Medical Diagnostic System Research Articles

Resilience-aware MLOps for AI-based medical diagnostic system.

The healthcare sector demands a higher degree of responsibility, trustworthiness, and accountability when implementing Artificial Intelligence (AI) systems. Machine learning operations (MLOps) for AI-based medical diagnostic systems are primarily focused on aspects such as data quality and confidentiality, bias reduction, model deployment, performance monitoring, and continuous improvement. However, so far, MLOps techniques do not take into account the need to provide resilience to disturbances such as adversarial attacks, including fault injections, and drift, including out-of-distribution. This article is concerned with the MLOps methodology that incorporates the steps necessary to increase the resilience of an AI-based medical diagnostic system against various kinds of disruptive influences. Post-hoc resilience optimization, post-hoc predictive uncertainty calibration, uncertainty monitoring, and graceful degradation are incorporated as additional stages in MLOps. To optimize the resilience of the AI based medical diagnostic system, additional components in the form of adapters and meta-adapters are utilized. These components are fine-tuned during meta-training based on the results of adaptation to synthetic disturbances. Furthermore, an additional model is introduced for post-hoc calibration of predictive uncertainty. This model is trained using both in-distribution and out-of-distribution data to refine predictive confidence during the inference mode. The structure of resilience-aware MLOps for medical diagnostic systems has been proposed. Experimentally confirmed increase of robustness and speed of adaptation for medical image recognition system during several intervals of the system's life cycle due to the use of resilience optimization and uncertainty calibration stages. The experiments were performed on the DermaMNIST dataset, BloodMNIST and PathMNIST. ResNet-18 as a representative of convolutional networks and MedViT-T as a representative of visual transformers are considered. It is worth noting that transformers exhibited lower resilience than convolutional networks, although this observation may be attributed to potential imperfections in the architecture of adapters and meta-adapters. The main novelty of the suggested resilience-aware MLOps methodology and structure lie in the separating possibilities and activities on creating a basic model for normal operating conditions and ensuring its resilience and trustworthiness. This is significant for the medical applications as the developer of the basic model should devote more time to comprehending medical field and the diagnostic task at hand, rather than specializing in system resilience. Resilience optimization increases robustness to disturbances and speed of adaptation. Calibrated confidences ensure the recognition of a portion of unabsorbed disturbances to mitigate their impact, thereby enhancing trustworthiness.

Development of a Diabetes Mellitus Diagnostic System Using Self-Organizing Map Algorithm: A Machine Learning Approach

Delivery of Health care services in developing nations has posed a huge problem to the world at large. The United Nations and the World Health Organization have been on the front burner sorting for ways of improving these problems to abate the yearly mortality rates which are caused largely by inadequate health facilities, poor technical know-how, and poor health care administration. One disease that has a high number of patients is diabetes. In Nigeria, out of a population of 200 million, diabetes kills over 2% yearly. To reduce this menace, early diagnosis and awareness are important. And automation of the medical diagnostic system is one of the sure ways of achieving these feet. This paper explores the potential of a self-organizing map algorithm; a machine learning technique in the development of a diabetes mellitus diagnostic system (DMDS). Data collected from 120 patients from the University of Port Harcourt Teaching Hospital (UPTH) was used in the training and validation of the model. The confusion matrix formula was used in testing the sensitivity and accuracy of the model which yielded 75.63% and 87.2% respectively which are within the accepted range, predefined by expert physicians. Keywords Artificial Intelligence, self-organizing map, diagnosis, neural networks, diabetes mellitus

Fully automated diagnosis of thyroid nodule ultrasound using brain-inspired inference

The interpretability of artificial intelligence (AI) based medical diagnostic systems is crucial to make the diagnosis adequately convincible. Deep learning has been extensively investigated and utilized in the area of medical assistance diagnosis in recent decades due to its outstanding performance and objective prediction. However, a huge semantic chasm dividing clinicians and unexplainable deep models emerges. Here we design a brain-inspired inference framework from medical images to explainable features, then to the final diagnostic conclusions. The fast thinking module is responsible for recognizing medical features in ultrasound (US) images, and the slow-thinking module builds a model for inferring from medical features to diagnostic results by constructing a knowledge graph of medical features with tensor decomposition. The whole model infers through intuition and thinking like a human being, and gives the recognized medical image features while inferring the diagnosis, which greatly improves the interpretability of the model. We conducted studies on thyroid cancer diagnoses using US images. The American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) characteristics are employed as medical features describing thyroid nodules. Our brain-inspired medical inference framework outperforms commonly used deep learning algorithms, with an AUC score of 0.963 (95% confidential interval (CI)=0.923–1.000) for thyroid US image diagnosis. Results indicate that our framework improves diagnostic objectivity and interpretability while providing performance that is better than deep models. Our proposed brain-inspired medical inference framework could improve the efficiency of diagnosis and our technique is performant, objective and interpretable.

A dynamic fuzzy rule-based inference system using fuzzy inference with semantic reasoning

The challenge of making flexible, standard, and early medical diagnoses is significant. However, some limitations are not fully overcome. First, the diagnosis rules established by medical experts or learned from a trained dataset prove static and too general. It leads to decisions that lack adaptive flexibility when finding new circumstances. Secondly, medical terminological interoperability is highly critical. It increases realism and medical progress and avoids isolated systems and the difficulty of data exchange, analysis, and interpretation. Third, criteria for diagnosis are often heterogeneous and changeable. It includes symptoms, patient history, demographic, treatment, genetics, biochemistry, and imaging. Symptoms represent a high-impact indicator for early detection. It is important that we deal with these symptoms differently, which have a great relationship with semantics, vary widely, and have linguistic information. This negatively affects early diagnosis decision-making. Depending on the circumstances, the diagnosis is made solo on imaging and some medical tests. In this case, although the accuracy of the diagnosis is very high, can these decisions be considered an early diagnosis or prove the condition is deteriorating? Our contribution in this paper is to present a real medical diagnostic system based on semantics, fuzzy, and dynamic decision rules. We attempt to integrate ontology semantics reasoning and fuzzy inference. It promotes fuzzy reasoning and handles knowledge representation problems. In complications and symptoms, ontological semantic reasoning improves the process of evaluating rules in terms of interpretability, dynamism, and intelligence. A real-world case study, ADNI, is presented involving the field of Alzheimer’s disease (AD). The proposed system has indicated the possibility of the system to diagnose AD with an accuracy of 97.2%, 95.4%, 94.8%, 93.1%, and 96.3% for AD, LMCI, EMCI, SMC, and CN respectively.

Lung Segmentation from Chest X-Ray Images Using Deeplabv3plus-Based CNN Model

As a result of technological advancements, a variety of medical diagnostic systems have grown rapidly to support the healthcare sectors. Over the past years, there has been considerable interest in utilizing deep learning algorithms for the proactive diagnosis of multiple diseases. In most cases, Coronavirus (COVID-19) and tuberculosis (TB) are diagnosed through the examination of pulmonary X-rays. Deep learning algorithms can identify tuberculosis with an almost medical-grade level of consistency by extracting the lung regions in the X-ray images. The probability of tuberculosis detection is increased when classification algorithms are applied to segmented lungs rather than the entire X-ray. The main focus of this paper is to execute lung segmentation from X-ray images using the deeplabv3plus CNN-based semantic segmentation model. In other CNN architectures, the feature resolution diminishes as the network becomes deeper due to the use of sequential convolutions with pooling or striding within the down-sampling stage. To tackle this drawback, deeplabv3plus incorporates "Atrous Convolution" in addition to modifying the pooling and convolutional striding components of the backbone. The experimental results were: an accuracy of 97.42%, a Jaccard index of 93.49%, and a dice coefficient of 96.63%. We also conduct an extensive comparison between the deeplabv3plus segmentation model and other benchmark segmentation architectures. The results prove the ability of the deeplabv3plus model to achieve precise lung segmentation from X-ray images.

An efficient hybrid extreme learning machine and evolutionary framework with applications for medical diagnosis

AbstractIntegrating machine learning techniques into medical diagnostic systems holds great promise for enhancing disease identification and treatment. Among the various options for training such systems, the extreme learning machine (ELM) stands out due to its rapid learning capability and computational efficiency. However, the random selection of input weights and hidden neuron biases in the ELM can lead to suboptimal performance. To address this issue, our study introduces a novel approach called modified Harris hawks optimizer (MHHO) to optimize these parameters in ELM for medical classification tasks. By applying the MHHO‐based method to seven medical datasets, our experimental results demonstrate its superiority over seven other evolutionary‐based ELM trainer models. The findings strongly suggest that the MHHO approach can serve as a valuable tool for enhancing the performance of ELM in medical diagnosis.

Using Artificial Intelligence (AI) Technology in the Health Sector has Several Goals

In a broad sense, artificial intelligence (AI) refers to any computer or system behavior that resembles human behavior. A subfield of artificial intelligence called "machine learning" enables computers to learn from data without explicit human programming. One of the most significant contemporary trends in global healthcare is the use of artificial intelligence (AI) technologies in medicine. Technologies based on artificial intelligence are profoundly transforming the world's healthcare system, enabling a dramatic reconstruction of the medical diagnostics system while simultaneously lowering healthcare expenditures. Identifying the class of diseases to which a disease belongs is crucial before treating it. It is feasible to categorize the type of disease based on the feature space of the condition. Algorithms for machine learning can address this issue.

ICSH review of internal quality control policy for blood cell counters.

This paper is a report of an ICSH review of policies and practices for internal quality control (IQC) policy for haematology cell counters among regulatory bodies, cell counter manufacturers and diagnostic laboratories. It includes a discussion of the study findings and links to separate ICSH guidance for such policies and practices. The application of internal quality control (IQC) methods is an essential pre-requisite for all clinical laboratory testing including the blood count (Full Blood Count, FBC, or Complete Blood Count, CBC). The ICSH has gathered information regarding the current state of practice through review of published guidance from regulatory bodies, a questionnaire to six major cell counter manufacturers (Abbott Diagnostics, Beckman Coulter, Horiba Medical Diagnostic Instruments & Systems, Mindray Medical International, Siemens Healthcare Diagnostics and Sysmex Corporation) and a survey issued to 191 diagnostic laboratories in four countries (China, Republic of Ireland, Spain and the United Kingdom) on their IQC practice and approach to use of commercial IQC materials. This has revealed diversity both in guidance and in practice around the world. There is diversity in guidance from regulatory organizations in regard to IQC methods each recommends, clinical levels to use and frequency to run commercial controls, and finally recommended sources of commercial controls. The diversity in practice among clinical laboratories spans the areas of IQC methods used, derivation of target values and action limits used with control materials, and frequency of running commercial controls materials. These findings and their implications for IQC Practice are discussed in this paper. They are used to inform a separate guidance document, which proposes a harmonized approach to address the issues faced by diagnostic laboratories.

CONCEPTS OF CREATING AN INTELLIGENT MEDICAL DIAGNOSTIC SYSTEM TO ASSIST IN THE WORK AND TRAINING OF DOCTORS BASED ON ARTIFICIAL INTELLIGENCE

Abstract. The article is devoted to the development of intelligent medical diagnostic systems using information systems and technologies. The article provides an overview of the current state and development of information systems, technologies and artificial intelligence in the medical field, analyzes existing intelligent diagnostic systems in medicine and cardiology in particular, and proves the need to create intelligent systems to help primary care physicians and cardiologists. The purpose of the publication was to investigate the current state of intelligent medical diagnostic systems, analyze the shortcomings of such systems, determine the feasibility of creating a new diagnostic system, and formulate principles and criteria for its operation. The scientific novelty of the article is a new approach for diagnosing and treating patients, as well as training medical professionals using an intelligent medical system based on artificial intelligence. Leveraging advanced AI algorithms, the system analyzes vast datasets encompassing patient records, medical literature, and real-time clinical data to provide accurate and timely diagnostic insights. The intelligent medical diagnostic system operates as a supportive tool, aiding doctors in the diagnostic process by offering refined suggestions, identifying potential anomalies, and recommending personalized treatment plans. By harnessing machine learning and deep learning techniques, the system continuously adapts and evolves, learning from each diagnostic scenario and refining its predictive accuracy over time. Through interactive simulations and case-based learning modules, aspiring and practicing doctors can engage in immersive, realistic scenarios, honing their diagnostic skills and expanding their knowledge base. The disadvantages of this article include only a theoretical approach to the formation of concepts and tasks for an intelligent medical system, and a review of existing systems.

Diseases diagnosis based on artificial intelligence and ensemble classification

BackgroundIn recent years, Computer Aided Diagnosis (CAD) has become an important research area that attracted a lot of researchers. In medical diagnostic systems, several attempts have been made to build and enhance CAD applications to avoid errors that can cause dangerously misleading medical treatments. The most exciting opportunity for promoting the performance of CAD system can be accomplished by integrating Artificial Intelligence (AI) in medicine. This allows the effective automation of traditional manual workflow, which is slow, inaccurate and affected by human errors. AimsThis paper aims to provide a complete Computer Aided Disease Diagnosis (CAD2) strategy based on Machine Learning (ML) techniques that can help clinicians to make better medical decisions. MethodsThe proposed CAD2 consists of three main sequential phases, namely; (i) Outlier Rejection Phase (ORP), (ii) Feature Selection Phase (FSP), and (iii) Classification Phase (CP). ORP is implemented to reject outliers using new Outlier Rejection Technique (ORT) that contains two sequential stages called Fast Outlier Rejection (FOR) and Accurate Outlier Rejection (AOR). The most informative features are selected through FSP using Hybrid Selection Technique (HST). HST includes two main stages called Quick Selection Stage (QS2) using fisher score as a filter method and Precise Selection Stage (PS2) using a Hybrid Bio-inspired Optimization (HBO) technique as a wrapper method. Finally, actual diagnose takes place through CP, which relies on Ensemble Classification Technique (ECT). ResultsThe proposed CAD2 has been tested experimentally against recent disease diagnostic strategies using two different datasets in which the first contains several diseases, while the second includes data for Covid-19 patients only. Experimental results have proven the high efficiency of the proposed CAD2 in terms of accuracy, error, precision, and recall compared with other competitors. Additionally, CAD2 strategy provides the best Wilcoxon signed rank test and Friedman test measurements against other strategies according to both datasets. ConclusionIt is concluded that CAD2 strategy based on ORP, FSP, and CP gave an accurate diagnosis compared to other strategies because it gave the highest accuracy and the lowest error and implementation time.

Design of Medical Diagnostic System Based on Artificial Intelligence

Design of Medical Diagnostic System Based on Artificial Intelligence

Acceleration of convolutional neural network based diabetic retinopathy diagnosis system on field programmable gate array

Diabetic retinopathy (DR) is one of the most common causes of blindness. The necessity for a robust and automated DR screening system for regular examination has long been recognized in order to identify DR at an early stage. In this paper, an embedded DR diagnosis system based on convolutional neural networks (CNNs) has been proposed to assess the proper stage of DR. We coupled the power of CNN with transfer learning to design our model based on state-of-the-art architecture. We preprocessed the input data, which is color fundus photography, to reduce undesirable noise in the image. After training many models on the dataset, we chose the adopted ResNet50 because it produced the best results, with a 92.90% accuracy. Extensive experiments and comparisons with other research work show that the proposed method is effective. Furthermore, the CNN model has been implemented on an embedded target to be a part of a medical instrument diagnostic system. We have accelerated our model inference on a field programmable gate array (FPGA) using Xilinx tools. Results have confirmed that a customized FPGA system on chip (SoC) with hardware accelerators is a promising target for our DR detection model with high performance and low power consumption.

An algorithm of blood typing using serological plate images

This paper describes an in vitro medical express diagnostic system designed to determine the blood group by analyzing the agglutination reaction (gluing of erythrocytes). The medical staff only needs to take a blood sample, put it on a serological plate, placing it in a special scanner for the blood group to be automatically determined. Data digitizing and machine-assisted plate identification allows two critical tasks to be addressed at once: storing the analysis results and controlling the human factor. The proposed recognition algorithm allows the alveolus boundaries to be accurately determined and the agglutination degree to be evaluated using a lightweight convolutional neural network. A unique dataset was collected with the independent assessment of agglutination degree conducted by medical experts. The agglutination estimation accuracy on the collected dataset of 3231 alveole was comparable to the accuracy of an average medical expert and equal to 0.98.

Applying AI in the Healthcare Sector: Difficulties

Artificial intelligence (AI) broadly speaking refers to any behavior shown by a computer or system that is similar to that of a person. Computers can learn from data without explicit human programming thanks to a kind of artificial intelligence known as "machine learning". The application of artificial intelligence (AI) technologies in medicine is one of the most important current trends in global healthcare. Artificial intelligence-based technologies are radically changing the global healthcare system by allowing for a drastic rebuilding of the medical diagnostics system and a corresponding decrease in healthcare costs. Prior to beginning treatment, an illness must be classified into which class of disorders it belongs. It is possible to classify the disease kind according to the feature space of the ailment. Machine learning algorithms can help with this problem.

A Comprehensive Review and Analysis of Deep Learning-Based Medical Image Adversarial Attack and Defense

Deep learning approaches have demonstrated great achievements in the field of computer-aided medical image analysis, improving the precision of diagnosis across a range of medical disorders. These developments have not, however, been immune to the appearance of adversarial attacks, creating the possibility of incorrect diagnosis with substantial clinical implications. Concurrently, the field has seen notable advancements in defending against such targeted adversary intrusions in deep medical diagnostic systems. In the context of medical image analysis, this article provides a comprehensive survey of current advancements in adversarial attacks and their accompanying defensive strategies. In addition, a comprehensive conceptual analysis is presented, including several adversarial attacks and defensive strategies designed for the interpretation of medical images. This survey, which draws on qualitative and quantitative findings, concludes with a thorough discussion of the problems with adversarial attack and defensive mechanisms that are unique to medical image analysis systems, opening up new directions for future research. We identified that the main problems with adversarial attack and defense in medical imaging include dataset and labeling, computational resources, robustness against target attacks, evaluation of transferability and adaptability, interpretability and explainability, real-time detection and response, and adversarial attacks in multi-modal fusion. The area of medical imaging adversarial attack and defensive mechanisms might move toward more secure, dependable, and therapeutically useful deep learning systems by filling in these research gaps and following these future objectives.

A MATHEMATICAL MODEL FOR AN AUTOMATED SYSTEM OF MEDICAL DIAGNOSTICS

One of the primary focuses of the Republic of Kazakhstan concerning sustainable and stable improvements in the well-being of its population is the advancement of the healthcare sector. A mathematical model for an automated medical diagnostics system integrates machine learning algorithms, statistical models, and decision trees to analyze patient data and facilitate accurate diagnoses. This model enables healthcare professionals to enhance the efficiency and reliability of medical diagnostics by leveraging advanced computational techniques.In diagnosing diseases, the differential diagnosis between primary biliary cirrhosis of the liver and active hepatitis with cholestatic syndrome poses a challenge due to their overlapping symptoms. Both conditions exhibit jaundice, pruritus, fatigue, and hepatomegaly. However, distinguishing features based on clinical and laboratory observations can aid in accurate differentiation between the two states. These distinguishing features can be incorporated by developing a mathematical model for diagnosing diseases, enabling precise identification and guiding appropriate treatment strategies.
 Predicting the progression of diseases is a crucial aspect of healthcare, enabling personalized interventions and improved patient outcomes. A mathematical approach can facilitate this prediction by monitoring changes in diagnostic results aligned with the severity of symptoms, which inherently vary over the observation period. By employing mathematical modeling techniques, healthcare professionals gain valuable insights into disease progression, supporting informed decision-making and tailored treatments.
 In conclusion, developing a mathematical model for an automated medical diagnostics system, incorporating machine learning algorithms, statistical models, and decision trees, significantly contributes to healthcare. These models enhance the accuracy, efficiency, and personalization of medical diagnoses. Additionally, mathematical models aid in the differential diagnosis of challenging conditions and provide predictions regarding disease progression, ultimately benefiting patient care and treatment outcomes.

Development of an Extended Medical Diagnostic System for Typhoid and Malaria Fever

In developing countries like Nigeria, malaria and typhoid fever are major health challenges in society today. The symptoms vary and can lead to other illnesses in the body which include prolonged fever, fatigue, nausea, headaches, and the risk of contracting infection occurring concurrently if not properly diagnosed and treated. There is a strong need for cost-effective technologies to manage disease processes and reduce morbidity and mortality in developing countries. Some of the challenging issues confronting healthcare are lack of proper processing of data and delay in the dissemination of health information, which often causes delays in the provision of results and poor quality of service delivery. This paper addressed the weaknesses of the existing system through the development of an Artificial Intelligence (AI) driven extended diagnostic system (EDS). The dataset was obtained from patients’ historical records from the Lagos University Teaching Hospital (LUTH) and contained two-hundred and fifty (250) records with five (5) attributes such as risk level, gender, symptom 1, symptom 2, and ailment type. The malaria and typhoid dataset was pre-processed and cleansed to remove unwanted data and information. The EDS was developed using the Naive Bayes technique and implemented using software development tools. The performance of the system was evaluated using the following known metrics: accuracies of true positive (TP), true negative (TN), false positive (FP), and false negative (FN). The performance of the EDS was substantially significant for both malaria and typhoid fevers.

HDSNE a new unsupervised multiple image database fusion learning algorithm with flexible and crispy production of one database: a proof case study of lung infection diagnose In chest X-ray images

Continuous release of image databases with fully or partially identical inner categories dramatically deteriorates the production of autonomous Computer-Aided Diagnostics (CAD) systems for true comprehensive medical diagnostics. The first challenge is the frequent massive bulk release of medical image databases, which often suffer from two common drawbacks: image duplication and corruption. The many subsequent releases of the same data with the same classes or categories come with no clear evidence of success in the concatenation of those identical classes among image databases. This issue stands as a stumbling block in the path of hypothesis-based experiments for the production of a single learning model that can successfully classify all of them correctly. Removing redundant data, enhancing performance, and optimizing energy resources are among the most challenging aspects. In this article, we propose a global data aggregation scale model that incorporates six image databases selected from specific global resources. The proposed valid learner is based on training all the unique patterns within any given data release, thereby creating a unique dataset hypothetically. The Hash MD5 algorithm (MD5) generates a unique hash value for each image, making it suitable for duplication removal. The T-Distributed Stochastic Neighbor Embedding (t-SNE), with a tunable perplexity parameter, can represent data dimensions. Both the Hash MD5 and t-SNE algorithms are applied recursively, producing a balanced and uniform database containing equal samples per category: normal, pneumonia, and Coronavirus Disease of 2019 (COVID-19). We evaluated the performance of all proposed data and the new automated version using the Inception V3 pre-trained model with various evaluation metrics. The performance outcome of the proposed scale model showed more respectable results than traditional data aggregation, achieving a high accuracy of 98.48%, along with high precision, recall, and F1-score. The results have been proved through a statistical t-test, yielding t-values and p-values. It’s important to emphasize that all t-values are undeniably significant, and the p-values provide irrefutable evidence against the null hypothesis. Furthermore, it’s noteworthy that the Final dataset outperformed all other datasets across all metric values when diagnosing various lung infections with the same factors.

Analysis of the application of expert systems

This article offers an overview of expert systems, identifies their advantages and disadvantages. The main advantages of expert systems are the ability to use expert knowledge to solve problems in automatic mode, which allows you to reduce the time and cost of solving a problem, as well as improve the quality of decisions made. However, there are also disadvantages, such as limitations in the use of only those knowledge that was embedded in the system at the stage of its development, which can lead to errors if the knowledge is outdated or insufficient. In addition, developing and supporting expert systems can be expensive and challenging. However, expert systems continue to develop and find application in various fields, which allows you to increase the efficiency and quality of solving problems, as well as increase the speed of solving them. This article reviewed modern expert systems as: “Intelligent medical differential diagnostic system based on expert systems,” “Dermatological expert system implementing ABCD,” “Hybrid recommendation system for recommendations of relevant films using an expert system” Their model construction, knowledge base formation and user interface are analyzed. The importance of their development, their effectiveness and the need for implementation as an information system are also discussed.

Toward a stable and low-resource PLM-based medical diagnostic system via prompt tuning and MoE structure

Machine learning (ML) has been extensively involved in assistant disease diagnosis and prediction systems to emancipate the serious dependence on medical resources and improve healthcare quality. Moreover, with the booming of pre-training language models (PLMs), the application prospect and promotion potential of machine learning methods in the relevant field have been further inspired. PLMs have recently achieved tremendous success in diverse text processing tasks, whereas limited by the significant semantic gap between the pre-training corpus and the structured electronic health records (EHRs), PLMs cannot converge to anticipated disease diagnosis and prediction results. Unfortunately, establishing connections between PLMs and EHRs typically requires the extraction of curated predictor variables from structured EHR resources, which is tedious and labor-intensive, and even discards vast implicit information.In this work, we propose an Input Prompting and Discriminative language model with the Mixture-of-experts framework (IPDM) by promoting the model’s capabilities to learn knowledge from heterogeneous information and facilitating the feature-aware ability of the model. Furthermore, leveraging the prompt-tuning mechanism, IPDM can inherit the impacts of the pre-training in downstream tasks exclusively through minor modifications. IPDM remarkably outperforms existing models, proved by experiments on one disease diagnosis task and two disease prediction tasks. Finally, experiments with few-feature and few-sample demonstrate that IPDM achieves significant stability and impressive performance in predicting chronic diseases with unclear early-onset characteristics or sudden diseases with insufficient data, which verifies the superiority of IPDM over existing mainstream methods, and reveals the IPDM can powerfully address the aforementioned challenges via establishing a stable and low-resource medical diagnostic system for various clinical scenarios.