| BACKGROUND:PARTI:Endoscopy is the most effective method for early detection of digestive tract diseases. Advanced endoscopic devices such as magnifying endoscopy, pigment endoscopy and narrow band imaging have made rapid progress in the diagnosis and detection of early carcinoma of digestive tract. The combination of magnifying endoscopy and staining endoscopy can improve the accuracy of biopsy through the subtle changes in the cell structure of the mucosal surface. However, the various endoscopic techniques are limited to the observation of the morphological changes of the mucosal surface of the digestive tract, and ultimately, the pathological biopsy of the lesion is still needed to be confirmed. Probe-based confocal laser endomicroscopy (pCLE), in recent years a new application in clinical endoscopic techniques, within conventional white light endoscopy, provide real-time digestive tract mucosa surface and submucosal 250μm transect of organization high resolution images. Confocal endomicroscopy is the tiny probe of cofocal laser scanning microscope integration in the head of the conventional endoscopy and based on observing sites in the digestive tract mucosa of single layer depth scanning, to display magnified 1000 times of high resolution histological and cytological image, to achieve the "optical biopsy" and instant tissue diagnosis, to improve the diagnosis accuracy of gastrointestinal mucosa lesions and reduce the number of blind biopsy.Most bulge and depressed type of the digestive tract mucosa lesions can be found by the conventional white light endoscopy, many flat lesions in the general endoscopic performance is more occult, easily be misdiagnosis or missed diagnosis. pCLE can be performed in the digestive tract mucosa in real time in the body of the histopathological diagnosis, and greatly improve the digestive endoscopic screening and the examination rate of the various gastrointestinal diseases. pCLE, a new type of endoscopic techniques in microstructure of real-time observation of digestive tract gland, mucosal epithelial cells and blood vessels, the implementation of target biopsy and obtain similar histopathological endoscope image, use tissue fluorescence imaging, which requires the organization in imaging with strong fluorescence groups. So we need observe in white light endoscopy and mucosal surface spraying or intravenous injection of fluorescent substance as a contrast agent. Due to its characters of easy to use, superior fluorescence characteristics, no malformation, cheaper prices, rare complication, fluorescein sodium is the most commonly used developer of the confocal endoscope and in eye surgery have been used for decades.Intravenous application of fluorescein sodium in the traditional dose is 10% fluorescein sodium 5-10ml. The dose can be guaranteed to get a high quality of clear images, but the patient’s skin and urine yellow dye is very obvious. The half life of fluorescein sodium is 24 hours, although the sodium fluorescein can be gradually metabolism by enough drinking water and yellow dye gradually subsided, the performance is still seriously affected the patient’s daily life and work in beauty. We in the pre-experiment found less than traditional dose of sodium fluorescein can get clear confocal endoscope image, such as 10% fluorescein sodium 2ml, lml even 0.5ml can sometimes clearly display organizational structure.PART Ⅱ:Gastric cancer is still the world’s second mortality rate of malignant tumors, the mortality rate for men 16.3/10 million,7.9/10 million women. China is a country with high incidence of gastric cancer. Although the incidence of gastric cancer has decreased slightly in recent years, it is still one of the most common malignant tumors in the digestive tract. The occurrence of gastric cancer is a multi-factor and multi-step process, which is generally considered to be related to environmental factors, Helicobacter pylori (Hp) infection and genetic factors. Gastric cancer development including the following processes: chronic superficial gastritis, atrophy gastritis, gastric mucosa with intestinal metaplasia (GIM), atypical hyperplasia, high grade intraepithelial neoplasia (HGIN), and ultimately the development of gastric cancer. This process takes about a few decades, in which atrophic gastritis, intestinal metaplasia and atypical hyperplasia are known as precancerous lesions.GIM is the gastric mucosal epithelium and (or) glandular epithelium instead by the intestinal mucosal epithelium and (or) gland epithelium in pathological conditions. GIM is a high risk factor for intestinal type gastric cancer, and it is considered to be the precancerous lesion of gastric cancer. The degree of intestinal metaplasia and gastric mucosa atrophy is parallel to the dysplasia frequency, which is widely regarded as a precancerous condition. It can be an effective way to prevent and treat gastric cancer if it can be identified and treated early in the endoscope to reverse the pathological changes.The clinical value of pCLE for the diagnosis of gastric cancer and intestinal metaplasia has been confirmed. Metaplasia epithelia in morphology is the absorptive cells, goblet cells, Pan’s cells, villous gastric foveolar epithelium or through mucus staining to identify, and normal conditions the cells and subcellular structures can be observed only under a microscope by tissue biopsy, but not under endoscope. With 1000 times height, pCLE can clearly observe the cellular and subcellular nuclear structures. Because the pathological sections of tissue lesions were observed by a series of process such as fixed, dehydrated, it has obviously different with that in vivo; therefore, pCLE has great advantages. Previous studies have shown that, according to the established diagnostic criteria, pCLE can be more accurate in the identification of gastric mucosal lesions in the presence of cancer. However, this study should be used to obtain the non-specific fluorescence imaging of gastric mucosa by intravenous injection of fluorescein sodium as a contrast agent.In part Ⅰ, we exclude the interference of various serious mucosa lesions and all patients are in upper gastrointestinal endoscopy. Gastrointestinal organs due to the influence of sodium fluorescein distribution and release of epithelial cell types, blood supply, mucosa thickness and permeability difference, the best sodium fluorescein dose also may be different. Is the optimal dose also applicable for gastric intestinal metaplasia, digestive ulcer, gastrointestinal cancer, severe gastritis and gastrointestinal tract in other parts of the body such as the colon and rectal mucosa?AIMS:(1) To optimize an appropriate dosage of fluorescein sodium which can provide a high-quality image to ensure a satisfactory endoscopy examination, while avoiding yellow discoloration of the skin, through a randomized blind controlled trial.(2) pCLE is an efficient and potentially important method for the diagnosis of gastric intestinal metaplasia (GIM) in vivo. To compare the diagnostic accuracy of GIM between using the optimized dosage and the conventional dosage of fluorescein sodium.METHODS:PART I:Patients those satisfied inclusion criteria who underwent upper gastrointestinal endoscopy from August to September 2015 were enrolled in this study. All participants were intravenously given 1ml of 1% fluorescein sodium for allergy test 20 minutes before the endoscopy. Patients with allergic reactions were excluded. The candidate dosages of fluorescein sodium were the ordinary O.lOml/kg and 3 lower dosages:0.05ml/kg,0.02ml/kg, and 0.01 ml/kg respectively. Once gastric antrum was clearly visible, the assistant promptly administered fluorescein sodium intravenously in accordance with the patient’s randomized group that was blind to the endoscopist. The two endoscopists with extensive experiences were responsible for the procedures. Both endoscopists were blinded to the groups of patients. The greater curvature of gastric antrum in a range of 3-4cm to pylorus was selected as the best observation spot for its good stability and convenience for touching the gastric mucosa with pCLE. After injecting fluorescein sodium, we recorded real-time videos for lminute, 5minutes, 10minutes,15minutes, and 20minutes.The pCLE videos of each patient were stored in a specific folder for later analysis. Each picture was rated for the quality according to the visualization of cellular and vessel structures, and epithelial cells by another two endoscopists experienced in pCLE. Before subjective assessment, the image sequence were randomly scrambled, irrespective of the groups of patients. The endoscopists were blinded to the dosage of fluorescein sodium and each other’s assessment. The pCLE images were classified into three grades:poor (almost invisible tissue and cellular structure); mild (visible tissue and cellular structure but with too dark vision); and good (very clearly visible tissue and cellular structure). The image was scored one, two, or three when at least two-thirds of it was classified as poor, mild, or good, respectively. To obtain the interobserver agreement of the four groups, another researcher who was also blinded to the patients’ groups and corresponding fluorescein sodium dosages reassessed all the images of each patient.The objective evaluation of the pCLE video quality was performed through calculating the signal-to-noise ratio (SNR) as a standard objective contrast parameter. SNR is an electrical engineering term, which is defined as the ratio of signal powers to the noise powers that corrupt the signal of a picture. A ratio higher than 1 indicates more signal than noise. The higher the ratio, the less corruption from the noise. We used the Matlab2013b software (Matlab, The Math Works, Inc. Natick, Massachusetts, USA) to calculate the SNR value directly. Together, this method enables effective assessment of the video and images even at a low quality. The highest value also means the best image quality.A transient yellowing of the skin and urine is the common adverse events of fluorescein sodium as mentioned above, although fluorescein sodium can be metabolized thoroughly after enough water intake for about 24-48 hours. Therefore, the jaundice values of the patients in the forehead, corner of the eye, and chest were measured by a neonatal transcutaneous jaundice detector (NJ33-wuhan, China) before and one hour after the endoscopy, respectively.PART II:From September to November 2015, consecutive patients were recruited into the study according to the inclusion and exclusion criteria. Enrolled patients were randomly given the conventional (0.10ml/kg, group A) or the lower (0.02ml/kg, group B) dosage of fluorescein sodium.The macroscopic lesions were found by endoscopy through which pCLE inserted. Targeted biopsies were only taken at the scanned sites where pCLE images revealed GIM. In addition, all macroscopic lesions were biopsied before the withdrawal of the endoscopy. The biopsy specimens were immediately immersed in formalin and then embedded in paraffin. Serial sections (4μm) were stained with hematoxylin and eosin for histopathological examination. All biopsy specimens were examined by an experienced pathologist who was blinded to the pCLE results. To assess the diagnostic accuracy, GIM was diagnosed when at least one single goblet cell was seen in sections of the biopsy specimens. Patients were diagnosed as GIM positive if GIM was present in at least one of the biopsy specimens.RESULTS:PART Ⅰ:A total of 63 patients met the inclusion criteria and were included in this study. There were no statistical differences in ages (P=0.449) and gender among four groups (P=0.635) A total of 1,575 pCLE images were acquired from 63 patients. For subjective assessment, the image quality significantly degraded with the reduced fluorescein sodium dosage (P<0.0001). Two independent samples tested between the adjacent dosages revealed that the dosage of group A (0.01 ml/kg) was the only one that resulted in a significantly decreased image quality compared with the other groups (group A vs. group B, P<0.0001). Agreement of four groups was substantial (P=0.465).In objective assessment, SNR appeared to be in the same distribution as the subjective assessment. SNR significantly decreased with the reduced fluorescein sodium dosage (P<0.05). The two independent samples tested between the adjacent dosages revealed that there was no significant difference between group C (0.05ml/kg) and group D (0.10ml/kg). Meanwhile, the dosage of group A (0.01 ml/kg) was the only one in which decreased SNR significantly compared to the adjacent higher dosage (group A vs. group B, P<0.0001; group B vs. group C, P=0.009; group C vs. group D, P=0.895). In other words, intravenously administered fluorescein sodium with a dose of 0.02ml/kg could obtain the clearest images. Skin discoloration value assessment revealed the same trend among the four fluorescein sodium dosages as the subjective assessment, indicating that no significant jaundice difference existed between groups A and B (P=0.148). We set the minimal influential units for obvious discoloration at 2. A t-test analysis showed that groups A and B had no significant obvious discoloration (average levels<2, P<0.05).PART Ⅱ:A total of 410 consecutive patients were screened and 249 patients (129 in group A 0.10ml/kg and 120 in group B 0.02ml/kg) were finally enrolled on the basis of the inclusion and exclusion criteria. There were no significant differences between groups A and B regarding patients’ ages and genders. Final histopathological findings were as follows:120 patients had GIM (group A,63 and group B,57). For a per-patient analysis, the diagnostic yields of GIM for groups A and B were 48.84% (63/129) and 47.5% (57/120), respectively (P>0.05). Regarding macroscopic lesions, a total of 318 were detected. In group A,152 macroscopic lesions were detected by pCLE and a total of 175 targeted biopsies were obtained. Among those areas, GIM was histologically diagnosed in 109 lesions in 63 patients. In group B, a total of 184 biopsies were obtained from 166 macroscopic lesions. Among these areas, GIM was detected in 117 macroscopic lesions in 57 patients. Therefore, for a per-macroscopic lesion analysis, pCLE-targeted biopsies obtained a higher diagnostic yield of GIM for both groups A (71.71%,109/152) and B (70.48%,117/166). The diagnostic accuracy of groups A and B were 96%(168/175) and 95.11%(175/184), respectively. The sensitivity, specificity, positive and negative predictive value of the final pCLE diagnosis of GIM for groups A and B were 96.40% and 95.73%,95.31% and 94.03%, 97.27% and 96.55%,92.42% and 92.65%, respectively. There were no statistical differences between groups A and B on pCLE diagnostic accuracy for GIM (P>0.05). The kappa value for the correlation between pCLE and histopathology for groups A and B was 0.867 (95%CI 0.782-0.952) and 0.891 (95%CI 0.811-0.971), respectively.CONCLUSIONS:PART I:1. Fluorescein-based pCLE of the upper gastrointestinal tract shows the best results using single-injection 10% fluorescein sodium 0.02ml/kg. This is the optimal dosage based on objective and subjective evaluations, and the result of no yellow discoloration of the skin.PART II:1. It is very efficient for the diagnosis of GIM using the dosage of fluorescein sodium 0.02ml/kg.SIGNIFICANCE:Using single-injection 10% fluorescein sodium 0.02ml/kg is the optimal dosage based on excellent image quality and with no yellow discoloration affect on the appearance. This lower dosage is also efficient for the diagnosis of GIM. This will make pCLE father widly used as an endoscopic device for in vivo real-time image of the gastrointestinal tract. It is of great clinical significance. |
PDF Full Text Request