A Novel Method of Virtual Histopathology Using Laser-Scanning Confocal Microscopy In-Vitro with Untreated Fresh Specimens from the Gastrointestinal Mucosa

• Introduction
• Materials and Methods
• Results
• Discussion
• Acknowledgments
• References

Background and Study Aims: Histopathological examination for superficial gastrointestinal lesions has been mainly based upon the light microscopic examination of thin-slice specimens with hematoxylin and eosin (H&E) staining. However, it takes at least a couple of days to create a slide-glass for microscopic study. In order to obtain immediate microscopic images for untreated specimens, the authors used laser-scanning confocal microscopy (LCM) to study fresh samples of gastrointestinal mucosa.

Materials and Methods: Fresh untreated mucosal specimens from the esophagus, stomach, and colon, obtained by endoscopic pinch biopsy, polypectomy, or endoscopic mucosal resection (EMR), were fixed in normal saline and examined by LCM collecting the reflective light of a 488-nm wavelength argon laser beam. Findings from the LCM image were compared with those of conventional H&E staining in all specimens. For objective evaluation of the similarity of both pictures, the nucleus-to-cytoplasm ratio (N/C) of normal mucosa and that of cancer of the esophagus were calculated and statistically analyzed. The overall diagnostic accuracy for cancer was evaluated.

Results: The average scanning time to obtain the LCM image of a specimen was 1.6 seconds. The LCM images acquired corresponded well to the conventional H&E light microscopic images in the esophagus, stomach, and colon. Cell wall, nucleus, cytoplasm, and tissue structural elements were simultaneously visualized by LCM scanning. A difference in N/C ratios between normal mucosa and cancer in the esophagus was statistically apparent when Welch's test (P = 0.05) was applied. The overall diagnostic accuracy of the LCM study for cancer was 89.7 %.

Conclusions: This novel method enables us to obtain an immediate serial virtual microscopic section through a fresh specimen, which has not actually been cut, although the resolution of the image obtained is still limited. These early results encourage us to develop imaging relevant to conventional histopathology alongside the development of LCM technology in the near future. We should aim at the in vivo application of LCM coupled to probes which can be introduced through the working channel of endoscopes.

Introduction

Conventional histopathology for gastrointestinal lesions has been mainly based upon the light microscopic examination of thin-slice specimens under hematoxylin and eosin (H&E) staining. It requires, however, a consecutive process beginning with formalin fixation of the specimen, cutting it into small columns, paraffin embedding, ultra-thin slicing, de-paraffinization, dye staining, slide-glass, mounting, and finally light microscopic observation. It usually takes, in principle, a few days at least to complete the whole process.

If a new technique to obtain an instant virtual section of fresh gastrointestinal mucosa is developed, it will eliminate the above-mentioned time-consuming and technically cumbersome process of conventional histopathology. Therefore the authors endeavored to develop a simple technique to acquire a virtual histopathological image needing no special treatment of the specimen.

Materials and Methods

Fresh gastrointestinal mucosal specimens of the esophagus, stomach, and colon, acquired by endoscopic pinch biopsy, polypectomy, or endoscopic mucosal resection (EMR), were first stretched and fixed on a rubber plate using fine needles and then bathed in normal saline. A total of 21 esophageal samples were taken, consisting of 17 biopsy specimens and four EMR specimens; from the stomach 11 biopsy specimens and one EMR specimen were included. From the colon six biopsy specimens and three polypectomy specimens were taken. All samples from normal mucosa were taken from the normal part of the surgically or endoscopically resected specimen. Histological diagnosis of the samples was as follows: in the esophagus 21 specimens examined were found to consist of 13 samples from normal mucosa and eight samples from squamous cell carcinoma (four biopsy samples and four EMR samples). In the stomach, 12 specimens were found to consist of 11 samples from normal mucosa and one sample of tubular adenocarcinoma. In the colon, nine specimens were found to consist of six samples from normal mucosa and three samples from focal mucosal cancer in an adenoma.

The Fluoview (Olympus Co., Tokyo, Japan) was employed for laser-scanning confocal microscopy (LCM); it scans a specimen with a 488-nm wavelength argon beam and analyzes the reflected light. An objective lens with a magnification of 40 × (Lumpel 40 ×, Olympus) was mounted on the LCM. The distance between virtual slices was fixed at 1 mm for horizontal sections and 0.5 mm for vertical sections. The scanning field was 400 μm × 300 μm. A fresh unfixed specimen was first scanned by LCM and then further studied by conventional light microscopy. The LCM image acquired was compared with the conventional light microscopic image. The light microscopic pictures for comparison with LCM were taken by biolight microscope (BX50, Olympus) with an objective lens (Uplapo 20 ×, Olympus). The final histological diagnosis of a resected specimen was carried out by two pathologists, who were blinded with regard to previous data, using conventional light microscopy under H&E staining. All images acquired were recorded on a digital camera (DP-10, Olympus). During the LCM scan serial horizontal-plane scanning was first carried out, and then a cross-sectional image was created with computer graphics, based on the digitally recorded information from the scanning.

In order to evaluate objectively the correlation between the LCM images and the conventional light microscopic images, the nucleus-to-cytoplasm ratio (N/C) in the esophageal epithelium was calculated for each image acquired using a personal computer (Power Macintosh 8100/80 with Adobe PhotoShop 3.0. software).

Results

All specimens were examined first by LCM and then by light microscopic study. The average time needed to scan each virtual section was 1.6 seconds, and the total time to complete a serial horizontal scan from the surface of a biopsy specimen to the bottom was 36 seconds on average. To acquire one vertical-section image, it took 9 seconds on average. In 93 % of the examined specimens clear LCM images were acquired. In two specimens of normal stomach and one specimen of normal colon no clear pictures were obtained.

In a LCM image of normal esophageal epithelium high-reflectivity of the laser beam was observed almost at the center of each low-reflex cellular compartment. The borders between the cellular compartments were highly reflective. On the other hand, H&E stained normal esophageal mucosa was characterized as follows.

A hematoxylin-stained basophilic nucleus was observed almost at the center of the eosinophilic cytoplasm. Cells were bordered by eosinophilic thin cell walls. These morphological characteristics shown by the two methods were considered to be similar (Figure [1]). In a LCM image of esophageal cancer each cell compartment was noted to be markedly small. A low-reflectivity spot was observed at the center of each high-reflectivity compartment, and the margin of each compartment was an obscure low-reflectivity band. This is a clear difference from the sharp margin of normal mucosa in the LCM image (Figure [2]).

In stomach samples donut-like patterns were observed in the LCM section. Low-reflectivity cellular compartments were attached together by relatively high-reflectivity walls and then created a circle. A round core with no reflectivity occupied the center of the donut. This core of the donut seemed to correspond to a gastric foveola. This pattern was constructively similar to the light microscopic image of H&E-stained normal gastric mucosa (Figure [3]). In colonic samples the foveolar pattern of the epithelium was also clearly demonstrated. As mentioned above, the LCM image obtained by simple observation of fresh gastrointestinal mucosa was observed to have a high morphological correspondence to the conventional light microscopic image under H&E-staining. For cancer-bearing mucosa in the esophagus, images obtained by LCM and by light microscopy also corresponded to each other.

In order to confirm the above-mentioned general impressions objectively, the N/C ratio was calculated on each acquired photograph in esophageal normal and cancerous mucosa, using a personal computer. There were no statistical differences, using Student's t-test (taking a level of P ≤ 0.05 as being significant), between N/C ratios in normal esophageal mucosa between the LCM photographs and the light microscopic photographs. There were no statistical differences using Student's t-test (taking a level of P ≤ 0.05 as being significant) between N/C ratios in cancer-bearing mucosa between the LCM photographs and the light microscopic photographs. On the pictures taken by the LCM the N/C ratio of the normal esophageal mucosa was 0.0523 whereas that of carcinoma was 0.3559. On the light microscopic pictures of the H&E-stained specimens the N/C ratio of normal esophageal mucosa was 0.0500; in contrast, that of carcinoma was 0.3850. Statistical differences between the N/C ratio of normal mucosa and that of cancerous mucosa were clearly demonstrated by Welch's t-test (taking a level of P ≤ 0.05 as being significant) in both LCM photography and light microscopic photography (Table [1]).

Out of 42 specimens no LCM images were obtained in three. The rate of acquisition of LCM images was 92.9 %. For 39 images acquired, the following parameters were calculated: the sensitivity of LCM for detection of cancer was 85.7 %; the specificity of LCM for detection of cancer was 92.0 %; the overall diagnostic accuracy was 89.7 % (Table [2]).

Discussion

In this paper the authors report the first experience of virtual microscopic observation of fresh and untreated gastrointestinal mucosal specimens, including both normal and cancerous tissue, using laser-scanning confocal microscopy (LCM).

Conventional histopathology using light microscopy has been based upon a consecutive management of specimens involving formalin fixation, embedding in paraffin, slicing of the specimen by microtome, de-paraffinization of the slice, dye-staining, and finally slide-glass mounting. The novel diagnostic technology using LCM presented here potentially eliminates all the above-mentioned procedures and offers an instant virtual histopathological picture of an untreated and uncut fresh specimen. Therefore the authors would like to name this virtual histopathological picture, using the reflective light of an argon beam, “virtual histology”.

So far some innovators have applied the LCM to microscopic analysis of body-surface structures such as the skin and the cornea by measuring intrinsic fluorescence or injected extrinsic fluorescence [1] [2] [3] [4] . Masters reported the LCM observation of rabbit and human cornea by measuring intrinsic fluorescence using an oil-immersion lens [4]. The first description of LCM observation of the digestive tract involves using injected fluorescence in the rat colon [5]. Special treatments such as fluorescence injection and the use of the oil-immersion lens are not well adapted to in vivo studies in clinical diagnostic settings. However, these early experiences of the application of LCM to human surface structures encouraged us to obtain virtual histopathological images in the digestive tract. Direct tissue observation using a reflective light has only been reported in the skin [6]. In the present paper the authors describe a novel application of LCM, using the reflective light of a 488-nm wavelength argon beam, to a fresh, untreated specimen of gastrointestinal mucosa, including both normal tissue and cancerous lesions.

In this study the acquired LCM picture was compared with conventional histopathology. The most characteristic features of “virtual histopathology”, which distinguishes it from conventional histopathology, are as follows: direct observation of untreated fresh specimens, instantaneous acquisition of serial sections in arbitrary directions [1], and subsequent examination by conventional light microscopic histology after completion of the LCM analysis.

In order to evaluate the histopathological diagnostic ability of LCM for gastrointestinal cancer, the N/C ratio was calculated as a representative and objective criterion of histological diagnosis. As a result, clear statistical differences were demonstrated.

Light microscopic images are, of course, obtained by actual slicing of the specimen followed by dye staining, but the application of modern technology potentially enables us to acquire images from unstained and uncut fresh specimens. This is the main advantage of using advanced technology. Optical coherent tomography (OCT) is another innovative technology which also potentially allows the virtual imaging of fresh specimens. OCT uses low coherent light to create images. The light penetration of OCT is deeper than the argon beam of LCM, but the resolution in OCT is not as good as that of LCM. OCT imaging is not capable of providing cell-level microscopic images; in fact an OCT image is almost similar to an ultrasonic probe image. The optimal range for tissue penetration in OCT varies from the surface to around 2 mm in depth. The OCT study is, therefore, better applied to more macroscopic and lower-magnification imaging than LCM. In contrast, LCM has much higher resolution than OCT although laser-beam penetration is limited to the epithelium, ranging from the surface to 200 μm. This is why LCM is well adapted to obtaining microscopic high-magnification images. So far, there have been a few reports about the application of OCT to gastrointestinal mucosa. Izatt et al. described the application of OCT to the colon [7], and Tearney et al. also used OCT with fresh specimens from the esophagus and colon [8]. Thus, OCT has actually been applied to the evaluation of normal mucosa, but its resolution is still limited to discriminating cancer from normal mucosa. Our report is the first to describe virtual imaging of a fresh specimen of gastrointestinal tract cancer, which can be discriminated from normal mucosa. The best application of LCM is the observation of the epithelium or superficial microstructure, which seems to equate to a light microscopy image of magnification 500 ×.

The problems of LCM that remain to be solved in the future are as follows. First, the penetration of the argon laser beam with a wavelength of 488 nm is still limited to a maximum infiltration depth of 500 μm, which is appropriate for the evaluation of bite biopsy specimens or changes of the epithelium but is not sufficient to observe the deeper layer of specimens. Secondly, resolution of the LCM is acceptable for the first generation of virtual histology but is still far behind the clear images of conventional light microscopy. A cross-sectional view of the specimen by LCM was also successfully reconstructed by the computer graphics program although the resolution of the acquired image was still limited. In three specimens in this series no clear LCM images were acquired, but at present the authors cannot identify specific reasons for this.

In this preliminary study the overall accuracy of LCM diagnosis was relatively high. One of the major reasons is that the authors selected either definite cancer or definite normal mucosa for this preliminary study. In other words, no inflammatory mucosa or dysplastic epithelium was included. Therefore, if all types of specimen were randomly included in the study, the overall accuracy might become worse.

The first problem would be solved by detecting the new appropriate laser beam which penetrates into deeper layers than the laser beam currently used. Switching between the beams would enable the operator to observe the superficial layer and the deeper layer in turn. The second problem would be solved by the development of a higher resolution sensor and a better arrangement of the devices currently available.

Although the clarity of the confocal microscopic photograph is still limited, these preliminary results encourage us to establish virtual histology of biopsy specimens alongside the mechanical sophistication of LCM in the future. By using this technology, we could obtain virtual histological images in the endoscopy unit soon after biopsy.

Furthermore, this novel technology may expand its in vivo application to diagnostic endoscopy if miniaturized sensor-probe technology is used. Actually Dickensheets and Kino have reported that a miniaturized probe of the LCM can indeed be produced, theoretically and mechanically [9]. If a miniaturized probe which can pass through the instrumental channel of an endoscope could be successfully introduced, an endoscopist could obtain a virtual histological image instantly during endoscopic examination simply by touching the target mucosa with that probe. A pathologist receiving transmitted digital LCM images could readily diagnose the histological character of the lesion. In other words, a miniaturized LCM probe could achieve true virtual biopsy in the near future.

Acknowledgments

The authors thank Hiroyuki Sangu PhD, Hiroki Hibino PhD, and Hitoshi Mizuno PhD from Olympus Co. Tokyo for their engineering contributions to this project.

A summary of this paper was presented at the annual meeting of the Digestive Disease Week which took place in Orlando, Florida, USA between 16 May and 19 May 1999.

References

• 1 Cavanagh H D, Petroll W M, Jester J V. The application of confocal microscopy to the study of living systems.  Neurosci Biobehav Rev. 1993;  17 483-489
  PubMedSearch in Google Scholar
• 2 Corcuff P, Leveque J L. In vivo vision of the human skin with the tandem scanning microscope.  Dermatology. 1993;  186 50-54
  PubMedSearch in Google Scholar
• 3 Rajadhyaksha M, Grossman M, Esterowitz D, et al. In vivo confocal scanning laser microscopy of human skin: melanin provides strong contrast.  J Invest Dermat. 1995;  104 946-952
  PubMedSearch in Google Scholar
• 4 Masters B R. Scanning slit confocal microscopy of the in vivo cornea.  Optical Engineering. 1995;  34 684-692
  PubMedSearch in Google Scholar
• 5 Delaney P M, Harris M R, King R G. Novel microscopy using fibreoptic confocal imaging and its suitability for subsurface blood vessel imaging in vivo.  Clin Exper Pharmac Physiol. 1993;  20 197-198
  PubMedSearch in Google Scholar
• 6 Masters B R, Gonnord G, Corcuff P. Three-dimensional microscopic biopsy of in vivo human skin: a new technique based on a flexible confocal microscope.  J Microscopy. 1997;  185 329-338
  PubMedSearch in Google Scholar
• 7 Izatt J A, Kulkarni M D, Wang H W, et al. Optical coherence tomography and microscopy in gastrointestinal tissues.  IEEE Trans Biomed Eng. 1996;  2 1017-1028
  PubMedSearch in Google Scholar
• 8 Tearney G J, Brezinski M E, Southern J F, et al. Optical biopsy in human gastrointestinal tissue using optical coherence tomography.  Am J Gastroenterol. 1997;  92 1800-1804
  PubMedSearch in Google Scholar
• 9 Dickensheets D L, Kino G S. Micromachined scanning confocal optical microscope.  Opt Lett. 1996;  21 764-767
  PubMedSearch in Google Scholar

H. Inoue, M.D.

First Dept. of Surgery Tokyo Medical and Dental University

1-5-45 Yushima Bunkyo-ku Tokyo 113-8519 Japan

Fax: Fax:+ 81-3-3817-4126

Email: E-mail: hiro.inoue.srg1@med.tmd.ac.jp

References

• 1 Cavanagh H D, Petroll W M, Jester J V. The application of confocal microscopy to the study of living systems.  Neurosci Biobehav Rev. 1993;  17 483-489
  PubMedSearch in Google Scholar
• 2 Corcuff P, Leveque J L. In vivo vision of the human skin with the tandem scanning microscope.  Dermatology. 1993;  186 50-54
  PubMedSearch in Google Scholar
• 3 Rajadhyaksha M, Grossman M, Esterowitz D, et al. In vivo confocal scanning laser microscopy of human skin: melanin provides strong contrast.  J Invest Dermat. 1995;  104 946-952
  PubMedSearch in Google Scholar
• 4 Masters B R. Scanning slit confocal microscopy of the in vivo cornea.  Optical Engineering. 1995;  34 684-692
  PubMedSearch in Google Scholar
• 5 Delaney P M, Harris M R, King R G. Novel microscopy using fibreoptic confocal imaging and its suitability for subsurface blood vessel imaging in vivo.  Clin Exper Pharmac Physiol. 1993;  20 197-198
  PubMedSearch in Google Scholar
• 6 Masters B R, Gonnord G, Corcuff P. Three-dimensional microscopic biopsy of in vivo human skin: a new technique based on a flexible confocal microscope.  J Microscopy. 1997;  185 329-338
  PubMedSearch in Google Scholar
• 7 Izatt J A, Kulkarni M D, Wang H W, et al. Optical coherence tomography and microscopy in gastrointestinal tissues.  IEEE Trans Biomed Eng. 1996;  2 1017-1028
  PubMedSearch in Google Scholar
• 8 Tearney G J, Brezinski M E, Southern J F, et al. Optical biopsy in human gastrointestinal tissue using optical coherence tomography.  Am J Gastroenterol. 1997;  92 1800-1804
  PubMedSearch in Google Scholar
• 9 Dickensheets D L, Kino G S. Micromachined scanning confocal optical microscope.  Opt Lett. 1996;  21 764-767
  PubMedSearch in Google Scholar

H. Inoue, M.D.

First Dept. of Surgery Tokyo Medical and Dental University

1-5-45 Yushima Bunkyo-ku Tokyo 113-8519 Japan

Fax: Fax:+ 81-3-3817-4126

Email: E-mail: hiro.inoue.srg1@med.tmd.ac.jp