Background: Several artificial intelligence systems based on large language models (LLMs) have been commercially developed, with recent interest in integrating them for clinical questions. Recent versions now include image analysis capacity, but their performance in gastroenterology remains untested. This study assesses ChatGPT-4’s performance in interpreting gastroenterology images. Methods: A total of 740 images from five procedures—capsule endoscopy (CE), device-assisted enteroscopy (DAE), endoscopic ultrasound (EUS), digital single-operator cholangioscopy (DSOC), and high-resolution anoscopy (HRA)—were included and analyzed by ChatGPT-4 using a predefined prompt for each. ChatGPT-4 predictions were compared to gold standard diagnoses. Statistical analyses included accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the curve (AUC). Results: For CE, ChatGPT-4 demonstrated accuracies ranging from 50.0% to 90.0%, with AUCs of 0.50–0.90. For DAE, the model demonstrated an accuracy of 67.0% (AUC 0.670). For EUS, the system showed AUCs of 0.488 and 0.550 for the differentiation between pancreatic cystic and solid lesions, respectively. The LLM differentiated benign from malignant biliary strictures with an AUC of 0.550. For HRA, ChatGPT-4 showed an overall accuracy between 47.5% and 67.5%. Conclusions: ChatGPT-4 demonstrated suboptimal diagnostic accuracies for image interpretation across several gastroenterology techniques, highlighting the need for continuous improvement before clinical adoption.

High-resolution anoscopy (HRA) plays a central role in the detection and treatment of precursors of anal squamous cell carcinoma (ASCC). Artificial intelligence (AI) algorithms have shown high levels of efficiency in detecting and differentiating HSIL from low-grade squamous intraepithelial lesions (LSIL) in HRA images. Our aim was to develop a deep learning system for the automatic detection and differentiation of HSIL versus LSIL using HRA images from both conventional and digital proctoscopes. A convolutional neural network (CNN) was developed based on 151 HRA exams performed at two volume centers using conventional and digital HRA systems. A total of 57,822 images were included, 28,874 images containing HSIL and 28,948 LSIL. Partial subanalyses were performed to evaluate the performance of the CNN in the subset of images acetic acid and lugol iodine staining and after treatment of the anal canal. The overall accuracy of the CNN in distinguishing HSIL from LSIL during the testing stage was 94.6%. The algorithm had an overall sensitivity and specificity of 93.6% and 95.7%, respectively (AUC 0.97). For staining with acetic acid, HSIL was differentiated from LSIL with an overall accuracy of 96.4%, while for lugol and after therapeutic manipulation, these values were 96.6% and 99.3%, respectively. The introduction of AI algorithms to HRA may enhance the early diagnosis of ASCC precursors, and this system was shown to perform adequately across conventional and digital HRA interfaces.

Aim: to evaluate the presence of subclinical HPV-induced anal lesions with anal cytology, High-Resolution Anoscopy (HRA) and HPV genotyping by polymerase chain reaction (PCR) in the follow-up of treated condylomata acuminata (CA). Methods: seventy-nine male patients were included. One month after anal CA eradication, the patients underwent brush samples collection for anal cytology and PCR, and HRA with biopsy of acetowhite lesions. These methods were compared within all patients and between groups, according to Human Immunodeficiency Virus (HIV) infection status: HIV-negative; HIV-positive with TCD4 count above and below 350 cells/mm3. Results: the most frequent HPV types were 6 and 16. HPV DNA was isolated in 92%. HIV infection was associated with a higher number of oncogenic HPV types (p=0.038). All patients with negative PCR had negative HRA and cytology. There were no differences in cytological, HRA or histopathological findings between groups. Conclusion: the association of the findings of cytopathology, HRA and genotyping of HPV refined the diagnosis of HPV-induced lesions. The degree of immunodeficiency was not associated with increase in remnant HPV-induced anal lesions.

Objective To compare proctological diseases in HIV-positive and -negative patients. Method Prospective study conducted in the proctology outpatient clinic at Instituto de Infectologia Emílio Ribas from 2013 through 2019. All of them underwent anamnesis and proctological examination and were divided into two groups (HIV+ and HIV−). The groups were then compared with regard to age, sex, diagnosis, and whether or not there was any indication for surgery. Results 485 HIV + patients (41.9%) (Group I) and 672 HIV− (58.1%) (Group II). The mean age in Group I was 42.8 years, and 404 (83.3%) of the patients were males. In Group II, the mean age was 48.3 years, and 396 (59%) patients were females. In comparing the groups, a statistically significant difference was found regarding sex, age, and diagnosis. The incidence of hemorrhoid and anal skin tags was higher in group II and that of condyloma acuminata was higher in in Group I (p < 0.001). In Group I, condyloma acuminata was more prevalent, but in those with CD4 levels above 500 cells/mm3, the diagnoses were more similar to those in the population without HIV. Conclusions Proctological diseases were found to be similar but have different incidences. The incidences of diseases in HIV− and + patients tend to approach each other as CD4 level increases.

The COVID-19 pandemic has changed the world health scenario, causing numerous problems related to the overload of services. In this scenario, the approach to oncological diseases becomes a challenge, considering the risk of progression of cancer disease and death due to delay in diagnosis and treatment. Faced with this exceptional situation, coloproctology services have been forced to change their routine to adapt to the new reality, considering risks and benefits in the conduct of these patients. Thus, the experience of the coloproctology service at the Santa Casa de São Paulo, São Paulo, state of São Paulo, Brazil, during the beginning of the COVID-19 pandemic is described, evaluating the possibility of maintaining elective oncological surgeries through selection by directed anamnesis. Method Retrospective cohort study with prospective collection comparing colorectal surgery for cancer before and during the pandemic from December 2019 to July 2020. Results In total, 81 patients were included. During the pandemic, 41 patients were operated on with 2 perioperative contaminations by COVID-19. Both length of stay and complications were not different between groups. Conclusion Since there was no increase in COVID-19 morbidity and mortality in elective oncology surgeries with patients screened for guided anamnesis, it is worth considering this method for maintaining surgical procedures even in the event of a pandemic.

Objectives To compare morphological abnormalities on anal colposcopy against histology to determine anal high-grade squamous intraepithelial lesions (HSILs). Methods This is a retrospective data assessment of HIV-negative and HIV-positive patients undergoing outpatient follow-up. The sample comprised 54 patients presenting acetowhite lesions on anal colposcopy. Acetowhite lesions were classified according to their morphology into punctation, verrucous, mosaic, ulcerated, or hypervascularized, and biopsies of these specimens were classified as anal HSIL, low-grade squamous intraepithelial lesion (LSIL), or normal. The data were analyzed using SPSS for Windows version 13.0 (SPSS Inc., Chicago, IL, USA). The results were analyzed using the nonparametric Mann-Whitney test, the Fisher exact test and the chi-squared parametric test. A 95% confidence interval (CI) was used and a level of significance < 5% was adopted for all statistical tests. Results Fifty-four patients (50 males, 80% HIV + ) with biopsied acetowhite lesions were assessed. There were 31 punctation lesions, 1 classified as HSIL (3.2%; 95%CI: 0–40.0), 17 verrucous lesions, 3 HSIL (17.7%; 95%CI: 0–10.7), and 1 ulcerated, classified as HSIL (100%), and 4 mosaic and 1 atypical vessel lesion, all classified as LSIL. The results showed no association of presence of anal HSIL with positivity for HIV infection or with counts above or below 500/µl in HIV+ patients. Statistical analysis was performed using the Mann-Whitney nonparametric test, the Fisher exact test, and the chi-squared parametric test. Conclusion The comparison of morphological findings on anal colposcopy against histology revealed no morphological pattern suggesting anal HSIL.