New AI Algorithm for Use on Eight Types of Cancer Trained on Seven Scanner Models
|
By LabMedica International staff writers Posted on 04 Oct 2023 |
|
A new artificial intelligence (AI) algorithm for use on one of the broadest trained environments is now available for commercial or research work.
Gestalt Diagnostics (Spokane, WA, USA) has released the Mitotic Counting algorithm which has been trained on more than 100,000 individual mitosis and across seven scanner models. The algorithm performs impressively with an F1-Score of 0.74, matching the performance of top-of-the-line methods in the field. The algorithm is currently available in the Gestalt AI Studio for use on a variety of cancers including cutaneous mast cell tumors, breast cancer, lymphoma, lung cancer, melanoma, neuroendocrine, colon cancer, and bladder carcinoma. The algorithm has been trained on several scanner models, including those from Hamamatsu, Leica, 3D Histech, and Aperio.
Key to the broad applicability of deep learning algorithms is their ability to adapt to new scanners and different types of tissue; a broad dataset is crucial to achieve this goal. Gestalt Diagnostics' dataset encompasses seven types of tissues spanning eight tumors and is digitized with seven different scanner models. Additionally, Gestalt Diagnostics also offers PathFlow, a smart and customizable workflow solution that offers genuine interoperability for optimal efficiency. This solution supports both digital and glass slides and allows for tailored rules that can be can be configured to meet specific needs and requirements.
"What does this mean for our customers? Confidence,” said Lisa-Jean Clifford, Gestalt COO & Chief Strategy Officer. “Gestalt is focused on providing solutions that provide a foundation of confidence in use as this industry continues to evolve, innovate, and incorporate products that are meant to enhance a pathologist's or scientist's daily life."
"We are extremely proud of the comprehensive approach we take to AI development as an organization,” added Brian Napora, Gestalt VP, AI Solutions. “Our position is to not only look at development and training with a small subset of images or on one or two scanners, that would not prove as effective or accurate as we feel our customers and the industry require. Therefore, we train on tens of thousands images, tumors cells, or in this case, mitosis and multiple scanner models across several scanner vendors to ensure the most accurate, clinically usable options for our customers."
Related Links:
Gestalt Diagnostics
Latest Pathology News
- AI Histopathology Tool Distinguishes Ewing Sarcoma for Diagnosis
- Rapid Mass Spectrometry Test May Aid Glioma Margin Decisions
- Genomic Classifier Predicts Benefit From Adding Hormone Therapy to Salvage Prostate Radiation
- Multimodal AI Supports Prostate Cancer Risk Assessment Across Diverse Cohorts
- New AI Model Maps Where Glioblastoma Could Return After Surgery
- Simple Test Could Reduce the Need for Invasive Uterine Cancer Checks
- Tumor Blood Vessel Features May Help Predict Colorectal Cancer Survival
- Multimodal AI Improves Breast Cancer Recurrence Risk Prediction Beyond Standard Genomic Testing
- Machine Learning Cytology Tool Improves Cancer Cell Identification
- CAP Updates Cancer Protocols to Refine Breast Cancer Biomarker Reporting and GIST Molecular Testing
- Tumor Budding Grading May Predict Chemotherapy Benefit in Resected Lung Squamous Cancer
- AI Spatial Analysis of Routine Slides Helps Predict Pancreatic Cancer Recurrence
- Hybrid Computational Imaging Method Improves Digital Pathology Resolution
- FDA Clearance Expands Digital Pathology Options for Clinical Laboratories
- Label-Free Imaging Reveals Hidden Immune Activity in Blood Samples
- AI Tool Improves Prediction of Immunotherapy Response in Lung Cancer
Channels
Clinical Chemistry
view channel
Blood Biomarker Detects Alzheimer’s Changes Decades Before Symptoms in Down Syndrome
Alzheimer’s disease can begin altering the brain long before clinical symptoms appear, creating a challenge for early-stage detection and research. People with Down syndrome face a particularly high age-related... Read more
Study Questions Broad Fasting Requirements Before Routine Blood Tests
Routine blood testing is central to diagnosis, monitoring, and disease risk assessment. Many patients are still asked to fast for 8–12 hours before phlebotomy, even though the requirement was originally... Read more
Elevated Lipoprotein(a) Linked to Long-Term Progression of Carotid Atherosclerosis
Lipoprotein(a), or Lp(a), is a genetically determined blood lipid and an important risk factor for cardiovascular disease. Its relationship to long-term carotid atherosclerosis has remained unclear, particularly... Read more
Molecular Diagnostics
view channel
Genomic Screening Expands Detection of Treatable Conditions in Newborns
Conventional newborn screening can miss conditions that lack biochemical biomarkers or present atypically. Initial hearing screens may also fail to detect hearing loss that is later identified through... Read more
Age-Based Genetic Testing May Miss Most Inherited Cancer Risk Variants
Inherited cancer risk can influence diagnosis, treatment planning, and family screening, yet current testing practices often depend on a patient’s age at diagnosis. Many patients undergo germline genetic... Read more
New NGS Suite Expands Clonality and MRD Research in Lymphoid Malignancies
Lymphoid malignancies require molecular tools that can distinguish clonal immune receptor rearrangements and track disease-related sequences over time. As immune receptor measurable residual disease analysis... Read more
Tumor-Informed ctDNA Test Shows Utility Across Gastrointestinal and Gynecologic Cancers
Natera announced new data on Signatera, its tumor-informed circulating tumor DNA test, that were presented in four oral presentations at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting.... Read moreHematology
view channel
New Donor Genetic Marker May Help Predict Stem Cell Transplant Success
Donor selection for hematopoietic stem cell transplantation plays a major role in relapse risk and survival for patients with blood cancers and other blood disorders. Despite advances in genotyping, uncertainty... Read more
Updated Ferritin Thresholds Improve Detection of Iron Deficiency
Iron deficiency is one of the most common health conditions worldwide, yet its nonspecific symptoms can delay diagnosis for months. Variation in testing practices and ferritin thresholds may contribute... Read more
Immunology
view channel
Blood-Based Immune Signal Could Reveal Lung Cancer Risk Before Tumors Develop
Lung cancer remains difficult to detect before symptoms appear, when advanced disease is harder to treat. In the U.K., about 32,800 people die from lung cancer each year, and only 11.1% of those diagnosed... Read more
Prototype Point-of-Care Test Aims to Improve Neonatal Sepsis Diagnosis
Neonatal sepsis remains difficult to diagnose because early signs can resemble other newborn conditions. Diagnostic delays may increase the risk of complications, death, long-term adverse outcomes, and... Read more
Microbiology
view channel
Rapid Urine Test Aids Diagnosis of Invasive Aspergillosis
Invasive aspergillosis is an uncommon mold infection in the general population but can pose serious risks for people with weakened immune defenses. Diagnosis can be difficult because existing approaches... Read more
Nanodroplet CRISPR Technology Supports Rapid, Multiplexed Mycobacterial Identification
Mycobacterial infections are difficult to diagnose because closely related species can have different clinical and therapeutic implications. Nontuberculous mycobacteria (NTM) are increasingly recognized... Read more
Genomic Workflow Identifies Fungal Pathogens Before Blood Cultures Turn Positive
Fungal bloodstream infections pose a major threat to hospitalized patients. Candida species cause most invasive fungal infections worldwide and are among the leading causes of hospital-acquired bloodstream... Read more
FDA-Cleared Multiplex PCR Test Detects 13 Respiratory Pathogens in a Single Sample
Respiratory tract infections can be difficult to distinguish at presentation because many cause overlapping, nonspecific symptoms and are initially grouped as influenza-like illnesses. Causes span a range... Read moreTechnology
view channel
Laser-Based Swab Analysis Shows Promise for Detecting Disease-Linked Odor Patterns
Disease-related changes in volatile organic compounds can alter body odor, producing measurable patterns in exhaled breath and bodily fluids. Current analytical methods can be complex, time-consuming,... Read more
Laser-Enhanced Assay Boosts Sensitivity for Colorectal Cancer Biomarker Detection
Colorectal cancer is the third most commonly diagnosed cancer and the second leading cause of cancer-related death worldwide. Early detection remains critical, but cancer biomarkers can produce only faint... Read more
Industry
view channel
Regulatory Milestone Expands Access to Blood-Based Neurology Biomarker Testing in China
Quanterix Corporation (Billerica, MA, USA) and Innovita Biological Technology Co., Ltd. (Beijing, China) announced regulatory approvals that expand access to Quanterix SIMOA technology and neurology biomarker... Read more
Collaboration Advances Automated Benchtop Platform for Routine Blood Testing
Routine blood testing is central to clinical decision-making, but access can vary across laboratory and healthcare settings. Broader use of automated benchtop platforms may help integrate testing more... Read more
Expanded Partnership Supports AI Biomarker Validation and Clinical Trial Deployment
CellCarta (Montreal, Canada) and Imagene AI (Miami, FL, USA) have expanded their collaboration to provide biopharma companies with a coordinated pathway for biomarker strategy, assay development, validation,... Read more