Hasil untuk "Computer applications to medicine. Medical informatics"

Jiayue Guo, MPH, Lili You, PhD, Lu Liu, MPH et al.

Objective: To summarize the key intervention characteristics and evaluate the effectiveness and safety of digital therapeutics (DTx) in patients receiving oral anticoagulation, with effectiveness evaluated using time in therapeutic range (TTR), thromboembolic events, and mortality, and safety evaluated based on bleeding events. Patients and Methods: We searched PubMed, Embase, Web of Science, and the Cochrane Library from inception to June 20, 2025, and identified 10 randomized controlled trials involving 7237 patients. The criteria required studies to assess software-based DTx supporting anticoagulation management and report effectiveness or safety outcomes. Study quality was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation framework, and random-effects models were applied. Results: Digital therapeutics interventions were associated with a lower incidence of major bleeding than usual care: no clear differences in TTR, thromboembolic events, or mortality. Evidence quality ranged from very low to high. Secondary analyses showed more international normalized ratio testing with DTx; rehospitalization rates did not differ significantly between the groups. Sensitivity analysis changed TTR effect after excluding a study with enhanced control, but other outcomes remained unchanged. Conclusion: Digital therapeutics interventions for anticoagulation management improve safety outcomes, particularly reducing major bleeding, and with greater monitoring intensity. Larger, long-term trials are needed to confirm the clinical benefits and evaluate cost-effectiveness. Trial Registration: PROSPERO Identifier: CRD420251107441.

Konstantinos Keremis, Eleni Vrochidou, George A. Papakostas

The ability of deep learning models to maintain consistent performance under image transformations-termed invariances, is critical for reliable deployment across diverse computer vision applications. This study presents a comprehensive empirical evaluation of modern convolutional neural networks (CNNs) and vision transformers (ViTs) concerning four fundamental types of image invariances: blur, noise, rotation, and scale. We analyze a curated selection of thirty models across three common vision tasks, object localization, recognition, and semantic segmentation, using benchmark datasets including COCO, ImageNet, and a custom segmentation dataset. Our experimental protocol introduces controlled perturbations to test model robustness and employs task-specific metrics such as mean Intersection over Union (mIoU), and classification accuracy (Acc) to quantify models’ performance degradation. Results indicate that while ViTs generally outperform CNNs under blur and noise corruption in recognition tasks, both model families exhibit significant vulnerabilities to rotation and extreme scale transformations. Notably, segmentation models demonstrate higher resilience to geometric variations, with SegFormer and Mask2Former emerging as the most robust architectures. These findings challenge prevailing assumptions regarding model robustness and provide actionable insights for designing vision systems capable of withstanding real-world input variability.

Atiya Mahmood, Farinaz Rikhtehgaran, Rojan Nasiri et al.

BackgroundThe prevalence of sensory, cognitive, and mobility disabilities in Canada underscores the need to address environmental barriers. This study adapts and validates the Stakeholders’ Walkability/Wheelability Audit in Neighborhoods (SWAN) tool to assess the challenges the built environment poses for individuals with disabilities, aiming to inform policy changes for accessibility and inclusivity. ObjectiveThis study aims to (1) adapt the SWAN tool for those with hearing, vision, or cognitive disabilities; (2) validate SWAN tool for researching environmental barriers for people with disabilities, including older adults; and (3) offer insights for policy changes in the built environment, contributing to literature and guiding future research. MethodsThe study uses a community-based research approach, carried out over 4 phases within an 18-month period in British Columbia. Phase 1 includes adapting and pilot-testing of the SWAN tool. In Phase 2, street intersections are identified for data collection using Geographic Information System tools and consultations with municipal officials. Phase 3 involves recruiting participants across four disability categories. The final phase includes analyzing the data and disseminating findings. ResultsData collection concluded in September 2024, involving 80 eligible participants across four streams in preidentified hotspots. The results are expected to be published in March 2025. To date, data collection is ongoing, and we are currently in the process of data analysis. ConclusionsThis study will contribute to the growing body of research on built environment accessibility by adapting the SWAN tool for individuals with diverse disabilities. By identifying key barriers in urban spaces, the study aims to inform policy changes that will lead to more inclusive, accessible, and safe urban environments for all individuals.

Daniel Wolf, Heiko Hillenhagen, Billurvan Taskin et al.

Clinical decision-making relies heavily on understanding relative positions of anatomical structures and anomalies. Therefore, for Vision-Language Models (VLMs) to be applicable in clinical practice, the ability to accurately determine relative positions on medical images is a fundamental prerequisite. Despite its importance, this capability remains highly underexplored. To address this gap, we evaluate the ability of state-of-the-art VLMs, GPT-4o, Llama3.2, Pixtral, and JanusPro, and find that all models fail at this fundamental task. Inspired by successful approaches in computer vision, we investigate whether visual prompts, such as alphanumeric or colored markers placed on anatomical structures, can enhance performance. While these markers provide moderate improvements, results remain significantly lower on medical images compared to observations made on natural images. Our evaluations suggest that, in medical imaging, VLMs rely more on prior anatomical knowledge than on actual image content for answering relative position questions, often leading to incorrect conclusions. To facilitate further research in this area, we introduce the MIRP , Medical Imaging Relative Positioning, benchmark dataset, designed to systematically evaluate the capability to identify relative positions in medical images.

Mutahar Safdar, Gentry Wood, Max Zimmermann et al.

Rapid and reliable qualification of advanced materials remains a bottleneck in industrial manufacturing, particularly for heterogeneous structures produced via non-conventional additive manufacturing processes. This study introduces a novel framework that links microstructure informatics with a range of expert characterization knowledge using customized and hybrid vision-language representations (VLRs). By integrating deep semantic segmentation with pre-trained multi-modal models (CLIP and FLAVA), we encode both visual microstructural data and textual expert assessments into shared representations. To overcome limitations in general-purpose embeddings, we develop a customized similarity-based representation that incorporates both positive and negative references from expert-annotated images and their associated textual descriptions. This allows zero-shot classification of previously unseen microstructures through a net similarity scoring approach. Validation on an additively manufactured metal matrix composite dataset demonstrates the framework's ability to distinguish between acceptable and defective samples across a range of characterization criteria. Comparative analysis reveals that FLAVA model offers higher visual sensitivity, while the CLIP model provides consistent alignment with the textual criteria. Z-score normalization adjusts raw unimodal and cross-modal similarity scores based on their local dataset-driven distributions, enabling more effective alignment and classification in the hybrid vision-language framework. The proposed method enhances traceability and interpretability in qualification pipelines by enabling human-in-the-loop decision-making without task-specific model retraining. By advancing semantic interoperability between raw data and expert knowledge, this work contributes toward scalable and domain-adaptable qualification strategies in engineering informatics.

Ye Du, Nanxi Yu, Shujun Wang

Vision-language foundation models (VLMs) have shown great potential in feature transfer and generalization across a wide spectrum of medical-related downstream tasks. However, fine-tuning these models is resource-intensive due to their large number of parameters. Prompt tuning has emerged as a viable solution to mitigate memory usage and reduce training time while maintaining competitive performance. Nevertheless, the challenge is that existing prompt tuning methods cannot precisely distinguish different kinds of medical concepts, which miss essentially specific disease-related features across various medical imaging modalities in medical image classification tasks. We find that Large Language Models (LLMs), trained on extensive text corpora, are particularly adept at providing this specialized medical knowledge. Motivated by this, we propose incorporating LLMs into the prompt tuning process. Specifically, we introduce the CILMP, Conditional Intervention of Large Language Models for Prompt Tuning, a method that bridges LLMs and VLMs to facilitate the transfer of medical knowledge into VLM prompts. CILMP extracts disease-specific representations from LLMs, intervenes within a low-rank linear subspace, and utilizes them to create disease-specific prompts. Additionally, a conditional mechanism is incorporated to condition the intervention process on each individual medical image, generating instance-adaptive prompts and thus enhancing adaptability. Extensive experiments across diverse medical image datasets demonstrate that CILMP consistently outperforms state-of-the-art prompt tuning methods, demonstrating its effectiveness. Code is available at https://github.com/usr922/cilmp.

Muhammad Zubair, Muzammil Hussai, Mousa Ahmad Al-Bashrawi et al.

Multi-modal medical image fusion (MMIF) is increasingly recognized as an essential technique for enhancing diagnostic precision and facilitating effective clinical decision-making within computer-aided diagnosis systems. MMIF combines data from X-ray, MRI, CT, PET, SPECT, and ultrasound to create detailed, clinically useful images of patient anatomy and pathology. These integrated representations significantly advance diagnostic accuracy, lesion detection, and segmentation. This comprehensive review meticulously surveys the evolution, methodologies, algorithms, current advancements, and clinical applications of MMIF. We present a critical comparative analysis of traditional fusion approaches, including pixel-, feature-, and decision-level methods, and delves into recent advancements driven by deep learning, generative models, and transformer-based architectures. A critical comparative analysis is presented between these conventional methods and contemporary techniques, highlighting differences in robustness, computational efficiency, and interpretability. The article addresses extensive clinical applications across oncology, neurology, and cardiology, demonstrating MMIF's vital role in precision medicine through improved patient-specific therapeutic outcomes. Moreover, the review thoroughly investigates the persistent challenges affecting MMIF's broad adoption, including issues related to data privacy, heterogeneity, computational complexity, interpretability of AI-driven algorithms, and integration within clinical workflows. It also identifies significant future research avenues, such as the integration of explainable AI, adoption of privacy-preserving federated learning frameworks, development of real-time fusion systems, and standardization efforts for regulatory compliance.

Simran Saggu, Hirad Daneshvar, Reza Samavi et al.

Abstract Background The proportion of Canadian youth seeking mental health support from an emergency department (ED) has risen in recent years. As EDs typically address urgent mental health crises, revisiting an ED may represent unmet mental health needs. Accurate ED revisit prediction could aid early intervention and ensure efficient healthcare resource allocation. We examine the potential increased accuracy and performance of graph neural network (GNN) machine learning models compared to recurrent neural network (RNN), and baseline conventional machine learning and regression models for predicting ED revisit in electronic health record (EHR) data. Methods This study used EHR data for children and youth aged 4–17 seeking services at McMaster Children’s Hospital’s Child and Youth Mental Health Program outpatient service to develop and evaluate GNN and RNN models to predict whether a child/youth with an ED visit had an ED revisit within 30 days. GNN and RNN models were developed and compared against conventional baseline models. Model performance for GNN, RNN, XGBoost, decision tree and logistic regression models was evaluated using F1 scores. Results The GNN model outperformed the RNN model by an F1-score increase of 0.0511 and the best performing conventional machine learning model by an F1-score increase of 0.0470. Precision, recall, receiver operating characteristic (ROC) curves, and positive and negative predictive values showed that the GNN model performed the best, and the RNN model performed similarly to the XGBoost model. Performance increases were most noticeable for recall and negative predictive value than for precision and positive predictive value. Conclusions This study demonstrates the improved accuracy and potential utility of GNN models in predicting ED revisits among children and youth, although model performance may not be sufficient for clinical implementation. Given the improvements in recall and negative predictive value, GNN models should be further explored to develop algorithms that can inform clinical decision-making in ways that facilitate targeted interventions, optimize resource allocation, and improve outcomes for children and youth.

Jethel Hernandez, Stephanie Batio, Rebecca Mullen Lovett et al.

Introduction: Missed visits have been estimated to cost the U.S. healthcare system $50 billion annually and have been linked to healthcare inefficiency, higher rates of emergency department visits, and worse outcomes. COVID-19 disrupted existing outpatient healthcare utilization patterns. In our study, we sought to examine the frequency of missed outpatient visits over the course of the COVID-19 pandemic and to examine patient-level characteristics associated with non-attendance. Methods: This study utilized data from a longitudinal cohort study (the Chicago COVID-19 Comorbidities (C3) study). C3 participants were enrollees in 1 of 4 active, “parent” studies; they were rapidly enrolled in C3 at the onset of the pandemic. Multiple waves of telephone-based interviews were conducted to collect experiences with the pandemic, as well as socio-demographic and health characteristics, health literacy, patient activation, and depressive and anxiety symptoms. For the current analysis, data from waves 3 to 8 (05/01/20-05/19/22) were analyzed. Participants included 845 English or Spanish-speaking adults with 1 or more chronic conditions. Results: The percentage of participants reporting missed visits due to COVID-19 across study waves ranged from 3.1 to 22.4%. Overall, there was a decline in missed visits over time. No participant sociodemographic or health characteristic was consistently associated with missed visits across the study waves. In bivariate and multivariate analysis, only patient-reported anxiety was significantly associated with missed visits across all study waves. Conclusion: Findings reveal that anxiety was consistently associated with missed visits during the COVID-19 pandemic, but not sociodemographic or health characteristics. Results can inform future public health initiatives to reduce absenteeism by considering patients’ emotional state during times of uncertainty.

Sufen Ren, Yule Hu, Shengchao Chen et al.

Medical image classification plays a crucial role in computer-aided clinical diagnosis. While deep learning techniques have significantly enhanced efficiency and reduced costs, the privacy-sensitive nature of medical imaging data complicates centralized storage and model training. Furthermore, low-resource healthcare organizations face challenges related to communication overhead and efficiency due to increasing data and model scales. This paper proposes a novel privacy-preserving medical image classification framework based on federated learning to address these issues, named FedMIC. The framework enables healthcare organizations to learn from both global and local knowledge, enhancing local representation of private data despite statistical heterogeneity. It provides customized models for organizations with diverse data distributions while minimizing communication overhead and improving efficiency without compromising performance. Our FedMIC enhances robustness and practical applicability under resource-constrained conditions. We demonstrate FedMIC's effectiveness using four public medical image datasets for classical medical image classification tasks.

Kishu Gupta, Vinaytosh Mishra, Aaisha Makkar

Healthcare has witnessed an increased digitalization in the post-COVID world. Technologies such as the medical internet of things and wearable devices are generating a plethora of data available on the cloud anytime from anywhere. This data can be analyzed using advanced artificial intelligence techniques for diagnosis, prognosis, or even treatment of disease. This advancement comes with a major risk to protecting and securing protected health information (PHI). The prevailing regulations for preserving PHI are neither comprehensive nor easy to implement. The study first identifies twenty activities crucial for privacy and security, then categorizes them into five homogeneous categories namely: $\complement_1$ (Policy and Compliance Management), $\complement_2$ (Employee Training and Awareness), $\complement_3$ (Data Protection and Privacy Control), $\complement_4$ (Monitoring and Response), and $\complement_5$ (Technology and Infrastructure Security) and prioritizes these categories to provide a framework for the implementation of privacy and security in a wise manner. The framework utilized the Delphi Method to identify activities, criteria for categorization, and prioritization. Categorization is based on the Density-Based Spatial Clustering of Applications with Noise (DBSCAN), and prioritization is performed using a Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS). The outcomes conclude that $\complement_3$ activities should be given first preference in implementation and followed by $\complement_1$ and $\complement_2$ activities. Finally, $\complement_4$ and $\complement_5$ should be implemented. The prioritized view of identified clustered healthcare activities related to security and privacy, are useful for healthcare policymakers and healthcare informatics professionals.

Davide Belluomo, Tiziana Calamoneri, Giacomo Paesani et al.

We present a new unified graph-based representation of medical data, combining genetic information and medical records of patients with medical knowledge via a unique knowledge graph. This approach allows us to infer meaningful information and explanations that would be unavailable by looking at each data set separately. The systematic use of different databases, managed throughout the built knowledge graph, gives new insights toward a better understanding of oncology medicine. Indeed, we reduce some useful medical tasks to well-known problems in theoretical computer science for which efficient algorithms exist.

Théo Sourget, Michelle Hestbek-Møller, Amelia Jiménez-Sánchez et al.

The development of larger models for medical image analysis has led to increased performance. However, it also affected our ability to explain and validate model decisions. Models can use non-relevant parts of images, also called spurious correlations or shortcuts, to obtain high performance on benchmark datasets but fail in real-world scenarios. In this work, we challenge the capacity of convolutional neural networks (CNN) to classify chest X-rays and eye fundus images while masking out clinically relevant parts of the image. We show that all models trained on the PadChest dataset, irrespective of the masking strategy, are able to obtain an Area Under the Curve (AUC) above random. Moreover, the models trained on full images obtain good performance on images without the region of interest (ROI), even superior to the one obtained on images only containing the ROI. We also reveal a possible spurious correlation in the Chaksu dataset while the performances are more aligned with the expectation of an unbiased model. We go beyond the performance analysis with the usage of the explainability method SHAP and the analysis of embeddings. We asked a radiology resident to interpret chest X-rays under different masking to complement our findings with clinical knowledge. Our code is available at https://github.com/TheoSourget/MMC_Masking and https://github.com/TheoSourget/MMC_Masking_EyeFundus

Ariana-Anamaria CORDOȘ

The Early Warning Scores (EWSs) are tools for bedside evaluation based on five physiological parameters: systolic pressure, pulse, respiratory rate, body temperature and AVPU (alert, voice, pain, unresponsive) score. EWSs have been used in many hospital departments, including general wards, intensive care units, or emergency rooms. Several iterations of EWSs have been developed with varying levels of sensitivity and specificity for use in different populations. The aim of this research was to understand the benefits of using these tools. This systematic review followed the Preferred Reporting Items for Systemic Reviews and Meta-Analyses (PRISMA) guidelines. This study included literature published in PubMed under the MeSH term “early warning score”. The search was performed on 12th July 2023. No restrictions on the types of articles were imposed. Considering the language limitations of the study investigators, only studies available in English were retained. A total of 392 items were retained. The articles’ titles and abstracts were screened to investigate whether the benefits of using early warning scores were the topic. It resulted in 286 relevant articles. Two major strength categories have been identified: the patient oriented outcome and the healthcare personnel oriented benefits. The patient oriented outcome indicators that these tools can predict are: transfers to the intensive care unit, sepsis, in hospital cardiac arrest, mortality, or disease specific clinical deterioration. Multiple healthcare personnel oriented strengths of these tools have been identified, including their simplicity and ability to standardize communication and reduce staff work burden, especially if they are continuously electronically recorded. The research highlights the importance of the integration of data-driven models into personalized care and represents an opportunity to inform biomedical and health informatics research on designing and evaluating EWS-based clinical interventions.

Xinrun Yang, Gaofei Jiang, Yaozhong Zhang et al.

Abstract Bacterial pathogens are one of the major threats to biosafety and environmental health, and advanced assessment is a prerequisite to combating bacterial pathogens. Currently, 16S rRNA gene sequencing is efficient in the open‐view detection of bacterial pathogens. However, the taxonomic resolution and applicability of this method are limited by the domain‐specific pathogen database, taxonomic profiling method, and sequencing target of 16S variable regions. Here, we present a pipeline of multiple bacterial pathogen detection (MBPD) to identify the animal, plant, and zoonotic pathogens. MBPD is based on a large, curated database of the full‐length 16S genes of 1986 reported bacterial pathogen species covering 72,685 sequences. In silico comparison allowed MBPD to provide the appropriate similarity threshold for both full‐length and variable‐region sequencing platforms, while the subregion of V3−V4 (mean: 88.37%, accuracy rate compared to V1−V9) outperformed other variable regions in pathogen identification compared to full‐length sequencing. Benchmarking on real data sets suggested the superiority of MBPD in a broader range of pathogen detections compared with other methods, including 16SPIP and MIP. Beyond detecting the known causal agent of animal, human, and plant diseases, MBPD is capable of identifying cocontaminating pathogens from biological and environmental samples. Overall, we provide a MBPD pipeline for agricultural, veterinary, medical, and environmental monitoring to achieve One Health.

Chloe Engler Hart, Daniel Ence, David Healey et al.

Abstract Better understanding the transcriptomic response produced by a compound perturbing its targets can shed light on the underlying biological processes regulated by the compound. However, establishing the relationship between the induced transcriptomic response and the target of a compound is non-trivial, partly because targets are rarely differentially expressed. Therefore, connecting both modalities requires orthogonal information (e.g., pathway or functional information). Here, we present a comprehensive study aimed at exploring this relationship by leveraging thousands of transcriptomic experiments and target data for over 2000 compounds. Firstly, we confirm that compound-target information does not correlate as expected with the transcriptomic signatures induced by a compound. However, we reveal how the concordance between both modalities increases by connecting pathway and target information. Additionally, we investigate whether compounds that target the same proteins induce a similar transcriptomic response and conversely, whether compounds with similar transcriptomic responses share the same target proteins. While our findings suggest that this is generally not the case, we did observe that compounds with similar transcriptomic profiles are more likely to share at least one protein target and common therapeutic applications. Finally, we demonstrate how to exploit the relationship between both modalities for mechanism of action deconvolution by presenting a case scenario involving a few compound pairs with high similarity.

Heather M. Whitney, Natalie Baughan, Kyle J. Myers et al.

Purpose: The Medical Imaging and Data Resource Center (MIDRC) open data commons was launched to accelerate the development of artificial intelligence (AI) algorithms to help address the COVID-19 pandemic. The purpose of this study was to quantify longitudinal representativeness of the demographic characteristics of the primary imaging dataset compared to the United States general population (US Census) and COVID-19 positive case counts from the Centers for Disease Control and Prevention (CDC). Approach: The Jensen Shannon distance (JSD) was used to longitudinally measure the similarity of the distribution of (1) all unique patients in the MIDRC data to the 2020 US Census and (2) all unique COVID-19 positive patients in the MIDRC data to the case counts reported by the CDC. The distributions were evaluated in the demographic categories of age at index, sex, race, ethnicity, and the intersection of race and ethnicity. Results: Representativeness the MIDRC data by ethnicity and the intersection of race and ethnicity was impacted by the percentage of CDC case counts for which data in these categories is not reported. The distributions by sex and race have retained their level of representativeness over time. Conclusion: The representativeness of the open medical imaging datasets in the curated public data commons at MIDRC has evolved over time as both the number of contributing institutions and overall number of subjects has grown. The use of metrics such as the JSD support measurement of representativeness, one step needed for fair and generalizable AI algorithm development.

Ahmad Wisnu Mulyadi, Heung-Il Suk

Extensive adoption of electronic health records (EHRs) offers opportunities for their use in various downstream clinical analyses. To accomplish this purpose, enriching an EHR cohort with external knowledge (e.g., standardized medical ontology and wealthy semantics) could help us reveal more comprehensive insights via a spectrum of informative relations among medical codes. Nevertheless, harnessing those beneficial interconnections was scarcely exercised, especially in the medication recommendation task. This study proposes a novel Knowledge-Induced Medicine Prescribing Network (KindMed) to recommend medicines by inducing knowledge from myriad medical-related external sources upon the EHR cohort and rendering interconnected medical codes as medical knowledge graphs (KGs). On top of relation-aware graph representation learning to obtain an adequate embedding over such KGs, we leverage hierarchical sequence learning to discover and fuse temporal dynamics of clinical (i.e., diagnosis and procedures) and medicine streams across patients' historical admissions to foster personalized recommendations. Eventually, we employ attentive prescribing that accounts for three essential patient representations, i.e., a summary of joint historical medical records, clinical progression, and the current clinical state of patients. We validated the effectiveness of our KindMed on the augmented real-world EHR cohorts, achieving improved recommendation performances against a handful of graph-driven baselines.

Hubert Baniecki, Bartlomiej Sobieski, Patryk Szatkowski et al.

Time-to-event prediction, e.g. cancer survival analysis or hospital length of stay, is a highly prominent machine learning task in medical and healthcare applications. However, only a few interpretable machine learning methods comply with its challenges. To facilitate a comprehensive explanatory analysis of survival models, we formally introduce time-dependent feature effects and global feature importance explanations. We show how post-hoc interpretation methods allow for finding biases in AI systems predicting length of stay using a novel multi-modal dataset created from 1235 X-ray images with textual radiology reports annotated by human experts. Moreover, we evaluate cancer survival models beyond predictive performance to include the importance of multi-omics feature groups based on a large-scale benchmark comprising 11 datasets from The Cancer Genome Atlas (TCGA). Model developers can use the proposed methods to debug and improve machine learning algorithms, while physicians can discover disease biomarkers and assess their significance. We hope the contributed open data and code resources facilitate future work in the emerging research direction of explainable survival analysis.

Junde Wu, Wei Ji, Yuanpei Liu et al.

The Segment Anything Model (SAM) has recently gained popularity in the field of image segmentation due to its impressive capabilities in various segmentation tasks and its prompt-based interface. However, recent studies and individual experiments have shown that SAM underperforms in medical image segmentation, since the lack of the medical specific knowledge. This raises the question of how to enhance SAM's segmentation capability for medical images. In this paper, instead of fine-tuning the SAM model, we propose the Medical SAM Adapter (Med-SA), which incorporates domain-specific medical knowledge into the segmentation model using a light yet effective adaptation technique. In Med-SA, we propose Space-Depth Transpose (SD-Trans) to adapt 2D SAM to 3D medical images and Hyper-Prompting Adapter (HyP-Adpt) to achieve prompt-conditioned adaptation. We conduct comprehensive evaluation experiments on 17 medical image segmentation tasks across various image modalities. Med-SA outperforms several state-of-the-art (SOTA) medical image segmentation methods, while updating only 2\% of the parameters. Our code is released at https://github.com/KidsWithTokens/Medical-SAM-Adapter.