Dataset Information

Deep learning-based transformation of H&E stained tissues into special stains.

ABSTRACT: Pathology is practiced by visual inspection of histochemically stained tissue slides. While the hematoxylin and eosin (H&E) stain is most commonly used, special stains can provide additional contrast to different tissue components. Here, we demonstrate the utility of supervised learning-based computational stain transformation from H&E to special stains (Masson's Trichrome, periodic acid-Schiff and Jones silver stain) using kidney needle core biopsy tissue sections. Based on the evaluation by three renal pathologists, followed by adjudication by a fourth pathologist, we show that the generation of virtual special stains from existing H&E images improves the diagnosis of several non-neoplastic kidney diseases, sampled from 58 unique subjects (P = 0.0095). A second study found that the quality of the computationally generated special stains was statistically equivalent to those which were histochemically stained. This stain-to-stain transformation framework can improve preliminary diagnoses when additional special stains are needed, also providing significant savings in time and cost.

SUBMITTER: de Haan K

PROVIDER: S-EPMC8361203 | biostudies-literature | 2021 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep learning-based transformation of H&E stained tissues into special stains.

de Haan Kevin K Zhang Yijie Y Zuckerman Jonathan E JE Liu Tairan T Sisk Anthony E AE Diaz Miguel F P MFP Jen Kuang-Yu KY Nobori Alexander A Liou Sofia S Zhang Sarah S Riahi Rana R Rivenson Yair Y Wallace W Dean WD Ozcan Aydogan A

Nature communications 20210812 1

Pathology is practiced by visual inspection of histochemically stained tissue slides. While the hematoxylin and eosin (H&E) stain is most commonly used, special stains can provide additional contrast to different tissue components. Here, we demonstrate the utility of supervised learning-based computational stain transformation from H&E to special stains (Masson's Trichrome, periodic acid-Schiff and Jones silver stain) using kidney needle core biopsy tissue sections. Based on the evaluation by th ...[more]

PMID: 34385460

Similar Datasets

Project description:The objective of the study was to investigate whether special stains can differentiate gastrointestinal stromal tumors (GISTs) and gastrointestinal leiomyomas (GILs). In this retrospective study, 39 cases of GISTs (diameter, 0.2-8.8 cm) and 75 cases of GILs (diameter, 0.2-4.5 cm) were recruited, all biopsy specimens were obtained by endoscopic submucosal dissection (ESD) and endoscopic mucosal resection (EMR) excision, and the depth of excision included the whole mucosa, mucosal myometria, and most submucosa. GISTs and GILs were the most common types of mesenchymal tumors found anywhere along the gastrointestinal (GI) tract, from the esophagus to the rectum. GISTs were often associated with a higher risk of malignancy. In this study, the gender, age of onset, size and sites of the lesions, together with the number of mucosal or lamina propria lesions all have significant differences, nevertheless, there was no significant difference in cell morphology of GISTs and GILs tested by hematoxylin eosin (H&E) stain, and all showed low echo areas by EUS examination. In this retrospective study, the GISTs and GILs had been diagnosed by immunohistochemistry combined with clinical morphology. Subsequently, special stains including Masson's trichrome (MT) stain, Alcian blue periodic acid-Schiff (AB-PAS) stain (pH 2.5), Wright-Giemsa (W-G) stain and periodic acid-Schiff (PAS) combined with diastase periodic acid-Schiff (D-PAS) stains were also applied in the diagnosis, the retrospective study results showed that 92.3% GISTs were stained blue with MT stain, 97.3% GILs were stained red with MT stain (P < 0.01), almost all GISTs were PAS-negative (light purple), in contrast, all GILs were PAS-positive (rose red) (P < 0.01), all of these experiments set control using the blood vessels stained by MT and AB-PAS stains. Nevertheless, there was no significant difference between GISTs and GILs stained by W-G stain. These obvious and meaningful differential results were also confirmed in the detection of new GISTs and GILs cases using MT and AB-PAS stains. In conclusion, MT and AB-PAS stains could also identify GISTs and GILs cases, particularly, AB-PAS was more sensitive and more specific, providing a more cost-effective, simple, and high sensitivity and specificity inspection methods, which should be noticed and widely used in the future, especially in resource-limited grass-roots testing institution or in cases with inconclusive immunostains or insufficient material.

Dataset Information

Deep learning-based transformation of H&E stained tissues into special stains.

Publications

Deep learning-based transformation of H&E stained tissues into special stains.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets