Granted Patent B2

US 12,362,038 · App. 16/739,667 · Granted Jul 15, 2025

Inventors: Rob Kimmerling (Cambridge, MA); Selim Olcum (Cambridge, MA); Clifford Reid (Pacifica, CA); Mark Stevens (Cambridge, MA)

Assignee: Travera, Inc.

G16B20/00G01N33/5011G01N33/5067G16H10/40G16H50/20G01N2800/52G01N2800/60G16H50/30

Abstract

Claims (37)

1. A method for selecting a cancer treatment, the method comprising:

training a machine learning system on a training data set, wherein the training data set includes patient information, therapeutic information, cancer features, and outcomes for patient, therapeutic, and cancer combinations;

operating the machine learning system to identify correlations between therapeutics, patients, cancers, and outcomes;

identifying, in silico, a list of therapeutics that meet predetermined criteria relating to one or more selected from the group of toxicology, efficacy, pharmacokinetics, side-effects, therapeutic interactions, patient compliance, and cost;

obtaining a biological sample containing cancer cells from a subject;

contacting the cancer cells with the identified therapeutics and measuring mass accumulation of the contacted cancer cells using at least one suspended microchannel resonator;

analyzing the contacted cancer cells after measuring mass accumulation by extracting nucleic acid or protein from the contacted cancer cells and performing a biomarker assay on the extracted nucleic acid or protein to identify a cancer biomarker;

selecting one of the identified therapeutics, based on the measured mass accumulation of the contacted cancer cells, for treating cancer in the subject;

feeding, into the machine learning system, the patient information of the subject, selected therapeutic, cancer features of the subject, and a clinical outcome for the subject; and

operating the machine learning system to update training of the machine learning system and update the identified correlations.

2. The method of claim 1 , wherein the biological sample is from a patient having received a prior therapy for cancer.

3. The method of claim 2 , wherein the predetermined criteria further comprises the prior therapy the patient has received.

4. The method of claim 1 , wherein the microchannel resonator includes a channel extending through a cantilever positioned between a light source and a photodetector, wherein the light source comprises a super-luminescent diode.

5. The method of claim 1 , further comprising isolating individual, live cells, from the biological sample before measuring the mass accumulation and performing the in vitro assay on the individual, live cells.

6. The method of claim 5 , wherein the biological sample comprises a tissue sample.

7. The method of claim 6 , wherein the tissue sample is a biopsy sample.

8. The method of claim 7 , where the biopsy sample is obtained by fine needle biopsy and comprises less than 50,000 cancer cells.

9. The method of claim 1 , wherein measuring the mass accumulation is performed within about 48 hours after obtaining the biological sample.

10. The method of claim 1 , wherein measuring the mass accumulation is performed within about 24 hours after obtaining the biological sample.

11. The method of claim 1 , wherein measuring the mass accumulation is performed within about 6 hours after obtaining the biological sample.

12. The method of claim 1 , wherein the suspended microchannel resonator further comprises a photodetector which is connected to a processing chip and wherein the photodetector converts an optical signal from the light source to a voltage signal sent the processing chip.

13. The method of claim 12 , wherein the suspended microchannel resonator is provided in one of an array of cantilevers and wherein the processing chip implements a plurality of parallel phase-locked loops that each both demodulate and drive one respective cantilever of the array of cantilevers.

14. The method of claim 13 , wherein the processing chip comprises a field-programmable gate array (FPGA).

15. The method of claim 13 , wherein the array comprises at least one piezo-ceramic actuator used for actuation of the cantilevers.

16. A method for selecting a cancer treatment, the method comprising:

training a machine learning system on a training data set, wherein the training data set includes patient information, therapeutic information, features of cancer cells, and outcomes for patient, therapeutic, and cancer combinations;

operating the machine learning system to identify correlations between therapeutics, patients, cancers, and outcomes;

identifying, in silico, a list of therapeutics that meet predetermined criteria relating to one or more selected from the group of toxicology, efficacy, pharmacokinetics, side-effects, therapeutic interactions, patient compliance, and cost;

obtaining a biological sample containing cancer cells from a subject;

contacting the cancer cells with the identified therapeutics and measuring mass accumulation of the contacted cancer cells using at least one suspended microchannel resonator;

selecting a first therapeutic from the identified therapeutics, based on the measured mass accumulation of the contacted cancer cells, for treating cancer in the subject;

feeding, into the machine learning system, patient information of the subject, therapeutic information of the subject, cancer features of the subject, a clinical outcome of the subject for the first therapeutic, and mass accumulation measurements from a second therapeutic from the identified therapeutics; and

operating the machine learning system to update training of the machine learning system and update the identified correlations.

17. A method for selecting a cancer treatment, the method comprising:

contacting cancer cells from a first patient with a panel of therapeutics and obtaining mass accumulation measurements of the cancer cells using a suspended microchannel resonator;

training a machine learning system with a data set that includes first patient information, features of the cancer cells from the first patient, and therapeutic information of the first patient as inputs, and the mass accumulation measurements and patient response data of the first patient as outcomes; and

operating the machine learning system to identify a therapeutic useful for treating cancer in a second patient based on second patient information and features of cancer cells from the second patient.

Assignments (2)

Continuity (2)

References Cited (272)