Prospective Study Protocol for Developing an Artificial Intelligence Model to Diagnose Lumbar Intervertebral Disc Prolapse

Data-based groundwater quality estimation and uncertainty analysis for irrigation agriculture

Real-World Surveillance of FDA-Cleared Artificial Intelligence Models: Rationale and Logistics.

Background: Trastuzumab deruxtecan has substantially changed the treatment of HER2-low breast cancer, emphasizing the need for accurate differentiation between HER2 immunohistochemistry (IHC) scores of 0 and 1+. However, the current accuracy of HER2 IHC 0 and 1+ scoring in real-world is inadequate. Artificial intelligence (AI) has emerged as a potential solution to improve interpretation accuracy. We developed an AI algorithm based on whole slide images (WSI) to quantitatively assess HER2 expression and its role in interpreting HER2 IHC scores of 0 and 1+. Methods: Our three-phase AI analysis framework involved segmenting tumor areas (excluding ductal carcinoma in situ), detecting and classifying tumor cells based on membrane staining patterns, and grading HER2 IHC scores according to the 2018 ASCO/CAP HER2 guideline. The AI tool was trained using 6012 patches annotated by experienced pathologists. Performance evaluation was conducted using a test dataset comprising 265 slides. A total of 141 HER2 IHC slides (27 IHC 0 and 114 IHC 1+) from patients diagnosed with invasive breast cancer at Fudan University Shanghai Cancer Center in 2021 were included. Two pretrained expert pathologists independently rescored the HER2 slides, followed by analysis using the AI tool. All inconsistent cases were also reviewed by a third senior pathologist. Interobserver agreement between the pathologists and concordance between the pathologists and the AI interpretation results were assessed. We also explored potential ranges where HER2 interpretation by pathologists exhibited inconsistency or inaccuracy. Results: The HER2 AI algorithm showed high performance in interpreting all levels of HER2 expression, with an overall kappa value of 0.85. For HER2 IHC 0 and 1+ cases, the AI model demonstrated high sensitivities (0.839 and 0.914) and specificities (0.983 and 0.890) (Table 1). After the rescoring process, 51 cases were reclassified as IHC 0 and 90 cases as IHC 1+. The overall agreement rate between the historical and re-scoring results was 73.05% (103/141). The agreement rate between the AI and re-scoring results was 91.49% (129/141) (Table 2), indicating comparable interpretation capabilities between the AI model and well-trained pathologists. The interobserver agreement between the two pathologists was 83.69% (118/141), with agreement rates of 88.23% (45/51) for IHC 0 and 81.11% (73/90) for IHC 1+. Analyzing the inconsistent cases (n=12) between IHC 0 and IHC 1+, we examined the percentage of weak and incomplete expression provided by the AI. In 9 out of 12 cases, the AI provided the same results as the third senior pathologist, suggesting that AI could assist in interpretation within this range. We identified a range of 2.83% to 24.98% as a "grey-zone," where even well-trained pathologists provided inconsistent results. Conclusion: Our study demonstrates the excellent performance of an AI-based tool for scoring HER2 IHC 0 and 1+. The AI model exhibits comparable capabilities to well-trained pathologists in interpreting HER2 expression. Additionally, we identified a "grey-zone" among pathologists, highlighting the limitations of manual subjective interpretation. AI assistance within this zone may help mitigate subjective variations. Table 1. Correlation between pathologist read and AI read of HER2 immunohistochemistry. Table 2. Correlation between pathologist rescoring and AI read of HER2 immunohistochemistry. Citation Format: Ming Li, Hong Lv, Yizhi Zhao, Chenglu Zhu, Hansheng Li, Mingzhen Lin, Wen-Tao Yang. The Advantage of Artificial-Intelligence in HER2 IHC 0 and 1+ Scoring in Breast Cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-26-09.

Artificial Intelligence: The Future of Maxillofacial Prognosis and Diagnosis?

This paper discusses the integration of artificial intelligence (AI) tools in education, delineating their potential to transform pedagogical practices alongside the challenges they present. Generative AI models like ChatGPT, had a disruptive impact on teaching and learning, due to their ability to create text, images, and sound, revolutionizing educational content creation and modification. However, nowadays the educational community is polarized, with some embracing AI for its accessibility and efficiency thus advocating it as an indispensable tool, while others cautioning against risks to academic integrity and intellectual development. This document is designed to raise awareness about AI tools and provide some examples of how they can be used to improve education and learning. From an educator’s perspective, AI is an asset for curriculum development, course material preparation, instructional design and student assessment, while reducing bias and workload. For students, AI tools offer personalized learning experiences, timely feedback, and support in various academic activities. The Turkish Biochemical Society (TBS) Academy recommends educators to embrace and utilize AI tools to enhance educational processes, and engage in peer learning for better adaptation while maintaining a critical perspective on their utility and limitations. The transfer of AI knowledge and methods to the teaching experiences should complement and not replace the educator’s creativity and critical thinking. The paper advocates for an informed embrace of AI, AI fluency among educators and students, ethical application of AI in academic settings, and continuous engagement with the evolving AI technologies, ensuring that AI tools are used to augment critical thinking and contribute positively to education and society.

Chapter 7 - Artificial intelligence in prediction of steatosis and fibrosis of nonalcoholic fatty liver disease

Will ChatGPT undermine ethical values in nursing education, research, and practice?

This exploratory study examines the transformative role of artificial intelligence (AI) in the Philippine media industry, particularly in news media, highlighting both opportunities and challenges. It utilizes a qualitative research method that includes desk reviews, key informant interviews, and focus group discussions with representatives of news media organizations, journalism and communication departments of universities, professional associations, a law firm, and an AI consulting firm. AI adoption in Philippine news media is a relatively recent phenomenon, with most news organizations adopting AI in the early 2020s and private entities leading the way. Some organizations have developed internal AI policies, while others are either crafting or refining their guidelines, indicating self-regulation initiatives. While AI is increasingly used across many newsroom tasks, human oversight and transparency are considered essential to maintain journalistic integrity. It is generally perceived as an augmentation tool rather than a replacement for human work. While there have been no reported job losses linked to AI, concerns persist. Key drivers of AI adoption include having young, tech-savvy staff, a culture of innovation, bottom-up adoption, cost-effectiveness, faster content production, digital competitiveness, and skills enhancement. However, concerns such as inaccuracies in AI-generated output, job displacement, subscription costs, and resistance from older staff and those who prefer traditional methods limit adoption. The perceived risks associated with AI are multifaceted and include hallucinations and biases in AI outputs, the exacerbation of misinformation, and job displacement. Other concerns include cybersecurity breaches, copyright issues, AI over-reliance that could diminish human capacity, the adverse impacts on the sustainability of news media, and environmental consequences. While AI improves operational efficiency, enhances productivity, and boosts audience engagement, it has also posed complex challenges that threaten the sustainability, integrity, and intellectual property of news media organizations. These challenges include declining revenues due to AI-powered platforms that deliver news summaries directly to audiences, diverting traffic away from news sites; the content scraping by platforms and AI tools, which raises serious copyright issues on top of the revenue losses experienced by news publishers; the low accountability of platforms and content creators for the content they circulate, especially when it involves copyrighted materials and false information; and the difficulties that small media organizations face to adapt to the digital age. To address these challenges and promote ethical and responsible use of AI, the study proposes several measures for AI governance. These include developing compensation policies for the media content used in AI training and on various platforms, exploring sustainable business and funding models, holding platforms accountable, enhancing the enforcement of existing laws, supporting small media organizations, developing localized AI models and tools, enhancing media self-regulation, making journalism education responsive to an AI-driven media landscape, enhancing media and information literacy, and addressing digital and AI gaps. Collaboration among all stakeholders is crucial to managing risks and leveraging AI’s potential to strengthen the Philippine news media industry’s role in society.

Purpose: This analysis aims to propose guidelines for artificial intelligence (AI) research ethics in scientific publications, intending to inform publishers and academic institutional policies in order to guide them toward a coherent and consistent approach to AI research ethics.Methods: A literature-based thematic analysis was conducted. The study reviewed the publication policies of the top 10 journal publishers addressing the use of AI in scholarly publications as of October 2024. Thematic analysis using Atlas.ti identified themes and subthemes across the documents, which were consolidated into proposed research ethics guidelines for using generative AI and AI-assisted tools in scholarly publications.Results: The analysis revealed inconsistencies among publishers’ policies on AI use in research and publications. AI-assisted tools for grammar and formatting are generally accepted, but positions vary regarding generative AI tools used in pre-writing and research methods. Key themes identified include author accountability, human oversight, recognized and unrecognized uses of AI tools, and the necessity for transparency in disclosing AI usage. All publishers agree that AI tools cannot be listed as authors. Concerns involve biases, quality and reliability issues, compliance with intellectual property rights, and limitations of AI detection tools.Conclusion: The article highlights the significant knowledge gap and inconsistencies in guidelines for AI use in scientific research. There is an urgent need for unified ethical standards, and guidelines are proposed for distinguishing between the accepted use of AI-assisted tools and the cautious use of generative AI tools.

The paper presents the results of the research study conducted on students at the Faculty of Humanities and Social Sciences in Zagreb (FHSS), Croatia, at the end of 2024, focusing on their experiences with artificial intelligence (AI) tools, and its use in academic activities, students' trust in AI, and the future use of AI by students. While over 81 % of respondents reported using AI tools for study purposes, more than 89% did not receive any education about AI and expressed a need for such education. Generative AI tools for text creation are the most frequently used type of AI tools, but students still require additional knowledge to use them effectively. AI tools are also utilized for seeking advice, finding information, being available 24/7, and assisting in writing student papers, though not yet on a massive scale. Teachers remain the most trusted source for information and for recommending AI tools. Despite known shortcomings of AI tools such as content errors, potential costs and ethical or privacy concerns, students expressed their intention to continue using AI tools in the future.

Study Design:Retrospective study of prospectively collected data.Objective:To evaluate learning curve of tubular microendoscopic discectomy (MED) in lumbar prolapsed intervertebral disc (PIVD) patients based on surgical and clinical parameters and delineate the challenges faced in early cases while practicing MED in large series of patients.Methods:This study was an institutional review board–approved retrospective study of the first 125 consecutive patients with single-level lumbar PIVD managed with tubular MED from 2008 to 2016 with a minimum 2-year follow-up. A total of 120 patients available at final follow-up were divided into quartiles (30 each) as per the date of surgery, with each consecutive group serving as a control for the previous group. Preoperatively and postoperatively clinical parameters (pain scores [Visual Analogue Scale; VAS], functional disability [Oswestry Disability Index; ODI] score, modified MacNab criteria), perioperative parameters (operative time, blood loss, hospital stay), technical issues (guide wire migration, tube docking–related problems, dural tear), and postoperative complications (postoperative leg pain, neural injury, infection, recurrence) were evaluated. Statistical analysis—logarithm curve-fit regression analysis and ANOVA test.Results:The sample consisted of 75 males and 45 females (mean age: 42.54 years) with no significant difference among the quartiles. There was significant difference (P < .005) noted in mean operative time (quartile 1, 87.33 minutes; quartile 2, 58.5 minutes) and mean blood loss (quartile 1, 76.33 mL; quartile 2, 32.66 mL) between quartile 1 and quartile 2, with no further significant reduction in quartile 3 and quartile 4. Significant difference (P < .005) in clinical parameters (VAS preoperative/postoperative 5.28/0.99; ODI preoperative/postoperative 32.18/12.08) were noted but was not associated with surgical experience. Overall, 90% (108 out of 120) of the patients had good to excellent results according to the modified MacNab criteria. The mean hospital stay did not show any significant difference among the quartiles. Guide wire migrated issues, neural injury, dural tear, and tube docking–related problems were significantly reduced after quartile 1. However, recurrence occurred at any phase. Infection occurred in one patient in quartile 1. Although blood loss and operative time showed a declining trend, it was not significant after quartile 2. So asymptote lay in quartile 1 and we recommend that novice surgeon should perform 25 to 30 cases to achieve mastery in this technique.Conclusion:For mastering the art of tubular MED for lumbar PIVD and to reduce its learning curve, novice surgeons can avoid the challenges and problems faced during initial cases with improvement in surgical skills by practicing on cadavers, wet labs, and bone-saw models following certain recommendations that we have after achieving asymptote. Familiarity with instrumentation, communication between surgical team, and defined expectations from radiology technicians are key to reduce the learning curve.

Background: Artificial intelligence (AI) tools have demonstrated success in US medical licensing examinations; however, their utility in infection prevention and control (IPC) remains unknown. Methods: The program of hospital epidemiology handles consultation calls and records each question and answer. Using 2022 data, we selected 31 frequently asked questions. We utilized four AI tools, including Chat GPT-3.5 and 4.0, Bing AI, and OpenEvidence, to generate answers. We predefined scales (Table 1) to capture responses by three reviewers, including two hospital epidemiologists and one infection preventionist. The mean score of ≥ 3 and ≥ 4 was considered acceptable in accuracy and completeness, respectively. We reported the percentage of responses with acceptable accuracy and completeness out of assessed questions for each category. Results: Among 31 questions, 16 were associated with isolation duration, 9 with healthcare personnel (HCP) exposure, 4 with cleaning contaminated rooms, and 2 with patient exposure. Regarding accuracy, most AI tools performed worse in questions about isolation duration, ranging between 75% and 93.8%. All AI tools, except OpenEvidence, had a 100% accuracy rate for HCP and patient exposure. All AI tools had a 100% accuracy rate for contaminated room handling. The highest overall acceptable accuracy rate was observed in Chat GPT-3.5. Regarding completeness, most AI tools performed worse in questions about isolation duration, ranging between 44% and 75%. All AI tools, except OpenEvidence, had a 100% completeness rate for contaminated rooms and patient exposure. The highest overall acceptable completeness rate was observed in Bing AI (Table 2). Conclusions: All AI tools provided reasonable answers to commonly asked IPC-related questions, although, there were variations among different tools used. AI could be used to supplement the infection control program, especially if resources are limited.

In the dynamic field of programming education, integrating artificial intelligence (AI) tools has started to play a significant role in enhancing learning experiences. This paper presents a case study conducted during a foundational programming course for first-year students in higher education, where students were encouraged to utilize generative artificial intelligence programming copilot extensions in their programming IDE and browser-based generative AI tools as supportive AI tools. The primary objective was to observe the impact of AI on the learning curve and the overall educational experience.Key findings suggest that the introduction of AI tools significantly altered the learning experience for students. Many who initially struggled with grasping elementary programming concepts found that AI support made understanding basic programming concepts much easier, enhancing their confidence and skills. This was particularly evident in the reduced levels of anxiety typically associated with early programming learning, as the AI copilot provided a non-judgmental, always-available source for clarifying doubts, including queries that students might hesitate to ask in a traditional classroom setting.Notably, some students leveraged the AI to generate similar exercise problems, reinforcing their understanding and skills. The AI's capability to address basic queries also freed up the instructor's time, allowing for more personalized student guidance in more advanced problems. This shift in the instructional dynamic further contributed to a learning environment where students felt more comfortable engaging with complex topics, thereby reducing the psychological barriers often linked with early-stage programming education.The course's structure, enriched by AI, enabled students to delve into more complex programming constructs earlier than traditional curricula would allow. For instance, students were tasked with simulating basic e-commerce operations, such as user registration, product browsing, and cart functionalities. These practical challenges naturally introduced advanced concepts like external data storage, unit testing, and user interface design, which are typically reserved for more advanced courses. With the help of generative AI programming copilot tools, students at any programming skill level were able to develop nearly functional complex structures. Interestingly, even when their projects were not fully functional, students remained motivated. Instead of feeling discouraged by these imperfect outcomes, they showed resilience and a keen interest in understanding and improving their code. This reaction is a significant shift from traditional learning settings, where unfinished or flawed projects often lead to increased anxiety or a drop in motivation.Furthermore, the AI's proactive suggestions inspired students to explore beyond the curriculum. Advanced learners delved into databases, cryptography libraries in Python, and even more advanced user interface design, ensuring that they remained engaged and challenged. This elementary course, enhanced by generative AI tools, also inspired students to learn other programming languages since they now learned that individual learning is more available with the aid of generative AI.In conclusion, the integration of AI in programming education offers a promising avenue for enhancing both the learning experience and outcomes. This case study underscores the potential of AI to revolutionize traditional teaching methodologies, fostering a more dynamic, responsive, and inclusive learning environment.This paper handles the results, possibilities and challenges of AI empowered education in programming. It also gives practical examples as well as future research perspectives.

Background Histological remission is an important target for Ulcerative Colitis (UC) treatment; however, scoring of histological images is time-consuming and prone to inter and intra-observer variability. Thus, a need exists for an accurate, reproducible, and reliable automated method. Previously, we demonstrated an Artificial Intelligence (AI) Tool using image processing and machine learning algorithms to measure histological disease activity using the Nancy index consistently and accurately.1 Here, we aim to enhance the capabilities of the AI Tool, by adding substantially more population-diversified training data while maintaining accuracy and robustness of results. Methods Eight global sites submitted 600 UC histological images. These were added to the 200 images previously used to train and validate the AI Tool. The 800-image dataset was divided into 2 groups: 90% used for training, 10% for testing. The novel AI algorithms were trained using state-of-the-art image processing and machine learning techniques based on deep learning and feature extraction. Cell and tissue regions of each training image were manually annotated, measured, and assigned a Nancy Index independently by 3 histopathologists, and used to further train the AI using over 43,000 characterisations. The AI Tool fully characterises histological images, identifying tissue types, cell types, cell numbers and locations, and automatically measures the Nancy Index for each image. Intra Class Correlation (ICC) and Confusion Matrix analyses were performed to evaluate the AI Tool and assess accuracy. Results The average ICC was 92.1% among the histopathologists and 91.1% between histopathologists and AI Tool, compared with 88.3% and 87.2% in the previous study.1 Confusion matrix analysis (Table 1) demonstrated the strongest correlation at the extremes of the Nancy Index, with 80% correlation between predicted and true labels for Nancy Scores of 0 or 4. When 2 adjacent scores were combined, correlations were stronger: 96% for a true Nancy score of 0 being predicted as 0 or 1, and 100% for a true Nancy score of 2 being predicted as 2 or 3. Conclusion By adding a larger number of images to the AI Tool training data, the robustness of the AI Tool was substantially improved while maintaining accuracy. The continued high correlation of AI Tool performance with the histopathologists reinforces the potential role for the AI Tool for IBD clinical applications. Fully characterising whole slides could standardise and validate an AI-driven scoring system for histology slides in IBD, eliminating the subjectivity of the human pathologist in assessment of disease activity.

Stereotactic radiosurgery (SRS) treatment planning requires accurate delineation of brain metastases, a task that can be tedious and time-consuming. Although studies have explored the use of convolutional neural networks (CNNs) in magnetic resonance imaging (MRI) for automatic brain metastases delineation, none of these studies have performed clinical evaluation, raising concerns about clinical applicability. This study aimed to develop an artificial intelligence (AI) tool for the automatic delineation of single brain metastasis that could be integrated into clinical practice. Data from 426 patients with postcontrast T1-weighted MRIs who underwent SRS between March 2007 and August 2019 were retrospectively collected and divided into training, validation, and testing cohorts of 299, 42, and 85 patients, respectively. Two Gamma Knife (GK) surgeons contoured the brain metastases as the ground truth. A novel 2.5D CNN network was developed for single brain metastasis delineation. The mean Dice similarity coefficient (DSC) and average surface distance (ASD) were used to assess the performance of this method. The mean DSC and ASD values were 88.34%±5.00% and 0.35±0.21 mm, respectively, for the contours generated with the AI tool based on the testing set. The DSC measure of the AI tool's performance was dependent on metastatic shape, reinforcement shape, and the existence of peritumoral edema (all P values <0.05). The clinical experts' subjective assessments showed that 415 out of 572 slices (72.6%) in the testing cohort were acceptable for clinical usage without revision. The average time spent editing an AI-generated contour compared with time spent with manual contouring was 74 vs. 196 seconds, respectively (P<0.01). The contours delineated with the AI tool for single brain metastasis were in close agreement with the ground truth. The developed AI tool can effectively reduce contouring time and aid in GK treatment planning of single brain metastasis in clinical practice.