Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Detecting volatile organic compounds (VOCs) could provide a rapid, noninvasive, and inexpensive screening tool for detecting cancer.

OBJECTIVE:

In this systematic review, we identified specific exhaled breath VOCs correlated with lung, colorectal, and breast cancer.

METHODS:

We identified relevant studies published in 2015 and 2016 by searching Pubmed and Web of Science. The protocol for this systematic review was registered in PROSPERO and the PRISMA guidelines were used in reporting. VOCs and performance data were extracted.

RESULTS:

Three hundred and thirty three records were identified and 43 papers were included in the review, of which 20 were review articles themselves. We identified 17 studies that listed the VOCs with at least a subset of statistics on detection cutoff levels, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and gradient.

CONCLUSIONS:

Breath analysis for cancer screening and early detection shows promise, because samples can be collected easily, safely, and frequently. While gas chromatography-mass spectrometry is considered the gold standard for identifying specific VOCs, breath analysis has moved into analyzing patterns of VOCs using a variety of different multiple sensor techniques, such as eNoses and nanomaterials. Further development of VOCs for early cancer detection requires clinical trials with standardized breath sampling methods.

Keywords

Exhaled breath volatile organic compounds (VOCs)biomarkers lung cancer colorectal cancer breast cancer

1. Introduction

Hippocrates, considered to be the father of medicine, taught his students to smell the breath of their patients as an indicator of disease in about 400 years BC [1]. In 1971, Linus Pauling and co-workers reported that normal human breath contains a complex mixture of several hundred volatile organic compounds (VOCs) [2] and recently scientists from the Max Planck Institute for Chemistry have observed differing VOC levels in movie audiences depending upon the emotional content of the film [3]. VOCs are a diverse group of chemicals emitted from the human body that often reflect the metabolic condition of the person [4]. Concentrations of specific organic compounds in the exhaled breath of cancer patients are thought to be the result of oxidative stress, abnormal metabolic processes, or an imbalance between the systematic manifestations of reactive oxygen species and a biological system’s ability to detoxify or repair damage [4, 5, 6]. Some VOCs are thought to originate from cancer cells as a result of abnormal metabolism that does not exist in healthy people [5, 7].

Lung, breast, and colorectal cancer are among the top five cancers in terms of both incidence and mortality. Lung cancer has the highest mortality rate of these; in 2016, it is estimated that 224,000 Americans will be diagnosed with lung cancer, and 158,000 will die of it. Breast cancer incidence is estimated at 249,000, with 41,000 deaths. And colorectal cancer is estimated to strike 134,000 Americans and to kill 49,000. Improved screening may reduce mortality for all of these cancers [8] by detecting cancer when it is small and more easily treated. Although there are a variety of techniques available to detect some cancers at an early stage, many are considered by patients to be uncomfortable and inconvenient and are often underutilized by large segments of the population for a variety of reasons, such as access and patient burden [9].

While cancer screening biomarkers have been developed using biological samples such as blood and urine, detecting specific levels of VOCs as cancer biomarkers holds the potential to provide a low patient burden, rapid, noninvasive, and inexpensive cancer screening tool [10, 11, 12]. Breath samples have elicited interest because the samples are easy to collect and analyze, and the literature in this area continues to grow [4, 5, 6, 7, 9, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55]. Testing exhaled breath is painless and noninvasive, and therefore suitable even for children and critically ill patients. Because exhaled breath tests can be performed frequently, they could also be helpful in monitoring cancer to catch recurrence or signs of progression early [9].

Van de Kant et al. provided a systematic review of VOCs in pulmonary diseases in 2012 [56]. In 2015, Krilaviciute et al. completed an extensive, detailed systematic review of cancer detection through exhaled breath [29], including 73 studies. This study found that analyzing exhaled breath is promising, but standardization of breath collection, sample storage and data handling are required in larger studies.

The science of detecting disease through exhaled breath has grown exponentially in the past few years with many technology teams developing new sensor techniques and devices. The aim of the following systematic review was to identify specific exhaled breath VOCs correlated with lung, colorectal, and breast cancer, along with reported detection levels, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and the concentration gradient in cancer patients. We focused on studies published after the Krilaviciute 2015 review [29].

2. Methods

2.1 Protocol and registration

The protocol for this systematic review was registered in the PROSPERO [57] database (CRD420160 45513) and the PRISMA guidelines were used in reporting the review [58]. Funding was received from Nano Engineered Applications (NEA) [59].

2.2 Eligibility criteria

Inclusion criteria were for studies of human breath VOCs related to breast, colorectal, or lung cancer, as discussed in either original articles or review papers. Duplicate articles, articles not directly relevant to the topic, articles without abstracts, and articles not focused on testing in humans were excluded. Two reviewers (Davis and Oakley-Girvan) independently screened the studies for eligibility. Disagreements were resolved by discussion.

2.3 Information sources and search process

On June 10, 2016, we searched both the PubMed and Web of Science databases of scientific articles using the following search string:

((cancer OR carcinoma OR adenocarcinoma OR tumor OR malignancy OR malignant disease) AND ((volatile AND (compound OR compounds OR marker OR markers OR biomarker OR biomarkers)) OR (VOC OR VOCs OR breathprint OR breath-print OR breath print) AND (breath OR exhaled OR air))) Filters: LANGUAGE: (English) Publication Years: 2015 AND 2016.

Figure 1.

Flow diagram for search strategy methods, reported according to the PRISMA statement. *outdoor/indoor air pollutants ( $n=$ 78), technical inventions not tested in humans ( $n=$ 41), not cancer ( $n=$ 22: diabetes, COPD, Helicobacter pylori, asthma, sleep apnea, ulcers, Crohn’s disease, tuberculosis, halitosis, arthritis, gastrointestinal illnesses, lower limb ulcers, kidney disease, bronchopulmonary, cystic fibrosis, chronic liver diseases, sclerosing cholangitis, pneumonia, caries, inflammatory bowel disease), not breath ( $n=$ 6), canines ( $n=$ 2), in vitro ( $n=$ 11), other cancers ( $n=$ 9: ovarian [73, 74, 75], esophageal [76, 77], gastric [78], thyroid [79], urological [80], malignant glottic lesions [81]), plants/herbs $n=$ 4, annual meeting summary $n=$ 1, no abstract $n=$ 17, other ( $n=$ 12: ICU VOCs, transplant rejection, facial deformities, fungal infections, hypoxia, radiation exposure, rats).

2.4 Study selection

In total, 333 records were identified, 99 fromPubMed, and 234 from Web of Science, of which 80 were duplicates. We reviewed the remaining 253 titles and abstracts to identify studies relevant to the topic. Articles excluded from the review were categorized as shown in Fig. 1. We reviewed 50 full-text papers for inclusion; seven papers were excluded after full-text review. A total of 43 papers were included in the systematic review, of which 20 were review articles themselves, and 23 were studies of VOCs in humans related to lung (19), colorectal (2), or breast cancer (2).

2.5 Data collection process

The relevant data were extracted from the 43 selected papers by two reviewers (Davis and Oakley-Girvan). We extracted the following information using a predesigned data extraction form: first author, publication year, number of cases, number of controls, names of volatile compounds tested, cut-off level, sensitivity, specificity, area under the receiver operating characteristic curve (AUC), and gradient. For those papers that did not include this information, we collected the first author, publication year, number of cases, a number of controls, type of study, and the accuracy statistics provided by the study comparing cancer cases with healthy controls where available. Review papers were searched for the same data elements.

Table 1
Lung cancer compounds, performance, and the concentration gradient in cancer patients compared with controls ${}^{\rm a}$

First author,publication year, N ${}^{\rm b}$	Volatile compounds;methods	Cut-off level	Sens	Spec	AUC	Gradient
Corradi, 2015 [21]N $=$ 71 casesN $=$ 67 with suspected cancer, but no dx	H ${}_{2}$ O ${}_{2}$ -EBCTrans-2-nonenal;SPME-GC-MS	0.21 $\mu$ M4.1 pM	63.460.6	61.262.7	0.630.64	UpUp
Fernandes, 2015 [26] (review)	AcetaldehydeFormaldehydeUndecaneIsopropeneMethanolEthylbenzeneAcetone;GC-MS	$<$ 2–10 ppb $<$ 2–10 ppb24 $\pm$ 4 ppb72.0–81.5 ppb $<$ 118.5 ppb145 $\pm$ 35 ppb $<$ 458.7 ppb	NP ${}^{\rm c}$	NP	NP	NP
Fuchs, 2010 [64]N $=$ 12 casesN $=$ 24 controls	PentanalHexanalOctanalNonanal;SPME-GC-MS	0.275 nmol/L1.208 nmol/L1.068 nmol/L8.433 nmol/L	75.0 8.358.333.3	95.891.791.795.8	NPNPNPNP	UpUpUpUp
Handa, 2014 [82]N $=$ 50 casesN $=$ 39 controls	Dodecaine;MCC/IMS	NP	70.0	89.7	NP	Up
Li, 2015 [30]N $=$ 85 casesN $=$ 34 benign nodulesN $=$ 85 healthy controls	C ${}_{4}$ H ${}_{8}$ O4-HHEC ${}_{4}$ H ${}_{8}$ O ${}_{2}$ C ${}_{2}$ H ${}_{4}$ O ${}_{2}$ 4-HNEC ${}_{5}$ H ${}_{10}$ O;FT-ICR-MS	2.255 nmol/L0.00672 nmol/L0.1595 nmol/L0.3875 nmol/L0.000255 nmol/L1.315 nmol/L	All comp96.0	All comp64.0–100	NP	Up
Ligor, 2015 [31]N $=$ 123 casesN $=$ 361 controls	ButaneButane, 2-methyl-Octane, 4-methyl-Propane2-pentanonePropanalHeptane, 2,4-dimethyl-Propene;SPME-GC-MS	NP	All comp62.5	All comp72.4	All comp0.65	Up
Schumer, 2015 [46]N $=$ 156 casesN $=$ 65 benignN $=$ 194 control	2-butanone,3-hydroxy-2-butanone, and2-hydroxyacetadehyde,4-hydroxyhexanal;FT-ICR-MS	All comp $\geqslant$ 1 SDs above mean in controls	All comp93.6	All comp85.6	NP	Up
Song, 2010 [71]N $=$ 43 casesN $=$ 41 controls	Butan-1-ol3-hydroxybutan-2-one;SPME-GC-MS	3.67 ng/L3.81 ng/L	95.393.0	85.492.7	0.940.96	UpUp
Wang, 2012 [83]N $=$ 47 casesN $=$ 42 controls	Hexadecanal;SPME-GC-MS	NP	96.5	89.2	0.949	–
Zou, 2014 [84]Training set: N $=$ 79 cases N $=$ 92 controlsValidation set: N $=$ 58 cases N $=$ 20 controls	5- (2-methylpropyl) nonane2, 6-di-tert-butyl-4-methyl-phenol2, 6, 11-trimethyldodecane hexadecanal8-hexylpentadecane;SPME-GC-MS	NPNPNPNPNP	NPNPNPNPNP	NPNPNPNPNP	0.8450.7240.8461.000.672	UpUpUpUpUp

${}^{\rm a}$ Data prior to 2015 from Krilaviciute, 2015 [29]. ${}^{\rm b}$ N $=$ number by case/control status. ${}^{\rm c}$ NP $=$ not provided/available in the published manuscript. Sens: sensitivity; Spec: specificity; AUC: area under the receiver operating characteristic curve; ppbv: parts per billion by volume; All comp: all compounds were tested together. Methods: GC $=$ gas chromatography; MS $=$ mass spectrometry; MCC/IMS $=$ multi-capillary column-ion mobility spectrometry; FT-ICR-MS $=$ Fourier transform ion cyclotron resonance mass spectrometry; SPME $=$ solid phase micro extraction.

Table 2

Breast cancer compounds, performance, and the concentration gradient in cancer patients compared with controls ${}^{\rm a}$

First author, publication year, N ${}^{\rm b}$	Volatile compound; method	Cut-off	Sens	Spec	AUC	Gradient
Barash, 2015 [16]N $=$ 80 cases (5 subtypes); N $=$ 46 controls (23 healthy, 13 benign breast condition, 10 DCIS ${}^{\rm d}$ )	Ethanol2-Propenoic acid, butyl esterAlpha-pinene; GC-MS	NP ${}^{\rm c}$	All comp 86.0	All comp70.0	All comp0.93	NP
Li, 2014 [85]N $=$ 22 casesN $=$ 24 controls	HexanalHeptanalOctanalNonanalEthenylbenzeneDecane; SPME/GC-MS	10.32 ppbv9.98 ppbv12.9 ppbv23.14 ppbv14.92 ng/L 1.64 ng/L	77.3 68.2 63.6 72.7 66.7 86.7	79.291.787.595.894.458.3	0.790.8230.7340.8320.8120.798	UpUpUpUpUpUp
Mangler, 2012 [66]N $=$ 10 casesN $=$ 10 controls	3-methylhexaneDec-1-eneCaryophylleneNaphthalenetrichloroethene;TD-GC-MS	$-$ 0.55 $\mu$ g/m ${}^{3}$ $-$ 0.125 $\mu$ g/m ${}^{3}$ $-$ 0.05 $\mu$ g/m ${}^{3}$ 0.05 $\mu$ g/m ${}^{3}$ 0.05 $\mu$ g/m ${}^{3}$	100.0100.0100.0 90.0 80.0	40.040.060.070.070.0	NPNPNPNPNP	DownDownDownDownUp

${}^{\rm a}$ Data prior to 2015 from Krilaviciute, 2015 [29]. ${}^{\rm b}$ N $=$ number by case/control status. ${}^{\rm c}$ NP $=$ not provided/available in the published manuscript. ${}^{\rm d}$ DCIS $=$ Ductal carcinoma in situ. Sens: sensitivity; Spec: specificity; AUC: area under the receiver operating characteristic curve; ppbv: parts per billion by volume; All comp: all compounds were tested together. Method: GC $=$ gas chromatography; MS $=$ mass spectrometry; SPME $=$ solid phase microextraction; TD $=$ thermal desorption.

3. Results

3.1 Summary

Most of the studies that identified individual VOCs used gas chromatography-mass spectrometry (GC-MS), which is considered the gold standard because it allows metabolite identification that is transferable between instruments with different manufacturers [5]. The GC-MS technique is not only easily reproducible, it has high sensitivity and robustness for individual VOCs [60].

Proton transfer reaction mass spectrometry (PTR-MS) is another method for measuring volatile signatures [61, 62]. Fienberg et al. looked at the volatile signature in lung cancer before and after oral glucose tolerance testing using this method and found differences between lung cancer and control patients [25], but is not included in the table because his findings are not focused on detecting lung cancer and he did not provide the requisite performance statistics.

Unfortunately, only the minority of original studies that we reviewed listed individual VOCs and provided specific information about their performance (N $=$ 4 for lung cancer, N $=$ 1 for breast cancer, and N $=$ 2 for colorectal cancer). Additional information was abstracted from review articles for 6 lung cancer articles, 2 breast cancer articles and 2 colorectal cancer articles. The majority of the 43 studies we reviewed utilized sensor-based electronic nose systems (eNoses) such as Cyranose [63] or SpiroNose [23], or described other types of experiments in which the VOCs were not listed.

Of the studies that did test specific VOCs, all were conducted in a case-control approach, in which diagnosed patients were compared with healthy controls, or controls with noncancerous conditions. Most studies enrolled small numbers of participants and several included multiple comparison groups. Overall, moderate to very good diagnostic performance was reported for distinguishing healthy controls from patients [5, 29]. However, reports to date indicate there is no single VOC that would clearly determine the presence of cancer with high performance. Instead, the occurrence of several VOCs in combination was used to increase both overall sensitivity and specificity of cancer diagnosis through breath testing [6, 29, 31, 42].

3.2 Lung cancer studies

Of the 19 lung cancer studies our literature review identified, only five included the names of specific VOCs identified in cancer patients, along with most of the statistics required to complete Table 1. In the study conducted by Corradi et al., subjects with suspected lung cancer were enrolled and VOCs were combined and compared for those with a subsequent diagnosis against those with no diagnosis [21]; Schumer and colleagues [46] and Li et al. [30] compared combined VOCs from lung cancer patients with those with benign disease and with healthy controls. Ligor et al. compared combined VOCs in lung cancer cases and controls [31]. Fernandes et al. provided a list of VOCs in his review article along with the cutoff levels, but without further details on individual articles [26]. The number of patients enrolled in the studies ranged from 12 [64] to 156 [46].

Table 3
Colorectal cancer compounds, performance, and the concentration gradient in cancer patients ${}^{\rm a}$

First author,publication year, N ${}^{\rm b}$	Volatile compound;method	Cut-off	Sens	Spec	AUC	Gradient
Altomare, 2013 [67]Training set: N $=$ 37 cases N $=$ 41 controlsValidation set: N $=$ 15 cases N $=$ 10 controls	Nonanal4-Methyl-2-pentanoneDecanal2-methylbutane1,2-pentadiene2-methylpentane3-methylpentaneMethylcyclopentaneCyclohexaneMethylcyclohexane1,3-dimethylbenzene4-methyloctane1,4-dimethylbenzeneA (4-methylundecane, RT $=$ 11 $\cdot$ 3)B (trimethyldecane, RT $=$ 13 $\cdot$ 2);TD-GC-MS	NP ${}^{\rm c}$	All comp86.0	All comp83.0	All comp0.852	NP
Altomare, 2015 [14]N $=$ 32 colorectal cancer patients disease free at follow-upN $=$ 55 controls	1,2-pentadiene Alkenes2-methylbutane Alkanes3-methylpentane AlkanesMethylcyclopentane AlkanesCyclohexane AlkanesDecanal AldehydesNonanal AldehydesMethylcyclohexane Alkanes4-methyl-2-pentanone Ketons1,4-dimethylbenzene Aromatics1,3-dimethylbenzene;TD-GC-MS	NP	All comp100.0	All comp97.92	All comp1	NP
Amal, 2016 [15]N $=$ 65 casesN $=$ 122 controls	AcetoneEthyl acetateEthanol4-methyl octane;GC-MS	999.6 ppb128.4 ppb464.7 ppb 19.1 ppb	All comp85.0	All comp94.0	NPNPNPNP	UpUpDownDown
Peng, 2010 [68]N $=$ 26 casesN $=$ 22 controls	1, 10- (1-butenylidene) bis benzene1,3-dimethyl benzene1-iodo nonaneAcetic acidBenzoic acid;SPME-GC-MS	NP	NP	NP	NP	NP

${}^{\rm a}$ References before 2015 from Markar, 2015 [32]. ${}^{\rm b}$ N $=$ number by case/control status. ${}^{\rm c}$ NP $=$ not provided/available in the published manuscript. Sens: sensitivity; Spec: specificity; AUC: area under the receiver operating characteristic curve; ppbv: parts per billion by volume; All comp: all compounds were tested together. Method: GC $=$ gas chromatography; MS $=$ mass spectrometry; SPME $=$ solid phase microextraction; TD $=$ thermal desorption.

Not all studies included the cut-off levels at which the VOC was considered to be abnormal, and not all included sensitivity, specificity, or AUC. A few did not indicate whether the VOC was higher or lower in cancer patients compared with healthy controls (gradient). The five studies prior to 2015 shown in Table 1 are from Krilaviciute et al. [29]. For some studies, the statistics on sensitivity, specificity and AUC was presented for all compounds combined, rather than each compound individually.

The studies not shown in Table 1 either did not provide a list of the VOCs identified, or did not provide any of the statistics required for this table. Four of these studies did list the VOCs [19, 38, 39, 65], but used different statistics to evaluate the accuracy of differences between patients and controls. Capuano et al. provided the percentage of identification for each of the VOCs [19], while Fu et al. utilized the P-value for the difference between lung cancer patients, controls, and benign nodules [65], Peralbo-Molina et al. used the P-value for the difference between lung cancer patients, smokers, and healthy controls [39], and Nardi-Agmon et al. provided the percentage of success in monitoring response to anticancer treatments, contrasting stable disease with partial response [38].

3.3 Breast cancer studies

In our review, along with the studies identified by Krilaviciute et al. [29], we identified an additional study evaluating individual VOCs for breast cancer as shown in Table 2. This study in 2015 by Barash [16] recruited 23 healthy women, 13 with benign breast conditions, 10 with ductal carcinoma in situ (DCIS), and 80 with malignant disease across five different subtypes. We report the VOC statistics for the healthy versus the malignant patients in Table 3. Study sizes for the breast cancer compounds ranged from 10 [66] to 80 cases [16].

One of the two studies we identified that reported on human breath compounds used to identify breast cancer did not list the VOCs, although they reported high sensitivity and specificity in their comparison of breast cancer patients with healthy controls [33].

3.4 Colorectal cancer studies

Both of the colorectal cancer studies identified in this review listed individual VOCs and provided statistics on performance (see Table 3). Krilaviciute et al.’s 2015 review [29] only identified the 2013 study from Altomare [67], which was also listed in a review by Markar [32]. The final study listed in Table 3 by Peng [68], was found in Markar et al.’s review [32]. All of the colorectal cancer compound studies contrasted patients diagnosed with colorectal cancer compared with healthy controls, except for Altomare et al., 2015 [14], which compared 32 colorectal cancer patients who were disease-free at follow-up with 55 healthy controls. Case size for the studies ranges from 26 [68] to 65 [15].

3.5 Review articles

Our review identified 20 systematic review or over- view articles, almost half of the 43 articles we identified through our search strategy. We read the full text of each review article, and identified any prior publications in the review article that included the VOCs for lung, colorectal or breast cancer. While several of the review articles provided VOCs found in the studies that they reviewed, only Krilaviciute et al. [29] and Markar et al. [32] named the VOCs as well as provided statistics on sensitivity, specificity, cutoffs, AUC, and gradient. The topics of the review articles are shown in Table 4.

4. Discussion

A challenge we encountered in conducting this review was that the majority of studies did not include full statistical details on the VOCs that were evaluated. We identified 19 studies through our search criteria that focused on lung cancer, but only five provided the VOCs with statistics on detection cutoff levels, sensitivity, specificity, AUC, and gradient, and some provided only a subset of these performance statistics. Only one of the two breast cancer studies provided these statistics, but both of the colorectal cancer studies listed the desired information.

Moreover, we encountered a variety of different methods for detecting and reporting on VOCs. Studies that utilized electronic noses (eNoses) to detect VOCs did not provide a list of the VOCs. The eNose experiments were based on the canine olfactory system, which seeks a pattern of organic chemicals, sometimes known as “breath prints”, rather than individual chemicals. ENoses are an assembly of nonspecific sensors which respond to different odors by producing a complex signal [6]. They have the advantage of being less expensive, easier to use, providing more rapid results, lower operating costs, and greater portability along with flexibility in the sensor array, which can be adjusted for specialized applications [54]. However, eNoses cannot identify individual compounds in complex mixtures [6, 69]. ENose studies also utilized a variety of statistics to verify the accuracy of their results, including the percentage correctly identified (e.g., [19]) or cross-validated values (e.g., [23]). Although GC-MS is considered the gold standard for identifying individual VOCs, this technique is also slow, cumbersome and expensive [6].

Table 4
Topics covered in review articles

First author,publication year	Type of review
Adiguzel, 2015 [13]	Aims to provide a rational basis for the development of breath sensors for lung cancer diagnostics. Reviews examples of many types of sensors, and discusses state of the art developments in the sensors field.
Boots, 2015 [5]	Aims to present the most recent concepts, data, and controversies regarding exhaled VOCs analysis in medicine. The currently available studies show moderate to good diagnostic accuracy, but require external validation.
Brodrick, 2015 [17]	Advances in ion mobility spectrometry have allowed rapid, reliable and robust analysis of exhaled breath. Presents a protocol for quickly and accurately collecting breath samples from 323 patients in 10 different secondary health care clinics.
Brooks, 2015 [18]	Focuses on: 1. The ability of canines to differentiate cancerous versus healthy individuals, 2. Electronic nose technology that uses clinical sensors to detect known biomarkers in exhaled breath.
Buljubasic, 2015 [4]	Summarizes publications from 2003 through 2013 related to the study of VOCs in human diseases.
Di Lena, 2016 [24]	Systematically reviews literature between 2000 and 2015 for VOCs using headspace gas chromatography as a potential screening tool for colorectal cancer. Concludes that the reliability of a metabolomic approach to screening for colorectal cancer is supported.
Fernandes, 2015 [26]	Reviews the various screening methods, including breath analysis using gold nanoparticles, to identify lung cancer. Finds this technique to be accurate and easy to operate. VOCs and cutoffs are listed in Table 1.
Krilaviciute, 2015 [29]	Systematic review of articles on detecting cancer through exhaled breath until April 30, 2015. Extremely thorough and detailed tables summarizing the studies and VOCs. Concludes that exhaled breath is promising, but standardization and larger studies are required.
Markar, 2015 [32]	Reviews VOC breath analysis in diagnosing and assessing endoluminal gastrointestinal disease, including colorectal cancer. Found that CRC could be distinguished from controls based on VOCs, but the variation in metabolites analyzed between studies prevented a reproducible diagnostic model.
Mathew, 2015 [35]	Reviews the technologies in the field of exhaled breath analysis. Discusses the various techniques to analyze VOCs, and discusses medical monitoring techniques for breath analysis.
Righettoni, 2015 [41]	Critically reviews advances in technologies for breath analysis. Summarizes breath markers related to different diseases. Discusses chemo-resistive gas sensors.
Saalberg, 2016 [42]	Performs a systematic literature search on VOC biomarkers in lung cancer. Provides tables: sampling and detection methods by author, identified biomarkers by author, and repeatedly identified biomarkers for lung cancer by number of occurrences. 33 VOCs were found by more than one research group. Five groups confirmed 2-butanone and 1-propanol as the best discriminators for lung cancer. No study considered a single substance sufficient.
Scarlata, 2015 [44]	Reviews the literature focusing on electronic nose technology. Discusses technical and methodological limitations in breath analysis.
Schmidt, 2015 [45]	Reviews the current challenges in the investigation of VOCs as potential cancer biomarkers. Discusses in vivo, ex vivo, and in vitro studies. Details techniques for VOCs collection, and focuses on VOCs identified both in cell tissue and in vitro.
Smith, 2015 [48]	Focuses on the experimental challenges presented by analysis of VOCs in exhaled breath. Discusses breath sampling and analysis procedures, instruments used for analysis, and data interpretation. Provides a table of individual VOCs from healthy people identified using SIFT-MS. Recommends using nose-exhaled breath to avoid contaminants.
Sun, 2016 [9]	Summarizes and critically reviews exhaled breath VOC sampling, collection, detection, and analytical methods. Provides an extensive list of VOCs from breath samples for various cancers and the techniques used in sampling (no performance statistics).
Trecate, 2016 [50]	Although a large number of potential breast cancer biomarkers have been discovered, their medical relevance and use in clinical settings remains to be validated.
Turner, 2016 [6]	Investigates techniques and issues in breath and clinical sample headspace analysis for diagnosing disease. Discusses sampling and handling considerations, techniques for analysis, background air contamination, and the need for standardization. Multiple compounds are more likely to be robust than individual ones.
Van der Schee, 2015 [52]	Overviews VOCs in lung disease, including lung cancer. Describes steps required prior to implementing VOC diagnostics in clinical practice, including considering comorbid conditions, standardizing breath collection and analysis procedures, and understanding the nature of VOC biomarkers.
Wilson, 2015 [54]	Summarizes recent developments in electronic-nose technologies. Provides a table of biomarker compounds associated with specific diseases (no performance statistics).

Several studies identified in our review focused on testing newly developed technology to distinguish the VOC pattern found in individuals with specific diseases compared with healthy people. For example, there are several reports that nanotechnology can be used to develop a simple, portable and effective screening device for lung cancer. Chemiresistors are one such example and they leverage monolayered gold nanoparticles for their biocompatibility and their ability to detect VOCs at very low concentrations [26]. Colorimetric sensors are another and use materials that can change color in response to their chemical environment. The sensors can be easily fabricated, and the colors can be read by taking a picture with a smart phone and using simple software. However, many VOCs could only be detected in the range of parts per million [9, 70]. Sun et al. and Adiguzel et al. provide excellent reviews of a variety of types of sensors [9, 13] that is beyond the scope of this review.

There are also several technical hurdles involved in taking, handling, storing and analyzing breath samples for VOC analysis [5, 6]. The method used for exhaling the breath can strongly affect the VOC mixtures [5] and the ambient air which the subject has been breathing may affect what chemicals are identified in their breath [5]. Patients diagnosed with lung cancer are frequently smokers. Therefore, tobacco smoke constituents are identified in breath [31]. There is currently no standard method for collecting breath, with some studies analyzing the ambient air and comparing it with the chemicals from the breath sample, and others using “lung washout” before breath sampling [29] which usually involves participant burden and clinical assistance. In addition, some breath analysis techniques require storing the sample for a short time prior to analysis. Using Tedlar bags, one of the most common storage techniques, can lead to some contaminants leaching into the breath sample if it is stored too long. If sample storage is required, thermal desorption tubes can help avoid confounding compounds [29].

Data analysis also varied widely between studies [6, 71, 72], making replication of results between studies difficult. In future work, internal validation by random sample split or leave-one-out cross-validation is recommended to get closer to estimates that are comparable [29].

Although we used a comprehensive search string in two independent databases, it is possible that we missed relevant studies. The wide variation in breath collection, storage, analysis and reporting in the studies makes summarization of data challenging. We included only studies that evaluated the exhaled breath of cancer patients compared with controls, excluding in vitro studies and studies with canines. We also focused our presentation and discussion on those studies that reported VOCs along with performance statistics, which was a subset of the studies reviewed.

5. Conclusion

Breath analysis for cancer screening and early detection shows promise, because technology measurement continues to advance and it allows breath samples to be collected easily, safely, and frequently [6, 9, 29]. This review focused on specific VOCs evaluated in human studies with respect to lung, breast, or colorectal cancer published in in the past five years. We noted specific VOCs correlated with these cancers, along with the cutoff levels at which they were detected, the sensitivity, specificity and AUC.

Gas chromatography-mass spectrometry (GC-MS) is considered the gold standard for identifying specific VOCs. However, the science of breath analysis has recently moved into analyzing patterns of VOCs using a variety of different multiple sensor techniques, such as eNoses and nanomaterials. Additional development of breath analysis for early cancer detection will require standardized breath sampling methods and innovative techniques for VOC detection and analysis. As technology advances and clinical studies are conducted breath analysis holds promise not only for detecting cancer early, but also for monitoring the response to cancer treatment and potentially early detection of recurrence when secondary treatment efficacy is at its highest. Although several authors felt that breath analysis requires further development and testing before integration into clinical practice, Brodrick et al. report on a protocol which was successfully used for quickly and accurately collecting breath samples from 323 respiratory patients recruited in 10 different secondary health care clinics [17]. Given the growing body of work and the compelling reasons to detect cancer and its recurrence early, further clinical work validating novel technologies and approaches is warranted.

Footnotes

Acknowledgments

This research was supported by Nano Engineered Applications (NEA).

References

Adams

, Hippocratic writings: Aphorisms IV, V, In the Internet Classic Archive, New York, New York (1994).

Pauling

et al., Quantitative analysis of urine vapor and breath by gas-liquid partition chromatography, Proc Nat Acad Sci USA 68 (1971), 2374–2376.

Williams

et al., Cinema audiences reproducibly vary the chemical composition of air during films, by broadcasting scene specific emissions on breath, Sci Rep 6 (2016), 25464.

Buljubasic

and Buchbauer

, The scent of human diseases: A review on specific volatile organic compounds as diagnostic biomarkers, Flavour and Fragrance Journal 30(1) (2015), 5–25.

Boots

A.W.

et al., Exhaled molecular fingerprinting in diagnosis and monitoring: Validating volatile promises, Trends in Molecular Medicine 21(10) (2015), 633–644.

Turner

, Techniques and issues in breath and clinical sample headspace analysis for disease diagnosis, Bioanalysis 8(7) (2016), 677–690.

Zonta

et al., Detection of colorectal cancer biomarkers in the presence of interfering gases, Sensors and Actuators B-Chemical 218 (2015), 289–295.

ACS, Cancer facts and figures 2016, Atlanta, Georgia (2016).

Sun

Shao

and Wang

, Detection of volatile organic compounds (VOCs) from exhaled breath as noninvasive methods for cancer diagnosis, Analytical and Bioanalytical Chemistry 408(11) (2016), 2759–2780.

10.

Hakim

et al., Volatile organic compounds of lung cancer and possible biochemical pathways, Chemical Reviews 112(11) (2012), 5949–5966.

11.

Boots

A.W.

et al., The versatile use of exhaled volatile organic compounds in human health and disease, Journal of Breath Research 6(2) (2012).

12.

Tisch

and Haick

, Arrays of chemisensitive monolayer-capped metallic nanoparticles for diagnostic breath testing, Reviews in Chemical Engineering 26(5–6) (2010), 171–179.

13.

Adiguzel

and Kulah

, Breath sensors for lung cancer diagnosis, Biosensors & Bioelectronics 65 (2015), 121–138.

14.

Altomare

D.F.

et al., Effects of curative colorectal cancer surgery on exhaled volatile organic compounds and potential implications in clinical follow-up, Annals of Surgery 262(5) (2015), 862–867.

15.

Amal

et al., Breath testing as potential colorectal cancer screening tool, International Journal of Cancer 138(1) (2016), 229–236.

16.

Barash

et al., Differentiation between genetic mutations of breast cancer by breath volatolomics, Oncotarget 6(42) (2015), 44864–44876.

17.

Brodrick

et al., Breath analysis: Translation into clinical practice, Journal of Breath Research 9(2) (2015).

18.

Brooks

S.W.

et al., Canine olfaction and electronic nose detection of volatile organic compounds in the detection of cancer: A review, Cancer Investigation 33(9) (2015), 411–419.

19.

Capuano

et al., The lung cancer breath signature: A comparative analysis of exhaled breath and air sampled from inside the lungs, Scientific Reports 5 (2015).

20.

Chatterjee

Castro

and Feller

J.F.

, Tailoring selectivity of sprayed carbon nanotube sensors (CNT) towards volatile organic compounds (VOC) with surfactants, Sensors and Actuators B-Chemical 220 (2015), 840–849.

21.

Corradi

et al., Exhaled breath analysis in suspected cases of non-small-cell lung cancer: A cross-sectional study, Journal of Breath Research 9(2) (2015).

22.

Daniel

D.A.P.

Thangavel

and Rajakeerthana

K.T.

, Empirical study on early detection of lung cancer using breath analysis, in: Innovations in Information, Embedded and Communication Systems (ICIIECS), 2015 International Conference, IEEE (2015), 1–6.

23.

de Vries

et al., Integration of electronic nose technology with spirometry: Validation of a new approach for exhaled breath analysis, Journal of Breath Research 9(4) (2015).

24.

Di Lena

Porcelli

and Altomare

D.F.

, Volatile organic compounds (VOC) as new biomarkers for colorectal cancer: A review, Colorectal Dis (2016).

25.

Feinberg

et al., Cancerous glucose metabolism in lung cancer-evidence from exhaled breath analysis, J Breath Res 10(2) (2016), 026012.

26.

Fernandes

M.P.

Venkatesh

and Sudarshan

B.G.

, Early detection of lung cancer using nano-nose – A review, The Open Biomedical Engineering Journal 9 (2015), 228–233.

27.

Nantz

and Bousamra

, Non-invasive method useful for detecting lung cancer disease, e.g., small cell lung cancer state, by obtaining exhaled breath, forming adducts, quantifying adduct, and comparing obtained value with healthy specimen value, in Fu

Nantz

and Bousamra

; Univ Louisville Res Found Inc.

28.

Gasparri

et al., Volatile signature for the early diagnosis of lung cancer, Journal of Breath Research 10(1) (2016).

29.

Krilaviciute

et al., Detection of cancer through exhaled breath: A systematic review, Oncotarget 6(36) (2015), 38643–38657.

30.

M.X.

et al., Breath carbonyl compounds as biomarkers of lung cancer, Lung Cancer 90(1) (2015), 92–97.

31.

Ligor

Pater

and Buszewski

, Application of an artificial neural network model for selection of potential lung cancer biomarkers, Journal of Breath Research 9(2) (2015).

32.

Markar

S.R.

et al., Exhaled breath analysis for the diagnosis and assessment of endoluminal gastrointestinal diseases, Journal of Clinical Gastroenterology 49(1) (2015), 1–8.

33.

Sinues

P.M.L.

et al., Secondary electrospray ionization-mass spectrometry and a novel statistical bioinformatic approach identifies a cancer-related profile in exhaled breath of breast cancer patients: A pilot study, Journal of Breath Research 9(3) (2015), 031001.

34.

Masuda

et al., SnO2 nanosheet/nanoparticle detector for the sensing of 1-nonanal gas produced by lung cancer, Scientific Reports 5 (2015).

35.

Mathew

T.L.

et al., Technologies for clinical diagnosis using expired human breath analysis, Diagnostics 5(1) (2015), 27–60.

36.

Mazzone

P.J.

et al., Progress in the development of volatile exhaled breath signatures of lung cancer, Annals of the American Thoracic Society 12(5) (2015), 752–757.

37.

McWilliams

et al., Sex and smoking status effects on the early detection of early lung cancer in high-risk smokers using an electronic nose, IEEE Transactions on Biomedical Engineering 62(8) (2015), 2044–2054.

38.

Nardi-Agmon

et al., Exhaled breath analysis for monitoring response to treatment in advanced lung cancer, Journal of Thoracic Oncology: Official Publication of the International Association for the Study of Lung Cancer 11(6) (2016), 827–837.

39.

Peralbo-Molina

et al., Metabolomics analysis of exhaled breath condensate for discrimination between lung cancer patients and risk factor individuals, Journal of Breath Research 10(1) (2016).

40.

Phillips

et al., Blinded validation of breath biomarkers of lung cancer, a potential ancillary to chest CT screening, Plos One 10(12) (2015).

41.

Righettoni

Amann

and Pratsinis

S.E.

, Breath analysis by nanostructured metal oxides as chemo-resistive gas sensors, Materials Today 18(3) (2015), 163–171.

42.

Saalberg

and Wolff

, VOC breath biomarkers in lung cancer, Clin Chim Acta 459 (2016), 5–9.

43.

Santini

et al., Electronic nose and exhaled breath NMR-based metabolomics applications in airways disease, Current Topics in Medicinal Chemistry 16(14) (2016), 1610–1630.

44.

Scarlata

et al., Exhaled breath analysis by electronic nose in respiratory diseases, Expert Review of Molecular Diagnostics 15(7) (2015), 933–956.

45.

Schmidt

and Podmore

, Current challenges in volatile organic compounds analysis as potential biomarkers of cancer, Journal of Biomarkers 2015 (2015), 981458.

46.

Schumer

E.M.

et al., High sensitivity for lung cancer detection using analysis of exhaled carbonyl compounds, Journal of Thoracic and Cardiovascular Surgery 150(6) (2015), 1517–1522.

47.

S.T.

et al., Application of SAW gas chromatography in the early screening of lung cancer, in: Piezoelectricity, Acoustic Waves, and Device Applications (SPAWDA), 2015 Symposium, IEEE (2015), 22–25.

48.

Smith

and Spanel

, Pitfalls in the analysis of volatile breath biomarkers: Suggested solutions and SIFT-MS quantification of single metabolites, Journal of Breath Research 9(2) (2015).

49.

Tzeng

T.H.

et al., A portable micro gas chromatography system for volatile compounds detection with 15 ppb of sensitivity, in: Solid-State Circuits Conference (ISSCC), IEEE (2015), 1–3.

50.

Trecate

Sinues

P.M.L.

and Orlandi

, Noninvasive strategies for breast cancer early detection, Future Oncology 12(11) (2016), 1395–1411.

51.

Tzeng

T.H.

et al., Portable micro gas chromatography system for lung cancer associated volatile organic compound detection, IEEE Journal of Solid-State Circuits 51(1) (2016), 259–272.

52.

van der Schee

M.P.

et al., Breathomics in lung disease, Chest 147(1) (2015), 224–231.

53.

van Hooren

M.R.

et al., Differentiating head and neck carcinoma from lung carcinoma with an electronic nose: A proof of concept study, Eur Arch Otorhinolaryngol (2016).

54.

Wilson

A.D.

, Advances in electronic-nose technologies for the detection of volatile biomarker metabolites in the human breath, Metabolites 5(1) (2015), 140–163.

55.

Zhao

et al., Detection of mixed volatile organic compounds and lung cancer breaths using chemiresistor arrays with crosslinked nanoparticle thin films, Sensors and Actuators B-Chemical 232 (2016), 292–299.

56.

van de Kant

K.D.

et al., Clinical use of exhaled volatile organic compounds in pulmonary diseases: A systematic review, Respir Res 13 (2012), 117.

57.

National Institute of Health Research, PROSPERSO, International Prospective Register of Systematic Reviews, Centre for Reviews and dissemination, University of York, York, UK.

58.

Moher

et al., Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement, Plos Med 6(7) (2009), e1000097.

59.

Crowd Company, Nano Engineered Applications, Inc., San Diego, California, 2017. www.neapplications.com, Accessed 9 January 2017.

60.

Van Berkel

J.J.

et al., A profile of volatile organic compounds in breath discriminates COPD patients from controls, Respir Med 104(4) (2010), 557–563.

61.

Kushch

et al., Compounds enhanced in a mass spectrometric profile of smokers’ exhaled breath versus non-smokers as determined in a pilot study using PTR-MS, J Breath Res 2(2) (2008), 026002.

62.

Rieder

et al., Analysis of volatile organic compounds: Possible applications in metabolic disorders and cancer screening, Wien Klin Wochenschr 113(5–6) (2001), 181–185.

63.

Sensigent, Sensigent intelligent sensing solutions, in.

64.

Fuchs

et al., Breath gas aldehydes as biomarkers of lung cancer, Int J Cancer 126(11) (2010), 2663–2670.

65.

X.A.

et al., Noninvasive detection of lung cancer using exhaled breath, Cancer Med 3(1) (2014), 174–181.

66.

Mangler

et al., Volatile organic compounds (VOCs) in exhaled breath of patients with breast cancer in a clinical setting, Ginekol Pol 83(10) (2012), 730–736.

67.

Altomare

D.F.

et al., Exhaled volatile organic compounds identify patients with colorectal cancer, Br J Surg 100(1) (2013), 144–150.

68.

Peng

et al., Detection of lung, breast, colorectal, and prostate cancers from exhaled breath using a single array of nanosensors, Br J Cancer 103(4) (2010), 542–551.

69.

Wilson

A.D.

and Baietto

, Applications and advances in electronic-nose technologies, Sensors (Basel) 9(7) (2009), 5099–5148.

70.

Mazzone

P.J.

et al., Exhaled breath analysis with a colorimetric sensor array for the identification and characterization of lung cancer, J Thorac Oncol 7(1) (2012), 137–142.

71.

Song

et al., Quantitative breath analysis of volatile organic compounds of lung cancer patients, Lung Cancer 67(2) (2010), 227–231.

72.

Hauschild

A.C.

Baumbach

J.I.

and Baumbach

, Integrated statistical learning of metabolic ion mobility spectrometry profiles for pulmonary disease identification, Genet Mol Res 11(3) (2012), 2733–2744.

73.

Amal

et al., Assessment of ovarian cancer conditions from exhaled breath, International Journal of Cancer 136(6) (2015), E614–E622.

74.

Van Le

, Assessment of ovarian cancer conditions from exhaled breath editorial comment, Obstetrical & Gynecological Survey 70(2) (2015), 89–91.

75.

Kahn

et al., Dynamic nanoparticle-based flexible sensors: Diagnosis of ovarian carcinoma from exhaled breath, Nano Letters 15(10) (2015), 7023–7028.

76.

Kumar

et al., Mass spectrometric analysis of exhaled breath for the identification of volatile organic compound biomarkers in esophageal and gastric adenocarcinoma, Annals of Surgery 262(6) (2015), 981–990.

77.

Zou

et al., Exhaled gases online measurements for esophageal cancer patients and healthy people by proton transfer reaction mass spectrometry, J Gastroenterol Hepatol (2016).

78.

Amal

et al., Detection of precancerous gastric lesions and gastric cancer through exhaled breath, Gut 65(3) (2016), 400–407.

79.

Guo

et al., Exhaled breath volatile biomarker analysis for thyroid cancer, Translational Research 166(2) (2015), 188–195.

80.

Aggio

R.B.M.

et al., The use of a gas chromatography-sensor system combined with advanced statistical methods, towards the diagnosis of urological malignancies, Journal of Breath Research 10(1) (2016).

81.

Shoffel-Havakuk

et al., Increased number of volatile organic compounds over malignant glottic lesions, Laryngoscope (2015).

82.

Handa

et al., Exhaled breath analysis for lung cancer detection using ion mobility spectrometry, Plos One 9(12) (2014), e114555.

83.

Wang

et al., The analysis of volatile organic compounds biomarkers for lung cancer in exhaled breath, tissues and cell lines, Cancer Biomark 11(4) (2012), 129–137.

84.

Zou

et al., Optimization of volatile markers of lung cancer to exclude interferences of non-malignant disease, Cancer Biomark 14(5) (2014), 371–379.

85.

et al., Investigation of potential breath biomarkers for the early diagnosis of breast cancer using gas chromatography-mass spectrometry, Clin Chim Acta 436 (2014), 59–67.

Breath based volatile organic compounds in the detection of breast,lung,and colorectal cancers: A systematic review

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Methods

2.1 Protocol and registration

2.2 Eligibility criteria

2.3 Information sources and search process

2.5 Data collection process

Table 1 Lung cancer compounds, performance, and the concentration gradient in cancer patients compared with controls a

3.1 Summary

3.2 Lung cancer studies

Table 3 Colorectal cancer compounds, performance, and the concentration gradient in cancer patients a

3.4 Colorectal cancer studies

3.5 Review articles

4. Discussion

Table 4 Topics covered in review articles

Footnotes

Acknowledgments

References

Table 1
Lung cancer compounds, performance, and the concentration gradient in cancer patients compared with controls ${}^{\rm a}$

Table 3
Colorectal cancer compounds, performance, and the concentration gradient in cancer patients ${}^{\rm a}$

Table 4
Topics covered in review articles