Poster Presentation at the April 2002 meetings of the Association for Chemoreception Sciences on: Me

1
Poster Presentation at the April 2001 meetings of
the Association for Chemoreception Sciences on 
Preliminary evidence of canine olfactory
detection of melanoma
Duane P. Pickel1, Armand B. Cognetta2, Glenda P.
Manucy3, Dianne B. Walker4, Sandra B. Hall5 and
James C. Walker4    1 VONPICKEL K-9 inc
Tallahassee FL USA2 Department of Dermatology
University of Florida Gainesville, FL USA3
Morninglo Goldens Tallahassee FL USA4 Sensory
Research Institute Florida State University
Tallahassee, FL USA
2
AbstractSince chemical markers of melanoma have
been reported in blood and urine, volatile
chemicals might be emitted from melanoma cells
(on the skin surface) in amounts sufficient to
allow accelerated diagnosis. Using methods
normally used in canine olfactory detection of
drugs and explosives, we demonstrated (in two
dogs) reliable localization of melanoma tissue
samples hidden on the skin of healthy volunteers.
One dog (1) also "confirmed" clinically
suspected (and subsequently biopsy-proven)
diagnoses of melanoma in five patients. In an
additional patient, this dog made a definitive
response to a skin location for which initial
pathological examination was negative, despite
clinical suspicion. Much more thorough
pathological examination in this individual then
confirmed melanoma in a fraction of the cells. In
a seventh case, in whom neither dog nor
dermatologist provided a definitive response,
pathology showed melanoma. Dog 2 searched four
of these seven patients in each case responses
agreed with those of dog 1. We suggest that
these findings warrant a much more refined and
multi-disciplinary study of the conditions under
which detection of melanoma might be enhanced by
biological or non-biological sampling of volatile
chemicals emanating from skin.
3
Introduction/Background! 
What led to this work? 47,000 new cases of
melanoma diagnosed annually in the US 8,000
deaths annually in the US from melanoma Many
deaths could be prevented if diagnoses could be
made earlier. Likelihood of a melanoma being
detected varies with such factors as stage and
type of melanoma and training of
physician. Current medical practice relies on
aided or (usually) unaided visual system of
physician. Over the past 15 years or so, a
number of urinary and blood biomarkers of
melanoma have been found (e.g., Wakamatsu Ito,
1990) A 1989 Letter (by Williams and Pembroke,
1989) to the Editor of Lancet described a patient
whose dog persisted in exploring a spot on the
patient's leg that was found to be melanoma. 
4
Collectively the above points indicate thatIt
is critically important to pursue possible
adjuncts to the present visual-only approach to
melanoma diagnosis.It is plausible that there
might be chemical markers of melanoma that appear
on the body surface and could, therefore, be used
to enhance diagnosisTwo questions were
posed1. Could dogs that were already highly
trained to perform in field scent discrimination
trials be trained to locate melanoma tissue
samples by smell?2. Could dogs detect melanoma
by smell in actual patients?
5
THE DOGS Two dogs, their history before the
study and efforts to train them to identify
melanoma-only or mixed B-S-M (basalsquamousmelan
oma) tissue samples.Dog 1 was a 4 year old
male Standard Schnauzer, CH OTCH tailgate George
von Pickel UDX ("George"), who had the following
training and titles
Served 2 1/2 years on the Tallahassee Police
force as a police working dog and a certified
bomb detection dog  Demo dog and consultant to
Navy Department and Pentagon  Achieved CD
(Companion Dog), CDX (Companion Dog Excellent),
UDX (Utility Dog Excellent) and OTCH (Obedience
Trial Champion) titles at AKC (American Kennel
Club) approved obedience trials.
Before tests were started with melanoma tissue
samples, "George" was given approximately 100
trials where he was shown a PVC retrieval tube
containing a mixed (B-S-M) set of tissue samples
and then sent to retrieve it.  Approximately
100 area search trials were performed in which
"George" had to locate by, smell, which of
several PVC retrieval tubes contained the B-S-M
sample.
6
THE DOGS Dog 2 was a 6 year old female Golden
Retriever, CH OTCH Colabaugh's Morninglo Breeze
UCX, MH, ("Breeze") received the following
training and titles prior to this study
Achieved CD, CDX, UDX and OTCH titles at AKC
approved obedience trials.  Achieved AKC title
of MH (Master Hunter)
Eleven area search trials were performed in which
"Breeze" was shown a tube containing the B-S-M
sample and then had to locate by smell which of
several PVC retrieval tubes contained this mixed
tissue sample.Seven Box training trials were
conducted in which "Breeze" had to locate by
smell which of 10 holes contained the PVC tube
with the B-S-M sample. (Box trials are described
in more detail below.)
7
Testing with Tissue Samples With both dogs box
trials and body searches were conducted but test
tube trials were run only with dog 1. Body
searches were conducted in which tissue samples
were taped to the skin of healthy volunteers.
A. Box tests The dog was led by the handler to
check all of the compartments of the box shown
here (all dimensions in cm). A valid response for
Dog 1 was to paw at the correct compartment and
then sit. For Dog 2, the correct response was to
mouth the correct compartment before sitting. No
time limit was employed.
8
(No Transcript)
9
Test Tube Tests B. Test tube tests These were
conducted only with Dog 1. The dog was led by
the handler to check all of the test tubes,
positioned in the custom-made rack shown below
(all dimensions in cm). A piece of gauze was
placed in each tube so that all tubes were the
same in appearance. Although odorous distractor
stimuli (typically at least 3) were on most
trials, this information was not recorded. Thus
the probability calculations were based on the
maximum possible number of possible responses
(n10).
10
(No Transcript)
11
Body Search Trials   C. Body search tests A
healthy volunteer wearing a two-piece swim suit
(if a woman) or a regular swimsuit (if a man) had
one adhesive bandage taped to the skin that held
a melanoma tissue sample as well as a varying
number of "empty" bandages. The volunteer lay on
the elevated platform shown schematically below.
On some trials, the volunteer also wore long
underwear or similar clothing to prevent the
adhesive bandages from being seen. The dog was
led by the handler to check all of the bandages.
Valid responses for both dogs were the same as
those used in previously described tests.
12
26 blank (no melanoma present) trials run with
dog 1 17 blank (no melanoma present) trials run
with dog 2 In no case did a false alert occur
13
Actual Patient Trials Dr. Armand Cognetta, a
research dermatologist, recruited patients. They
were instructed to wear either bathing suits,
underwear or trunks and a t-shirt. They were also
asked to bathe with ivory soap the night before
and not to wear any perfume or aftershave. An
assisting nurse placed a variable number of
adhesive bandages and/or 2x2 gauze pads with tape
were placed on the patient's skin.
14
(No Transcript)
15
(No Transcript)
16
CONCLUSIONS AND NEXT STEPS   1.    Research
approaches not typically employed by academic
chemosensory scientists (but commonly used with
success to locate drugs, explosives or criminals)
have provided preliminary evidence that there are
volatile chemical signals from melanoma cells. 
These have the potential to ultimately enhance
medical practice by allowing earlier diagnosis of
this form of skin cancer.2.    Blanket
conclusions as to whether it is feasible to use
canine olfactory ability in the clinical setting
are, we suggest, premature.  Under some sets of
assumptions and conditions, this may be the ideal
tool.  Under others, it may be appropriate to
view canine olfaction as simply the tool that was
used to suggest possible use of chemical signals
in diagnosis. 
17
3.    Ideally, four parallel lines of work will
be stimulated by these results. A.        It
would be valuable for other groups to attempt an
independent replication of our basic observation
that canine olfaction may rival (and conceivably
exceed in some cases) current visual-only
approaches to melanoma detection.B.      
Assuming our observations are confirmed in future
work, it would also be valuable to compare
comprehensively (using a wide variety of
patients) the sensitivity of canine olfaction and
current medical practice.  (It seems plausible to
hypothesize that there will be some cases where
one is superior, others where the other is
superior and still others where senitivity is
similar.)C.       Identification of the
compounds present in the "headspace" over a
melanoma (while it is in place on the surface of
the skin), and determining the relative
abundances of these compounds, would have obvious
benefits in terms of improved diagnosis.D.      
In this and many other areas (see Brisbin et
al., 2000) where important decisions are
contemplated based, in whole or in part, on
olfactory-guided behavior of the dog, it would be
extremely useful if a technique for quantifying
olfactory prowess were available which was
readily understood and accepted by both academic
chemosensory scientists and those that use canine
olfaction to solve "messy" real-world problems.
18
References   Brisbin, I.L., S. Austad S.K.
Jacobson (2000) Canine detectives the nose
knows-or does it? Science 290 Nov. 10, pg.
1093.   Wakamatsu, K. S. Ito.  (1990)
Identification of ester glucronide and sulfate
conjugates of 5-hydroxy-5-methoxyindole-2-carboxyl
ic acid and 6-hydroxy-5-methoxyindole-2-carboxylic
acid in melanoma urine.  Journal of
Dermatological Science 1 253-260.   Williams, H.
A. Pembroke (1989) Sniffer dogs in the melanoma
clinic? The Lancet.  pg. 734.
19
Acknowledgements   Kim Edwards Janice Tatum
Jennifer Taylor Mary Burch Jack Manucy Julie
Tullos Karen Moore Miriam Henrikson Duncan
Moore Tallahassee Memorial Hospital Dermatology
Associates