Predicting personalized cancer metastasis routes, biological mediators of metastasis and metastasis blocking therapies

Abstract

Predicting the metastasis of cancer in a patient from one tissue to another is disclosed. A computer-implemented method for predicting metastasis may comprise receiving an indication of at least one disrupted gene of the cancer; searching for data representing a gene-to-gene or protein-to-protein interaction network to determine the position of a received gene, wherein the data representing a gene-to-gene or protein-to-protein interaction network includes data that expresses the gene or the protein as a node of the network and expresses functional or physical interactions among the genes or proteins as an edge of the network; traversing the data representing a gene-to-gene or protein-to-protein interaction network specific for a type of the cancer type from a position of the received genein the network to a position of at least one gene involved in metastasis for a tissue type, organ or body part; determining at least one shortest path in the network between the received gene and theat least one gene involved in metastasis for the tissue type, organ or body part; generating a prediction of metastasis to the tissue type based on the at least one determined path; and generating anoutput display indicating a likelihood of spread of cancer to the tissue type, organ or body part.

Description

Background technique [0001] The present invention relates to techniques for predicting the spread (metastasis) of a patient's cancer from one tissue to another. [0002] Many methods for predicting the spread of cancer in a patient provide a prognostic prediction, such as whether the cancer is likely to spread to certain other tissues and increase the patient's risk of death or expected survival. However, conventional methods cannot predict whether a cancer will spread to a particular tissue or organ. Such conventional approaches may rely on correlations (comorbidity of cancers), cancers that tend to co-occur in patients based on medical records are considered more likely to spread in the same way. [0003] However, due to a lack of understanding of the molecular basis of metastasis, such conventional methods for predicting cancer prognosis or survival often do not provide sufficient information to prevent cancer from spreading to other tissues. Likewise, existing methods as...

AI Technical Summary

Problems solved by technology

[0003] However, due to a lack of understanding of the molecular basis of metastasis, such conventional methods for predicting cancer prognosis or survival often do not provide sufficient information to prevent cancer from spreading to other tissues

Likewise, existing methods assume that metastasis from one tissue to another is the same from patient to patient

Furthermore, existing methods, as well as those in development, may require assaying many genes to predict prognosis, which is expensive and requires significant effort and expense in the clinical validation of new diagnoses

Images