How to Diagnose Cancer Better Using Multiple Blood Markers and Patient Data
This patent describes a method for improving cancer diagnosis by combining measurements of several biological markers with patient-specific information like age or body mass, transforming them into a special score, and then comparing that score to a model of known cancer cases.
Patent Number
US 20230274838
Status
Active
Filing Date
May 5, 2023
Grant Date
—
Expiration
May 5, 2043
Claims
22
Assignee
OTRACES
Inventors
Keith LINGENFELTER, Mohsen Marefat, Galina KRASIK
Citations
0 forward · 1 backward
What it covers
This method aims to improve cancer diagnosis and treatment by analyzing a subject's biological samples. First, it receives concentrations of at least three specific biological markers, called analytes, from a patient's sample (Claim 1a). Then, it selects additional patient details, known as meta-variables, such as age, body mass, or menopausal status, which are known to vary in the general population (Claim 1b, Claim 2). These analyte concentrations are then transformed, using both population data and the selected meta-variables, into a special unitless number called a "pseudo-concentration" for each analyte (Claim 1c). These pseudo-concentrations are scored against a pre-built model of pseudo-concentrations from people known to have or not have the disease (Claim 1d). Finally, based on this score, the method diagnoses whether the subject has cancer, provided the diagnostic model performs with at least 75% accuracy in avoiding false positives and negatives, and then suggests administering anti-cancer drugs (Claim 1e). For example, a doctor could use this system to get a more accurate cancer diagnosis for a patient by inputting their blood test results and their age, and the system would output a diagnosis and treatment recommendation.
What it doesn't cover
- —Does not cover diagnostic methods that use fewer than three predetermined analytes.
- —Does not cover diagnostic methods that do not incorporate meta-variables specific to the subject's population, like age or body mass.
- —Does not cover diagnostic methods where the raw analyte concentrations are not transformed into unitless, clinically predictive pseudo-concentrations.
- —Does not cover diagnostic models with a false negative and false positive performance worse than 75%.
- —Does not cover diagnostic methods that only provide a diagnosis without also including the administration of anti-cancer drugs.
The clever bit
The truly clever part is transforming raw analyte concentrations into "pseudo-concentrations" by factoring in patient-specific meta-variables and population distribution characteristics. This normalizes the data, making the diagnostic score more robust and less prone to variations caused by individual differences, thus improving diagnostic accuracy.
Why it matters
Accurate and early cancer diagnosis is critical for effective treatment. This patent matters because it proposes a way to make diagnostic tests more precise by accounting for individual patient characteristics and population differences, rather than relying solely on raw biomarker levels. By integrating meta-variables, it aims to reduce misdiagnoses, which can lead to better patient outcomes and more targeted treatment plans. This approach could help personalize medicine by making diagnostic results more relevant to each unique patient.
Real-world examples
- 1.Personalized cancer diagnostic panels
- 2.AI-powered medical diagnostic systems
- 3.Advanced blood tests for early disease detection
- 4.Precision oncology treatment planning tools
Generated by PatentBrief · Not legal advice · patentbrief.org
US 20230274838 · 2026