Clinical applications of optical coherence tomography in urology

Abstract

Since optical coherence tomography (OCT) was first demonstrated in 1991, it has advanced significantly in technical aspects such as imaging speed and resolution, and has been clinically demonstrated in a diverse set of medical and surgical applications, including ophthalmology, cardiology, gastroenterology, dermatology, oncology, among others. This work reviews current clinical applications in urology, particularly in bladder, urether, and kidney. Clinical applications in bladder and urether mainly focus on cancer detection and staging based on tissue morphology, image contrast, and OCT backscattering. The application in kidney includes kidney cancer detection based on OCT backscattering attenuation and non-destructive evaluation of transplant kidney viability or acute tubular necrosis based on both tissue morphology from OCT images and function from Doppler OCT (DOCT) images. OCT holds the promise to positively impact the future clinical practices in urology.

Principle and Instrumentation of OCT

OCT is an emerging medical imaging technology which enables cross-sectional imaging of tissue microstructure in situ and in real-time.¹ OCT can achieve 1–10 µm resolutions and 1–2 mm penetration depths, approaching those of standard excisional biopsy and histopathology, but without the need to remove and process tissue specimens.² OCT is analogous to ultrasound imaging, except that imaging is performed by measuring the echo time delay and intensity of backscattered light rather than sound. OCT imaging can be performed fiber-optically using delivery devices such as hand-held probes, endoscopes, catheters, laparoscopes, and needles which enable non-invasive or minimally-invasive internal body imaging.^3,4

Figure 1A shows a schematic of time-domain OCT. Measurements are performed using a Michelson interferometer with a low coherence length (broadband) light source. One arm of the interferometer illuminates the light on the tissue and collects the backscattered light (typically referred to as “sample arm”). Another arm of the interferometer has a reference path delay which is scanned as a function of time (typically referred to as “reference arm”). Optical interference between the light from the sample and reference arms occurs only when the optical delays match to within the coherence length of the light source.

Figure 1. Schematics of (A) time-domain, (B) spectral / Fourier-domain, and (C) swept-source / Fourier-domain OCT systems. (D) A Clinical OCT system and endoscopic probes from Imalux Corporation. (E) A clinical vascular OCT system and the fiber optic probe from LightLab Imaging (now St. Jude Medical, Inc.).

Alternatively, OCT interference signals can be detected in frequency or Fourier domain. In Fourier-domain OCT, the reference mirror position is fixed, and echoes of light are obtained by Fourier transforming the interference spectrum. These techniques are somewhat analogous to Fourier transform spectroscopy and have a significant sensitivity and speed advantage compared with time-domain OCT because they measure the optical echo signals from different depths along the entire axial scan simultaneously rather than sequentially. Fourier-domain detection enables 10–100 folds improvement in detection sensitivity and speed over the time-domain configuration.^5-7 These advances greatly improve the performance of OCT, enabling three-dimensional OCT (3D-OCT) imaging in vivo.

Fourier-domain OCT can be performed using two complementary techniques, known as spectral/Fourier-domain OCT and swept-source/Fourier-domain OCT (SS-OCT, also known as Optical Frequency Domain Imaging, OFDI). Spectral/Fourier-domain detection uses a spectrometer and a high speed line scan camera to measure the interference spectrum in parallel (see Fig. 1B).^8,9 In contrast, swept-source/Fourier-domain OCT uses a frequency-swept laser light source and a photodetector to measure the interference spectrum (see Fig. 1C).^10-12 Three-dimensional imaging of biological tissue in vivo enabled by Fourier-domain OCT promises to have a powerful impact in disease diagnosis^13,14 and therapy monitoring.^15,16 Up to date, many clinical applications using OCT have demonstrated in a diverse set of medical and surgical specialties. Several commercially-available devices have received US Food and Drug Administration (FDA) clearance to be sold in the market,¹⁷ such as Imalux Corporation (Fig. 1D) whose OCT system is based on time-domain mechanism for endoscopic imaging, and LightLab Imaging (now part of St. Jude Medical, Inc.) (Fig. 1E) that adapts frequency-domain mechanism for their OCT system in cardiovascular imaging.

To image internal organs, miniaturized catheter/endoscope imaging devices have been developed for intraluminal and intravascular imaging.^18,19 Other imaging devices such as laparoscopes^20,21 and needle imaging device have been developed to enable solid organ imaging.^22-27 Nowadays, various OCT imaging probes have been developed for different clinical applications.^23,28-43 Development of such devices facilitates the translation of OCT to clinical applications and allows clinicians to use the enhanced imaging capabilities of this technique to benefit the patients.

Figure 2A shows the schematic of a representative OCT catheter/endoscope device consisting of a hollow cable carrying a single-mode (SM) optical fiber. The beam from the distal end of the fiber is focused by a gradient-index (GRIN) microlens and is directed perpendicular to the catheter axis by a microprism or micromirror. The distal optics is encased in a transparent housing. The beam can be scanned either circumferentially (by rotating the cable) or linearly (by translating the cable) to form a cross-sectional OCT image. The outer diameter of the catheter/endoscope can be made small enough to image inside a human coronary artery (see Figure 2B). Figure 2C shows the schematic of a catheter based OCT (from St. Jude Medical, Inc.) combined with a modified vacuum-pumped biopsy needle.²⁵ This modified core-needle biopsy device includes the addition of a transparent front window for real-time OCT guidance, the addition of a long steel/plastic tube through which the OCT catheter is inserted, and a Y-valve to allow both linear access for the OCT catheter and the vacuum/pressure tube connection. Figure 2D depicts a custom laparoscopic OCT device imaging the ovaries in patients undergoing oophorectomy.²¹

Figure 2. Fiber-optic probes for catheter/endoscope, biopsy needle, and laparoscope. (A) Schematic of the distal end of an OCT probe. (B) Photograph of an intravascular imaging catheter (0.4 mm in diameter). (C) Schematic of a modified core-needle biopsy device with a catheter-based OCT probe (figures are adapted from reference²⁵ with permission) and photographs of modified tip. (D) The photograph of a custom laparoscopic OCT probe for imaging human ovary. Figures are adapted from reference 21 with permission.

Clinical Applications of OCT

Since its invention in 1991, OCT has rapidly developed as a non-invasive biomedical imaging modality that enables cross-sectional visualization of tissue microstructures in vivo.^44-47 The resolution of OCT is one to two orders of magnitude higher than conventional ultrasound, approaching that of histopathology, thereby allowing architectural morphology to be visualized in situ and in real-time. OCT enables imaging of structures in which biopsy would be hazardous or impossible, and promise to reduce the sampling errors associated with excisional biopsy. OCT has been translated from bench to various clinical applications including ophthalmology,⁴⁸ cardiology,^49-52 gastroenterology,^28,53-67 dermatology,^68-70 dentistry,^71-73 urology,^74-77 gynecology,^78-80 among others.⁵² The most developed clinical OCT applications are those focusing on ophthalmic, cardiovascular, and oncologic imaging. For the application in oncology, many cancers arise from the epithelial layers, and demonstrate disruption of normal architectural morphology of tissues. The resolution and imaging field-of-view of OCT is approaching those of standard biopsy and histopathology, therefore OCT represents a potential method for “optical biopsy” of the tissue in situ, which can guide the excision biopsy to improve the sampling accuracy. OCT has shown promises in detecting structural alterations associated with malignancies including those arising in the breast,^81-85 bladder,^77,86-89 brain,^90-92 gastrointestinal,^65,66,93,94 respiratory⁹⁵ and reproductive tracts,^96,97 skin,⁹⁸ larynx,^99,100 and oral cavity.^101,102

Clinical applications of OCT in ophthalmology,^48,52 cardiology,^52,103 and gastroenterology^28,53-67 have been reviewed extensively elsewhere. In this review, we focus on clinical OCT applications in urology, particularly in bladder, ureter, and kidney.

Bladder

Bladder cancer originates in the urothelium and is curable if diagnosed and treated early, but has a high mortality rate in advanced stages.¹⁰⁴ However, early diagnosis of bladder cancer remains a clinical challenge. The other problem is its high recurrence rate resulting in lifelong follow-up and possible repeated treatments, which make bladder cancer one of the most expensive cancers to manage. Currently, white light cystoscopy (WLC) is the standard for initial bladder cancer diagnosis with several shortcomings such as flat carcinoma in situ (CIS) is difficult to visualize. OCT and several other optical imaging techniques (such as fluorescence imaging) have been developed to better identify and characterize bladder lesions beyond what is possible with standard WLC.

Over the last decade, both ex vivo^88,105,106 and in vivo studies^{76,77,89,107-109} have been conducted on the ability of OCT to detect bladder cancer by resolving the changes of bladder wall layers in urothelium, lamina propria, and muscularis propria and/or the corresponding backscattering. A 32 patient study showed that OCT has high detection accuracy for real-time imaging and staging of bladder cancer adjunct to WLC (90% sensitivity and 89% specificity for tumor confined to the mucosa, and 100% sensitivity and 90% specificity for muscle-invasive tumors).⁸⁹ Another clinical study based on OCT imaging with 24 patients reported an overall sensitivity of 100%, specificity of 89%, and diagnostic accuracy of 92% for superficial bladder transitional-cell carcinoma (TCC).⁷⁷ A 56 patient study showed that the overall specificity of cystoscopic OCT (81%) was comparable to voided cytology (88.9%, P = 0.49), but significantly higher than WLC (62.5%, P = 0.02) in TCC diagnosis.¹⁰⁷ Figure 3 illustrates in vivo WLC, OCT, and H&E images of normal human bladder (Fig. 3A–C) and TCC (Fig. 3D–F).¹⁰⁷ TCC exhibited enhanced urothelial heterogeneity as indicated by the arrows shown in Figure 3E. Furthermore, the same work also demonstrated better tumor margin detection using OCT to guide transurethral resection (TUR), which is commonly used for non-muscle-invasive bladder cancer such as TCC that attributes to approximately 75% of all bladder cancer,¹¹⁰ and to enhance re-TUR cases where the scar or necrosis induced by previous TUR may make it difficult to identify residual or recurrent tumors by WLC. Figure 4 shows in vivo WLC, OCT, and H&E images of TCC post-TUR (Fig. 4A–C) and carcinoma in situ (CIS) (Fig. 4D–F).¹⁰⁷ It demonstrated that OCT image can differentiate recurrent TCC from scar or necrosis (Fig. 4B). CIS has low diagnostic sensitivity and specificity (e.g., 30–60%) under routine WLC and remains a critical clinical problem.^111,112 Its OCT image showed characteristics including no obvious urothelial thickening, slightly decreased backscattering in urothelium, and drastically diminished backscattering in lamina propria layer (Fig. 4E). Finally, Zagaynova et al. evaluated 28 cases with OCT during TUR to discriminate between muscle-invasive and non-muscle-invasive tumors with a sensitivity of 100% and specificity of 77%.¹⁰⁸ Table 1 summarizes the performance of OCT in clinical diagnosis of urological diseases.

Figure 3. In vivo surface (A and D), cross-sectional OCT (B and E), and H&E-stained histologic images (C and F) of normal human bladder (A–C) vs a papillary TCC (D–F). The morphologic details of normal bladder (U, urothelium; LP, lamina propria; M, upper muscularis) were clearly delineated by OCT, but those of papillary TCC diminished. Solid arrows: subsurface blood vessels; dashed arrows: papillary features; dashed circle: TCC, identified by OCT based on increased urothelial heterogeneity; dashed line: boundary with adjacent normal bladder. Figures and captions are adapted from reference 107 with permission.

Figure 4. In vivo surface (A and D), cross-sectional OCT (B and E), and H&E-stained histologic images (C and F) of a recurrent TCC post-TUR bladder tumor (A–C) and a CIS (D–F). Yellow and white arrows: papillary features and scar or necrotic lesions. OCT differentiation of TCC (left circle) vs scar was based on low-scattering and papillary features in TCC vs ultrahigh superficial scattering with abruptly diminished underlying architecture in scar or necrotic lesion, which was nonspecific under surface image (A). Arrows in (E and F): blood vessels. The morphology (e.g., lamina propria [LP] and muscularis [M]) under CIS (U*) diminished. CIS was low backscattering and identified by OCT based on increased urothelial heterogeneity and less distinguishable U-LP interface. Figures and captions are adapted from reference 107 with permission.

Table 1. Summary of OCT Patient Studies in diagnosis of urological diseases.

Reference number	Number of Patients	Sensitivity(%), Specificity(%), and others	Determination of diseases by OCT	OCT image features
Bladder
89	32	90, 89, 90(PPV), 89(NPV)a 75, 97, 75(PPV), 97(NPV) 100,90,70(PPV), 100 (NPV)	Tumor confined to the mucosa (Ta); Lamina propria-invasive tumor (T1); Muscle-invasive tumor (T2)	Enhanced urothelial heterogeneity and/or urothelial thickening
77	24	100, 89, 75 (PPV), 100 (NPV), 92 (accuracy) 90 (PPV)	Bladder carcinoma including 16 papillary Transitional-cell carcinoma (TCC) and 5 flat lesions; Tumor invasion	Enhanced urothelial heterogeneity and/or urothelial thickening
107	56	94, 81	CIS and TCC (Ta, T1, and T2)	TCC: enhanced urothelial heterogeneity and/or urothelial thickening; CIS: no obvious urothelial thickening, slightly decreased backscattering in urothelium, and drastically diminished backscattering in lamina propria layer
108	164 28	85, 68 100, 77	Bladder carcinoma; Differentiation between muscle-invasive and non-muscle-invasive TCC during TUR	Enhanced urothelial heterogeneity and/or urothelial thickening for carcinoma
106	142b	83.8, 78.1	Malignant bladder	Enhanced urothelial heterogeneity and/or urothelial thickening for carcinoma
Ureter
134	8	Not available (NA)	Invasive tumor (pT1); Noninvasive tumor (pTa); CIS; Grade 2 vs. grade 3	Interruption of the basement membrane; Basement membrane appeared as a thin dark line below the tumor; Flat, broadened urothelial layer with low reflectivity without interruption of the basement membrane; Grade 3 lesions have higher calculated attenuation coefficient
Kidney
135	16	NA	Malignant tumor	Malignant tumor has higher calculated attenuation coefficient than normal parenchyma
122	20	NA	Angiomyolipoma; Transitional cell carcinomas	Fat elements appear as lucent dark areas easily visible on OCT; Unique and layered papillary architecture
126	29	NA	Ischemic kidney	Close kidney parenchyma tubular, less blood flow from vessels and glomeruli

^a PPV, positive predictive value; NPV, negative predictive value. ^bThe number of fresh human bladder tissue samples.

Computer-aided recognition of bladder cancer using OCT and texture analysis is under investigation to improve the clinical utility of OCT.⁸⁸ Higher OCT axial resolution demonstrated the ability to differentiate healthy urothelial tissue, CIS, and TCC from 142 fresh human bladder tissue samples.¹⁰⁶ The reported sensitivity and specificity to detect malignant bladder are 83.8% and 78.1%, respectively. Recently, real-time 3D-OCT imaging was demonstrated in 3 clinical cases with bladder/ureter carcinoma to show the contrast of muscle-invasive carcinoma area, the scar tissue area from normal bladder wall, and ureter with three distinguishable layers, including the urothelium, lamina propria, and muscularis layer.¹¹³

Similar to other techniques, OCT has some limitations in bladder cancer detection.^110,114 One is false-positives that may be induced by scarring¹⁰⁷ or inflammation of the mucosa.⁸⁹ More clinical studies are needed to confirm the reported results in detecting bladder cancer. The other limitation is the limited field-of-view (FOV) in both lateral and depth directions. OCT was compared with high-resolution ultrasound (i.e., 40 MHz high frequency ultrasound, HFUS) in a rat bladder cancer model.⁸⁷ Results showed that OCT could differentiate inflammatory lesions and TCC based on characterization of urothelial thickening and enhanced backscattering or heterogeneity, which HFUS failed due to insufficient image resolution and contrast. On the other hand, HFUS was able to stage large T2 tumors that OCT failed due to limited imaging depth. Multimodality cystoscopy combining OCT and HFUS, or the combination of OCT with larger lateral FOV technique such as WLC, narrow band imaging, and photodynamic diagnosis may help improve diagnosis and staging.^{87,110,114,115}

Ureter

Few OCT studies have been conducted in ureter, which has somewhat similar mucosal morphology as bladder that the tissue surface is covered with urothelial cells. Early detection of ureteral cancer, as well as accurate tumor staging and grading, is also critical to reduce the mortality of the disease and help making the optimal treatment decisions.¹¹⁶ The staging and grading of urothelial carcinoma in ureter is challenging because the narrow caliber makes biopsy difficult and unreliable. Endoscopic OCT (EOCT) is necessary to access the layer structures of the ureteral wall with sufficient resolution to stage early ureteral cancer. Several ex-vivo studies in porcine ureter have demonstrated to clearly distinguish anatomical layers particularly the urothelium and lamina propria layers^117,118 with better differentiation ability than endoluminal ultrasound.¹¹⁷ Bus et al. reported the intraluminal OCT identification of anatomical layers of the healthy human ureter in vivo and the results for grading and staging upper urinary track (UUT) urothelial carcinoma using OCT.¹¹⁹ They identified several unique features by OCT although this study does not have enough patients to provide information on OCT’s sensitivity and specificity of UUT diagnosis. Their study demonstrated that OCT can: (1) distinguish healthy tissues from tumors; (2) differentiate invasive and non-invasive tumors; (3) differentiate grade 2 and 3 lesions by quantifying OCT backscattering attenuation and, thus, has the potential to provide intraoperative real-time histological information on stage and grade during minimally-invasive procedures. Figure 5 shows representative OCT images of healthy ureter with identified urothelium, lamina propria, and muscularis layers. Figure 6 shows representative OCT images of invasive tumor (namely stage T3G3 urothelial carcinoma) where distinction among anatomical layers was not possible.

Figure 5. OCT of healthy ureter. (A) Individual OCT images obtained from volumetric OCT data set. Inset, higher magnification reveals normal ureter urothelium (pond sign), lamina propria (asterisk), and muscularis (dollar sign). (B) 520-frame volumetric data set across 52 mm trajectory along probe in approximately 5.2 s, resulting in 52 mm long by 10 mm diameter total scanned cylindrical volume. Figures and captions are adapted from reference 119 with permission.

Figure 6. (A and B) Cross-sectional OCT images of proximal ureter show interruption (white asterisk) of thin dark line (white pound sign) suggesting invasive tumor. Distinction among anatomical layers was not possible. Corresponding histology reveled T3G3 urothelial carcinoma (black arrow). (C) 3D pullback of OCT built from 520 individual cross-sectional images over 5.2 cm length. Figures and captions are adapted from reference 119 with permission.

Kidney

OCT studies in clinical kidney diseases include applications in kidney cancer^120-122 and non-destructive evaluation of transplant kidney viability or acute tubular necrosis (ATN).^123-126 Barwari et al. conducted both an ex vivo study with 14 patients and an in vivo study with 16 cases. They demonstrated the capability of OCT to distinguish normal renal parenchyma from malignant renal tumors based on the backscattering properties. Both studies measured higher backscattering property in malignant tumors (measured from the surface or measured directly in the internal tumors) than normal parenchyma. The averaged backscattering value of three benign tumors reported in the in vivo study is between the value from normal and malignant tumor but it did not show significant difference from that of normal renal parenchyma and tumors. Linehan et al. imaged fresh surgical resected tissues of normal renal parenchyma and neoplasm using a laboratory OCT system with lateral resolution of 10 μm and axial resolution of 4 μm.¹²² They found angiomyolipoma and transitional cell carcinomas^127,128 can be distinguished from normal parenchyma. However, higher resolution OCT is necessary to distinguish clear-cell tumors and other renal carcinoma subtypes from normal parenchyma and between carcinoma subtype themselves, which had a heterogeneous appearance on OCT. Figure 7 shows OCT image and corresponding light microscopy of renal carcinoma, chromophobe subtype (top panel) and papillary subtype, grade 4 (bottom panel). Some defining features such as collections of large polygonal cells arranged in trabeculae in chromophobe renal carcinoma and elements of cuboidal cells surrounding a fibrovascular stalk in papillary renal carcinoma were not clearly evident on corresponding OCT images.

Figure 7. OCT image and corresponding light microscopy of renal carcinoma, chromophobe subtype (top panel) and papillary subtype, grade 4 (bottom panel). In the chromophobe subtype (top panel), collections of large polygonal cells arranged in trabeculae are seen as areas of intermediate brightness with intervening dark spaces on OCT. In the papillary subtype (bottom panel), elements of cuboidal cells surrounding a fibrovascular stalk were seen on light microscopy but not visible on the OCT image. Bars are 500 μm. Figures and captions are adapted from reference 122 with permission.

Acute tubular necrosis (ATN) is the most common insult to donor kidneys destined for transplantation.^129,130 ATN is caused by a lack of oxygen to the kidney (ischemia of the kidneys), and is one of the most common causes of kidney failure. Both ex vivo,^123-126 and in vivo¹²⁶ studies demonstrated the capability of OCT to visualize kidney parenchyma morphology and function (i.e., tubular morphology, blood flow from vessels and glomeruli) that provide information to kidney ischemic damage. Figure 8 shows the hand-held OCT imaging device used in the operating room (Fig. 8A–C). Figure 8D depicts representative in vivo kidney OCT images after kidney transplant showing cross-sectional profiles of superficial proximal tubules below the renal capsules. The openness of tubule lumens labeled in Figure 8D reflects a functioning post-transplanted kidney. Figure 8E shows the combination of morphological imaging with OCT and functional imaging with DOCT for one patient that displayed good tubular morphology and blood flow. Fairly densely packed uriniferous tubules are observed with several cortical blood vessels indicating re-perfusion. Finally, Video S1 shows combined OCT and DOCT real-time images of the living kidney following its transplant as would be seen while imaging the kidney in the operation room.

Figure 8. (A) Transplant surgeons used the sterilized hand-held OCT probe shown in (B) to image a transplanted human donor kidney in the operating room. Both surgeons are looking at real-time images of the functioning kidney. The OCT probe and associated wires are covered with a sterile camera sleeve. (C) The OCT imaging probe covered with transparent Tegaderm (small arrow). The cords leading to the probe are covered with a sterile camera sleeve (large arrow). (D) In vivo OCT imaging of human kidneys following transplantation showing open uriniferous tubules below the renal capsule. Tubules appear to be fairly open and round with some degree of homogeneity throughout the images. Scale bar is 500 μm. (E) In vivo human kidney showing open tubules and cortical blood flow. Open tubules appear round and relatively uniform across all images. Also, a larger blood vessel is seen. Scale bar is 500 μm. D and E are from reference 126 with permission.

Summary

OCT is a powerful medical imaging technology that can reveal microstructure and blood flow in biological tissues in a non-invasive fashion and in real-time. Current technology improvements enable 3D-OCT imaging in real-time,^131-133 thereby dramatically reducing the motion artifacts during image acquisition when accurate quantification of OCT/DOCT image is essential for disease diagnosis and decision making. In addition, higher resolution might also help to enhance the classification of imaging parameters for disease diagnosis. With continued technology development and clinical translation, OCT promises to enhance current clinical practice in urology.

Glossary

Abbreviations:

OCT

optical coherence tomography

SS-OCT

swept source/OCT

OFDI

optical frequency domain imaging

DOCT

doppler optical coherence tomography

UHR

ultrahigh-resolution

TCC

transitional-cell carcinoma

CIS

carcinoma in situ

HFUS

high-frequency ultrasound

10.4161/intv.28770

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.