Artificial Intelligence Neuropathologist
Study Details
Study Description
Brief Summary
CNS tumor requires biopsy for pathological diagnosis, which is known as the "golden standard". We would like to achieve automated classification of brain tumors based on deep learning in digital histopathology images and molecular pathology results. We expect to develop an assistant system (including software and hardware), to help pathologists during their diagnosis for CNS tumor.
Condition or Disease | Intervention/Treatment | Phase |
---|---|---|
|
Detailed Description
The aim of the study is to develop an automated pathological diagnosis system for CNS tumors based on deep learning technique. It is designed to firstly develop the best deep learning model for pathological diagnosis of CNS tumors, in order to improve the accuracy of pathological diagnosis. Then to be used clinically, reduce the workload and stress of neuropathologists and obtain the benefits for CNS tumor patients.
Different CNS tumors including meningioma, glioma, lymphoma and other various tumors have their own different treatment principles and plans. For example, high grade glioma requires operational resection and post-operational chemo-radiotherapy. However, operational resection is not significant for improving prognosis in lymphoma patients, systematic chemotherapy will be performed after specific diagnosis based on biopsy. Therefore, in this study, an automated CNS tumor pathological diagnosis system will be developed to classify the different type of those tumors.
At present, pathological diagnosis of CNS tumors is based on histopathological characteristics and molecular information after a systematic analyzed by pathologists. The accuracy of the diagnosis very much relies on the experience of the pathologists. However, to become a experienced and qualified pathologist requires years of training. Pathologists may give completely different diagnose outcome for the same patient. Thus, it is essential to develop a system that can assist pathologists.
Deep learning is one of the most advanced techniques of artificial intelligence. In particular, the ability of image recognition is extremely powerful. Therefore, we are able to develop a model for histopathological section images based on deep learning. WHO Classification of CNS Tumors 2016 has included molecular markers as the important part of diagnosis. Hence, there will be an additional model of molecular pathology to be added to the system.
Huashan Hospital has one of the largest CNS tumor biobank in China, which is the key part for deep learning, as it needs large amount of data. The case load of this study is able to show the representative and authoritative of those data.
There will be three stages of the study. Stage 1 and 2 are supervised learning process. Stage 1 is to develop the best deep learning model for histopathological diagnosis of CNS tumors, we anticipate the accuracy for the first model to achieve at least 70%. The training data (pathological sections) will be provided by Huashan Hospital CNS tumor biobank. In the mean time, a micro-positioning platform is under investigation for the use of image collection. At the end of stage 1, we anticipate to integrate the model (software) and the platform (hardware) as the whole diagnose system for histopathological images. Stage 2 is to design a model for molecular pathological diagnosis for CNS tumors. The model will be trained by numerous amount of related molecular information extracted from those pathological sections. At the end of stage 2, we anticipate to combine stage 1 system and stage 2 model as the primary prototype. Stage 3 is known as the unsupervised learning process. By using the prototype developed after previous stages, the system will be used clinically. With the incoming of more patients and data, together with pathologists in the hospital, it will give its diagnosis. By comparing the results with pathologists, it will be able to self-learn and improve the accuracy as the time goes on. By the end of stage 3, we anticipate to have the system ready for independent clinical pathological diagnosis ability with the accuracy greater than 90%.
Study Design
Arms and Interventions
Arm | Intervention/Treatment |
---|---|
CNS Tumor All patients age from 18-75 years with CNS tumors are included and count as one group |
Outcome Measures
Primary Outcome Measures
- Automated histopathological diagnosis outcome (software development) [Nov,2018 - Nov,2019]
After supervised training, the software of the histopathological diagnosis of CNS tumor achieve at least 70% accuracy
- Positioning platform for microscope (hardware development) [Nov,2018 - Nov,2019]
Hardware investigation for pathology section image collection, to automatically scan the section images.
- Combine automated molecular pathological diagnosis [Nov,2019 - Jun,2020]
Molecular information being added to the histopathological diagnosis regarding to WHO 2016 CNS Tumor guide. Combine histopathology and molecular to give final diagnosis
Secondary Outcome Measures
- Unsupervised training with more cases to improve the system [Nov,2019 - Nov,2022]
Improve diagnosis accuracy of the system by continuous collection with a large number of CNS tumor cases from Huashan Hospital, which can be done without supervision.
Eligibility Criteria
Criteria
Inclusion Criteria:
The participants diagnosed with brain cancer by diagnosis of WHO 2016 classification of CNS tumors.
Exclusion Criteria:
Voluntarily quit
Contacts and Locations
Locations
Site | City | State | Country | Postal Code | |
---|---|---|---|---|---|
1 | Hushan Hospital, Fudan University | Shanghai | Shanghai | China | 200040 |
Sponsors and Collaborators
- Huashan Hospital
- United Imaging Healthcare
Investigators
- Study Chair: Jinsong Wu, Ph.D. & M.D, Huashan Hospital
Study Documents (Full-Text)
None provided.More Information
Publications
- Ehteshami Bejnordi B, Veta M, Johannes van Diest P, van Ginneken B, Karssemeijer N, Litjens G, van der Laak JAWM; the CAMELYON16 Consortium, Hermsen M, Manson QF, Balkenhol M, Geessink O, Stathonikos N, van Dijk MC, Bult P, Beca F, Beck AH, Wang D, Khosla A, Gargeya R, Irshad H, Zhong A, Dou Q, Li Q, Chen H, Lin HJ, Heng PA, Haß C, Bruni E, Wong Q, Halici U, Öner MÜ, Cetin-Atalay R, Berseth M, Khvatkov V, Vylegzhanin A, Kraus O, Shaban M, Rajpoot N, Awan R, Sirinukunwattana K, Qaiser T, Tsang YW, Tellez D, Annuscheit J, Hufnagl P, Valkonen M, Kartasalo K, Latonen L, Ruusuvuori P, Liimatainen K, Albarqouni S, Mungal B, George A, Demirci S, Navab N, Watanabe S, Seno S, Takenaka Y, Matsuda H, Ahmady Phoulady H, Kovalev V, Kalinovsky A, Liauchuk V, Bueno G, Fernandez-Carrobles MM, Serrano I, Deniz O, Racoceanu D, Venâncio R. Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women With Breast Cancer. JAMA. 2017 Dec 12;318(22):2199-2210. doi: 10.1001/jama.2017.14585.
- Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017 Feb 2;542(7639):115-118. doi: 10.1038/nature21056. Epub 2017 Jan 25. Erratum in: Nature. 2017 Jun 28;546(7660):686.
- Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy A, Venugopalan S, Widner K, Madams T, Cuadros J, Kim R, Raman R, Nelson PC, Mega JL, Webster DR. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016 Dec 13;316(22):2402-2410. doi: 10.1001/jama.2016.17216.
- Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, van der Laak JAWM, van Ginneken B, Sánchez CI. A survey on deep learning in medical image analysis. Med Image Anal. 2017 Dec;42:60-88. doi: 10.1016/j.media.2017.07.005. Epub 2017 Jul 26. Review.
- Louis DN, Perry A, Reifenberger G, von Deimling A, Figarella-Branger D, Cavenee WK, Ohgaki H, Wiestler OD, Kleihues P, Ellison DW. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol. 2016 Jun;131(6):803-20. doi: 10.1007/s00401-016-1545-1. Epub 2016 May 9. Review.
- Schmidhuber J. Deep learning in neural networks: an overview. Neural Netw. 2015 Jan;61:85-117. Epub 2014 Oct 13. Review.
- Wen PY, Huse JT. 2016 World Health Organization Classification of Central Nervous System Tumors. Continuum (Minneap Minn). 2017 Dec;23(6, Neuro-oncology):1531-1547. doi: 10.1212/CON.0000000000000536. Review.
- Yu KH, Zhang C, Berry GJ, Altman RB, Ré C, Rubin DL, Snyder M. Predicting non-small cell lung cancer prognosis by fully automated microscopic pathology image features. Nat Commun. 2016 Aug 16;7:12474. doi: 10.1038/ncomms12474.
- KY2017-340