Prospective replacement of magnifying endoscopy by a newly developed endocytoscope, the 'GIF-Y0002'

The development and clinical application of microscopic endoscopy for in vivo optical biopsies: Endocytoscopy and confocal laser endomicroscopy

Endoscopies are crucial for detecting and diagnosing diseases in gastroenterology, pulmonology, urology, and other fields. To accurately diagnose diseases, sample biopsies are indispensable and are currently considered the gold standard. However, random 4-quadrant biopsies have sampling errors and time delays. To provide intraoperative real-time microscopic images of suspicious lesions, microscopic endoscopy for in vivo optical biopsy has been developed, including endocytoscopy and confocal laser endomicroscopy. This article reviews recent advances in technology and clinical applications, as well as their shortcomings and future directions.

Accuracy of artificial intelligence-assisted detection of esophageal cancer and neoplasms on endoscopic images: A systematic review and meta-analysis

OBJECTIVE

To investigate systematically previous studies on the accuracy of artificial intelligence (AI)-assisted diagnostic models in detecting esophageal neoplasms on endoscopic images so as to provide scientific evidence for the effectiveness of these models.

METHODS

A literature search was conducted on the PubMed, EMBASE and Cochrane Library databases for studies on the AI-assisted detection of esophageal neoplasms on endoscopic images published up to December 2020. A bivariate mixed-effects regression model was used to calculate the pooled diagnostic efficacy of AI-assisted system. Subgroup analyses and meta-regression analyses were performed to explore the sources of heterogeneity. The effectiveness of AI-assisted models was also compared with that of the endoscopists.

RESULTS

Sixteen studies were included in the systematic review and meta-analysis. The pooled sensitivity, specificity, positive and negative likelihood ratios, diagnostic odds ratio and area under the summary receiver operating characteristic curve regarding AI-assisted detection of esophageal neoplasms were 94% (95% confidence interval [CI] 92%-96%), 85% (95% CI 73%-92%), 6.40 (95% CI 3.38-12.11), 0.06 (95% CI 0.04-0.10), 98.88 (95% CI 39.45-247.87) and 0.97 (95% CI 0.95-0.98), respectively. AI-based models performed better than endoscopists in terms of the pooled sensitivity (94% [95% CI 84%-98%] vs 82% [95% CI 77%-86%, P < 0.01).

CONCLUSIONS

The use of AI results in increased accuracy in detecting early esophageal cancer. However, most of the included studies have a retrospective study design, thus further validation with prospective trials is required.

The Impact of Artificial Intelligence in the Endoscopic Assessment of Premalignant and Malignant Esophageal Lesions: Present and Future

In the gastroenterology field, the impact of artificial intelligence was investigated for the purposes of diagnostics, risk stratification of patients, improvement in quality of endoscopic procedures and early detection of neoplastic diseases, implementation of the best treatment strategy, and optimization of patient prognosis. Computer-assisted diagnostic systems to evaluate upper endoscopy images have recently emerged as a supporting tool in endoscopy due to the risks of misdiagnosis related to standard endoscopy and different expertise levels of endoscopists, time-consuming procedures, lack of availability of advanced procedures, increasing workloads, and development of endoscopic mass screening programs. Recent research has tended toward computerized, automatic, and real-time detection of lesions, which are approaches that offer utility in daily practice. Despite promising results, certain studies might overexaggerate the diagnostic accuracy of artificial systems, and several limitations remain to be overcome in the future. Therefore, additional multicenter randomized trials and the development of existent database platforms are needed to certify clinical implementation. This paper presents an overview of the literature and the current knowledge of the usefulness of different types of machine learning systems in the assessment of premalignant and malignant esophageal lesions via conventional and advanced endoscopic procedures. This study makes a presentation of the artificial intelligence terminology and refers also to the most prominent recent research on computer-assisted diagnosis of neoplasia on Barrett’s esophagus and early esophageal squamous cell carcinoma, and prediction of invasion depth in esophageal neoplasms. Furthermore, this review highlights the main directions of future doctor–computer collaborations in which machines are expected to improve the quality of medical action and routine clinical workflow, thus reducing the burden on physicians.

Chondromodulin-1 and vascular endothelial growth factor-A expression in esophageal squamous cell carcinoma: accelerator and brake theory for angiogenesis at the early stage of cancer progression

BACKGROUND

Magnifying endoscopy has demonstrated dramatic morphologic changes in the surface microvasculature of superficial esophageal squamous cell carcinoma (ESCC) according to the depth of invasion. We investigated the mechanism of angiogenesis in early-stage ESCC by examining the expression of vascular endothelial growth factor (VEGF)-A and chondromodulin (ChM)-1.

METHODS

Using 41 samples of superficial esophageal cancer (EP and LPM 19 cases, MM or deeper 22 cases) and 7 samples of regenerative squamous epithelium, the expression of VEGF-A and ChM-1 was examined in relation to the histological grade or morphology of the surface microvasculature demonstrated by magnifying endoscopy (types A, B, and C correspond to types A, B1, and B2 and B3 of the magnifying endoscopic classification of the Japan Esophageal Society, respectively). We also investigated the correlation between CD31-positive microvessel density (MVD) and VEGF-A or ChM-1 expression.

RESULTS

In normal squamous epithelium, regenerative squamous epithelium, EP and LPM cancer, and MM or deeper cancer, the positivity rates for VEGF-A and ChM-1 were 0%, 85.7%, 52.6% and 90.9%, respectively, and 48.5%, 71.4%, 73.7% and 23.8%, respectively. The VEGF-A and ChM-1 positivity rates in type B or type C vasculature were 70.0% and 76.2%, respectively, and 75.0% and 19.0%, respectively. The expression of neither VEGF-A nor ChM-1 in cancer cells was correlated with MVD (P = 0.19 and 0.68, respectively), whereas that of VEGF-A in stromal mononuclear cells (SMCs) was significantly correlated with MVD (P = 0.04).

CONCLUSION

Angiogenesis at the early stage of ESCC progression is configured by the balance between accelerator (angiogenic factors from both cancer cells and SMCs) and brake (angiogenic inhibitor) factors.

Diagnosis using deep-learning artificial intelligence based on the endocytoscopic observation of the esophagus

BACKGROUND AND AIMS

The endocytoscopic system (ECS) helps in virtual realization of histology and can aid in confirming histological diagnosis in vivo. We propose replacing biopsy-based histology for esophageal squamous cell carcinoma (ESCC) by using the ECS. We applied deep-learning artificial intelligence (AI) to analyse ECS images of the esophagus to determine whether AI can support endoscopists for the replacement of biopsy-based histology.

METHODS

A convolutional neural network-based AI was constructed based on GoogLeNet and trained using 4715 ECS images of the esophagus (1141 malignant and 3574 non-malignant images). To evaluate the diagnostic accuracy of the AI, an independent test set of 1520 ECS images, collected from 55 consecutive patients (27 ESCCs and 28 benign esophageal lesions) were examined.

RESULTS

On the basis of the receiver-operating characteristic curve analysis, the areas under the curve of the total images, higher magnification pictures, and lower magnification pictures were 0.85, 0.90, and 0.72, respectively. The AI correctly diagnosed 25 of the 27 ESCC cases, with an overall sensitivity of 92.6%. Twenty-five of the 28 non-cancerous lesions were diagnosed as non-malignant, with a specificity of 89.3% and an overall accuracy of 90.9%. Two cases of malignant lesions, misdiagnosed as non-malignant by the AI, were correctly diagnosed as malignant by the endoscopist. Among the 3 cases of non-cancerous lesions diagnosed as malignant by the AI, 2 were of radiation-related esophagitis and one was of gastroesophageal reflux disease.

CONCLUSION

AI is expected to support endoscopists in diagnosing ESCC based on ECS images without biopsy-based histological reference.

Thymidine phosphorylase and angiogenesis in early stage esophageal squamous cell carcinoma

BACKGROUND

The relationship between thymidine phosphorylase (TP) and angiogenesis at the early stage of esophageal squamous cell carcinoma has been unclear.

METHODS

Using 14 samples of normal squamous epithelium, 11 samples of low-grade intraepithelial neoplasia, and 64 samples of superficial esophageal cancer, microvessel density (MVD) was estimated using immunostaining for CD34 and CD105. TP expression was also evaluated in both cancer cells and stromal monocytic cells (SMCs). We then investigated the correlation between MVD and TP expression in both cancer cells and SMCs.

RESULTS

On the basis of the above parameters, MVD was significantly higher in cancerous lesions than in normal squamous epithelium. In terms of CD34 and CD105 expression, MVD showed a gradual increase from normal squamous epithelium, to low-grade intraepithelial neoplasia, and then to M1 and M2 cancer, and M3 or deeper cancer. M1 and M2 cancer showed overexpression of TP in both cancer cells and SMCs. There was no significant correlation between TP expression in cancer cells and MVD estimated from CD34 (rS = 0.16, P = 0.21) or CD105 (rS = 0.05, P = 0.68) expression. Significant correlations were found between TP expression in SMCs and CD34-related (rS = 0.46, P < 0.001) and CD105-related (rS = 0.34, P < 0.01) MVD. In M3 or deeper cancers, there were no significant correlations between TP expression in cancer cells or SMCs and venous invasion, lymphatic invasion, and lymph node metastasis.

CONCLUSION

TP expression is activated in both cancer cells and stromal monocytic cells at the very early stage of ESCC progression. TP expression in SMCs, rather than in cancer cells, is significantly correlated with angiogenesis.

On endocytoscopy and posttherapy pathologic staging in esophageal cancers, and on evidence-based methodology

The following, from the 12th OESO World Conference: Cancers of the Esophagus, includes commentaries on the value of endocytoscopy to replace biopsy histology for squamous cell carcinoma and the clinical significance of posttherapy pathologic stage in patients with esophageal adenocarcinoma following preoperative chemoradiation; a short discussion of evidence-based methodology is also included.

Endocytoscopic observation of various esophageal lesions at ×600: can nuclear abnormality be recognized?

Endocytoscopy (ECS) is a novel endoscopic technique that allows detailed diagnostic examination of the gastrointestinal tract at the cellular level. We previously reported that use of ECS at ×380 magnification (GIF-Y0002) allowed a pathologist to diagnose esophageal squamous cell carcinoma (ESCC) with high sensitivity (94.9%) but considerably low specificity (46.7%) because this low magnification did not reveal information about nuclear abnormality. In the present study, we used the same magnifying endoscope to observe various esophageal lesions, but employed digital 1.6-fold magnification to achieve an effective magnification of ×600, and evaluated whether this improved the diagnostic accuracy in distinguishing neoplastic from non-neoplastic lesions.We examined the morphology of surface cells using vital staining with toluidine blue and compared the histological features of 40 cases, including 19 case of ESCC and 21 non-neoplastic esophageal lesions (18 cases of esophagitis, 1 case of glycogenic acanthosis, 1 case of leiomyoma, and 1 case of normal squamous epithelium). One endoscopist classified the lesions using the type classification, and we consulted one pathologist for judgment of the ECS images as 'neoplastic', 'borderline', or 'non-neoplastic'. At ×600 magnification, the pathologist confirmed that nuclear abnormality became evident, in addition to the information about nuclear density provided by observation at ×380. The overall sensitivity and specificity with which the endoscopist was able to predict neoplastic lesions using the type classification was 100% (19/19) and 90.5% (19/21), respectively, in comparison with values of 94.7% (18/19 cases) and 76.2% (16/21), respectively, for the pathologist using a magnification of ×600. The pathologist diagnosed two non-neoplastic lesions and one case of ESCC showing an apparent increase of nuclear density with weak nuclear abnormality as 'borderline'. Among the 21 non-cancerous lesions, two cases of esophagitis that were misdiagnosed by the endoscopist were also misinterpreted as 'neoplastic' by the pathologist. We have shown, by consultation with a pathologist, that an ECS magnification of ×600 (on a 19-inch monitor) is adequate for recognition of nuclear abnormality. We consider that it is feasible to diagnose esophageal neoplasms on the basis of ECS images, and that biopsy histology can be omitted if a combination of increased nuclear density and nuclear abnormality is observed.

Angiogenesis in superficial esophageal squamous cell carcinoma: assessment of microvessel density based on immunostaining for CD34 and CD105

OBJECTIVE

The esophagus is the only organ where changes in the superficial microvasculature from normal squamous epithelium to invasive cancer are evident by magnifying endoscopy. We investigated in detail the features of angiogenesis in early-stage esophageal cancer using CD34 and CD105 immunostaining, and also the correlation between angiogenesis and mononuclear cell infiltration.

MATERIALS AND METHODS

Using 10 samples of normal squamous epithelium, 7 samples of low-grade intraepithelial neoplasia, and 45 samples of superficial esophageal cancer, we determined the microvessel density at hot spots showing positive staining for CD34 and CD105. We observed the histological features of CD34- and CD105-positive microvessels that corresponded to observations made by magnifying endoscopy. We then investigated the correlation between microvessel density and each histological situation or the grade of mononuclear cell infiltration.

RESULTS

The histological features of CD34- and CD105-positive microvessels were able to explain the morphological changes in the microvasculature during cancer progression observed by magnifying endoscopy. The microvessel density for CD34 or CD105 was significantly correlated with each of the histological types (P < 0.001, rS = 0.51 and 0.76, respectively). Mononuclear cell infiltration at CD105 hot spots was most frequent in M1 and M2 cancer (94.7%). The correlation between the degree of mononuclear cell infiltration and microvessel density for CD105 staining was also significant (P < 0.001, rS = 0.49).

CONCLUSIONS

The microvessel density based on CD34 and CD105 immunostaining can be used to corroborate observations of superficial esophageal squamous cell carcinoma made by magnifying endoscopy. Mononuclear cell infiltration may play an important role in angiogenesis at the early stage of cancer progression.

Detailed features of palisade vessels as a marker of the esophageal mucosa revealed by magnifying endoscopy with narrow band imaging

The palisade vessels present at the distal end of the esophagus are considered to be a landmark of the esophagogastric junction and indispensable for diagnosis of columnar-lined esophagus on the basis of the Japanese criteria. Here we clarified the features of normal palisade vessels at the esophagogastric junction using magnifying endoscopy. We prospectively studied palisade vessels in 15 patients undergoing upper gastrointestinal endoscopy using a GIF-H260Z instrument (Olympus Medical Systems Co., Tokyo, Japan). All views of the palisade vessels were obtained at the maximum magnification power in the narrow band imaging mode. We divided the area in which palisade vessels were present into three sections: the area from the squamocolumnar junction (SCJ) to about 1 cm orad within the esophagus (Section 1); the area between sections 1 and 3 (Section 2); and the area from the upper limit of the palisade vessels to about 1 cm distal within the esophagus (Section 3). In each section, we analyzed the vessel density, caliber of the palisade vessels, and their branching pattern. The vessel density in Sections 1, 2, and 3 was 9.1 ± 2.1, 8.0 ± 2.6, and 3.3 ± 1.3 per high-power field (mean ± standard deviation [SD]), respectively, and the differences were significant between Sections 1 and 2 (P= 0.0086) and between Sections 2 and 3 (P < 0.0001). The palisade vessel caliber in Sections 1, 2, and 3 was 127.6 ± 52.4 µm, 149.6 ± 58.6 µm, and 199.5 ± 75.1 µm (mean ± SD), respectively, and the differences between Sections 1 and 2, and between Sections 2 and 3, were significant (P < 0.0001). With regard to branching form, the frequency of branching was highest in Section 1, and the 'normal Y' shape was observed more frequently than in Sections 2 and 3. Toward the oral side, the frequency of branching diminished, and the frequency of the 'upside down Y' shape increased. The differences in branching form were significant among the three sections (P < 0.0001). These results indicate that the density of palisade vessels is highest near the SCJ, and that towards their upper limit they gradually become more confluent and show an increase of thickness. Within a limited area near the SCJ, observations of branching form suggest that palisade vessels merge abruptly on the distal side. We have demonstrated that palisade vessels are a useful marker for endoscopic recognition of the lower esophagus.

Current status and limitations of the newly developed endocytoscope GIF-Y0002 with reference to its diagnostic performance for common esophageal lesions

OBJECTIVES

To investigate both neoplastic and non-neoplastic lesions of the esophagus and to clarify the features of the surface cell morphology using a newly developed endocytoscope, the GIF-Y0002.

METHODS

The surface cell morphology was examined with toluidine blue staining, and histological features of 53 patients with 54 lesions, including 39 patients with esophageal squamous cell carcinoma (ESCC) and 14 patients with 15 non-neoplastic esophageal lesions, were compared. One endoscopist classified the lesions using type classification, and we consulted one pathologist to evaluate the endocytoscopy pictures with regard to neoplasia or non-neoplasia.

RESULTS

The overall sensitivity for ESCC of the findings by the endoscopist and pathologist based on GIF-Y0002 observation were 100.0% and 94.9%, respectively; while the specificity was 80.0% and 46.7%. For the 3 cases of low-grade intraepithelial neoplasia, 2 were diagnosed as Type 2 and one case as suspected neoplasia by the endoscopist while the pathologist considered 2 cases to be neoplastic. Among the 9 cases of esophagitis, the endoscopist diagnosed 2 cases as Type 2 or 3, which was suggestive of neoplasia, whereas the pathologist diagnosed 6 cases to be neoplastic.

CONCLUSION

The low percentage of specificity for the pathologist's diagnosis was considered to be attributed to the low magnification power of the GIF-Y0002. A further increase in the magnifying power of this instrument will be necessary to broaden its clinical applications.

Angiogenesis in superficial esophageal squamous cell carcinoma: magnifying endoscopic observation and molecular analysis

Observations of esophageal squamous cell carcinoma using magnifying endoscopy have now been carried out extensively and, as a result, it has become clear that the morphology of the microvessels evident at the tumor surface reflects the depth of tumor invasion. In M1 and M2 cancer, the surface microvasculature reveals dilation and elongation of the intrapapillary capillary loops (IPCL). However, at this stage, some immature capillaries resembling IPCL also arise inside the tumor and, therefore, the view of the microvasculature should be described as one showing 'intermixing of modified IPCL and IPCL-like immature capillaries (IPCL-like abnormal capillary)'. As cancer invades into the muscularis mucosa (M3 or deeper), an obviously dilated and irregularly branched tumor-specific vasculature, more accurately described as 'neovasculature', can be observed. From our magnifying endoscopy observations and studies of the molecular profile of early esophageal cancer, we conclude that two major angiogenic steps exist in precancerous and M3 lesions in the early phase of cancer progression. In addition, it is now possible to study cell morphology using an endocytoscope with a much higher magnification (×400-×1000) than magnifying endoscopes currently on the market. The histology revealed in this way may reduce the need for conventional biopsy histology in the future.