«Novel Biophotonic Imaging Techniques for Assessing Women’s Reproductive Health by Tyler Kaine Drake Department of Biomedical Engineering Duke ...»
From these early experiments, it was concluded that multiplexing, or 22 recording LCI data over several parallel channels, could improve tissue coverage without the need for a mechanical scanner.
Lou et al.
demonstrated the feasibility of a similar parallel optical coherence tomography system which used eight channels to increase scanning area without a mechanical scanning galvanometer.49 This system was in the time- domain and failed to take advantage of SNR gain and improved speed of a Fourier- domain multiplexed device.
2.3 Cervical Cancer 2.3.1 Cervical Intraepithelial Neoplasia and risk factors Cervical Intraepithelial Neoplasia (CIN), which is also known as cervical dysplasia, is defined by the presence of abnormal growth of squamous cells in the epithelium of the cervix, and is the precursor lesion to cervical cancer.50 Not all cases of CIN progress to cervical cancer;
they may remain stable or be eliminated by the immune system of the body.
CIN progresses in severity of time from CIN1 (mild epithelial dysplasia) to CIN2 (moderate dysplasia) to CIN3 (severe dysplasia).
While not all CIN lesions become invasive carcinomas, the risk increases with progression as CIN1 lesions have a 1% chance and CIN3 lesions have a 12% chance of becoming cervical carcinomas.50, 51 Therefore, early detection is paramount for successful therapeutic intervention.
While almost all cases of cervical cancer are caused by infection with human papillomavirus (HPV)52, HIV and HPV share common risk factors.
HIV- infected women have a higher cervical cancer prevalence than HIV- negative women.53 HIV- infected 23 women also have a five- fold increase in risk of developing CIN lesions.53 This co- infection risk increase could help explain why although HPV is necessary for cervical cancers and their CIN precursors, other co- factors are also involved which increases the likelihood of CIN progression.
Therefore, with widespread use of HPV vaccination still in its early stages, the need for cervical screening remains essential for effective cervical cancer management.
2.3.2 Cervical cancer screening Current screening techniques for cervical cancer are based on technologies that were developed in the early 1900s.3 Cervical cytology, or Papanicolaou test, has been an effective screening tool by detecting the possible presence of CIN and other abnormalities in the endocervix and endometrium, and has had success in reducing morbidity.3 Pap tests are limited in that they are subjective and time consuming technique with a wide range of sensitivity and specificity values, of 68- 80% and 75- 90%, respectively.33 They are also limited in tissue coverage because they only sample randomly exfoliated cells.
If a woman has a positive Pap test, she returns at least 1- 2 weeks later for a colposcopy examination and/or physical biopsy.
Colposcopy is the current standard of care for diagnosis of dysplasia in the ectocervix region.
However, as with cervical cytology, it is a subjective technique which limits its accuracy in diagnosis because the skill of the colposcopist affects efficacy.
An acetic acid wash is often used with colposcopy to create contrast between normal tissue 24 and lesions, which stain a white color from the acetowhite effect.54 Subjectivity and variability limits the sensitivities and specificities of colposcopy as shown by large ranges in performance, 50- 96% and 43- 86%, respectively.30- 32 Furthermore, in colposcopy, a challenge exists in differentiating dysplastic and metaplastic tissues, since cervical cancer commonly arises is the metaplastic transformation zone.
Metaplastic tissue and low grade dysplasias stain the same opaque white color due to the acetowhite effect, making them difficult to differentiate visually.54 If colposcopy identifies suspicious tissue, it will be biopsied, processed, and examined by a pathologist.
If dysplasia is found (CIN2 or worse), the woman will return for treatment.
The current screening process is troubled at several stages.
Pap tests often suffer from inadequate sampling of cells and insufficient time for full screening.
Colposcopy is time consuming and subjective.
Repeat visits are necessary throughout the process, as it typically will take 2- 8 weeks from a positive Pap test to eventual treatment, making it time consuming and expensive.
The introduction of two FDA- approved HPV vaccines was important to the primary prevention of cervical cancer, as these have 90% efficacy in preventing CIN- 2+ caused by vaccine types HPV- 16 and HPV- 18.
The 9- valent vaccine is under development but its impact is yet unknown.
However, the critical need for more effective secondary prevention with improved cervical screening and diagnosis persists, for several reasons.
First, vaccination does not alter the cervical screening schedule since 25 the vaccine does not treat pre- existing infections, let alone all types of HPV.55, 56 Second, to achieve high efficacy, the 3- dose vaccine must be given at an HPV- naïve age, typically 11- 12 yrs.57 Unfortunately, U.S.
vaccine uptake remains low and has been complicated by issues of parental and provider acceptance, and overall less contact with healthcare providers during adolescence.58, 59 In 2011, only 53% of girls aged 13- 17 years received 1 or more doses and only 32% received all 3.60 Third, vaccine access is extremely limited in many developing countries due to cost and lack of experts in the field.
2.3.3 Optical detection techniques Optical imaging techniques have recently been developed which can be used for screening for cervical cancer and CIN in vivo.61 Dysplasia causes a number of changes to tissue, such as alterations of gross architecture, cell morphology, vasculature, and metabolic activity.
These changes, in turn, alter the optical properties of tissue which can be exploited for detection.
Wide- field imaging relies on changes in autofluorescence and reflectance at multiple wavelengths to create contrast between normal and dysplastic tissues.
It uses wide area sensors to image large areas of the cervix.
Wide- field imaging can achieve resolutions of approximately 50- 100 µ;m, but can only image the surface of the tissue.61 A study of wide- field reflectance images of the cervix from 29 women showed that high- grade pre- cancers could be detected, using with an automated image analysis algorithm with a sensitivity of 79% and specificity of 88% when compared to pathological 26 assessment.61, 62 Several larger studies with wide- field imaging using reflectance and autoflourescence have been completed with detected dysplasia with sensitivities from 92- 95% and specificities from 50- 70%.63- 65 Optical spectroscopy can be used to detect fluorescence and reflectance spectra from a selected point on the cervical epithelium with a fiber- optic probe.
It measures the intensity of light reflected or emitted as a function of wavelength which can provide information about changes in the epithelium structure.
It has a spatial resolution on the order of 1 mm, and a depth penetration of 0.3- 1 mm.
Optical spectroscopy typically utilizes a high spectral resolution of 1- 5 nm, which can provide information on molecules in the tissue.61 Several studies have been performed which found optical spectroscopy capable of detecting CIN with sensitivities ranging from 83- 92% and specificities of 80- 90%.66, 67 High resolution imaging approaches, such as confocal microscopy, have also been developed to image cervical epithelium.
These microscopes are capable of a spatial resolution of 1- 2 µ;m, but they are limited with depth penetration of 1mm.61 Acetic acid is often applied to the cervical tissue and nuclear to cytoplasmic ratio is calculated in order to differentiate dysplastic tissues.
This method was used in two studies Collier et al.
and a sensitivity of 100% with a specificity of 91- 100% was found.68, 69 While these optical techniques are an advance over traditional cervical screening, they are still subjective because images must be interpreted by a trained professional.
27 The process remains time consuming and subject to human error in interpretation.
These modalities also lack the depth penetration and depth resolution to identify tissue of interest, as dysplasia typically originates in the basal layer.40, 43 There exists a need for a quantitative and immediate method for CIN detection.
2.3.4 Angle-resolved low coherence interferometry Angle- resolved low coherence interferometry (a/LCI) is a light scattering technique that provides depth resolved measurements of nuclear morphology in vivo.42 a/LCI combines the capabilities of light scattering spectroscopy in measuring morphological changes of cell nuclei with the depth- resolved power of Fourier- domain LCI.
a/LCI uses nuclear diameter as a surrogate biomarker for dysplastic change in tissue.
The cell nucleus is treated as a Mie scatterer and variations in angular scattering indicate nuclear size.
The collected angular distribution of elastically scattered light from cell nuclei in localized layers of targeted tissue is compared to a Mie Theory database, and the size match is reported as the mean nuclear size for a tissue depth.
Using this method, a/LCI obtains depth- resolved nuclear morphology measurements with submicron accuracy, without the use of exogenous contrast agents.
a/LCI is a quantitative technique, which reduces variability and subjectivity, and can identify diseased tissues of interest.
a/LCI has been successful in identifying dysplasia in human and animal epithelial tissue models with high sensitivity and specificity.34- 40 More specifically, a/LCI 28 has been successful in detecting dysplasia in the human esophagus and colonic epithelium, indicating that CIN detection may also be possible.36- 38 Table 2.1 shows a breakdown of a/LCI studies on detecting dysplasia in human or animal epithelial models to date.
In a clinical evaluation, 46 subjects undergoing endoscopic monitoring of Barrett’s esophagus were scanned in vivo with a/LCI.
Depth resolved nuclear morphology measurements enabled dysplastic tissue to be identified with 100% sensitivity and 87% specificity, over 172 biopsy sites.38 The a/LCI technique successfully detected enlargement of nuclear size in the basal layer of esophageal tissues that were confirmed to be dysplastic via histopathological analysis of biopsy specimens obtained from the same sites.
The a/LCI approach has shown the ability to detected dysplasia in a wide range of epithelial tissues, such as a rat model of esophageal carcinogenesis with 100% sensitivity and 80% specificity (n = 42)40, a hamster model of tracheal carcinogenesis with sensitivity of 78% and specificity of 91% (n = 20)35, and most recently 29 an ex vivo study of human colon epithelium with sensitivity of 93% and specificity of 84% (n = 81)37.
These results indicate that a/LCI shows great promise as a clinical tool for diagnosis of dysplasia in a variety of epithelial tissues, and suggests it may be applicable to detecting cervical dysplasia.
Chapter2 presented the background information which is necessary to understand the technologies and applications of the imaging methods presented in this
(1) dual- modality optical imaging and (2) a/LCI.
The chapter first focused on microbicide gels, the target of the dual- modality optical imaging instrument measurements.
The importance of measuring in vivo microbicide gel coating thickness distributions was discussed, as well as, the effects of dilution on gel behavior.
Remote sensing imaging techniques were presented in Section 2.2.
3, while Section 2.2.
4 and 2.2.5 detail the early proof- of- concept studies of single channel LCI in measuring microbicide gel layers.
Section 2.3 discussed the background of the second imaging modality described in this dissertation, a/LCI.
CIN, the target disease of the a/LCI study, and several co- risk factors such as HPV and HIV were discussed.
The current cervical cancer screening method was then examined in Section 2.3.2, and several optical detection techniques that have been applied to cervical dysplasia, including wide- field imaging, spectroscopy, and confocal high- resolution imaging were detailed.
The chapter concluded with an 30 introduction to a/LCI and a breakdown of relevant a/LCI studies that have been performed to date.
31 3 Instrumentation
Chapter2 provided background information on microbicide gels and techniques for measuring their thickness distributions.
Background information of low coherence interferometry (LCI) was provided in Section 2.2.
4, and two initial LCI imaging setups were described.
The first common- path system showed the ability of LCI in measuring microbicide gels and this system was expanded to full Michelson geometry which allowed simpler data interpretation of complex samples.
Here, the evolution from a single channel LCI device to a fully integrated, dual- modality optical imaging instrument is presented.
The first benchtop mLCI device, detailed in Section 3.2.1, was initially built to increase tissue coverage during imaging, and validation experiments were performed with this device.
To refine this device for application to measuring microbicide gel thickness distributions, mLCI was combined with an existing fluorimetric device used by Henderson et al.
for simultaneous, dual- modality optical imaging.27, 28 Calibration and optical performance data are presented in Section 3.2.2, as the system was completely characterized.
Initial in vivo human microbicide gel thickness measurements are presented in Section 3.2.3, while Section 3.2.4 concludes with recent improvements to the clinical dual- modality optical imaging instrument that were made to increase data fidelity and ease of operation of the instrument in the clinical setting.
3.2 Dual-modality optical imaging instrument 3.2.1 Benchtop multiplexed low coherence interferometry (mLCI) instrument To demonstrate the utility of multiplexed LCI for gel thickness measurements, a benchtop mLCI system was constructed.