Immune checkpoint blockade (ICB) has improved outcomes for patients with head and neck squamous cell carcinoma (HNSCC), but predictive biomarkers remain limited. Here, we use a time-resolved, multi-omic approach in a murine HNSCC model to characterize peripheral immune responses to ICB. Single-cell transcriptomics and T/B cell receptor analyses reveal early on-treatment expansion of effector memory T and B cell repertoires in responders, preceding tumor regression. These dynamic immune features inform a composite transcriptional signature that accurately predicts ICB response in independent human HNSCC cohorts. LiBIO outperforms existing biomarkers and generalizes to melanoma, non-small cell lung cancer, and breast cancer without retraining. These findings suggest that early treatment-induced changes in circulating immune repertoires reflect the host's capacity to mount an effective antitumor response. This work provides a framework for leveraging transient peripheral immune dynamics to develop non-invasive, high-fidelity biomarkers for response to immunotherapy across cancer types.
AI is steadily changing the way banks work. The technology can sift through massive amounts of data, calculate risks, and handle routine tasks at speeds people can’t match. Now, Malaysia has entered that space with the launch of Ryt Bank, billed as the first AI-powered bank created in the country.
The new venture, led by YTL Group in partnership with Sea Limited, arrives just ahead of Merdeka. “Ryt Bank demonstrates that groundbreaking innovation can be imagined, built, and led right here in Malaysia,” said Dato’ Seri Yeoh Seok Hong, Managing Director of YTL Power International. “By combining homegrown AI with the values and diversity of our people, we’ve created a bank that Malaysians can proudly call their own – one that speaks our languages, understands our culture, and sets a new standard for how banking should feel.”
Banking for Malaysians
Ryt Bank has been designed to work in the languages most Malaysians use every day. Its app is already available in Bahasa Malaysia and English, with Mandarin support scheduled to arrive by next month (September 2025). By offering multilingual access, the bank aims to make financial services more inclusive and easy to use for people in many different communities.
The centrepiece of the AI-powered bank is Ryt AI, a digital assistant powered by ILMU, Malaysia’s first locally-developed large language model. Ryt AI can understand natural conversation – in Bahasa Malaysia, English, or a mixture of both – and act on requests instantly.
The AI assistant can read and pay bills, track spending, and explain financial basics in plain terms. The idea is to blend convenience with cultural familiarity, while maintaining enterprise-grade security.
Everyday AI banking in one app
Ryt Bank is designed to pull together multiple financial needs into one platform. Customers can use the AI bank app to save, spend, borrow, and pay bills, with Ryt AI making the process more conversational and personal.
Personal banking with Ryt AI
Send money or pay bills through text chat, with support for DuitNow and JomPAY.
Snap and upload bills or receipts for instant payment.
Access guides and simple financial explanations as you bank.
All actions are encrypted and verified for security.
New users can claim a small launch reward of up to RM5 when they try Ryt AI.
Growing money
Customers earn up to 4% interest per year, credited daily.
Withdraw funds anytime, with no lock-in requirements.
Ryt PayLater
Access instant credit of up to RM1,499.
0% interest if paid back in a month.
No late fees and no paperwork.
Earn cashback on DuitNow QR payments and extra rewards with select partners.
Ryt Card
Switch between debit and credit card models in the app.
Accepted worldwide through Visa.
No foreign transaction or ATM fees in Malaysia.
Cashback on spending, plus partner offers like Shopee vouchers and dining discounts at YTL Hotels.
The AI bank is licensed by Bank Negara Malaysia and covered by PIDM, which protects deposits up to RM250,000 per customer. Security features include biometric login, layered encryption, and real-time fraud alerts.
A step forward for Malaysia’s banking sector
Ryt Bank’s launch shows how AI is being used to rethink traditional banking. It aims to make financial services more accessible and satisfy regulatory and security standards by supporting local languages and developing its own AI assistant.
Constructing effective representation of lesions is essential for disease classification and localization in medical image analysis. Prototype-based models address this by leveraging visual prototypes to capture representative lesion patterns, yet effectively handling the complexity of diverse lesion characteristics remains a critical challenge, as they typically rely on single-level, fixed-size prototypes and suffer from prototype redundancy. In this paper, we present HierProtoPNet, a new prototype-based framework designed to handle the complexity of lesions in medical images. HierProtoPNet leverages hierarchical visual prototypes across different semantic feature granularities to effectively capture diverse lesion patterns. To prevent redundancy and increase utility of the prototypes, we devise a novel prototype mining paradigm to progressively discover semantically distinct prototypes, offering multi-level complementary analysis of lesions. Also, we introduce a dynamic knowledge distillation strategy that allows transferring essential classification information across hierarchical levels, thereby improving generalisation performance. Comprehensive experiments show that HierProtoPNet achieves state-of-the-art classification performances in three benchmarks: binary breast cancer screening, multi-class retinal disease diagnosis, and multi-label chest X-ray classification. Quantitative assessments also illustrate HierProtoPNet's significant advantages in weakly-supervised disease localisation and segmentation.
In this work, we propose a radiomics-guided neural network, XRadNet, for breast cancer molecular subtype prediction. XRadNet is a two-head neural network, with one for predicting molecular subtypes and the other for approximating radiomic features. In addition, a training scheme with radiomics guidance is proposed to improve performance. First, we conduct a series of experiments to test the radiomic feature learning capacity of different neural networks, which determines the backbone of XRadNet. Moreover, significant radiomic features are also determined according to radiomics and prior knowledge. XRadNet is subsequently pretrained in a self-supervised manner. The pretraining uses synthetic samples to train the backbone and radiomic feature regression head. This mitigates the impact of an insufficient number of samples. Finally, XRadNet is fine-tuned with a downstream real-world dataset by enabling all heads. Furthermore, a logistic regression is built with radiomic features and learned features, which provides a new way to interpreting the trained model with concepts familiar to radiologists. The experimental results show that XRadNet effectively predicts the four molecular subtypes of breast cancer. These results also demonstrate that the proposed training scheme yields better or competitive performance than those models pretrained on ImageNet or medical datasets.
Functional magnetic resonance imaging (fMRI) has demonstrated significant potential in the early diagnosis and study of pathological mechanisms of Alzheimer's disease (AD). To fit subtle cross-spatiotemporal interactions and learn pathological features from fMRI, we propose a fine-grained spatiotemporal graph neural network with self-supervised learning (SSL) for diagnosis and biomarker extraction of early AD. First, considering the spatiotemporal interaction of the brain, we design two masks that leverage the spatial correlation and temporal repeatability of fMRI. Afterwards, temporal gated inception convolution and graph scalable inception convolution are proposed for the spatiotemporal autoencoder to enhance subtle cross-spatiotemporal variation and learn noise-suppressed signals. Furthermore, a spatiotemporal scalable cosine error with high selectivity for signal reconstruction is designed in SSL to guide the autoencoder to fit the fine-grained pathological features in an unsupervised manner. A total of 5,687 samples from four cross-population cohorts are involved. The accuracy of our model was 5.1% higher than the state-of-the-art models, which included four AD diagnostic models, four SSL strategies, and three multivariate time series models. The neuroimaging biomarkers were precisely localized to the abnormal brain regions, and correlated significantly with the cognitive scale and biomarkers (P$
The advancement of single-cell sequencing technology has promoted the generation of a large amount of single-cell transcriptional profiles, providing unprecedented opportunities to identify drug-resistant cell subpopulations within a tumor. However, few studies have focused on drug response prediction at single-cell level, and their performance remains suboptimal. This paper proposed scAdaDrug, a novel multi-source domain adaptation model powered by adaptive importance-aware representation learning to predict drug response of individual cells. We used a shared encoder to extract domain-invariant features related to drug response from multiple source domains by utilizing adversarial domain adaptation. Particularly, we introduced a plug-and-play module to generate importance-aware and mutually independent weights, which could adaptively modulate the latent representation of each sample in element-wise manner between source and target domains. Extensive experimental results showed that our model achieved state-of-the-art performance in predicting drug response on multiple independent datasets, including single-cell datasets derived from both cell lines and patient-derived xenografts (PDX) models, as well as clinical tumor patient cohorts. Moreover, the ablation experiments demonstrated our model effectively captured the underlying patterns determining drug response from multiple source domains.
Artificial intelligence techniques play a pivotal role in the accurate identification of drug-drug interaction (DDI) events, thereby informing clinical decisions and treatment regimens. While existing DDI prediction models have made significant progress by leveraging sequence features such as chemical substructures, targets, and enzymes, they often face limitations in integrating and effectively utilizing multi-modal drug representations. To address these limitations, this study proposes a novel multi-modal feature fusion model for DDI event prediction: MMDDI-SSE. Our approach integrates drug sequence modality with DDI graph representations through a novel architecture that employs static subgraph generation to capture structural properties. The model utilizes a graph autoencoder architecture to learn both local and global topological features from these subgraphs, while simultaneously processing diverse sequence-based characteristics including semantically enhanced pharmacodynamic features, chemical substructures, target proteins, and enzyme information. Through comprehensive evaluation on two distinct datasets, MMDDI-SSE demonstrates superior predictive performance compared to state-of-the-art baselines. Ablation studies further validate the effectiveness of each architectural component in enhancing DDI prediction accuracy.
CircRNA-miRNA interaction (CMI) plays a crucial role in the gene regulatory network of the cell. Numerous experiments have shown that abnormalities in CMI can impact molecular functions and physiological processes, leading to the occurrence of specific diseases. Current computational models for predicting CMI typically focus on local molecular entity relationships, thereby neglecting inherent molecular attributes and global structural information. To address these limitations, we propose a multi-feature fusion prediction model based on the transformer and graph attention network, named EGATCMI. Specifically, EGATCMI combines the transformer architecture with Word2vec to pre-train the sequence of circRNA and miRNA, capturing their sequence feature representation and sequence similarity. By leveraging the self-attention mechanism, EGATCMI extracts global structural feature from the CMI network. EGATCMI effectively integrates the obtained multi-feature for prediction, achieving AUC values of 0.9106 and 0.9470 on the CMI-9905 and CircBank datasets, respectively, outperforming existing methods. In case studies that the prediction of interactions between three miRNAs that are closely related to diseases and circRNAs, 8 out of 10 pairs were accurately predicted and validated. Extensive experimental results demonstrate the potential of EGATCMI as a reliable tool for candidate screening in biological investigations.
Background: Chatbots have demonstrated promising capabilities in Medicine, scoring passing grades for board examinations across various specialties. However, their tendency to express high levels of confidence in their responses, even when incorrect, poses a limitation to their utility in clinical settings. Objective: To examine whether token probabilities outperform chatbots' Expressed Confidence levels in predict-ing the accuracy of their responses to medical questions. Methods: Seven large languages models (LLMs), comprising both commercial (GPT-3.5, GPT-4 and GPT-4o) and open-source (Llama 3-8b, Llama 3-70b, Phi-3-Mini, and Phi-3-Medium), were prompted to respond to a set of 2,522 questions from the US Medical Licensing Examination (MedQA database). Addition-ally, the models rated their confidence from 0 to 100 and the token probability of each response was extracted. The models’ success rates were measured, and the predictive performances of both Ex-pressed Confidence and Response Token Probability in predicting response accuracy were evaluated using Area Under the Receiver Operating Characteristic Curve (AUROC), Adapted Calibration Error (ACE) and Brier score. Sensitivity analyses were conducted using additional questions sourced from other databases in English (MedMCQA, n=2,797), Chinese (MedQA Main-land China, n=3,413 and Taiwan, n=2,808), and French (FrMedMCQA, n=1,079). Results: Overall, mean accuracy ranged from 52.7%[50.8-54.7] for Phi-3-Mini to 87.6%[86.2-88.9] for GPT-4o. Across the US Medical Licensing Examination questions, all chatbots consistently expressed high levels of confidence in their responses (ranging from 90[90-90] for Llama 3-70B to 100[100–100] for GPT-3.5). However, Expressed Confidence failed to predict response accuracy (AUROC ranging from 0.52[0.50-0.53] for Phi 3 Mini to 0.68[0.65-0.71] for GPT-4o). In contrast, the Response Token Probability consistently outperformed Expressed Confidence for predicting response accuracy (AU-ROC ranging from 0.67[0.65-0.69] for Phi-3-Mini to 0.83[0.81-0.85] for Llama 3-70B, all p-values
Electronic health records are often extracted and combined with environmental data to conduct research or public health surveillance. However, to date, electronic health record systems do not integrate environmental data to aid real-time decisions that could mitigate the health impacts of environmental hazards, including the impacts of climate change. Pursuing this goal requires the enhancement of health record systems and the modification of financial incentives driving healthcare innovation and delivery.
Trained and validated on multimodal data from 14.5 million images from multicountry datasets, a foundation model is shown to increase diagnostic and referral accuracy of clinicians when used as an assistant in a trial involving 16 ophthalmologists and 668 patients.
Glioma is an aggressive brain tumor that requires challenging treatments. Tumor Treating Fields (TTFields), an FDA-approved therapy for glioblastoma (GBM), pleural mesothelioma, and platinum-refractory metastatic nonsmall cell lung cancer (in combination with PD-1/PD-L1 inhibitors or docetaxel), employs specific frequency electric fields to disrupt cell division and enhance treatment efficacy. However, their molecular mechanisms remain unclear. This study aimed to elucidate these mechanisms and optimize the therapeutic potential of TTFields through quantitative proteomics, phosphoproteomics, and glycoproteomics. Pathway analysis of the proteomics revealed that TTFields impact the cell cycle, DNA repair, autophagy, and DNA replication. Phosphoproteomic studies further demonstrated a marked decline in the activity of key kinases ABL1 and PDK1, while glycoproteomics highlighted disruptions in cell adhesion and ECM-receptor interactions. Notably, proteomic analysis identified an upregulation of PARP1 and BRD4 protein levels, suggesting a previously unrecognized resistance mechanism. Consistently, combining TTFields with inhibitors targeting these proteins significantly enhanced the treatment efficacy in U87 cells. Thus, this study uncovers comprehensive molecular mechanisms underlying TTFields' effects on GBM cells and supports the development of concomitant therapies to enhance treatment efficacy.
Elias Nogueira shares a strategic approach to modern microservice testing, detailing solutions for three core challenges. He discusses parallel execution for multiple databases using Testcontainers, creating reliable, sharable mock environments with WireMock, and handling asynchronous event delays with Awaitility, offering a blueprint for senior engineers.
Login
