Article Figures & Data
Figures
- Figure 1. SCNAs can be detected in the CSF of glioma patients and provide additional information about tumor heterogeneity
A. Concentration of total cell‐free DNA (cfDNA) in 13 CSF samples. Samples in which SCNAs were detected are shown in dark blue and tended to have higher levels of total cfDNA.
B. Relative copy number estimation in 20 genes of interest, and 18 large genomic regions, determined by sWGS of CSF from 13 glioma patients. Genes are ordered by genomic position, and their chromosomal locations are indicated in parentheses. Amplifications are shown in dark blue, deletions are in orange, and copy number neutral regions are in gray. The top green bar indicates the tumor largest length dimension (< 42 mm vs. > 42 mm), and the top violet bar indicates the glioma subtype. SCNAs were more frequently detected in CSF from patients with large glioblastoma tumors.
C. SCNAs determined by sWGS in four tumor subparts (T1 − T4) and the CSF sample, collected from patient G1. Amplifications are shown in dark blue, deletions are in orange, and copy number neutral regions are in dark gray. sWGS from plasma and urine samples collected at the same time as the CSF sample showed no SCNAs.
- Figure 2. Detection of SCNAs in CSF is influenced by tumor grade, cfDNA concentration, and contact between the tumor and CSF
A. Heat map summarizing the SCNAs detected by sWGS of 28 genes of interest in tumor biopsies and CSF from patient G1 (four tumor subparts and one CSF sample). Amplifications are shown in dark blue, deletions are in orange, and copy number neutral regions are in light gray.
B. Heat map summarizing detection of EGFR and PTEN alterations in tumor tissue and in CSF samples. Shared detection in tissue and CSF is indicated in green, detection of the alteration only in tissue in orange, and non‐detection in blue. The top bars indicate the cfDNA concentration (copies/ml; in a range of purples), the size of the tumors (in a range of browns), the type of glioma (in a range of blues), and whether the tumor was in direct contact with the CSF or not (based on MRI, green or red, respectively). Samples are ranked from the left to right by decreasing concentration of cfDNA (copies/ml).
- Figure 3. Distribution of DNA fragment sizes in the CSF is related to the presence of tumor DNA
A. Fragment size distribution, determined by sWGS, in CSF (blue line), plasma (red dashed line), and urine (black dashed line) from patient G1. The three samples were collected simultaneously before initiation of treatment.
B. Fragment size distribution of cfDNA, determined by sWGS, in CSF from 13 glioma patients. The samples contained a high fraction of reads from DNA fragments with lengths corresponding to ˜145 bp and ˜167 bp. The fragment size profile of cfDNA from the plasma sample from patient G1 is shown in red. All samples were collected simultaneously before initiation of treatment.
C. Cumulative frequency analysis of the average density of fragment size in CSF (blue) and plasma (red). The vertical dashed line represents 167 bp.
D. Mean fragment size density in patients with detected SCNAs in the CSF (yellow), and those with no detected SCNAs (dark blue).
E. Log2 ratio of the difference in cfDNA fragment sizes between CSF samples with detected SCNAs and those samples in which no SCNAs were detected.
F. The amplitude of the 10 bp periodic peaks in fragment size is a signature for non‐tumor DNA in the CSF. The amplitude of the 10 bp periodic peaks in fragment length distribution (in the range between 75 and 150 bp, see Materials and Methods) when compared to the mean copy number alteration calculated from sWGS revealed a significant negative correlation (Pearson −0.85; P = 0.0002).
Supplementary Materials
