Search Results (37)

Background/Objectives: This study evaluates the performance of a large language model (lLm) in providing geographically tailored colorectal cancer screening and surveillance recommendations to gastrointestinal surgeons. Methods: Fifty-four patient cases, varying by age and family history, were developed based on colorectal cancer guidelines. Standardized prompts with predefined query terms were used to query ChatGPT-4.5 on 18 April 2025, from four locations: Canada, Italy, Romania, and the United Kingdom. Responses were classified as “Correct,” “Partially Correct,” or “Incorrect” based on clinical guidelines and expert recommendations for each country. Outcomes were analyzed using descriptive statistics. Results: ChatGPT provided recommendations on screening eligibility, test interpretation, the management of positive results, and surveillance intervals. Correct recommendations were given for 50.0% (27/54) of cases in Canada, 63.0% (34/54) of cases in Italy, 40.7% (22/54) of cases in Romania, and 55.6% (30/54) of cases in the United Kingdom. Queries in Italian yielded correct guidance for 64.8% (35/54) of cases, while Romanian queries were accurate for 40.7% (22/54) of cases. Notably, Romania and Italy lacked detailed guidelines for polyp management and post-test surveillance. A key finding was the inconsistency between ChatGPT-generated titles and corresponding recommendations, which may impact its reliability in clinical decision-making. Conclusions: ChatGPT-4.5’s performance varies by country and language, highlighting inconsistencies in geographically tailored recommendations. This study highlights limitations associated with the training data cutoff and the potential biases introduced by model-generated responses. Healthcare professionals should recognize these limitations and the possible gaps in guideline availability, particularly for high-risk screening, polyp management, and surveillance in certain European countries. Full article

Background: Constipation is one of the most common gastrointestinal complaints among women, with a variety of contributing factors. We aimed to assess the role of colonoscopy in evaluating young women with constipation. Methods: A multi-center, large cohort, retrospective study included all data from colonoscopies performed between 2016 and 2023 in seven endoscopy departments. The indications and findings of the procedures were collected, and findings of young women aged ≤40 y with constipation as an indication were compared to older women and men of the same age groups. Results: The cohort comprised 377,795 patients, including 198,629 (52.6%) females and 179,166 (47.4%) males. In total, 7872 females underwent colonoscopy for constipation and other indications (Cohort 1). In addition, 6852 women were referred for a colonoscopy for constipation only (Cohort 2). In sum, 75% of colonoscopies in women <40 y were normal in both cohorts. In Cohort 1, inflammatory bowel diseases (IBD) were significantly higher in women <40 y with Ulcerative Colitis (UC) (1.2%) and Crohn’s disease (CD) (0.7%), p < 0.001). The rate of IBD was lower but still significant in Cohort 2. In both cohorts, diverticulosis and polyp rates exponentially increased with age >40 y, p < 0.001. Higher rates of diverticulosis and polyps were found among males <40 y in Cohort 1. One case (0.1%) of Colorectal cancer (CRC) was found in <40 y women. Similar IBD and CRC rates were found in males and females of all ages, p > 0.05. Conclusions: The diagnostic yield of colonoscopy for investigating isolated constipation in young females is not significant. Diagnostic work-up should be guided by accurate clinical understanding. Full article

The lack of extensive, varied, and thoroughly annotated datasets impedes the advancement of artificial intelligence (AI) for medical applications, especially colorectal cancer detection. Models trained with limited diversity often display biases, especially when utilized on disadvantaged groups. Generative models (e.g., DALL-E 2, Vector-Quantized Generative Adversarial Network (VQ-GAN)) have been used to generate images but not colonoscopy data for intelligent data augmentation. This study developed an effective method for producing synthetic colonoscopy image data, which can be used to train advanced medical diagnostic models for robust colorectal cancer detection and treatment. Text-to-image synthesis was performed using fine-tuned Visual Large Language Models (LLMs). Stable Diffusion and DreamBooth Low-Rank Adaptation produce images that look authentic, with an average Inception score of 2.36 across three datasets. The validation accuracy of various classification models Big Transfer (BiT), Fixed Resolution Residual Next Generation Network (FixResNeXt), and Efficient Neural Network (EfficientNet) were 92%, 91%, and 86%, respectively. Vision Transformer (ViT) and Data-Efficient Image Transformers (DeiT) had an accuracy rate of 93%. Secondly, for the segmentation of polyps, the ground truth masks are generated using Segment Anything Model (SAM). Then, five segmentation models (U-Net, Pyramid Scene Parsing Network (PSNet), Feature Pyramid Network (FPN), Link Network (LinkNet), and Multi-scale Attention Network (MANet)) were adopted. FPN produced excellent results, with an Intersection Over Union (IoU) of $0.64$, an F1 score of $0.78$, a recall of $0.75$, and a Dice coefficient of 0.77. This demonstrates strong performance in terms of both segmentation accuracy and overlap metrics, with particularly robust results in balanced detection capability as shown by the high F1 score and Dice coefficient. This highlights how AI-generated medical images can improve colonoscopy analysis, which is critical for early colorectal cancer detection. Full article

Colorectal cancer (CRC) prevention depends on effective colonoscopy; yet variability in adenoma detection rates (ADRs) and missed lesions remain significant hurdles. Artificial intelligence-powered computer-aided detection (CADe) systems offer promising advancements in enhancing polyp detection. This review examines the role of CADe in improving ADR and reducing adenoma miss rates (AMRs) while addressing its broader clinical implications. CADe has demonstrated consistent improvements in ADRs and AMRs; largely by detecting diminutive polyps, but shows limited efficacy in identifying advanced adenomas or sessile serrated lesions. Challenges such as operator deskilling and the need for enhanced algorithms persist. Combining CADe with adjunctive techniques has shown potential for further optimizing performance. While CADe has standardized detection quality; its long-term impact on CRC incidence and mortality remains inconclusive. Future research should focus on refining CADe technology and assessing its effectiveness in reducing the global burden of CRC. Full article

Background/Objectives: Endoscopic resection with lift polypectomy using submucosal injection (SI) for large non-pedunculated colorectal polyps is recommended to facilitate complete mucosal resection and decrease the risk of perforation; however, there are no studies comparing the safety and efficacy of large polypectomies with and without lift polypectomy. We aimed to evaluate the feasibility and safety of the polypectomy technique without SI compared to the routine use of SI. Methods: We performed a single tertiary center retrospective study evaluating all consecutive large non-pedunculated colorectal polyps (≥20 mm) referred to expert endoscopists in polypectomy from 2018 through 2021. We collected and analyzed data on demographics, polyp characteristics, resection technique, complications, and polyp recurrence in 6–12 months. Results: In 200 large non-pedunculated colonic polyp resections, 110 (55%) were performed with SI. The median polyp size was slightly larger in SI (30 mm IQR [20, 40] vs. 25 mm IQR [20, 30] in no-SI, p < 0.05), with a range of 20 to 130 mm. There were no differences in delayed bleeding rates. No perforation was noted in the no-SI group, and five perforations occurred in the SI group, without a statistically significant difference between groups. There was no statistically significant difference in the polyp recurrence rate at 6–12 months between the no-SI and SI groups (12% vs. 8% in no-SI vs. SI, respectively, p = 0.48). Conclusions: Complete removal of large non-pedunculated polyps without SI is feasible and safe in this large series. This approach had similar rates of clinically significant post-polypectomy bleeding and a non-significant difference in perforation rate compared to polyp resection with routine use of SI. Randomized trials are warranted to further assess the safety and efficacy of this approach. Full article

Inflammatory bowel diseases (IBDs), encompassing Ulcerative Colitis (UC) and Crohn’s Disease (CD), are chronic inflammatory disorders affecting the gastrointestinal tract. The association between IBD and colorectal cancer (CRC) is well-documented. Multiple factors have been identified as contributors to the risk of developing CRC in patients with IBD, including duration of disease, disease extension, family history of CRC, co-existance of primary sclerosing cholangitis (PSC), and potentially the presence of post-inflammatory polyps (PIPs). PIPs, often referred to as pseudopolyps, are polypoid structures that emerge as a result of severe mucosal inflammation. While their presence has been linked to greater disease severity, the role of PIPs in increasing CRC risk remains controversial. Increasing evidence suggests an association between post-inflammatory polyps (PIPs) and the risk of colorectal neoplasia, with PIPs potentially serving as an indicator of this risk through a history of enhanced inflammation. PIPs may also be linked to a distinct patient phenotype, including the presence of other known risk factors. More recent studies suggest that the risk burden (characterized by a high number or by large polyps) may be important. However, the evidence remains inconsistent, with some studies showing no clear association between PIPs and CRC risk after adjusting for other factors, including histological inflammation. In contrast, the data suggest a low rate of malignant transformation of the PIPs themselves. This narrative review aims to summarize the latest evidence regarding the relationship between PIPs and CRC in IBD, with a focus on UC. While some studies suggest that PIPs may serve as markers of higher disease severity and inflammation, their direct contribution to CRC risk remains unclear. Further research is needed to explore the inflammatory and carcinogenic pathways in patients with PIPs to better understand their role in colorectal cancer (CRC) development. Full article

(1) Background: Familial adenomatous polyposis (FAP) is a hereditary condition characterized by the development of numerous adenomas in the large intestine, often necessitating colectomy due to an elevated risk of colorectal cancer. Despite surgical intervention, adenomas frequently recur, underscoring the importance of ongoing surveillance. This study evaluates the outcomes of a 12-year endoscopic follow-up after colectomy and gastrointestinal reconstruction for FAP. (2) Methods: A retrospective analysis was conducted on 41 FAP patients who underwent at least one postoperative endoscopic examination. Assessments of the pouch or rectum were performed every 12–18 months following ileorectal anastomosis and every 18–24 months after ileal pouch–anal anastomosis. Follow-up biopsies were assessed using the adopted Spigelman classification. (3) Results: Postoperative pathology revealed invasive colorectal cancer in three patients. Abdominoperineal resection was performed in two cases due to secondary invasive carcinoma, and one T1 tumor was radically removed with ESD. One patient underwent radical pouch excision following a nodal pelvic recurrence of rectal cancer. Over a 12-year observation period, the mean Spigelman score increased by 2 points, and the proportion of patients with low-grade polypoid lesions decreased. The quantity or size of polyps increased in 24 patients, decreased in 8 patients, and remained stable in 9 patients. In four patients, granular, laterally spreading tumors were discovered in the rectal stump. (4) Conclusions: Regular endoscopic surveillance in FAP patients facilitates early identification of neoplastic and inflammatory changes. The downstaging potential highlights the effectiveness of early interventions. While the Spigelman classification assessed polyps well, it did not predict cancer occurrence. A notable number of patients had invasive cancer at the time of surgery, underscoring the importance of early surgical qualification, which is particularly crucial for identifying upstaging or secondary cancer. Full article

Background. The large number of lesions detected via high-definition (HD) imaging during colonoscopy calls for the reliable real-time histological characterization of polyps, especially diminutive and small ones, to permit tailored management based on the neoplastic risk, such as a “resect-and-discard” or a “diagnose-and-leave” strategy for low-risk adenomas and hyperplastic polyps (HPs). The Kudo classification of glandular pit pattern is currently used for predicting polyp histology. Aim. The aim in this study was to assess whether Kudo’s glandular pit pattern, assessed via HD digital chromoendoscopy (i-Scan) without magnification and optical enhancement, reliably predicts polyp histology and differentiates neoplastic lesions (NLs) from non-neoplastic lesions (non-NLs) during routine colonoscopy. Methods. Consecutive colorectal lesions recorded in a database over 12 months, with Kudo’s glandular pit pattern classification, were retrospectively compared with histology. The diagnostic accuracy and negative predictive value (NPV) for adenomatous histology of Kudo’s pit patterns were assessed separately for diminutive (≤5 mm) and small (6–9 mm) polyps, accordingly to the American Society for Gastrointestinal Endoscopy (ASGE) Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI), and in large (≥10 mm) lesions. Results. A total of 2230 lesions were recorded: 898 diminutive, 704 small, and 628 large. Kudo’s type II pit pattern was prevalent in diminutive polyps and recognized mostly in HPs (83.27%); it was also found in 38.8% of adenomas. In the right colon, Kudo’s type II pit pattern was prevalent in adenomas (70.04% vs. 20.74% in HPs); among the serrated lesions, it was evenly distributed between HPs and adenomas. Kudo’s type IIIL/IIIs/IV pit pattern was prevalent in NLs (61% vs. 8.37% of non-NLs) in diminutive polyps, evenly distributed between non-NLs and NLs in small polyps, and found only in NLs in large polyps. Kudo’s type Vi/Vn pit pattern correctly identified all but one adenocarcinoma. The NPV for adenomatous histology did not reach the recommended 90% PIVI threshold for differentiation between NLs and non-NLs in diminutive polyps showing Kudo’s type II pit pattern and in small polyps showing type IIIL/IIIs/IV pit pattern. Conclusions. Kudo’s pit pattern classification carried out with digital chromoendoscopy (i-Scan) during routine colonoscopy does not allow the reliable differentiation between non-NLs and NLs in diminutive and small polyps, so a “diagnose-and-leave” strategy for diminutive polyps may leave undetected adenomas, while a “resect-and-discard” strategy could miss lesions requiring closer follow-up. Full article

Background/Objectives: Colorectal cancer incidence in Bahrain occurs at a ratio of 13.4–18.8 per 100,000 persons after age standardization. This study aims to explore the relationship between colorectal cancer/abnormalities and different lifestyle factors. Secondly, it aims to explore the association between f-Hb levels, colonoscopy findings, and lifestyle factors in a FIT-positive population in Bahrain. Method: A retrospective cross-sectional study was performed for patients positive for FIT and who had a colonoscopy. Different dietary and demographic factors as well as f-Hb levels were assessed. Results: A total of 559 (M: 330; F: 229) subjects were enrolled in this study. Subjects with CRC had significantly higher f-Hb concentrations (median: 1269 μg/mg) when compared with subjects of other groups. Higher percentages of CRC as well as large and small polyps were recorded in males. However, there was no significant difference in f-Hb concentration between males and females (p = 0.90). Higher median levels were found for f-Hb in patients with Q3 (higher red meat consumption) compared to Q1 and Q2 in the category with CRC, despite there being no statistically significant differences among the groups (p = 0.742). Similar results for coffee consumption and f-Hb concentrations in the different groups have been recorded (p = 0.697). A higher quartile of red meat consumption was associated with an increase in CRC risk of 79.9%. Coffee consumption reflected a lower risk of CRC by −47% moving from Q1 to Q2, while BMI was found to be a risk factor (+44%) for CRC. Conclusion: This study highlighted that high f-Hb concentration can be used as a predictive biomarker of CRC. Full article

Ensuring precise segmentation of colorectal polyps holds critical importance in the early diagnosis and treatment of colorectal cancer. Nevertheless, existing deep learning-based segmentation methods are fully supervised, requiring extensive, precise, manual pixel-level annotation data, which leads to high annotation costs. Additionally, it remains challenging to train large-scale segmentation models when confronted with limited colonoscopy data. To address these issues, we introduce the general segmentation foundation model—the Segment Anything Model (SAM)—into the field of medical image segmentation. Fine-tuning the foundation model is an effective approach to tackle sample scarcity. However, current SAM fine-tuning techniques still rely on precise annotations. To overcome this limitation, we propose WSPolyp-SAM, a novel weakly supervised approach for colonoscopy polyp segmentation. WSPolyp-SAM utilizes weak annotations to guide SAM in generating segmentation masks, which are then treated as pseudo-labels to guide the fine-tuning of SAM, thereby reducing the dependence on precise annotation data. To improve the reliability and accuracy of pseudo-labels, we have designed a series of enhancement strategies to improve the quality of pseudo-labels and mitigate the negative impact of low-quality pseudo-labels. Experimental results on five medical image datasets demonstrate that WSPolyp-SAM outperforms current fully supervised mainstream polyp segmentation networks on the Kvasir-SEG, ColonDB, CVC-300, and ETIS datasets. Furthermore, by using different amounts of training data in weakly supervised and fully supervised experiments, it is found that weakly supervised fine-tuning can save 70% to 73% of annotation time costs compared to fully supervised fine-tuning. This study provides a new perspective on the combination of weakly supervised learning and SAM models, significantly reducing annotation time and offering insights for further development in the field of colonoscopy polyp segmentation. Full article

Colorectal cancer (CRC) outcomes in terms of incidence and mortality are significantly worse in African Americans than other Americans. While differences in primary preventions for neoplasia (diet, obesity remediation, aspirin prophylaxis) are being elucidated, genetic mutations affecting premalignant lesions and immune response mechanisms may possibly also explain the increased incidence and mortality, particularly from right-sided disease. Objective: Our team therefore examined colonic segments seeking to test the hypothesis that the immune response and somatic genetic profiles of the colonic anatomic segments may vary and thus account for variations in neoplasia risk among the various colonic segments revealing an antigenic relationship with precancerous lesions. The p87 antigenic field effect is recognized via Adnab-9 antibody immunohistochemistry to be significantly less in the right colon in African Americans, particularly in the cecum. Method: Since small high-grade dysplastic adenomas (SHiGDA) likely missed by CRC screening may progress to cancer, we used Ion Torrent™ sequencing of DNA extracted from four normal colonic segments (two left-sided and two right) of patients with SHiGDAs. We also contrasted unique mutational fields in one patient with a large HiGDA (APC with unique mutations) and one patient who prospectively developed a SHiGDA (JAK3). Result: The SHiGDA (small high-grade dysplastic polyp) patient was p87 negative for any extracted stool, saliva, or colonic effluent via ELISA (enzyme linked immunoadsorbant assay). Furthermore, mean values of expression in segments from the right colon were reduced with respect to the means obtained from the left segments in 233 patients evaluated for a p87 field effect. This has recently been shown to be the case in a large cohort of AA and Caucasian 2294 patients, possibly explaining the right-sided CRC disparity in African Americans and the subsequent increase in mortality. This field effect disparity is also true for two cancers contracted by the SHiGDa patient (lung and prostate). Conclusion: Thus, this pilot study suggests that the reduction in p87 in the right colon is possibly correlated with JAK3 mutations. If confirmed, JAK3 mutations, known to be associated with immune aberrations, may provide a mechanistic explanation for the lack of a p87 (protein 87 kilodaltons) field in some patients with HGD polyps who might benefit from possible intervention such as more intensive screening. Limited microbiome studies were also performed on two patients with familial cancer syndromes and these compared favorably with controls available from the literature. Full article

Automatically segmenting polyps from colonoscopy videos is crucial for developing computer-assisted diagnostic systems for colorectal cancer. Existing automatic polyp segmentation methods often struggle to fulfill the real-time demands of clinical applications due to their substantial parameter count and computational load, especially those based on Transformer architectures. To tackle these challenges, a novel lightweight long-range context fusion network, named LightCF-Net, is proposed in this paper. This network attempts to model long-range spatial dependencies while maintaining real-time performance, to better distinguish polyps from background noise and thus improve segmentation accuracy. A novel Fusion Attention Encoder (FAEncoder) is designed in the proposed network, which integrates Large Kernel Attention (LKA) and channel attention mechanisms to extract deep representational features of polyps and unearth long-range dependencies. Furthermore, a newly designed Visual Attention Mamba module (VAM) is added to the skip connections, modeling long-range context dependencies in the encoder-extracted features and reducing background noise interference through the attention mechanism. Finally, a Pyramid Split Attention module (PSA) is used in the bottleneck layer to extract richer multi-scale contextual features. The proposed method was thoroughly evaluated on four renowned polyp segmentation datasets: Kvasir-SEG, CVC-ClinicDB, BKAI-IGH, and ETIS. Experimental findings demonstrate that the proposed method delivers higher segmentation accuracy in less time, consistently outperforming the most advanced lightweight polyp segmentation networks. Full article

Colorectal cancer is one of the three most common cancers worldwide. Polyps are complex and have unclear boundaries, often leading to inaccurate boundary segmentation and missed detections. To address these challenges, we propose a boundary constraint multi-level feature aggregation framework called BMFA-Net to precisely segment polyps. The framework comprises four key modules. First, the parallel partial decoder is introduced to aggregate high-level features within the network to generate a globally informative semantic map serving as the initial guidance region for reverse erasing the foreground. Second, we propose an efficient atrous convolution attention module to effectively aggregate local and global contextual information over multiple levels. Additionally, a multi-level feature aggregation mechanism is designed and placed among the efficient atrous convolution attention modules to enable the network to capture a large amount of semantic structure while preserving intricate details. Finally, a boundary constraint reverse attention module is proposed to perform the boundary constraint while removing the foreground to improve the quality of boundary segmentation. Extensive experiments demonstrated the superiority and versatility of our framework compared with state-of-the-art methods; specifically, it achieved a mean Dice score of 0.922 on the CVC-300 dataset. Full article

Gastrointestinal (GI) tract disorders are a significant public health issue. They are becoming more common and can cause serious health problems and high healthcare costs. Small bowel tumours (SBTs) and colorectal cancer (CRC) are both becoming more prevalent, especially among younger adults. Early detection and removal of polyps (precursors of malignancy) is essential for prevention. Wireless Capsule Endoscopy (WCE) is a procedure that utilises swallowable camera devices that capture images of the GI tract. Because WCE generates a large number of images, automated polyp segmentation is crucial. This paper reviews computer-aided approaches to polyp detection using WCE imagery and evaluates them using a dataset of labelled anomalies and findings. The study focuses on YOLO-V8, an improved deep learning model, for polyp segmentation and finds that it performs better than existing methods, achieving high precision and recall. The present study underscores the potential of automated detection systems in improving GI polyp identification. Full article

Fecal calprotectin (FC) has been largely recognized as a surrogate marker of intestinal neutrophilic inflammation, very reliable in distinguishing between inflammatory bowel diseases and functional gastrointestinal (GI) disorders. Aging has been suggested to influence FC results and their diagnostic accuracy; however, no studies are specifically targeted on this focus. In a retrospective study, we evaluated the eventual age-differences of the diagnostic accuracy of FC in discriminating between organic-inflammatory GI diseases and functional GI disorders. In 573 younger and 172 older (≥65 years) subjects undergoing an FC assay, we found that the latter showed higher median FC values, 72 (25–260) µg/g vs. 47 (25–165) µg/g (p < 0.01). Younger patients were more commonly affected by IBDs, while colorectal cancer and high-risk polyps, infective colitis, and diverticular disease represented the most common findings in the older subgroup. However, the estimated optimum FC threshold in discriminating between organic-inflammatory GI diseases and functional GI disorders was quite similar between the two groups (109 μg/g for the younger subgroup and 98 μg/g for the older subgroup), maintaining a very high specificity. In conclusion, we show that FC also represents a very specific test for intestinal inflammation in older patients, at similar threshold levels to younger subjects. Full article