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
. 2023 Jun 28;11(3):638-648.
doi: 10.14218/JCTH.2022.00309. Epub 2023 Feb 2.

Breath Biopsy® to Identify Exhaled Volatile Organic Compounds Biomarkers for Liver Cirrhosis Detection

Giuseppe Ferrandino  1 , Giovanna De Palo  1 , Antonio Murgia  1 , Owen Birch  1 , Ahmed Tawfike  1 , Rob Smith  1 , Irene Debiram-Beecham  2 , Olga Gandelman  1 , Graham Kibble  2 , Anne Marie Lydon  2 , Alice Groves  2 , Agnieszka Smolinska  1   3 , Max Allsworth  1 , Billy Boyle  1 , Marc P van der Schee  1 , Michael Allison  4   5 , Rebecca C Fitzgerald  6 , Matthew Hoare  4   5   7 , Victoria K Snowdon  5
Affiliations

Breath Biopsy® to Identify Exhaled Volatile Organic Compounds Biomarkers for Liver Cirrhosis Detection

Giuseppe Ferrandino et al. J Clin Transl Hepatol. .

Abstract

Background and aims: The prevalence of chronic liver disease in adults exceeds 30% in some countries and there is significant interest in developing tests and treatments to help control disease progression and reduce healthcare burden. Breath is a rich sampling matrix that offers non-invasive solutions suitable for early-stage detection and disease monitoring. Having previously investigated targeted analysis of a single biomarker, here we investigated a multiparametric approach to breath testing that would provide more robust and reliable results for clinical use.

Methods: To identify candidate biomarkers we compared 46 breath samples from cirrhosis patients and 42 from controls. Collection and analysis used Breath Biopsy OMNI™, maximizing signal and contrast to background to provide high confidence biomarker detection based upon gas chromatography mass spectrometry (GC-MS). Blank samples were also analyzed to provide detailed information on background volatile organic compounds (VOCs) levels.

Results: A set of 29 breath VOCs differed significantly between cirrhosis and controls. A classification model based on these VOCs had an area under the curve (AUC) of 0.95±0.04 in cross-validated test sets. The seven best performing VOCs were sufficient to maximize classification performance. A subset of 11 VOCs was correlated with blood metrics of liver function (bilirubin, albumin, prothrombin time) and separated patients by cirrhosis severity using principal component analysis.

Conclusions: A set of seven VOCs consisting of previously reported and novel candidates show promise as a panel for liver disease detection and monitoring, showing correlation to disease severity and serum biomarkers at late stage.

Keywords: Biomarker; Breath Biopsy; Cirrhosis; Liver function test; Non-invasive.

PubMed Disclaimer

Conflict of interest statement

Giuseppe Ferrandino, Giovanna De Palo, Antonio Murgia, Owen Birch, Rob Smith, Ahmed Tawfike, Olga Gandelman, Max Allsworth, and Billy Boyle, are/were employees of Owlstone Medical Ltd.

Figures

Fig. 1
Fig. 1. Volcano plot of exhaled VOCs.
The X-axis represents the log2 mean ratio fold-change of the relative abundance of each VOC between cirrhosis and controls. The Y-axis represents the p-value of each VOC. Compounds with fold-change >2 and p<0.05 are highlighted in blue. Limonene and 2-pentanone were elevated in the breath of patients with cirrhosis and dimethyl selenide was reduced, as expected. VOC, volatile organic compound.
Fig. 2
Fig. 2. Box plots of discriminatory VOCs between cirrhosis and controls.
A total of 29 on-breath VOCs were found significantly different (p<0.05, Mann-Whitney U-test) between control and cirrhosis groups. VOCs, volatile organic compounds.
Fig. 3
Fig. 3. Receiver operating characteristic plots of the four top single VOCs comparing cirrhosis vs. controls.
The top 4 ROC plots for on-breath VOCs were calculated to explore their discriminatory performance. 2-pentanone, limonene, and dimethyl selenide were found among them. ROC, receiver operating characteristic; VOCs, volatile organic compounds.
Fig. 4
Fig. 4. Classification performance of combined VOCs.
(A) ROC plot and confidence interval obtained for the training set. (B) ROC plot and confidence interval obtained for the test set. (C) Corresponding confusion matrix generated using the Youden index as threshold. (D) Improvements of classification performance by addition of VOCs to the model. ROC, receiver operating characteristic; VOCs, volatile organic compounds.
Fig. 5
Fig. 5. Correlation of breath VOCs with blood metrics of liver function in cirrhosis subjects.
(A) Correlation plot of identified VOCs and serum bilirubin, albumin, and INR. Blue indicates a negative and red a positive correlation. Circle size and color intensity show the magnitude of the correlation. (B) CCA score plot using the first canonical variates of selected sets of VOCs and blood metrics of liver function. Each projected data point represents the combined information of breath VOCs and blood metrics of one cirrhotic patient. The CCA analysis revealed significant correlations, with R2=0.842. CCA, canonical correlation analysis; INR, international normalized ratio; VOCs, volatile organic compounds.
Fig. 6
Fig. 6. Canonical loadings of set of variables.
Canonical loadings represent the correlation between a variable and its canonical variate and express the contribution of each variable to the overall correlation. (A) Loadings for blood variables, albumin has opposite correlation than bilirubin and INR, as expected. (B) Loadings for breath VOCs, limonene and 2-pentanone had a positive contribution. Dimethyl selenide (VOC4), which was downregulated in the breath of patients with cirrhosis, had a negative correlation. INR, international normalized ratio; VOCs, volatile organic compounds.
Fig. 7
Fig. 7. VOC alterations in relation to cirrhosis severity.
Projected data points of the first two components of a PCA calculated using breath variables with a BH adjusted p<0.1. PC1 explains 10.1% and PC2 4.4% of variance and shows that separation of cirrhotic patients based on their CP score. About 50% of the patients with a CP=5 cluster with controls. BH, Benjamin-Hochberg; CP, Child-Pugh; PCA, principal component analysis; VOC, volatile organic compound.
See this image and copyright information in PMC

References

    1. Holzhutter HG, Wuensch T, Gajowski R, Berndt N, Bulik S, Meierhofer D, et al. A novel variant of the (13)C-methacetin liver function breath test that eliminates the confounding effect of individual differences in systemic CO2 kinetics. Arch Toxicol. 2020;94(2):401–415. doi: 10.1007/s00204-020-02654-0. - DOI - PubMed
    1. Tsochatzis EA, Bosch J, Burroughs AK. Liver cirrhosis. Lancet. 2014;383(9930):1749–1761. doi: 10.1016/S0140-6736(14)60121-5. - DOI - PubMed
    1. Soresi M, Giannitrapani L, Cervello M, Licata A, Montalto G. Non invasive tools for the diagnosis of liver cirrhosis. World J Gastroenterol. 2014;20(48):18131–18150. doi: 10.3748/wjg.v20.i48.18131. - DOI - PMC - PubMed
    1. Murgia A, Ahmed Y, Sweeney K, Nicholson-Scott L, Arthur K, Allsworth M, et al. Breath-Taking Perspectives and Preliminary Data toward Early Detection of Chronic Liver Diseases. Biomedicines. 2021;9(11):1563. doi: 10.3390/biomedicines9111563. - DOI - PMC - PubMed
    1. Gaude E, Nakhleh MK, Patassini S, Boschmans J, Allsworth M, Boyle B, et al. Targeted breath analysis: exogenous volatile organic compounds (EVOC) as metabolic pathway-specific probes. J Breath Res. 2019;13(3):032001. doi: 10.1088/1752-7163/ab1789. - DOI - PubMed