Oncologists Presenting Successful Collaborations with Pangaea at Bio-IT World 2022
Leading US and UK oncologists will be presenting results from two independent studies through which they applied Pangaea’s novel AI to extract 26 features from EHRs with 97% accuracy in a privacy preserving manner and to find 6x more cancer patients with cachexia.
“Novel Unsupervised AI Extracts Intelligence from EHRs in a Privacy Preserving Manner with 97% Accuracy, to Create Research Quality Data” Presented by Dr. VK Gadi and Dr. Rachel Yung
“Breakthrough AI to Find 6x More Undiagnosed and Miscoded Cancer Patients with Cachexia at Scale” Presented by Dr. Judith Sayers
Visit us at booth 202.
Novel Unsupervised AI Extracts Intelligence from EHRs in a Privacy Preserving Manner with 97% Accuracy, to Create Research Quality Data
Date & time: 12:25 – 12:55PM ET, May 4th
Track: AI for Drug Discovery & Development
Abstract: Understanding the impact of precision medicine on medical practice, patient care and clinical outcomes is essential for advancing cancer care. However, extracting tumor genomic testing (TGT) from EHRs is challenging. This presentation will review a pilot study, conducted between leading US-based clinicians and Pangaea, to assess the ability for Natural Language Processing (NLP) algorithms to convert unstructured text data and PDF-formatted TGT results into research quality data. Results showed that Pangaea’s AI-driven product, PIES, was proven to extract 26 variables (for demographics, genomic testing results and social indicators), with an average accuracy of 97.3% (100% for 14 variables), to create research quality data.
Breakthrough AI to Find 6x More Undiagnosed and Miscoded Cancer Patients with Cachexia at Scale
Date & time: 3:10 – 3:40PM ET, May 4th
Track: AI for Oncology, Precision Medicine, and Health
Abstract: This presentation will demonstrate how Pangaea’s novel unsupervised AI has been found to discover clinical features characterizing cachexia in cancer patients, which helped with earlier detection of 6x more cancer patients with cachexia, who were undiagnosed, miscoded or at risk. These findings have the potential to reduce treatment costs by 50% and to save $1 billion annually, in the UK.
A Scalable Workflow to Build Machine Learning Classifiers with Clinician-in-the-Loop to Identify Patients in Specific Diseases
Automatic and Accurate Medical Regulatory Report Generation with Clinician-in-the-Loop
Automatic Clinical Trial Matching at Scale based on Patient Profiles
Predicting Length of Stay and Mortality Risk with Unstructured Clinical Notes in Intensive Care Units (ICU)