1
|
Xie Y, Sang Q, Da Q, Niu G, Deng S, Feng H, Chen Y, Li YY, Liu B, Yang Y, Dai W. Improving diagnosis and outcome prediction of gastric cancer via multimodal learning using whole slide pathological images and gene expression. Artif Intell Med 2024; 152:102871. [PMID: 38685169 DOI: 10.1016/j.artmed.2024.102871] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/25/2023] [Revised: 03/08/2024] [Accepted: 04/15/2024] [Indexed: 05/02/2024]
Abstract
For the diagnosis and outcome prediction of gastric cancer (GC), machine learning methods based on whole slide pathological images (WSIs) have shown promising performance and reduced the cost of manual analysis. Nevertheless, accurate prediction of GC outcome may rely on multiple modalities with complementary information, particularly gene expression data. Thus, there is a need to develop multimodal learning methods to enhance prediction performance. In this paper, we collect a dataset from Ruijin Hospital and propose a multimodal learning method for GC diagnosis and outcome prediction, called GaCaMML, which is featured by a cross-modal attention mechanism and Per-Slide training scheme. Additionally, we perform feature attribution analysis via integrated gradient (IG) to identify important input features. The proposed method improves prediction accuracy over the single-modal learning method on three tasks, i.e., survival prediction (by 4.9% on C-index), pathological stage classification (by 11.6% on accuracy), and lymph node classification (by 12.0% on accuracy). Especially, the Per-Slide strategy addresses the issue of a high WSI-to-patient ratio and leads to much better results compared with the Per-Person training scheme. For the interpretable analysis, we find that although WSIs dominate the prediction for most samples, there is still a substantial portion of samples whose prediction highly relies on gene expression information. This study demonstrates the great potential of multimodal learning in GC-related prediction tasks and investigates the contribution of WSIs and gene expression, respectively, which not only shows how the model makes a decision but also provides insights into the association between macroscopic pathological phenotypes and microscopic molecular features.
Collapse
Affiliation(s)
- Yuzhang Xie
- Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China
| | - Qingqing Sang
- Department of General Surgery, Shanghai Key Laboratory of Gastric Neoplasm, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China
| | - Qian Da
- Department of Pathology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, 200025, China
| | - Guoshuai Niu
- Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China
| | - Shijie Deng
- Department of Pathology, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, 200025, China
| | - Haoran Feng
- Department of General Surgery, Shanghai Key Laboratory of Gastric Neoplasm, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China
| | - Yunqin Chen
- Shanghai-MOST Key Laboratory of Health and Disease Genomics, Shanghai Institute for Biomedical and Pharmaceutical Technologies, Shanghai 200237, China; Shanghai Engineering Research Center of Pharmaceutical Translation, Shanghai, 201203, China
| | - Yuan-Yuan Li
- Shanghai-MOST Key Laboratory of Health and Disease Genomics, Shanghai Institute for Biomedical and Pharmaceutical Technologies, Shanghai 200237, China; Shanghai Engineering Research Center of Pharmaceutical Translation, Shanghai, 201203, China
| | - Bingya Liu
- Department of General Surgery, Shanghai Key Laboratory of Gastric Neoplasm, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China.
| | - Yang Yang
- Department of Computer Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200240, China.
| | - Wentao Dai
- Department of General Surgery, Shanghai Key Laboratory of Gastric Neoplasm, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, China; Shanghai-MOST Key Laboratory of Health and Disease Genomics, Shanghai Institute for Biomedical and Pharmaceutical Technologies, Shanghai 200237, China; Shanghai Engineering Research Center of Pharmaceutical Translation, Shanghai, 201203, China.
| |
Collapse
|
2
|
An Increase in Plasma Adipsin Levels Is Associated With Higher Cumulative Dust Exposure and Airway Obstruction in Foundry Workers. J Occup Environ Med 2023; 65:203-209. [PMID: 36730948 DOI: 10.1097/jom.0000000000002736] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 02/04/2023]
Abstract
OBJECTIVE The aim of the study was to assess whether plasma adipokine levels (adipsin, adiponectin, leptin, and resistin) are associated with pulmonary function in foundry workers. METHODS We examined 65 dust-exposed foundry workers and 40 nonexposed controls and analyzed their lung function and plasma adipokine levels at baseline and after approximately 7 years of follow-up. RESULTS A higher increase in plasma adipsin was associated with the development of airway obstruction in exposed subjects during follow-up after adjusting for body mass index changes during the follow-up period. Furthermore, the increase in adipsin levels was positively associated with cumulative dust exposure even after adjusting for smoking and body mass index changes during follow-up ( P = 0.015). CONCLUSION The results suggest that plasma adipsin is involved in the pathogenesis of subclinical airway inflammation and the development of chronic obstruction and is induced by occupational dust exposure.
Collapse
|
3
|
Lung T, Sakem B, Risch M, Nydegger U. Convalescent blood plasma (CBP) donated by recovered COVID-19 patients - A comment. Transfus Apher Sci 2021; 60:103108. [PMID: 33678561 PMCID: PMC7923855 DOI: 10.1016/j.transci.2021.103108] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/21/2021] [Revised: 02/18/2021] [Accepted: 02/25/2021] [Indexed: 11/28/2022]
Affiliation(s)
- Th Lung
- Center for Laboratory Medicine Dr Risch, Vaduz, Principality of Liechtenstein, University of Bern, Switzerland
| | - B Sakem
- Center for Laboratory Medicine Dr Risch, Vaduz, Principality of Liechtenstein, University of Bern, Switzerland
| | - M Risch
- Center for Laboratory Medicine Dr Risch, Vaduz, Principality of Liechtenstein, University of Bern, Switzerland
| | - U Nydegger
- Center for Laboratory Medicine Dr Risch, Vaduz, Principality of Liechtenstein, University of Bern, Switzerland.
| |
Collapse
|
4
|
Lung T, Sakem B, Risch L, Würzner R, Colucci G, Cerny A, Nydegger U. The complement system in liver diseases: Evidence-based approach and therapeutic options. J Transl Autoimmun 2019; 2:100017. [PMID: 32743505 PMCID: PMC7388403 DOI: 10.1016/j.jtauto.2019.100017] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/21/2019] [Revised: 09/05/2019] [Accepted: 09/10/2019] [Indexed: 12/12/2022] Open
Abstract
Complement is usually seen to largely originate from the liver to accomplish its tasks systemically - its return to the production site has long been underestimated. Recent progress in genomics, therapeutic effects on complement, standardised possibilities in medical laboratory tests and involvement of complosome brings the complement system with its three major functions of opsonization, cytolysis and phagocytosis back to liver biology and pathology. The LOINC™ system features 20 entries for the C3 component of complement to anticipate the application of artificial intelligence data banks algorythms of which are fed with patient-specific data connected to standard lab assays for liver function. These advancements now lead to increased vigilance by clinicians. This reassessment article will further elucidate the distribution of synthesis sites to the three germ layer-derived cell systems and the role complement now known to play in embryogenesis, senescence, allotransplantation and autoimmune disease. This establishes the liver as part of the gastro-intestinal system in connection with nosological entities never thought of, such as the microbiota-liver-brain axis. In neurological disease etiology infectious and autoimmune hepatitis play an important role in the context of causative viz reactive complement activation. The mosaic of autoimmunity, i.e. multiple combinations of the many factors producing varying clinical pictures, leads to the manifold facets of liver autoimmunity.
Collapse
Affiliation(s)
- Thomas Lung
- Labormedizinisches Zentrum Dr. Risch, Lagerstrasse 30, CH-9470, Buchs, Switzerland
| | - Benjamin Sakem
- Labormedizinisches Zentrum Dr. Risch, Waldeggstrasse 37, CH-3097, Liebefeld bei Bern, Switzerland
| | - Lorenz Risch
- Labormedizinisches Zentrum Dr. Risch, Waldeggstrasse 37, CH-3097, Liebefeld bei Bern, Switzerland
| | - Reinhard Würzner
- Medical University Innsbruck, Division of Hygiene & Medical Microbiology, Department of Hygiene, Microbiology and Public Health, Schöpfstrasse 41, A-6020, Innsbruck, Austria
| | - Giuseppe Colucci
- Clinica Luganese Moncucco, Lugano, Via Moncucco, CH-6900, Lugano, Switzerland
- Faculty of Medicine, University of Basel, Basel, Switzerland
| | - Andreas Cerny
- Epatocentro Ticino, Via Soldino 5, CH-6900, Lugano, Switzerland
| | - Urs Nydegger
- Labormedizinisches Zentrum Dr. Risch, Waldeggstrasse 37, CH-3097, Liebefeld bei Bern, Switzerland
| |
Collapse
|