Identification and use of circulating nucleic acid tumor markers

Abstract

Methods for creating a selector of mutated genomic regions and for using the selector set to analyze genetic alterations in a cell-free nucleic acid sample are provided. The methods can be used to measure tumor-derived nucleic acids in a blood sample from a subject and thus to monitor the progression of disease in the subject. The methods can also be used for cancer screening, cancer diagnosis, cancer prognosis, and cancer therapy designation.

1 - 21 . (canceled) 
     
     
         22 . A method of producing a selector set for a cancer comprising:
 (a) identifying genomic regions comprising mutations in one or more subjects from a population of subjects suffering from the cancer;   (b) ranking the genomic regions based on a Recurrence Index (RI), wherein the RI of the genomic region is determined by dividing the number of subjects or tumors with mutations in the genomic region by the size of the genomic region; and   (c) producing a selector set based on the RI.   
     
     
         23 . The method of  claim 22 , wherein at least a subset of the genomic regions are exon regions, intron regions, untranslated regions, or a combination thereof. 
     
     
         24 . The method of  claim 22 , wherein producing the selector set based on the RI comprises selecting genomic regions that have a recurrence index in the top 70 th , 75 th , 80 th , 85 th , 90 th , or 95 th  or greater percentile. 
     
     
         25 . The method of  claim 22 , wherein producing the selector set comprises applying an algorithm to a subset of the ranked genomic regions. 
     
     
         26 . The method of  claim 22 , wherein producing the selector set comprises selecting genomic regions that:
 (i) maximize a median number of mutations per subject of the selector set,   (ii) minimize the total size of the genomic regions, or   (iii) minimize the total size of the genomic regions.   
     
     
         27 - 52 . (canceled) 
     
     
         53 . The method of  claim 22 , further comprising the step of determining a statistical significance of a selector set, the method comprising:
 (I) detecting a presence of one or more mutations in one or more samples from a subject, wherein the one or more mutations are based on the selector set of  claim 22 ;   (II) determining a mutation type of the one or more mutations present in the sample; and   (III) determining a statistical significance of the selector set by calculating a ctDNA detection index based on a p-value of the mutation type of mutations present in the one or more samples.   
     
     
         54 . The method of  claim 53 , wherein if a rearrangement is observed in two or more samples from the subject, then the ctDNA detection index is 0. 
     
     
         55 . The method of  claim 54 , wherein at least one of the two or more samples is a plasma sample or a tumor sample. 
     
     
         56 - 57 . (canceled) 
     
     
         58 . The method of  claim 53 , wherein if one type of mutation is present, then the ctDNA detection index is the p-value of the one type of mutation. 
     
     
         59 . The method of  claim 53 , wherein if: (i) two or more types of mutations are present in the sample; (ii) the p-values of the two or more types mutations are less than 0.1; and (iii) a rearrangement is not one of the types of mutations, then the ctDNA detection is calculated based on the combined p-values of the two or more mutations. 
     
     
         60 . The method of  claim 59 , wherein the p-values of the two or more mutations are combined according to Fisher's method. 
     
     
         61 . The method of  claim 59 , wherein one of the two or more types of mutations is a SNV. 
     
     
         62 . The method of  claim 61 , wherein the p-value of the SNV is determined by Monte Carlo sampling. 
     
     
         63 . The method of  claim 59 , wherein one of the two or more types of mutations is an indel. 
     
     
         64 . The method of  claim 53 , wherein if: (i) two or more types of mutations are present in the sample; (ii) a p-value of at least one of the two or more types of mutations are greater than 0.1; and (iii) a rearrangement is not one of the types of mutations, then the ctDNA detection is calculated based on the p-value of one of the two or more types mutations. 
     
     
         65 . The method of  claim 64 , wherein one of the two or more types of mutations is a SNV. 
     
     
         66 . The method of  claim 65 , wherein the ctDNA detection index is calculated based on the p-value of the SNV. 
     
     
         67 . The method of  claim 64 , wherein one of the two or more types of mutations is an indel. 
     
     
         68 - 96 . (canceled) 
     
     
         97 . The method of  claim 22 , wherein the identifying genomic regions comprises selecting genes known to be drivers in the cancer of interest to generate a pool of known drivers. 
     
     
         98 . The method of  claim 22 , wherein step (b) further comprises:
 (i) selecting exons from known drivers with the highest recurrence index that identify at least one new patient that is not from the population of subjects of step (a); and repeating until no further exons meet these criteria;   (iii) identifying remaining exons of known drivers with a Recurrence Index≥30 and with SNVs covering ≥3 patients in a relevant database that result in the largest reduction in patients with only 1 SNV; and repeating until no further exons meet these criteria;   (iv) repeating step (b) using Recurrence Index≥20;   (v) adding in all exons from additional genes previously predicted to harbor driver mutations; and   (vi) adding for known recurrent rearrangement the introns most frequently implicated in the fusion event and the flanking exons, wherein steps (i)-(vi) are embodied as a program of instructions executable by computer and performed by means of software components loaded into the computer.