1
Detection of Single Ring Stage P. falciparum in
Human Thin Film Blood Smears Using FTIR
Microspectroscopy and Differentiation of
Plasmodium Positive from Plasmodium Negative Red
Blood Cells
by McKale Santin, Dr. Bryan Holmes, Dr. Adam
Hunt, and Kenneth A. Puzey of QuantaSpec, Inc.
Contact kpuzey_at_quantaspec.com
Spectra collected from a 13x13 ?m area of a
sing-cell layer of each thin film smear (1-3
RBCs). Mid-IR spectra collected from 4000cm-1 to
600cm-1 at a spectral resolution of 2cm-1. 100
scans/measurement. Background measurements taken
from an adjacent, blank 13x13 ?m area of each
slide, interactively subtracted from sample
spectra using Opus 6.5 software. Data Processing.
Raw spectral data organized into two
classification groups based on sample identity
Plasmodium-positive and Plasmodium-negative. Data
imported into Excel?, 1st derivative calculated
by taking the slope of the raw data. 1st
derivative data imported into JMP? software and
multivariate discriminant analysis performed on
all spectra. Mahalanobis distances calculated for
each classification group for each replicate.
Algorithm Development. Identification algorithm
developed to determine Plasmodium spp. infection
based on processed IR spectra. The algorithm
consists of a set of vectors that are multiplied
with the first derivative spectra of an unknown
sample to be identified. For algorithm
development, the full set of absorbance values
(all optical frequencies) is replaced by a much
smaller subset of data containing 350 key optical
frequencies for identification.
ABSTRACT Currently, rapid diagnostic
tests for malaria infection perform poorly at low
parasite loads, are degraded by severe
temperatures, and contain reagents, which
contribute to their costs. The overall objective
of this study was to perform a preliminary
evaluation of the utility of FTIR
microspectroscopy for in vitro diagnosis of thin
film blood smears for malaria infection. FTIR
microspectroscopy has potential advantages in
detecting low parasite loads, is not affected by
temperature, and does not require any reagents.
Giemsa-stained thin film blood smear slides were
analyzed in this study. 240 slides with ring
stage P. falciparum infected human blood were
prepared from culture. P. falciparum negative
controls included 80 clinical P. vivax slides
(collected and verified by expert microscopy
(EM), 40 slides with Salmonella- infected human
blood (prepared from culture), and 40 uninfected
human blood slides. Infrared spectra were
measured from a small area of each slide (13
microns x 13 microns) usually containing only one
red blood cell. Algorithms were written to
differentiate Plasmodium positive spectra from
Plasmodia negative spectra and tested by
cross-validation. The sensitivity was 98.8 to
100 and the specificity was 95.4 to 100 for
Plasmodia positive samples with a 95 confidence
interval. These results suggest that further
study of FTIR spectroscopy as an automated
reagent-less diagnostic method with potential for
detection of single parasites is warranted.
Infrared spectroscopy could radically lower
marginal test costs by eliminating the need for
expensive consumables.
Figure 3. Visual Images, 740X magnification.
Ring-stage P. falciparum infected red blood cells
(left, red arrow) and uninfected red blood cells
(right). Each sample was visually located, then
the microscope was switched to IR mode to collect
reflectance-absorbance spectra from the center
square.
INTRODUCTION Malarial infection is
a major global health problem. A key part of
malaria control strategies is early case
detection with in vitro diagnostics. The current
gold standard for malaria diagnosis is expert
microscopy of Giemsa-stained blood smears,
however this method has many limitations. It is
labor-intensive, requires consistent, quality
staining, and requires diagnosis by a trained,
expert microscopist, which are in short supply.
Furthermore, most malaria patients are treated in
peripheral healthcare facilities that do not have
access to quality microscopy. Simple rapid
diagnostic tests (RDTs) based on detection of
parasite antigens have been introduced to try and
provide an alternative to diagnosis with
microscopy, yet these diagnostics also have many
limitations. RDTs are poor at detecting
low-level parasitemia, have a limited shelf life,
and contain reagents, which contribute to their
cost. Our present study investigates
the feasibility of using infrared (IR)
microspectroscopy as an alternative diagnostic
approach that can overcome the limitations
inherent to diagnosis based on analysis of a
visible image or by reagent-based assays. Fourier
Transform Infrared (FTIR) microspectroscopy can
probe the entire chemistry of an intact
biological cell with IR light instead of
reagents. The spectral signatures of biological
cells vary depending on the molecular components
of the cell, and the chemical alterations that
accompany infection provide the basis for this
detection technology. The goal of this research
was to evaluate FTIR microspectroscopy for
automatically differentiating Plasmodium-positive
from Plasmodium-negative red blood cells in thin-
film blood smears.
Figure 4. Absorbance Spectra. 40 absorbance
spectra from P. falciparum strain 3D7 infected
blood (left) and 40 absorbance spectra from
uninfected human blood (right).
RESULTS Visual images of uninfected and
ring-stage P. falciparum infected red blood cells
at 740X magnification are shown in Figure 3.
Spectra were taken from the center square all
other IR light was blocked off by perpendicular
apertures. Spectra in the mid-IR region of
Plasmodium-positive (P. falciparum strain 3D7)
and Plasmodium-negative (uninfected blood) are
shown in Figure 4. From the computed Mahalanobis
distances, it was found that the longest
within-group distances were small (103) when
compared to the shortest across-group distances
(1012). Cross validation testing was used to
evaluate the accuracy of the developed algorithm
for Plasmodium spp. detection. The algorithm
correctly identified 280 out of 280 true
positives and 80 out of 80 true negatives. The
sensitivity of the developed algorithm was
98.8-100 (95CI) and the specificity was
95.4-100 (95CI).
Table 1. Sample Characterization.
Plasmodium-positive controls contained P.
falciparum and P. vivax infected human blood.
Plasmodium-negative controls contained
Salmonella-infected human blood and uninfected
human blood.
MATERIALS and METHODS Positive and Negative
Controls. Uninfected negative controls human
blood, 5 Hematocrit. Salmonella-infected
negative controls uninfected human blood spiked
with Salmonella SL1344. P. falciparum uninfected
blood spiked with strains 7G8, D6 (MR4) and
3D7,1776, HB3, Dd2 (NYU School of Medicine). P.
vivax samples were prepared from clinical cases
in India Parasite and bacterial counts can be
found in Table 1. Sample Preparation. 40
thin film blood smears prepared per control group
(Table 1). All samples prepared on low-e
microscope slides (transparent in the visible
region but highly reflective in the IR).
Samples were fixed and stained with a 10 Giemsa
solution. Spectral Data Collection. A Bruker
Hyperion 1000?infrared microscope and a Bruker
Tensor? 27 FTIR spectrometer were used to
collect the spectral data (Figures 1 2). This
system uses a glowbar IR source and a MCT
detector. The microscope was modified with a
high power reflective objective for an
overall magnification of 740X.
CONCLUSION Initial results indicate FTIR
microspectroscopy can be used as a rapid
identification tool for the detection of
Plasmodia in human thin film blood smears with
high sensitivity and specificity. All 320
replicates were correctly identified as either
malaria positive (240) or malaria negative (80),
supporting the hypothesis that FTIR
microspectroscopy can be used to detect
ring-stage P. falciparum infection. This
research study has also demonstrated the
potential for infrared microspectroscopy to
detect low-level parasitemia, as single parasites
were detected. We are currently working on
collecting a larger clinical sample set, and
increasing the number of red blood cells that can
be diagnosed simultaneously.
ACKNOWLEDGEMENTS This work is supported by
the U.S Army Medical Research and Materiel
Command under contract No.W81XWH- 09-C-0019. The
views, opinions and/or findings contained in this
report are those of the authors and should not be
construed as an official Department of the Army
position, policy or decision unless so designated
by other documentation. In the conduct of
research where humans are the subjects, the
investigator(s) adhered to the policies regarding
the protection of human subjects as prescribed by
Code of Federal Regulations (CFR) Title 45,
Volume 1, Part 46 Title 32, Chapter 1, Part 219
and Title 21, Chapter 1, Part 50 (Protection of
Human Subjects).
Figure 2. Bruker Hyperion 1000 IR microscope and
Tensor 27 FTIR spectrometer.
Figure 1. Hyperion 1000 optical beam path.
2
Automated Reagent-less Differentiation of P.
falciparum from P. vivax in Human Thin Film Blood
Smears With FTIR Microspectroscopy
by Kenneth A. Puzey, Dr. Bryan Holmes, Dr. Adam
Hunt, and McKale Santin of QuantaSpec, Inc.
Contact kpuzey_at_quantaspec.com
MATERIALS and METHODS Positive and Negative
Controls. Uninfected negative controls human
blood, 5 Hematocrit. Salmonella-infected
negative controls uninfected blood spiked with
Salmonella SL1344. P. falciparum uninfected
blood spiked with strains 7G8, D6 (MR4) and
3D7,1776, HB3, Dd2 (NYU School of Medicine). P.
vivax samples were prepared from clinical cases
in India. Sample Preparation. 40 thin-film blood
smears prepared per control group. All samples
prepared on low-e microscope slides (transparent
in the visible region but highly reflective in
the IR). Samples were fixed and stained with a
10 Giemsa solution. Spectral Data Collection. A
Bruker Hyperion 1000?infrared microscope and a
Bruker Tensor? 27 FTIR spectrometer were used to
collect the spectral data (Figures 1 2). This
system uses a glowbar IR source and a liquid
nitrogen-cooled MCT detector. The microscope was
modified with a high power reflective objective
for an overall magnification of 740X. Spectra
collected from a 13x13 ?m area of a sing-cell
layer of each thin film smear (1-3 RBCs). Mid-IR
spectra collected from 4000cm-1 to 600cm-1 at a
spectral resolution of 2cm-1. 100
scans/measurement. Background measurements taken
from an adjacent, blank 13x13 ?m area of each
slide, interactively subtracted from sample
spectra using Opus 6.5 software. Data Processing.
Raw spectral data organized into 3 classification
groups based on the identity of the spectral
sample P. falciparum-positive, P.
vivax-positive, and Plasmodia-negative. Data
imported into Excel?, 1st derivative calculated
by taking the slope of the raw data. 1st
derivative data imported into JMP? software and
multivariate discriminant analysis performed on
all spectra. Mahalanobis distances calculated for
each classification group for each replicate.
Algorithm Development. Identification algorithms
developed to determine P. falciparum infection,
P. vivax infection, or no infection based on
processed IR spectra. The algorithms consists of
a set of vectors that are multiplied with the
first derivative spectra of an unknown sample to
be identified. For algorithm development, the
full set of absorbance values (all optical
frequencies) is replaced by a much smaller subset
of data containing 350 key optical
frequencies for identification.
ABSTRACT In malaria cases species of
infection affects course of treatment.
Differentiation of P. falciparum from P. vivax by
RDTs requires multiple antibodies, which
increases test costs. Furthermore, RDTs are
subject to reader error. Speciation by visual
microscopy is dependent on the skill and
availability of an expert microscopist. The
objective of this study was to evaluate the
utility of FTIR microspectroscopy for automatic
reagent-less differentiation of P. falciparum
from P. vivax infected human red blood cells.
Geimsa-stained thin film blood smear slides were
analyzed in this study. For P. falciparum
positive controls, 240 slides with ring stage P.
falciparum were prepared from culture. For P.
vivax positive controls, 40 clinical P. vivax
slides were collected and verified by expert
microscopy (EM). For negative controls, 40
slides with Salmonella-infected blood (prepared
from culture) and 40 uninfected blood slides were
prepared. Infrared spectra were measured from a
small area of each slide (13 microns x13
microns) typically containing only one red blood
cell. Algorithms were written to differentiate
red blood cells infected with P. falciparum, red
blood cells infected with P. vivax, red blood
cells infected with Salmonella and uninfected red
blood cells based on their infrared spectra.
Algorithms were tested by cross-validation. For
P. falciparum sensitivity was 98.4 to 100 and
specificity was 97.7 to 100 (95 CI). For P.
vivax the sensitivity was 95.4 to 100 and the
specificity was 98.8 to 100 (95 CI). These
results suggest that FTIR spectroscopy may be
useful for automated reagent-less differentiation
of malaria infection. In high throughput
settings spectroscopy testing may be lower
cost because it does not require consumables.
INTRODUCTION Over 3 billion people
worldwide are at risk of malaria, representing
almost half of the worlds population. Prompt and
correct diagnosis of malarial infection is a
primary part of malaria control and is essential
for saving patient lives. In regions where both
Plasmodium falciparum and Plasmodium vivax are
present, effective diagnosis requires not only
detecting malaria infection but also determining
the species of infection, as different species
respond to different chemotherapeutic treatments.
Expert microscopy remains the gold standard for
distinguishing different species of malarial
infection, but unfortunately high-quality expert
microscopy is difficult to maintain in
resource-poor settings where the majority of
malaria diagnosis is being performed. Rapid
diagnostic tests (RDTs) based on detection of
species-specific antigens such as pLDH (parasite
lactate dehydrogenase) have been introduced to
provide an alternative to diagnosis with
microscopy. However, RDTs have many limitations.
They are poor at detecting low-level
parasitemia, have a limited shelf life, and
contain reagents, which contribute to their cost.
Our present study investigates the feasibility
of using infrared (IR) microspectroscopy as an
alternative diagnostic approach that can overcome
the limitations inherent to diagnosis based on
analysis of a visible image or by reagent-based
assays. Fourier Transform Infrared (FTIR)
microspectroscopy can probe the entire chemistry
of an intact biological cell with IR light
instead of reagents. The spectral signatures of
biological cells vary depending on the molecular
components of the cell, and the chemical
alterations that accompany infection provide the
basis for this detection technology. The goal of
this research was to evaluate FTIR
microspectroscopy for automatically
differentiating Plasmodium falciparum from
Plasmodium vivax infected red blood cells
in thin- film human blood smears.
Figure 3. Visual Images, 740X magnification. P.
vivax infected red blood cells (left, blue arrow)
and P. falciparum infected red blood cells
(right, red arrow). Each sample was visually
located, then the microscope was switched to IR
mode to collect reflectance-absorbance spectra
from the center 13x13 ?m square.
RESULTS Visual images of P. vivax and P.
falciparum infected red blood cells at 740X
magnification are shown in Figure 3. Spectra
were taken from the center square all other IR
light was blocked off by perpendicular apertures.
From the FTIR absorbance spectra, P. vivax and P.
falciparum cannot be visually distinguished. From
multivariate analysis, Mahalanobis distances were
calculated for each replicate to every other
replicate. It was found that the longest
within-group distances are small (103) when
compared to the shortest across-group distances
(1012). Cross validation testing was used to
evaluate the accuracy of the developed algorithm
for Plasmodium spp. detection. The algorithm
correctly identified 240 out of 240 true
positives and 80 out of 80 true negatives. The
sensitivity of the P.f. identification algorithm
was 98.4-100 (95CI) and the specificity was
97.7-100 (95CI). The sensitivity of the P.v.
identification algorithm was 95.4-100 (95CI)
and the specificity was 98.8-100(95CI).
? Camera
CONCLUSION Initial results indicate that
RBCs infected with P.f. can be differentiated
from RBCs infected with P.v. in thin film blood
smears using FTIR microspectroscopy with high
sensitivity and specificity. This method is
reagent-less and automated (results provided by
computer) and is capable of detecting a single
Plasmodia parasite. Further study with slides
from both clinical P.f. and clinical P.v. from a
larger number of cases will be needed to
determine the clinical utility of FTIR
microspectroscopy for diagnosis and such a study
is underway. Equipment modifications to examine a
large number of RBCs in parallel are also
underway to improve diagnostic throughput.
ACKNOWLEDGEMENTS This work is supported by
the U.S Army Medical Research and Materiel
Command under contract No.W81XWH- 09-C-0019. The
views, opinions and/or findings contained in this
report are those of the authors and should not be
construed as an official Department of the Army
position, policy or decision unless so designated
by other documentatio In the conduct of
research where humans are the subjects, the
investigator(s) adhered to the policies regarding
the protection of human subjects as prescribed by
Code of Federal Regulations (CFR) Title 45,
Volume 1, Part 46 Title 32, Chapter 1, Part 219
and Title 21, Chapter 1, Part 50 (Protection of
Human Subjects).
Figure 2. Bruker Hyperion 1000 IR microscope and
Tensor 27 FTIR spectrometer.
Figure 1. Hyperion 1000 optical beam path.
3
Automated Reagent-less Differentiation of Three
Drug Susceptible Strains of P. falciparum from
Three Drug Resistant Strains of P. falciparum in
Human Thin Film Blood Smears Using FTIR
Microspectroscopy
by Kenneth A. Puzey, Dr. Bryan Holmes, Dr. Adam
Hunt, and McKale Santin of QuantaSpec, Inc.
Contact kpuzey_at_quantaspec.com
microscope slides (transparent in the visible
region but highly reflective in the IR). Samples
were fixed and stained with a 10 Giemsa
solution. Spectral Data Collection. A Bruker
Hyperion 1000?infrared microscope and a Bruker
Tensor? 27 FTIR spectrometer were used to collect
the spectral data (glowbar IR source/ liquid
nitrogen-cooled MCT detector). 74X reflective
objective used to collect spectra from a 13x13 ?m
area of a sing-cell layer of each thin film smear
(1-3 RBCs). Mid-IR spectra collected from
4000cm-1 to 600cm-1 at a spectral resolution of
2cm-1. 100 scans/measurement. Background
measurements taken from an adjacent, blank 13x13
?m area of each slide, subtracted using Opus 6.5
software. Data Processing. Raw spectral data
organized into 3 classification groups based on
the identity of the spectral sample drug
susceptible P.f.-positive, drug-resistant
P.f.-positive, and P.f.-negative. Data imported
into Excel?, 1st derivative calculated by taking
the slope of the raw data. 1st derivative data
imported into JMP? software and multivariate
discriminant analysis performed on all spectra.
Mahalanobis distances calculated for each
classification group for each replicate.
Algorithm Development. Identification algorithms
developed to differentiate drug-susceptible
P.f.-positive from drug- resistant P.f.- positive
from P.f.-negative infection based on processed
IR spectra. The algorithm consists of a set of
vectors that are multiplied with the first
derivative spectra of an unknown sample to be
identified. For algorithm development, the full
set of absorbance values (all optical
frequencies) is replaced by a much smaller subset
of data containing 350 key optical
frequencies for identification.
ABSTRACT In some regions of the world
malaria parasite drug resistance is present in
50 of cases. Unfortunately, tests to determine
drug resistance are not clinically available
forcing health ministries and doctors to make
difficult choices. An economical clinical test
for drug resistance would enable doctors to
administer less expensive chloroquine to
susceptible cases, lowering health costs and
slowing the spread of resistance to newer drugs.
The objective of this study was a preliminary
evaluation of the utility of FTIR
microspectroscopy for differentiating red blood
cells infected with drug resistant strains and
drug susceptible strains of P. falciparum. 120
Geimsa-stained thin film blood smear slides were
prepared with drug-susceptible ring stage P.
falciparum from culture (40 slides strain 3D7, 40
slides strain 1776, 40 slides D6), and 120
Geimsa-stained thin film blood smear slides were
prepare with drug-resistant ring stage P.
falciparum from culture (40 slides strain HB3, 40
slides strain Dd2, 40 slides strain 7G8).
Negative controls included 40 Geimsa-stained thin
film blood smear slides of uninfected human blood
as well as human blood infected with Salmonella
from culture (40 slides). Additional P.
falciparum negative controls included 40 clinical
Geimsa-stained P. vivax slides collected and
verified by expert microscopy (EM). Infrared
spectra were measured from a small area of each
slide (13 microns x13 microns) typically
containing only one red blood cell. Algorithms
were written to differentiate red blood cells
infected with P. falciparum, red blood cells
infected with P. vivax, red blood cells infected
with Salmonella and uninfected red blood cells
based on their infrared spectrum. Algorithms
were tested by cross-validation. For drug
susceptible strains, sensitivity was 97 to 100
and specificity was 98.7 to 100 (95 CI). For
drug resistant strains sensitivity was 97 to
100 and specificity was 98.7 to 100 (95 CI).
These results suggest that FTIR spectroscopy may
be useful for automated reagent-less
differentiation of drug resistant and drug
susceptible strains of P. falciparum in thin film
blood smears. This capability could enable more
cost effective case management and reduce
the spread of drug resistance to newer drugs.
Figure 1. Absorbance Spectra. 40 absorbance
spectra from drug susceptible P. falciparum D6
(left) and 40 absorbance spectra from drug
resistant P. falciparum Hb3 (right).
INTRODUCTION Each year, there are
an estimated 250 million malaria cases and
approximately 1 million malaria-related deaths.
Fundamental to reducing the burden of malaria
infection and improving patient outcome is rapid
and accurate diagnosis. Field diagnosis and
treatment of malarial infection in
malaria-endemic regions remains a problem, and is
becoming increasingly difficult due to malaria
parasite drug resistance. Major methods for
malaria diagnosis (expert microscopy and rapid
diagnostic tests) are unable to detect drug
resistance prior to treatment, and instead are
used to monitor for treatment failure. This
method is time consuming, with prolonged periods
of patient follow-up, and is costly due to the
need for multiple tests. Regardless of the
species of malaria, all drug resistance
mechanisms involve genetic and chemical
differences in the parasite. Our present study
investigates the feasibility of using infrared
(IR) microspectroscopy as an alternative
diagnostic approach that can overcome the
limitations inherent to diagnosis based on
analysis of a visible image or by reagent-based
assays. Fourier Transform Infrared (FTIR)
microspectroscopy can probe the entire chemistry
of an intact biological cell with IR light
instead of reagents. The chemical differences
between strains provides the basis for this
detection technology. The goal of this research
was to evaluate FTIR microspectroscopy for
automatically differentiating drug susceptible P.
falciparum from drug resistant P. falciparum in
thin-film blood smears.
Figure 2. 1st Derivative of Absorbance Spectra.
40 1st derivatives from drug susceptible P.
falciparum D6 spectra (left) and 40 1st
derivatives from drug resistant P. falciparum Hb3
spectra (right).
RESULTS Spectra in the mid-IR region of drug
susceptible P. falciparum D6 and drug resistant
P. falciparum Hb3 are shown in Figure 1. From
the FTIR absorbance spectra, drug susceptible and
drug resistant species cannot be visually
distinguished. Figure 2 shows the 1st derivative
of the absorbance data for drug susceptible P.f
D6 and drug resistant P.f Hb3. Table 1 shows the
computed Mahalanobis distances using all optical
frequencies. The chart shows within-group maximum
distances (green, diagonal values) and
between-group minimum distances (red values). It
was found that the longest within-group distances
are small (103) when compared to the
across-group shortest distances (1012). Cross
validation testing was used to evaluate the
accuracy of the developed algorithms for drug
susceptible P.f positive, drug resistant P.f
positive, and P.f negative. The algorithm
correctly identified 120/120 drug susceptible P.f
replicates, 120/120 drug resistant P.f replicates
and 120/120 P.f negatives. The sensitivity of the
developed algorithm for drug-susceptible P.f. was
97-100 (95CI) and the specificity was 98.7-100
(95CI). The sensitivity of the developed
algorithm for drug-resistant P.f. was
97-100(95CI) and the specificity was 98.7-100
(95CI).
CONCLUSION Initial results indicate that
FTIR IR microscopy can differentiate
drug-susceptible and drug-resistant strains of
P.f. (for the strains that were studied).
Additional studies with additional strains would
be of interest to determine if the initial
results are valid for a wider variety of strains.
In addition, it was found that IR spectra could
be used to accurately differentiate all six
strains from each other as well as from P.v. and
negative controls with sensitivity of
91.19-100(95CI) and with a specificity of
98.98-100(95CI). Therefore, further study of
FTIR microspectroscopy as an alternative
diagnostic method is warranted.
Table 1. Calculated Mahalanobis Distances. The
vertical column represents the actual identity of
the sample and the horizontal row represents the
comparison group. Values were calculated using
all optical frequencies measured from
4000-600cm-1. Values in green represent the
furthest distance between samples in the same
identity group, and values in red represent the
closest distance between sample from different
identity groups.
MATERIALS and METHODS Positive and Negative
Controls. Uninfected negative controls human
blood. Salmonella/ P. vivax infected negative
controls uninfected blood spiked with Salmonella
SL1344, P. vivax clinical cases from India. P.
falciparum drug susceptible uninfected blood
spiked with strains 3D7, or 1776, or D6. P.
falciparum drug resistant uninfected blood
spiked with strains HB3, or Dd2, or 7G8. Sample
Preparation. 40 thin-film Blood smears prepared
per group. All samples prepared on low-e
ACKNOWLEDGEMENTS This work is supported by
the U.S Army Medical Research and Materiel
Command under contract No.W81XWH- 09-C-0019. The
views, opinions and/or findings contained in this
report are those of the authors and should not be
construed as an official Department of the Army
position, policy or decision unless so designated
by other documentation. In the conduct of
research where humans are the subjects, the
investigator(s) adhered to the policies regarding
the protection of human subjects as prescribed by
Code of Federal Regulations (CFR) Title 45,
Volume 1, Part 46 Title 32, Chapter 1, Part 219
and Title 21, Chapter 1, Part 50 (Protection of
Human Subjects).