Artificial intelligence and pathology: From principles to practice and future applications in histomorphology and molecular profiling

Alber, Maximilian; Allgäuer, Michael; Jurmeister, Philipp; Bockmayr, Michael; Budczies, Jan; Lennerz, Jochen K.; Eschrich, Johannes; Kazdal, Daniel; Schirmacher, Peter; Wagner, Alex H.; Tacke, Frank; Capper, David; Müller, Klaus‐Robert; Klauschen, Frederick

doi:10.1016/j.semcancer.2021.02.011

Cited by 57 publications

(44 citation statements)

References 121 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We also propose four other possible ways in which AI model could be incorporated in pathology workflow; it could act as first reader, second reader, triage, and pre-screening (Fig. 5) 10,30,31 .…”

Section: Discussionmentioning

confidence: 99%

A promising deep learning-assistive algorithm for histopathological screening of colorectal cancer

Zhao

Chen

et al. 2022

Sci Rep

View full text Add to dashboard Cite

Colorectal cancer is one of the most common cancers worldwide, accounting for an annual estimated 1.8 million incident cases. With the increasing number of colonoscopies being performed, colorectal biopsies make up a large proportion of any histopathology laboratory workload. We trained and validated a unique artificial intelligence (AI) deep learning model as an assistive tool to screen for colonic malignancies in colorectal specimens, in order to improve cancer detection and classification; enabling busy pathologists to focus on higher order decision-making tasks. The study cohort consists of Whole Slide Images (WSI) obtained from 294 colorectal specimens. Qritive’s unique composite algorithm comprises both a deep learning model based on a Faster Region Based Convolutional Neural Network (Faster-RCNN) architecture for instance segmentation with a ResNet-101 feature extraction backbone that provides glandular segmentation, and a classical machine learning classifier. The initial training used pathologists’ annotations on a cohort of 66,191 image tiles extracted from 39 WSIs. A subsequent application of a classical machine learning-based slide classifier sorted the WSIs into ‘low risk’ (benign, inflammation) and ‘high risk’ (dysplasia, malignancy) categories. We further trained the composite AI-model’s performance on a larger cohort of 105 resections WSIs and then validated our findings on a cohort of 150 biopsies WSIs against the classifications of two independently blinded pathologists. We evaluated the area under the receiver-operator characteristic curve (AUC) and other performance metrics. The AI model achieved an AUC of 0.917 in the validation cohort, with excellent sensitivity (97.4%) in detection of high risk features of dysplasia and malignancy. We demonstrate an unique composite AI-model incorporating both a glandular segmentation deep learning model and a classical machine learning classifier, with excellent sensitivity in picking up high risk colorectal features. As such, AI plays a role as a potential screening tool in assisting busy pathologists by outlining the dysplastic and malignant glands.

show abstract

Section: Discussionmentioning

confidence: 99%

A promising deep learning-assistive algorithm for histopathological screening of colorectal cancer

Zhao

Chen

et al. 2022

Sci Rep

View full text Add to dashboard Cite

show abstract

“…The complexity of the biological tests will become associated with patient phenotype, genetic and epigenetic analyses, and an extension of therapeutic strategies. It is without any doubt that approaches based on algorithms that integrate developments in artificial intelligence will rapidly be envisaged in the domain of care of early stage NSCLC patients [123].…”

Section: Discussionmentioning

confidence: 99%

EGFR Status Assessment for Better Care of Early Stage Non-Small Cell Lung Carcinoma: What Is Changing in the Daily Practice of Pathologists?

Hofman

2021

Cells

View full text Add to dashboard Cite

The recent emergence of novel neoadjuvant and/or adjuvant therapies for early stage (I-IIIA) non-small cell lung carcinoma (NSCLC), mainly tyrosine kinase inhibitors (TKIs) targeting EGFR mutations and immunotherapy or chemo-immunotherapy, has suddenly required the evaluation of biomarkers predictive of the efficacy of different treatments in these patients. Currently, the choice of one or another of these treatments mainly depends on the results of immunohistochemistry for PD-L1 and of the status of EGFR and ALK. This new development has led to the setup of different analyses for clinical and molecular pathology laboratories, which have had to rapidly integrate a number of new challenges into daily practice and to establish new organization for decision making. This review outlines the impact of the management of biological samples in laboratories and discusses perspectives for pathologists within the framework of EGFR TKIs in early stage NSCLC.

show abstract

“…This bias-free computational approach yields insights into the global nature of brain aging as well as pathomechanisms. Finally, due to its generalizability, this approach can be broadly applied across clinical neuroscience, galvanizing the generation of data-driven hypotheses and boosting its applications in personalized medicine (Binder et al, 2021; Esteva et al, 2021; Stenzinger et al, 2021) .…”

Section: Discussionmentioning

confidence: 99%

Towards the Interpretability of Deep Learning Models for Multi-modal Neuroimaging: Finding Structural Changes of the Ageing Brain

Beyer

Lapuschkin

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Brain-age (BA) estimates based on deep learning are increasingly used as neuroimaging biomarker for brain health; however, the underlying neural features have remained unclear. We combined ensembles of convolutional neural networks with Layer-wise Relevance Propagation (LRP) to detect which brain features contribute to BA. Trained on magnetic resonance imaging (MRI) data of a population-based study (n=2637, 18-82 years), our models estimated age accurately based on single and multiple modalities, regionally restricted and whole-brain images (mean absolute errors 3.38-5.07 years). We find that BA estimates capture aging at both small and large-scale changes, revealing gross enlargements of ventricles and subarachnoid spaces, as well as lesions, iron accumulations and atrophies that appear throughout the brain. Divergence from expected aging reflected cardiovascular risk factors and accelerated aging was more pronounced in the frontal lobe. Applying LRP, our study demonstrates how superior deep learning models detect brain-aging in healthy and at-risk individuals throughout adulthood.

show abstract

Artificial intelligence and pathology: From principles to practice and future applications in histomorphology and molecular profiling

Cited by 57 publications

References 121 publications

A promising deep learning-assistive algorithm for histopathological screening of colorectal cancer

A promising deep learning-assistive algorithm for histopathological screening of colorectal cancer

EGFR Status Assessment for Better Care of Early Stage Non-Small Cell Lung Carcinoma: What Is Changing in the Daily Practice of Pathologists?

Towards the Interpretability of Deep Learning Models for Multi-modal Neuroimaging: Finding Structural Changes of the Ageing Brain

Contact Info

Product

Resources

About