First results of endoscopic applications of optical coherence tomography for in vivo studies of human mucosa in respiratory, gastrointestinal, urinary and genital tracts are presented. A novel endoscopic OCT (EOCT) system has been created that is based on the integration of a sampling arm of an all-optical-fiber interferometer into standard endoscopic devices using their biopsy channel to transmit low-coherence radiation to investigated tissue. We have studied mucous membranes of esophagus, larynx, stomach, urinary bladder, uterine cervix and body as typical localization for carcinomatous processes. Images of tumor tissues versus healthy tissues have been recorded and analyzed. Violations of well-defined stratified healthy mucosa structure in cancered tissue are distinctly seen by EOCT, thus making this technique promising for early diagnosis of tumors and precise guiding of excisional biopsy.
© 1997 Optical Society of America
Optical coherence tomography as a novel method of biotissue imaging extends our capabilities in structural analysis of human organs and systems. It is now well known of clinical applications of OCT in ophthalmology and dermatology.[1 – 4] Here we report the first results on endoscopic OCT for in vivo studies of human mucosa.
Monitoring of epithelial and subepithelial layers of internal organs in the near infrared spectral region to the depth of 1.5-2 mm is one of the challenging problems targeted by modern methods of bioimaging since its solution promotes noninvasive early diagnosis for a wide range of inflammatory and tumor processes. It has already been confirmed in a number of in vitro OCT studies of human mucosa in GI tract, lungs, urinary bladder, etc. In our early paper [5,6] we have demonstrated first OCT images of cancered human intestine in vitro. Similar tomograms have been obtained later by others groups.[7,8] Recently, at the CLEO'97 conference, first results on in vivo EOCT have been presented by our group for human studies in gynecology  and by the group from MIT for experiments with animals. In this paper we present the description of our EOCT system and results of its clinical applications in gastroenterology, urology, otorhinolaryngology and gynecology.
2. Implementation of endoscopic OCT device
The main problem to be solved for endoscopic implementation of OCT is to provide a reliable and convenient access of probing low-coherence radiation to the surface of internal organs. This problem includes several optical, engineering and biomedical aspects such as creation of an OCT interferometer with a flexible arm, development of a miniaturized lateral scanning probe with remote control, acquisition of OCT data in parallel with common endoscopic imaging. The idea we put in the basis of our EOCT system was integration of a sampling arm of an all-optical-fiber interferometer into standard endoscope using their biopsy channel to transmit low-coherence radiation to investigated tissue. It has resulted in creation of a whole family of diagnostic devices suitable for studying different internal organs.
To probe the surface of an internal organ we have developed a miniaturized electromechanical unit controlling the lateral scanning process. This unit terminates the sampling arm of the fiber interferometer and has the size to fit the diameter and the curvature radius of the biopsy channel of endoscopes. In Fig. 1a the head of an endoscope for GI investigations with the integrated OCT scanner is demonstrated. Fig. 1b is a schematic diagram presenting the geometry of the scanning unit and its positioning against a studied tissue. The probe beam swinging along the tissue surface approaches 2 mm. The beam deviation system embodies the galvanometric principle, and the voltage with the maximum of 5V is supplied to the distal end of the endoscope. The distance between the output lens and a sample is 5 - 7 mm, the focal spot diameter is 20 μm. The scanning unit and the extended part of the fiber interferometer inserted in the endoscope are sealed which allows common cleaning procedure of the device in clinical applications.
Implementation of an extended flexible arm of the OCT interferometer to provide the access to the studied tissue is based on the usage of polarization maintaining fibers for transportation of the low-coherence light. By this we eliminate polarization fadings connected with polarization distortions at bending of the endoscope arm. Using high-quality fiber polarizers and couplers, the “single-frame” dynamic range of our OCT scheme, determined as the maximum variation of the reflected signal power within a single image frame, is 35-40 dB, whereas the “traditional” dynamic range (incident power divided by the minimal detected backscattered signal) is 90 dB.
As a source of radiation we employ a superluminescent diode with the central wavelength 830 nm, the bandwidth 30 nm, and the power 1.5 mW. The in-depth scanning is performed with an integrated piezo-optical modulator  that introduces the difference in the length of the interferometer arms. With the scanning rate of 30 cm/s and the image depth of 3 mm (in free space units), an OCT picture of 200x200 pixel size is acquired for approximately 1 s. This rate is sufficient to eliminate the effect of internal organs movement on the image quality.
The combination of the OCT system with the standard endoscopic equipment has proved to be quite convenient for clinical studies allowing doctors to perform usual observation of internal organs and in the case of interest to make simultaneously the noninvasive optical biopsy for any amount of tissue locations.
3. Materials and methods
In the present-day medical practice, the final cancer or precancer diagnosis is based on the result of biopsy. The location of a biopsy specimen is relatively random and sometimes erroneous which leads to ambiguous conclusions. Optical diagnostic techniques can help to guide biopsy for superficial tissues, i.e. skin and mucous membranes that are mostly suffered of carcinomatous degeneration because of intensive proliferative processes. According to the World Health Organization experts classifying carcinomas in TNM system , the critical localization depth of the early cancer stages with favorable prognosis after operative treatment is 600-1000 μm in the mucous membrane of different organs. Thus, the depth of OCT imaging of mucous membranes is sufficient for detection of early stages of most widely spread carcinomas of hollow organs.
Our study has been focused on mucous membranes of hollow organs contacting with the outer environment. We have chosen mucous membranes of upper and middle portions of gastrointestinal tract (esophagus and stomach), upper respiratory tract (larynx), urinary tract (bladder and urethra) as well as mucosa of uterine as typical localization for carcinomatous processes. EOCT studies in vivo have been conducted during routine fibrogastroscopic examinations of four patients (three of them had exophytic tumor formations in stomach and esophagus, suspicious to carcinoma). EOCT study of urinary tract has been performed in three patients during the endoscopic surgery of tumors of urinary bladder and prostate gland. The reason for the hysteroscopic examination of the uterine cavity by OCT in three patients has been tumor-like processes in endo- and myometrium. In eight female patients with the uterine cervix cancer, previously diagnosed using biopsy, in vivo imaging of the diseased area has been performed. All of them have been treated surgically afterwards.
In all endoscopic and surgical procedures, the biopsy material of suspicious formations has been taken to produce histological preparations. Then a morphometry have been performed and compared further with corresponding tomograms. In addition to the pathological regions, healthy areas have been imaged by OCT in all patients. The tomograms have then been identified with histological images of the corresponding specimens. Unfortunately, the procedure of taking the biopsy material after OCT recording, that has been available in clinics, does not ensure yet detailed quantitative comparison between the histology and optical images. Therefore, the search of correlation and the identification of images have been performed by the analysis of tissue layers (their thickness, form and sequence) and large-scale features (glands and large blood vessels, their characteristic sizes and mutual arrangement).
4. In vivo OCT imaging of healthy human mucosa
It is well known that mucous membranes of different organs with different functions possess many common features. The necessary components of mucous membranes are epithelium that has a protective function (stratified squamous and transitional epithelium where a traumatization risk is sufficient, or simple columnar epithelium where a secretory function is dominant) and lamina propria that is a crumbly connective tissue containing specific glands, blood vessels and nervous texture. These two different components are delimited by the basal membrane. The layered structure of mucosa promotes the contrast of tomographic imaging of internal organs.
Among the organs, we have studied by OCT, the stratified squamous and transitional epithelium is present in mucosa of larynx, esophagus, uterine cervix and urinary bladder. By their location and functioning in the corresponding systems, these organs are in the neighbourhood of regions contacting with the outer environment, which requires a high protecting potential provided by this type of epithelium. Fig. 2 shows in vivo images of healthy mucosa of larynx (a), esophagus (b), uterine cervix (c) and urinary bladder (d). Owing to the different backscattering intensity we can identify different mucosa layers, so that the stratified squamous (or transitional) epithelium (E) is seen as a weakly scattering layer, while lamina propria (LP) - as a bright (mostly scattering) one. A peculiarity of the esophagus structure, associated with the food movement, consists in the presence of a muscular plate, so called - muscularis mucosa (MM), that is observed in the tomogram as a minimally backscattering layer, i.e. as a dark stripe. Besides, one can differentiate deeper layers that are submucosa (SM) (more backscattering) and muscular layers (ML) (less backscattering) in esophagus and urinary bladder. In lamina propria and submucosa one can find blood vessels (BV) and glands (e.g. pyloric glands (PG) in stomach), visible as circular or elongated formations.
The sizes of layers and structures in optical images correspond to the values, determined in the histological study. The optically imaged depth for the stratified nonkeratinized (or transitional) epithelium varies within 30-300 μm in different organs, for lamina propria, within 100-150 μm, for muscularis mucosa, 80-100 μm. The internal diameter of blood vessels is 20-40 μm, while the diameter of different types of glands is 30-300 μm. All these data are in agreement with the values, obtained by morphometry.
Contrary to the above cases, mucosa of several internal organs or of their parts is covered by the simple columnar epithelium that is much thinner as compared to the squamous epithelium. Therefore the capability of our OCT system may be enough to locate but not to resolve and measure the epithelial layer. Fig. 3a shows an OCT picture of normal mucosa of the pyloroantral portions of stomach in vivo. The mucous membrane structure of this stomach zone comprises the simple columnar epithelium and is peculiar for the presence of papillaes in the broad lamina propria, forming foveae gastrici with 300-400 μm thickness. Lamina propria contains pyloric glands with wide lumen and blood and lymphatic vessels.
All these features can be observed in OCT images. A surface of the simple columnar epithelium (whose thickness is of the order of spatial in-depth resolution) appears as a bright broken border framing foveae gastrici and, thus, going as deep as 300-400 μm. At the 100-300 μm depth some pyloric glands can be seen as weakly scattering cavities. Muscularis mucosa is presented as a weakly scattering strip with 100-120 μm thickness at the depth of 500-600 μm.
Summarizing the EOCT observation of normal mucosa, we can state that the obtained images of stomach, larynx, uterine cervix and urinary bladder are characterized by the regular stratification over all specimens, which, due to different scattering properties of different tissue components, is well resolved in our tomograms.
5. In vivo OCT imaging of cancered human mucosa
Irrespectively of its localization in the organ, a carcinomatous process is more often represented in histology by continuous fields of atypical epithelioid cells. It destroys the regular structure of mucous membranes and does not permit to differentiate the classical sequence of its layers. This is clearly demonstrated in Fig 4 showing tomograms of mucosa of larynx (a), esophagus (b), uterine cervix (c) and urinary bladder (d) with tumors. A corresponding tomogram of a tumor in stomach is presented in Fig. 3b. The tumor tissue in vivo is characterized by a higher homogeneity and typically higher than in epithelium backscattering. A higher vascularization degree is also typically present in tomograms of cancered tissues.
6. In vivo OCT imaging of pathological states in female genital tract
As it is known, the survival prognosis for patients with different types of cancer is directly connected with the invasion depth of the tumor in the tissue. In gynecology the “threshold” of invasion can in many cases be considered as the attainability of the basal membrane by the tumor. The results we have obtained in monitoring of human mucosa allow us to discuss the potentialities of EOCT in diagnostics of precancer states and preinvasive forms of cancer in the female genital tract. We will first illustrate it by an example of a concrete clinical case.
A female patient of 25 years old has been examined by the usual colposcopy and by EOCT. We have clearly observed structural changes of the stratified squamous epithelium in vivo with our tomographic device and have found different states of the basal membrane. In Fig.5 an image of the uterine cervix is presented that is characterized by destruction of the basal membrane seen as breaks in the bright stripe and its disappearance. Further, the uterine cervix has been excised and the final histological diagnosis has been formulated as CIN III with transition to cancer in situ and the presence of areas of microcarcinoma. Based on the above conclusion, we believe that in this case EOCT has been capable to catch the border of invasion (i.e. of destruction of the basal membrane), which is a challenging problem for any method of early diagnostics.
The OCT data characterizing endometrium are of a special interest. In vivo tomography of endometrium in parallel with hysteroscopy permits not only to estimate a functional state of the endometrium but it may also be useful in differential diagnosis of intrauterine formations.
The endometrium state is known to depend on the age and a phase of the menstrual cycle. In the menopause an atrophied endometrium is typical that is characterized by a small number of endometrial glands whose lumen is not large (usually, of the order of 50, maximum 100 μm). The compact layer of endometrium is practically absent, the vascularization degree is low, and stroma of the lamina propria is sclerosed. Thus the stroma is shown in the tomogram as an intensively scattering medium (Fig. 6).
At the stage of secretion (14-28 day of the menstrual cycle) endometrium is magnificent and substantially thickened. Its compact (CL) and spongy (SL) layers are clearly distinguished by the presence of endometrial glands (EG). The compact layer is about 500 μm thick and is well vascularized. The spongy layer contains a large number of broad secretion-filled glands, whose diameter can reach 250-500 μm. This clearly correlates with OCT images (see Fig. 7).
In Fig. 8 we demonstrate the case when, being performed during the standard hysteroscopy, EOCT allowed to provide a differential diagnosis between submucosal myoma and glandular fibrous polyp of endometrium with cyst expansion of many glands. It became possible due to peculiarities of the polyp structure. It is characterized by a sclerosed (well backscattering) stroma and by the presence of cyst-expanded glands (CEG) seen as weakly scattering cavities with the diameter up to 600-900 μm. Other histological correlations can also be found in analyzing the image in Fig. 8.
In conclusion, we have demonstrated first, to our knowledge, applications of OCT for in vivo studies of human mucous membranes in respiratory, gastrointestinal, urinary and genital tracts. It became possible due to the construction of a novel endoscopic OCT system based on our previously developed compact all-optical-fiber tomograph technology and the idea of the integration of a sampling arm of the OCT interferometer into standard endoscopic devices. We have acquired a number of images of mucous membranes in larynx, esophagus, stomach, urinary bladder, uterine cervix and body. Images of tumor tissues versus healthy tissues have been recorded and analyzed. We have demonstrated that violations of well-defined stratified healthy mucosa structure in cancered tissue is distinctly seen by EOCT, thus making this technique promising for early diagnosis of tumors and precise guiding of excisional biopsy.
The authors thank the Russian Basic Research Foundation for the financial support of this work. The authors are also grateful to Prof. Ya.I.Khanin for fruitful discussions and support and to the staff and management of Regional Hospital for the possibility to conduct this research study.
References and links
1. D. Huang, J. Wang, C. P. Lin, J. S. Shuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, “Optical coherence tomography,” Science 254, 1178 (1991). [CrossRef] [PubMed]
2. A. F. Fercher, C. K. Hitzenberger, W. Drexler, G. Kamp, and H. Sattmann, “In vivo optical coherence tomography,” Am. J. Ophthalmol. 116, No. 1, 113–114 (1993). [PubMed]
3. E. A. Swanson, J. A. Izatt, M. R. Hee, D. Huang, C. P. Lin, J. Shutz, C. A. Puliafito, and J. G. Fujimoto, “In vivo retinal imaging by optical coherence tomography,” Opt. Lett. 18, 1864 (1993). [CrossRef] [PubMed]
4. A. Sergeev, V. Gelikonov, G. Gelikonov, F. Feldchtein, K. Pravdenko, D. Shabanov, N. Gladkova, V. Pochinko, V. Zhegalov, G. Dmitriev, I. Vazina, G. Petrova, and N. Nikulin, “In Vivo Optical Coherence Tomography Of Human Skin Microstructure,” in: “Biomedical optoelectronic devices and systems II,” Proc. SPIE 2328, 144 (1994). [CrossRef]
5. V. M. Gelikonov, G. V. Gelikonov, R. V. Kuranov, K. I. Pravdenko, A. M. Sergeev, F. I. Feldchtein, Ya. I. Khanin, D. V. Shabanov, N. D. Gladkova, N. K. Nikulin, G. A. Petrova, and V. V. Pochinko, “Coherent optical tomography of microscopic inhomogeneties in biological tissues,” JETP Lett. 61, 149 (1995).
6. A. Sergeev, V. Gelikonov, G. Gelikonov, F. Feldchtein, K. Pravdenko, R. Kuranov, N. Gladkova, V. Pochinko, G. Petrova, and N. Nikulin, “High-spatial-resolution optical coherence tomography of human skin and mucous membranes,” in Digest of Conference on Lasers and Electro-Optics (Optical Society of America, Washington, D.C., 1995) p.349.
7. J. G. Fujimoto, M. E. Brezinski, G. J. Tearney, S. A. Boppart, B. E. Bouma, M. R. Hee, J. F. Southern, and E. A. Swanson, “Optical biopsy and imaging using optical coherence tomography,” Nat.Med. 1, 970 (1995). [CrossRef] [PubMed]
8. J. A. Izatt, H.-W. Wang, M. Kulkarni, K. Kobayashi, M. I. Canto, and M. V. Sivak, “Optical coherence tomography and microscopy in gastrointestinal tissues,” Trends in Optics & Photonics 2, 203 (1996).
9. G. Gelikonov, V. Gelikonov, F. Feldchtein, J. Stepanov, A. Sergeev, I. Antoniou, J. Ioannovich, D. Reitze, and W. Dawson, “Two-color-in-one-interferometer OCT system for bioimaging,” in Digest of Conference on Lasers and Electro-Optics (Optical Society of America, Washington, D.C., 1997) p.210.
10. G. Tearney, S. A. Boppart, B. E. Bouma, C. Pitris, M. E. Brezinski, J. F. Southern, E. A. Swanson, and J. G. Fujimoto, “High-speed catheter/endoscopy optical coherence tomography for the optical biopsy of in vivo tissues,” in Digest of Conference on Lasers and Electro-Optics (Optical Society of America, Washington, D.C., 1997) p.211.
11. F. I. Feldchtein, G. V. Gelikonov, V. M. Gelikonov, N. D. Gladkova, V. I. Leonov, and A. M. Sergeev, “Optical fiber interferometer and piezoelectric modulator,” International application PCT/RU96/00045 (1995).
12. American Joint Committee on Cancer: Manual for Staging of Cancer, 3rd edition (JB Lippincott Company, Philadelphia, USA1988).