Figure 1 from Serial Plasma Comprehensive Genomic Profiling Captures Therapy Resistance and Guides Management of Non–Small Cell Lung Cancer

Impact of plasma genotyping vs. tissue on patient management. A–D, We considered circumstances in which pCGP was informative and concurrent tissue biopsy uninformative and then the converse, situations in which tissue biopsy was informative and concurrent pCGP uninformative. This latter analysis was limited to second and third assays of genotyping. The total number of patients with tissue biopsy was 31, of whom 22 (71%) had tissue NGS performed as well. Tumor fraction was not calculated as part of the pCGP assays. A, Instances when pCGP was informative and tissue uninformative, clustered according to the iteration of pCGP. For two patients marked as “other,” one had imaging changes that could represent NSCLC recurrence or Mycobacterium avium infection. The patient was hesitant to proceed with tissue biopsy, and the treating physician ordered pCGP, which identified a KRAS G12R mutation. This was considered to increase the likelihood of NSCLC recurrence and to justify the risks of tissue biopsy to obtain confirmation. In the second case, a frail patient had a BRAF V600E alteration on tissue NGS at first diagnosis and progressed through several lines of chemo- and immunotherapy. The treating physician performed repeat pCGP at progression and found no BRAF V600E. Given the toxicities of therapy and concern that the BRAF alteration was no longer present and circulating, the physician was guided not to pursue combined BRAF/MEK inhibition. B, Instances when tissue was informative and contemporaneous pCGP uninformative. C, In the EGFRm population subset, instances when pCGP was informative, whereas tissue NGS was uninformative. D, In the EGFRm population subset, instances when tissue was informative and contemporaneous pCGP uninformative. Although the number of cases was relatively small (n = 15), we found that concurrent tissue biopsy could guide management in more than half of the cases (57%) through information not available on pCGP, including histologic confirmation of radiographically equivocal sites of progression and the diagnosis of small cell histologic transformation. E, Sankey diagram illustrating the mechanisms through which pCGP influenced patient management at the first and second iterations of testing. pCGP was uniquely informative to patient management in the first and second iterations of testing, with benefits accruing to some patients more than once. At the first testing, it outperformed alternative methods for practical reasons, such as rapid turnover and permitting testing in patients who might not otherwise be fit, but it also identified some alterations that were not present on tissue NGS. It continued to be beneficial in the second iteration for patients in which tissue testing was not possible, which may reflect declining fitness for biopsy in this population.