Tempus Labs has been granted a patent for a system and method that analyzes de-identified patient data to generate dynamic user interfaces for predicting treatment responses. The system uses automated analysis of patterns in clinical, molecular, phenotypic, and response data to help clinicians evaluate large datasets and potentially discover therapeutic insights. The method involves receiving user selections of datasets, generating notebook interfaces, and allowing users to modify and analyze the data using preconfigured elements. GlobalData’s report on Tempus Labs gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on Tempus Labs, Personalized treatment planning was a key innovation area identified from patents. Tempus Labs's grant share as of September 2023 was 22%. Grant share is based on the ratio of number of grants to total number of patents.
Method for disease progression analysis using interactive notebooks
A recently granted patent (Publication Number: US11769572B2) describes a method for disease progression analysis using an interactive portal displayed on a computer. The method involves receiving a user selection of a dataset representing a cohort of patients, each diagnosed with one or more cancers. The dataset includes clinical information and biomarker information related to tumor specimens. The method further involves identifying pre-defined notebooks, each containing templates for implementing specific types of disease progression analysis. These notebooks can be displayed in the interactive portal as cells filled with preconfigured elements such as analytics, models, visualizations, or reportings.
The method allows users to generate new notebooks and select pre-defined notebooks from the interface. Users can drag and drop preconfigured elements into cells of the notebook, populating them with code for generating associated analytics, models, visualizations, or reportings. The notebooks can be modified by the user, including modifying code, inputs for statistical models or analyses, and generating draggable user interface elements.
The patent also describes the ability to model or analyze subsets of the dataset using statistical models or analyses selected by the user. Visualizations and reportings can be applied to the modeled or analyzed dataset, providing results to the user through the interactive portal. The visualizations and reportings include representations of each cell of the notebook corresponding to the user-selected notebook user interface element.
The method can be used for various disease progression analyses, including metastases events, recurrence, progression measures for drugs, tumor size, and enriched measurements. It can also reflect different lines of therapy or clinical trials. The statistical models or analyses can be previously trained machine learning models, machine learning models trained on the fly, or unsupervised machine learning models. These models can include linear regression models, non-linear regression models, logistic regression models, classification models, bootstrap resampling models, subset selection models, dimensionality reduction models, or tree-based models.
The patent also mentions the ability to generate visualizations or reportings based on user selections, including templates for text, images, or graphs. The notebooks, analysis results, or datasets can be dynamically shared with other users or groups of users provisioned with access. Provisioning can be based on the existence of an account or access to a linked dataset. The dataset can be obtained from a subscribed database or purchased for the user.
Overall, this patent describes a method and system for disease progression analysis that provides a user-friendly interface for creating, modifying, and analyzing notebooks containing preconfigured elements and customizable code.