The Outpost is a comprehensive collection of curated artificial intelligence software tools that cater to the needs of small business owners, bloggers, artists, musicians, entrepreneurs, marketers, writers, and researchers.
© 2025 TheOutpost.AI All rights reserved
Curated by THEOUTPOST
On Fri, 25 Oct, 8:02 AM UTC
3 Sources
[1]
New AI tool revolutionizes detection of rare gastrointestinal diseases
Ludwig-Maximilians-Universitaet Muenchen (LMU)Oct 25 2024 Researchers from LMU, TU Berlin, and Charité have developed a new AI tool that uses imaging data to also detect less frequent diseases of the gastrointestinal tract. Already used in many areas of medicine, AI has tremendous potential when it comes to helping doctors diagnose diseases with the help of imaging data. However, AI models have to be trained with large numbers of examples, which are generally available in sufficient quantities only for common diseases. It's as if a family doctor only had to diagnose coughs, runny noses, and sore throats. The actual challenge is to also detect the less common diseases, which current AI models often overlook or misclassify." Professor Frederick Klauschen, Director of the Institute of Pathology at LMU Together with the group of Professor Klaus-Robert Müller from TU Berlin/BIFOLD and colleagues from the Charité - Universitätsmedizin Berlin, Klauschen has developed a novel approach that overcomes this limitation: As the scientists report in the journal New England Journal of Medicine AI (NEJM AI), their new model only needs training data from common findings to also reliably detect the less frequent diseases. This could significantly improve the diagnostic accuracy and ease the workloads of pathologists in future. Learning from normality The new approach is based on anomaly detection: From the very precise characterization of normal tissue and findings from frequent diseases, the model learns to recognize and flag deviations, without having to be specifically trained for these rarer cases. For their study, the researchers collected two large datasets of microscopic images of tissue sections from gastrointestinal biopsies with the corresponding diagnoses. In these datasets, the ten most common findings - including normal findings and common diseases such as chronic gastritis - account for around 90 percent of cases, whereas the remaining 10 percent contained 56 disease entities - including many cancers. For the training and evaluation of their model, the researchers used a total of 17 million histological images from 5,423 cases. "We compared various technical approaches and our best model detected with a high degree of reliability a broad range of rarer pathologies of the stomach and colon, including rare primary or metastasizing cancers. To our knowledge, no other published AI tool is capable of doing this," says Müller. Using heatmaps, moreover, the AI can indicate in color the position of anomalies in the tissue section. Significantly easing the diagnosis workload By identifying normal findings and frequent diseases and detecting anomalies, the new AI model, which will be further improved over time, could provide critical support to doctors. Although the identified diseases still need to be confirmed by pathologists, "doctors can save a lot of time, because normal findings and a certain proportion of the diseases can be automatically diagnosed by AI. This applies to around a quarter to a third of cases," says Klauschen. "And in the remaining cases, AI can facilitate case prioritization and reduce missed diagnoses. This would represent huge progress." Ludwig-Maximilians-Universitaet Muenchen (LMU)
[2]
AI in medicine: New approach for more efficient diagnostics
Researchers from LMU, TU Berlin, and Charité have developed a new AI tool that uses imaging data to also detect less frequent diseases of the gastrointestinal tract. Already used in many areas of medicine, AI has tremendous potential when it comes to helping doctors diagnose diseases with the help of imaging data. However, AI models have to be trained with large numbers of examples, which are generally available in sufficient quantities only for common diseases. "It's as if a family doctor only had to diagnose coughs, runny noses, and sore throats," says Professor Frederick Klauschen, Director of the Institute of Pathology at LMU. "The actual challenge is to also detect the less common diseases, which current AI models often overlook or misclassify." Together with the group of Professor Klaus-Robert Müller from TU Berlin/BIFOLD and colleagues from the Charité -- Universitätsmedizin Berlin, Klauschen has developed a novel approach that overcomes this limitation: As the scientists report in the journal New England Journal of Medicine AI (NEJM AI), their new model only needs training data from common findings to also reliably detect the less frequent diseases. This could significantly improve the diagnostic accuracy and ease the workloads of pathologists in future. Learning from normality The new approach is based on anomaly detection: From the very precise characterization of normal tissue and findings from frequent diseases, the model learns to recognize and flag deviations, without having to be specifically trained for these rarer cases. For their study, the researchers collected two large datasets of microscopic images of tissue sections from gastrointestinal biopsies with the corresponding diagnoses. In these datasets, the ten most common findings -- including normal findings and common diseases such as chronic gastritis -- account for around 90 percent of cases, whereas the remaining 10 percent contained 56 disease entities -- including many cancers. For the training and evaluation of their model, the researchers used a total of 17 million histological images from 5,423 cases. "We compared various technical approaches and our best model detected with a high degree of reliability a broad range of rarer pathologies of the stomach and colon, including rare primary or metastasizing cancers. To our knowledge, no other published AI tool is capable of doing this," says Müller. Using heatmaps, moreover, the AI can indicate in color the position of anomalies in the tissue section. Significantly easing the diagnosis workload By identifying normal findings and frequent diseases and detecting anomalies, the new AI model, which will be further improved over time, could provide critical support to doctors. Although the identified diseases still need to be confirmed by pathologists, "doctors can save a lot of time, because normal findings and a certain proportion of the diseases can be automatically diagnosed by AI. This applies to around a quarter to a third of cases," says Klauschen. "And in the remaining cases, AI can facilitate case prioritization and reduce missed diagnoses. This would represent huge progress."
[3]
AI-based anomaly detection offers more efficient clinical-grade histopathological diagnostics
Already used in many areas of medicine, AI has tremendous potential when it comes to helping doctors diagnose diseases with the help of imaging data. However, AI models have to be trained with large numbers of examples, which are generally available in sufficient quantities only for common diseases. "It's as if a family doctor only had to diagnose coughs, runny noses, and sore throats," says Professor Frederick Klauschen, Director of the Institute of Pathology at LMU. "The actual challenge is to also detect the less common diseases, which current AI models often overlook or misclassify." Together with the group of Professor Klaus-Robert Müller from TU Berlin/BIFOLD and colleagues from the Charité -- Universitätsmedizin Berlin, Klauschen has developed a novel approach that overcomes this limitation. As the scientists report in the journal NEJM AI, their new model only needs training data from common findings to also reliably detect the less frequent diseases. This could significantly improve the diagnostic accuracy and ease the workloads of pathologists in future. Learning from normality The new approach is based on anomaly detection: From the very precise characterization of normal tissue and findings from frequent diseases, the model learns to recognize and flag deviations, without having to be specifically trained for these rarer cases. For their study, the researchers collected two large datasets of microscopic images of tissue sections from gastrointestinal biopsies with the corresponding diagnoses. In these datasets, the 10 most common findings -- including normal findings and common diseases such as chronic gastritis -- account for about 90% of cases, whereas the remaining 10% contained 56 disease entities -- including many cancers. For the training and evaluation of their model, the researchers used a total of 17 million histological images from 5,423 cases. "We compared various technical approaches and our best model detected with a high degree of reliability a broad range of rarer pathologies of the stomach and colon, including rare primary or metastasizing cancers. To our knowledge, no other published AI tool is capable of doing this," says Müller. Using heatmaps, moreover, the AI can indicate in color the position of anomalies in the tissue section. Significantly easing the diagnosis workload By identifying normal findings and frequent diseases and detecting anomalies, the new AI model, which will be further improved over time, could provide critical support to doctors. Although the identified diseases still need to be confirmed by pathologists, "doctors can save a lot of time, because normal findings and a certain proportion of the diseases can be automatically diagnosed by AI. This applies to around a quarter to a third of cases," says Klauschen. "And in the remaining cases, AI can facilitate case prioritization and reduce missed diagnoses. This would represent huge progress."
Share
Share
Copy Link
Researchers from LMU, TU Berlin, and Charité have developed a novel AI tool that can detect rare gastrointestinal diseases using imaging data, potentially improving diagnostic accuracy and easing pathologists' workloads.
Researchers from Ludwig-Maximilians-Universitaet Muenchen (LMU), Technical University of Berlin (TU Berlin), and Charité - Universitätsmedizin Berlin have developed a groundbreaking AI tool that promises to revolutionize the detection of rare gastrointestinal diseases. This innovative approach addresses a significant limitation in current AI diagnostic models, which typically excel at identifying common conditions but struggle with less frequent diseases [1][2][3].
Professor Frederick Klauschen, Director of the Institute of Pathology at LMU, explains the core issue: "It's as if a family doctor only had to diagnose coughs, runny noses, and sore throats. The actual challenge is to also detect the less common diseases, which current AI models often overlook or misclassify" [1]. This limitation stems from the fact that AI models require large datasets for training, which are often only available for common diseases.
The new AI tool employs an innovative approach based on anomaly detection. Instead of requiring extensive training data for rare conditions, the model learns to recognize deviations from normal tissue and common findings. This allows it to flag potential rare diseases without specific training for each condition [2].
For their study, the researchers utilized:
The model demonstrated a high degree of reliability in detecting a broad range of rare pathologies of the stomach and colon, including rare primary or metastasizing cancers. Professor Klaus-Robert Müller from TU Berlin/BIFOLD states, "To our knowledge, no other published AI tool is capable of doing this" [3].
The AI tool goes beyond mere detection by utilizing heatmaps to visually indicate the position of anomalies in tissue sections. This feature enhances the tool's utility for medical professionals [1][2][3].
By automating the diagnosis of normal findings and common diseases, which account for about a quarter to a third of cases, the AI model could significantly reduce pathologists' workloads. For the remaining cases, it aids in prioritization and helps reduce missed diagnoses [2].
While pathologists will still need to confirm the AI's findings, this tool represents a significant step forward in medical diagnostics. As Professor Klauschen notes, "This would represent huge progress" in the field of pathology and disease detection [3]. The model's ability to learn from common cases to identify rare conditions could pave the way for more efficient and accurate diagnostic processes in various medical fields.
Reference
[1]
[2]
[3]
Medical Xpress - Medical and Health News
|AI-based anomaly detection offers more efficient clinical-grade histopathological diagnostics
Researchers at Washington State University have developed a deep learning AI model that can identify pathologies in animal and human tissue images faster and more accurately than human experts, potentially revolutionizing disease research and medical diagnostics.
4 Sources
Researchers develop a generative AI model that creates virtual colorations of cancer tissue, enhancing diagnostic capabilities in clinical settings. This breakthrough could lead to more accurate and efficient cancer detection and treatment.
2 Sources
Researchers at the University of Washington have developed BiomedParse, an AI model capable of analyzing nine different types of medical images to predict systemic diseases, potentially revolutionizing medical diagnostics.
2 Sources
A revolutionary AI model, similar to ChatGPT, demonstrates potential in detecting multiple types of cancer and improving treatment decisions. This advancement could significantly impact cancer care and patient outcomes.
5 Sources
A new international study shows that AI-based models can surpass human experts in identifying ovarian cancer from ultrasound images, potentially improving diagnosis accuracy and reducing unnecessary referrals.
4 Sources
