Analytical and unconventional methods of cancer detection using odor

Trends in Analytical Chemistry, Vol. 38, 2012
Trends
Analytical and unconventional
methods of cancer detection
using odor
Bogusław Buszewski, Joanna Rudnicka, Tomasz Ligor, Marta Walczak,
Tadeusz Jezierski, Anton Amann
Finding a non-invasive, painless and simple screening method for early detection of cancer is a desirable goal. There is evidence
that volatile organic compounds (VOCs) detectable in exhaled air and producing specific breath odor could be taken into
consideration as possible cancer markers. Chemical analysis of VOCs in the breath that utilizes gas chromatography-mass
spectrometry or an array of specific sensors (an electronic nose) could be useful for cancer screening. Preliminary reports show
that canines, due to their extraordinary sense of smell and ability to perform well with operant conditioning, could also be used in
the future as biological screeners for different forms of cancer. However, the question remains open whether specific VOCs or
breath odors appear at very early (sub-clinical) stages of cancer disease or only at advanced stages of the disease during tumor
decomposition.
ª 2012 Elsevier Ltd. All rights reserved.
Keywords: Breath analysis; Cancer marker; Cancer screening; Carcinoma; Electronic nose; Exhaled air; Gas chromatography-mass spectrometry
(GC-MS); Odor; Sniffer dog; Volatile organic compound (VOC)
1. Introduction
Bogusław Buszewski*, Joanna Rudnicka, Tomasz Ligor
Department of Environmental Chemistry and Bioanalytics, Faculty of Chemistry,
Nicolaus Copernicus University,
7 Gagarin St., 87-100 Toruń, Poland
Marta Walczak, Tadeusz Jezierski
Institute of Genetics and Animal Breeding of Polish Academy of Sciences,
Department of Animal Behaviour, Jastrze˛biec,
05-552 Wólka Kosowska, Poland
Anton Amann
Breath Research Institute of the Austrian Academy of Sciences, Rathausplatz 4,
A-6850 Dornbirn, Austria
Department of Anesthesiology and General Intensive Care, Innsbruck Medical University,
Anichstrasse 35, A-6020 Innsbruck, Austria
*
Corresponding author.
Tel.: +48 566114308;
Fax: +48 566114837;
E-mail: [email protected]
Early diagnosis of cancer with effective
screening methods is crucial for successful
therapy. Ultimately, diagnosis of cancer
disease can be made only on the basis of
biopsy and histopathological examination
of the tissue or cells, whereas cancer
screening, which is based on different
imaging methods [e.g., X-ray, ultrasound,
computer tomography (CT), positron
emission tomography (PET), mammography, or detection of cancer markers) can
only select the patients suspected of having cancer, who need to be examined
further by biopsy and histopathology.
In recent years, many studies on different screening methods have been undertaken. A desirable screening method
should be non-invasive, painless, inexpensive and easily accessible to a large
number of patients. Yet, above all, to be
useful, a screening method should be
reliable and facilitate diagnosis of an earlystage cancer, rendering use of additional
invasive methods unnecessary. Breath
analysis, which has numerous advantages
0165-9936/$ - see front matter ª 2012 Elsevier Ltd. All rights reserved. doi:http://dx.doi.org/10.1016/j.trac.2012.03.019
1
Trends
in comparison with traditional screening methods, has
thus come into the spotlight [1–3].
In the 1970s, Linus Pauling found around 200 volatile
organic compounds (VOCs) in exhaled breath at concentrations as low as parts per million by volume to parts per
trillion by volume (ppmv–pptv) [4,5]. VOCs are products of
metabolic processes in the human body. However, biochemical pathways of most of the VOCs detected in the
human breath still have not been precisely elucidated.
There may be different sources for organic compounds
found in the breath; they may appear due to endogenous
metabolic processes or be of external origin (e.g., pollutants
inhaled from the environment, or compounds ingested in
the food and released within the gastrointestinal tract) [4].
This process involves various coupled reactions (e.g.,
enzymatic oxidation or autooxidation), all of which have
an impact on the metabolic pathways of producing some
compounds, which are often precursors of free radicals.
It is known that the composition of breath varies
depending on the disease. For example, a sweetish, acetone-like odor indicates diabetes, while the odor of rotten
eggs caused by organic sulfide and thiol compounds
suggests liver problems [6–8]. Currently, intensive studies
aiming to identify compounds that could be markers of
cancer are being carried out. Substances suitable for
cancer markers may be low-mass molecules [e.g., VOCs
(e.g., hydrocarbons, alcohols, aldehydes, ketones, or
nitrogen- and sulfur-containing compounds), macromolecules of proteins, proteins with carbohydrate or lipid
components, glycolipids and nucleic acids can be taken
into consideration as well [2,9–11]. VOCs are usually
present in breath at very low concentrations, so it is
necessary to enrich them before analysis. The most
common method of enrichment of VOCs is solid-phase
microextraction (SPME) [12–14] and sorption on solid
sorbents followed by thermal desorption (TD) [15,16].
Enriched VOCs can be subsequently analyzed by gas
chromatography (GC) or GC-mass spectrometry (GC-MS)
[17,18]. Other promising methods are selected ion flow
tube MS (SIFT-MS) [19,20], proton-transfer MS (PTR-MS)
[21,22] as well as sensor technology [an electronic nose
(e-nose)] [23,24]. Although chromatographic methods
seem to be the leading tool in the analysis of breath as they
obtain a lot of information about the molecular composition of the exhaled breath, some unconventional
methods for detection of VOCs that could be cancer
markers are worth considering.
In this article, we describe relevant instrumental
methods for identification of VOCs and an unconventional biodetection method utilizing trained sniffer dogs.
2. Carcinomas as a human health problem
Currently, the frequency of cancer malignancies is
increasing worldwide. Cardiovascular problems and
2
http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 38, 2012
cancer diseases are the two most common causes of
death. In men, more than 50% of all cancers diagnosed
are lung, prostate and colorectal cancers. In women,
breast, lung and colorectal cancers prevail, estimated to
account for more than 50% of carcinomas. In both men
and women, the greatest mortality rate is for lung cancer
with 31% and 26%, respectively, of all cases in 2008
[25]. Of particular importance is that currently no
effective screening method is available for lung cancer.
Although, the CT scan of pulmonary parenchyma might
be an effective screening method, it is quite expensive.
It is estimated that cigarette smoke has been responsible for lung cancer in 85-90% of sufferers. This is not
surprising since more than 4000 chemicals, including
nearly 200 mutagenic and carcinogenic substances,
were found in cigarette smoke. Apart from carbon
dioxide and carbon monoxide that are the major components of the cigarette smoke, highly carcinogenic
polycyclic aromatic hydrocarbons, phenols, aldehydes,
ketones, aromatic amines, volatile alkenes, N-nitrosamines, metals and other organic compounds were
identified. The N-nitrosamines found in the cigarette
smoke can be divided into two groups:
(1) tobacco-specific non-volatile nitrosamines, present
both in tobacco and in smoke (from the decomposition of alkaloids); and,
(2) non-volatile and semi-volatile tobacco-specific
nitrosamines found in steam as well as in smoke.
These compounds may be precursors for free radicals,
which contribute qualitatively and quantitatively to
oxidative stress [26,27].
Another common initiating cause of lung cancer is
exposure to radon, cadmium, arsenic, beryllium or
asbestos [27–29].
Clinical cancers are divided into benign and malignant
tumors. Benign tumors comprise slow-growing, diverse
cells, whereas malignant tumors are characterized by
rapid growth and formation of metastases. Transition
from normal to tumor cells is a multistage process. In the
first stage, the initiation of cancer, the DNA of a single
cell is damaged. Environmental factors [e.g., chemicals
(carcinogens)], physical factors (UV radiation) or oncogenic viruses influence progress of the disease. In the
promotion stage, uncontrolled growth of cancer cells
occurs. Tumor progression is the irreversible stage
leading to tumor formation [30,31].
In recent years, intensive research was performed on
utilizing potential volatile markers for early detection of
cancer. This was done by applying chromatographic
and direct mass-spectrometric (MS) methods for identification of these compounds. Methods based on volatile markers in breath may detect early-stage cancer.
These are non-invasive, painless and harmless to the
patient. Table 1 lists typical analytical methods that
can be applied for determination of specific volatile
markers.
Trends in Analytical Chemistry, Vol. 38, 2012
Trends
Table 1. Exemplary data of VOCs analysis, depending on the sample, detection methods, and limits of detection (LOD) and quantification (LOQ)
Compounds
Sample/Type of cancer Technique for detection of VOCs
Benzene
Styrene
Propyl benzene
Decane
Undecane
Propanal
Butanal
Pentanal
Hexanal
Heptanal
Octanal
Nonanal
Pentane
Octane
Nonane
Decane
Undecane
Acetone
Benzene
Isoprene
Hexanal
Heptanal
Propyl benzene
Acetone
Hexanal
Heptanal
Nonane
5-Methyltridecane
3-Methylundecane
6-Methylpentadecane
2-Methylpropane
3-Methylnonadecane
4-Methyldodecane
2-Methyloctane
Butane
3-Methyltridecane
7-Methyltridecane
4-Methyloctane
3-Methylhexane
Heptane
2-Methylhexane
Pentane
5-Methyldecane
LOD
LOQ
Ref.
0.25 ng/mL
1.26 ng/mL
0.067 ng/mL
0.012 ng/mL
0.027 ng/mL
1 pmol/L
0.84 ng/mL
4.20 ng.mL
0.23 ng/mL
0.04 ng/mL
0.08 ng/mL
3 pmol/L
[5]
0.56 lg/L
0.62 lg/L
0.34 lg/L
0.28 lg/L
0.23 lg/L
1.96 lg/L
0.81 lg/L
0.91 lg/L
1.29 lg/L
1.31 lg/L
6.50 lg/L
ns
ns
ns
ns
[33]
ns
[11]
Breath/lung
SPME-GC-FID
Breath/lung
SPME-GC-MS
Breath/lung
HS-SPMEGC-TOFMS
Blood/lung
HS-SDME-GC-MS
Breath/breast
TD-GC-MS
0.18 lg/L
0.20 lg/L
0.11 lg/L
0.09 lg/L
0.08 lg/L
0.65 lg/L
0.27 lg/L
0.30 lg/L
0.43 lg/L
0.43 lg/L
2.15 lg/L
0.62 nmol/L
0.24 nmol/L
0.32 nmol/L
ns
Breath/lung
TD-GC-MS
ns
[32]
[34]
[35]
ns, not specified.
3. Analytical methods
3.1. Sampling
Breath could be collected into canisters, syringes, sorbent
tubes or Tedlar bags. Canisters (1000–3000 dm3) are
made of stainless steel and the surface of canister is inert.
Tedlar bags are made of polyvinyl fluoride (PVF). This
material is chemically rather inert to a wide range of
compounds and is relatively resistant to adsorption
of molecules on its surface. Nevertheless, permeation of
certain substances (notably water) should not be
underestimated. Tedlar bags can be reused for most
applications. Prior to reuse, the bags must be evacuated
and thoroughly cleaned and flushed with purified air or
nitrogen [6,7,19].
The collection of exhaled air is a significant issue in
breath analysis. Human breath comprises a mixture
alveolar air, which has been in contact with blood inside
alveoli, and ambient air, which is retained in the dead
space, including mouth, nose, pharynx, trachea and
bronchi. During sample collection, dead-space air is pulled
into the sampling system, diluting the alveolar gas. The
concentrations of endogenous compounds in alveolar air
are greater than found in mixed expiratory samples and
they will increase at the end of expiration when the end
tidal pressure of expired CO2 reaches a plateau [6,7,36].
http://www.elsevier.com/locate/trac
3
Trends
Trends in Analytical Chemistry, Vol. 38, 2012
3.2. Methods of preconcentrating VOCs in breath
samples
The most frequently applied method of enriching VOCs
in breath analysis is solid-phase extraction (SPE) followed by TD. The analytes can be released from the trap
and identified by GC [12,17,37–41]. The multisorbent
tubes have to be arranged in order of increasing strength
during sampling, and the flow has to be reversed for
desorption in order to prevent contamination [39].
Among adsorbents in collection traps, the most frequently used are carbon molecular sieves, different types
of graphitized carbon and organic polymers [6,40].
Sorption traps are applied for analysis of VOCs in human
breath [38,39].
SPME on coated fused-silica fibers was developed by
Arthur and Pawliszyn in 1990 [42]. This technique is
used for enrichment and separation of analytes from
various matrices. It is a fast, inexpensive, solvent-free
method offering the possibility of automation. SPME is
applied in the analysis of polar, non-polar, semi-volatile
and volatile compounds, which subsequently can be
detected by GC [42]. This method is used to enrich
samples for the identification of compounds in human
breath {e.g., isoprene [11], acetone [12], alkanes or
aromatic hydrocarbons (benzene, styrene) [5]}.
SPME involves the extraction of the analyzed compounds from the gas or liquid phase to the adsorption or
absorption layer deposited on the extraction fiber. Subsequently, it is immersed directly in the research medium
for a specified period, from a few to a dozen-plus minutes.
After completion of the exposure, the device is removed,
and the fiber is introduced into the injector of the gas
chromatograph. Afterwards the analytes are released by
TD, and are introduced in the stream of carrier gas into
the column [43–51].
Selection of the fiber must take into account the
properties of the compound (e.g., polarity, molecular
mass or estimated range of concentrations). The most
commonly applied sorbents are carboxen/poly-
dimethylsiloxane (Car/PDMS), 75 lm, to determination
of isoprene [11], polydimethylsiloxane/divinylbenzene
(PDMS/DVB), 65 lm, to identification of acetone [12],
and polydimethylsiloxane (PDMS), 100 lm, used for
analysis of alkanes and aromatic hydrocarbons in
human breath [32].
As an alternative to commercially-available SPME
coatings, fiber-based custom materials can be used.
Conductive polymers have found application in gas
sensors and biochemical analyses (e.g., identifying acetone in the breath of patients with diabetes) [52–54].
Polymers are synthesized during electrochemical or
chemical polymerization. For the first method, counter
and reference electrodes (e.g., made of platinum) are put
into solution containing monomer and electrolyte (the
dopant) diluted in a solvent. After applying a suitable
voltage, the polymer film immediately starts to form on
the working electrode. In the second method, chemical
polymerization, monomers react with an excess amount
of an oxidant in a suitable solvent (acid). The polymerization takes place spontaneously and requires constant
stirring [54,55].
An alternative method to SPE and SPME is the needletrap device (NTD), which is an extraction trap containing a sorbent (Tenax, Carboxen) within a needle. The
main advantages of NTDs include sensitivity, stability,
ease and rapidity of analysis, and very low amounts of
sorbent (less than 1 mg DVB or Carboxen) [56–61]. The
NTD is applied to identify VOCs in human breath [60,61]
or water [62,63]. Various preconcentration methods of
VOCs are presented in Table 2.
3.3. GC-MS
GC-MS is the method most widely applied for the determination of mixtures of VOCs and semi-VOCs in different
matrices. The technique is characterized by high sensitivity and enables qualitative analysis and use of retention time and Kovats index as well as quantitative
analysis [12,64,65]. MS is one of the techniques
Table 2. Application of various preconcentration techniques to breath analysis
Compounds
Isoprene
Benzene
Isoprene
Acetone
Isoprene
Dimethyl sulfide
2-Butenal
Hexane
Pentane
Hexanal
Pentanal
Acetone
Type of sample
Breath
TD-GC
Breath
SPME-GC-MS
Breath
NTD-GC-MS
ns, not specified.
4
Technique for detection of VOCs
http://www.elsevier.com/locate/trac
LOD
ns
0.3 lg/m3
0.25 nmol/L
0.049 ppb
0.4 ng/L
0.5 ng/L
0.9 ng/L
1.0 ng/L
1.2 ng/L
2.3 ng/L
5.3 ng/L
8.3 ng/L
LOQ
Ref.
ns
ns
ns
ns
ns
[16]
[37]
[13]
[14]
[60]
Trends in Analytical Chemistry, Vol. 38, 2012
Trends
characterized by low limit of detection (LOD), which allows identification and quantitation of VOCs present in
trace amounts in human breath [66–69].
In MS, typically used are the quadrupole analyzer, the
ion-trap analyzer and the time-of-flight (TOF) analyzer.
The quadrupole analyzer is characterized by durability,
small size, low cost and ease of use [20]. The drawbacks
are that its range of masses separated is approximately
1000 Da and that it is able to separate ions with masses
differing by a unit [69]. Another example is the ion-trap
analyzer, which is very similar to the quadrupole analyzer. The advantage of this analyzer, when compared
with the quadrupole, is higher sensitivity and tandem
MS (MS2) experiments, especially multiple reaction
monitoring (MRM), since it operates at low pressure of
helium (about 0.1 Pa) in the chamber traps [69]. The
disadvantage of ion traps are the spectra often being
modified as a result of chemical ionization, and difficulties in interpreting the spectra for polar compounds (e.g.,
alcohols, aldehydes, and ketones), which are present in
the exhaled air [6].
The TOF analyzer identifies molecules in a sample by
measuring their TOF through a field-free TOF tube. TOFMS provides fast data through multiple scans, and is
characterized by high resolution. Deconvolution can be
used for chromatographic separation of co-eluted substances and obtaining the spectrum for each compound
[6,69].
3.4. Real-time measurements of VOCs in biological
samples
Selected ion-flow tube-MS (SIFT-MS) is another technique for the rapid detection and quantitative determination of gas concentrations in exhaled air. SIFT-MS
uses different primary ions (H3O+, NO+ and O2+) to detect various volatile compounds (e.g., acetone, ethanol,
acetaldehyde, ammonia or water vapor), measured at
parts per billion (ppb) concentration levels [70–73]. In
human breath, acetonitrile (in smokers and non-smokers) [19], formaldehyde, acetaldehyde and propanol
were detected as potential cancer markers [74]. In
SIFT-MS, positive ions are formed in the ion source by
electron ionization or microwave discharge. An auxiliary
quadrupole mass filter selects only primary ions with a
specific mass-to-charge ratio (H3O+, NO+ or O2+), which
are then placed in the fast-flowing inert carrier gas,
usually helium. The sample is added to the inert gas
containing the precursor ions moving along the drift
tube. Ionic products are formed by the reaction of the
primary ions with trace amounts of gases present in the
sample [70–73]. At the end of the tube is a (second)
quadrupole mass spectrometer, which enables detection
of product ions formed by chemical ionization [72]. SIFTMS is characterized by short analysis time of about
several seconds and provides analysis of samples with
high humidity content [73].
Proton transfer reaction-MS (PTR-MS) is widely applied for measuring trace amounts of VOCs in exhaled
air. The technique may be used for real-time monitoring
of VOCs down to breath-to-breath resolution. Examples
are the investigation of breath during exercising on a
stationary bicycle or during sleep [75,76,4,77–81]. In
PTR-MS, H3O+ ions are formed in the ion source by
hollow-cathode discharge in water vapor. H3O+ ions in
the drift tube are mixed with the continuously flowing
sample of air. Particles with proton affinity greater than
water will accept a proton. Modern PTR-MS devices can
also use NO+ or O2+ as primary ions (similar to SIFTMS). Ions leaving the drift tube through small slits get
into the lens system, where they are transferred into the
mass spectrometer [73].
The analysis of breath by PTR-MS has several
advantages:
(1) the gas mixture can be identified without previous
preconcentration and separation;
(2) compounds occurring at high concentrations (e.g.,
N2, CO2, O2, or H2O) do not interfere in the measurements; and,
(3) PTR-MS has a very high sensitivity.
SIFT-MS and PTR-MS characterize the substances
according to their mass-to-charge ratio, whereas chemical identification must be provided by other techniques
Table 3. Real-time measurements of VOCs in human breath with PTR-MS, SIFT-MS and IMS
Compounds
Formaldehyde
Isopropanol
Formaldehyde
Acetone
Ammonia
Ethanol
Isobutanol
Isoprene
Pentane
Type of sample
Technique for detection VOCs
LOD
LOQ
Ref.
Breath/lung
PTR-MS
ns
ns
[112]
Urine/bladder and prostate
Breath/lung
infection
SIFT-MS
MCC-63Ni-IMS
ns
50 pg/L
10 pg/L
1 lg/L
200 pg/L
8 pg/L
5 lg/L
ns
ns
[113]
[85]
ns, not specified.
http://www.elsevier.com/locate/trac
5
Trends
(e.g., GC-MS) [75,76,4,82]. Table 3 shows the results of
real-time measurement that can be applied to identify
volatile markers.
Ion-mobility spectrometry (IMS) is a fast, high-sensitivity technique, with LODs down to ng/L–pg/L and
ppbv–pptv, with no pre-concentration necessary for
identification of trace amounts of substances in a gasphase sample. The device detects compounds with high
proton affinities (aldehydes, ketones, amines) [83–85].
IMS comprises an ionization chamber, ion-molecule
injection shutter, an ion-drift tube and an ion collector
(Faraday plate) [85,86].
Different methods are used to ionize the gas. The most
common is application of a radioactive source {e.g., 63Ni
[85,87]}, but, an alternative way is by using UV light
[88]. A carrier gas, usually air or nitrogen, introduces
analyte molecules into the ion source, where they undergo ionization. Afterwards, the ions are injected into
the drift region by opening an ion shutter, and they are
separated according to differences in ion mobility
[85,86]. IMS is often coupled with standard GC columns
[88] or multi-capillary columns (MCCs) [85]. MCCs are
characterized by comparatively high flow rate and high
sample capacity in comparison to single narrow columns
[85]. IMS is commonly used to detect chemical-warfare
agents and illegal drugs [86,89–91], and is applied in
biological analysis [92,93], medical diagnostics to
determine human metabolites, sarcoidosis or biomarkers
in exhaled breath [85,87,94–96] and testing food
quality [97,98].
An e-nose is a device containing semiconductor sensors or chemical sensors. It responds selectively to some
properties of the analytes, which may be typical odor
molecules [99,100]. The apparatus often comprises
many different sensors, which result in a combined
information pattern. In most cases, the individual sensors are not selective for a particular compound. The
computer collects signals from all the sensors and forms
an electronic model of smell, which is then compared
with a database containing information about the patterns of chemical composition that serves to identify
them. The e-nose is relatively cheap, provides a short
analysis time and enables the determination of analytes
in trace quantities, qualitative and quantitative analysis
of mixtures without earlier separation, and testing a
large number of odors. The e-nose rapidly returns to its
initial state and has high sensitivity. Repeatability,
however, may present some problems connected to signal drift. The disadvantages of the e-nose are the possibility of contamination, the sensitivity of certain sensors
to polar compounds and water, and change in sensor
response over time [100].
Data obtained from chemical sensors can be analyzed
using statistical calculations [e.g., principal component
analysis, discriminant function analysis and factor
analysis or structural algorithms (neural networks)] to
6
http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 38, 2012
distinguish and to identify the odor substances in the
samples [99,101]. An e-nose has a wide area of application:
(1) production of cosmetics [102];
(2) food industry to control the freshness and quality of
the products of fermentation processes or in food
production [103–105];
(3) medicine to distinguish between patients with lung
cancer and healthy persons or to diagnose urinary
tract infections [106,107];
(4) forensic research [108]; and,
(5) testing for pollution of the environment [100,109].
The types of e-nose include [110,111]:
(1) optical – fluorescence, reflective, absorption;
(2) thermal pellistor;
(3) electrochemical – chemiresistive (conducting polymers, metal oxide semiconductors), potentiometric,
and amperometric; and,
(4) gravimetric – bulk acoustic wave, surface acoustic
wave, and flexural plate wave.
4. Unconventional methods for detection of
carcinoma
Domestic dogs (Canis familiaris) are kept by humans or
accompany human settlements all over the world. Due
to having been domesticated about 14,000 years ago
and subsequent selective breeding, dogs are now used by
humans in multiple ways. One of the most important
features making dogs suitable for various tasks is their
ability to learn and to cooperate with humans.
In most uses of dogs, their excellent sense of smell
plays a crucial role. The sense of smell generally plays a
fundamental role in animalsÕ life (e.g., in finding food,
alerting to predatorsÕ approach, identifying individuals of
their own species, reproduction, and communication
between parents and offspring) [114]. Trained dogs are
used for many purposes {e.g., tracking [115,116],
detection of various substances – mainly drugs [117–
120], explosives [121–123], smuggled food, cigarettes,
alcohol [121] – identification of persons on the basis of
individual scent [124,125], finding human victims of
disasters (e.g., people buried after earthquakes or avalanches) [126], or searching for human remains [127]}.
Currently some new, quite unusual areas of canine
employment have been reported, including detection of
mold in buildings [128], and moths, termites, bark beetles or bedbugs [129].
The hypothesis about the ability of dogs to detect
cancer in humans based on the specific odor was put
forward for the first time by Williams and Pembroke
[130], who described the case of an obvious contribution
of an untrained pet dog to detection of melanoma on the
leg of its owner. Another similar case of was reported by
Church and Williams [131]. In both these cases, the
Trends
Figure 1. Canine sense of smell [141,142].
Trends in Analytical Chemistry, Vol. 38, 2012
http://www.elsevier.com/locate/trac
7
Trends
dogs showed a clear interest in one particular spot (a
mole) on the skin of their ownerÕs leg. This spot was
frequently sniffed, even through clothing, and licked; the
dog was trying to bite off the mole. This unusual
behavior prompted the owner to visit a dermatologist,
who, on the basis of histopathological examination,
diagnosed a melanoma located in the spot that interested
the dog. After surgery, the dogs showed no more interest
in the area from which the melanoma had been removed
[130,131].
These two reports stimulated further studies on special
training of several dogs to detect different cancers using
various odor samples (breath, urine, cancer tissue)
[132–138]. Although dogsÕ sense of smell is in many
aspects still superior to analytical devices, we can suppose that, in the future, development and improvement
of instrumental methods, especially GC-MS and sensors,
may effectively replace application of dogs in cancer
screening based on emitted odors.
5. Canine sense of smell
In dogs and other macrosmatic animal species (e.g., rats
or pigs) that are characterized by a well-developed sense
of smell, the area of olfactory epithelium in the nasal
cavity is much larger than in microsmatic species,
including humans. The ambient air containing odor
molecules enters the dogÕs nasal cavity during sniffing
through a pair of nostrils (Fig. 1). The characteristic
shape of canine nostrils with their alar folds ensures
that, during sniffing, sufficient odor molecules go with
the air flow into the nasal cavity to be analyzed by the
olfactory organ. The olfactory receptors are located in
the nasal epithelium, covered with mucus. Until recently, the role of olfactory receptors was ascribed to the
olfactory cells located in the olfactory epithelium. Currently, what are considered to be olfactory receptors are
the proteins found in the membrane of olfactory cells.
Odor molecules that enter the nasal cavity during sniffing dissolve in the mucus lining and are bound to
odorant-binding proteins [139,140]. The axons of the
olfactory cells reach the olfactory bulb, within which
they converge in structures, called glomeruli, located in
the outer layer of the olfactory bulb. In the inner layer of
the olfactory bulb, the so-called mitral cells form
glomeruli with the axons of olfactory receptor cells and
send the stimulation caused by odor via nerves within
the olfactory tract to the olfactory cortex where multiple
signals are processed. One olfactory cell expresses only
one functional odor receptor, which can be responsive to
various odor molecules. Due to the axon convergence at
the level of an olfactory bulb and highly complex processing of the olfactory signal, each individual odor
produces a specific spatial map of excitation in the
olfactory bulb. Through spatial encoding, the brain is
8
http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 38, 2012
able to distinguish specific odors. The chemical nature of
the odors is important, as there may be a chemotropic
map in the brain showing specific activation pattern for
specific odors.
The acuity of the sense of smell depends on the size of
the olfactory epithelium and on the number of olfactory
cells. For example, in dogs (German Shepherds), there
are more than 200 million olfactory cells on an area of
about 170 cm2, while, in humans, there are about 5
million cells on about 5 cm2 of olfactory epithelium. The
ability to discriminate numerous compounds depends
also on the proportion of active/inactive genes of the
olfactory receptor proteins. Olfactory receptors are
stimulated only by gaseous substances in the air, which,
by diffusion, get in contact with the mucous membrane
and dissolve in the thin mucus layer covering the
membrane [143].
Dogs are able to sniff out about half-a-million odorous
compounds at trace concentrations (e.g., amyl acetate at
the concentration of 2 ppt, which is imperceptible to a
human nose). An excellent sense of smell and the ability
to learn by operant conditioning make dogs good biodetectors for different kinds of odors that may be markers
of cancers.
Most of the results of using trained dogs to detect
cancers are promising; however, this must be confirmed
in further in-depth studies to exclude possible artifacts.
For example, Willis et al. [132] trained six dogs of various breeds (including a mongrel, a Labrador and a
Cocker Spaniel) to detect bladder cancer on the basis of
the odor of urine from patients and healthy persons. For
this study, the dogs were trained for a period of
7 months, 1–2 h/day, 5 times a week. Of 54 samples
taken from patients with bladder cancer, the six dogs
correctly indicated 22 samples (41% correct positive
responses), which is significantly better than by chance.
These authors applied a line-up of seven samples, for
which the probability of correct indications by chance
was 14% [132].
Less satisfactory results were obtained by Gordon et al.
[133], who used six trained dogs for detection of breast
cancer and four dogs for detection of prostate cancer. In
their study, samples of urine odor were used in dog
training, which took 12–13 months, 15–30 samples per
day, 2–7 times a week [133]. The detection sensitivity
for patients was only 22% and 17% for breast and
prostate cancer, respectively. Of 10 dogs, only two
achieved detection sensitivity significantly better than by
chance and none of the dogs demonstrated detection
specificity better than random choice.
In the research conducted by Pickel et al. [134], two
dogs were used to locate samples of melanoma tissue put
on the body of healthy volunteers. The first dog correctly
found melanoma tissue in 6 out of 7 persons used for the
experiment. In one case, the dog indicated a sample in
which only a fraction of cells showed neoplastic changes,
Trends in Analytical Chemistry, Vol. 38, 2012
Trends
Figure 2. Classification compounds by PCA [144].
and, in one case, the dog did not indicate a sample in
which dermatological examination was inequivocal but
histopathological examination was positive. With the
second dog, only 4 out of 7 persons were tested, and all
indications were consistent with the indications of the
first dog. Pickel et al. [134] suggested that VOCs may be
released from cancer cells on the surface of the skin in an
amount that allows early detection by dogs.
One dog (Riesenschnauzer) was trained for 12 months
and used by Horvath et al. [135] to study ovarian tumor
tissues from 20 patients and 80 samples of healthy tissues; the detection sensitivity in this test was 100%,
whereas the specificity was 97.5%.
Also, McCulloch et al. [136] used five trained dogs and
samples of exhaled breath of patients with lung and
breast cancer tested versus samples of exhaled air of
healthy volunteers to achieve a very high detection
sensitivity of 99% and 88% for lung and breast cancer,
respectively. The respective detection specificity was 99%
and 98%.
More recently, Cornu et al. [137] conducted another
experiment on detection of prostate cancer by using odor
of urine samples. One trained dog in their study showed
a detection sensitivity and specificity of 91%.
A study conducted by Sonoda et al. [138] demonstrated that a trained dog is able to detect colorectal
cancer in humans on the basis of both exhaled breath
and feces odor at the detection sensitivity of 91% and
97% for breath and feces samples, respectively, while the
specificity was 99% for both kinds of samples.
The papers published so far demonstrate that dogs,
after appropriate training, are able to discriminate
breath, urine, feces or tumor-tissue samples of patients
with lung, breast, prostate and ovarian cancers from
respective samples taken from healthy humans with
sensitivity (the true positive) and specificity (the true
negativity) exceeding 80%. The canine indications are
easily interpretable in terms of calculating the detection
sensitivity and specificity, but no information can be
obtained on what chemical compounds dogs are
responding to or the quantity of those compounds.
McCulloch et al.[136] and Sonoda et al. [138] applied
the chi-square (Chi2) test or the Welch test to evaluate
differences in the dogsÕ indications between pattern and
control samples. However, the statistical analysis was
carried out on a small number of samples with a type of
cancer.
Other methods that could find correlation between dog
indications and chemical analyses are PearsonÕs correlation and principal component analysis (PCA). The
PearsonÕs correlation is used to indicate a relationship
between two variables. Strength of the correlation is
defined as the value in the range -1 to 1. If the correlation value is positive, it means that increasing values of
one variable increases the values of another variable
(decrease of one variable - decrease the other). If the
http://www.elsevier.com/locate/trac
9
Trends
Trends in Analytical Chemistry, Vol. 38, 2012
correlation value is negative, then increasing the value
of one variable decreases the value of the other variable
(decrease of one variable - increase in the second).
Buszewski et al. [144] tried to find correlation between
VOCs in human breath and dog indications. For the
patient group, a positive PearsonÕs correlation indicated
a significant positive or negative tendency. Ethyl acetate
and 2-pentanone correlated with the dog indication
positively (r = 0.85 and r = 0.97, respectively), whilst
the concentrations for acetonitrile, propanal and 1-propanol were negatively correlated with the dog indications
(r = 0.78,
r = 0.87
and
r = 0.98,
respectively). Analogously to a simple correlation, the
signs of loading explain the role of compounds in the dog
indications in the same way as the simple correlations.
Hence, the first factor could be named as indifferent for
the dog, whereas the second factor was concerned with
the dog indications. PCA could also be applied for classification to show the relations between VOCs and the
dog indications towards two components (Fig. 2) [144].
6. Summary
Chemical analysis of exhaled air can provide essential
information about human health. The composition of
exhaled breath can vary depending on the types of
illnesses (e.g., diabetes or lung diseases producing specific odors). Analysis of breath constituents may offer
assistance in identifying potential markers, which may
be useful in cancer screening and thus contribute to
early diagnosis and successful treatment. Analysis of
odor samples by GC-MS carried out simultaneously with
tests using trained dogs may therefore elucidate the
question of whether and which VOCs could be markers
of the cancer disease that dogs respond to. Such an approach and improvement in analytical methods may in
the future be crucial for the development of new methods
of cancer screening.
Acknowledgements
This work was supported by Nicolaus Copernicus University in Toruń Grant Faculty of Chemistry no. 456,
project UDA-POKL.04.01.01-00-081/10-00 and Grant
N N 204 026238 (2010-2013).
References
[1] C. Prado, P. Martin, J.F. Periago, J. Chromatogr., A 1011 (2003)
125.
[2] M. Libardoni, P.T. Stevens, J. Waite, R. Sacks, J. Chromatogr., B
842 (2006) 13.
10
http://www.elsevier.com/locate/trac
[3] T. Ligor, J. Szeliga, M. Jackowski, B. Buszewski, J. Breath Res. 1
(2007) 1.
[4] A. Amann, G. Poupart, S. Telser, M. Ledochowski, A. Schmid, S.
Mechtcheriakov, Int. J. Mass Spectrom. Ion Processes 239
(2004) 227.
[5] H. Yu, L. Xu, P. Wang, J. Chromatogr. B 826 (2005) 69.
[6] B. Buszewski, M. Ke˛sy, T. Ligor, A. Amann, Biomed. Chromatogr.
21 (2007) 553.
[7] W. Miekisch, J.K. Schubert, G.F.E. Noeldge-Schomburg, Clin.
Chim. Acta 347 (2004) 25.
[8] D. Smith, P. Spanel, A.A. Foyer, F. Hanna, G.A.A. Ferns, J.
Breath Res. 5 (2011) 022001.
[9] E. Kopczyńska, Przydatność markerów nowotworowych w diagnostyce onkologicznej, Bydgoszcz, Poland, 2004.
[10] M. Phillips, K. Gleeson, J.M.B. Hughes, J. Greenberg, R.N.
Cataneo, L. Baker, W.P. McVay, Lancet 353 (1999) 1930.
[11] M. Phillips, R.N. Cataneo, A.R.C. Cumin, A.J. Gagliardi, J.
Greenberg, R.A. Maxfield, W.N. Rom, Chest 123 (2003) 2115.
[12] F. Di Francesco, R. Fuoco, M.G. Trivella, A. Ceccarini, Microchem. J. 79 (2005) 405.
[13] R. Hyspler, S. Crhova, J. Gasparic, Z. Zadak, M. Cizkova, V.
Balasova, J. Chromatogr., B 739 (2000) 183.
[14] C. Deng, J. Zhang, Y. Xiaofeng, W. Zhang, X. Zhang, J.
Chromatogr., B 810 (2004) 269.
[15] A. Hryniuk, B.M. Ross, Int. J. Mass Spectrom. Ion Processes 285
(2009) 26.
[16] A.W. Jones, V. Lagesson, C. Tagesson, J. Chromatogr., B 672
(1995) 1.
[17] W. Mueller, J. Schubert, A. Benzing, K. Geiger, J. Chromatogr., B
716 (1998) 27.
[18] J.J.B.N. van Berkel, J.W. Dallinga, G.M. Moller, R.W.L. Godschalk, E. Moonen, E.F.M. Wouters, F.J. van Scooten, J. Chromatogr., B 861 (2008) 101.
[19] M.S. Abbott, B. Elder, P. Spanel, D. Smith, Int. J. Mass Spectrom.
Ion Processes 228 (2003) 655.
[20] P. Spanel, P. Rolfe, B. Rajan, D. Smith, Ann. Occup. Hyg. 40
(1996) 615.
[21] C. Warneke, J. Kuczyński, A. Hansel, A. Jordan, W. Vogel, W.
Lindinger, Int. J. Mass Spectrom. Ion Processes 154 (1996) 61.
[22] A. Hansel, A. Jordan, R. Holzinger, P. Prazeller, W. Vogel, W.
Lindinger, Int. J. Mass Spectrom. Ion Processes 149 (150) (1995)
609.
[23] A. DÕAmico, G. Pennazza, M. Santonico, E. Martinelli, C.
Roscioni, G. Galluccio, R. Paolesse, C. Di Natale, Lung Cancer
68 (2010) 170.
[24] S.M. Chao, Y.J. Kim, G.S. Heo, S.M. Shin, Sens. Actuators, B 117
(2006) 50.
[25] A. Jemal, R. Siegel, E. Ward, Y. Hao, J. Xu, T. Murray, M.J. Thun,
CA Cancer J. Clin. 58 (2008) 71.
[26] M. Borgerding, H. Klus, Exp. Toxicol. Pathol. 57 (2005) 43.
[27] R.R. Baker, Prog. Energy Combust. Sci. 32 (2006) 373.
[28] A. McWilliams, S. Lam, Curr. Oncol. Rep. 4 (2002) 487.
[29] V. Kumar, A.K. Abbas, N. Fausto, R. Mitchell, Robbins Basic
Pathology (Eighth edition), Saunders/Elsevier, Amsterdam, The
Netherlands, 2007, pp. 528–540.
[30] P.A. Oliviera, A. Colaco, R. Chaves, H. Guedes-Pint, P.L.F. De-LaCruz, C. Lopes, An. Acad. Bras. Cienc. 79 (2007) 593.
[31] H.C. Pitot, Y.P. Dragon, FASEB J. 5 (1991) 2280.
[32] D. Poli, M. Goldoni, M. Corradi, O. Acampa, P. Carbognani, E.
Internullo, A. Casalini, A. Mutti, J. Chromatogr. B 878 (2010)
2643.
[33] E.M. Gaspar, A.F. Lucena, J. Duro da Costa, H. Chaves das Neves,
J. Chromatogr., A 1216 (2009) 2749.
[34] N. Li, C. Denga, N. Yao, X. Shen, X. Zhang, Anal. Chim. Acta
540 (2005) 317.
Trends in Analytical Chemistry, Vol. 38, 2012
[35] M. Phillips, R.N. Cataneo, B.A. Ditkoff, P. Fisher, J. Greenberg, R.
Gunawardena, C.S. Kwon, F. Rahbari-Oskoui, C. Wong, Breast J.
9 (2003) 184.
[36] H.P. Chan, C. Lewis, P.S. Thomas, Lung Cancer 63 (2009) 164.
[37] J. Hrivnak, E. Kralovicova, P. Tolgyessy, Petrol. Coal 50 (2008)
10.
[38] J.M. Sanchez, R.D. Sacks, Anal. Chem. 75 (2003) 2231.
[39] J.M. Sanchez, R.D. Sacks, Anal. Chem. 78 (2006) 3046.
[40] B. Buszewski, M. Szultka, Crit. Rev. Anal. Chem. (accepted).
[41] M. Clement, S. Arzel, B. Le Bot, R. Seux, M. Millet, Chemosphere
40 (2000) 49.
[42] N.H. Snow, J. Chromatogr., A 885 (2000) 445.
[43] Z. Mester, R. Strugeon, J. Pawliszyn, Spectrochim. Acta, Part B
56 (2001) 233.
[44] M.F. Musteata, J. Pawliszyn, J. Biochem. Biophys. Methods 70
(2007) 181.
[45] J. Pawliszyn, Anal. Chem. 75 (2003) 2543.
[46] J. Namieśnik, B. Zygmunt, A. Jastrze˛bska, J. Chromatogr., B 885
(2000) 405.
[47] L. van der Wal, C.A.M. van Gestel, J.L.M. Hermens, Chemosphere
54 (2004) 561.
[48] X. Song, J. Li, L. Chen, Z. Cai, C. Liao, H. Peng, H. Xiong, J. Braz.
Chem. Soc. 23 (2012) 132.
[49] M.S.S. Curren, J.W. King, J. Agric. Food Chem. 49 (2001) 2175.
[50] R. Baciocchi, M. Attina, G. Lombardi, M.R. Boni, J. Chromatogr.,
A 911 (2001) 135.
[51] J.-B. Yu, H.-G. Byun, M.-S. So, J.-S. Huh, Sens. Actuators, B 108
(2005) 305.
[52] V. Vickackaite, V. Ciuvasovaite, Cent. Eur. J. Chem. 5 (2007)
727.
[53] B. Sari, M. Talu, Turk. J. Chem. 22 (1998) 301.
[54] J. Huang, Pure Appl. Chem. 78 (2006) 15.
[55] M.H. Harun, E. Saion, A. Kassim, N. Yahya, E. Mahmud, J. Am.
Stat. Assoc. 2 (2007) 63.
[56] Y. Gong, I.-Y. Eom, D.-W. Lou, D. Hein, J. Pawliszyn, Anal.
Chem. 80 (2008) 7275.
[57] A. Wang, F. Fang, J. Pawliszyn, J. Chromatogr., A 1072 (2005)
127.
[58] I.-Y. Eom, J. Pawliszyn, J. Sep. Sci. 31 (2008) 2283.
[59] I.-Y. Eom, A.-M. Tugulea, J. Pawliszyn, J. Chromatogr., A 1196–
1197 (2008) 3.
[60] M. Mieth, S. Kischkel, J.K. Schubert, D. Hein, W. Miekisch, Anal.
Chem. 81 (2009) 5851.
[61] M. Mieth, J.K. Schubert, T. Groger, B. Sabel, S. Kischkel, P.
Fuchs, D. Hein, R. Zimmermann, W. Miekisch, Anal. Chem. 82
(2010) 2541.
[62] P. Tolgyessy, J. Hrivnak, E. Kralovicova, D. Barlokova, Chromatographia 71 (2010) 953.
[63] H. Jurdakova, R. Kubinec, M. Jurcisinova, Z. Krkosova, J. Blasko,
I. Ostrovsky, L. Sojak, V.G. Berezkin, J. Chromatogr., A 1194
(2008) 161.
[64] C. Wang, P. Sahay, Sensor 9 (2009) 8230.
[65] L. Mondello, P. Quinto Tranchida, P. Dugo, G. Dugo, Mass
Spectrom. Rev. 27 (2008) 101.
[66] S.E. Stein, J. Am. Soc. Mass Spectrom. 10 (1999) 770.
[67] D.E. Garcia, E.E. Baidoo, P.I. Benke, F. Pingitore, Y.J. Tang, S.
Villa, J.D. Kesling, Curr. Opin. Microbiol. 11 (2008) 233.
[68] A. El-Aneed, A. Cohen, J. Banou, Appl. Spectrosc. Rev. 44
(2009) 210.
[69] S.E. Van Bramer, An Introduction to Mass Spectrometry,
Widener University, Pennsylvania, USA, 1997.
[70] P. Spanel, D. Smith, J. Breath Res. 2 (2008) 046003.
[71] P. Spanel, D. Smith, Mass Spectrom. Rev. 30 (2011) 236.
[72] D. Smith, P. Spanel, Eur. J. Mass Spectrom. 13 (2007) 77.
[73] P. Spanel, K. Dryahina, D. Smith, Int. J. Mass Spectrom. Ion
Processes 249–250 (2006) 230.
Trends
[74] A. Amann, D. Smith (Editors), Breath Analysis for Clinical
Diagnosis and Therapeutic Monitoring, World Scientific, Singapore, 2005.
[75] A. Jordan, A. Hansel, R. Holzinger, W. Lindinger, Int. J. Mass
Spectrom. Ion Processes 148 (1995) L1.
[76] A. Critchley, T.S. Elliott, G. Harrisom, C.A. Mayhew, J.M.
Thompson, T. Worthington, Int. J. Mass Spectrom. Ion Processes
239 (2004) 235.
[77] H. Koc, J. King, G. Teschl, K. Unterkofler, S. Teschl, P.
Mochalski, H. Hinterhuber, A. Amann, J. Breath Res. 5 (2011)
037102.
[78] J. King, H. Koc, K. Unterkofler, P. Mochalski, A. Kupferthaler, G.
Teschl, S. Teschl, H. Hinterhuber, A. Amann, J. Math. Biol. 63
(2011) 959.
[79] J. King, H. Koc, K. Unterkofler, P. Mochalski, A. Kupferthaler, G.
Teschl, S. Teschl, H. Hinterhuber, A. Amann, J. Theoret. Biol.
267 (2010) 626.
[80] J. King, P. Mochalski, A. Kupferthaler, K. Unterkofler, H. Koc, W.
Filipiak, S. Teschl, G. Teschl, H. Hinterhuber, A. Amann, Physiol.
Meas. 31 (2010) 1169.
[81] J. King, A. Kupferthaler, K. Unterkofler, H. Koc, S. Teschl, G.
Teschl, W. Miekisch, J. Schubert, H. Hinterhuber, A. Amann, J.
Breath Res. 3 (2009) 1.
[82] P. Spanel, D. Smith, J. Breath Res. 2 (2008) 1.
[83] J.I. Baumbach, Anal. Bioanal. Chem. 384 (2006) 1059.
[84] J. Rudnicka, P. Mochalski, A. Agapiou, M. Statheropoulos, A.
Amann, B. Buszewski, Anal. Bioanal. Chem. 398 (2010) 2031.
[85] V. Ruzsanyi, J.I. Baumbach, S. Sielemann, P. Litterst, M.
Westhoff, L. Freitag, J. Chromatogr., A 1084 (2005) 145.
[86] G.A. Eiceman, Z. Karpas, Ion Mobility Spectrometry, Taylor and
Francis Group, Boca Raton, FL, USA, 2005.
[87] M. Westhoff, P. Litterst, L. Freitag, J.I. Baumbach, J. Physiol.
Pharmacology 58 (2007) 739.
[88] Z. Xie, S. Sieleman, H. Schmidt, F. Li, J.I. Baumbach, Anal.
Bioanal. Chem. 372 (2002) 606.
[89] L.M. Matz, H.H. Hill, Anal. Chem. 73 (2001) 1664.
[90] A.H. Lawrence, Anal. Chem. 61 (1989) 343.
[91] T. Keller, A. Miki, P. Regenscheit, R. Dirnhofer, A. Schneider, H.
Tsuchihashi, Forensic Sci. Int. 94 (1998) 55.
[92] H.H. Hill, C.H. Hill, G.R. Asbury, C. Wu, L.M. Matz, T. Ichiye, Int.
J. Mass Spectrom. Ion Processes 219 (2002) 23.
[93] V. Ruzsanyi, J.I. Baumbach, G.A. Eiceman, Int. J. Ion Mobility
Spectrom. 6 (2003) 53.
[94] V. Ruzsanyi, J.I. Baumbach, Int. J. Ion Mobility Spectrom. 8
(2005) 5.
[95] B. Bödeker, W. Vautz, J.I. Baumach, Int. J. Ion Mobil. Spectrom.
11 (2008) 89.
[96] A. Bunkowski, B. Bödeker, S. Bader, M. Westhoff, P. Litterst, J.I.
Baumbach, Int. J. Ion Mobil. Spectrom. 12 (2009) 73.
[97] G.M. Bota, P.B. Harrington, Talanta 68 (2006) 629.
[98] Z. Karpas, B. Tilman, R. Gdalevsky, A. Lorber, Anal. Chim. Acta
463 (2002) 155.
[99] R.F. Machaldo, D. Laskowski, O. Deffenderfer, T. Burch, S. Zheng,
P.J. Mazzone, T. Mekhail, C. Jennings, J.K. Stoller, J. Pyle, J.
Duncan, R.A. Dweik, S.C. Erzurum, Am. J. Respir. Crit. Care Med.
171 (2005) 1286.
[100] A. Branca, P. Simonian, M. Ferrante, E. Novas, R.M. Negri, Sens.
Actuators, B 92 (2003) 222.
[101] J. Brezmes, B. Ferreres, E. Llobet, X. Vilanova, X. Correig, Anal.
Chin. Acta 348 (1997) 503.
[102] A. Carrasco, C. Saby, P. Bernadet, Flavour Fragrance J. 13
(1998) 335.
[103] I. Lundström, T. Ederth, H. Kariis, H. Sundgren, A. Spetz, F.
Winquist, Sens. Actuators, B 23 (1995) 127.
[104] C. Li, G.W. Krewer, P. Ji, H. Scherm, S.J. Kays, Postharvest Biol.
Technol. 55 (2010) 144.
http://www.elsevier.com/locate/trac
11
Trends
Trends in Analytical Chemistry, Vol. 38, 2012
[105] R. Saja, J. Souto, M.L. Rodriguez-Mendez, J.A. Saja, Mater. Sci.
Eng., C 8–9 (1999) 565.
[106] A. Dragonieri, J.T. Annema, R. Schot, M.P.C. van der Schee, A.
Spanevello, P. Carratu, O. Resta, K.F. Rabe, P.J. Sterk, Lung
Cancer 64 (2009) 166.
[107] A.K. Pavlou, N. Magan, C. McNulty, J.J. Meecham, D. Sharp, J.
Brown, A.P.F. Turner, Biosens. Bioelectron. 17 (2002) 893.
[108] G.A. Bakken, G.W. Kauffman, P.C. Jurs, K.J. Albert, S.S. Stitzel,
Sens. Actuators, B 79 (2001) 1.
[109] P. Mielle, Trends Food Sci. Technol. 7 (1996) 432.
[110] D.J. Strike, M.G.H. Meijerink, M. Koudelka-Hep, FreseniusÕ J.
Anal. Chem. 364 (1999) 499.
[111] D. James, S.M. Scott, Z. Ali, T. OÕHare, Microchim. Acta 149
(2005) 1.
[112] A. Wehinger, A. Schmid, S. Mechtcheriakov, M. Ledochowski, C.
Grabmer, G.A. Gastl, A. Amann, Int. J. Mass Spectrom. Ion
Processes 265 (2007) 49.
[113] P. Spanel, D. Smith, T.A. Holland, W. Al Singary, J.B. Elder,
Rapid Commun. Mass Spectrom. 13 (1999) 1354.
[114] A.M. Elgier, A. Jakovcevic, G. Barrera, A.E. Mustaca, M.
Bentosela, Behav. Process. 81 (2009) 402.
[115] A. Thesen, J.B. Steen, K.B. Doving, J. Exp. Biol. 180 (1993) 247.
[116] P.G. Hepper, D. Wells, Chem. Senses 30 (2005) 291.
[117] M. Maejima, M. Inoue- Murayama, K. Tonosaki, N. Matsuura, S.
Kato, Y. Saito, A. Weiss, Y. Murayama, S. Ito, Appl. Anim.
Behav. Sci. 107 (2007) 287.
[118] N. Lorenzo, T.L. Wan, R.J. Harper, Y.-L. Hsu, M. Chow, S. Rose,
K.G. Furton, Anal. Bioanal. Chem. 376 (2003) 1212.
[119] A.M. Rouhi, Chem. Eng. News 75 (1997) 24.
[120] G.J. Adams, K.G. Johnson, Appl. Anim. Behav. Sci. 41 (1994)
115.
[121] K.G. Furton, L.J. Myers, Talanta 54 (2001) 487.
[122] I. Gazit, J. Terkel, Appl. Anim. Behav. Sci. 82 (2001) 65.
[123] R. Fjellanger, E.K. Andersen, I.G. McLean, Int. J. Comp. Psychol.
15 (2002) 277.
[124] G.A.A. Schoon, Appl. Anim. Behav. Sci. 49 (1996) 257.
[125] R.H. Settle, B.A. Sommerville, J. McCormick, D.M. Broom, Anim.
Behav. 48 (1994) 1443.
[126] L. Lit, C.A. Crawford, Appl. Anim. Behav. Sci. 98 (2006) 277.
12
http://www.elsevier.com/locate/trac
[127] V. Fenton, J. Wilderness Med. 3 (1992) 292.
[128] E. Kauhanen, M. Harri, A. Nevalainen, T. Nevalainen, Environ.
Int. 28 (2002) 153.
[129] S.E. Brooks, F.M. Oi, P.G. Koehler, J. Econ. Entomol. 95 (2003)
1259.
[130] H. Williams, A. Pembroke, Lancet 333 (1989) 734.
[131] J. Church, H. Williams, Lancet 358 (2001) 930.
[132] C.M. Willis, S.M. Church, C.M. Guest, W.A. Cook, M. McCarthy,
A.J. Bransbury, M.R.T. Church, J.C.T. Church, Br. Med. J. 329
(2004) 712.
[133] R.T. Gordon, C.B. Schatz, L.J. Myers, M. Kosty, C. Gonczy, J.
Kroener, M. Tran, P. Kurtzhals, S. Heath, J.A. Koziol, N. Arthur,
M. Gabriel, J. Hemping, G. Hemping, S. Nesbitt, L. Tucker-Clark,
J. Zaayer, J. Altern. Complemen. Med. 14 (2008) 61.
[134] D.P. Pickel, G.P. Manucy, D.B. Walker, S.B. Hall, J.C. Walker,
Appl. Anim. Behav. Sci. 89 (2004) 107.
[135] G. Horvath, G.A. Jarverud, S. Jarverud, I. Horvath, Integr.
Cancer. Ther. 7 (2008) 76.
[136] M. McCulloch, T. Jezierski, M. Broffman, A. Hubbard, K. Turner,
T. Janecki, Integr. Cancer. Ther. 5 (2006) 30.
[137] J.N. Cornu, G. Cancel-Tassin, V. Ondet, C. Girardet, Eur. Urol. 59
(2010) 197.
[138] H. Sonoda, S. Kohnoe, T. Yamazato, Y. Satoh, G. Morizono, K.
Shikata, M. Morita, A. Watanabe, M. Morita, Y. Kakeji, F. Inoue,
Y. Maehara, Gut (2011) (DOI: 10.1136/gut.2010.218305).
[139] P. Quignon, E. Kirkness, E. Cadieu, N. Touleimat, R. Guyon, C.
Renier, C. Hitte, C. André, C. Fraser, F. Galibert, Genome, Biol. 4
(2003) 80.1.
[140] B. Malnic, J. Hirono, T. Sato, L.B. Buck, Cell 96 (1999) 713.
[141] K.D. Budras, P.H. McCarthy, W. Fricke, R. Richter, Anatomy of
the Dog (Fifth edition), Schlütersche Verlagsgesellschaft mbH &
Co. KG, Hanover, Germany, 2007.
[142] T. Jezierski, M. Walczak, D. Glanc, A. Górecka, M. Dziubińska,
Zmysł we˛chu psów i jego praktyczne wykorzystanie, Instytut
Genetyki i Hodowli Zwierza˛t PAN, Jastrze˛biec, Poland, 2008.
[143] C. Browne, K. Stafford, R. Fordham, Irish Vet. J. 2 (2006) 97.
[144] B. Buszewski, T. Ligor, T. Jezierski, A. Wenda-Piesik, M. Walczak,
J. Rudnicka, Anal. Bioanal. Chem. (2012) (DOI: 10.1007/
s00216-012-6102-8).