Realistic in Generation of HEp-2 Cell Images Using Latent Diffusion Models: a Multi-center Visual Turing Test

NCT ID: NCT06542783

Last Updated: 2024-08-07

Basic Information

• Recruitment Status: NOT_YET_RECRUITING
• Total Enrollment: 300 participants
• Study Classification: OBSERVATIONAL
• Study Start Date: 2024-09-30
• Study Completion Date: 2026-06-30

Brief Summary

The objective of this prospective observational study is to rigorously examine the feasibility and efficacy of utilizing latent diffusion models for data augmentation in anti-nuclear antibody (ANA) Hep-2 cell immunofluorescence images. The main question it aims to answer is:

Can the application of such models potentially enhance the data quality, increase sample diversity, or improve the accuracy and efficiency of subsequent analytical processes (like disease diagnosis and classification) when utilized with ANA-related images?

Detailed Description

A fundamental problem in biomedical research is the low number of observations available, mostly due to a lack of available biosamples, prohibitive costs, or ethical reasons. Augmenting few real observations with generated in silico samples could lead to more robust analysis results and a higher reproducibility rate. Here, The investigators propose to use unsupervised learning with latent diffusion models for the realistic generation of ANA-IIF image data.

The investigators hypothesize that the the generation of ANA-IIF image will be realistic if it is hard to differentiate them (fake) from real (true) . To test this hypothesis, the investigators present a Multi-center Visual Turing tests (https://turing.rednoble.net/) in order to evaluate the quality of the generated (fake) images.

This experimental setup allows the investigators to validate the overall quality of the generated ANA-IIF images, which can then be used to (1) train cytopathologists for educational purposes, and (2) generate realistic samples to train deep networks with big data.

Conditions

Anti-nuclear Antibody; Visual Turing Tests; Artifical Intelligence

Study Design

• Observational Model Type: COHORT
• Study Time Perspective: PROSPECTIVE

Study Groups

experts

with over 20 years of experience in ANA-IIF reading

referring to the results of AI model output

Intervention Type: BEHAVIORAL

determining the ANA pattern type with or without referring to the results of AI model output.

junior cytopathologists

less than 5 years of academic medical experience

referring to the results of AI model output

Intervention Type: BEHAVIORAL

determining the ANA pattern type with or without referring to the results of AI model output.

Interventions

referring to the results of AI model output

determining the ANA pattern type with or without referring to the results of AI model output.

Intervention Type: BEHAVIORAL

Eligibility Criteria

Inclusion Criteria

• Originating from reputable medical institutions
• Possessing relevant certification and qualifications
• Having over one year of experience in interpreting anti-nuclear antibody (ANA) patterns within a laboratory setting

Exclusion Criteria

• Lacking relevant professional certification and qualifications
• Without experience in interpreting ANA patterns
• Unwilling to accept the rules and informed consent of the visual Turing test

• Eligible Sex: ALL
• Accepts Healthy Volunteers: Yes

Sponsors

Xinhua Hospital, Shanghai Jiao Tong University School of Medicine

OTHER

Sponsor Role: lead

Responsible Party

Responsibility Role: SPONSOR

Principal Investigators

Guangyu Chen, PhD

Role: STUDY_DIRECTOR

Xinhua Hospital, Shanghai Jiao Tong University School of Medicine

Central Contacts

Junxiang Zeng, Dr

Role: CONTACT

zengjunxiang@xinhuamed.com.cn

+8613162232879

References

Zeng J, Gao X, Gao L, Yu Y, Shen L, Pan X. Recognition of rare antinuclear antibody patterns based on a novel attention-based enhancement framework. Brief Bioinform. 2024 Jan 22;25(2):bbad531. doi: 10.1093/bib/bbad531.

Reference Type: BACKGROUND

PMID: 38279651 (View on PubMed)

Rahman S, Wang L, Sun C, Zhou L. Deep learning based HEp-2 image classification: A comprehensive review. Med Image Anal. 2020 Oct;65:101764. doi: 10.1016/j.media.2020.101764. Epub 2020 Jul 7.

Reference Type: BACKGROUND

PMID: 32745976 (View on PubMed)

Hobson P, Lovell BC, Percannella G, Vento M, Wiliem A. Benchmarking human epithelial type 2 interphase cells classification methods on a very large dataset. Artif Intell Med. 2015 Nov;65(3):239-50. doi: 10.1016/j.artmed.2015.08.001. Epub 2015 Aug 13.

Reference Type: BACKGROUND

PMID: 26303104 (View on PubMed)

Niehues JM, Muller-Franzes G, Schirris Y, Wagner SJ, Jendrusch M, Kloor M, Pearson AT, Muti HS, Hewitt KJ, Veldhuizen GP, Zigutyte L, Truhn D, Kather JN. Using histopathology latent diffusion models as privacy-preserving dataset augmenters improves downstream classification performance. Comput Biol Med. 2024 Jun;175:108410. doi: 10.1016/j.compbiomed.2024.108410. Epub 2024 Apr 4.

Reference Type: BACKGROUND

PMID: 38678938 (View on PubMed)

Selim M, Zhang J, Brooks MA, Wang G, Chen J. DiffusionCT: Latent Diffusion Model for CT Image Standardization. AMIA Annu Symp Proc. 2024 Jan 11;2023:624-633. eCollection 2023.

Reference Type: BACKGROUND

PMID: 38222387 (View on PubMed)

Marouf M, Machart P, Bansal V, Kilian C, Magruder DS, Krebs CF, Bonn S. Realistic in silico generation and augmentation of single-cell RNA-seq data using generative adversarial networks. Nat Commun. 2020 Jan 9;11(1):166. doi: 10.1038/s41467-019-14018-z.

Reference Type: BACKGROUND

PMID: 31919373 (View on PubMed)

Provided Documents

Document Type: Study Protocol and Statistical Analysis Plan

View Document

Related Links

https://turing.rednoble.net/

Our platform for the Visual Turing Test

Other Identifiers

XH-24-007

Identifier Type: -

Identifier Source: org_study_id