Transcript Presentation

Existing Methodology
Question:
Will we be able to do this?
Question:
What is “normal” tissue?
Surrogate Anatomic Sites for Evaluating
Cancer Risk
Vadim Backman, Ph.D.
Biomedical Engineering Department
Northwestern University
Colorectal Cancer (CRC)

#2 cause of cancer deaths, ~155K cases annually
and ~57K deaths – stagnant for 20 years

90%+ survival rate if caught early but today the
majority (61%) are later stage

Colonoscopy detects and removes a precursor to
colon cancer, adenomatous polyp, thus decreasing
future occurrence of CRC by 75-90%
Why Colon Cancer Screening Program
Does Not Work?
Problem:
 Current guidelines: everybody over age 50 is recommended to undergo
colonoscopy at least once every 10 years.
 There are >90 million Americans over age 50.
 70-80% of colonoscopies are negative and unnecessary
 Colonoscopic screening of this entire eligible population is impossible
due to
- expense (annual cost would be ~$50 billion)
- insufficient number of endoscopists
- patient reluctance (hate prep!)
- complication rate.
 85% of the population receives no colonoscopic CRC screening.
Solution: Develop a minimally invasive technology to identify patients who
are at risk for CRC and would benefit from colonoscopy.
Precedent: Cervical cancer screening with Pap smear in the last 50 years
reduced mortality from #1 cancer in women to #13!
Factors for A Successful Screening Test:
Inexpensive, No Prep, & Sensitive
Technique
Expensive?
Bowel
preparation
required?
Colonoscopy
Yes, >$1,000
Yes
FOBT
No
No
10%
Fecal DNA
Yes, $700
No
18%
Virtual
colonoscopy
Yes, ~$1,000
Yes
55-80%
Imaging
capsule
Yes, >$2,000
Yes
Not yet approved
for colon imaging
Sensitivity*
100%**
Requirements for a
population screening test:
• No bowel preparation
• Performed by a PCP
• Sensitive
• Inexpensive
*Sensitivity relative to colonoscopy for precancerous advanced adenomas (polyps >1cm).
Precancerous polyps, not cancer, is the most clinically relevant for the screening population.
**Gold standard with assumed sensitivity of 100%. Colonoscopy is estimated to miss up to 10% of
advanced lesions (>1cm) and up to 35% of smaller lesions
Exploiting the Field Effect
Field Effect: the genetic/environmental milieu that results in a neoplastic
lesion in one area of the colon exists throughout the organ.
• Conventional wisdom: tissue in a
tumor is abnormal, tissue
surrounding tissue is normal.
• This is only an approximation
• New methodology: detection of
carcinogenesis by analysis of
normal appearing cells in an
accessible part of an organ.
Commonly Used Biomarkers of Field
Effect in CRC
• Focal Neoplastic Lesions to predict Proximal Neoplasia
• Adenoma on flexible sigmoidoscopy
(Arch Intern Med. 2004 1881-7).
• Rectal aberrant crypt foci
(Takayama et al N Engl J Med. 1998; 339:1277.)
• Diffuse Alterations in the Histologically Normal Mucosa
• Decreased apoptosis
(Bernstein et al. Cancer Res. 1999 59:2353-7)
• Increased proliferation
(Ponz de Leon, et al Cancer Res. 1988 ;48:4121).
Novel Biomarkers of Field Effect in CRC
Microarray evidence
Proteomic evidence
TUMOR
Chen et al. Cancer Res 2004
Polley et al. Cancer Res 2006
Optically-detectable Markers of the
Field Effect
• Our approach: sensing changes in
tissue that CANNOT be detected by
histopathology, spatially outside the
extent of a neoplastic lesion.
• Tissue physiology: Increased mucosal microvascular blood supply
• Tissue morphology: Alterations in tissue fractal microarchitecture
• Intracellular morphology: Alterations in intracellular nanoscale architecture
Lessons from Animal Studies
 AOM-treated rat and MIN-mouse models
 Early increase in microvascular blood supply (EIBS) and alterations in tissue
micro and nano-architecture develop prior to ACFs and microadenomas.
 These alterations can be detected at a distance from the neoplastic focus.
Pericryptal capillary network
 EIBS is caused in part by iNOS
upregulation.
Gastroenterology, 126, 1071-1081 (2004)
Clinical Cancer Research, 19, 961-968 (2006)
FEBS, 581, 3857-3862 (2007)
Gut, 54, 654-660 (2005)
Epithelium
Lamina
propria
Muscularis
mucosa
Submucosa
 Depth-resolution is crucial: changes are only
in the mucosa (top 100 m).
Mucosa
 Diffusely present in ~90% of tissue sites.
Muscularis
propria
Arterioles
Mesenteric artery
Optically-detectable Markers of the
Field Effect
• Tissue physiology: Increased microvascular blood supply
• Tissue morphology: Alterations in tissue microarchitecture
• Intracellular morphology: Alterations in intracellular nanoscale architecture
EIBS: In Vivo Clinical Validation
• Technology: polarization-gated spectroscopy sensitive to mucosal
microcirculation
• Design: in vivo, during colonoscopy
• Patient characteristics:
• 220 average risk screening patients
• 51 with adenomas: 30 non-advanced adenomas, 9 multiple nonadvanced adenomas,12 advanced
• 169 patients with no neoplasia including 26 with hyperplastic polyps
colonoscope
EIBS fiber-optic probe
Oxygenated Hb concentration
2
*
*
1.6
*
1.2
0.8
Adenoma
<10 cm
from
adenoma
<30 cm
from
adenoma
>30 cm
from
adenoma
3
2.5
P-value=N.S.
2
1.5
1
0.5
Hyperplastic
Adenoma
polyp
1.4
<10 cm
10-30 cm >30 cm
No adenoma or
from
from
from
hyperplastic
hyperplastic hyperplastic hyperplastic
polyp
Distance from hyperplastic polyps
polyp
polyp
polyp
*
50
40
*
*
30
*
1
*
1.2
2
*
3
4
5
*
1
0.8
Control
Adenoma
Distance from adenomas
(c)
(b)
0
(b)
Oxygenation, %
* Indicates significantly different from control group
(a)
(d)
Deoxygenated blood content
Deoxygenated
Hb concentration
(normalized)
Deoxygenated blood content in
Superficial
DeOxy-Hb
Deoxygenated
Hb
concentration
superficial
tissue,
a.u.
Deoxygenated Hb concentration
EIBS: In Vivo Clinical Validation
3
<10 cm
from
adenoma
<30 cm
from
adenoma
>30 cm
from
adenoma
Control
Distance from adenomas
p<0.01
all p-values<0.01
2
1
0
Advanced
Diminutive
10-30
10-30
1
2
3 cm
4 cm
adenoma adenoma
from
from
advanced diminutive
adenoma adenoma
5
Normal
control
5
*
*
hemoglobin
oxyRectal "superficial"
EIBS
Rectal
Rectal EIBS is Indicative of Presence of
Adenomas Throughout the Colon
4
Patients with no dysplasia vs.
patients with advanced adenomas
3
Area under ROC curve =90%
2
No
Dysplasia
Single nonadvanced
adenomas
Multiple
nonadvanced
adenomas
Advanced
adenomas
EIBS reading: rectum only
Adenoma location: throughout the colon
Sensitivity = 100%
Specificity = 75%
Optically-detectable Markers of the
Field Effect
• Tissue physiology: Increased microvascular blood supply
• Tissue morphology: Alterations in tissue (fractal) microarchitecture
• Intracellular morphology: Alterations in intracellular nanoscale architecture
Alterations in Mucosal
Microarchitecture: Clinical Study Results
• Technology: low-coherence enhanced backscattering (LEBS)
spectroscopy
• Design: rectal biopsy
• Patient characteristics:
• 233 patients undergoing screening colonoscopy
• Mean age 56.8 ±10.7
• 47% female
• 60 with adenomas (17 advanced adenomas)
• 9 with previous h/o adenomas but none on present colonoscopy
• 158 with no current, prior or family history of adenomas
Human Clinical Study Results
2
*
1
AUC = 0.8947
(Advanced Adenoma)
0.8
1
*
Sensitivity
LEBS Marker
ANOVA p-value = 5 x 10-6
0
0.6
AUC = 0.7066
(Any Adenoma)
0.4
-1
0.2
-2
No
Dysplasia
0-4mm
Adenoma
0
0
5-9mm
Advanced
Adenoma Adenoma
LEBS FOBT**
0.2
Sensitivity*
100%
11%
Fecal
DNA**
18%
Specificity*
88%
80%
95%
94%
PPV
42%
NPV
100%
0.4
0.6
1 - Specificity
0.8
1
Potential Confounding Factors Do Not
Appear to Affect LEBS Diagnosis
Normal
Hyperplastic
Polyp
Hemorrhoids
Diverticulosis
Diverticulitis
LEBS Marker
0
ANCOVA
p-value
-1
-2
presence of
neoplasia
0.0000
smoking history
0.0209
race
0.2235
alcohol history
0.0443
gender
0.5847
BMI
0.5857
medication history
0.9278
age
0.6526
ANOVA p-value = 0.91
-3
age
BMI
correlation
coefficient
0.10
-0.038
regression
p-value
0.20
0.64
Optically-detectable Markers of the
Field Effect
• Tissue physiology: Increased microvascular blood supply
• Tissue morphology: Alterations in tissue microarchitecture
• Intracellular morphology: Alterations in cell nanoscale architecture
Alterations in Epithelial
Nanoarchitecture: Clinical Study Results
• Technology: partial wave spectroscopic (PWS) microscopy
• Design: rectal mucosal brushings
• Patient characteristics:
• 35 patients
• 21 with no neoplasia
• 14 with adenomas, 4 advanced
1 µm
1 µm
Human Clinical Study Results
Disorder strength Ld (mm)
x 10-7
13
12
* p-value < 0.0001
Sensitivity
100%
Specificity
95%
*
no neoplasia
11
neoplasia
*
10
9
8
7
Control
Adenoma
Advanced
adenoma
Disorder in nanoscale density fluctuations in endoscopically-normal rectal
mucosa is increased in patients with sporadic adenomas
Colonoscopy-free Screening for
Colon Cancer Using Optical
Detection of the Field Effect
Annual population screening by PCP’s during an annual exam
without colonoscopy and preparation
+
Colonoscopy
_

 Only patients with adenomas receive colonoscopies
 Most (all) patients with adenomas are screened
 Patients are more compliant
 Better allocation of colonoscopic resource
LEBS probe
Does LEBS Work in Other Organs?
Does LEBS Work in Other Organs?
Example: Pancreatic Cancer
• No existing technique is capable of
accurate diagnosis of pancreatic
carcinogenesis in preinvasive (PanIN) or
resectable stage.
• 95% mortality within a year after
diagnosis.
• Problem: pancreatic duct exam is not
suitable for screening due to a high rate of
complications including acute pancreatitis
(~5-20%).
• Solution: PWS analysis of duodenal
periampullary cells brushed during upper
endoscopy.
Prospect of Pancreatic Cancer Screening
LEBS (Microarchitecture):
204 patients total
84 Healthy control
26 Family History
29 Cyst
45 Pancreatic Adenocarcinoma
20 Other diseases
Testing set:
Sensitivity = 85%
Specificity = 80%
PWS (Nanoarchitecture):
35 patients total
26 Healthy control
9 Pancreatic Adenocarcinoma
Sensitivity = 90%
Specificity = 81%
Example III: Lung Cancer
• >80% of lung cancer patients had altered nuclear texture features
Us-Krasovec et al., Anal Quant Cytol Histol. 2005 Oct;27(5):254-62
•Automated Quantitative cytology of buccal nuclei correlated with lung
cancer
• 66% sensitivity and 70% specificity for lung cancers
• 61% sensitivity for stage 1
Turic et al, Chest 2005 (abstract)
• Increased incidence of head and neck cancer in patients with lung cancer
Johnson et al., B. J. Natl. Cancer Inst., 1998
• Genetic changes in the histologically normal large-airway epithelial cells
obtained at bronchoscopy.
Guo, M. et al. Clin. Cancer Res. 2004.
• 80% sensitivity and 84% specificity
Spira et al., Nat Med 2006.
Lung Cancer Screening by PWS
Analysis of Buccal Cells
Number of Patients: 108







Normal - Nonsmokers
Cancer - Nonsmokers
COPD (smokers)
COPD – Family Hx
Lung cancer (smokers)
Other cancers
Smokers (no COPD or cancer)
-
5
3
31
7
53
4
5
PWS
PWS
Bright field
Non-cancer
(COPD) patients
Disorder strength (Ld), um
Bright field
Lung cancer
patients
PWS Images are Different for Noncancer and Cancer Patients
1.00
Disorder in Nanoarchitecture is
Increased in Buccal Cells in Lung
Cancer Patients
4.E-06
0.25
3.E-06
0.50
0.75
P<0.001
Sensitivity
2.E-06
COPD-all
Area under ROC curve = 84%
Cancer
0.00
Disorder strength (Ld ) (um)
5.E-06
Control
(COPD)
Lung
cancer
0.00
0.25
Area under ROC curve = 0.8393
Sensitivity = 90%
Specificity = 77%
0.50
1 - Specificity
0.75
1.00
Potential Confounding Effects
• Can the differences in cell nanoarchitecture be simply due to
difference in age among COPD and lung cancer patients?
• Can the differences be due to different smoking history?
Age
Pack years
Control
71±9
78±38
Cancer
70±12
59±46
Demographic factor
P-value (Effect on Ld)
P-value (Effect on SDLd)
Age
0.56
0.45
Smoking
0.65
0.48
Race
0.60
0.40
Gender
0.79
0.93
Conclusions
• What we call “histologically normal tissue” is not entirely normal in patients
with neoplasia
• Not only neoplastic lesions but also tissue outside neoplastic lesions is
abnormal
• Biophotonics can detect field effects associated with carcinogenesis in the
colon, pancreatic and lung
• Optically-detectable markers of the field effect include increased mucosal
blood supply, micro and nano-architectural changes in the mucosa
• Potential for colon, pancreatic and lung cancer screening/risk-stratification
through biophotonics detection of the field effect
Acknowledgements
Northwestern University
Vladimir Turzhitsky
Andrew Gomez
Hariharan Subramanian
Sarah Ruderman
Jeremy Rogers, Ph.D.
Young Kim, Ph.D.
Yang Liu, Ph.D.
Prabhakar Pradhad, Ph.D.
Xu Li, Ph.D.
Alexei Kromine
Evanston Hospital
Hemant Roy, M.D.
Ramesh Wali, Ph.D.
Randall Brand, M.D.
M. Goldberg, M.D.
Funding
National Institutes of Health
R01CA128641, R01 CA109861,
R01 EB003682, R01 CA112315,
R01 CA118794, R33CA122017,
R21 EB006742, U01 CA11125
National Science Foundation
CBET- 0733868, CBET-0238903
V Foundation
Coulter Foundation
AACR