A Scoping Review of Endoscopic Assessment Considering Disease Extent in Ulcerative Colitis: Insights for the Artificial Intelligence Era Authors: Yasuharu Maeda, Shin-ei Kudo, Takanori Kuroki, Yurie Kawabata, Katsuro Ichimasa, Masashi Misawa, Noriyuki Ogata, Kazuo Ohtsuka Published: 2026 ABSTRACT Conventional endoscopic indices for ulcerative colitis (UC) primarily assess the most severely affected segment, potentially underestimating the cumulative inflammatory burden across the colon. Consequently, there is increasing interest in assessment methods that incorporate both disease severity and spatial extent, including emerging artificial intelligence (AI)–based approaches. The aim of this study was to map and critically evaluate current evidence on endoscopic indices that integrate disease extent in UC, including both conventional and AI-assisted approaches. Original studies in which patients with UC were evaluated via endoscopic approaches for disease activity, severity, or extent and clinically relevant outcomes were reported. Non-original articles and studies of other diseases were excluded. PubMed/MEDLINE and the Cochrane Central Register of Controlled Trials were searched for studies published between January 2000 and December 2025. Two reviewers independently screened studies and extracted data. Findings were synthesized narratively due to heterogeneity. Twenty-one studies were included: eight addressed index development or validation, 11 evaluated clinical outcomes, and four investigated AI-assisted assessments, with some studies contributing to multiple categories. Extent-based indices showed stronger correlations with biomarkers and histological activity than conventional focal scores, although their ability to predict relapse was inconsistent. AI-based systems enabled automated assessment of inflammatory distribution, improving reproducibility and showing promising clinical associations, but external validation remains limited. Extent-integrated endoscopic assessment provides a more comprehensive evaluation of disease burden in UC patients and may improve risk stratification and treatment monitoring. However, methodological heterogeneity and limited prospective validation hinder clinical implementation. Further standardization and multicenter studies are needed. INTRODUCTION Endoscopic evaluation is fundamental to the assessment and management of ulcerative colitis (UC), serving as an objective measure of mucosal inflammation and a key outcome in clinical trials and treat-to-target strategies.1,2 The Mayo Endoscopic Subscore (MES)3 and Ulcerative Colitis Endoscopic Index of Severity (UCEIS)4 remain the most widely used indices in both clinical practice and research; however, these indices primarily reflect the most severely affected segment of the colon and do not account for the longitudinal extent of inflammation. This limitation is clinically relevant, as UC is characterized by continuous mucosal involvement of variable length, and the cumulative inflammatory burden across the colon may influence symptoms, biomarker levels, therapeutic response, and long-term prognosis. To overcome these shortcomings, several endoscopic indices incorporating disease extent have been developed.5-7 These indices assess inflammation on a segmental basis and aggregate severity across the colon using summation, weighting, or extent-adjusted calculations. By integrating both severity and distribution of disease, such scores aim to provide a more comprehensive representation of global mucosal activity than conventional point-based assessments. In parallel, recent advances in artificial intelligence (AI) have enabled not only automated,8-10 quantitative endoscopic assessment11-13 but also approaches that explicitly incorporate disease extent.14,15 AI-assisted systems can analyze endoscopic images or videos, localize inflammation across colonic segments, and generate reproducible Cumulative Disease Scores (CDS).15,16 Such approaches may reduce interobserver variability and improve prognostic accuracy, but their role relative to established extent-based indices has not been comprehensively evaluated. Therefore, we conducted a scoping review to comprehensively summarize and critically appraise endoscopic indices that incorporate disease extent in UC, including both conventional and AI-assisted approaches. We evaluated their methodological characteristics, associations with biomarkers and histology, and evidence for predicting clinical outcomes and treatment response to clarify their current role and inform future directions in global endoscopic assessment of UC (Fig. 1). METHODS This study is a scoping review summarizing the existing literature on endoscopic assessment incorporating disease extent in UC, including both conventional indices and emerging AI-assisted approaches. The review was performed and is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR). The completed PRISMA-ScR checklist is provided in the Supplementary Table 1. A scoping review approach was selected because the available evidence is heterogeneous in terms of study design, scoring systems, endpoints, and analytical methods. The literature consists largely of methodological studies, observational cohorts, and emerging AI-based approaches rather than a uniform body of comparative trials suitable for quantitative synthesis. Accordingly, this review aimed to map existing methodologies, summarize current evidence, and identify knowledge gaps. The review protocol was not prospectively registered. 1. Literature search strategy A comprehensive literature search was conducted in PubMed/MEDLINE and the Cochrane Central Register of Controlled Trials (CENTRAL). On January 17, 2026, we performed a search of these databases for studies published between January 1, 2000, and December 31, 2025 (Supplementary Table 2). The search strategy was designed to identify studies evaluating endoscopic assessment of disease activity, severity, and extent; prognostic outcomes; and AI-based approaches in patients with UC. Controlled vocabulary terms, where applicable, were combined with free-text keywords related to UC, endoscopy or colonoscopy, endoscopic scoring systems, disease activity or severity, extent of inflammation, relapse or prognostic outcomes, and AI, including machine learning and deep learning. To maximize the search sensitivity, no restrictions on the study design were applied. 2. Study selection All retrieved records were imported into reference management software, and duplicate records were removed. Two independent reviewers (Y.M. and T.K.) screened titles and abstracts for eligibility. Full-text articles were retrieved and assessed when the relevance could not be determined from the abstract alone. Any disagreements regarding study eligibility were resolved through discussion and in consultation with a third reviewer (N.O.) when necessary. Reference lists of included studies were manually screened to identify any further relevant publications. 3. Eligibility criteria Studies were considered eligible if they included patients diagnosed with UC and evaluated endoscopic disease activity, severity, or extent using either conventional endoscopic scoring systems or AI-based approaches. Eligible studies were required to assess clinically relevant outcomes, such as disease relapse, prognosis, treatment failure, or disease monitoring. Only original articles published in peer-reviewed journals were included. Studies were excluded if they focused exclusively on Crohn’s disease or indeterminate colitis; were published as case reports, editorials, letters, or conference abstracts lacking sufficient data; or were not original articles. Animal and in vitro studies were also excluded. 4. Data extraction and synthesis Data were extracted using a standardized data collection form, which included study design, population characteristics, endoscopic assessment methods, use of AI, evaluated outcomes, and key findings. Data on the diagnostic performance of endoscopic assessment in relation to disease extent were extracted, and contingency tables were constructed to include sensitivity, specificity, positive predictive value, and negative predictive value. When a study reported multiple contingency tables for the same or different algorithms, each was treated as an independent dataset. Given the heterogeneity of study designs, assessment methods, and outcome measures, the results were synthesized narratively. A formal assessment of study quality was not performed because of the methodological nature of the included studies. 5. Ethical considerations This study did not involve the direct participation of human subjects or animals. Therefore, informed consent, institutional review board approval, and ethical clearance were not required. RESULTS 1. Search results and study selection The systematic literature search identified 4,328 records from the PubMed/MEDLINE and CENTRAL databases. After the removal of 1,596 duplicates, 2,732 records were screened based on titles and abstracts. Full-text review was performed for 103 articles, of which 82 were excluded because of ineligibility. The remaining 21 studies met the inclusion criteria and were included in the qualitative synthesis. The study selection process is summarized in a flow diagram (Fig. 1). Of the included studies, eight described the development or validation of endoscopic indices incorporating disease extent (Table 1), 11 evaluated the association between extent-based endoscopic assessment and clinical outcomes, including relapse prediction and treatment response (Table 2), and four investigated AI-assisted endoscopic assessment incorporating disease extent (Table 3). Because some studies addressed multiple research objectives, two studies were included in more than one category. 2. Data synthesis Considerable variations in the study populations, modalities, comparisons, and outcomes (definitions of histologic remission and clinical relapse) were identified among the included studies. A quantitative meta-analysis could not be conducted because of this heterogeneity. 3. Development and validation of endoscopic scores considering disease extent A total of seven endoscopic scores that take disease extent into account were used to assess the inflammatory burden in patients with UC. For segmental severity assessment, four scores were based on MES. The remaining three scores used their originally predefined endoscopic variables (Table 1).7,17,18 The Ulcerative Colitis Colonoscopic Index of Severity (UCCIS) reported by Thia et al.17 is calculated as a weighted linear sum of four endoscopic mucosal features (vascular pattern, granularity, ulceration, and bleeding/friability) scored across five colonic segments. Segmental scores are summed and multiplied by regression-derived coefficients: UCCIS = 3.1×Σ(Vascular pattern) + 3.6×Σ(Granularity) + 3.5×Σ(Ulceration) + 2.5×Σ(Bleeding/Friability) Scores range from 0 to 162, with higher scores indicating greater colon-wide endoscopic severity. In a prospective validation of 50 patients using blinded ratings by eight expert endoscopists, moderate correlations were observed with Simple Clinical Colitis Activity Index (r=0.62, p<0.0001)19 and Clinical Activity Index (r=0.52, p<0.001),20,21 as well as with C-reactive protein (CRP) (r=0.56), albumin (r=–0.55), and hemoglobin (r=–0.39).22 The Modified MES (MMES) developed by Lobatón et al.5 is calculated by summing the MES across five colonic segments (ascending, transverse, descending, sigmoid colon, and rectum). The Extended Modified Score is obtained by multiplying this sum by the maximal extent of inflammation. The MMES is then calculated by dividing the Extended Modified Score by the number of segments with active inflammation. In a multicenter prospective study including 171 patients with UC,23 the MMES showed significant correlations with the Partial Mayo Score (r=0.535), CRP (r=0.238), fecal calprotectin (FCP) (r=0.730), and the histological Geboes score (r=0.615) (all p<0.001). The Ulcerative Colitis Segmental Endoscopic Index (UCSEI) proposed by Suzuki et al.18 is calculated as the sum of severity subscores in each of five colonic segments (cecum/ascending colon, transverse colon, descending colon, sigmoid colon, and rectum). Disease severity is evaluated based on four endoscopic parameters (erythema, vascular pattern, friability, and ulceration), each graded on a scale from 0 to 10. To date, no studies have compared UCSEI with other indices of inflammatory disease activity. The Sum of Mayo Endoscopic Subscores (S-MES) reported by Kawashima et al.24 is calculated by summing the MES values across five colonic segments (ascending colon, transverse colon, descending colon, sigmoid colon, and rectum). In an analysis of 136 complete colonoscopies from 102 patients with UC, S-MES showed a stronger correlation with FCP (r=0.86, p<0.001) compared with MES (r=0.79, p<0.001).24 The Pancolonic Modified Mayo Score (PanMayo) reported by Bálint et al.25 is calculated by summing the subscores of MES across five colorectal segments. When active inflammation is present (MES ≥2 in any segment), the total score is multiplied by a factor of 3, yielding a range of scores from 0 (normal) to 45 (most severe). This score showed significant correlations with disease extent (r=0.30, p<0.001) and the Partial Mayo Score (r=0.70, p<0.001). In addition, PanMayo demonstrated a strong association with histological disease activity, as assessed by the histological Riley Score (r=0.70, p<0.001).26 Furthermore, PanMayo correlated with CRP (r=0.50, p<0.001), fecal matrix metalloproteinase-9 (r=0.50, p<0.001), and FCP (r=0.35, p=0.043), as well as with leukocyte count (r=0.30, p=0.009) and platelet count (r=0.30, p=0.003). Negative correlations were observed with hemoglobin (r=−0.30, p=0.001), hematocrit (r=−0.30, p=0.004), and serum iron levels (r=−0.50, p<0.001). The Degree of Ulcerative Colitis Burden of Luminal Inflammation (DUBLIN) score developed by Rowan et al.6 is calculated as the product of the MES and disease extent (E1–E3), as defined by the Montreal classification.27 In a cohort of 70 patients with UC, the DUBLIN score showed a significant correlation with FCP (r=0.394, p<0.01). Receiver operating characteristic analysis using an FCP threshold of >50 µg/g demonstrated a higher area under the curve (AUC) for the DUBLIN score (AUC=0.76) compared with the MES (AUC=0.73). The Toronto IBD Global Endoscopic Reporting (TIGER) score proposed by Zittan et al.7 is calculated by evaluating five colonic segments (ascending, transverse, and descending colon; sigmoid colon; and rectum). In each segment, five predefined endoscopic variables are scored: general mucosal appearance, ulcer or erosion size, percentage of ulcerated surface, percentage of affected surface, and degree of narrowing. Segmental scores are summed, and segments with moderate-to-severe activity (segmental score ≥5) receive an additional 100-point bonus per segment. The total TIGER score reflects the overall endoscopic disease severity and extent. Overall, the per-segment agreement between the MES and the TIGER score was 84%, with a kappa (κ) coefficient of 0.60. In addition, the TIGER score showed a significant correlation with FCP (p<0.0001). 4. Endoscopic assessment considering disease extent to predict future clinical relapse and assess the response to treatment Eleven studies evaluated the prognostic value of endoscopic assessment considering disease extent in UC. Two were single-center, open-label, prospective studies, and nine were retrospective (Table 2). In 157 UC patients in clinical remission, S-MES and UCCIS showed comparable accuracy for predicting relapse over 24 months (AUC: 0.718 vs 0.727). Relapse rates were significantly higher in patients with S-MES ≥1 or UCCIS ≥9.8.28 Silva et al.29 conducted a retrospective cohort study of 107 patients undergoing total colonoscopy with serial biopsies to assess the association between endoscopic and histological activity and the therapeutic failure rate. The AUCs were 0.74 for MES, 0.78 for the DUBLIN score, and 0.84 for the Nancy histological index,30 indicating significant predictive value for all indices. Another retrospective study included 110 patients with UC, of whom 87 completed 12 months of follow-up.31 DUBLIN scores were significantly higher in patients with moderate-to-severe disease than in those with mild disease (p<0.05), but no association was observed between the DUBLIN score and long-term remission. Bacsur et al.32 compared the prognostic utility of the PanMayo and DUBLIN scores. At baseline, 157 patients (62.8%) were in clinical remission. Higher PanMayo scores were associated with an increased risk of clinical flare, both for any mucosal inflammation (>0; p=0.001) and for extensive severe disease (>12; p=0.003). A PanMayo score >12 was associated with the initiation of biologic therapy, biologic dose escalation, and inflammatory bowel disease-related hospitalization, while any PanMayo activity (>0) predicted subsequent systemic corticosteroid use. Similarly, a baseline DUBLIN score >0 was associated with an increased risk of clinical flare (p=0.005). Higher DUBLIN scores were associated with the initiation of biologic therapy, inflammatory bowel disease-related hospitalization, and corticosteroid requirement, indicating that both indices capture clinically meaningful inflammatory burden, albeit with different thresholds. The prognostic performance of MMES and the DUBLIN score was compared in a retrospective cohort of 300 UC patients.33 All indices correlated with inflammatory biomarkers. The DUBLIN score showed superior discrimination for UC-related admission and medication escalation, while no differences were observed for predicting acute severe UC, disease extension, or surgery. Across predefined cutoffs, higher scores consistently predicted acute severe UC, with the DUBLIN score showing the strongest association. In a retrospective analysis of 166 patients, the TIGER score outperformed UCEIS and MES in predicting recent advanced therapy (AUC=0.848) and 1-year readmission (AUC=0.700).34 A TIGER score ≥317 was independently associated with the need for advanced UC therapy. Consistently, in a subsequent prospective study of 53 patients, those with baseline TIGER scores ≥100 were significantly more likely to require corticosteroids or advanced therapy and to develop steroid dependence or refractoriness within 12 months.7 Treatment responsiveness was assessed in a multicenter study of 51 patients receiving mesalazine.18 Both the Ulcerative Colitis Disease Activity Index35 and the UCSEI improved significantly over 8 weeks, with the latter showing a 57.6% reduction, supporting the utility of extent-based endoscopic indices in monitoring therapeutic response. In a retrospective cohort of 178 hospitalized UC patients, the DUBLIN score correlated strongly with endoscopic disease severity, increasing across mild, moderate, and severe categories.36 However, UCEIS demonstrated superior predictive accuracy for treatment escalation and 1-year readmission, whereas MES showed limited prognostic utility. Only UCEIS differentiated steroid-responsive from steroid-dependent or -refractory disease. Three studies evaluated the prognostic significance of endoscopic disease extent in UC patients in clinical remission by assessing inflammatory extent based on established endoscopic scoring systems without introducing novel scoring tools. Asonuma et al.37 conducted a multicenter retrospective study of 161 patients with extensive UC in clinical remission, comparing complete endoscopic remission with residual short-segment distal inflammation. Over a 2-year follow-up, residual distal inflammation was associated with a higher rate of minor relapse but did not increase the risk of major relapse requiring treatment escalation or hospitalization. These findings suggest that limited distal inflammation may not necessitate treatment intensification when proximal mucosal healing is achieved. Similarly, Shimodaira et al.38 retrospectively analyzed 83 UC patients undergoing total colonoscopy and focused on small mucosal lesions (≤3 cm) equivalent to an MES of 0. Clinical relapse occurred more frequently in patients with MES 1, whereas relapse rates were comparable between patients with MES 0 and those with MES 0 plus a low level of mucosal inflammation. Multivariable analysis confirmed that limited residual inflammation did not independently predict relapse, supporting the concept that minimal inflammatory burden may be clinically acceptable. Kaneshiro et al.39 prospectively evaluated 325 UC patients in clinical remission using three endoscopic strategies: original UCEIS, worst-affected UCEIS, and pancolonic UCEIS. Pancolonic assessment demonstrated the highest predictive accuracy for 1-year clinical relapse (AUC=0.852). A pancolonic UCEIS >3 and histological activity (Geboes score >3.0) were independent risk factors for relapse, and their combination achieved high sensitivity and specificity for relapse prediction. This study highlights the value of whole-colon assessment in risk stratification during remission. Overall, it remains controversial whether extent-based indices that capture the cumulative inflammatory burden of the colon outperform or merely complement conventional focal severity scores in predicting clinical relapse, treatment escalation, hospitalization, and steroid dependence. Importantly, the available evidence remains heterogeneous across studies, with differences in study design, patient populations, endoscopic scoring methods, outcome definitions, and follow-up duration, which may partly explain the inconsistent findings reported across studies. Nevertheless, these indices appear to reflect therapeutic responsiveness and may be useful for prognostication and longitudinal disease monitoring. Further large-scale prospective studies are warranted to confirm these observations. 5. AI-assisted endoscopic assessment considering disease extent Four studies evaluated the utility of AI-assisted endoscopic assessment considering disease extent in UC (Table 3). One study was based on clinical practice data, and three were based on clinical trials. Fan et al.40 introduced a new AI-driven scoring system that incorporates the distribution of inflammation from video-based assessments. In their approach, the colon was partitioned into 85 predefined areas (cecum: 20 areas; transverse colon: 20; descending colon: 20; sigmoid colon: 15; rectum: 10), and the scoring system automatically evaluated the inflammatory severity in these areas for each video and produced a comprehensive visual output depicting overall intestinal inflammatory activity. Focusing on automated and quantitative measures of the global endoscopic disease severity and CDS, Stidham et al.41 introduced an AI-assisted CDS. The CDS is determined by summing the squared MES values at 50 evenly distributed segments in the left colon, including the rectum, and consequently, it concentrates on left-sided disease but does not evaluate the entire colon. Within the UNIFI clinical trial cohort, comprising 748 induction and 348 maintenance UC patients, the CDS correlated significantly with both the MES (p<0.001) and all clinical components of the Partial Mayo Score (p<0.001). When patients were stratified according to their baseline CDS, ustekinumab showed superior efficacy compared with placebo (p<0.001), with a more pronounced effect in severe disease relative to mild disease (–85.0 vs –55.4; p<0.001). Although the AI-derived CDS proposed by Stidham et al.41 was designed to provide a quantitative measure of cumulative inflammatory burden, it should be noted that this approach was limited to the left colon, including the rectum, and did not capture inflammation throughout the entire colon. Consequently, its applicability to patients with extensive or right-sided disease may be inherently constrained. In a separate large-scale study,42 Stidham et al. applied a similar AI-assisted cumulative assessment framework to endoscopic video data from the phase 3 TrueNorth trial of ozanimod.43 Despite the continued focus on left-sided colonic evaluation, this study demonstrated that the CDS showed stronger agreement with symptomatic remission and health-related quality of life than the conventional MES. Importantly, even among patients with moderate-to-severe endoscopic activity by MES, lower CDS values were associated with improved patient-reported outcomes, reinforcing the clinical relevance of cumulative disease burden beyond maximal focal severity. Similarly, Gutierrez-Becker et al.44 developed and evaluated the UC Severity Classification and Localized Extent system, which adopted a camera localization algorithm that maps each frame to a specific location within the colon. UC Severity Classification and Localized Extent was trained and validated using 4,326 sigmoidoscopy videos from phase III etrolizumab clinical trials and yielded high concordance with central readings at the colon section level (κ=0.80), and strong agreement between central and local readings (κ=0.84), thereby indicating that its inter-rater reliability is comparable to that of experienced readers. 6. Current limitations of endoscopic assessment considering disease extent Despite growing interest in endoscopic indices incorporating disease extent in UC, several limitations hinder their widespread adoption and definitive positioning in clinical practice. First, substantial heterogeneity exists across the proposed indices with respect to conceptual frameworks, scoring algorithms, colonic segmentation strategies, and the weighting of severity versus extent. Some indices rely on a simple summation of the segmental MES,5,6 whereas others apply regression-derived coefficients, multiplicative extent adjustments, or bonus-point systems.7,18 This methodological diversity complicates direct comparison between scores and precludes harmonization across studies. Second, most available evidence derives from single-center observational cohorts with relatively small sample sizes and variable follow-up durations. Randomized controlled trials incorporating extent-based endoscopic indices as primary or secondary endpoints are lacking. In addition, definitions of clinical outcomes—such as relapse, treatment escalation, or hospitalization—vary considerably across studies, further limiting generalizability and preventing robust quantitative synthesis. External validation across independent cohorts and standardized outcome frameworks therefore remains insufficient for most indices. Third, although extent-based indices often correlate more strongly with FCP, histological activity, and systemic inflammatory markers than focal severity scores, their incremental prognostic value over established indices such as MES or UCEIS remains inconsistent. Some studies report improved prediction of relapse or treatment escalation,37,39 whereas others report comparable or even inferior performance.29,33,36 These discrepancies likely reflect differences in patient populations, baseline disease severity, therapeutic context, and analytical thresholds. Finally, practical considerations may limit routine clinical implementation. Several extent-inclusive indices require detailed segmental scoring and manual calculations, which can be time-consuming and may reduce reproducibility outside specialized centers. Simplified yet robust scoring systems that balance comprehensiveness with feasibility remain an unmet need. 7. Opportunities and challenges of AI-assisted assessment AI-assisted approaches may help address several of these limitations while introducing new challenges (Fig. 2). Early studies suggest that AI-based systems can provide automated, quantitative evaluation of inflammatory distribution and global disease burden, potentially improving reproducibility and reducing interobserver variability. However, important barriers remain. Most current AI models have been trained on selected datasets derived from clinical trials or single institutions, and their external validity across different endoscopic platforms and real-world practice settings remains uncertain.15 In addition, the “black box” nature of some algorithms raises concerns regarding interpretability, regulatory approval, and clinician acceptance.45,46 Despite these challenges, AI-assisted assessment may offer an opportunity to integrate both inflammatory severity and spatial distribution into a single quantitative framework. Such approaches could complement existing treat-to-target strategies, which currently rely primarily on focal severity-based indices, by enabling a more comprehensive assessment of cumulative inflammatory burden. Preliminary evidence supports this concept. In our recent work, automated video-based analysis using a deep learning–based system (QUAD) quantified both inflammatory severity and spatial distribution from full-length colonoscopy recordings in patients with UC in clinical remission.47 Relapse prediction improved when the analysis focused on distal colonic segments rather than whole-colon assessment, suggesting that regional inflammatory burden may carry prognostic information beyond maximal focal severity. Although derived from a single prospective cohort and requiring external validation across diverse populations and endoscopic platforms, these findings highlight the potential value of integrating spatial inflammatory metrics into endoscopic evaluation. Looking ahead, future research should prioritize standardized, multicenter prospective studies directly comparing extent-based indices, conventional severity scores, and AI-assisted approaches within unified outcome frameworks. Integration of spatial endoscopic assessment with histological findings, biomarkers, and patient-reported outcomes may enable more precise treat-to-target strategies.48-51 Importantly, the goal is not to increase scoring complexity but to develop reliable, interpretable, and clinically actionable measures of cumulative inflammatory burden that can be seamlessly incorporated into routine clinical workflows. However, given the methodological heterogeneity and limited external validation across existing studies, the clinical role of AI-assisted extent-based assessment should be interpreted cautiously and requires confirmation in large multicenter prospective studies. 8. Future directions toward the standardization of extent-aware endoscopic indices Despite increasing interest in extent-aware endoscopic assessment, the lack of standardized segmentation frameworks, weighting strategies, and clinical endpoints currently limits cross-study comparability and clinical implementation. To facilitate generalization and eventual standardization, several key research priorities should be addressed. First, consensus is needed regarding a practical and reproducible segmentation framework for assessing disease extent. Current studies employ different segment definitions and scoring schemes, which limits cross-study comparability. Establishing standardized segmental assessment—potentially aligned with routine colonoscopic reporting—would improve consistency across studies. Second, further methodological work is required to determine how the severity and spatial extent of inflammation should be integrated. Existing indices differ in how segmental scores are aggregated or weighted, and it remains unclear which approach best reflects cumulative inflammatory burden. Comparative studies evaluating alternative weighting strategies may help identify models that most accurately predict clinically meaningful outcomes. Third, candidate extent-aware indices require prospective validation in large multicenter cohorts using standardized endpoints. In particular, their ability to predict outcomes such as relapse, hospitalization, need for corticosteroids, biologic escalation, or colectomy should be evaluated. Such validation would help determine whether extent-integrated scoring provides incremental prognostic value beyond conventional indices focused on the worst segment. Finally, emerging AI-based endoscopic analyses may offer an opportunity to objectively quantify both severity and disease distribution across the entire colon. AI-assisted pancolonic assessment could enable automated calculation of extent-integrated scores and facilitate their implementation in both clinical trials and routine practice. Addressing these priorities will be essential for translating the concept of extent-aware endoscopic assessment from methodological proposals into standardized tools applicable to clinical research and patient management. Ultimately, international consensus efforts, similar to those that led to the development of UCEIS, may be required to establish standardized extent-integrated endoscopic scoring systems. CONCLUSIONS Endoscopic indices that incorporate disease extent provide a more holistic assessment of cumulative inflammatory burden in UC than focal severity scores alone. Emerging evidence suggests that these approaches can improve correlations with biomarkers, capture treatment responsiveness, and refine risk stratification for relapse and disease progression. However, the available evidence remains heterogeneous, and most indices lack large-scale prospective validation. Advances in AI-assisted endoscopic analysis may facilitate objective, whole-colon assessment of inflammatory distribution and support the development of standardized extent-integrated scoring systems. Future multicenter prospective studies will be essential to determine whether incorporating disease extent can meaningfully improve clinical decision-making and long-term disease management in UC.