Otitis media (OM), an inflammatory disease of the middle ear, is one of the most common diagnoses for pediatric patients and often follows upper respiratory tract infections (Danishyar and Ashurst, 2017). It is one of the leading causes of outpatient antibiotic use for children (Palms et al., 2019), resulting in more than 25 million physician visits and $4.3 billion in public healthcare costs, annually, in the United States alone (Tong et al., 2018). Approximately 80-90% of children present with some form of OM by the age of three. Thus, an appropriate diagnosis and treatment regimen is crucial to minimize the risk of delayed speech development and socialization skills that may result from the long-term reduction of auditory acuity during chronic states of the disease. OM involves the accumulation of a middle ear effusion (MEE) caused by Eustachian tube dysfunction. In the absence of acute infection, this condition is termed OM with effusion (OME). However, planktonic bacteria from the upper respiratory tract can migrate into the middle ear through the Eustachian tube and colonize within the nutrient-rich MEE. This migration may lead to the onset of an acute OM (AOM) infection by one or more pathogens that presents with physical symptoms (e.g., otalgia, otorrhea, fever, irritability, restlessness, and poor feeding) and tympanic membrane (TM) bulging or erythema in children. Cases of OME can resolve spontaneously within months, and many AOM cases typically resolve within 1-2 weeks with appropriate administration of antibiotics. However, identifying and appropriately treating a single persistent/chronic or multiple recurrent infections is still an ongoing challenge as most of these cases do not respond to antibiotic therapy and surgical intervention is advised (Hoberman et al., 2011).

Evidence has pointed towards the importance of biofilm formation in the occurrence and persistence of chronic OM. Bacterial biofilms are self-assembling microenvironments that provide a protective habitat for microorganisms to colonize and grow. The biopolymeric matrix adheres to tissue structures and protects the pathogens from external agents like host immune cells or antimicrobial drugs. Researchers have reported findings up to 92% of middle ear mucosal biopsies sampled from patients with chronic OM showed the presence of biofilms (Kaya et al., 2013; Gu et al., 2014). OM-related biofilms usually contain one or multiple species of the primary otopathogenic bacteria, including: Streptococcus pneumoniae, non-typeable Haemophilus influenzae (NTHI), Moraxella catarrhalis, Pseudomonas aeruginosa, and Staphylococcus aureus (Fergie et al., 2004; Thornton et al., 2011). Although targeted therapies like species-specific antibiotic agents can be more effective than broad-spectrum antibiotics when pathogens are in free-floating (i.e., planktonic) form, biofilms exhibit unique antibiotic resistance mechanisms that affect drug treatment efficacy (Sharma et al., 2019). Therefore, detection and characterization of infection in a planktonic or biofilm state are critical diagnostic criteria needed to monitor the progression of chronic forms of OM. Consideration of these aspects of OM pathology would not only improve procedures for prescribing antibiotic therapies, but allow for tracking of drug treatment response in chronic OM.

Another challenge in the administration of treatment is identifying the cause of OM. Although a large number of AOM cases are associated with bacterial pathogens, in approximately 5-30% of AOM cases in children, the causative pathogen is viral or a combination of both (Shoichi et al., 2019). With viral infections, the administration of antibiotics is unnecessary, or in combined viral/bacterial infections, monitoring of antibiotic response over a longer period of time may be needed. This is because the bacterial infection could have resolved, but symptoms relating to a viral presence persisted. There are also instances where the presence of an effusion in the middle ear is serous (i.e., OME) without signs of infection (Vanneste and Page, 2019). In these cases, antibiotics treatments are not recommended under current guidelines, and “watchful waiting” is recommended (Lambert, 2016). Therefore, an OM diagnostic tool capable of real-time identification of bacterial presence and pathogen speciation is needed to help guide clinical decision making for the most appropriate treatment regimen.

Visual or digital otoscopy is the current clinical standard for the diagnosis of OM, but can be subjective amongst practitioners. Alternative methods like pneumatic otoscopy or tympanometry examine the motility of the TM in response to pressure oscillations in the ear canal to determine the presence of MEE, but lack the specificity of pathogen identification and the ability to identify or track the evolution of the effusion or biofilm over the course of treatment and progression/regression of the disease. Tympanocentesis, a minor surgical procedure to aspirate a MEE from the middle ear cavity with a small gauge needle, remains the gold standard for typing of OM-causing bacterial species, but is only used in complex cases that do not respond to first-line antibiotics like amoxicillin or penicillin. However, in some cases (~40%) the extracted middle ear fluid is culture-negative, and more time consuming and complex tools like polymerase chain reaction (PCR) and DNA sequencing are needed to determine speciation (Xu et al., 2011). These limitations imposed by current diagnostic methods have resulted in treatment guidelines for OM recommending the use of broad-spectrum antibiotics that often adds to the risk of antibiotic resistance (Center for Disease Control and Prevention, 2019). This risk includes increases in the incidence of adverse effects such as disruption of the gut microbiome (Francino, 2016), upset stomach, or diarrhea (Gerber et al., 2017), which can lead to other poor health outcomes and diseases (Holmes et al., 2011; Chen et al., 2021). Thus, advances in more specific and less labor-intensive diagnostic tools are essential to improve clinical treatment decisions and patient outcomes.

Building off of existing visual and digital otoscopy platforms, optical methods have garnered interest in OM diagnostics as they are inherently non-invasive, fast, can be readily applied in vivo, and offer a wide range of compositional or structural information about microbes and the microenvironment they inhabit. Optical coherence tomography (OCT) is an interferometric imaging modality that provides micron-scale resolution of tissue morphology. This optical modality has been investigated for application in OM to detect the presence of middle ear effusions and biofilms (Nguyen et al., 2012; Monroy et al., 2015; Won et al., 2021b), correlate TM thickness to a chronic state of the disease (Dsouza et al., 2018), and study the longitudinal effects of antibiotic treatment on biofilm growth and MEEs (Won et al., 2021a). To date, OCT is the only technique that can provide direct visualization of biofilms adherent on the TM in vivo. Raman spectroscopy (RS) is a vibrational spectroscopic modality that excels in its ability to probe molecular composition through the inelastic scattering of light. It has previously been reported for use in OM diagnostics by identifying causative bacterial species, grown on agar in vitro, with 97% classification accuracy (Ayala et al., 2017). Also, RS has been used to characterize compositional differences in non-infected MEEs in vitro, where thicker mucoid MEE could be discriminated from thinner serous effusions with 92% accuracy (Pandey et al., 2018). This is significant because mucoid MEEs are less easily cleared by middle ear ciliary action and are associated with more chronic episodes of OM (Matković et al., 2007). In a recent study, a combined Raman coupled low-coherence interferometry (LCI) probe was developed for the real-time detection and differentiation of microbial pathogens (i.e., Pseudomonas aeruginosa and Streptococcus pneumoniae) in the middle ear (Zhao et al., 2016). Therein, RS was employed to identify the planktonic bacterial species, and LCI was utilized to determine the depth-resolved structural information from the TM and to visually guide the placement of the RS probing beam. Thus, this work demonstrated the advantages of combining spectroscopy with imaging to provide a more informed description of the nature of bacterial infection within the middle ear.

The two modalities, OCT and RS, offer very different but complementary information about effusion scattering properties and biofilm microstructure, and cellular biochemical composition, respectively. While OCT can identify MEEs and biofilm formation, track film thickness, and investigate optical properties like refractive index, attenuation, or other scattering properties (Tearney et al., 1995; Balaev et al., 2003; Chang and Bowden, 2019), it lacks chemical specificity which could be used to determine which species are responsible for any persistent disease like chronic OM. On the other hand, RS provides a comprehensive biomolecular identification of the bacterial, biofilm, and MEE, but cannot provide information about the structural organization of functional characteristics of tissue when applied in vivo. Thus, there is a growing interest in exploring a combined multimodal optical system that utilizes the advantages of both techniques. In the context of acute and chronic OM, a multimodal RS-OCT platform would be able to detect MEE and biofilm formation and identify the causative bacterial pathogen(s) to better inform the clinical guidance of appropriate treatment, and allow for longitudinal observation of a patient’s response to the antibiotic therapy.

The goal of this study is to characterize OM-causative bacterial species, grown in different in vitro environments, through the independent techniques of RS and OCT. This will facilitate an understanding of the key biochemical and structural differences between bacteria found on agar, in a fluidic environment (i.e., planktonic), and within established biofilms. In particular to OM, this study aims to identify discriminatory markers that can be used to distinguish the four most common causative bacterial species: Haemophilus influenzae, Streptococcus pneumoniae, Moraxella catarrhalis, and Pseudomonas aeruginosa.

To characterize differences between OM-causing bacterial species when grown as colonies on brain heart infusion (BHI) agar, in a fluidic environment, or a biofilm matrix, RS and OCT measurements were acquired from four of the most common otopathogenic OM bacterial species: H. influenzae, S. pneumoniae, M. catarrhalis, and P. aeruginosa under all three growth conditions.

An OM diagnostic tool that can provide additional information about the bacterial otopathogen, including detecting and characterizing the bacterial state (i.e., effusions (planktonic) versus biofilms), would help better inform treatment; including the administration of antibiotics. Herein, the detection capabilities of both RS and OCT were explored independently but under similar experimental conditions to highlight the potential for these two modalities to serve as a future dual-modality tool capable of achieving this diagnostic goal. Both light-based technologies offer different but complementary mechanisms to characterize differences in growth states of OM-causing bacteria and, in most cases, offer a means for differentiation across bacterial species.

First, the differences in the biochemical composition of four OM-causative bacterial species studied herein can be seen using RS. Furthermore, with above 93% sensitivity and specificity, RS can detect feature-specific chemical and structural changes that occur during the growth of the OM-causing bacteria from colonies to planktonic form and finally, biofilms. In particular, P. aeruginosa is known to secrete toxins such as pyocyanin as a defense mechanism to safeguard itself within biofilms and promote extracellular DNA (eDNA) release (Das and Manefield, 2012). The Raman signatures of pyocyanin have been well studied by other researchers who noted strong bands at ~1354 and 1610 cm^-1 (Bodelón et al., 2016; Turasan, 2020). Biofilms of P. aeruginosa also display this prominent peak at 1355 cm^-1, as seen in Figure 1D, indicating the secretion of this toxin within three days of the biofilm growth. Also seen, across the three growth forms, is a shift in the Amide I (C=O stretch) Raman band between 1640 – 1680 cm^-1 (Maquelin et al., 2000; AlMasoud et al., 2021). In particular, the colonies and planktonic form of P. aeruginosa Amide I band peak reside at ~1670 cm^-1. However, for the biofilms, this peak is shifted to ~1659 cm^-1. The ~1610 cm^-1 Raman peak of pyocyanin could be contributing to this peak shift associated with the P. aeruginosa biofilm formation. Moreover, the chemical secretion, and its corresponding Raman peak at 1355 cm^-1, provides a means to distinguish P. aeruginosa biofilms from its colony and planktonic form. This key Raman peak also allows for differentiation from the other OM-causing bacterial species studied herein, as seen in Figure 2, with 100% specificity and sensitivity.

For S. pneumoniae, the key differences between the biofilm and planktonic forms are seen in Figure 1B as changes in the protein/lipid peak ratio at 1095 cm^-1 (C-N vibrational bonds) and the skeletal acyl backbone of lipid Raman peak at ~1124 cm^-1 (C-C: transconformation) (He et al., 2014). The planktonic form exhibits a stronger Raman intensity at 1095 cm^-1 compared to the biofilms. Also, a conformational change of the protein/lipid is observed, where the colonies exist in a -’cis’ configuration (1660 cm^-1) while the planktonic and biofilms exist in the ‘trans’ (1670 cm^-1) configuration (Talari et al., 2015). These changes suggest that one potential mechanism in the formation of S. pneumoniae biofilms is associated with membrane lipid structural changes as the bacteria adapts to biofilm growth. Furthermore, the relative redistribution in the intensity of the nucleic acid content (~740–832 cm^-1) of the S. pneumoniae biofilms provides a discriminatory factor for this bacterium from the other three OM-causing bacteria. This redistribution is seen in Figure 2 as a higher relative Raman intensity between 800-830 cm^-1, and potentially relates to higher levels of extracellular DNA (eDNA) release compared to the other bacteria. The release of eDNA is important for the initial adhesion and aggregation of bacterial cells toward biofilm formation. This biomolecule is reported to be released spontaneously during the formation of S. pneumoniae biofilms (Moscoso et al., 2006). Also, S. pneumoniae is a gram-positive bacterium and the only one of the four studied herein. Therefore, differences are seen in the Raman bands of its agar colonies between 710- 730 cm^-1 could relate to its cell wall components (Figure 4). Unlike the other species whose nearest Raman peak resides at ~725 cm^-1 and relates to DNA content (Guicheteau et al., 2006), the Raman spectra of this bacterium contain a distinct Raman peak at 716 cm^-1 suggesting the presence of phospholipid-containing choline biochemical signatures (Krafft et al., 2005). Researchers have reported that this bacterium cell wall includes lipoteichoic acid-choline moieties (Meiers et al., 2014). Therefore, the distinct peak at 716 cm^-1, suggests the Raman spectral feature of the choline contained within its cell walls, which is not present in the cell walls of the other bacteria. In addition to the 716 cm^-1 spectral feature, other differences in the phospholipid signatures can be seen in the RS regions of 1094-1126 cm^-1 and 1294-1304 cm^-1 in Figure 4. These differences are potentially related to the bio-composition of the outer membrane of gram-positive versus gram-negative bacteria. Altogether, these three spectral differences allow for the differentiation of S. pneumoniae colonies, grown on BHI agar, from the other three bacterial species.

Similar to S. pneumoniae, M. catarrhalis biofilms also demonstrates lipid structural changes as shown in the reduction in relative Raman intensity at 1098 cm^-1, but maintains relatively similar Raman intensities at 1124 cm^-1 compared to its planktonic form (Figure 1C). Also, in differentiating the agar colonies from the other environments, the Amide I Raman peak resides at 1660 cm^-1 for the colonies compared to 1670 cm^-1 for the planktonic form and biofilms. However, the main distinguishing feature of this bacterium to differentiate the planktonic form and agar colonies from its biofilms relates to the Raman peak at ~1554 cm^-1. This peak potentially relates to elevated levels of tryptophan release during DNase production (Hussein, 2016) and can be used to differentiate M. catarrhalis biofilms from the other OM-causing bacterial species (Figure 2).

Raman spectroscopy is also able to differentiate all three growth forms of H. influenzae. The differences can be seen in Figure 1A in the Raman regions between 1643-1650 cm^-1, 1342-1373 cm^-1, and ~666 cm^-1. The biofilms of H. influenzae differ from its planktonic form in the Raman spectral regions of 666 cm^-1 and 1342-1373 cm^-1 relating to the redistribution of amino acid and proteins in the 1500-1700 cm^-1 region (Guicheteau et al., 2006). One potential cause of these redistributions may be the regulation of proteins in the formation of biofilms. A proteomic study conducted by Post et al. reported that there is a downregulation of proteins associated with energy metabolism and protein synthesis, but those involved in fatty acids, DNA metabolism, and transcription are upregulated when H. influenzae exists in biofilm form (Post et al., 2014). These protein biochemical changes can be detected via RS and allow for the differentiation of H. influenzae from the other bacterial species (Figure 2).

Altogether, RS can characterize the biochemical differences that occur between growth forms of the same species as well as the differences in bio-compositions across species. Across species, RS can differentiate the four causative species from each with other with specificities greater than 96% in all three growth forms. This is important in the case of OM detection where the different causative species and growth states (i.e., planktonic form within effusion versus biofilms) may cause varying response to antibiotics. Thus, knowing more information about the disease can better inform treatment. However, the biochemical “fingerprint” is only one aspect of bacterial characterization. Its morphological structure, as it inhabits an environment, can also be used for characterization and differentiation. In particular, biofilms are known to be inherently heterogeneous (Hall-Stoodley et al., 2008; Hall-Stoodley and Stoodley, 2009). Therefore, OCT, an imaging modality, has the added benefit of visualizing and spatially assessing the morphological changes of the bacteria in its colony, planktonic form, and biofilm forms, with fast acquisition times, to better inform species and growth identification.

OCT is a non-invasive and non-destructive label-free technique that requires no sample preparation. Herein, it was utilized to assess whether morphological differences can be detected via the RIs and/or ACs for each of the four OM-causing bacterial species. As the OM-causing bacterial species are grown from colonies to biofilms, there are observed differences in the light scattering optical properties based on varying morphological differences. It is expected, that samples with thicker layers or membranes, denser or clustered cells, or ones which contains higher amounts of biochemical secretions will result in larger light scattering events. In general, single colony bacteria are highly dense clusters of cells, and factors such as how tightly packed these colonies are as well as what biochemical they secrete can impact its optical properties. Planktonic cells are free-floating cells in a liquid matrix, and are responsible for initiating biofilm formation in response to a nutrient-rich environment by undergoing various metabolic response pathways, also known as quorum sensing (O’Toole et al., 2000) which can impact its optical properties based on the secretory biochemicals and the density of cells. Whereas, biofilms are dynamic but represent a stable point in the bacterial lifecycle and contain distributed structures with water pores and channels; and thus, contributing to a much lower density (Wang et al., 2011) across the extracellular polymeric substance (EPS) matrix. Therefore, biofilms are expected to produce less scattering than the denser planktonic form or colonies (Hou et al., 2019); due to the presence of a mixed environment of bacterial aggregates and an EPS matrix with high water content (also known as a highly hydrated gel) (Flemming and Wingender, 2010) in the biofilm. Within this study, these structural differences were measured via RI and AC, where RI and AC provide information about how light is transmitted and scattered, respectively, through a sample. Herein, a decrease in RI of the bacterial colonies compared to its biofilms’ counterparts is observed. Moreover, the measured RI for each species’ biofilms ranged from ~1.33-1.34 which is similar to the RI of water (~1.33), which further confirms that the biofilm matrix contains high water content. Also, the measured RIs of the biofilms in this study, particularly for P. aeruginosa, is comparable to previous work shown by Bakke et al. who reported a mean RI of ~1.35 for P. aeruginosa biofilms (Bakke et al., 2001). The measured RIs for the colonies also correspond well with the reference RI values of bacterial colonies in a prior study (Balaev et al., 2003). However, the presence of water in the planktonic and biofilm stages minimized the capabilities of RI to differentiate across all growth environments and species. Thus, this study show that the optical scattering differences quantified by the AC measurements can more easily distinguish across growth environments and species in comparison to the RI measurements as shown in Figure 6B.

In general, the biofilms exhibit lower AC values compared to the planktonic or colony forms, due to the lower cell density as can be seen in Figure 5. P. aeruginosa displays the lowest AC in both colony and planktonic forms and H. influenzae has highest AC values across all species in both forms. As seen in the Figure 5, compared to all other species’ colonies, H. influenzae produce smaller and more compact colonies which result in more light scattering events than P. aeruginosa bacterial colonies. On the other hand, the 3-days old H. influenzae biofilms show the smallest AC across all species, with the 3-days old P. aeruginosa biofilms generating the highest AC. All biofilms were grown in the nutrient-rich BHI medium that contained proteose peptone and infusions from the calf brain and beef heart which provided carbon, nitrogen, essential growth factors, amino acids, and vitamins for biofilm formation. P. aeruginosa is reported to be one of the greatest biofilm producers (Pericolini et al., 2018) and produces more secretions of a highly diffusible nitrogen-containing compound, known as pyocyanin toxin metabolites (Das et al., 2016), at an early stage (~ within 14 h) of biofilm formation. It has also been reported that pyocyanin promotes eDNA release (Das and Manefield, 2012) in P. aeruginosa biofilms by inducing cell-lysis to bacteria cells mediated by the production of hydrogen peroxide (Allesen-Holm et al., 2006), which enhances the P. aeruginosa biofilm formation. Therefore, the increase in the mean AC for the P. aeruginosa biofilm suggests the presence of these secreted molecules, also confirmed by the RS data, as well as the formation of dense gels of the EPS biofilm matrix, which resulted in more scattering events. Another possible explanation for the dense EPS matrix of P. aeruginosa biofilm can be the increase of cellular cyclic-di-GMP (Guttenplan and Kearns, 2013) levels. This increase causes the reduction of the flagella-assisted swarming motility (Shrout et al., 2006; Guttenplan and Kearns, 2013) during the early stages of biofilm formation by transitioning motile P. aeruginosa to non-motile bacteria in the biofilm state.

The 2D and 3D images of biofilms from each species also provide more information about the change in signal intensities in relation to the spatial distribution of the matrix. The 3D OCT images are constructed by combining multiple cross-sectional 2D OCT images; thus, these images provide better visualization of the complex internal heterogeneous structures such as bacterial cell aggregates, pores, and water channels inside the biofilm, as shown in Supplementary Figure 2. In particular, the obtained images show that the biofilms contain mostly water-filled pores. The pores are noted as relatively dark (low-scattering) regions in the OCT images. It has been theorized, by other research groups, that these channels and pores supply localized nutrients to the growing biofilm (Flemming and Wingender, 2010). Another advantage of OCT imaging is that the AC coefficient maps can characterize the rate of change of OCT signals as a function of depth, which is important given that biofilms have a heterogeneous structure and often grow to thicknesses of hundreds of microns or more. These thicknesses are well beyond the imaging depth of fluorescence and confocal microscopes. The 2D images also show differences in the heterogeneity of the biofilms across species. In particular, S. pneumoniae biofilms show less heterogeneity than the other species, while both M. catarrhalis and H. influenzae show layered biofilms in the bottom layer. On the other hand, the top surface of P. aeruginosa is very distinct from the other species and show pores with water channels and morphological differences. Currently, we are performing texture analysis to identify these variations of OCT images among three growth forms and across species.

In summary, OCT provides visualization of the bacterial species in each environment via 2D and 3D reconstruction. This visualization and complementary optical properties values provide a quantitative metrics for differentiating across growth environment and species. In particular, the mean attenuation coefficients of bacterial colonies, planktonic form, and biofilms allowed for significant differentiation of each species in the three different growth environments and in most cases across species. Overall, a major advantage of OCT is the depth imaging of each species which allow for better understanding of the bacteria grows within its growth environment.

This study has additional considerations that are worth mentioning. First, in terms of sample preparation and growth methodology, it should be noted that bacterial growth largely depends on the growth media and environments such as CO₂, temperature, pH, humidity, and substrates. In literature, various culture media, agars, and environmental conditions have been used to grow these bacteria. The experiments in this study followed the standard protocols as mentioned in the Methods section for the optimal growth of bacteria in colonies, planktonic, and biofilm forms. Briefly, nutrient-rich BHI broth and agar were used as culture media for all bacterial species except for H. influenzae. For H. influenzae, BHI media/agar supplemented with hemin and NAD was used. All the samples were incubated at 37 °C, 5% CO₂. Also, the initial attachment of planktonic bacteria on the substrate surface is the primary step for growing a bacterial biofilm. Herein, biofilms were grown on calcium fluoride slides (for RS measurements) and poly-d-lysine coated culture dishes (for OCT measurements) for initial adhesion of bacteria and growth of the biofilms. Calcium fluoride was necessary to minimize substrate background noise in the Raman spectra. We know that other growth media and conditions could potentially be used to culture the different species in the various growth forms, which could impact both the biochemical and scattering properties of the cells. Therefore, the media and growth conditions were standardized in this study to minimize the influence of these external factors. Also, planktonic bacteria in BHI media were cultured from the overnight growth of bacteria colonies on BHI agar. The RS and OCT measurements were acquired on dense spun-down pellets obtained from these planktonic bacteria to improve the signal-to-noise of the spectroscopic measurement.

Other factors to consider relate to the methodology used to quantify the RI and AC values, as several methods can be used to calculate these values. Herein, the RI measurement was based on the ratio of the optical thickness and the geometric thickness of the sample from a single OCT B-scan frame as shown in the flow diagram of Figure 7 (Step - 5). The bulk RI of the sample can be effectively measured with our current SDOCT setup, and the sensitivity of RI is ~0.01. For the AC measurements, a depth-resolved single-scattering algorithm was implemented to determine the intensity variations of adjacent voxels and to recover the amount of attenuation that occurs pixel-wise. The algorithm assumes the complete attenuation of light through the sample and the linear relationship between backscattering and attenuation, which overestimates the AC value at the greater imaging depths of the OCT B-scan images. The thicknesses of our samples were comparatively small (~200 µm), allowing for the overestimation error at greater depths to be negligible, but for thick samples, these errors were mitigated by fitting an exponential curve on averaging the last few pixels of the bottom of the sample (OCT B-scan), also known as the optimized depth-resolved estimation (ODRE) algorithm (Liu et al., 2019).

Overall, these results demonstrate the advantages provided by both Raman spectroscopy and OCT to detect and characterize differences in the biochemical composition and scattering properties of the four primary OM-causing bacterial species. Moreover, the ability to differentiate these species in both their planktonic and biofilm forms provide a valuable dual contrast approach, based on the biochemical and morphological differences, in assessing the presence and growth environment of each bacterial species. With the current clinical standards requiring complex secondary tools such as PCR and nucleic acid sequencing for the identification of bacterial species, and the lack of currently-used tools capable of detecting effusions and biofilms, in real-time, a dual RS-OCT approach has the potential to be a useful multi-modal tool for future in vivo detection and assessment of OM. However, translating such technologies for OM diagnosis in vivo requires resolving some physiological constraints and design challenges. For instance, acquiring Raman spectra of effusions or biofilms from directly behind the delicate TM with a fiber optic RS probe is challenging. Therefore, we recently developed a prototype of a multimodal RS-OCT system and successfully tested it in ex vivo ear models (i.e., clinical effusion under chinchilla TM). This provided an understanding of the specific physiological limitations (improved positioning feedback via OCT for RS measurements) and design challenges (i.e. maximum permissible exposure and optical safety limits) for signal collection within the middle ear (Monroy et al., 2022). Currently, we are designing and implementing our 2^nd generation RS-OCT for the clinical studies of in vivo effusions and biofilms. With additional study and refinement, the diagnostic capabilities of this combined RS-OCT platform may positively impact in vivo detection and assessment of OM in the future.

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

AKL and FRZ were involved in designing, performing, and analyzing the experiments and results, and drafting the manuscript. KS was involved in performing the experiments and analyzing the results. SF was involved in drafting and editing the manuscript. GLM, HC, and JW were involved in editing the manuscript. AM-J. and SAB were involved in designing the experiments, editing the manuscript, and acquiring funding. AKL and FRZ are co-first authors. All authors contributed to the article and approved the submitted version.

The authors appreciate the support from the National Institutes of Health (NIBIB: R01EB028615).

The authors would like to acknowledge support from the Imaging Technology Group and Microscopy Suite of the Beckman Institute for Advanced Science and Technology for generating CLSM images from their confocal laser scanning microscope (CLSM) system (Leica TCS SP8 CLSM, Leica Microsystems Heidelberg GmbH, Manheim, Germany).

SAB, GLM and JW have disclosed and patented intellectual property with the University of Illinois Urbana-Champaign related to the optical detection and characterization of OM. AM-J has disclosed and patented intellectual property with Vanderbilt University related to the optical detection and characterization of bacteria. SAB is a co-founder and Chief Medical Officer of PhotoniCare, Inc., which is developing OCT for the detection and diagnosis of OM.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.