We use cookies to understand how you use our site and to improve your experience. This includes personalizing content and advertising. To learn more, click here. By continuing to use our site, you accept our use of cookies. Cookie Policy.

Features Partner Sites Information LinkXpress hp
Sign In
Advertise with Us

Download Mobile App

AI Accurately Predicts Cancer Outcomes from Tissue Samples

By LabMedica International staff writers
Posted on 27 Dec 2023
Print article
Image: The AI model analyzes cell spatial organization to make patient diagnoses and prognoses (Photo courtesy of UT Southwestern)
Image: The AI model analyzes cell spatial organization to make patient diagnoses and prognoses (Photo courtesy of UT Southwestern)

Patient tissue samples are commonly examined on slides by pathologists, a process integral to diagnosis. This traditional method, while effective, is notably time-intensive and subject to variability in interpretations among different pathologists. Moreover, some subtle details in pathology images might escape human observation but could hold critical insights into a patient's health status. Over recent years, several artificial intelligence (AI) models have been developed to undertake certain tasks typically performed by pathologists, such as classifying cell types or gauging cellular interactions based on proximity. Nevertheless, these models have not fully captured the more intricate aspects of tissue image analysis that pathologists conduct, including recognizing complex cell spatial arrangements and filtering out irrelevant image 'noise' that could distort interpretations. Addressing this gap, researchers have now introduced an innovative AI model that is capable of examining the spatial organization of cells within tissue samples, offering precise predictions about cancer patient outcomes and offering new prospects for AI-assisted cancer prognosis and tailored treatment plans.

Dubbed Ceograph, this AI tool, created by researchers at UT Southwestern Medical Center (Dallas, TX, USA), mimics the approach taken by pathologists for examining tissue slides. It begins by identifying cells and their respective positions within the image. It then classifies cell types and delineates their shapes and spatial distributions, creating a comprehensive map where the arrangement, dispersion, and interactions among cells are detailed for analysis. The team validated Ceograph in three clinical scenarios using pathology slides. In one instance, Ceograph distinguished between two lung cancer types: adenocarcinoma and squamous cell carcinoma. In another, it gauged the progression risk of potentially cancerous oral conditions to malignancy. Finally, it identified lung cancer patients most likely to benefit from epidermal growth factor receptor inhibitors.

In each scenario, Ceograph's performance in predicting patient outcomes surpassed traditional methods. Notably, the cell spatial organization insights provided by Ceograph are not only interpretable but also shed light on the biological implications of varying individual cell-cell spatial interactions. These developments highlight the increasingly vital role AI can play in healthcare, particularly in enhancing the precision and efficiency of pathology analyses. This technology promises to refine preventive strategies for individuals at high risk and tailor treatment choices to meet each patient's unique needs.

“Cell spatial organization is like a complex jigsaw puzzle where each cell serves as a unique piece, fitting together meticulously to form a cohesive tissue or organ structure,” said study leader Guanghua Xiao, Ph.D. “This research showcases the remarkable ability of AI to grasp these intricate spatial relationships among cells within tissues, extracting subtle information previously beyond human comprehension while predicting patient outcomes.”

Related Links:
UT Southwestern Medical Center 

Platinum Member
Flu SARS-CoV-2 Combo Test
OSOM® Flu SARS-CoV-2 Combo Test
Magnetic Bead Separation Modules
Anti-Cyclic Citrullinated Peptide Test
GPP-100 Anti-CCP Kit
Gold Member
Magnetic Bead Separation Modules

Print article


Clinical Chemistry

view channel
Image: PhD student and first author Tarek Eissa has analyzed thousands of molecular fingerprints (Photo courtesy of Thorsten Naeser / MPQ / Attoworld)

Screening Tool Detects Multiple Health Conditions from Single Blood Drop

Infrared spectroscopy, a method using infrared light to study the molecular composition of substances, has been a foundational tool in chemistry for decades, functioning similarly to a molecular fingerprinting... Read more

Molecular Diagnostics

view channel
Image: Protein ‘signatures’ obtained via a blood sample can be used to predict the onset of 67 diseases (Photo courtesy of Queen Mary University of London)

Protein Signatures in Blood Can Predict Risk of Developing More Than 60 Diseases

Measuring specific proteins to diagnose conditions like heart attacks, where troponin is tested, is a well-established clinical practice. Now, new research highlights the broader potential of protein measurements... Read more


view channel
Image: The Truvian diagnostic platform combines clinical chemistry, immunoassay and hematology testing in a single run (Photo courtesy of Truvian Health)

Automated Benchtop System to Bring Blood Testing To Anyone, Anywhere

Almost all medical decisions are dependent upon laboratory test results, which are essential for disease prevention and the management of chronic illnesses. However, routine blood testing remains limited worldwide.... Read more


view channel
Image: The Simplexa C. auris direct kit is a real-time polymerase chain reaction assay run on the LIAISON MDX instrument (Photo courtesy of Diasorin)

Novel Molecular Test to Help Prevent and Control Multi Drug-Resistant Fungal Pathogen in Healthcare Settings

Candida auris (C. auris) is a rapidly emerging multi drug-resistant fungal pathogen that is commonly found in healthcare environments, where it presents a challenge due to its ability to asymptomatically... Read more


view channel
Image: Beckman Coulter will utilize the ALZpath pTau217 antibody to detect key biomarker for Alzheimer\'s disease on its DxI 9000 immunoassay analyzer (Photo courtesy of Beckman Coulter)

Beckman Coulter Licenses Alzpath's Proprietary P-tau 217 Antibody to Develop Alzheimer's Blood Test

Cognitive assessments have traditionally been the primary method for diagnosing Alzheimer’s disease, but this approach has its limitations as symptoms become apparent only after significant brain changes... Read more
Copyright © 2000-2024 Globetech Media. All rights reserved.