ChatGPT M.D.: Is there any room for generative AI in neurology?

ChatGPT, a general artificial intelligence, has been recognized as a powerful tool in scientific writing and programming but its use as a medical tool is largely overlooked. The general accessibility, rapid response time and comprehensive training database might enable ChatGPT to serve as a diagnostic augmentation tool in certain clinical settings. The diagnostic process in neurology is often challenging and complex. In certain time-sensitive scenarios, rapid evaluation and diagnostic decisions are needed, while in other cases clinicians are faced with rare disorders and atypical disease manifestations. Due to these factors, the diagnostic accuracy in neurology is often suboptimal. Here we evaluated whether ChatGPT can be utilized as a valuable and innovative diagnostic augmentation tool in various neurological settings. We used synthetic data generated by neurological experts to represent descriptive anamneses of patients with known neurology-related diseases, then the probability for an appropriate diagnosis made by ChatGPT was measured. To give clarity to the accuracy of the AI-determined diagnosis, all cases have been cross-validated by other experts and general medical doctors as well. We found that ChatGPT-determined diagnostic accuracy (ranging from 68.5% ± 3.28% to 83.83% ± 2.73%) can reach the accuracy of other experts (81.66% ± 2.02%), furthermore, it surpasses the probability of an appropriate diagnosis if the examiner is a general medical doctor (57.15% ± 2.64%). Our results showcase the efficacy of general artificial intelligence like ChatGPT as a diagnostic augmentation tool in medicine. In the future, AI-based supporting tools might be useful amendments in medical practice and help to improve the diagnostic process in neurology.

Comparison of risk of malignancy and predictive value of diagnostic categories defined by Papanicolaou Society of Cytopathology system and WHO reporting system for pancreaticobiliary cytopathology in solid pancreatic lesions.

Background: The standardized diagnostic categories defined by the World Health Organization (WHO) reporting system support the interdisciplinary interpretation of cytological findings in the management of pancreatic cancer. Objective: To compare this classification to the Papanicolaou Society of Cytopathology (PSC) system in terms of predictive value and risk of malignancy (ROM) in solid pancreatic lesions. Design: Retrospective cohort study. Methods: All consecutive patients with solid pancreatic lesions who underwent endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) sampling at the University of Szeged from 2014 to 2021 were retrospectively enrolled. The predictive value and ROM of cytological findings were determined with comparison to histologic outcome and/or clinical follow-up. Results: A total of 521 EUS-FNAs were performed with a malignancy rate of 81.76%. In both classification systems, the absolute ROM of "non-diagnostic," "negative for malignancy," "atypical," "suspicious for malignancy," and "malignant" categories were 48.2%, 2.3%, 78.1%, 100.0%, and 99.4%, respectively. Despite the heterogeneous nature of the "neoplastic: other" category of the PSC system, the absolute ROM for solid lesions was 100%. Pancreatic neoplasm: high-risk/grade category including only two endosonographically solid cases of high-grade intraductal papillary mucinous neoplasms showed 100% ROM. There were no differences between PSC and WHO systems in sensitivity, specificity, and negative and positive predictive values: excluding the "atypical" category, these were 99.7%, 95.6%, 97.7%, and 99.5%, respectively. The "atypical" category considered benign resulted in a higher decrease in validity and negative predictive value, compared to "atypical" considered true malignant (93.6% vs 97.7% and 65.8% vs 97.7%). Conclusion: For solid pancreatic lesions, the WHO system was identical to the PSC system in terms of ROM and predictive values.

Long-term outcomes after endoscopic removal of malignant colorectal polyps: Results from a 10-year cohort.

BACKGROUND: Choosing an optimal post-polypectomy management strategy of malignant colorectal polyps is challenging, and evidence regarding a surveillance-only strategy is limited. AIM: To evaluate long-term outcomes after endoscopic removal of malignant colorectal polyps. METHODS: A single-center retrospective cohort study was conducted to evaluate outcomes after endoscopic removal of malignant colorectal polyps between 2010 and 2020. Residual disease rate and nodal metastases after secondary surgery and local and distant recurrence rate for those with at least 1 year of follow-up were investigated. Event rates for categorical variables and means for continuous variables with 95% confidence intervals were calculated, and Fisher's exact test and Mann-Whitney test were performed. Potential risk factors of adverse outcomes were determined with univariate and multivariate logistic regression models. RESULTS: In total, 135 lesions (mean size: 22.1 mm; location: 42% rectal) from 129 patients (mean age: 67.7 years; 56% male) were enrolled. The proportion of pedunculated and non-pedunculated lesions was similar, with en bloc resection in 82% and 47% of lesions, respectively. Tumor differentiation, distance from resection margins, depth of submucosal invasion, lymphovascular invasion, and budding were reported at 89.6%, 45.2%, 58.5%, 31.9%, and 25.2%, respectively. Residual tumor was found in 10 patients, and nodal metastasis was found in 4 of 41 patients who underwent secondary surgical resection. Univariate analysis identified piecemeal resection as a risk factor for residual malignancy (odds ratio: 1.74; P = 0.042). At least 1 year of follow-up was available for 117 lesions from 111 patients (mean follow-up period: 5.59 years). Overall, 54%, 30%, 30%, 11%, and 16% of patients presented at the 1-year, 3-year, 5-year, 7-year, and 9-10-year surveillance examinations. Adverse outcomes occurred in 9.0% (local recurrence and dissemination in 4 patients and 9 patients, respectively), with no difference between patients undergoing secondary surgery and surveillance only. CONCLUSION: Reporting of histological features and adherence to surveillance colonoscopy needs improvement. Long-term adverse outcome rates might be higher than previously reported, irrespective of whether secondary surgery was performed.

Risk Factors and Interpretation of Inconclusive Endoscopic Ultrasound-Guided Fine Needle Aspiration Cytology in the Diagnosis of Solid Pancreatic Lesions.

BACKGROUND: The inconclusive cytological findings of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) remain a major clinical challenge and often lead to treatment delays. METHODS: Patients who had undergone EUS-FNA sampling for solid pancreas lesions between 2014 and 2021 were retrospectively enrolled. The "atypical" and "non-diagnostic" categories of the Papanicolaou Society of Cytopathology System were considered inconclusive and the "negative for malignancy" category of malignancy was suspected clinically. We determined the frequency and predictors of inconclusive cytological finding. RESULTS: A total of 473 first EUS-FNA samples were included, of which 108 cases (22.83%) were inconclusive. Significant increases in the odds of inconclusive cytological findings were observed for lesions with a benign final diagnosis (OR 11.20; 95% CI 6.56-19.54, p < 0.001) as well as with the use of 25 G FNA needles (OR 2.12; 95% CI 1.09-4.01, p = 0.023) compared to 22 G needles. Furthermore, the use of a single EUS-FNA technique compared to the combined use of slow-pull and standard suction techniques (OR 1.70; 95% CI 1.06-2.70, p = 0.027) and less than three punctures per procedure led to an elevation in the risk of inconclusive cytology (OR 2.49; 95% CI 1.49-4.14, p < 0.001). Risk reduction in inconclusive cytology findings was observed in lesions between 2-4 cm (OR 0.40; 95% CI 0.23-0.68, p = 0.001) and >4 cm (OR 0.16; 95% CI 0.08-0.31, p < 0.001) compared to lesions ≤2 cm. CONCLUSIONS: The more than two punctures per EUS-FNA sampling with larger-diameter needle (19 G or 22 G) using the slow-pull and standard suction techniques in combination may decrease the probability of inconclusive cytological findings.

[Endoscopic ultrasound in the lower gastrointestinal tract]. / Az alsó tápcsatornai endoszkópos ultrahangvizsgálat.

Endoscopic ultrasound is a minimally invasive modality that combines endoscopy with ultrasound, providing a possibility to visualize the wall of the gastrointestinal tract and adjacent tissues and organs. Since the development of the modality in the 1980s, advancements in endoscopic ultrasound technology have led to increasingly broadening indications: besides diagnostic indications, therapeutic indications have also expanded greatly. According to recent guidelines regarding rectal cancer staging, rectal ultrasonography is mainly considered to be a secondary imaging modality compared to magnetic resonance imaging. With the use of forward-viewing echoendoscopes and ultrasound miniprobes that can be inserted through the working channel of the endoscope, endoscopic ultrasound technology can be expanded to proximal, colonic areas as well. Rectal ultrasonography can also play an important role in the differential diagnosis of subepithelial lesions, in the detection of rectal varices, in the diagnosis of inflammatory bowel diseases as well as perianal complications. Diagnostic accuracy can further be improved with the addition of ultrasound-guided sampling in certain cases. Currently, therapeutic indications are more like promising possibilities, than part of everyday clinical practice, but this might change in the near future. The purpose of this review is to summarize the current indications of rectal ultrasound in the clinical practice, including diagnostic and therapeutic ones as well. Orv Hetil. 2023; 164(30): 1176-1186.