Method and process for predicting and analyzing patient cohort response, progression, and survival

Patent metadata

Field	Value
Publication number	US-11037685-B2
Application number	US-202017139929-A
Country	US
Kind code	B2
Filing date	Dec 31, 2020
Priority date	Dec 31, 2018
Publication date	Jun 15, 2021
Grant date	Jun 15, 2021

Abstract

A system and method for analyzing a data store of de-identified patient data to generate one or more dynamic user interfaces usable to predict an expected response of a particular patient population or cohort when provided with a certain treatment. The automated analysis of patterns occurring in patient clinical, molecular, phenotypic, and response data, as facilitated by the various user interfaces, provides an efficient, intuitive way for clinicians to evaluate large data sets to aid in the potential discovery of insights of therapeutic significance.

First claim

What is claimed is:

1. A method for identifying an outlier group of patients within a cohort of patients, comprising:
a) generating the cohort of patients by selecting a plurality of clinical and molecular characteristics from patient data, wherein patients included in the cohort of patients satisfy the selection of the plurality of clinical and molecular characteristics, wherein each patient in the cohort of patients has been diagnosed with cancer, wherein the molecular characteristics are characteristics of a respective cancer, and wherein the clinical characteristics are characteristics of a respective patient;
b) generating a plurality of analytical characteristics from the clinical and molecular characteristics associated with the patient data from each patient within the cohort of patients;
c) calculating an associated health measurement based at least in part on a deviation of a health measurement between the plurality of analytical characteristics and each other analytical characteristic of the plurality of analytical characteristics, comprising, for each analytical characteristic:
c.i) dividing the cohort of patients into a first subgroup satisfying a threshold of the analytical characteristic and a second subgroup not satisfying the threshold of the analytical characteristic;
c.ii) determining the deviation in the health measurement between the first subgroup and the second subgroup;
c.iii) storing the analytical characteristic of the plurality of analytical characteristics having the largest deviation in the health measurement as a characteristic of the outlier group of patients;
c.iv) removing the stored analytical characteristic from the plurality of analytical characteristics;
c.v) identifying the outlier group of patients as patients of the cohort of patients satisfying each stored characteristic;
c.vi) identifying an associated health measurement of the outlier group of patients; and
c.vii) repeating steps c.i) through c.vi) until either:
a maximum number of analytical characteristics have been removed from the plurality of analytical characteristics, or
a minimum number of patients are identified within the outlier group of patients; and

d) storing the identified outlier group and the associated health measurement.

2. The method of claim 1 , further comprising:
identifying a plurality of alternate outlier groups of patients having a first outlier group and a second outlier group by repeating steps c.i) through c.vii) from each identified outlier group of c.v), wherein the first outlier group comprises a respective cohort of patients satisfying the respective stored characteristic and the second outlier group comprises a respective cohort of patients not satisfying the respective stored characteristic; and
generating an interactive user interface visually depicting each first outlier group and second outlier group of the plurality of alternate outlier groups.

3. The method of claim 2 , wherein the visual depiction of each alternate outlier group of the plurality of alternate outlier groups is presented in a first region of the user interface, and wherein the user interface includes a second region including a control panel for modifying the presentation of the alternate outlier groups of the plurality of alternate groups in the first region.

4. The method of claim 2 , further comprising:
receiving a user selection of an alternate outlier group of the plurality of alternate groups; and
generating a user interface object presenting specific information regarding the subgroup represented by the selected alternate outlier group of the plurality of alternate groups.

5. The method of claim 4 , wherein the user interface includes a central node reflecting a health measurement of the cohort of patients.

6. The method of claim 4 , wherein the user interface object also presents comparative information with regard to a second, larger cohort of patients.

7. The method of claim 4 , wherein the specific information includes a comparison of one or more of the characteristics attributable to the selected alternate outlier group of the plurality of alternate groups as compared to values of the one or more characteristics for the cohort of patients.

8. The method of claim 1 , wherein the health measurement is selected from a measurement of progression free survival, a measurement of observed survival, a measurement of an outcome, or a measurement of an adverse reaction.

9. The method of claim 1 , wherein the plurality of analytical characteristics comprise the clinical and molecular characteristics commonly represented within the patient data of the cohort of patients.

10. The method of claim 9 , wherein the threshold of the analytical characteristic is selected to:
identify presence or absence of the analytical characteristic, or
identify satisfaction of a numeric threshold of a value by the analytical characteristic.

11. The method of claim 1 , further comprising:
identifying a common anchor point in time from a set of anchor points associated with the plurality of patients, wherein the health measurement is calculated relative to the common anchor point.

12. The method of claim 1 , wherein the health measurement is determined using a predictive algorithm built on survival rates of the plurality of patients in the cohort.

13. The method of claim 1 , wherein the health measurement is a measurement of progression free survival (PFS) and wherein the health measurement is determined using an external source for PFS prediction.

14. The method of claim 13 , wherein the external source is an FDA published PFS for the cancer.

15. The method of claim 1 , wherein one of the molecular characteristics is a genetic marker.

16. The method of claim 1 , wherein one of the clinical characteristics is a procedure performed.

17. The method of claim 1 , wherein one of the clinical characteristics is a pharmaceutical treatment.

18. The method of claim 1 , wherein one of the clinical characteristics is an age at diagnosis.

19. The method of claim 1 , wherein one of the clinical characteristics is an age at treatment.

20. The method of claim 1 , wherein one of the clinical characteristics is a lifestyle indicator.

21. The method of claim 1 , wherein the health measurement is whether a patient is a smoker.

22. The method of claim 1 , wherein the health measurement is presence or absence of a genetic mutation.

23. The method of claim 22 , wherein the genetic mutation is a KRAS mutation.

24. The method of claim 1 , wherein the health measurement is an age separation value.

25. The method of claim 1 , wherein the health measurement is a gender.

26. A method for identifying an outlier group of patients within a cohort of patients, comprising:
a) generating the cohort of patients by selecting a plurality of clinical and molecular characteristics from patient data, wherein patients included in the cohort of patients satisfy the selection of the plurality of clinical and molecular characteristics, wherein each patient in the cohort of patients has been diagnosed with cancer, wherein the molecular characteristics are characteristics of a respective cancer, and wherein the clinical characteristics are characteristics of a respective patient;
b) generating a plurality of analytical characteristics from the clinical and molecular characteristics associated with the patient data from each patient within th

Assignees

• Tempus Labs Inc

Inventors

• Lefkofsky Hailey
• Hafez Ashraf
• Habib Julian
• Fishel Carin
• Epstein Caroline

Classifications

• G16H10/20 Primary

  for electronic clinical trials or questionnaires · CPC title

• G16H50/70 Primary

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G06F18/24323

  Tree-organised classifiers · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06F18/2115

  by evaluating different subsets according to an optimisation criterion, e.g. class separability, forward selection or backward elimination · CPC title