An Imaging-Based Approach to the Evaluation of Xerostomia

Background and Objective— Goal was to evaluate the potential of in vivo optical coherence tomography (OCT) imaging to determine the response of patients with xerostomia to a dry mouth toothpaste versus fluoride tooth-paste placebo. Study Design/Materials and Methods— Ten subjects with xerostomia participated in this double-blind, crossover, placebo-controlled study. After examination and OCT imaging, subjects used the first product for 15 days, followed by a 7-day washout period, and then they used the second product for 15 days. Data were acquired at 5-day intervals, also before and after the washout. Results— Visual examination and tongue blade adhesion test did not reflect response to the product. Two imaging-based markers were identified: (i) In OCT images, epithelial thickness increased significantly ( P < 0.05) after use of the dry mouth toothpaste, but did not change significantly ( P > 0.05) after the use of a fluoride toothpaste and (2) Optical backscattering data showed progressive characteristic changes from baseline with use of the active product. Conclusions— In this pilot study using in vivo OCT imaging, it was possible to detect and measure oral epithelial response to the dry mouth product versus placebo in patients with xerostomia. Clinical Implications— This approach may permit site-specific assessment of xerostomia, individualized treatment planning and monitoring, and sequential mucosal mapping in patients with dry mouth.


INTRODUCTION
Dry mouth is a common oral condition, with an estimated prevalence in 20-40% of the population [1,2]. Higher levels are documented in the elderly and in women [3]. Causes include the following: • A common side effect of many prescription and non-prescription drugs, including drugs used to treat depression, anxiety, pain, allergies, colds, obesity, acne, epilepsy, hypertension, diarrhea, nausea, psychotic disorders, urinary incontinence, asthma, and Parkinson's disease. Muscle relaxants and sedatives, as well as tobacco use, are also implicated. The list of drugs includes >400 agents [1].
• Salivary gland damage, for example, from radiation to the head and neck, or from chemotherapy for cancer [1].
• A result of nerve damage to the head and neck area from an injury or surgery [1].
Saliva serves important functions in the mouth including remineralization of teeth [4,5], lubricating the oral mucosa of the mouth, facilitating speech, eating, swallowing, and preventing mechanical injury to the surfaces of the mouth. Saliva and salivary flow also help prevent the accumulation of microorganisms in the mouth [4,5] including cavity-causing bacteria. Salivary flow and constituents, such as amylase, initiate digestion of foods and help dissolve and remove food particles from the mouth and salivary components have antibacterial, antifungal, and antiviral properties [6]. Current standard of care for evaluating dry mouth includes mucosal appearance, tongue blade adhesion test, saliva collection over a fixed time period and patient self-evaluation using a standard questionnaire [1,2,7]. These approaches, although generally accepted and frequently used have significant limitations. They are subjective in nature and often require a large patient sample size in order to maximize chances of obtaining useful results [1,2]. Additionally, current methodologies are not sensitive to small, subtle changes that may take place within the oral mucosa at the microstructural level, as a result of xerostomia and/or during the course of its treatment.
The inability to assess xerostomia on an individual basis, in a quantifiable and nonsubjective manner, to individualize treatment planning or to evaluate the effectiveness of dry mouth interventions using existing conventional approaches is a significant clinical problem and a limiting factor in the validation of outcome measures in interventional clinical studies. Although some useful information may be obtained using current evaluation approaches with a large enough patient sample size, methodologies that enable the assessment and quantification of small changes within the microstructure of the oral mucosa are desired, in order to improve upon the limitations of current approaches by allowing accurate and reproducible assessment of xerostomia and treatment effect on an individual basis.
Optical coherence tomography (OCT) is an emergent imaging modality that generates highresolution microstructural images, which can also be analyzed for site-specific backscattering data based on coherence-domain optical technology. OCT has most often been compared to ultrasound imaging. Both technologies employ backscattered signals reflected from different layers within the tissue to reconstruct structural images, with the latter measuring sound rather than light. The resulting OCT image is a two-dimensional representation of the optical reflection within a tissue sample. Cross-sectional images of tissues are constructed in real time, at near histologic resolution (approximately 5-15 μm with current technology). These images can be stacked to generate 3D reconstructions of the target tissue, such as epithelial and subepithelial structures.
Several OCT systems have received FDA approval for clinical use, and OCT is deemed by many as an essential imaging modality in ophthalmology. In vivo image acquisition is facilitated through the use of a flexible fiber-optic OCT probe, which is placed on the surface of the tissue to generate real-time, immediate surface and sub-surface images of tissue microanatomy.
Goal of this pilot study was to (i) evaluate the potential of in vivo OCT imaging and (ii) identify potential markers to determine the response of patients with clinically diagnosed moderate to severe dry mouth to the use of a dry mouth toothpaste versus a placebo.

Protocol
This research was executed in full compliance with UCI IRB approval #2002-2805. Informed consent was obtained from all subjects prior to their enrollment in the study. Ten subjects previously diagnosed with moderate to severe dry mouth (7 female, 3 male; age from 37 to 53 with a mean age of 48) were enrolled in this randomized, double-blind, crossover, placebo-controlled study. Two products that were packaged identically and coded for content were tested. One product was a dry mouth tooth-paste and the other was a placebo (regular fluoride tooth-paste). After a baseline clinical exam consisting of photographs and OCT imaging, subjects used their first assigned dentifrice to brush their teeth for 15 days. A 7-day wash out period using Colgate Cavity Protection toothpaste was followed by a second 15-day period using the second assigned dentifrice. The sequence of product use was randomized. Standard toothbrushes and floss were also provided to the subjects; no other form of oral healthcare (rinses, gels, gum, etc.) was permitted. Subjects were required to inform the investigators about any potential new medications that might interfere with the study, including but not limited to antibiotics, antiseptics, decongestants, and antihistamines. OCT imaging and a clinical examination, a dry mouth questionnaire and photographs were acquired at 5-day intervals during use of the investigational dentifrices, and at the beginning and end of the 7-day washout period.

Clinical Examination
Adhesion of a standard wooden tongue blade to the left and right buccal mucosa, the dorsal and ventral surfaces of the tongue and the lip was determined by one blinded clinician (PWS) at each time point. Adhesion was scored as "yes" (score 1) or "no" (score 0) for each site; the lower the cumulative score, the more moist was the mouth, and the higher the score, the drier was the mouth. Scores were added up for each time point to generate an overall semi-quantitative tongue blade adhesion score on each evaluation day for each subject.

Self-Evaluation Questionnaire
At each time point of the study, subjects' responses to a standard dry mouth questionnaire [7] were semi-quantified on a scale of 0-5, with 5 representing the most severe perception of the specific symptom, and 0 representing the absence of each specific symptom. Questionnaire scores were added up for each time point to generate an overall semiquantitative dry mouth score for each evaluation day; the higher the score, the worse were the perceived symptoms of dry mouth.

OCT Imaging
All subjects were imaged using the same commercially available Niris ® OCT console and imaging probe ( Fig. 1; Imalux Corporation, Cleveland, OH), which allows real-time video rate imaging speed, simultaneous OCT and CCD imaging channels, 3D volumetric imaging and surface profiling capability at an imaging depth of up to 50 μm.

Imaging Data
From each image, at four standardized locations per image, epithelial thickness was measured, by superimposing a grid on the OCT image and measuring at 1 mm intervals across the width of the scan. A backscattering signal was determined directly from the optical scan data by using proprietary software from the Imalux ® system to quantify the optical signal at 1 pixel increments on individual images. Optical measurements were made at the same locations as the epithelial thickness determinations. All measurements were performed by two blinded scorers pre-standardized to 95% accuracy from our OCT image databank. This process was repeated after 1 month to evaluate intra-scorer and inter-scorer variability.

Analysis
The predefined primary efficacy variable was the change in thickness of the epithelium at four specific landmark locations for each imaging site, at each time point. A secondary variable for analysis was the optical reflectance changes at the epithelial surface and at depth intervals of 1 pixel down to 2 mm, measured at the four standardized locations. For microstructural measurements, data for each patient were averaged across all scan sites for each time point. For the optical data, ratios between specific points on the line graphs of intensity versus depth were used (Fig. 2). Data were compared using repeated measures ANOVA with one within-group factor (before vs. after) and one between-group factor (treatment). After adjusting for any between-subject differences, the significance of the interaction factor for time by treatment (F-test with 1 df) provides a test for differences over time between the active and placebo groups.

Clinical Evaluation
The tongue blade adhesion test did not detect changes in oral dryness in a meaningful way. Improvements in dryness were detected in 4/10 subjects after use of the dry mouth toothpaste and in 5/10 subjects after use of the placebo and/or washout. Thus, upon evaluation of a sample size of ten patients, measurements were not sufficiently consistent to be useful in evaluating a specific dry mouth intervention.

Self-Evaluation Questionnaire
Dry mouth questionnaire responses among the 10 patients were not consistent and therefore did not provide a meaningful evaluation of patient response to dry mouth treatments. In one subject, based on the dry mouth questionnaire, the active product and the washout toothpaste were deemed to be ineffective, whereas the placebo product was perceived to progressively worsen the symptoms of dry mouth, especially after >10 days of use. No trends could be identified based on subject questionnaire data in six subjects. Throughout the course of the study, a feeling of progressive improvement was expressed by two subjects, whereas a perceived progressive worsening was expressed in another subject. In these three subjects, this observation was unrelated to the product used, be it active, placebo, or washout product.

OCT Imaging Data
The OCT imaging required <1 second per scan. Using a simple fiberoptic probe that resembles a pen, images were produced almost immediately and the procedure was very well tolerated by patients.
A. Based on the OCT images, average, epithelial thickness increased significantly (P < 0.05) after use of the dry mouth toothpaste for 15 days (Table 1), but did not change significantly (P > 0.05) after use of a regular fluoride toothpaste placebo or washout product. In the OCT images, epithelial and subepithelial tissues subjectively appeared thicker, denser, and plumper after the use of the dry mouth toothpaste than after regular fluoride toothpaste used as washout and placebo (Fig.  2).

B.
In the baseline log plots of backscattering intensity (y-axis) versus tissue depth (xaxis) images, there was at first an intense backscattered signal at the air-epithelial interface, with the rest of the signal becoming progressively less intense with increasing depth into the sample (Fig. 3). With use of the dry mouth product, optical scattering data showed progressive and characteristic changes from baseline. The signal ratio between epithelial surface and overlying air (Ratio B: A) decreased significantly (P < 0.05) after 5 days use of the active product, with another significant decrease (P < 0.05) after 15 days use. A similar, but somewhat weaker trend was observed for the ratio between epithelial surface and subsurface epithelium (Ratio B:C). Treatment-related changes in backscatter were confined to the most superficial 750-1,000 μm of the oral mucosa. Figure 4 is a compilation of representative data for subject 1 showing a measureable increase in epithelial thickness and a considerable reduction in the B/A ratio in response to the active dry mouth toothpaste. However, no discernible responses to the toothpastes are apparent using self-scoring and tongue blade test.

DISCUSSION
Current techniques are limited in their ability to diagnose and quantitatively measure dry mouth with the necessary precision and reproducibility. Furthermore, these current approaches cannot quantify, map and track specific mucosal changes during the course of xerostomia or its treatment [1,2]. Results from this study were in agreement with the literature, determining that standard dry mouth questionnaires and clinical examinations are not well suited to quantify oral response to dry mouth treatments on an individual basis.
Using in vivo OCT imaging, we were able to establish two markers for response to dry mouth intervention: mucosal microarchitecture and the depth-resolved backscattering signal intensity from the mucosa. The progression of specific microstructural events occurring at specific sites to a depth of 2 mm within the oral mucosa was successfully quantified and mapped in these subjects. OCT images showed reversal of xerostomia-induced mucosal thinning including a more consistent surface keratinized layer of the epithelium, a thicker epithelial layer and a denser subepithelial layer resulting from treatment using dry mouth toothpaste (the active). Additionally, characteristic, quantifiable, simple optical measures of mucosal rehydration were identified and validated using backscattering data extracted from OCT images. One possible explanation for the change in optical backscatter seen as a result of dry mouth treatment may be that the more hydrated tissues absorb more of the OCT light at 1,310 nm [35], resulting in a smaller signal that is returned to the detection fiber of the OCT system.
Because this was a very limited feasibility study, we did not analyze the data for genderbased differences, nor did we evaluate the hormonal status of the four younger women enrolled in this study. It is known that hormonal status can influence epithelial thickness, including oral epithelia and this will be an interesting concept to pursue in future, larger studies.
Recent studies have indicated that complex microstructural and vascular components of xerostomia-related mucosal damage exists [36][37][38]. However, there is little specific in vivo information regarding pathways of mucosal damage in xerostomia patients. The approach outlined in these in vivo studies for the first time provides a means of tracking these events in patients to gain a better understanding of the process, which can lead to the development of superior, more specific interventions for specific categories of xerostomia. Important applications for clinical research include cancer-therapy-induced xerostomia, medicationinduced dry mouth, as well as a diverse range of medical conditions (such as diabetes, Sjögren's syndrome, HIV/AIDS, rheumatoid arthritis, hypertension, Parkinson's disease, stroke, and mumps).

CONCLUSIONS
In this pilot study, the authors identified markers and established the basis for developing and validating a novel, non-invasive, in vivo imaging approach to clinically (i) evaluating dry mouth, (ii) quantifying the effectiveness of specific interventions for this condition, and (iii) mapping at high resolution the surface and subsurface mucosal effects of dry mouth and its mitigation. Using in vivo OCT imaging, the authors were able to detect and measure oral epithelial response to the use of a dry mouth tooth-paste in patients diagnosed with moderate to severe xerostomia.

CLINICAL IMPLICATIONS
Current techniques are not able to quantify adequately the effects of dry mouth on the oral mucosa during the course of xerostomia and its treatment. This pilot study demonstrated a novel in vivo imaging approach, which will potentially permit: • Accurate site-specific assessment of dry mouth.
• Individualized treatment planning and monitoring.
• In vivo evaluation of the effects of dry mouth and of interventional effectiveness in individuals and specific populations using two sets of indicators.
• Optical indices of hydration.
• Microstructural analysis of the oral mucosa.
• A sequential site-specific high-resolution map of mucosal and events as components of mucosal damage and potential therapy-induced mitigation in patients with xerostomia.
Thus, these preliminary studies provide the basis for a clinical imaging approach that may potentially overcome (i) our current inability to evaluate and quantify the effects of dry mouth accurately and in a non-subjective manner, (ii) our very limited ability to validate outcome measures in interventional clinical studies of xerostomia, and (iii) an important barrier to clinical research. Representative OCT images of right buccal mucosa in Subject 3 throughout the study showing improved epithelial surface integrity, and greater epithelial and subepithelial thickness, density and plumpness after use of the active dry mouth product. The placebo product and regular Colgate toothpaste had no measurable effect. [Color figure can be seen in the online version of this article, available at http://wileyonlinelibrary.com/journal/lsm] Representative log plots of backscattering intensity (y-axis) versus tissue depth (x-axis) in Subject 7. Plots after using placebo did not differ in essence form baseline. However, after using active dry mouth product the plot differed markedly from baseline. Treatment-related changes in backscatter were confined to the most superficial 750-1,000 μm of the oral mucosa. [Color figure can be seen in the online version of this article, available at http:// wileyonlinelibrary.com/journal/lsm] Representative collated data from Subject 1. No discernible responses to the tooth-pastes are apparent using self-scoring and tongue blade test. However, a measureable increase in epithelial thickness and a considerable reduction in the B/A ratio were determined in response to the active dry mouth toothpaste. [Color figure can be seen in the online version of this article, available at http://wileyonlinelibrary.com/journal/lsm]