1.

Introduction

Middle-ear infections, commonly known as otitis media (OM), affect more than 80% of children.¹ Current studies estimate the diagnostic accuracy for OM by an expert physician using standard otoscopy at 60% to 70%.² Existing technology used to diagnose OM is limited primarily to illumination and magnification devices,³ such as otoscopes or surgical microscopes, which only provide a surface view of tissue. Diagnosis using these tools relies on accurately interpreting visual features of the tympanic membrane (TM) or the accurate detection of the presence of any middle-ear effusion (MEE). Pneumatic otoscopy, while considered the gold standard for OM diagnosis, is difficult to perform correctly and is often not utilized in daily clinical practice.⁴ Other acoustic techniques, such as tympanometry and acoustic reflectometry, provide additional insight into the functional behavior of the TM and middle-ear mobility. While tympanometry can help to infer middle-ear dysfunction and provide slightly improved diagnostic accuracy over otoscopy alone, tympanometry results are still recommended to be considered alongside otoscopy findings.⁵ Persistent or recurrent cases of OM are treated with tympanostomy tube (TT) placement, one of the most commonly performed surgeries in children requiring general anesthesia.^6,7 While $\sim 6\%$ of all children in the US will receive TT before the age of 2, there are reports that physicians often inconsistently recommend TT placement when verified against treatment guidelines,⁸ although varying subpopulations and differing treatment implementation strategies across the US explain this variance.^8–12

This study is motivated by previous efforts that used optical coherence tomography (OCT), a noninvasive and high-resolution medical imaging technology, to visualize the TM and middle ear. Past studies have used OCT in surgical settings,^13–15 although its use has previously been limited mainly to ex vivo^16–19 tissue due to technical challenges and physical limitations of light delivery to the ear. Our group has addressed many of these challenges and performed both animal^7,20 and human^6,21–25 studies. Recently, one study²¹ identified and visualized additional infection-related structures affixed to the TM in nearly every case of chronic OM, which led to our current interest to understand longitudinal changes in the observed structures following surgical intervention.

The pathogenesis of these additional infection-related structures observed in OCT images is unclear. Yet, the presence of biofilms during infection in the respiratory tract and upper airway, including the adenoids and middle-ear cavity (MEC), is understood and well established.^20,26–34 There remains, however, a lack of understanding of the specific role that biofilms play during “healthy” or “infected” conditions in the middle ear, as previously it has not been possible to regularly and longitudinally observe these biofilms in vivo. Biofilms are communities of bacteria enclosed in a self-produced extracellular polymeric substance.³⁵ Biofilms protect bacteria from host immune response mechanisms^36–43 and antibiotic treatments. Given the possible virulence, persistence, and/or recurrence of biofilm-associated infection in the ear, understanding the mechanisms that drive the infection process was recently declared a grand challenge in pediatric otolaryngology.⁴⁴

In this study, an OCT system combined with video-based surface imaging of the TM was used to prospectively observe patients through consecutive preoperative, surgical, and postoperative visits to quantitatively characterize and assess the efficacy of surgical TT placement to clear middle-ear infection. It is expected that a patient with a successful surgical intervention would have no further clinical symptoms, and OCT images would be clear of any image-based features that would be associated with OM, including any MEE or biofilm. In the presence of persistent or recurrent clinical symptoms of OM, we expect to find OCT image evidence of persistent or recurrent MEE and/or middle-ear biofilm.

2.

Materials and Methods

2.2.

Handheld Probe and Portable OCT System

A custom-built handheld probe and portable OCT system (Fig. 1) were used to image subjects. The system utilized a broadband optical source (Superlum, Ireland) centered at 860 nm with an approximate bandwidth of 135 nm at full width at half maximum, which provided an axial resolution of $\sim 2.4\mu \mathrm{m}$ and a transverse resolution of $15\mu \mathrm{m}$, in air. A line scan camera-based spectrometer (Wasatch Photonics) provided data to process cross-sectional B-mode images. The OCT system emitted an optical power of 2.5 mW onto the TM, which is well below ANSI standard safety limits.⁴⁸ A separate, commercial video otoscope (Welch Allyn) was used to document surface images of the TM. Additional details regarding the system can be found in a prior publication.⁴⁹

Fig. 1

Portable optical imaging system and handheld OCT probe (a) within the specialist’s clinic and (b) within the operating theater. (c) Handheld probe utilizes standard interchangeable ear specula tips that are replaced between patients.

3.

Results

Twenty-five subjects (Table 1) were imaged in this study, which concluded without any adverse events. Patient history was available for any performed examination or procedure. Risk factors indicate an increased likelihood of recurrent OM. Subjects were split into two groups: the “C” (complete) group, and “P” (partial) group. Six subjects in the C group (C1 to C6) were observed at all three time points: preoperative consultation, intraoperatively, and at the follow-up exam. Nineteen subjects in the P group (P1 to P19) were observed at the preoperative consultation visit and then either did not receive surgery (5/19), did not return for follow-up exam after surgery (5/19), or were unable to be imaged given scheduling conflicts (9/19), which is clarified in the “observed visits” column. The recurrence column indicates treatment outcome through any OM-related visit after surgery to the primary care or otolaryngology clinic, which are further described in the footnote of Table 1.

Table 1

Summary of imaging subjects and clinical history.

A simplified visual breakdown of subjects enrolled in this study is presented in Fig. 2, organized by the day of observation. All subjects presented initially with a history of recurrent OM or Eustachian tube dysfunction (ETD) and were observed throughout their treatment with OCT. On their day of preoperative consultation, 92% (23/25) of subjects had a biofilm identified with OCT. On the day of surgery, 11 subjects were observed with OCT, and 91% (10/11) of subjects had a biofilm identified with OCT immediately prior to surgical intervention. At the follow-up examination, six subjects returned for OCT imaging. These subjects were compliant with the study schedule and comprised the C (complete) group, with OCT data and clinical reports available from each time point. Of the subjects in the C group, two of six subjects had additional structures identified with OCT (subjects C4 and C5), and subject C5 reported recurrence issues of OM. In 67% (4/6) of subjects, surgery was effective at clearing symptoms of OM and had no biofilm identified with OCT. For the remaining 19 subjects that were noncompliant with the study schedule, only partial data was able to be collected. The P (partial) group included clinical reports and OCT data from the day of consultation and reports if/when care was received at a later unscheduled visit. While 8 of 19 subjects returned for all care, 88% (7/8) of subjects later reported persistent OM symptoms and had biofilm-like structures previously identified with OCT at an earlier visit. Six of 19 subjects did not return for follow-up care after receiving TT surgery during the study observation period. While follow-up visits and care are recommended as standard practice and included in the cost of surgery, one review found that only 25% of patients return for follow-up care, despite guidelines that recommend patients regularly follow-up with otolaryngologists until TT has extruded.⁵² While dependent on the type of TT used, most TTs safely extrude on their own within a year,⁵³ though follow-up care may still be needed in the future if complications occur. Five of 19 subjects did not receive TT surgery or have a follow-up visit scheduled.

Fig. 2

Graphical breakdown of subjects enrolled in this study. On the day of preoperative consultation, 23 out of 25 subjects had a biofilm identified with OCT. On the day of surgery, 11 subjects were observed with OCT, and 10 out of 11 subjects had biofilms identified with OCT prior to surgical intervention. On the final day of follow-up, 6 subjects returned for OCT imaging. These subjects comprised the complete group identified in Table 1, with OCT data and clinical reports available from each time point. Two out of 6 subjects had additional structures visible in OCT images at follow-up, and subject C5 reported persistent OM symptoms. Four out of 6 subjects had no reported OM symptoms and had no biofilm identified with OCT. Only partial data were able to be collected from the remaining 19 subjects who were noncompliant with the study schedule. The partial group included OCT data from the day of consultation and patient reports if/when care was received at an unscheduled visit. Eight out of 19 subjects that returned for unscheduled visits reported persistent OM symptoms, and 7 out of 8 of these subjects had a biofilm previously identified at the day of consultation. Six out of 19 subjects did not return for follow-up care after receiving TT surgery, and 5 out of 19 did not receive TT surgery or have a follow-up visit scheduled. OM: otitis media and TT: tympanostomy tube.

In the C group, five out of six subjects (83%) had no recurrence of symptoms after TT surgery. In four out of six subjects, cross-sectional OCT images at the postoperative follow-up visit showed that the MEC was free of biofilm and MEE, although a biofilm had been visualized with OCT during the preoperative visit and intraoperatively during TT surgery. Figure 3 highlights a representative case where subject C3 presented with chronic OME and ETD and was prospectively imaged with OCT to identify changes following TT surgery. Figure 3(a) shows data from the first visit to the otolaryngology clinic (patient evaluation and preoperative consultation). Cross-sectional OCT shows a bright, scattering biofilm structure (green arrow) affixed to the medial mucosal surface of the TM, as well as evidence of suspended, optically scattering particles in the MEC, indicative of an MEE (white arrow, dashed). The corresponding video otoscope surface image shows a dull and relatively featureless TM (inset). Here, and in subsequent figures, yellow and green bars denote the location and thickness of the TM and biofilm, respectively, as determined using a previous method.²¹ Figure 3(b) shows intraoperative image data from this same subject and ear (second visit) prior to myringotomy, aspiration of the MEC, and TT placement. OCT imaging confirmed the persistence of infection-related image-based features, including the biofilm (green arrow, green line) and MEE (white arrow, dashed) imaged at the consultation visit, although scanned at a slightly different TM location within the light reflex region.

Fig. 3

Representative longitudinal interventional data from subject C3. Cross-sectional OCT images were acquired at (a) the preoperative evaluation and surgical consultation, (b) the day of surgery prior to myringotomy, and (c) at the postoperative follow-up. Arrows denote the additional infection-related components. Digital otoscopy images of the ear taken at the surgical consultation and follow-up are shown as insets, where white-dashed lines denote where the cross-sectional OCT images were acquired. (a and b) Biofilms (green arrows) are visualized as being affixed to the TM, and effusions (containing many point-like optically scattering particles in suspension) (white arrows, dashed) are visualized within the MEC. (c) Images from the follow-up show a cleared MEC with an image acquired near the umbo, a thicker region of the TM when compared to the light reflex. Earwax is also present on the outer surface of the TM (orange arrows, orange line), and the TT is visible deeper within the MEC (blue arrows, dashed). Surgical intervention appears to have cleared any biofilm and effusion components visible under OCT from this MEC. Scale bars represent 100 μm in depth. Yellow, green, and orange bars indicate the positions and thicknesses of the TM, biofilm, and earwax, respectively.

Approximately one month later, the same subject was imaged at the third visit (postoperative follow-up) following standard postsurgical care. Here, the microstructural effects of surgical intervention are visualized. In Fig. 3(c), the MEC is free of any MEE and biofilm. A portion of the TT (blue arrows, dashed) deeper inside the MEC is visualized. The TM (yellow line) appears thicker due to the presence of residual earwax (orange arrows, orange line) on the outer surface of the TM and because the image was acquired closer to the umbo, where the TM is naturally thicker.⁵⁴ The surface image (inset) shows that some transparency has returned to the TM. Otoscopic metrics provide visual confirmation of surgical outcomes and, in this case, indicate a reduction in inflammation and a fluid-free middle-ear space compared to presurgery time points. The TT is observed in place and functionally venting the MEC.^47,55 At the end of the records-based observation period, this subject did not report any symptoms of OM recurrence.

When an infection occurred after postoperative follow-up or perhaps when surgery was ineffective at clearing infection-related components, additional biofilm- or MEE-related structures were observed affixed to the TM and/or within the MEC, respectively. Figure 4 highlights subject C5, who was presented with a history of recurrent episodes of acute OM that persisted after TT placement. Visible biofilm structures detected in OCT cross-sectional images persisted throughout all observation time points, including pre-TT placement [Figs. 4(a) and 4(b), green arrows] and post-TT placement [Fig. 4(c), green arrows]. Yellow and green bars indicate the location and thickness of the TM and biofilm, respectively. Longer-term follow-up was possible given a recurrence of the infection in this subject, who required additional outpatient visits. Approximately six months after the postoperative visit, a patchy or spatially inconsistent biofilm was observed affixed to the TM [Fig. 4(d)], while in some regions no biofilm was present [Fig. 4(e)]. The identification of both seemingly normal and pathological regions in these OCT images highlighted here was unexpected and found in only one ear from this one subject. However, this finding demonstrates the complex dynamics of biofilm formation in OM and regression after treatment, indicating that further investigation and longitudinal tracking at more time points may better capture these processes. The other ear of this subject and all other study subjects had consistent findings from each time point. No digital otoscopy image was available due to patient discomfort. An indisputable and unavoidable limitation in studies involving human subjects, especially those in active clinical and surgical settings, is the priority of patient needs and well-being over data collection, which in some cases presented a challenge for acquiring complete datasets. This subject was provided with antibiotics based on the physical exam and was scheduled for re-evaluation at a later date.

Fig. 4

Longitudinal imaging of the TM and infection-related components from study participant C5. Cross-sectional OCT images of the middle ear are shown at various time points in the progression of infection and treatment of subject C5, along with inset digital otoscopy images of the ear: (a) patient evaluation and surgical consultation, (b) day of surgery prior to myringotomy, (c) postoperative follow-up, and (d and e) at later, long-term follow-up visits. Arrows denote persistent additional, infection-related structures. Digital otoscopy imaging for (d) and (e) was not possible due to patient temperament and discomfort. Subject C5 received TTs to treat recurrent acute otitis media (AOM) and hearing loss. After a one-month recovery from TT placement (c), additional middle-ear structures were visible (green arrows). However, (d) and (e) show two images from long-term follow-up exams roughly six months after (c), after the subject revisited both the primary care office and specialists clinic for OM-related complaints. While (d) shows additional structure (green arrows), (e) appears normal, perhaps indicating a persistent, but spatially inconsistent and patchy middle-ear biofilm within the light reflex region of this subject. Scale bars represent 100 μm in depth. Yellow and green bars indicate the position and thickness of the TM and biofilm, respectively.

Available data from the P dataset was typically limited to the first consultation visit and included OCT images, digital otoscopy images, and patient history. Imaging data from subsequent visits were not available and were lost to follow-up, a common problem in otolaryngology,⁵² as imaging subjects in this observational study were free to seek, reschedule, or not seek care as per their wishes. Almost all subjects in the P group had a reported history of recurrent OM (18/19, 95%) and mostly (17/19, 90%) were found with OCT to have biofilm-like structures affixed to the TM during the first visit, as noted in our previous studies. A total of 20 out of 25 subjects (80%) underwent TT surgery. However, 8 out of 25 (32%) subjects reported recurrence of infection during the postoperative observation period, which may have resulted from local recurrence of the original infection and biofilm or from a new, discrete OM infection. While TT surgery was performed depending on the clinical circumstances of each patient, it is notable that 3 out of 19 (16%) in our study had a history of multiple TT surgeries, with 2 of these subjects having biofilm-related structures present in OCT images. Additional figures are presented in the Appendix that demonstrates the diverse presentation of recurrent or chronic OM encountered in this study.

4.

Discussion

Additional infection-related structures were initially observed in OCT images from most subjects (92%, 23/25) diagnosed with recurrent or chronic OM and scheduled to undergo TT surgery, as per their standard-of-care. Through record-based follow-up of all subjects, surgical intervention was effective at clearing the infection and reducing or eliminating symptoms in 15 out of 20 subjects (75%) that received TT surgery. If observed with OCT, image analysis confirmed the absence of a middle-ear biofilm in these subjects. However, 5 out of 20 subjects (25%) that received surgery had persistent clinical symptoms of OM infection following TT surgery, and if observed with OCT, infection-related structures were present. These structures were similarly identified in subjects that had multiple surgical interventions and had returned for repeated treatments. Therefore, based on OCT imaging results, if a subject did not respond to surgical intervention, there was likely an insufficient clearance of infectious material from the MEC or a recurrence of the bacterial population and subsequent biofilm regrowth. A postsurgical OCT confirmation of the absence of a biofilm could then perhaps serve as a secondary diagnostic indicator or predictor for long-term success of the surgical intervention. A future study could similarly explore the effectiveness of pharmacological interventions to clear symptoms of OM in subjects identified with a middle-ear biofilm.

Surgical intervention, in this case myringotomy and TT placement, results in both short- and long-term effects that influence several changes in an infected MEC. With myringotomy, a renewed immune and healing response is stimulated⁵⁶ along with aspiration of the MEC and insertion of a TT. These interventions aid in clearing middle-ear fluid and provide a means of aerating the MEC. When considered in the context of biofilm-related infections, mechanical clearance and aspiration of the MEC disturb any biofilm and its growth environment both structurally and biochemically. Ototopical drops placed into the MEC also provide a higher and more effective concentration of antibiotics directly to the site of infection^57,58 when compared to oral antibiotics.⁵⁹ Successful clearance of the MEC and reduction of OM-related symptoms may indicate these structures are cleared by a combination of these factors.

Regardless of any potential new diagnostic information gained with OCT, considering recurrent or chronic OM as a biofilm-mediated disease would also require investigating new or re-evaluating existing treatment options.^60,61 Methods that attempt to noninvasively weaken, disrupt, or clear a biofilm^62–64 could be employed before invasive means are considered or even employed in conjunction with surgical treatment for a more aggressive approach. However, if TT placement is found to be one of the most effective therapeutic interventions to treat biofilm-derived recurrent or chronic OM, then it will be necessary to weigh the related risks and outcomes associated with presumably more patients receiving tubes at a younger age. Serious consideration should be given to the fact that a majority of patients tend to outgrow complications and/or susceptibility to recurrent OM, perhaps due to improved function and increased diameter of the Eustachian tube with age⁶⁵ or improved immune function with age.⁶⁶ Conversely, it is also important to consider that many children are unnecessarily hearing impaired during this time, which delays learning and speech development at a crucial age in young subjects. Further investigation is undeniably necessary to observe broader subject population demographics, as well as biochemical and genetic factors, to more fully understand the dominant factors of middle-ear disease pathogenesis, as well as any physical or secondary indicators that can be used for diagnosis and monitoring. In this regard, tracking the reduction or clearance of biofilm biomass with OCT after surgical or pharmacological treatment may provide further insight into mechanisms for effective treatment of OM.

To suggest any potential use of this OCT data for clinicians, further study is still needed and should include a randomized clinical trial with a larger subject population and a wider geographic and demographic distribution to capture a broader view of OM. With this longer-term study, diagnostic features in OCT images may be identified that predict a need for surgical intervention. This may allow for a quantitative and evidenced-based referral to surgery at earlier time points in the pathogenesis of chronic OM, potentially shortening the duration of disease symptoms and related morbidity prior to treatment and successful resolution. In this regard, OCT-based treatment guidelines for OM will need to be proposed, established, and evaluated, as was done for clinical cardiovascular⁶⁷ OCT and clinical ophthalmic⁶⁸ OCT clinical applications.

There are several limitations in this initial observational study. With the spectral-domain OCT system used in these investigations, it was not technically possible to image deeper to the medial mucosal surfaces of the MEC. However, swept-source OCT systems can provide deeper imaging to over a centimeter,⁶⁹ which can provide additional information on the location, extent, and severity of any middle-ear biofilm or MEE simultaneously on both the interior mucosal surface of the TM as well as the mucosal surface throughout the MEC. While the light reflex region of the TM was interrogated in this study and used as a consistent anatomical landmark for longitudinal image collection, the remaining areas of the TM may provide further information related to the biofilm or other middle-ear diseases. Techniques for rapidly generating three-dimensional images of the entire TM, while minimizing motion artifacts and coregistering these datasets among longitudinal visits are in development. The presence of specific otopathogens may ultimately govern whether surgery is effective.⁷⁰ However, OCT is not sensitive to specific bacterial species. Ongoing research in our group is investigating the use of Raman spectroscopy for determining microbial content of the middle ear,⁷¹ which has previously been used to differentiate bacterial strains in culture.⁷²

Finally, additional studies are needed to verify longer-term outcomes of study participants with OCT and the clinical relevance of the presence or absence of a middle-ear biofilm in patients with chronic OM. Another ongoing study is aimed at physically sampling, identifying, and verifying the infection-related structures and microbiological content that has been observed on the TM with OCT. Results will help to better understand and correlate the image-based findings with the presence of biofilms that were previously detected on middle-ear mucosa during chronic OM.²⁶

5.

Conclusion

In this study, a broad subject population was prospectively observed using OCT to assess the efficacy of surgical intervention to clear OM-related middle-ear fluid or biofilms and associated clinical symptoms. OCT image findings suggest that surgical intervention, consisting of myringotomy followed by TT placement, provides a means to clear the middle ear of infection-related biofilm structures and components. OCT shows promise to provide a clinically viable diagnostic and monitoring platform capable of providing physicians with new image-based information about the disease state of the middle ear. With further investigation, OCT image features may aid in the diagnosis and management of OM, especially when interpreted alongside physical exam, otoscopy findings, and patient history.