Cedars-sinai Medical Center

University of Southern California

Applied Invention, LLC

Genentech, Inc.

Applied Proteomics, Inc.

Agus, David B.

David B Agus

Rishi Raghav Rawat

Daniel Lee Ruderman

William Daniel Hillis

Paul Thomas Macklin

W Daniel Hillis

Daniel E H Afar

David H Mack

Fei Sha

Darryl Shibata

Histologic classification of pathology specimens through machine learning is a nascent field which offers tremendous potential to improve cancer medicine. Its utility has been limited, in part because of differences in tissue preparation and the relative paucity of well-annotated images. We introduce tissue recognition, an unsupervised learning problem analogous to human face recognition, in which the goal is to identify individual tumors using a learned set of histologic features. This feature set is the 'tissue fingerprint.' Because only specimen identities are matched to fingerprints, constructing an algorithm for producing them is a self-learning task that does not need image metadata annotations. Here, we provide an algorithm for self-learning tissue fingerprints, that, in conjunction with color normalization, can match hematoxylin and eosin stained tissues to one of 104 patients with 93% accuracy. We applied this identification network's internal representation as a tissue fingerprint for use in predicting the molecular status of an individual tumor (breast cancer clinical estrogen receptor (ER) status). We describe a fingerprint-based classifier that predicts ER status from whole-slides with high accuracy (AUROC=0.90), and is an improvement over traditional transfer learning approaches. The use of tissue fingerprinting for digital pathology as a concise but meaningful histopathologic image representation enables a new range of machine learning algorithms leading to increased information for clinical decision making in patient management.

Histopathology classification through machine self-learning of 'tissue fingerprints'

A method assessing tissue morphology using machine learning includes a step of training a machine learnable device to predict the status of a diagnostic feature in stained tissue samples. The machine learnable device is trained with a characterized set of digital images of stained tissue samples. Each digital image of the characterized set has a known status for the diagnostic feature and an extracted feature map provides values for a extracted feature over an associated 2-dimensional grid of spatial locations. A step of inputting the set of extracted feature maps is inputted into the machine learnable device to form associations therein between the set of extracted feature maps to and the known status for the diagnostic feature to form a trained machine learnable device. The status for the diagnostic feature of a stained tissue sample of unknown status for the diagnostic feature is predicted from the trained machine learnable device.

Machine learning for digital pathology

Described herein are methods and systems to measure dynamics of disease progression, including cancer growth and response, at multiple scales by multiple techniques on the same biologic system. Methods and systems according to the invention permit personalized virtual disease models. Moreover, the invention allows for the integration of previously unconnected data points into an in silico disease model, providing for the prediction of disease progression with and without therapeutic intervention.

Multi-scale complex systems transdisciplinary analysis of response to therapy

The present application describes the use of ErbB antagonist, especially ErbB2 antibodies such as rhuMAb 2C4, for treating pain.

ErbB antagonists for pain therapy

This invention relates to methods and kits for positive selection of species of interest based on peptide/protein sequence from a biological sample. The species of interest may be proteins and/or peptides of interest which may be placed through a mass spectrometer to obtain a blood peptide/protein signature. The blood peptide/protein signature may be used in proteomic analysis. The techniques include but are not limited to the use of collectors comprising nucleic acid molecules to extract a composition that has a lower concentration of a high abundance species of interest from a sample. This limits the level of influence that any collectors species may have on the results of a mass spectra.

Positive selection of serum proteins for proteomic analysis

The present invention relates to a system and method for mass spectrometry () that allows for bi-directional introduction of collections of charged particles into the magnetic field of a mass spectrometer. More particularly, the present invention includes a system for mass spectrometry () (e.g., an FTMS mass spectrometer) with a cylindrical magnet () configured to receive and measure the cyclotron frequencies () of charged particles that are introduced () into the cylindrical magnet () from either of the two axial ends thereof. Methods of the invention relate to performing mass spectrometry analysis on collections of charged particles that are introduced (), serially, simultaneously or both, into a cylindrical magnet () from opposing axial ends thereof. The present invention exhibits significantly increased magnet throughput relative to currently available devices, by allowing flow in the opposite direction to a second detector, e.g., during ion processing time of a first detector.

Bi-directional system for mass spectrometry

Described herein are methods for treating cancer and other disease conditions in individuals who have either developed a resistance to conventional tyrosine kinase inhibitor (TKI) therapy or who are non-responsive ab initio to conventional TKI therapy. In various embodiments, the methods include administering to a patient a resistance-surmounting quantity of a TKI on a weekly or semi-weekly basis. Alternate embodiments of the present invention include a diagnostic method for assessing an individual's probability of being resistant to TKI therapy, based upon an expression level of epithelial membrane protein-1 (EMP-1); one of the genes believed to be responsible for TKI resistance. The methods of the present invention may be particularly useful in the treatment of lung, breast, prostate, ovarian, brain and colon cancers. The methods of the present invention may be effective in blocking the HER-2 kinase domain either in addition to or in lieu of blocking the EGFR kinase domain.

Method of screening for sensitivity to kinase inhibitor therapy

Described herein are genes whose expression are up-regulated or down-regulated in prostate cancer. Also described are such genes whose expression is further up-regulated or down-regulated in drug-resistant prostate cancer cells. Related methods and compositions that can be used for diagnosis and treatment of prostate cancer are disclosed. Also described herein are methods that can be used to identify modulators of prostate cancer.

Methods of diagnosis of androgen-dependent prostate cancer, prostate cancer undergoing androgen withdrawal, and androgen-independent prostate cancer

Disclosed herein is a method for treating and preventing prostate cancer, and particularly androgen-independent prostate cancer, the method including administering to a mammal a benzothiopene having the formula or pharmaceutically acceptable salts or prodrugs thereof, wherein R and Rare each independently selected from the group consisting of hydrogen, &#x2014;COR, &#x2014;COR, and R; Ris selected from the group consisting of hydrogen, C1-C14 alkyl, C1-C3 chloroalkyl, C1-C3 fluoroalkyl, C5-C7 cycloalkyl, C1-C4 alkoxy, and phenyl; Ris phenyl with at least one substitution selected from the group consisting of C1-C4 alkyl, C1-C4 alkoxy, hydroxy, nitro, chloro, fluoro, trichloromethyl, and trifluoromethyl; Ris selected from the group consisting of C1-C4 alkyl, C5-C7 cycloalkyl, and benzyl; and Ris selected from the group consisting of oxygen and &#x2014;C&#x2550;O. The method may further include the administration of an estrogen lowering drug to enhance efficacy of the compound of the present invention.

Growing community of inventors

Beverly Hills, CA, United States of America

David B Agus