Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Aug;11(8):e499.
doi: 10.1002/ctm2.499.

Tri-modal liquid biopsy: Combinational analysis of circulating tumor cells, exosomes, and cell-free DNA using machine learning algorithm

Jiyoon Bu  1   2 , Tae Hee Lee  3   4 , Michael J Poellmann  1   2 , Piper A Rawding  1   2 , Woo-Jin Jeong  1   2   5 , Rachel S Hong  1 , Sung Hee Hyun  4 , Hyuk Soo Eun  6 , Seungpyo Hong  1   2   7
Affiliations

Tri-modal liquid biopsy: Combinational analysis of circulating tumor cells, exosomes, and cell-free DNA using machine learning algorithm

Jiyoon Bu et al. Clin Transl Med. 2021 Aug.
No abstract available

PubMed Disclaimer

Figures

FIGURE 1
FIGURE 1
Schematic illustration of the machine learning‐based multimodal analysis of the triple tumor biomarkers – CTCs, exosomes, and cfDNA: (A) A graphical abstract of the multimodal liquid biopsy analysis. (B) Schematic illustration depicting the experimental and analytical procedures for the isolation of CTCs, exosomes, and cfDNA using functionalized alginate beads. (C) Clustering of the CTC counts, exosome NA amounts, and plasma cfDNA concentrations based on the machine learning algorithm; (D) Establishment of MMLBScore for each cohort based on PCA. (E) Validation of the clinical utility of the MMLB analysis
FIGURE 2
FIGURE 2
The diagnostic capability of the new bead‐based system for (A) CTCs, (B) exosomes, and (C) cfDNA tested using the human colorectal cancer cell line, SW480 cells. (D‐F) Target specificity of the bead‐based system, validated by comparing CTCs, exosomes, and cfDNA captured on each type of functionalized beads to those captured on bare alginate beads. Note that the capture of all three biomarkers was not prominent on the bare alginate beads, implying that the alginate itself does not contribute to the capture of tumor biomarkers. (G) Number of CTCs, (H) amount of exosome NA, and (I) concentration of plasma cfDNA quantified from a cohort consisting of 72 patients with malignant tumors, 14 patients with benign tumors, and 14 healthy individuals. For exosomes and cfDNA, the bead‐based system was compared with commercially available kits, ExoQuick kit, and Qiagen mini kit, respectively. (J‐L) ROC curves demonstrating the diagnostic capability of the new bead‐based system for distinguishing the patients with malignant tumors from healthy individuals, patients with benign tumors, and overall non‐cancer cohorts (healthy individuals + patients with benign tumors)
FIGURE 3
FIGURE 3
Machine learning‐based clustering of the three tumor biomarkers (CTCs, exosomes, and cfDNA) to establish MMLBScore and validate its diagnostic potential: (A) A 3D scatter plot demonstrating CTC count, exosome NA amount, and plasma cfDNA level of each patient depending on the status (malignant, benign, and healthy) of the cohorts. (B) A k‐means clustering of the cohorts based on the expression levels of the three tumor biomarkers. A total of 41, 18, 20, 8, and 13 samples were designated to each cluster, denoted as A1, A2, A3, A4, and A5, respectively. (C‐E) CTC count, exosome NA amount, and plasma cfDNA level for each cluster. (F) PCA applied to reduce the complexity of the 3D plot (CTCs, exosomes, and cfDNA) into the arbitrary 2D plot, with the x‐ and y‐axes consisted of two best linear approximations for stratifying the clusters. The x‐axis in the 2D scatterplot was determined as MMLBScore, which minimized the mean‐squared reconstruction error of the clusters and demonstrated a strong correlation with the status of the cohorts. (G and H) MMLBScore depending on the status of the cohorts. (I and J) The ROC curve and a heatmap of AUC‐ROC values demonstrating the enhanced diagnostic capability of MMLBScore compared to any of the single tumor biomarkers used in this study. (K) A heatmap showing expression levels of the three biomarkers compared to MMLBScore for each cohort
FIGURE 4
FIGURE 4
MMLB analysis for determining the pathological features of a tumor, estimating the survival outcomes, and detecting KRAS mutation: (A) A heatmap representing MMLBCancer for each patient, along with their TNM stage, tumor biomarker expressions, and serum antigen expressions. (B) MMLBCancer for each patient depending on their T stage. (C‐E) ROC curves demonstrating the diagnostic capability of MMLBCancer (black), CTCs (red), exosomes (blue), and cfDNA (green) for differentiating advanced T stage patients. (F) A heatmap of AUC‐ROC values demonstrating the capability of MMLBCancer for differentiating patients depending on the pathological features of the tumor, including its size (T stage), the existence of nodal metastasis (N stage), the prevalence of distant metastasis (M stage), and LVI status. (G‐N) Kaplan–Meier survival analysis for (G‐J) DFS and (K‐N) OS between the patients with high (dark) and low (bright) MMLBCancer, CTC counts, exosome NA level, and cfDNA expressions. The median value for each biomarker (or score) was determined as a threshold for dividing the high versus low groups. (O) A heatmap of HR values demonstrating the prognostic capability of MMLBCancer was compared with the individual tumor biomarkers and serum antigens. (P‐U) The clinical capability of MMLB analysis to determine the tissue KRAS mutation. (P) A k‐means clustering of patients based on CTC counts, miR‐100 expression in exosomes, and the KRAS MAF in cfDNA. The size of the sphere is proportional to the fraction of KRAS mutant allele found in tissue. (Q) The KRAS MAF in tissue for each cluster. (R‐U) 2D scatterplots representing a correlation with the tissue KRAS MAF for MMLBKRAS and the single tumor biomarkers
See this image and copyright information in PMC

References

    1. Siravegna G, Marsoni S, Siena S, Bardelli A. Integrating liquid biopsies into the management of cancer. Nat Rev Clin Oncol. 2017;14:531–548. - PubMed
    1. Bu J, Kang Y‐T, Lee Y‐S, et al. Lab on a fabric: mass producible and low‐cost fabric filters for the high‐throughput viable isolation of circulating tumor cells. Biosens Bioelectron. 2017;91:747–755. - PubMed
    1. Myung JH, Eblan MJ, Caster JM, et al. Multivalent binding and biomimetic cell rolling improves the sensitivity and specificity of circulating tumor cell capture. Clin Cancer Res. 2018;24:2539–2547. - PMC - PubMed
    1. Finotti A, Allegretti M, Gasparello J, et al. Liquid biopsy and PCR‐free ultrasensitive detection systems in oncology (Review). Int J Oncol. 2018;53:1395–1434. - PMC - PubMed
    1. Punnoose EA, Atwal S, Liu W, et al. Evaluation of circulating tumor cells and circulating tumor DNA in non–small cell lung cancer: association with clinical endpoints in a phase II clinical trial of pertuzumab and erlotinib. Clin Cancer Res. 2012;18:2391. - PubMed

Substances

LinkOut - more resources