Faecal Volatile Organic Compounds to Detect Colorectal Neoplasia in Lynch Syndrome-A Prospective Longitudinal Multicentre Study

Abstract

Background: Non-invasive biomarkers may reduce post-colonoscopy colorectal cancer (CRC) rates and colonoscopy overuse in Lynch syndrome. Unlike faecal immunochemical test (FIT), faecal volatile organic compounds (VOCs) may accurately detect both advanced and non-advanced colorectal neoplasia.

Aim: The aim of this study was to evaluate the potential of faecal VOCs-separately and with FIT-to guide optimal colonoscopy intervals in Lynch syndrome.

Methods: Prospective longitudinal multicentre study in which individuals with Lynch syndrome collected faeces before and after high-quality surveillance colonoscopy. VOC-patterns were analysed using field asymmetric ion mobility spectrometry (FAIMS) and gas chromatography-ion mobility spectrometry (GC-IMS) followed by machine learning pipelines, and combined with FIT at 2.55 μg Hb/g faeces. Gas chromatography time-of-flight mass spectrometry analysed individual VOC abundance.

Results: Among 200 included individuals (57% female, median 51 years), 62 had relevant neoplasia at colonoscopy: 3 CRC, 6 advanced adenoma (AA), 3 advanced serrated lesion (ASL), and 50 non-advanced adenoma (NAA). Respective sensitivity and negative predictive value for CRC and AA (and also ASL in case of FAIMS) were 100% and 100% using FAIMS (54% specificity), and 89% and 99% using GC-IMS (58% specificity). Respective sensitivity and specificity for any relevant neoplasia were 88% and 44% (FAIMS) and 84% and 28% (GC-IMS); accuracy did not significantly improve upon VOC-FIT. VOC-patterns differed before and after polypectomy (AUC 0.70). NAA showed decreased faecal abundance of butanal, 2-oxohexane, dimethyldisulphide and dimethyltrisulphide.

Conclusions: In Lynch syndrome, faecal VOCs may be a promising strategy for postponing colonoscopy and for follow-up after polypectomy. Our results serve as a stepping stone for large validation studies.

Trial registration: NL8749.

Keywords: Lynch syndrome; biomarkers; colorectal cancer; faecal immunochemical test; surveillance; volatile organic compounds.

FIGURE 1

Flow diagram showing the different analyses conducted, including numbers of patients analysed, to evaluate the performance of faecal volatile organic compounds for detection of colorectal neoplasia (upper red box) and for intra‐individual follow‐up after polypectomy (lower blue box). Abbreviations: CRC = colorectal cancer; FAIMS = field asymmetric ion mobility spectrometry; FIT = faecal immunochemical test; GC‐IMS = gas chromatography—ion mobility spectrometry; GC‐TOF‐MS = gas chromatography time‐of‐flight mass spectrometry. ^aDue to insufficient faecal sample, FAIMS could not be evaluated for neoplasia detection in 68 individuals nor for neoplasia follow‐up in any individual. ^bDue to financial constraints, GC‐TOF‐MS was performed in a random selection of 13 non‐advanced adenomas and 14 controls. ^cFor sensitivity analysis, individuals with and without colorectal neoplasia were matched 1:1 on possible confounders: Gender, age, body mass index, smoking habits and dietary habits. Only n = 8 (GC‐IMS) and n = 6 (FAIMS) non‐advanced adenomas could not be matched and therefore were excluded. Moreover, the analysis ‘CRC + advanced adenomas’ could not be performed for FAIMS due to insufficient number of CRC and advanced adenomas in this group.

FIGURE 2

Receiver operating characteristic curves to detect colorectal neoplasia in Lynch syndrome by faecal volatile organic compounds, as measured with gas chromatography—ion mobility spectrometry (GC‐IMS) and field asymmetric ion mobility spectrometry (FAIMS).

FIGURE 3

The performance of faecal volatile organic compounds, as analysed with gas chromatography—ion mobility spectrometry, for intra‐individual follow‐up after polypectomy in Lynch syndrome. Samples collected before and after normal colonoscopy were analysed (‘controls’) as well as samples collected before and after complete removal of adenomas and advanced serrated lesions (‘polypectomy patients’). ^aOf the 52 polypectomy patients, 5 (10%) had advanced adenomas, 3 (6%) advanced serrated lesions and 44 (85%) non‐advanced adenomas. All neoplasia were resected en‐bloc, except from one lesion which was resected piecemeal. ^bSensitivity analysis included 46 controls and 46 polypectomy patients, of which 4/46 (9%) had advanced adenomas, 2/46 (4%) advanced serrated lesions and 40/46 (87%) non‐advanced adenomas.

FIGURE 4

Fagan's nomograms illustrating the probability of relevant neoplasia (colorectal cancer, advanced serrated lesions and all adenomas), following positive (blue line) or negative (red line) results of (A) FIT at threshold 2.55 μg Hb/g faeces^a, (B) GC‐IMS analysis in FIT‐negatives and (C) FAIMS analysis in FIT‐negatives. The grey line represents the situation when the post‐test probability remains unchanged from the pre‐test probability, which represents the relevant neoplasia prevalence. Abbreviations: FAIMS = field asymmetric ion mobility spectrometry; FIT = faecal immunochemical test; GC‐IMS = gas chromatography—ion mobility spectrometry; IQR = interquartile range. ^aFigure 4A Is adapted from our previous study [17]. This cohort consisted of 217 individuals with Lynch syndrome, in which FIT detected 4/4 colorectal cancers, 4/5 advanced adenomas, 1/4 advanced serrated lesions and 9/57 non‐advanced adenomas. As such, relevant neoplasia were present in 28% of FIT‐negatives further tested for VOC patterns: 0.5% had advanced adenomas, 1.5% advanced serrated lesions, and 26% non‐advanced adenomas.

FIGURE 5

Acceptability of faeces collection and surveillance colonoscopy by 200 individuals with Lynch syndrome on a scale from extremely burdensome [0] to not burdensome [10]. Median patient acceptability was 7 (IQR 6–9) for faeces collection and 6 (IQR 4–8) for colonoscopy, p‐value < 0.001.