Transforming Cancer Nanotechnology Data Analysis and User Experience. Part I: Current Challenges and Solutions Provided by caNanoLab

Photodynamic therapy (PDT) has been used in the treatment of cancers and other benign diseases for several years in clinic. However, the hypoxia of tumors and the penetration limitation of excitation light to tissues can dramatically reduce the efficacy of PDT to cancers. To overcome these drawbacks, various assembled nanocarriers such as nanoparticles, nanocapsules, nanocrystals, and so on were introduced. The assembled nanocarriers have the ability of loading photosensitizers, delivering O2 into tumors, generating O2 in situ in tumors, as well as turning near-infrared (NIR) light, X-rays, and chemical energy into ultraviolet or visible light. Therefore, it is easy for the nanocarriers to improve the hypoxia microenvironment or increase the treatment depth of cancers, which will improve the efficiency of PDT to some degree. In recent years, a number of investigations were focused on these subjects. We will summarize the advances of nanocarriers in PDT, especially in O2 introduction PDT and deep PDT. The perspectives, challenges, and potential in translation of PDT will also be discussed. This article is categorized under: Therapeutic Approaches and Drug Discovery > Emerging Technologies Biology-Inspired Nanomaterials > Lipid-Based Structures Therapeutic Approaches and Drug Discovery > Nanomedicine for Oncologic Disease.

Dean Ho, PhD, a professor in the division of oral biology and the division of advanced prosthodontics as well as codirector of The Jane and Jerry Weintraub Center for Reconstructive Biotechnology at the University of California at Los Angeles (UCLA) School of Dentistry, is excited about the promise nanotechnology holds in better treating and diagnosing cancer. “A major challenge in cancer therapy is drug resistance, and our group and others are finding that when drugs are modified with nanoparticles, they are less toxic and stay in the tumor longer,” he says. “You also can add targeting agents and antibodies to focus drug delivery more effectively.” Dr. Ho's work involves nanoscale, diamond-like particles called nanodiamonds. They resemble mini soccer balls and are made from a carbon material that is well tolerated in living organisms. He and his colleagues have used nanodiamonds to deliver doxorubicin, a highly toxic cancer drug. The synthesis of the 2, known as NDX, has been studied in a variety of cancers, including those of the breast, blood, liver, and brain, generating promising results. In the realm of imaging, Dr. Ho's research team also has investigated the use of nanodiamonds in helping to better locate tumors and significantly reduce the toxicity of dyes that are injected into patients for magnetic resonance imaging. Dr. Ho began his research on nanodiamonds while at Northwestern University in 2007. “Taking this sustainable biomaterial made from mining and refining operations and watching it go from just a proof of concept to showing it's safe in large animals has been a very interesting clinical journey,” he says. “Hopefully, in the next 9 to 12 months we'll have a small clinical study.” His work is just one example of the many research projects being performed across the country that are developing nanomedical approaches to combatting cancer. The first nano drugs approved by the US Food and Drug Administration (FDA) for the treatment of patients with cancer were doxorubicin in 1995 and paclitaxel in 2005. Since that time, the field has continued to evolve. The National Cancer Institute (NCI) Alliance for Nanotechnology in Cancer has been leading funding efforts in the use of nanotechnology since 2004. The Alliance will start its third funding cycle next year and at press time was about to issue its latest grants solicitation. “Round 1 was very open, and in round 2 we suggested focusing on 4 tumor types with low survival rates: ovarian, pancreatic, lung, and brain,” says Piotr Grodzinski, PhD, director of the Alliance. “We are continuously encouraging translational work and having academics involved in the program reach out for alternative funding to move the technology to the translational and even commercial phase.” Toward that end, the Alliance funds 9 centers of excellence at academic medical centers across the country charged with developing nanotechnology applications for clinical oncology. One example of a technology that is already in clinical trials was developed by Robert Langer, ScD, codirector of the MIT (Massachusetts Institute of Technology)- Harvard Center of Cancer Nanotechnology Excellence in Cambridge, Massachusetts. Dr. Langer, who began working on nanotechnologies in the 1970s, helped to lay much of the groundwork for the field. He and his Harvard Medical School colleague Omid Farokhzad, MD, founded BIND Therapeutics in 2007 to develop innovative pharmaceutical products based on nanomedicine. Their second-generation therapeutics, known as accurins, are polymeric nanoparticles designed to deliver a payload to the tumor site and have prolonged circulation in the bloodstream. Accurins are designed to target tumors at the tissue, cellular, and molecular levels. They also have been engineered for controlled and timely release. BIND Therapeutics has collaborated or is collaborating with Amgen, Pfizer, AstraZeneca, and Roche to develop these therapeutics. The company is conducting 2 phase 2 trials of 40 patients each in the second-line treatment of non-small cell lung cancer and the first-line treatment of metastatic, castrate-resistant prostate cancer. They expect to report top-line results from these studies later this year. Other investigators who have developed promising nanotechnologies for cancer include Mark Davis, PhD, of the California Institute of Technology (Caltech) in Pasadena, California, who has developed a nanoparticle for the delivery of chemotherapeutics and small interfering RNA (siRNA), which is currently being tested in clinical trials; Chad Mirkin, PhD, of Northwestern University in Chicago, Illinois, who has designed a wide range of diagnostic devices for the detection of biomarkers; and James Heath, PhD, also of Caltech, who has developed devices that monitor levels of protein in the blood and in single cells, which are being used in immunotherapy clinical trials. “We have a good mix of diagnostic and therapeutic discoveries that are becoming more practical,” Dr. Grodzinski says, noting that in therapeutics, promising new nanoparticle platforms not only appear to reduce side effects but also enable greater accumulations of drugs at the tumor site. He points to 2 other NCI Alliance nanotechnology initiatives that are helping to advance the field. The first, known as the Nanotechnology Characterization Lab (NCL), performs preclinical efficacy and toxicity testing of nanoparticles. The NCL serves as a resource for cancer researchers investigating nanoscale particles and devices and is a collaboration between the NCI, the National Institute of Standards and Technology, and the FDA. The second is a public-private industry partnership known as Translation of Nanotechnology in Cancer. The consortium brings together government entities as well as pharmaceutical, biotechnology, and other health care-related companies to expose them to and seek their input on future directions of nanotherapeutics. Dr. Grodzinski hopes to see more clinicians involved in the early development of cancer nanotechnologies well before the clinical trial stage. “We need to have a broader exposure of these technologies to practicing clinicians so they can guide the selection of the application and become more involved in the design process,” he says. Drs. Grodzinski, Langer, and Ho agree that nanotechnology faces the same challenges as other new cancer treatments and diagnostics in terms of funding, regulatory pathways, and weighing risks versus benefits. Nevertheless, they anticipate seeing FDA approvals for some of these new nanoparticles and devices within the next 5 years. Although it is not completely clear how these materials might affect patients in the long term, that should not discourage scientists and clinicians from working to bring them into the clinic, Dr. Ho says. “People should embrace [nanotechnology] and learn as much as they can about it, because the benefits are there.” Our group and others are finding that when drugs are modified with nanoparticles, they are less toxic and stay in the tumor much longer. — Dean Ho, PhD Below are 2 recent nanotechnology drug developments in the treatment of cancer. In recent years, scientists have developed nanoparticles that deliver 1 or 2 chemotherapy drugs, but they have been unable to make particles that can deliver a larger number of drugs in a precise ratio. Recently, chemists at MIT in Cambridge, Massachusetts, have accomplished that goal. Senior author Jeremiah Johnson, PhD, an assistant professor of chemistry at MIT, and colleagues published an article in the Journal of the American Chemical Society describing how they were able to load their particles with 3 drugs commonly used to treat ovarian cancer.1 Dr. Johnson says he believes their research is the first example of a nanoparticle that delivers a precise ratio of 3 drugs and can release them in response to 3 different triggering mechanisms. The particles could potentially be designed to deliver more than 3 drugs, which would enable new treatments that could more effectively kill cancer cells while avoiding the side effects of traditional chemotherapy, he says. The investigators reported that the nanoparticles were able to more effectively kill ovarian cancer cells than particles with only 1 or 2 drugs. They have begun testing in tumors in animals. The work of encapsulating small drug molecules inside of, or chemically attaching them to, the nanoparticles becomes increasingly difficult with each new drug that is added. Dr. Johnson and his colleagues created a new type of particle that would enable the loading of many different drugs. Rather than building the particle and attaching drug molecules, they created building blocks that already include the drug and can be joined together in a very specific structure. In that way, scientists can control how much of each drug is loaded into the particle. The building block includes the drug molecule, a linking unit that can connect to other blocks, and a chain of polyethylene glycol that helps protect the protein from being broken down in the body. Hundreds of these molecules can be linked using an approach the researchers developed called “brush-first polymerization.” The study involved creating particles that carried cisplatin, doxorubicin, and camptothecin, which are used alone or in combination to treat ovarian cancer. The particles contained a ratio of the maximum tolerated dose of each drug, each of which had its own release mechanism. Cisplatin and camptothecin were designed to be released quickly whereas doxorubicin was designed to be released only when ultraviolet light shines on the particle. Once the drugs are released, polyethylene glycol is easily biodegradable, they say. The team is now working on particles that carry 4 drugs as well as ways with which to tag the particles with molecules, thereby allowing them to target tumors by interacting with proteins found on cell surfaces. Dr. Johnson adds that the nanoparticles could enable large-scale testing of potential new cancer treatments. A drug delivery system using nanomedicine technologies can precisely target cancer cells in the bone and increase bone strength to prevent the disease's progression, according to a study published recently in the Proceedings of the National Academy of Sciences.2 Researchers from Brigham and Women's Hospital and the Dana-Farber Cancer Institute in Boston, Massachusetts, developed stealth nanoparticles made of a combination of clinically validated biodegradable polymers and alendronate, which is a therapeutic agent that belongs to the bisphosphonate class of drugs. Bisphosphonates bind to calcium, which is highly concentrated in bones. The surface of the nanoparticles is decorated with alendronate, which can kill the tumor cells and stimulate new bone growth. Co-lead author Michaela Reagan, PhD, of the Dana-Farber Cancer Institute Center for Hematologic Oncology, notes that few treatment options exist for patients with bone cancer. At the same time, her colleague and co-lead author, Archana Swami, PhD, of the Brigham and Women's Hospital Laboratory of Nanomedicine and Biomaterials, says that bone is a favorable microenvironment for cancer cells that migrate from tumors such as those from the breast, prostate, and blood. The team tested the nanoparticles in mice with multiple myeloma. The mice were initially pretreated with nanoparticles loaded with the anticancer drug bortezomib before being injected with myeloma cells. The treatment led to slower cancer growth and longer survival. The drug also enhanced the strength and volume of the bone. The study authors add that their work provides proof of concept that targeting the bone marrow niche can prevent or delay bone metastasis and will pave the way for future clinical trials.

Background and Aims Large language models (LLMs) have gained significant attention in the field of natural language processing (NLP), marking a shift from traditional techniques like Term Frequency-Inverse Document Frequency (TF-IDF). We developed a traditional NLP model to predict arteriovenous fistula (AVF) failure within next 30 days using clinical notes. The goal of this analysis was to investigate whether LLMs would outperform traditional NLP techniques, specifically in the context of predicting AVF failure within the next 30 days using clinical notes. Method We defined AVF failure as the change in status from active to permanently unusable status or temporarily unusable status. We used data from a large kidney care network from January 2021 to December 2021. Two models were created using LLMs and traditional TF-IDF technique. We used “distilbert-base-uncased”, a distilled version of BERT base model [1], and compared its performance with traditional TF-IDF-based NLP techniques. The dataset was randomly divided into 60% training, 20% validation and 20% test dataset. The test data, comprising of unseen patients’ data was used to evaluate the performance of the model. Both models were evaluated using metrics such as area under the receiver operating curve (AUROC), accuracy, sensitivity, and specificity. Results The incidence of 30 days AVF failure rate was 2.3% in the population. Both LLMs and traditional showed similar overall performance as summarized in Table 1. Notably, LLMs showed marginally better performance in certain evaluation metrics. Both models had same AUROC of 0.64 on test data. The accuracy and balanced accuracy for LLMs were 72.9% and 59.7%, respectively, compared to 70.9% and 59.6% for the traditional TF-IDF approach. In terms of specificity, LLMs scored 73.2%, slightly higher than the 71.2% observed for traditional NLP methods. However, LLMs had a lower sensitivity of 46.1% compared to 48% for traditional NLP. However, it is worth noting that training on LLMs took considerably longer than TF-IDF. Moreover, it also used higher computational resources such as utilization of graphics processing units (GPU) instances in cloud-based services, leading to higher cost. Conclusion In our study, we discovered that advanced LLMs perform comparably to traditional TF-IDF modeling techniques in predicting the failure of AVF. Both models demonstrated identical AUROC. While specificity was higher in LLMs compared to traditional NLP, sensitivity was higher in traditional NLP compared to LLMs. LLM was fine-tuned with a limited dataset, which could have influenced its performance to be similar to that of traditional NLP methods. This finding suggests that while LLMs may excel in certain scenarios, such as performing in-depth sentiment analysis of patient data for complex tasks, their effectiveness is highly dependent on the specific use case. It is crucial to weigh the benefits against the resources required for LLMs, as they can be significantly more resource-intensive and costly compared to traditional TF-IDF methods. This highlights the importance of a use-case-driven approach in selecting the appropriate NLP technique for healthcare applications.

Living cells have become ideal therapeutic agents for cancer treatment owing to their innate activities, such as efficient tumor targeting and delivery, easy engineering, immunomodulatory properties, and fewer adverse effects. However, cell agents are often fragile to rigorous tumor microenvironment (TME) and limited by inadequate therapeutic responses, leading to unwanted treatment efficacy. Cell nanomodification, particularly the cell surface-nanoengineering has emerged as reliable and efficient strategy that not only combines cell activity properties with nanomaterials but also endows them with extra novel functions, enabling to achieve remarkable treatment results. In this review, we systematically introduce two major strategies have been adopted to develop cell surface engineering with nanomaterials, mainly including living cell nano-backpacks and cell membrane-mimicking nanoparticles (NPs). Based on various functional NPs and cell types, we focus on reviewing the cell-surface nanoengineering for targeted drug delivery, immune microenvironment regulation, and precisely antitumor therapy. The advances and challenges of cell surface-nanoengineered antitumor agents for cancer therapy applications are further discussed in future clinical practice. This review provides an overview of the advances in cell surface-engineering for targeting immunoregulation and treatment and could contribute to the future of advanced cell-based antitumor therapeutic applications. This article is categorized under: Therapeutic Approaches and Drug Discovery > Nanomedicine for Oncologic Disease Therapeutic Approaches and Drug Discovery > Emerging Technologies Nanotechnology Approaches to Biology > Cells at the Nanoscale.

AI-enabled language models (LMs) to large language models (LLMs) and multimodal large language models (MLLMs) in drug discovery and development

JupyterLab is a web-based interactive development platform that is widely used in the Earth science community. Using Jupyter Notebooks, it is possible to perform data analysis tasks, annotate and visualize results in a way that is easy to reproduce, present and share with others. JupyterLab allows the use of “extensions”, which add functionality to the platform. One of these is Jupyter-AI [1], which allows the use of Large Language Models (LLMs), such as ChatGPT, Claude Sonnet and Ollama, within the JupyterLab environment, through  a chat interface or directly within notebooks. By integrating LLMs into JupyterLab, it is possible to leverage their code generation capabilities to assist a user to translate their analysis tasks from an idea to actual executable code in an efficient manner. One drawback of using these LLMs in tasks involving spatio–temporal data is that the models typically do not have access to the data necessary for the analysis task and will often resort to generating fictional data or using placeholders in the code that they create. This requires the user to adapt the provided code to their data, which removes some of the utility provided by the LLM.In this context we make use of FrevaGPT, an approach for using LLMs in climate data analysis that allows for quick, complex and reproducible analyses of data sets, such as decadal climate model forecasts. Leveraging LLM’s capability to write code and using few-shot prompting (in-context learning) allows the LLM to utilize Freva [2,3] (Free Evaluation Framework), a data search and analysis platform, which provides a standardised interface to spatio-temporal datasets hosted on an HPC cluster [4].  FrevaGPT integrates seamlessly into Jupyter-AI and, by making use of the Freva library, combines the code-generating capabilities of LLMs with contextual understanding of how to access relevant datasets on the HPC cluster. This in addition with FrevaGPT’s ability to execute generated code in an isolated environment on an HPC node, annotating and explaining any intermediate results, as well as automatically correcting errors encountered along the way, could serve as a starting ramp for researchers to efficiently produce new analysis products based on spatio-temporal climate data. This PICO will include examples of using FrevaGPT within JupyterLab to analyse spatio-temporal datasets from the climate of the past, as well seasonal to decadal climate predictions.  

e23161 Background: Data retrieval is challenging in clinical research and traditional methods for data collection are often time-consuming and may be error-prone. Large Language Models (LLMs) have shown zero-shot capabilities in converting unstructured clinical text into structured data. These technologies could support the retrieval stage of clinical trials by leveraging the information reported in Electronic Health Records (EHRs) without relying any longer on manual curation. APOLLO 11 Consortium (NCT05550961) is a multicentric Italian trial which leverages a federated infrastructure for the analysis of advanced lung cancer patient data across Italy. Methods: We conducted a pilot study using Llama 3.1 8B on 358 Non-Small Cell Lung Cancer patients from the IRCCS Istituto Nazionale dei Tumori, leader of the APOLLO 11 Consortium. Anonymized EHRs have been analyzed within the LLM pipeline for feature extraction by Wiest et al. A combination of zero/few shot prompting techniques both in English and Italian languages was used. We selected smoking, histology, PD-L1 and staging as multiclass variables and bone/brain/liver metastases as binary variables. The ground truth collection involved a first Manual Data Entry (1-MDE) and a final full-revised MDE (2-MDE). The LLM accuracy was calculated only for the comparison LLM vs 2-MDE. In addition, we calculated the percentage of Missing Information (% MI) in 1-MDE, 2-MDE and LLM extraction. Results: Compared to 2-MDE, LLM achieved feature-specific accuracies of 0.78 for PD-L1, 0.85 for BONE METASTASIS, 0.83 for BRAIN METASTASIS, 0.89 for LIVER METASTASIS and 0.96 for TUMOUR STAGING. For smoking and staging, LLM extraction also reduced % MI relative to 1-MDE (Table 1). Only for PD-L1, we further analyzed the 12.8% of MI and found that 91.3% resulted from hallucinations (i.e., PD-L1 was misclassified as missing). Evaluations using English prompts confirmed the pipeline’s adaptability and high tasks accuracy. Conclusions: This study confirms the feasibility of LLMs for data retrieval in clinical trials demonstrating strong performance across diverse clinical features with minimal prompt optimization. LLMs could assist clinicians and data entry personnel in the 1-MDE process, streamlining initial data structuring and saving time. The 2-MDE step can remain as a quality check to address any discrepancies. Further improvements could focus on prompt optimization and integrating human feedback to reduce hallucination rates. Clinical trial information: NCT05550961 . %MI in 1-MDE, 2-MDE and LLM extraction. Accuracy refers only to LLM vs 2-MDE. Histology and metastasis sites were collected only in 2-MDE. NA = not available. Smoking PD-L1 Histology Bone Met Brain Met Liver Met T N M Stage % MI 1-MDE 6.4 8.9 NA NA NA NA 22.5 22.5 23.11 98.3 % MI 2-MDE 6.6 3 0 0 0 0 0 0 0 0 % MI LLM 2.7 12.8 10.3 0 0 0 0 0 0 6.9 % accuracy (LLM vs 2-MDE) 67 78 91 85 83 89 39 52 70 96

The rapid deployment of large language models (LLMs) on mobile devices has introduced significant privacy concerns, particularly regarding data collection, user profiling, and compliance with evolving AI regulations such as the GDPR and the AI Act. While these on-device LLMs promise improved latency and user experience, their potential to inadvertently leak sensitive information remains understudied. This paper proposes a red teaming framework to systematically assess the privacy risks of phone-based LLMs, simulating adversarial attacks to identify vulnerabilities in model behavior, data storage, and inference processes. We evaluate popular mobile LLMs under scenarios such as prompt injection, side-channel exploitation, and unintended memorization, measuring their compliance with strict privacy-by-design principles. Our findings reveal critical gaps in current safeguards, including susceptibility to context-aware deanonymization and insufficient data minimization. We further discuss regulatory implications, advocating for adaptive red teaming as a mandatory evaluation step in AI governance. By integrating adversarial testing into the development lifecycle, stakeholders can preemptively align phone-based AI systems with legal and ethical privacy standards while maintaining functional utility.

Nanoparticles are the only materials small enough to target cells in the body, and therefore are crucial to targeted drug delivery. Issues with the synthesis, consistency, and bioactivity of these molecules are still being addressed, but base on current proof of concept studies there is a reason to believe that the 'holy grail' of targeted drug delivery might someday be achieved using nanoparticle-based systems.

Escalating medical expenses due to infectious diseases are causing huge socioeconomic pressure on mankind globally. The emergence of antibiotic resistance has further aggravated this problem. Drug-resistant pathogens are also capable of forming thick biofilms on biotic and abiotic surfaces to thrive in a harsh environment. To address these clinical problems, various strategies including antibacterial agent delivering matrices and bactericidal coatings strategies have been developed. In this review, we have discussed various types of polymeric vehicles such as hydrogels, sponges/cryogels, microgels, nanogels, and meshes, which are commonly used to deliver antibiotics, metal nanoparticles, and biocides. Compositions of these polymeric matrices have been elaborately depicted by elucidating their chemical interactions and potential activity have been discussed. On the other hand, various implant/device-surface coating strategies which exploit the release-active mechanism of bacterial killing are discussed in elaboration. This article is categorized under: Therapeutic Approaches and Drug Discovery > Nanomedicine for Cardiovascular Disease Implantable Materials and Surgical Technologies > Nanomaterials and Implants Therapeutic Approaches and Drug Discovery > Nanomedicine for Infectious Disease.

Synthetic vectors for therapeutic nucleic acid delivery are currently competing significantly with their viral counter parts due to their reduced immunogenicity, large payload capacity, and ease of manufacture under GMP-compliant norms. The approval of Onpattro, a lipid-based siRNA therapeutic, and the proven clinical success of two lipid-based COVID-19 vaccines from Pfizer-BioNTech, and Moderna heralded the specific advantages of lipid-based systems among all other synthetic nucleic acid carriers. Lipid-based systems with diverse payloads-plasmid DNA (pDNA), antisense oligonucleotide (ASO), small interfering RNA (siRNA), microRNA (miRNA), small activating RNA (saRNA), and messenger RNA (mRNA)-are now becoming a mature technology, with growing impact in the clinic. Research over four decades identified the key factors determining the therapeutic success of these multi-component systems. Here, we discuss the main nucleic acid-based technologies, presenting their mechanism of action, delivery barriers facing them, the structural properties of the payload as well as the component lipids that regulate physicochemical properties, pharmacokinetics and biodistribution, efficacy, and toxicity of the resultant nanoparticles. We further detail on the formulation parameters, evolution of the manufacturing techniques that generate reproducible and scalable outputs, and key manufacturing aspects that enable control over physicochemical properties of the resultant particles. Preclinical applications of some of these formulations that were successfully translated from in vitro studies to animal models are subsequently discussed. Finally, clinical success and failure of these systems starting from 1993 to present are highlighted, in a holistic literature review focused on lipid-based nucleic acid delivery systems. This article is categorized under: Therapeutic Approaches and Drug Discovery > Emerging Technologies Therapeutic Approaches and Drug Discovery > Nanomedicine for Oncologic Disease Toxicology and Regulatory Issues in Nanomedicine > Toxicology of Nanomaterials.

Photodynamic therapy (PDT) is a minimally invasive technique which has proven to be successful in the treatment of several types of tumors. This relatively simple method exploits three inseparable elements: phototoxic compound (photosensitizer [PS]), light source, and oxygen. Upon irradiation by light with specified wavelength, PS generates reactive oxygen species, which starts the cascade of reactions leading to cell death. The positive therapeutic outcome of PDT may be limited due to several aspects, including low water solubility of PSs, hampering their effective administration and blood circulation, as well as low tumor specificity, inefficient cellular uptake and activation energies requiring prolonged illumination times. One of the promising approaches to overcome these obstacles involves the use of carrier systems modulating pharmacokinetics and pharmacodynamics of the PSs. In the present review, we summarized current in vitro and in vivo studies regarding the use of nanoparticles as potential delivery devices for PSs to enhance their cellular uptake and cytotoxic properties, and thus-the therapeutic outcome of PDT. This article is categorized under: Therapeutic Approaches and Drug Discovery > Emerging Technologies Nanotechnology Approaches to Biology > Nanoscale Systems in Biology Therapeutic Approaches and Drug Discovery > Nanomedicine for Oncologic Disease.

This paper presents a system design for automating incident update responses in telecommunication service providers using Large Language Models (LLMs). We address key challenges in current incident management practices: tailoring customer-faced communications while also maintaining cost efficiency and user experience. The proposed system design implements a caching mechanism to reduce operational costs and response latency. Rather than using semantic similarity search, our system employs a strict key-match caching approach that ensures consistent responses while minimizing computational overhead. We selected DeepSeek-V3 API as our primary LLM based on comprehensive cost-performance analysis. Our prompt engineering approach incorporates persona definition, chain-of-thought reasoning, and constitutional AI principles to generate tailored responses. Experimental results demonstrate significant efficiency gains, validated through quantitative analysis. Eliminating redundant LLM calls avoids linear cost growth, while reduced latency improves speed by 99%. Additionally, LLM response quality successfully passed User Acceptance Test. This work contributes to the practical implementation of LLMs in Customer Relationship Management systems by balancing cost efficiency, response quality, and customer experience.

This review explores the frontiers of large language models (LLMs) in psychological applications. Psychology has undergone several theoretical changes, and the current use of artificial intelligence (AI) and machine learning, particularly LLMs, promises to open up new research directions. We aim to provide a detailed exploration of how LLMs are transforming psychological research. We discuss the impact of LLMs across various branches of psychology—including cognitive and behavioral, clinical and counseling, educational and developmental, and social and cultural psychology—highlighting their ability to model patterns, cognition, and behavior similar to those observed in humans. Furthermore, we explore the ability of such models to generate coherent, contextually relevant text, offering innovative tools for literature reviews, hypothesis generation, experimental designs, experimental subjects, and data analysis in psychology. We emphasize the importance of addressing technical and ethical challenges, including data privacy, the ethics of using LLMs in psychological research, and the need for a deeper understanding of these models’ limitations. Researchers should use LLMs responsibly in psychological studies, adhering to ethical standards and considering the potential consequences of deploying these technologies in sensitive areas. Overall, this review provides a comprehensive overview of the current state of LLMs in psychology, exploring the potential benefits and challenges. We hope it can serve as a call to action for researchers to responsibly leverage LLMs’ advantages while addressing the associated risks.

11026 Background: Increasingly more people use large language models (LLMs) to find information about medical treatments. However, there are notable concerns about the accuracy and completeness of information generated by these models. We assessed whether LLMs accurately summarized uncertainties about the benefits and harms of new cancer drugs. Methods: We identified the 10 cancer drugs approved by the US Food and Drug Administration (FDA) between 2019 and 2022 with the highest Medicare spending in 2022 (5 in Part B and 5 in Part D). We then searched FDA review documents to extract information about the uncertainties with each drug’s clinical trial evidence that were identified by FDA reviewers at the time of approval. Uncertainties with clinical trial evidence were assigned to mutually exclusive categories. We evaluated the extent to which 4 state-of-the-art LLMs (OpenAI’s ChatGPT-4, Google’s Gemini 1.5 Pro, Meta’s Llama 3.1, and Anthropic’s Claude 3.5 Sonnet) provided information about FDA-identified uncertainties when queried for information about the drugs using two prompts: (1) how well does [drug] work for [condition]? ; (2) is there anything uncertain about how well [drug] works for [condition]? . Results: For the 10 recently approved cancer drugs with the highest Medicare spending in 2022, FDA reviewers identified a total of 38 uncertainties with the clinical trial evidence. For 9 of 10 drugs, FDA reviewers identified uncertainties related to the generalizability of the evidence. Other common uncertainties included bias related to the measurement of the outcome and the use of single arm trial designs. When the LLMs were prompted about how well these 10 cancer drugs worked, the models rarely provided information about the uncertainties identified by FDA reviewers: GPT-4 (4/38, 11%), Gemini 1.5 Pro (3/38, 8%), Llama 3.1 (3/38, 8%), and Claude Sonnet (2/38, 5%). The proportion of FDA-identified uncertainties reported by the models improved marginally when specifically prompted for uncertainties about the drugs. Qualitative assessment of the information generated by the LLMs showed that most of the models tended to report similar, non-specific uncertainties for every drug with little variation. Conclusions: LLMs do not provide adequate information about uncertainties related to the benefits and harms of recently approved cancer drugs, despite the availability of this information in the public domain. There is a need to improve LLMs to accurately report such information, so that patients can make informed decisions about cancer treatments. Performance of LLMs compared to uncertainties identified by the FDA. Prompt GPT-4 Gemini 1.5 Pro Llama 3.1 Claude 3.5 Sonnet Average How well does [drug] work for [condition]? 4/38 (11%) 3/38 (8%) 3/38 (8%) 2/38 (5%) 3/38 (8%) Is there anything uncertain about how well [drug] works for [condition]? 8/38 (21%) 5/38 (13%) 6/38 (16%) 5/38 (13%) 6/38 (16%)