In Vivo Optical Coherence Tomography: The Role of the Pathologist

Optical coherence tomography (OCT) is a nondestructive, high-resolution imaging modality that generates cross-sectional images with endogenous contrast based on mismatches in index of refraction (Figure 1, A).^1–3  Optical coherence tomography is similar in principle to ultrasound, but, rather than measuring the sound echoes reflected back from the tissue, OCT measures the amount of backscattered near-infrared light. This results in significantly higher resolutions (<10 μm axial, 20–40 μm lateral) than can be achieved with ultrasound. However, the rapid attenuation of the OCT signal in tissue, because of the scattering and absorption of light, results in shallower depth penetration which is typically limited to 1 to 3 mm.⁴ 

Because of the rapid speed of light propagation in tissue, the delay of the optical echoes is measured with low-coherence interferometry. Light from a broadband near-infrared source is split into 2 paths: one path travels to a sample arm, where light is focused onto the tissue sample, and the other travels to a reference arm, which typically consists of a mechanically scanning reference mirror.^1–3  When the optical path length of light backscattered from the sample and light reflected from the reference mirror are within the coherence length of the light source, constructive interference occurs, and a signal is generated. When the light returning from the sample and reference arms are not within the coherence length of the light source, destructive interference occurs, and no signal is generated. This allows for precise detection of backscattered signal associated with a specific tissue depth. To generate an axial depth profile, or A-line, the reference mirror is mechanically scanned to alter the optical path length of the reference arm and, therefore, the location within the tissue. Two-dimensional and 3-dimensional images can be generated by systematically scanning the OCT light source across the tissue.

Optical coherence tomography is limited in its imaging speed to only a few frames per second. Optical frequency domain imaging (OFDI), also termed frequency domain optical coherence tomography or swept source optical coherence tomography, is a second-generation OCT imaging technology capable of significantly more rapid image acquisition rates (100 times faster, up to 200 frames/s) without compromising image quality (Figure 1, A).^5–12  This increased image acquisition speed allows for high-resolution imaging over larger tissue volumes, providing a method of in vivo comprehensive microscopy. Both OCT and OFDI systems are compatible with benchtop imaging and a variety of catheter-based imaging systems (Figure 1, A through C). Throughout the rest of this article, the term OCT refers to both traditional time-domain OCT and OFDI.

The nondestructive nature and high-resolution inherent to OCT imaging is ideal for imaging tissue during standard endoscopic procedures. Optical coherence tomography is particularly useful when tissue cannot be removed for pathologic assessment, such as in assessment of coronary artery disease. In clinical scenarios where tissue biopsy samples are collected to assess pathology but suffer from sampling errors, such as the assessment of intestinal metaplasia and dysplasia of the esophagus, OCT can aid in guiding biopsy-site selection to reduce the likelihood of missed diagnoses. In pulmonary pathology, tissue biopsy may suffer from small sample size as well as sampling error. Optical coherence tomography can aid in guiding biopsy-site selection to ensure accurate targeting of lesional tissue and to increase diagnostic yield of tissue biopsy. The large volumes of tissue assessed by second-generation OCT also provide a form of additional virtual in vivo tissue, which could be used for pathologic assessment, in addition to the physical tissue collected. As the imaging features of various tissue pathologies are developed and validated, OCT may be able to provide an in vivo optical biopsy. The pathologist is well suited to play a role in the development and implementation of OCT in clinical practice, interpreting high-resolution OCT images as a complement to standard tissue pathology. Although OCT has been conducted in many clinical applications, we will briefly touch on intracoronary, esophageal, and pulmonary OCT imaging.

The vulnerable plaque, termed the thin-cap fibroatheroma, has been identified as the culprit lesion in approximately 80% of sudden cardiac deaths.^13–19  The thin-cap fibroatheroma typically consists of a minimally occlusive, atheromatous lesion with the following histologic features: (1) thin, fibrous cap (<65 μm); (2) large lipid pool; and (3) activated macrophages near or within the fibrous cap.^13–19  Clearly, biopsy of coronary arteries in patients to evaluate these features is not feasible. The ability to assess vulnerable plaques nondestructively with near-histologic resolution during in vivo, intracoronary imaging may significantly enhance the evaluation and management of patients with coronary artery disease.

Intravascular optical coherence tomography has been studies by many groups.^12,20–38  Thin, flexible intracoronary OCT catheters have been developed that are compatible with standard percutaneous coronary intervention techniques (Figure 1, A). These catheters facilitate imaging of the coronary arteries during proximal balloon occlusion or flushing of the vessel to temporarily displace the highly light-scattering blood from the imaging field of view. Large scale, ex vivo studies have been performed to develop and validate OCT diagnostic criteria for atheromatous lesions, including the identification of features of thin-cap fibroatheromas (Table 1).^20–30,37 

Fibrous plaques were characterized by homogeneous, signal-rich regions; fibrocalcific plaques by signal-poor regions with sharp borders; and lipid-rich plaques by signal-poor regions with diffuse borders (Figure 2, A through F). These OCT imaging criteria demonstrated high sensitivities and specificities for various plaque morphologies (71%–79% and 97%–98% for fibrous plaques, 95%–96% and 97% for fibrocalcific plaques, and 90%–94% and 90%–92% for lipid-rich plaques, respectively) in blinded assessments.²⁴  Calcific nodules appear as signal-poor regions with a sharp delineation between the calcific nodule and the surrounding tissue. Optical coherence tomography imaging of cholesterol crystals demonstrated linearly oriented, highly reflecting structures within atheromatous plaques.²⁴  The high refractive index contrast results in strong optical signals from macrophages. Studies evaluating the ability to quantify macrophage density in fibrous caps yielded greater than 90% sensitivity and specificity for identifying caps containing more than 10% CD68 staining.³²  Optical coherence tomography identification of all 3 components of thin-cap fibroatheromas with high sensitivity and specificity provides nondestructive, in vivo assessments of vulnerable plaques during standard percutaneous coronary intervention procedures. Intracoronary OCT is becoming a widely used tool in interventional cardiology with the potential to affect clinical management of coronary artery disease.

Barrett's esophagus, or specialized intestinal metaplasia (SIM), is a major risk factor for the development of esophageal adenocarcinoma. For patients with known SIM, periodic endoscopic surveillance is recommended to screen for high-grade dysplasia and intramucosal carcinoma.^39–44  Current guidelines for surveillance recommend 4-quadrant biopsies at 1- to 2-cm increments along the axial length of the glandular mucosa of the distal esophagus; however, the efficacy of surveillance endoscopy is limited by sampling error.^40–44  The ability to provide a more targeted approach to surveillance endoscopic biopsy via in vivo OCT may decrease sampling error and increase diagnostic yield.

The use of OCT for differentiating esophageal pathology relevant to screening and surveillance of Barrett's esophagus has been conducted by many groups.^45–53  A novel balloon-centering OCT catheter was developed to conduct volumetric imaging of the entire distal esophagus (6 cm) (Figure 1, B).⁵³  Squamous mucosa is characterized by the presence of layering without epithelial glands. Gastric cardia is characterized by the identification of at least 2 of the following features: vertical pit structure; well-defined, epithelial surface reflectivity; relatively poor image penetration; and/or broad, regular foveolar regions or rugae. Specialized intestinal metaplasia is characterized by the presence of glands in a layered epithelium, or at least 2 of the following features: lack of layered or vertical pit architecture; heterogeneous scattering; or an irregular surface (Figure 3, A through F). The OCT diagnostic criteria are summarized in Table 2.⁵³  The potential of OCT to diagnose SIM has been demonstrated in prospective studies, with sensitivities of 81% to 97% and specificities of 57% to 92%.^45–53  Additionally, studies have demonstrated successful grading of dysplasia with OCT. Reported sensitivities and specificities for detecting high-grade dysplasia/intramucosal carcinoma range from 54% to 83% and 72% to 75%, respectively.^45–53  Currently, studies are being conducted to evaluate real-time endoscopic OCT-guided biopsy with laser marking in patients.^53–55 

Pathology diagnosis of lung biopsies and fine-needle aspirations can be limited by small tissue volumes and/or sampling errors. Larger quantities of tissue can improve diagnostic yield but are often difficult to obtain. The ability to assess larger tissue volumes by high-resolution OCT during bronchoscopy could potentially enhance biopsy site selection of targeted lesions and provide considerably more architectural information via “virtual tissue,” which may be used to aid diagnosis.

Helical-scanning OCT catheters can be used to generate cross-sectional or volumetric images of the airways and are ideal for endobronchial imaging (Figure 1). Second-generation OCT (OFDI) is capable of imaging entire bronchial segments, spanning multiple airway generations, and obtaining comprehensive microstructural evaluations in 3 dimensions. Optical coherence tomography has been used to assess the pulmonary airways and parenchyma in animal models^56–61  and in vivo human airway.^62–68  Optical coherence tomography imaging of the normal bronchial wall (Figure 4, A through E) reveals the fine, layered features of the airway, including the epithelium, basement membrane, lamina propria, salivary-type glands and ducts, vessels, and cartilage with surrounding perichondrium. In more distal airways, airway layering and attached lattice-like, signal-void alveoli can be appreciated. In vivo OCT imaging of airway-based carcinomas reveals architectural disorganization of the bronchial layering, segments of mucosa where normal surface maturation is lost, and increased surface-image intensity when compared with the underlying tissue.^64,66,67 

The ability of OCT to discern preinvasive cancers of the bronchial mucosa has been assessed.^67,68  Lam et al⁶⁸  used OCT to evaluate bronchial mucosal lesions identified by autofluorescence bronchoscopy in a group of high-risk smokers. A total of 281 OCT images were evaluated from 148 patients (145 normal/hyperplasia, 61 metaplasia, 39 mild dysplasia, 10 moderate dysplasia, 6 severe dysplasia, 7 carcinoma in situ, and 13 invasive carcinomas). Normal respiratory epithelium or hyperplasia was characterized by 1 or 2 cell layers above a highly scattering basement membrane and upper lamina propria. Thickening of the epithelial cell layer was observed in metaplasia, various grades of dysplasia, and carcinoma in situ. Quantitative evaluation of epithelial thickness revealed that invasive carcinoma appears to have a significantly thicker epithelium than does carcinoma in situ (P = .004). Dysplasia was found to have significantly thicker epithelium than was seen in metaplasia or hyperplasia (P = .002). The basement membrane was still intact in carcinoma in situ, but was discontinuous or no longer visible in invasive cancer.⁶⁸ 

Solitary pulmonary nodules frequently require biopsy to determine malignant potential. Peripheral lesions may be assessed via transthoracic fine-needle aspiration, which has reasonable diagnostic yields but has an increased rate of pneumothorax (pneumothorax rate of approximately 25%, of which, at least 15% require chest tube insertion).⁶⁹  Transbronchial needle aspiration has a much lower risk of pneumothorax, but the diagnostic yield can be as low as 14%, even with the acquisition of 4 to 8 serial tissue specimens.^69–71  The diagnostic yield of transbronchial needle aspiration does increase when performed in conjunction with biopsy-guidance techniques, such as endobronchial ultrasound⁷² ; however, the yield is still unacceptably low, particularly for lesions smaller than 2 cm in diameter (<33%).^{69–71,73,74}  Recently developed, needle-based OCT catheters provide a method to image peripheral lung (Figure 1, C).^59,60  In particular, a novel, needle-based OCT catheter, developed to be compatible with standard bronchoscopic transbronchial biopsy/needle aspiration, has the potential to access solitary, peripheral nodules, assess needle location just before tissue acquisition, and potentially increase diagnostic yield.⁶⁰ 

Preliminary studies have demonstrated the potential of OCT for identifying pulmonary pathology in vivo. However, before OCT can make a significant impact, in vivo OCT-image interpretation criteria for pulmonary pathology need to be developed and validated, similar to the validated OCT criteria for intracoronary and esophageal imaging. To appropriately characterize OCT imaging features of normal airway, parenchyma, and pulmonary pathology, images need to be correlated one-to-one with histopathology, which can only be performed in the ex vivo setting. In a recent ex vivo study, initial OCT-image feature interpretations of common, benign, and malignant lung lesions and interstitial fibrosis were described by precisely correlating OCT images with histopathology (Table 3)⁵⁷ :

• •

  Carcinomas displayed architectural disarray with loss of normal airway/alveolar structure and rapid light attenuation.

• •

  Squamous cell carcinomas showed nested architecture.

• •

  Atypical glandular formation was seen in adenocarcinomas.

• •

  Uniform trabecular gland formation was seen in salivary gland carcinomas.

• •

  Mucinous adenocarcinomas (previously know as mucinous bronchioloalveolar carcinomas) showed alveolar wall thickening with intra-alveolar mucin (Figure 5, A through F).

• •

  Cartilaginous hamartomas showed well-circumscribed, lobulated architecture with evenly dispersed, fine-scale regions of high signal intensity.

• •

  Interstitial fibrosis was visualized as signal-dense tissue, with an interstitial distribution in mild interstitial fibrotic disease and a diffuse, subpleural pattern of signal-dense tissue with cystic space formation in usual interstitial pneumonitis (Figure 6, A through E).⁵⁷ 

The acquisition of large, volumetric OCT data sets, both endobronchially and transbronchially, may provide additional architectural information (comparable to low power bright field microscopy) to accompany standard biopsies. For instance, OCT could be particularly useful in instances where small tissue biopsies are needed for multiple pathology assessments, such as in biopsies of carcinoma, where tissue must be used for both histologic tumor subclassification and molecular analyses.^75–81  Optical coherence tomography could both increase tumor yield by intraprocedural guidance of biopsy-site selection and provide a form of additional, “virtual” tissue for the pathologist to assess as a complement to the tissue biopsy.

In cases of interstitial pneumonitis, patients often must undergo thoracoscopic wedge resections of multiple lung lobes to obtain diagnostic tissue for histopathology. This form of surgical assessment can be problematic because these patients often have severely compromised respiratory systems, and both the surgical procedure and the loss of lung tissue may pose a significant risk for respiratory failure. The ability to visualize the small, subpleural cystic spaces in a background of subpleural fibrosis with OCT is very promising (Figure 6). These features could possibly be used to identify areas of diagnostic tissue for biopsy (to avoid areas consisting predominantly of cystic spaces [honey combing]) and to decrease the amount of tissue needed to make the histologic diagnosis. However, for OCT to be used in these types of settings, the imaging features of a variety of pulmonary diseases would have to be established and validated in large-scale studies.

Optical coherence tomography provides high-resolution, cross-sectional, architectural information that is homologous to low power (×4) microscopy. In clinical scenarios in which there are few diagnostic possibilities with somewhat-straightforward imaging features, such as in coronary and esophageal imaging, image interpretation will likely be performed by the interventionalist at the time of procedure. In coronary imaging, OCT provides a means of tissue assessment where a biopsy is not obtainable for pathology evaluation. When performed and assessed in real time, intracoronary OCT could aid in plaque assessment and patient management. The nondestructive nature of OCT also allows for serial imaging, which could aid in increasing our understanding of the pathophysiology and natural history of the vulnerable plaque.

Regular Barrett's esophagus surveillance protocols for the detection of dysplasia and adenocarcinoma suffer from sampling errors. Volumetric OCT of the entire distal esophagus performed during endoscopy could be used to identify regions most suspicious for high-grade dysplasia and/or carcinoma to guide biopsy-site selection and to improve the performance of endoscopic biopsies. Similar to intracoronary OCT, this form of OCT image assessment would likely be performed by the endoscopist during the procedure.

Optical coherence tomography has numerous, potential clinical applications in pulmonary medicine and may be used during bronchoscopy to guide biopsy-site selection in both the conducting airways and the parenchyma. Similar to coronary and esophageal imaging, both pulmonary nodules and adjacent normal lung have distinct imaging features and could be evaluated by the pulmonologist in real-time during bronchoscopy to guide biopsy-site selection. Additionally, OCT could also be used to identify regions of fibrosis and/or necrosis within nodules, which could then be avoided by the pulmonologist during biopsy acquisition. This would likely result in increased diagnostic yield of both endobronchial and transbronchial biopsies.

However, in contrast to coronary and esophageal pathology, the differential diagnosis of pulmonary nodules can be quite broad, often including reactive, infectious, benign, and/or malignant etiologies. The ability to distinguish these entities requires knowledge of the histomorphologic features characteristic of each entity, which is a skill set inherent to pathologists. Thus, volumetric data sets as a form of “virtual” tissue would be best evaluated by the pathologist interpreting the accompanying physical tissue biopsy. Features seen in OCT imaging are likely related to the histopathologic architecture characteristics of each entity. For example, squamous cell carcinomas frequently display signal-intense nests, which correspond to nests of malignant squamous cell cells, whereas atypical gland formation is seen in many adenocarcinomas. Given the strong parallels between histopathology and OCT imaging, trained pathologists should be ideal interpreters of OCT images. Pulmonary pathologists have been involved in the development of the initial imaging features described in Table 3. However, for OCT to be used as a diagnostic modality complementary to tissue biopsy, OCT imaging criteria of a variety of pulmonary pathologies will have to be established and validated in large-scale studies and will require continued involvement of pathologists.