Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

Lituiev, Dmytro; Jik, Sung; Chin, Aaron; Glicksberg, Benjamin S.; Bishara, Andrew; Dobi, Deján; Cheng, Ruizhe; Sohn, Jae Ho; Lászik, Zoltán; Hadley, Dexter

doi:10.1101/19002634

Cited by 3 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…inflammation scoring while using algorithms fitting the most current Banff classification to decrease inter‐observer variability. Because personalized medicine has been a major focus in recent years, utilizing algorithms and archetype analysis has the potential to combine biopsy results with clinical and laboratory data and to provide a more complete pathologic and prognostic picture, especially in the era of expanding electronic medical records 31 . Furthermore, A. Loupy presented the capability of natural language processing (NLP) to help such algorithms assist clinicians and reduce human errors.…”

Section: Role Of Artificial Intelligence Data Integration and Machimentioning

confidence: 99%

The Banff 2019 Kidney Meeting Report (I): Updates on and clarification of criteria for T cell– and antibody-mediated rejection

Loupy

Haas

Roufosse

et al. 2020

American Journal of Transplantation

559

545

View full text Add to dashboard Cite

The XV. Banff conference for allograft pathology was held in conjunction with the annual meeting of the American Society for Histocompatibility and Immunogenetics in Pittsburgh, PA (USA) and focused on refining recent updates to the classification, advances from the Banff working groups, and standardization of molecular diagnostics. This report on kidney transplant pathology details clarifications and refinements to the criteria for chronic active (CA) T cell–mediated rejection (TCMR), borderline, and antibody‐mediated rejection (ABMR). The main focus of kidney sessions was on how to address biopsies meeting criteria for CA TCMR plus borderline or acute TCMR. Recent studies on the clinical impact of borderline infiltrates were also presented to clarify whether the threshold for interstitial inflammation in diagnosis of borderline should be i0 or i1. Sessions on ABMR focused on biopsies showing microvascular inflammation in the absence of C4d staining or detectable donor‐specific antibodies; the potential value of molecular diagnostics in such cases and recommendations for use of the latter in the setting of solid organ transplantation are presented in the accompanying meeting report. Finally, several speakers discussed the capabilities of artificial intelligence and the potential for use of machine learning algorithms in diagnosis and personalized therapeutics in solid organ transplantation.

show abstract

Section: Role Of Artificial Intelligence Data Integration and Machimentioning

confidence: 99%

The Banff 2019 Kidney Meeting Report (I): Updates on and clarification of criteria for T cell– and antibody-mediated rejection

Loupy

Haas

Roufosse

et al. 2020

American Journal of Transplantation

559

545

View full text Add to dashboard Cite

show abstract

“…The model was trained using a binary cross-entropy loss and the Adam optimizer with learning rate 10 -5 and default parameters otherwise. Implementation is available on Github 22 .…”

Section: Model Architecture and Trainingmentioning

confidence: 99%

Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

Hadley¹

2022

CPRJ

View full text Add to dashboard Cite

Allograft rejection is a major concern in kidney transplantation. Inflammatory processes in patients with kidney allografts involve various patterns of immune cell recruitment and distributions. Lymphoid aggregates (LAs) are commonly observed in patients with kidney allografts and their presence and localization may correlate with severity of acute rejection. Alongside with other markers of inflammation, LAs assessment is currently performed by pathologists manually in a qualitative way, which is both time consuming and far from precise. Here we present the first automated method of identifying LAs and measuring their densities in whole slide images of transplant kidney biopsies. We trained a deep convolutional neural network based on U-Net on 44 core needle kidney biopsy slides, monitoring loss on a validation set (n=7 slides). The model was subsequently tested on a hold-out set (n=10 slides).We found that the coarse pattern of LAs localization agrees between the annotations and predictions, which is reflected by high correlation between the annotated and predicted fraction of LAs area per slide (Pearson R of 0.9756). Furthermore, the network achieves an auROC of 97.78 ± 0.93% and an IoU score of 69.72 ± 6.24 % per LAcontaining slide in the test set. Our study demonstrates that a deep convolutional neural network can accurately identify lymphoid aggregates in digitized histological slides of kidney. This study presents a first automatic DL-based approach for quantifying inflammation marks in allograft kidney, which can greatly improve precision and speed of assessment of allograft kidney biopsies when implemented as a part of computer-aided diagnosis system.

show abstract

“…The model was trained using a binary cross-entropy loss and the Adam optimizer with learning rate 10 -5 and default parameters otherwise [22]. Implementation is available on Github [23].…”

Section: Model Architecture and Trainingmentioning

confidence: 99%

Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

2021

Microbiol Immunol Pathol

View full text Add to dashboard Cite

high performance in image segmentation tasks [6,7], in which each pixel of an image is assigned a discrete class. By assigning each image pixel a discrete class label, the image can be segmented into distinct regions of interest. The use of CNNs have thus recently gained traction in biomedical segmentation [6] and digital pathology tasks, such as the detection and localization of breast cancer and its metastases in lymph nodes [8][9][10]. Based on such successful applications, a number of DL-based approaches have been approved for clinical use in the USA and other countries and are already contributing to higher precision and efficiency in many domains of diagnostics and treatment, such as diabetes care, oncology, cardiology, and radiology to name a few [11]. DL has been successfully applied for several routine nephrologyrelated evaluation procedures [12][13][14], which promises to improve speed

show abstract

Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

Cited by 3 publications

References 17 publications

The Banff 2019 Kidney Meeting Report (I): Updates on and clarification of criteria for T cell– and antibody-mediated rejection

The Banff 2019 Kidney Meeting Report (I): Updates on and clarification of criteria for T cell– and antibody-mediated rejection

Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

Automated Localization and Segmentation of Mononuclear Cell Aggregates in Kidney Histological Images Using Deep Learning

Contact Info

Product

Resources

About