 17.8% did not correspond to a publication. Scenario variability had the greatest impact on ratings, followed by prompting strategy. Finally, latent knowledge across the fields showed minimal variation. In conclusion, ChatGPT provided useful responses in one-third of cases, but the number of errors and variability highlight that it is not yet adequate for everyday diagnostic practice and should be used with discretion as a support tool. The lack of thoroughness in providing references also suggests caution should be employed even when used as a self-learning tool. It is essential to recognize the irreplaceable role of human experts in synthesizing images, clinical data and experience for the intricate task of histopathological diagnosis.

show abstract

Section: Introduction -Background and Significancementioning

confidence: 99%

Section: Introduction - Background and Significancementioning

confidence: 99%

Section: Introduction -Background and Significancementioning

confidence: 99%

See 1 more Smart Citation

Evaluation of ChatGPT’s Usefulness and Accuracy in Diagnostic Surgical Pathology

Guastafierro,

Corbitt,

Bressan

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…The advent of high-throughput multi-omics technologies like next-generation sequencing (NGS), high-resolution radiological and histopathology imaging, and the rapid digitization of medical records has led to an explosion of diverse, multimodal data [ 8 ]. This data deluge has been a boon for machine learning, where abundant training data has directly enabled significant breakthroughs [ 9 , 10 ]. For example, the rise of large general-purpose datasets like Common Crawl for natural language processing (NLP) has fueled advances in language models and Artificial Intelligence (AI) assistants [ 11 ].…”

Section: Introductionmentioning

confidence: 99%

Building Flexible, Scalable, and Machine Learning-Ready Multimodal Oncology Datasets

Tripathi,

Waqas,

Venkatesan

et al. 2024

Sensors

Self Cite

View full text Add to dashboard Cite

The advancements in data acquisition, storage, and processing techniques have resulted in the rapid growth of heterogeneous medical data. Integrating radiological scans, histopathology images, and molecular information with clinical data is essential for developing a holistic understanding of the disease and optimizing treatment. The need for integrating data from multiple sources is further pronounced in complex diseases such as cancer for enabling precision medicine and personalized treatments. This work proposes Multimodal Integration of Oncology Data System (MINDS)—a flexible, scalable, and cost-effective metadata framework for efficiently fusing disparate data from public sources such as the Cancer Research Data Commons (CRDC) into an interconnected, patient-centric framework. MINDS consolidates over 41,000 cases from across repositories while achieving a high compression ratio relative to the 3.78 PB source data size. It offers sub-5-s query response times for interactive exploration. MINDS offers an interface for exploring relationships across data types and building cohorts for developing large-scale multimodal machine learning models. By harmonizing multimodal data, MINDS aims to potentially empower researchers with greater analytical ability to uncover diagnostic and prognostic insights and enable evidence-based personalized care. MINDS tracks granular end-to-end data provenance, ensuring reproducibility and transparency. The cloud-native architecture of MINDS can handle exponential data growth in a secure, cost-optimized manner while ensuring substantial storage optimization, replication avoidance, and dynamic access capabilities. Auto-scaling, access controls, and other mechanisms guarantee pipelines’ scalability and security. MINDS overcomes the limitations of existing biomedical data silos via an interoperable metadata-driven approach that represents a pivotal step toward the future of oncology data integration.

show abstract

“…Nongenerative, also known as task-specific, AI has been used for decades. Nongenerative AI involves selecting a particular task such as learning to play chess or counting mitosis on a pathology slide image, for example 12 . The nongenerative AI model does not create something new, but rather learns how to best perform a task.…”

mentioning

confidence: 99%

Enabling Personalized Medicine in Orthopaedic Surgery Through Artificial Intelligence

Huffman,

Pasqualini,

Khan

et al. 2024

JBJS Reviews

View full text Add to dashboard Cite

» The application of artificial intelligence (AI) in the field of orthopaedic surgery holds potential for revolutionizing health care delivery across 3 crucial domains: (I) personalized prediction of clinical outcomes and adverse events, which may optimize patient selection, surgical planning, and enhance patient safety and outcomes; (II) diagnostic automated and semiautomated imaging analyses, which may reduce time burden and facilitate precise and timely diagnoses; and (III) forecasting of resource utilization, which may reduce health care costs and increase value for patients and institutions.» Computer vision is one of the most highly studied areas of AI within orthopaedics, with applications pertaining to fracture classification, identification of the manufacturer and model of prosthetic implants, and surveillance of prosthesis loosening and failure.» Prognostic applications of AI within orthopaedics include identifying patients who will likely benefit from a specified treatment, predicting prosthetic implant size, postoperative length of stay, discharge disposition, and surgical complications. Not only may these applications be beneficial to patients but also to institutions and payors because they may inform potential cost expenditure, improve overall hospital efficiency, and help anticipate resource utilization.» AI infrastructure development requires institutional financial commitment and a team of clinicians and data scientists with expertise in AI that can complement skill sets and knowledge. Once a team is established and a goal is determined, teams (1) obtain, curate, and label data; (2) establish a reference standard; (3) develop an AI model; (4) evaluate the performance of the AI model; (5) externally validate the model, and (6) reinforce, improve, and evaluate the model's performance until clinical implementation is possible.» Understanding the implications of AI in orthopaedics may eventually lead to wide-ranging improvements in patient care. However, AI, while holding tremendous promise, is not without methodological and ethical limitations that are essential to address. First, it is important to ensure external validity of programs before their use in a clinical setting. Investigators should maintain high quality data records and registry surveillance, exercise caution when evaluating others' reported AI applications, and increase transparency of the methodological conduct of current models to improve external validity and avoid propagating bias. By addressing these challenges and responsibly embracing the potential of AI, the medical field may eventually be able to harness its power to improve patient care and outcomes.

show abstract

Revolutionizing Digital Pathology With the Power of Generative Artificial Intelligence and Foundation Models

Cited by 27 publications

References 96 publications

Evaluation of ChatGPT’s Usefulness and Accuracy in Diagnostic Surgical Pathology

Evaluation of ChatGPT’s Usefulness and Accuracy in Diagnostic Surgical Pathology

Building Flexible, Scalable, and Machine Learning-Ready Multimodal Oncology Datasets

Enabling Personalized Medicine in Orthopaedic Surgery Through Artificial Intelligence

Contact Info

Product

Resources

About