Artificial intelligence in histopathology: enhancing cancer research and clinical oncology

Nature Cancer, Journal Year: 2022, Volume and Issue: 3(9), P. 1026 - 1038

Published: Sept. 22, 2022

Language: Английский

---

Towards a general-purpose foundation model for computational pathology

Nature Medicine, Journal Year: 2024, Volume and Issue: 30(3), P. 850 - 862

Published: March 1, 2024

Language: Английский

---

Computational pathology in 2030: a Delphi study forecasting the role of AI in pathology within the next decade

EBioMedicine, Journal Year: 2023, Volume and Issue: 88, P. 104427 - 104427

Published: Jan. 4, 2023

Artificial intelligence (AI) is rapidly fuelling a fundamental transformation in the practice of pathology. However, clinical integration remains challenging, with no AI algorithms to date routine adoption within typical anatomic pathology (AP) laboratories. This survey gathered current expert perspectives and expectations regarding role AP from those first-hand computational experience.Perspectives were solicited using Delphi method 24 subject matter experts between December 2020 February 2021 anticipated by year 2030. The study consisted three consecutive rounds: 1) an open-ended, free response questionnaire generating list items; 2) Likert-scale scored analysed for consensus; 3) repeat items not reaching consensus obtain further consensus.Consensus opinions reached on 141 180 (78.3%). Experts agreed that would be routinely impactfully used laboratory pathologist workflows High was 100 across nine categories encompassing impact (1) key performance indicators (KPIs) (2) workforce specific tasks performed (3) pathologists (4) lab technicians, as well (5) applications their likelihood use 2030, (6) AI's integrated diagnostics, (7) likely fully automated AI, (8) regulatory/legal (9) ethical aspects pathology.This systematic details expected short-to-mid-term practice. These findings provide timely relevant information future care delivery raise practical, ethical, legal challenges must addressed prior successful implementation.No funding provided this study.

Language: Английский

---

Generalizable biomarker prediction from cancer pathology slides with self-supervised deep learning: A retrospective multi-centric study

Cell Reports Medicine, Journal Year: 2023, Volume and Issue: 4(4), P. 100980 - 100980

Published: March 22, 2023

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL also other biomarkers with high performance and predictions generalize to external patient populations. Here, we acquire CRC tissue samples two large multi-centric studies. We systematically compare six different state-of-the-art architectures pathology slides, including MSI mutations in BRAF, KRAS, NRAS, PIK3CA. Using a validation cohort provide realistic evaluation setting, show that models using self-supervised, attention-based multiple-instance consistently outperform previous approaches while offering explainable visualizations the indicative regions morphologies. While prediction BRAF reaches clinical-grade performance, mutation PIK3CA, NRAS was clinically insufficient.

Language: Английский

---

Adversarial attacks and adversarial robustness in computational pathology

Nature Communications, Journal Year: 2022, Volume and Issue: 13(1)

Published: Sept. 29, 2022

Abstract Artificial Intelligence (AI) can support diagnostic workflows in oncology by aiding diagnosis and providing biomarkers directly from routine pathology slides. However, AI applications are vulnerable to adversarial attacks. Hence, it is essential quantify mitigate this risk before widespread clinical use. Here, we show that convolutional neural networks (CNNs) highly susceptible white- black-box attacks clinically relevant weakly-supervised classification tasks. Adversarially robust training dual batch normalization (DBN) possible mitigation strategies but require precise knowledge of the type attack used inference. We demonstrate vision transformers (ViTs) perform equally well compared CNNs at baseline, orders magnitude more At a mechanistic level, associated with latent representation categories ViTs CNNs. Our results line previous theoretical studies provide empirical evidence learners computational pathology. This implies large-scale rollout models should rely on rather than CNN-based classifiers inherent protection against perturbation input data, especially

Language: Английский

---

Deep learning generates synthetic cancer histology for explainability and education

npj Precision Oncology, Journal Year: 2023, Volume and Issue: 7(1)

Published: May 29, 2023

Artificial intelligence methods including deep neural networks (DNN) can provide rapid molecular classification of tumors from routine histology with accuracy that matches or exceeds human pathologists. Discerning how make their predictions remains a significant challenge, but explainability tools help insights into what models have learned when corresponding histologic features are poorly defined. Here, we present method for improving DNN using synthetic generated by conditional generative adversarial network (cGAN). We show cGANs generate high-quality images be leveraged explaining trained to classify molecularly-subtyped tumors, exposing associated state. Fine-tuning through class and layer blending illustrates nuanced morphologic differences between tumor subtypes. Finally, demonstrate the use augmenting pathologist-in-training education, showing these intuitive visualizations reinforce improve understanding manifestations biology.

Language: Английский

---

The current state of digital cytology and artificial intelligence (AI): global survey results from the American Society of Cytopathology Digital Cytology Task Force

Journal of the American Society of Cytopathology, Journal Year: 2024, Volume and Issue: 13(5), P. 319 - 328

Published: April 16, 2024

The integration of whole slide imaging (WSI) and artificial intelligence (AI) with digital cytology has been growing gradually. Therefore, there is a need to evaluate the current state cytology. This study aimed determine landscape via survey conducted as part American Society Cytopathology (ASC) Digital Cytology White Paper Task Force. A 43 questions pertaining practices experiences WSI AI in both surgical pathology was created. sent members ASC, International Academy (IAC), Papanicolaou (PSC). Responses were recorded analyzed. In total, 327 individuals participated survey, spanning diverse array practice settings, roles, around globe. majority responses indicated routine scanning slides (n = 134; 61%) fewer respondents 150; 46%). primary challenge for faster cost minimization, whereas image quality top issue WSI. tools are not widely utilized only 16% participants using samples 13% practice. Utilization limited laboratories compared pathology. However, more willing implement near future establishment practical clinical guidelines needed.

Language: Английский

---

Direct prediction of genetic aberrations from pathology images in gastric cancer with swarm learning

Gastric Cancer, Journal Year: 2022, Volume and Issue: 26(2), P. 264 - 274

Published: Oct. 20, 2022

Computational pathology uses deep learning (DL) to extract biomarkers from routine slides. Large multicentric datasets improve performance, but such are scarce for gastric cancer. This limitation could be overcome by Swarm Learning (SL).Here, we report the results of a retrospective study SL prediction molecular in We collected tissue samples with known microsatellite instability (MSI) and Epstein-Barr Virus (EBV) status four patient cohorts Switzerland, Germany, UK USA, storing each dataset on physically separate computer.On an external validation cohort, SL-based classifier reached area under receiver operating curve (AUROC) 0.8092 (± 0.0132) MSI 0.8372 0.0179) EBV prediction. The centralized model, which was trained all single computer, similar performance.Our findings demonstrate feasibility In future, used collaborative training and, thus, performance these biomarkers. may ultimately result clinical-grade generalizability.

Language: Английский

---

Cutting-edge technology and automation in the pathology laboratory

Virchows Archiv, Journal Year: 2023, Volume and Issue: 484(4), P. 555 - 566

Published: Nov. 6, 2023

One of the goals pathology is to standardize laboratory practices increase precision and effectiveness diagnostic testing, which will ultimately enhance patient care results. Standardization crucial in domains tissue processing, analysis, reporting. To innovative technologies are also being created put into use. Furthermore, although problems like algorithm training data privacy issues still need be resolved, digital artificial intelligence emerging a structured manner. Overall, for field advance improved, standard must adopted. In this paper, we describe state-of-the-art automation laboratories order lead technological progress evolution. By anticipating needs demands, aim inspire innovation tools processes as positively transformative support operators, organizations, patients.

Language: Английский

---

Standardized Classification of Lung Adenocarcinoma Subtypes and Improvement of Grading Assessment Through Deep Learning

American Journal Of Pathology, Journal Year: 2023, Volume and Issue: 193(12), P. 2066 - 2079

Published: Aug. 5, 2023

Language: Английский

---