Endocytoscopy for histological evaluation of inflammation in ulcerative colitis

Abstract

BACKGROUND

Histological remission in ulcerative colitis (UC) can be assessed traditionally by microscopic examination of endoscopic biopsies. Recently, advanced endocytoscopy has been increasingly used for real-time evaluation of the colonic mucosa in UC.

AIM

To develop a novel scoring system based on endocytoscopic (ECS) images to ascertain the function of ECS score (ECSS) in the evaluation of endoscopic remission and histological remission among patients with UC.

METHODS

In this retrospective study, ECS images were analyzed and ECSS was developed to differentiate between active and inactive UC. ECSS was correlated with the Mayo endoscopic score and the pathological Geboes score (GS).

RESULTS

A total of 72 UC patients were included in the study. With reference to the GS, a value of 14.5 on the newly developed ECSS served as the cut-off value. Patients were divided into two groups based on this cut-off value: Active or inactive UC. The ECSS showed a strong correlation with the more stringent pathological GS (< 2B for inflammatory inactive) (kappa value: 0.838) and differences from the pathological GS (< 3.2 for inflammatory inactive). In contrast, white light endoscopy showed differences from the more stringent pathological GS (< 2B for inflammation) and only general agreement with the relatively low standard GS (< 3.2 for inflammation) (kappa value: 0.496). There was high consistency in the ECSS among endoscopists (intraclass correlation coefficient value of 0.790).

CONCLUSION

The endocytoscopy technique has high interobserver consistency and good reproducibility. The newly developed ECSS system has a strong correlation with the histological changes associated with patients with UC. It can be used to stratify patients based on inflammatory activity and mucosal healing without biopsies, and warrants further investigation.

Key Words: Ulcerative colitis; Endocytoscopy; Endocytoscopic score; Geboes score; Histological remission; Mucosal healing

Core Tip: Histological remission in ulcerative colitis (UC) can be assessed traditionally by microscopic examination of endoscopic biopsies. Recently, advanced endocytoscopy has been increasingly used for real-time evaluation of the colonic mucosa in UC. The current study developed a novel scoring system based on endocytoscopic images called as endocytoscopic score (ECSS) to evaluate endoscopic and histological remission in UC. ECSS showed strong correlation with histological changes. ECSS has high interobserver consistency and good reproducibility. ECSS can be used to stratify patients based on inflammatory activity and mucosal healing without biopsies, providing a new direction for real-time non-invasive assessment of UC.

INTRODUCTION

Ulcerative colitis (UC) is a chronic inflammatory disease of the bowel. Induction and maintenance of disease remission is crucial for treatment. According to previous studies, about 40% of patients in clinical remission have active lesions on endoscopy[1]. Hence, endoscopic remission (ER) is considered to be remission as it is associated with a low recurrence rate and better prognosis[2]. The TREAT-TO-TARGET strategy of the Selecting Therapeutic Targets in Inflammatory Bowel Disease-II initiative recommended complete endoscopic mucosal healing (MH) as the best therapeutic target for the management of UC. However, subsequent studies showed that even with MH (a Mayo Endoscopic Score of 0), 12% to 17% of patients still experience clinical relapse within 12 months[1-3]. At the same time, increasing evidence suggests tissue inflammation in patients with ER may be responsible for this recurrence and poor prognosis. Hence, ER is not equivalent to histological remission (HR). The difference between HR and ER stems from the drawbacks of conventional endoscopy, which is unable to accurately detect submucosal inflammation. Several studies have demonstrated that HR is strongly associated with decreased steroid use, reduced risk of complications, reduced hospitalization, and decreased incidence of colorectal cancer in patients with UC, and is therefore an important endpoint in assessing treatment outcomes[4-6]. Although a consensus has not been formally delineated, MH is increasingly understood to include both ER and HR. The definition of remission has evolved from clinical remission alone, to also now include ER and further to HR, with the treatment goals becoming increasingly ambitious[1]. By incorporating HR in the definition of MH, the concept of deep remission follows[4,7].

Currently, there are two main methods to assess HR: Endoscopic biopsy followed by pathological examination and advanced microendoscopy. The traditional method relies on the pathological examination of biopsy samples, whereas the absence of inflammatory cell infiltrates, such as neutrophils, in the mucosal epithelium or lamina propria is considered to be HR. Among the various pathological tools, the Geboes score (GS) is the most widely accepted scoring system. A GS of less than 2B or 3.2 both indicate HR with the criteria of 2B.0 being the most stringent. However, there are limitations of this traditional method, including the need for multiple biopsies, tissue processing time, and restriction of the bowel assessment to the biopsied area. The second technique to determine HR is to use an advanced microendoscope with unrestricted real-time observation and perform simultaneous histological evaluation[1,8]. Currently, advanced endoscopic techniques are coming closer to histological assessment, by introducing new concepts of mucosal and vascular healing patterns such as virtual electron dye endoscopy, high definition (HD) imaging, etc.[4,9,10], and the use of ultra-magnification optical microscopy or endocytoscopy. Among them, the microendoscope can directly observe the microstructure of the crypts (such as crypt shape, and spacing) and the morphology of the cell nucleus. It can assess the histological activity of UC in real-time[11], and is highly reproducible among endoscopists[4]. The current study was conducted to determine the role of endocytoscopic (ECS) score in the assessment of ER and HR in patients with UC.

MATERIALS AND METHODS

Patients

This retrospective study was conducted in the Department of Small Intestinal Endoscopy at the Sixth Affiliated Hospital of Sun Yat-sen University, from February 2022 to August 2022. A total of 72 patients with UC were included. The study was based on endocytoscopic images and a newly developed ECS score (ECSS), aiming to measure inflammatory activity to distinguish between active and inactive UC, and to assess the consistency between the ECSS and the GS, as well as the consistency between endoscopists and pathologists. The inclusion criteria were as follows: (1) Patients met the diagnostic criteria of the 2023 Xi’an Consensus on Inflammatory Bowel Disease and Treatment[12]; (2) Three months period from the time of diagnosis; and (3) Mayo score between 0 and 2 (Table 1)[12]. Exclusion criteria were: (1) Mayo score of 3; and (2) Previous history of colonic resection, pregnancy, self-discontinuation of medications during follow-up, inability to provide informed consent, presence of coagulation dysfunction or severe comorbidities, and/or incomplete baseline data. All patients included in this study signed an informed consent form. Approval from the Ethics Committee of the Sixth Affiliated Hospital of Sun Yat-sen University, approval No. E2025172 was obtained and the study was performed in accordance with the relevant regulations of the Helsinki Declaration.

Table 1 Mayo endoscopic score for ulcerative colitis.

Number	Characteristics
0	Normal mucosa or inactive disease
1	Mild disease: Mucosal erythema, decreased vascular pattern, mild friability
2	Moderate disease: Marked mucosal erythema, absent vascular pattern, friability, erosions
3	Severe disease: Spontaneous mucosal bleeding, mucosal ulceration

Endoscopic assessment

All patients participating in this study underwent colonoscopy with a hyperexpanded cell endoscope (ECS; CF-H290ECI, a prototype from Olympus, Japan). ECS is a contact optical microendoscope system that offers ultra-magnification capabilities, with a magnification power of up to 520. In this study, we employed two independent observation modes of the ECS: A conventional mode and an ultra-magnification mode. The ultra-magnification mode facilitated cellular-level observation, such as the examination of crypt structures, nuclei, and other fine structures, thus capturing ultra-magnified images and identifying areas of inflammation and healing (Figure 1). To ensure consistency, all examinations were conducted by two endoscopists with extensive experience in ECS and in optical diagnosis of inflammatory bowel disease. Endoscopy was performed under local anesthesia on an outpatient basis. However, in some cases, patients required admission and/or conscious sedation with propofol and/or remimazolam.

Figure 1 Endocytoscopic scoring among patients with endoscopic remission (Mayo endoscopic score of 0). A: Conventional endoscopy image showing remission; B: The corresponding endocytoscopic score (ECSS) image: Round crypts, normal distance between adjacent crypts with no inflammatory cell infiltration (ECSS score of 0, inactive inflammation); C: Conventional endoscopy image showing remission; D: The corresponding ECSS image, widened distance between the adjacent crypts with inflammatory cell infiltration (ECSS score of 1, active inflammation).

Conventional endoscopic examination began with HD white light endoscope (WLE) up to the terminal ileum, followed by withdrawal of the colonoscope for observation. The inflammatory activity was quantified using the white light endoscopy scoring system [Mayo endoscopy score (MES)]. ER was defined as a MES score of 0 or 1. The mucosa was then irrigated with water and acetic acid, followed by staining with a 5 mL solution (containing 0.4 mL crystal violet, 2 mL methylene blue, and 2.6 mL 0.9% sodium chloride). After 1-2 minutes of staining, ECS was performed. To ensure standardization in the quality of the ECS images for analysis, the following criteria was employed by the endoscopist: (1) Clear and focused images; (2) Appropriate image brightness; and (3) No exudates or residual matter interfering with the image. Tissue biopsy was taken from the corresponding image area (including regions showing active inflammation, suspected active inflammation, or endoscopic healing). The endoscopist scored each criterion according to the ECSS system during the procedure (Table 2), and the scores were reconfirmed and added after the procedure. In this study, the ECSS system included observing the five most inflammatory representative crypts in each field of view, examining each crypt opening (0 = normal, 1 = deformed, 2 = absent), crypt spaces (0 = normal, 1 = widened, 2 = lymphocyte infiltration, 3 = neutrophil infiltration), and inflammatory cell infiltration within the crypts (0 = none, 1 = lymphocytes, 2 = neutrophils) (Figure 1). Each observation criterion was scored on a scale of 0 to 2 or 3. The final ECSS was calculated by summing the scores of three indicators observed in all crypts. All images captured during HD WLE and ECSS were digitally stored using the Olympus IMH-20 recording system for subsequent analysis.

Table 2 Endoscopic scoring system.

Parameter		ECSS (1 + 2 + 3)
Dyeing effect: A: Good; B: Difference
ESCC criteria		Crypt 1	Crypt 2	Crypt 3	Crypt4	Crypt 5	Crypt score (1 + 2 + 3 + 4 + 5)
Crypt orifice	Normal	-	-	-	-	-	Total
	Deformed
	Disappeared
Crypt space	Normal	-	-	-	-	-	Total
	Widened
	Lymphocytes
	Neutrophils
Ductule	Normal	-	-	-	-	-	Total
	Lymphocyte infiltration
	Neutrophil infiltration

ECSS: Endoscopic scoring.

Histological assessment

All biopsy samples were fixed in a 10% formalin solution, followed by paraffin embedding, serial sectioning, and staining with hematoxylin and eosin. A histological examination was conducted by two senior pathologists who were blinded to the clinical and endoscopic data. Histological evaluation was based on the GS (Table 3)[13]. This scoring system is comprised of six grades; each further subdivided into 4 to 5 subgroups. Based on the final score, patients are categorized into two groups. The first group of patients with inactive disease consisted of individuals with a GS less than 2B.0 or 3.2, indicating that the disease was in an inactive state with no neutrophil infiltration in the epithelial and lamina propria layers. The second group of patients with active disease was composed of individuals with a GS greater than or equal to 2B.0 or 3.2, indicating neutrophil infiltration in the epithelial layer. According to previous studies, this index is a marker of inflammation and a predictor of recurrence[8], suggesting that the disease is in an active state.

Table 3 Geboes scoring system.

Grade	Characteristics
Grade 0 architectural changes	0.0 No abnormality
	0.1 Mild abnormality
	0.2 Mild/moderate diffuse or multi-focal abnormalities
	0.3 Severe diffuse or multifocal abnormalities
Grade 1 chronic inflammatory infiltrates	1.0 No increase
	1.1 Mild but unequivocal increase
	1.2 Moderate increase
	1.3 Marked increase
Grade 2A: Eosinophils in lamina propria	2A.0 No increase
	2A.1 Mild but unequivocal increase
	2A.2 Moderate increase
	2A.3 Marked increase
Grade 2B: Neutrophils in lamina propria	2B.0 No increase
	2B.1 Mild but unequivocal increase
	2B.2 Moderate increase
	2B.3 Marked increase
Grade 3 neutrophils in epithelium	3.0 None
	3.1 < 5% crypts involved
	< 50% crypts involved
	3.3 > 50% crypts involved
Grade 4 Crypt destruction	4.0 None
	4.1 Probable: Local excess of neutrophils in part of the crypts
	4.2 Probable: Marked attenuation
	4.3 Unequivocal crypt destruction
Grade 5 Erosions and ulcerations	5.0 No erosion, ulceration or granulation tissue
	5.1 Recovering epithelium + adjacent inflammation
	5.2 Probable erosion: Focally stripped
	5.3 Unequivocal erosion
	5.4 Ulcer or granulation tissue

Blinding of researchers

In this study, double blinding was used to avoid bias. Pathologists were unaware of the clinical (symptoms and laboratory, imaging findings) and endoscopic results, while endoscopists were unaware of the clinical (symptoms and laboratory, imaging findings) and pathological examination results.

Statistical analysis

SPSS 25.0 statistical software was used for data analysis. Continuous data are presented as the mean ± SD, while categorical data are expressed as frequency and percentage. For measurement data that conformed to normal distribution and homogeneity of variance, t-tests or one-way analysis of variance were used for comparison. For data that did not conform to normal distribution, non-parametric tests were used for data analysis. Categorical data analysis was conducted using the χ² test. The criterion for statistical significance was set at P < 0.05.

In this study, a Spearman correlation coefficient of > 0.8, 0.6-0.8, 0.4-0.6 and < 0.4 indicated a very strong, strong, moderate, and weak correlation, respectively. For the consistency evaluation of ordinal variables, kappa was used. Kappa values > 0.75, 0.4-0.75 and < 0.4 indicated strong, moderate, and poor consistency, respectively. For the consistency evaluation of continuous variables, intraclass correlation coefficient (ICC) was used. ICC values > 0.75, 0.4-0.75 and < 0.4 indicated strong, moderate, and poor consistency, respectively. Referencing the Geboes histological scoring system, the cutoff value of the intracellular endoscopic score was determined by plotting the receiver operating characteristic curve, and its sensitivity and specificity for histological healing (HR) were assessed to explore its predictive ability for HR.

RESULTS

Patient characteristics

A total of 72 patients with UC were included in the study, comprising 36 males and 36 females with a male-to-female ratio of 1:1. The mean age was 38.3 ± 14.7 years. Patient lesions were located in the rectum, left-side colon, and extensive colon in 19 patients, 28 patients, and 25 patients, respectively. Most of the patients had a Mayo score of 0 (Table 4). The median disease course was 39.0 months (interquartile range 18.0, 73.5).

Table 4 Comparison of baseline characteristics between endoscopic scoring active and inactive groups, n (%).

Patient characteristics	Total (n = 72)	ECSS-active (n = 23)	ECSS-inactive (n = 49)	P value	P value (Bonferroni)
Sex
Female	36 (50.00)	11 (47.83)	25 (51.02)	0.800	-
Male	36 (50.00)	12 (52.17)	24 (48.98)	0.800	-
Age, median (range), years	38.3 ± 14.7	36.05 ± 3.03	41.70 ± 14.7	0.128	-
Disease duration, median (range), years	38.3±14.7	30 (18.0, 51.0)	5 (18.0, 78.0)	0.108	-
Extent of colitis				0.029	-
Pancolitis (E3)	25 (34.72)	13 (56.52)	12 (24.49)		0.021 (E3 vs E2)
Left-sided colitis (E2)	28 (38.89)	6 (26.09)	22 (44.90)		1.000 (E2 vs E1)
Proctitis (E1)	19 (26.39)	4 (17.39)	15 (30.61)		0.037 (E1 vs E3)
MES				0.002	-
Mayo 0	55 (76.39)	12 (52.17)	43 (87.76)		0.048 (Mayo 0 vs Mayo 1)
Mayo 1	9 (12.50)	5 (21.74)	4 (8.16)		0.620 (Mayo 1 vs Mayo 2)
Mayo 2	8 (11.11)	6 (26.09)	2 (4.08)		0.005 (Mayo 2 vs Mayo 0)
Mayo 3	0	0	0	-	-
Use of medication for UC
Biological therapy	21 (29.17)	9 (39.13)	12 (24.49)	1.000	-
Infliximab	5 (6.94)	2 (8.70)	3 (17.39)
Vedolizumab	14 (19.44)	6 (26.09)	8 (17.39)
Ustekinumab	2 (2.78)	1 (4.35)	1 (17.39)
Non-biological therapy	49 (68.05)	23 (100)	26 (53.06)	0.703	-
Mesalazine	47 (65.28)	23 (100)	24 (48.98)
Steroids	4 (5.56)	3 (13.04)	1 (4.35)
Immunosuppressants	3 (4.17)	2 (8.70)	1 (4.35)
CRP mean (range) mg/dL	1.9 (0.9, 6.9)	5.17 (0.78, 20.27)	1.44 (0.79, 3.33)	0.042	-

ECSS: Endoscopic scoring; MES: Mayo endoscopy score; CRP: C-reactive protein; UC: Ulcerative colitis.

Comparison between ECSS active and inactive UC patients

There were no significant differences in the age, sex distribution, and disease duration between the two groups. The ECSS-active group had a wider extent of colitis and higher endoscopic scores. The specific demographic details are shown in Table 4.

From the perspective of ECSS, whether it is the total score or the total scores of each scoring indicator (the scoring indicators include crypt opening, crypt space, and the degree of inflammatory cell infiltration in the crypt), the ECSS-active group had significantly higher scores compared to ECSS-inactive group (Table 5 and Figure 2).

Figure 2 Comparison of endocytoscopic findings. ECSS: Endoscopic scoring.

Table 5 Comparison of endocytoscopic findings.

Parameter	ECSS-active (n = 23)	ECSS-inactive (n = 49)	P value (Mann-Whitney U)
ECSS, median (range)	18.00 (16.00, 20.00)	7.00 (4.50, 9.50)	< 0.001
Crypt orifice, median (range)	5.00 (4.50, 5.50)	0.00 (0.00, 2.00)	< 0.001
Crypt space, median (range)	9.00 (7.50, 10.50)	5.00 (3.25, 6.75)	< 0.001
Inflammatory cell infiltration, median (range)	5.00 (4.00, 6.00)	0.00 (0.00, 2.00)	< 0.001

ECSS: Endoscopic scoring.

Comparing the number of patients corresponding to the different scores under the rating indicators, the ECSS-active group showed significantly higher numbers of patients with inflammatory changes. However, the two groups exhibited no significant difference in the number of patients whose crypt openings had disappeared or with widened crypt spacing. This may be attributed to the relatively small number of cases demonstrating disappearance in crypt opening, and the fact that assessment of the widening of crypt intervals is mainly based on visual estimation (Table 6).

Table 6 Comparison of endocytoscopic findings, n (%).

Characteristics	ECSS-active (n = 23)	ECSS-inactive (n = 49)	P value
Total score of ECSS	413	346	-
Crypt orifice (number of patients)
0	10 (43.48)	42 (85.71)	0.000
1	22 (95.65)	15 (30.61)	0.000
2	3 (13.04)	1 (2.04)	0.177
Crypt space (number of patients)
0	3 (13.04)	25 (51.02)	0.002
1	14 (60.87)	31 (63.27)	0.845
2	22 (95.65)	26 (53.06)	0.001
3	3 (13.04)	1 (2.04)	0.177
Ductules (number of patients)
0	7 (30.43)	42 (85.71)	0.000
1	22 (95.65)	21 (42.86)	0.000
2	10 (43.48)	1 (2.04)	0.000

ECSS: Endoscopic scoring.

Correlation between endocytoscopy and pathology

The Spearman correlation coefficient between the ECSS and GS less than 2B was 0.737 (P < 0.001), indicating a strong positive correlation. The Spearman correlation coefficient between the ECSS and GS less than 3.2 was 0.61 (P < 0.001), indicating a strong positive correlation. However, the ECSS exhibited a higher correlation with a GS pathological grade of less than 2B.

Endocytoscopy score, receiver operating characteristic curve, and critical value

The diagnostic accuracy of ECSS based on pathology (dividing patients into active or inactive inflammation based on the GS < 2B) was 0.961 (P < 0.001, 95%CI 0.91-1.00). Using the cut-off value of 14.5, the sensitivity was 0.909, and the specificity was 0.94 (Figure 3A).

Figure 3 Receiver operating characteristic curves of endoscopic scoring for predicting histological inflammation stratified by different Geboes score cut-offs. A: Geboes score < 2B; B: Geboes score < 3.2. AUC: Arear under curve; CI: Confidence interval.

The diagnostic accuracy of cell endoscopy based on pathology (dividing patients into active or inactive inflammation based on the GS < 3.2) was 0.950 (P < 0.001, 95%CI: 0.903-0.996). Using the cut-off value of 14.5, the sensitivity of ECSS was 1.000, and the specificity was 0.845 (Figure 3B).

Consistency between endocytoscopy and pathology

To evaluate the consistency between endocytoscopy and histopathological assessment, a kappa value analysis was performed. It was observed that there was significant agreement between the two methods (endocytoscopy and histopathology, classifying patients into active or inactive inflammation based on whether the GS was < 2B), with a kappa value of 0.838, indicating a high degree of agreement between the two assessments. However, for the classification based on whether the GS was < 3.2, the P value was 0.004, indicating a statistically significant difference, and further assessment of consistency could not be made.

Consistency between white light endoscopy and pathology

To assess the concordance between white light endoscopy and histopathology, a kappa value analysis was conducted. The concordance kappa value between white light endoscopy and pathology, with patients categorized into active or inactive inflammation based on the GS < 3.2, was 0.496 indicating only moderate agreement between the methods. However, when patients were categorized based on a more stringent GS of < 2B, the P value was 0.001, indicating a statistically significant difference. Consequently, further assessment of concordance could not be performed. Combined with the above data, there is a considerable gap between white-light endoscopy and endocytoscopy in terms of histological assessment.

Consistency between endocytoscopy and white light endoscopy

To assess the agreement between endocytoscopy and white light endoscopy, a kappa value analysis was conducted. In this analysis, patients were categorized into groups with active or inactive inflammation based on the MES for WLE and ECSS similarly for endocytoscopy, where patients were classified according to whether their endocytoscopy score was less than 14.5. A P value of 0.001 was found when comparing white light endoscopy with cell endoscopy, signifying a statistically significant difference. Consequently, it is not feasible to further evaluate the consistency between the two methods.

Assessment of pathologists’ consistency

The consistency among pathologists was evaluated by analyzing the assessment results of two pathologists using kappa coefficient analysis. The kappa value was 0.798, indicating a high degree of consistency.

Assessment of endoscopists’ consistency

To evaluate the consistency between endoscopists, the intracellular endoscopic scores of two endoscopists were analyzed using ICC. The ICC value was 0.790, indicating a high degree of consistency.

DISCUSSION

HR is an important therapeutic goal in the treatment of patients with UC in order to achieve long-term clinical remission. Typically, histological examination of colonoscopic biopsies is performed to determine HR, however, it is a time-consuming procedure and does not allow real time assessment of the colon. With the emergence of advanced endoscopic technologies, such as HD and virtual chromoendoscopy endoscopy, there has been continuous improvement in the assessment of the intestinal mucosa, making it possible to visualize the microstructural features of the mucosa within the body, thus reducing the differences between endoscopic ER and histological HR. Studies on the predictors of clinical recurrence in UC have been conducted, with most studies assessing prognosis through changes in crypt structure and the microvascular system[2,4,14]. Only a few studies have mentioned the potential use of ECSS, which is believed to be able to evaluate a broader range of intestinal mucosal areas more accurately. However, due to the varying experimental methods, subjects, and scoring criteria in these studies, the conclusions lack persuasiveness[8,11,15]. In our study, we specifically referenced the Geboes histological scoring system to refine and optimize ECSS, revealing its significant potential in assessing histological activity. We found that a high correlation between ECSS and GS was particularly important in determining whether patients with an MES (Mayo Endoscopic Score) of 0 or 1 have histological inflammatory activity. Endocytoscopy is highly accurate in determining HR, showing significant consistency with the stricter HR criteria of a GS of < 2B.0. It was significantly superior to WLE in determining HR. In this study, the ECSS cut-off value corresponding to HR was 14.5 as mentioned in the results section previously. However, the ECSS system is an integer-based evaluation system, so to facilitate clinical decision-making, we recommend adjusting the cutoff value of 14.5 points to 15 points. Hence, ECSS exceeding 15 points corresponds to the active phase of histological GS, while a score below 15 points corresponds to the remission phase. Regardless of whether the patient is in an inflammatory active state or a state of remission ECS can detect subtle inflammatory changes in patients with UC and assess HR without relying on biopsy specimens. It provides a simple and objective basis for optimizing UC treatment strategies, making endoscopic assessment of histological healing possible.

This study selected a significant number of patients who experienced symptom relief with treatment and had a MES of 0. The histological inflammatory state of these patients was rigorously assessed by endocytoscopy. The ECSS criteria of this study focused on inflammatory cells and crypt structures, observing at least five crypts in each field of view at a time, with each evaluation criterion for each crypt scored from 0-2 or 3, providing a more extensive and detailed assessment, which is the innovation of this study. In this study, we utilized artificial staining as although it is time-consuming, is not influenced by intestinal conditions and allows for the observation of numerous indicators, such as the shape and distance of crypts, infiltration of inflammatory cells, and provides a basis for further subcategorization of patients with an MES of 0. Previous studies by Nishiyama et al[16] and Ueda et al[17] also reported that ECS could predict clinical recurrence, but due to the limited number of patients in the studies (less than 30 cases) and the observation area being restricted to the rectum, the results could not be generalized. This study followed 72 patients and expanded the scope of observation to verify the research findings. The process of using ECSS to identify the ultrastructural characteristics of mucosal ER is easy and the endoscopic experts can effectively use it with minimal training time. There is high consistency in judgments among endoscopists, as well as high consistency with pathologists, and the assessment has high repeatability. The next step is to transition to assessments powered by artificial intelligence, thereby further enhancing work efficiency. All the above findings are improvements developed in this study, compared to previous ones[4,8,11,14].

This study has several limitations. First, the patient sample size was relatively small and it was a single-center study. The cutoff value derivation and model performance evaluation in this study were both based on the same cohort of 72 patients. Although this design is commonly used in preliminary exploratory studies, it may introduce the risk of overfitting. Future prospective multicenter studies are required to validate the findings of this study. Second, there was no unified treatment plan for the study patients. Although this study did not evaluate the efficacy of treatment, the impact was relatively minor. However, stratified analysis should be considered in subsequent studies involving treatment efficacy. Third, the position for image acquisition was determined by the endoscopist, and the assessment of potential inflammation in the intestinal mucosa may vary depending on the location. Further studies must employ standardized sampling protocols (e.g., fixed positioning for each segment of the colon). Fourth, the images were obtained in vivo, but the total ECSS score was calculated after the procedure[2,8,11]. The assessment of deep recess space primarily relies on visual estimation, with the future introduction of image analysis software for quantification. Fifth, basal plasmacytosis typically occurs at deeper levels, and it is challenging to comprehensively assess inflammatory cells throughout the entire lamina propria using ECSS[18,19]. Sixth, lack of long-term follow-up data and subsequent follow-up studies could be conducted to validate the prognostic value of ECSS. All these areas are ones that we need to focus on in future studies.

CONCLUSION

In conclusion, the endoscopic evaluation of MH has entered a new stage of development. Compared with traditional histological scoring methods, advanced endoscopic techniques, such as ECS, can effectively stratify the clinical risk of recurrence in patients with UC who show sustained clinical remission and endoscopic healing under safer and simpler operating conditions. As a bridge between endoscopy and histological assessment, the MH revealed by endocytoscopy is different from previous MH concepts, and it may become an important new focus in the management of disease in patients with UC. Although endocytoscopy technology shows great potential and expectation, further research and validation are still necessary[1,4].