Digital Frontiers in Healthcare: Integrating mHealth, AI, and Radiology for Future Medical Diagnostics

Abstract

In an era where technology and healthcare increasingly intertwine, we stand on the precipice of a seismic shift in the future of healthcare and medicine. This chapter unravels the confluence of mHealth, artificial intelligence (AI), and radiology as it navigates the labyrinth of these advancements, dissecting their unique qualities, historical evolution, and projected trajectories. From the democratising potential of mHealth to AI’s analytical might, and the transformative journey of medical imaging from film to digital—the chapter offers a deep dive into current realities and future horizons. Further, the intersection of these domains is explored, illuminating AI’s revolutionary role in enhancing mHealth capabilities through advances in medical imaging. An exhaustive review of cutting-edge applications and the ethico-regulatory conundrums they pose, forms a substantial part of the discourse, followed by a foresight into anticipated technological breakthroughs, their potential impacts, and the critical role of policymakers and health leaders in this odyssey. The chapter culminates in a holistic synthesis, tying together the strands of the preceding sections to underscore the transformative potential of this technological triumvirate. The text is designed as a captivating exploration, a reflective critique, and a roadmap for the future as we collectively navigate towards a technologically empowered healthcare era.

1. Introduction

In the rapidly evolving landscape of healthcare, the merging of mobile health (mHealth), artificial intelligence (AI), and medical imaging signifies a critical transformation, offering a multidimensional, patient-centred, data-driven model optimised for enhanced diagnostic and therapeutic outcomes.

mHealth, a subset of eHealth, capitalises on the pervasive nature of mobile technologies to augment the reach and efficacy of healthcare services [1]. The omnipresence of smartphones has spawned an era of ubiquitous health monitoring and services. No longer tethered by geographical constraints, we are now able to furnish real-time health data, diminishing healthcare disparities and enabling timely interventions [2].

AI, the broad umbrella under which machine learning (ML), deep learning (DL), and neural networks reside, has woven itself into the fabric of healthcare. It plays an integral role in predicting, diagnosing, and managing diseases [3, 4, 5]. The sheer power of AI to unravel intricate health datasets is ushering in a new era of personalised care that’s nothing short of extraordinary [5].

This transformative potential of AI is particularly evident in the realm of medical imaging, an indispensable pillar of modern medicine. While traditional imaging techniques hold undeniable value, they are contingent on the clinician’s expertise and are not impervious to human error. Enter AI algorithms, particularly those rooted in machine learning. These tools swiftly dissect complex imaging data with extraordinary precision, minimising human error [4]. Equipped with vast image databases, these algorithms discern nuanced patterns beyond the reach of human cognition and lay the groundwork for predictive analytics in medicine [4, 5, 6].

When we interweave mHealth, AI, and medical imaging, we unlock the potential to revolutionise healthcare delivery. AI uncovers the diagnostic treasures hidden in medical images, while mHealth ensures their accessibility, thereby accelerating diagnoses, optimising treatment plans, and improving patient outcomes [7].

mHealth, AI, and medical imaging, in isolation, boast formidable capabilities. However, their combined potential is the spark that may ignite a full-scale transformation in healthcare. As we navigate the industry’s shift towards a preventative, personalised, and precision-focused model, the integration of these three domains becomes all the more critical. This chapter delves deep into the advantages, challenges, and future implications of this promising coalescence for the future of healthcare.

2. Understanding mHealth

2.1 Definition and key features of mHealth

mHealth, a derivative of eHealth, represents the utilisation of mobile and wireless technologies to support the attainment of health objectives [8]. Defined by the Global Observatory for eHealth as “medical and public health practice supported by mobile devices, such as mobile phones, patient monitoring devices, personal digital assistants (PDAs), and other wireless devices”, mHealth has opened new avenues for the delivery of healthcare services [9].

Key features of mHealth include the facilitation of data collection in real-time, provision of access to healthcare services irrespective of geographic location, ensuring timely healthcare interventions, and improving patient adherence to medical advice and prescriptions [10, 11]. The ability to incorporate a variety of applications, ranging from telemedicine, electronic health records (EHRs), and healthcare analytics, to wearable technology and health-related IoT applications, underscores the flexibility of mHealth [2].

mHealth promotes an individualised and patient-centric approach to healthcare, affording a comprehensive health perspective of each patient through a combination of real-time and longitudinal data [12]. Furthermore, by enabling remote patient monitoring and diagnostics, mHealth minimises unnecessary hospital visits, thereby reducing the strain on healthcare systems [11]. With the integration of artificial intelligence and machine learning algorithms, mHealth has the potential to revolutionise healthcare by providing personalised, predictive, and preventive care. Despite its transformative capacity, mHealth implementation must negotiate challenges related to privacy, security, interoperability, and regulatory considerations.

2.2 Tracing the path: historical progression and evolution

The inception of mHealth in the early 2000s, coinciding with the mobile telecommunications boom, signified the harnessing of mobile and wireless technologies for health objectives [8]. Initially, mHealth primarily facilitated health information dissemination and communication among healthcare providers, reaching even remote areas [1].

The debut of the first iPhone in 2007 significantly broadened the mHealth landscape. These ‘smartphones,’ essentially handheld computers with advanced processing capacities and sophisticated software, evolved beyond mere communication tools to potent health platforms [13].

The subsequent introduction of app stores ushered in a wave of health-related applications, drastically expanding mHealth’s scope [13]. These apps ranged from fitness trackers promoting healthy habits to telemedicine interfaces enabling virtual consultations, thus empowering individuals to actively manage their health.

The late 2000s saw the emergence of wearable technologies and the Internet of Things (IoT) [14]. Wearables, equipped with biometric sensors, facilitated real-time health monitoring, promoting proactive and preventative care. Concurrently, the IoT interconnected these devices, creating a synchronised health data ecosystem [14].

The transition from 3G to 4G and 5G further propelled mHealth. Enhanced data transfer rates and reduced latencies of these mobile networks enabled real-time transmission of large, complex data, such as medical imaging, improving care quality and speed [2].

Over the past decade, mHealth has evolved from a technological novelty to a crucial component of global healthcare delivery systems. This swift integration of mHealth has been propelled by shifting societal habits, relentless technological advancements, and dynamic healthcare demands [15]. The COVID-19 pandemic underscored the necessity for remote healthcare delivery and patient monitoring, further elevating mHealth’s relevance and urgency.

2.3 mHealth today and beyond: current status and future implications

At the confluence of the technological era and healthcare, mHealth emerges as a dynamic catalyst. The current state is characterised by broad utilisation across diverse applications: chronic disease management, patient monitoring, health education, and telehealth consultations, among others [11]. mHealth technologies facilitate improved patient outcomes by fostering proactive health management and personalised healthcare.

Pervasive use of smartphones and wearable devices, combined with growing health consciousness, has driven mHealth integration into routine healthcare. Telehealth services, catalysed by the COVID-19 pandemic, remain an exemplary adaptation of mHealth, bridging patient-provider geographical disparities [16].

mHealth’s future appears laden with potential. An anticipated trajectory involves the marriage of mHealth and AI, facilitating predictive analytics and personalised medicine. The wealth of data procured by mHealth devices could be harnessed to predict health outcomes and prompt preventative interventions, amplifying the healthcare paradigm shift towards prevention over cure.

Furthermore, incorporating genomics and precision medicine within the mHealth spectrum could revolutionise patient management, transitioning from symptom-based to molecular-based treatment [10, 15].

Yet, the path forward also raises concerns: ethical, privacy, and cybersecurity issues must be scrupulously addressed. As mHealth continues its journey towards seamless integration into healthcare, continuous innovation, stringent regulation, and user acceptance remain vital elements influencing its future prospects. Thereby, mHealth stands on the precipice of defining the next healthcare epoch, embodying a blend of ubiquitous computing, personalised medicine, and predictive analytics.

3. Harnessing artificial intelligence and machine learning in healthcare

3.1 Decoding AI and ML

Artificial Intelligence, in its simplest form, refers to the capability of machines to emulate human cognitive functions such as understanding, learning, problem-solving, and decision-making [17]. This complex field of computer science was designed with the intention of creating intelligent machines that can behave, react, and work like human beings. They can execute a wide array of tasks with precision and efficiency, ranging from performing simple tasks to solving intricate problems.

There are two main types of AI—Narrow AI, which is designed to perform a narrow task such as voice recognition, and General AI, which is a type of machine intelligence that has the potential to perform any intellectual task that a human being can do [3]. The present era predominantly features Narrow AI, where AI technologies specialise in particular domains such as customer service through chatbots, voice recognition in virtual assistants, or image interpretation in medical imaging.

Machine learning, on the other hand, is an application of AI that grants systems the ability to automatically learn and improve from experience without explicit programming [17]. ML operates under the fundamental principle that machines should access data and learn autonomously.

ML is typically categorised into three types: supervised learning, where algorithms are trained on labelled data; unsupervised learning, where algorithms learn from unlabelled data to discover hidden patterns or intrinsic structures; and reinforcement learning, where an agent learns to make decisions by interacting with its environment, receiving rewards for positive actions and penalties for negative actions [3].

A further extension of ML is deep learning, which is an algorithm-based model that aims to mimic the human brain. DL algorithms utilise artificial neural networks with multiple layers—sometimes consisting of hundreds—that process the data, each layer providing a distinct interpretation [18]. This layered interpretation allows DL models to handle complex data and draw accurate conclusions or make predictions effectively. Figure 1 provides a brief overview of AI and its various subsets.

Figure 1.

An overview of artificial intelligence, machine learning, deep learning, and natural language processing.

3.2 AI and ML in contemporary medicine

AI and ML are revolutionising healthcare, impacting a spectrum of areas from diagnosis to administration. A fundamental application of these technologies lies in managing copious amounts of patient data, encompassing laboratory results, imaging studies, genetic profiles and comprehensive electronic health records, thereby facilitating the extraction of invaluable insights [6].

In the sphere of personalised medicine, ML underpins the derivation of significant patterns from patient data [19]. This empowers the provision of bespoke therapeutic approaches, tailored to individual genetic, environmental, and lifestyle factors, heralding a departure from traditional methodologies. ML models, when applied to large-scale biomedical datasets, can illuminate associations between genetic variations and disease susceptibility, thereby enriching our comprehension of disease pathogenesis and progression [20].

Integral to contemporary healthcare are AI-driven clinical decision support systems. These sophisticated systems decode complex patient data, providing clinicians with essential insights that inform diagnostic and treatment decisions. Furthermore, predictive analytics, underpinned by AI, represent a potent tool for early disease detection, identifying patterns within longitudinal patient data to predict disease risks, thus optimising patient outcomes [19].

In the domain of medical imaging, AI introduces transformative enhancements. Notably, computer vision, an AI sub-discipline, automates the interpretation of a myriad of imaging modalities, such as radiographs, computed tomography (CT) scans, and magnetic resonance imaging (MRI), often equalling or surpassing the accuracy of human interpretation [21]. This automation supports radiologists by mitigating their workload and reducing error incidence.

AI and ML play a pivotal role in drug discovery, streamlining the process and curtailing both costs and timeframes. Through utilising ML, researchers can forecast interactions between potential drug candidates and specific diseases, predicated upon extensive, multi-dimensional biological datasets, thus expediting the drug discovery trajectory [4, 5].

AI-powered robotics have ushered in unprecedented precision and consistency within surgical practice. Robotic surgical systems, guided by AI algorithms, can execute complex surgical tasks with minimal invasiveness, resulting in reduced postoperative discomfort and accelerated recovery [22].

3.3 Case studies: practical applications of ML in mHealth

3.3.1 Diabetic retinopathy

Diabetic retinopathy (DR), a severe diabetes complication, significantly contributes to vision impairment and blindness globally. The prevention of vision loss hinges on early detection and prompt intervention, both of which require regular retinal screening. However, such screenings are often unachievable or unavailable in many regions, particularly those with limited resources. This is where ML and AI technologies present a robust solution, automating the analysis of retinal images for precise and efficient DR detection [23]. One such transformative application of ML is in the development and deployment of a DL model for detecting DR from retinal images (Figure 2).

Figure 2.

Machine learning applications in diabetic retinopathy.

A striking example of this advancement is the DL model developed by Gulshan et al. [24]. By employing a dataset of 128,175 retinal images and convolutional neural networks (CNNs), this model demonstrated remarkable success in the automatic detection of DR. The model’s superior performance was reflected in an impressive area under the receiver operating characteristic curve (AUC-ROC) of 0.991 for DR detection, a clear improvement over previous methodologies and a testament to the potential of AI and ML in DR screening [24].

The application of AI and ML in mHealth, as epitomised by this case study, offers several advantages over traditional DR screening methods. The primary benefit is a significant reduction in healthcare professionals’ workload, enabling concentration on complex cases and direct patient care [25]. AI-driven systems also deliver swift, accurate retinal image analysis, facilitating timely diagnosis and intervention to prevent vision loss [23]. Furthermore, the automation of DR screening using AI and ML can enhance healthcare access in underserviced areas, where the availability of specialist ophthalmologists is often scarce [26].

Nonetheless, it’s important to recognise the challenges accompanying the implementation of AI and ML for DR detection in mHealth. The success of AI models hinges on their generalisability and robustness across diverse populations and varied retinal image quality [27]. Additionally, addressing ethical considerations surrounding data privacy, informed consent, and algorithmic bias is essential for the responsible use of AI in mHealth applications [28].

3.3.2 Chronic obstructive pulmonary disease (COPD)

This case study examines a practical use of ML models in mHealth for COPD management and monitoring. It provides crucial insights and challenges encountered during this process, offering valuable guidance for future AI and ML healthcare integration efforts [29].

COPD is a persistent respiratory disorder characterised by continuous symptoms and airflow limitation due to airway and/or alveolar anomalies [30]. mHealth applications could revolutionise COPD care by enabling remote patient monitoring and delivering real-time, personalised health feedback [31].

The development of the COPD mHealth application encountered various obstacles (Figure 3).

1. Data collection and quality: the collection of high-quality data is essential for the effective deployment of AI and ML in mHealth [29]. Data collection challenges included incomplete data and quality discrepancies due to variations in data collection protocols across healthcare providers. Data preprocessing techniques, such as imputation and normalisation, were used to overcome these issues.

2. Algorithm selection and model performance: choosing the right AI and ML algorithms is critical for precise prediction and improved patient outcomes [32]. Various algorithms, including support vector machines (SVM), random forest (RF), and DL models, were assessed, with convolutional neural networks (CNNs) proving most effective for predicting patient outcomes and providing personalised feedback [29].

3. Security and privacy: preserving patient confidentiality and protecting sensitive health data was a priority [33]. Robust security measures, such as secure data transmission protocols and access control systems, were implemented to address these concerns.

4. Integration with clinical workflows: the incorporation of the mHealth application into existing healthcare systems and clinical workflows required strategic partnerships with healthcare providers to tailor the application for smooth integration with electronic health records (EHRs) and other clinical tools [34].

5. User adoption and engagement: encouraging user adoption and engagement with the mHealth application was critical [35]. A user-centric design approach and the inclusion of feedback from COPD patients and healthcare providers during the design and development stages resulted in an intuitive application with high user adoption rates.

Figure 3.

Machine learning applications in COPD.

3.4 A glimpse into the future: potential applications

AI and ML are progressively transforming the healthcare ecosystem with a myriad of applications that span diagnostics, therapeutics, patient care management, and health system optimisation, presenting a vision of future healthcare marked by individualised and efficient care [5].

• Predictive analytics and disease modelling: AI/ML algorithms, meticulously trained on expansive datasets, possess the capability to discern patterns that elude human cognitive processes. For instance, sophisticated DL models can prognosticate the progression of chronic ailments such as diabetes or chronic kidney disease, thereby allowing for proactive interventions to mitigate disease progression [10].

  Unintrusive wearable IoT devices persistently collate and transmit integral patient health data, including heart rate variability, blood glucose metrics, and sleep patterns, into a designated ML model. This dynamic model, continually learning and evolving, has the potential to recognise early indications of disease progression, substantially reducing the burden of chronic diseases by facilitating proactive health management [36].

• Radiomics and imaging diagnostics: AI/ML has made significant strides in radiomics, an area dedicated to extracting large volumes of features from radiographic images via data characterisation algorithms. ML models, specifically CNNs, have demonstrated superior diagnostic accuracy in identifying conditions such as lung cancer from CT scans and breast cancer from mammograms [37]. As these models evolve, their predictive accuracy is likely to improve, leading to a decrease in both false positives and negatives, curtailing the need for invasive diagnostic procedures, and consequently enhancing patient outcomes. Additionally, the incorporation of federated learning—an approach allowing ML across multiple datasets at their original location—could augment model generalisability without jeopardising patient data privacy [38].

• Personalised medicine and therapeutics: AI/ML algorithms, through the comprehensive analysis of both genotypic and phenotypic patient data, have the capability to discern the most effective therapeutic intervention for an individual, thereby reducing the dependency on the often-inherent trial-and-error methodology in treatment strategies [39].

  Through the use of a single blood examination and genomic sequencing, an AI algorithm can propose the most suitable pharmacological regimen for a patient with complex comorbidities. The AI model integrates factors such as drug interactions, individual metabolic responses, and potential adverse effects into its decision-making process. This degree of precision in therapeutics has the potential to significantly improve patient outcomes and elevate the standard of patient care [39].

• Health systems improvement: AI/ML holds the capability to enhance health systems, streamlining administrative tasks, optimising patient flow, and improving resource allocation. Future application of natural language processing (NLP) algorithms could sift through electronic health records, automate billing, facilitate appointment scheduling, and augment workflow efficiency, thus allowing healthcare professionals to devote more time to patient care [40].

4. Medical imaging: a journey from radiography to superintelligence

4.1 Overview of traditional medical imaging techniques

The genesis of modern medical imaging can be traced back to the discovery of X-rays by Wilhelm Conrad Roentgen in 1895 (Figure 4). The initial technique, radiography, leveraged the differential absorption of X-ray radiation among tissues, with denser structures, such as bones, appearing white due to increased absorption [41]. Further sophistication was introduced through the utilisation of contrast agents, enhancing the perceptibility of specific structures in procedures such as angiography or barium swallow examinations [41].

Figure 4.

A historical timeline of the key discoveries in medical imaging.

Fluoroscopy, conceived in the early twentieth century as a dynamic extension of radiography, facilitated real-time visualisation of internal body structures. This capability has proved invaluable for executing minimally invasive procedures like catheter placement and joint injections, solidifying its irreplaceable role in the clinical scenario [42]. Thereafter, the 1940s heralded the introduction of ultrasound imaging, a modality leveraging high-frequency sound waves to generate real-time images of internal organs (Edler). Embodying safety, portability, and cost-effectiveness, ultrasound has emerged as a cornerstone in diverse fields such as obstetrics, cardiology, and emergency medicine (Bushberg).

The 1970s witnessed another breakthrough with the introduction of CT scanning by Sir Godfrey Hounsfield. CT employs multiple X-ray measurements taken from distinct angles, yielding cross-sectional images or ‘slices’ of the body [43]. The three-dimensional vistas offered by CT scans supersede conventional X-rays in visualising intricate structures such as the brain, lungs, and abdominal organs.

Medical imaging took another significant leap forward in the early 1980s with the clinical adoption of MRI. Devoid of ionising radiation, MRI creates high-resolution images by harnessing the principles of nuclear magnetic resonance [44, 45]. Particularly efficacious for imaging soft tissues and organs, MRI is an unrivalled tool for studying structures such as the brain, spinal cord, and musculoskeletal system [44, 45].

Nuclear medicine techniques of the contemporary era, such as Positron Emission Tomography (PET) and Single-Photon Emission Computed Tomography (SPECT), afford functional imaging by tracking radiation from an introduced radioactive tracer [46, 47]. These techniques impart critical insights into the physiological function and metabolic processes of organs, facilitating diagnostic and therapeutic applications in fields as diverse as cardiology, neurology, and oncology [46, 47].

Nevertheless, the remarkable benefits accrued from traditional medical imaging are somewhat offset by inherent limitations. These include exposure to ionising radiation (pertinent to X-rays and CT scans), variable image resolution, dependence on operator skill (in ultrasound imaging), significant economic implications, and protracted scanning times (particularly in MRI). The transition to digital imaging, bolstered by the integration of AI, promises to counter many of these challenges, a perspective explored in subsequent sections.

4.2 The digital revolution in imaging

The digital era has heralded pivotal developments in medical science, presenting an array of advantages whilst concurrently introducing novel complexities. Digital medical imaging, essentially a transformation of images into numerical data for computational interpretation, has facilitated a period of enhanced diagnostic precision, efficient data archiving, and accelerated transmission of medical imagery, encapsulating X-rays, CT scans, MRI, and ultrasounds [48].

The transition to digital has primarily eclipsed analogue imaging methodologies, underscoring its potential to revolutionise diagnostic efficacy and accuracy. The intricacy afforded by digital images, attributable to superior resolution and contrast, has been instrumental in augmenting the sensitivity and specificity of disease detection [49]. This aspect is indispensable across numerous clinical scenarios, including the discernment of diminutive tumours or the delineation of nuanced stroke lesions. Additionally, digital images facilitate manipulations for optimal visualisation without compromising the image integrity, offering adjustable contrast, brightness, and magnification parameters tailored to individual patient contexts and specific clinical requisites.

Another salient advantage of digital imaging is its capacity for streamlined storage and retrieval. The physical archiving of analogue images poses significant challenges, occupying substantial space and risking degradation over time. Conversely, digital images can be effectively preserved within electronic databases such as Picture Archiving and Communication Systems (PACS), thus ensuring sustained image integrity [50]. The integration of digital imaging with EHRs further optimises retrieval and bolsters the continuity of patient care.

The capability for instantaneous and remote transmission of digital images represents a key benefit, particularly relevant in the burgeoning field of telemedicine. Physicians can access diagnostic images virtually anywhere, encouraging global collaboration irrespective of geographical barriers [51]. Nevertheless, the digitisation of medical imaging is not devoid of challenges. Foremost, the substantial costs associated with the acquisition and maintenance of digital imaging equipment and software can impose a significant financial encumbrance on numerous healthcare providers.

While digital imaging expedites storage, the vast increase in data consequent of this transition presents a formidable challenge. This voluminous data surge necessitates robust data management strategies to ensure data integrity, security, and accessibility. Confidentiality and security concerns also underscore the hurdles inherent to digital imaging. As medical images become an integral component of an interconnected digital health ecosystem, potential breaches of confidentiality and unauthorised access emerge, thereby necessitating rigorous cybersecurity protocols. Lastly, the assimilation of digital imaging into established healthcare workflows may necessitate substantial change management efforts, encompassing training healthcare professionals to operate innovative technologies, interpret digital images, and adapt to evolving workflow dynamics [52].

4.3 The impact of AI on the transformation of radiology

Within the ambit of medical imaging, AI has substantially expedited the evolution from traditional to digital techniques, reinventing diagnostics and prognostics by providing more accurate, rapid, and efficient tools that redefine clinical decision-making.

The transformative capacity of AI in medical imaging is primarily grounded in its sophisticated pattern recognition ability [53]. As such, a gamut of AI algorithms, particularly those harnessing DL approaches, are making substantial strides in medical imaging interpretation. For instance, CNNs, a type of DL model inspired by the visual cortex, has been especially impactful in the realm of radiology, demonstrating exceptional competence in identifying complex patterns in voluminous image data sets.

CNNs can process both 2D and 3D images, discerning nuanced details that may elude the human eye, thereby extending the capacity of radiologists. By learning from vast data sets and continuously improving its accuracy with each data interaction, ML can potentially uncover novel imaging biomarkers, facilitating earlier detection, better prognostication, and personalised treatment.

AI is also instrumental in expediting the image acquisition and reconstruction processes in imaging modalities like CT and MRI [54]. For instance, AI has proven adept at producing high-quality images from undersampled data, mitigating the traditionally extended MRI scan times and enhancing patient comfort. Additionally, it aids in reducing exposure to ionising radiation in CT imaging by enabling superior image quality from lower dose acquisitions.

The significant progress in AI-powered medical imaging extends beyond diagnostics. The predictive potential of AI, paired with imaging, holds promise in enhancing patient management by forecasting disease progression and therapeutic response [55]. This predictive capacity is particularly relevant in chronic conditions, such as cardiovascular diseases and cancer, where timely and accurate assessment can significantly influence therapeutic outcomes. Moreover, AI’s integration in interventional imaging promotes the development of image-guided therapies, thus further optimising patient care.

Furthermore, AI can be a potent tool in managing the substantial data generated in medical imaging [56]. With digital imaging systems producing a tremendous amount of data daily, traditional methods of data handling prove increasingly inadequate. AI, with its capacity to analyse and extract meaningful insights from large-scale imaging data, offers an effective solution to this challenge. This data-driven approach can potentially foster a move towards precision medicine, tailoring therapies based on individual patient profiles.

However, while AI’s potential in medical imaging is substantial, it should be viewed as a tool to augment, not replace, clinicians. Ideally, AI serves as a secondary reader, assisting clinicians in improving efficiency and reducing diagnostic errors, whilst maintaining human oversight to validate the final clinical decision.

5. Intersection of mHealth and AI in medical imaging

5.1 Real-world applications: mHealth meets AI

The capability of mHealth platforms to collect and share healthcare data provides an unprecedented avenue for AI to influence patient care. These applications broadly fall under two categories: diagnosis assistance and image analysis. Both these functionalities, powered by AI and delivered through mHealth, can significantly contribute to streamlined and efficient medical imaging processes [57].

Firstly, AI has been applied in numerous diagnostic imaging applications, where its machine learning algorithms can enhance the accuracy of interpreting images by leveraging CNNs adept at processing image data. A notable example is Google’s DeepMind, which demonstrated its potential in diagnosing age-related macular degeneration and diabetic retinopathy from retinal scans, performing at par or better than human experts [58]. Its deployment via mHealth platforms could provide timely and accurate diagnosis, thereby improving patient outcomes, particularly in remote or resource-poor settings.

Another breakthrough has been the use of AI algorithms in analysing mammograms to detect breast cancer. For instance, ScreenPoint Medical’s Transpara™ employs AI to interpret mammograms and tomosynthesis exams, helping radiologists identify potential anomalies with improved accuracy [59]. This mHealth-accessible tool allows healthcare professionals to utilise advanced AI-assisted diagnosis, even on their handheld devices.

Furthermore, the intersection of mHealth and AI is revolutionising ultra-sound imaging. Butterfly Network’s Butterfly iQ , for instance, uses a handheld, pocket-sized ultrasound device that connects to a smartphone and uses AI to capture and interpret images [60]. This democratises the use of ultrasound technology, particularly in rural or low-resource settings where conventional ultrasound machines are not readily available.

Shifting our focus to AI-assisted image analysis, AI algorithms are now capable of providing real-time insights, making them vital tools during surgical procedures. A case in point is the AI software developed by Aidoc, designed to flag acute conditions like strokes, pulmonary embolisms, and cervical spine fractures in CT scans [61]. By integrating such AI software into mHealth platforms, the potential for immediate, round-the-clock analysis could transform acute care.

It is noteworthy to highlight the advent of smart wearable devices and biosensors that collate and interpret an abundance of physiological data. Companies such as Apple and Fitbit are spearheading this trend with AI-powered interpretation of ECG data to detect irregular cardiac rhythms, an established risk factor for cerebrovascular conditions [62]. Similarly, Google’s Project Nightingale strives to anticipate patient deterioration and optimise corresponding treatment plans through real-time data analytics [63].

5.2 The synergy of mHealth and AI in radiology: key benefits

The synthesis of mobile health with AI within the context of medical imaging proffers a cornucopia of benefits, engendering a fundamental shift in the paradigm of healthcare delivery. In order to comprehend these advantages, it is critical to understand the unique attributes each technology contributes to this potent amalgamation.

The integration of mHealth and AI within the realm of medical imaging offers an array of benefits, facilitating a substantive shift in the landscape of healthcare delivery. To fully appreciate these merits, it is crucial to understand the unique characteristics that each technology contributes to this powerful fusion.

• Accessibility and expediency: a primary and compelling merit of mHealth is its capacity to democratise healthcare through enhanced accessibility. This is critically vital in remote and underserved regions where the scarcity of medical resources and expertise exacerbates health inequities [4, 64]. By integrating mHealth with AI in medical imaging, we enable remote diagnostics and patient monitoring, thereby negating the requirement for physical proximity to advanced medical facilities. Additionally, AI-driven algorithm within mHealth applications can significantly hasten the diagnostic process and turnaround time of imaging interpretation. In emergency scenarios, this swift response may constitute the difference between life and death, underscoring the immense potential of this symbiosis.

• Precision and consistency: AI algorithms demonstrate remarkable proficiency in extracting salient features from medical images, outperforming humans in some instances [4]. When harnessed in mHealth applications, these AI tools offer the dual advantage of precision and consistency. They can effectively eliminate human errors arising from fatigue or oversight, which are not uncommon in traditional, human-dependent diagnostic processes. AI’s prowess in detecting minute aberrations in medical images that may be overlooked by the human eye could potentially enhance early disease detection, improving prognostic outcomes.

• Patient empowerment and engagement: mHealth solutions are inherently patient-centred, promoting active patient involvement in managing their health [64]. The incorporation of AI in mHealth for medical imaging further empowers patients by providing them with actionable insights. Patients can monitor health trends, understand potential health risks, and participate more effectively in shared decision-making processes. This paradigm shift from a reactive to a proactive healthcare model cultivates a more engaged, informed patient populace.

• Efficient resource utilisation: The healthcare sector, particularly in developing regions, is often beleaguered by resource constraints. The combination of mHealth and AI in medical imaging can engender more efficient utilisation of resources [4]. AI-powered tools, with their automated image analysis capabilities, can alleviate the workload of overburdened healthcare professionals, enabling them to focus on more critical tasks. In the context of growing patient numbers and a limited healthcare workforce, such efficiency gains are indispensable.

• Data-driven decision making: AI’s capability to process and learn from vast datasets can imbue mHealth with powerful predictive and prescriptive analytics [64]. These data-driven insights can inform both population health strategies and individual patient care plans. By identifying patterns and trends in medical imaging data, AI can help predict disease progression or response to treatment, enabling healthcare providers to tailor therapeutic interventions for optimal outcomes.

5.3 Pioneering advances of AI-driven radiology

AI/ML have revolutionised medical imaging, facilitating a swift shift from traditional reactive medical practices to more proactive and predictive healthcare. Its integration into medical imaging has yielded a panoply of innovative applications that have enhanced both the accuracy and speed of diagnostics.

One of the most groundbreaking applications of AI in medical imaging pertains to computer-aided detection and diagnosis (CAD) [65]. AI algorithms, through DL techniques, are capable of detecting subtle changes in imaging scans that may elude the human eye. For instance, CNNs have been utilised to identify anomalous patterns in images, successfully aiding in the diagnosis of conditions ranging from cancerous tumours to cardiovascular diseases. Research shows that CNNs have demonstrated exemplary performance in detecting lung nodules in CT scans, enhancing early-stage lung cancer detection and ultimately improving patient prognoses [65, 66] (Figure 5). Additionally, CNNs demonstrate profound capabilities in detecting intracranial infarcts or haemorrhages and in the identification of large vessel occlusion. Their incorporation into these sectors has markedly revolutionised the paradigm of stroke treatment, as showcased in Figure 6.

Figure 5.

Machine learning applications in oncological imaging.

Figure 6.

Machine learning applications in neuroradiology.

Another notable application is the use of AI in image segmentation. Traditionally, the demarcation of regions of interest (ROIs) in medical images was performed manually, a laborious process often fraught with variability and inconsistencies. The emergence of AI has streamlined this process, ensuring higher precision and reproducibility [67]. For instance, in radiation oncology, the accurate segmentation of tumour regions and adjacent healthy tissues in CT or MRI scans is crucial for treatment planning [68]. AI algorithms can automate this process, minimising human error and enabling more accurate and targeted treatments.

AI has also proven instrumental in predictive modelling in medical imaging [56]. By leveraging machine learning algorithms, it’s possible to predict patient outcomes based on imaging data. For example, the progression of neurodegenerative diseases such as Alzheimer’s can be predicted by analysing structural changes in brain MRI scans over time using AI algorithms. Similarly, the risk of cardiovascular events can be predicted by analysing coronary artery calcium scoring in CT scans.

AI is also being deployed in automated quality control in medical imaging. In an environment where the acquisition of high-quality images is paramount for accurate diagnosis and treatment, AI can play a pivotal role in identifying sub-optimal images and suggesting adjustments to improve quality [21]. This application of AI not only enhances the accuracy of diagnoses but also improves the efficiency of the imaging workflow.

In the realm of real-time imaging, AI is increasingly being used to facilitate immediate analysis and interpretation of images during procedures such as ultrasounds, potentially expediting diagnoses and enabling more immediate therapeutic interventions [40]. The application of AI in real-time imaging could reduce the need for specialist involvement during image acquisition, thereby streamlining workflows and improving access to healthcare services.

Furthermore, AI-aided radiogenomics represents another cutting-edge application [65]. This field aims to correlate imaging features with genomic data, thus providing a non-invasive means of determining the genetic profile of tumours. This method can potentially guide more personalised therapeutic strategies, enhancing treatment efficacy and patient outcomes.

6. Forecasting the future: mHealth, AI, and medical imaging

6.1 The technological horizon: projected advancements

The impending healthcare revolution is forecasted to be driven by major advancements in mHealth, AI/ML, and medical imaging technologies, all enabled by the rapid progress in AI algorithms, computation, sensor technology, and network connectivity (Figure 7).

Figure 7.

Current and future applications of artificial intelligence and machine learning in healthcare.

ML models and AI algorithms are set to undergo significant refinement, with increasingly prevalent application of DL networks for complex tasks like medical image-based disease diagnosis. These models are expected to evolve, taking on multifaceted tasks such as comprehensive treatment plan formulation and patient outcome prediction with heightened precision. AI-empowered predictive analytics will facilitate accurate forecasting of disease trajectories, potentially even before symptom manifestation [69].

Medical imaging technologies are also on a trajectory of considerable improvement. Development is underway for imaging modalities capable of elucidating biological processes at cellular or molecular level. Hybrid imaging technologies like PET/MRI and PET/CT are expected to gain wider acceptance due to their combined strengths [67]. Simultaneously, advancements in imaging resolution will provide hitherto unattainable anatomical and physiological details. The symbiosis of AI and medical imaging is projected to strengthen, with machine learning models proficiently interpreting imaging data, identifying patterns, and making accurate predictions (Rajpurkar; Salameh). Through exposure to real-world data, these algorithms will progressively improve diagnostic precision, reducing misdiagnosis and enhancing patient outcomes [70].

mHealth technologies are set for remarkable advancement, with the next generation of wearable and implantable devices likely to offer broader real-time health monitoring [71]. The immense data generated by these devices will be harnessed by sophisticated AI systems for personalised health guidance [72]. Additionally, the roll-out of 5G and anticipated 6G networks will significantly enhance the capacity of mHealth applications for data management and real-time stakeholder synchronisation [73]. This will support superior remote patient monitoring and telemedicine services, improving healthcare accessibility and efficiency [74].

The integration of AI and blockchain technologies offers promising solutions for data security and patient privacy in mHealth, with blockchain’s decentralised and tamper-proof structure providing secure platforms for storing and sharing sensitive health data. Moreover, virtual and augmented reality are poised to revolutionise medical training, surgical planning, rehabilitation, and patient education, further facilitating remote patient care and contributing to healthcare democratisation [75].

While these advancements promise transformative impacts, they also raise challenges related to data privacy, ethics, and regulatory frameworks. The subsequent sections will delve into the implications of these advancements on healthcare delivery and patient outcomes, along with the role of policymakers and healthcare leaders in shaping this future.

6.2 Impact on healthcare delivery and patient outcomes

The impending era of digital healthcare is anticipated to profoundly reshape patient care, healthcare delivery, and clinical outcomes. It is essential to comprehensively explore the potential implications of this transformation.

One key impact will be increased healthcare accessibility. The integration of AI and mHealth liberates medical imaging from the restrictive, stationary hospital equipment, enabling portable imaging devices enhanced by AI algorithms for immediate interpretations [76]. This shift could decentralise diagnostic services, particularly benefiting remote or underserved regions and potentially reducing healthcare disparities [77].

Simultaneously, the precision and speed of diagnostics are set to improve substantially. AI algorithms excel in extracting meaningful information from complex imaging data, often outperforming human capacity for pattern recognition [78]. This advanced capacity could promote early disease identification and personalised care, refining treatment plans and enhancing patient outcomes. Furthermore, AI’s machine learning models promise standardisation in image interpretation, mitigating inter-observer variability and human error [79].

In addition, patient monitoring is poised for significant advancement. AI-embedded wearable technology provides continuous, real-time health monitoring, transforming chronic disease management and early detection of complications [80]. This ongoing surveillance could streamline interventions, optimise outcomes, and reduce hospital admissions.

Telemedicine is also expected to evolve, driven by AI, mHealth, and medical imaging. This transition from hospital-focused to home-based care may enhance patient safety and satisfaction, supporting a more patient-centred model [81]. AI’s predictive capacity could guide remote patient management and preempt potential health issues, reducing emergency hospitalisations.

Advanced language models like ChatGPT are emerging as vital tools in this changing healthcare landscape [82]. Acting as an interface between complex AI algorithms and users, these models can synthesise large volumes of medical data into actionable insights, and provide immediate responses to patient queries, education, and psychological support [83]. They also aid research by scanning vast literature to identify novel findings and research trends [83].

At a broader scale, AI-enhanced medical imaging can offer population-level insights into disease trends, informing public health strategies and resource allocation [84]. The aggregation of health data from mHealth apps creates a rich repository for AI to uncover new medical knowledge, accelerating research and clinical innovation [85].

The synergy between AI, mHealth, and medical imaging stands to deliver substantial economical advantages through increased diagnostic precision, timely interventions, and enhanced preventive measures [83]. However, this necessitates the development of regulatory and ethical frameworks to safeguard patient confidentiality and autonomy, ensuring these technologies are oriented towards maximising patient benefit [86]. The fundamental role of healthcare leaders in directing this course underlines the importance of preserving human empathy within the healthcare system.

6.3 Policymakers and health leaders: their role in shaping the future

Policymakers and health leaders hold the reins in guiding this anticipated seismic shift in healthcare, their strategies and directives having the potential to significantly determine the future trajectory and impact of these emerging technologies.

The junction of AI and mHealth brings an array of opportunities for medical imaging, encompassing enhanced diagnostic accuracy, optimisation of workflow processes, and the personalisation of healthcare delivery. Conversely, this technological advancement introduces intricacies related to privacy, equity, regulation, and professional roles. Policymakers and health leaders are tasked with steering this complex landscape, delicately balancing between fostering innovation and ensuring regulation, privacy, and safety [87, 88, 89].

Understanding that innovation does not transpire in isolation is crucial in appreciating the role of these leaders. The intricacies of AI-driven mHealth applications must be contextualised and regulated within societal, legal, and ethical frameworks. The ability to comprehend the potential and perils of these technologies is paramount in formulating balanced and forward-thinking regulations that protect interests without inhibiting innovation [88, 89].

One potential regulatory hurdle pertains to data privacy. AI and mHealth technologies function on the collection, storage, and processing of extensive patient data, creating potential privacy and security risks. Policymakers are tasked with modernising privacy laws and regulations, demanding rigorous data security measures while empowering patients with the autonomy over their health data [90].

The question of algorithmic transparency and accountability also arises. Given the often ‘black-box’ nature of AI algorithms, unintended, discriminatory, or harmful outcomes may occur. Ensuring transparency and explicability of AI systems in healthcare is critical, allowing for accountability during adverse events [91].

From an equity standpoint, these technologies pose a risk of widening the healthcare divide. High-tech solutions frequently necessitate considerable resources, potentially exacerbating disparities between affluent and impoverished, urban and rural. Health leaders are obligated to strive for equitable distribution of these technologies, ensuring their benefits permeate all demographics [92].

In terms of professional roles, AI-driven mHealth applications may catalyse significant shifts in the healthcare workforce. Policymakers and health leaders are required to predict and navigate these changes, aiding healthcare professionals in developing the requisite skills to collaborate with AI, and ensuring that these technologies augment, rather than substitute, the human element in healthcare [93].

Health leaders are further mandated to cultivate a culture of change and innovation within healthcare institutions. Their role includes advocating for AI-driven mHealth applications, propelling adoption, and managing resistance to change among healthcare providers.

7. Challenges and strategies in AI integration

7.1 Technical obstacles in implementation

The marriage of AI/ML and mHealth inaugurates an unprecedented aeon in healthcare, offering personalised, accessible, and streamlined care [94]. Despite this, the deployment of these advanced technologies presents a multitude of complex issues. The obstacles encountered in this integration demand careful contemplation and tactical planning to ensure success (Figure 8).

Figure 8.

Opportunities and challenges of AI integration into clinical practice.

The primary hurdle is data compatibility and interoperability. AI algorithms require extensive and high-quality annotated data for efficient training and accurate results [95]. mHealth applications, conversely, yield diverse data types such as electronic health records, real-time sensor data, images, and text, often in varying formats and standards. This heterogeneity can result in compatibility issues, hampering effective data utilisation by AI algorithms. Consequently, harmonising these disparate data forms for AI applications represents a significant technical challenge.

Simultaneously, scalability and integration present their unique challenges. The exponential number of mHealth users concomitantly generates a growing volume and variety of data. AI systems must possess scalability to accommodate this growth, necessitating substantial computational resources and robust algorithms that can manage increasing complexity [77]. Integrating AI systems into established healthcare information systems, known for their complexity and structured workflows, demands careful strategising to avert disruption and ensure seamless and efficient AI tool integration.

The inherent complexity of AI algorithms further complicates the process. Particularly, DL models comprise intricate networks with potentially millions of parameters. Training these models requires a data abundance, significant computational power, and technical expertise [96]. Coupled with their ‘black box’ characteristic, making the prediction process opaque, these complexities can impede the broader adoption of AI in mHealth, as healthcare professionals might hesitate to utilise tools that are not transparent.

The reliability and validity of AI-powered mHealth applications present another major hurdle. Given their crucial role in assisting health decisions, consistent accuracy and reliability of outputs are paramount [85]. This necessitates rigorous testing and validation methods, ensuring performance across diverse scenarios and patient demographics. Additionally, these algorithms must be resilient to changes and anomalies in the input data, safeguarding against erroneous outputs when confronted with unfamiliar data.

To surmount these technical obstacles, it is proposed that potential solutions include the development of standardized data protocols and comprehensive testing methodologies [97]. Enhanced collaboration among AI developers, mHealth creators, healthcare professionals, and policymakers could expedite the creation of efficient integration strategies and scalability solutions. Ongoing training and education in AI for healthcare professionals could help demystify the ‘black box’, fostering trust in AI-driven mHealth applications [98]. These challenges can only be overcome through collective effort, enabling the full potential of AI and mHealth integration in healthcare.

7.2 Regulatory challenges and how to navigate them

The advent and adoption of artificial intelligence and mHealth in medical imaging introduces multifaceted regulatory considerations, shaped by the converging factors of data privacy, algorithmic accountability, clinical responsibility, and policy enforcement [99]. Successfully manoeuvring this regulatory landscape calls for a considered and rigorous strategy.

Data protection stands out as a key challenge, as AI and mHealth involve handling significant volumes of personal health information. Safeguarding this data in accordance with health information privacy regulations, including Australia’s Privacy Act 1988, is an imposing undertaking [100]. Effective navigation of this terrain may be achieved through implementing robust data governance protocols, stringent encryption techniques, anonymisation processes, and rigorous controls on data access and sharing. This will not only ensure regulatory compliance but will also fortify user trust and confidence, thereby enhancing adoption and endorsement [99].

The interpretation of regulations concerning software as a medical device introduces another challenge. Regulatory authorities, traditionally accustomed to assessing the safety and efficacy of static, unchanging medical devices, now grapple with the dynamic and continually evolving nature of AI [101]. The Therapeutic Goods Administration (TGA) in Australia has taken strides towards clarifying regulatory expectations for AI-imbued medical devices, indicating that ongoing engagement with the TGA, awareness of emerging regulations, and pre-submission consultations may serve as valuable strategies to traverse this regulatory challenge [102].

The lack of global regulatory harmonisation for AI applications, due to the expansive potential of mHealth solutions and AI in medical imaging, presents additional difficulties. Participation in international initiatives like the International Medical Device Regulators Forum (IMDRF) can aid in the development of globally agreed standards, easing the regulatory burden associated with AI integration in mHealth [103]. Lastly, clarification of clinical responsibility and liability linked to AI-mediated medical decisions is crucial. The growing influence of AI tools in healthcare introduces ambiguity concerning accountability [95]. Addressing this necessitates fostering a comprehensive understanding of the roles of AI systems and demarcating clinical responsibilities in healthcare delivery.

Ultimately, effectively navigating the regulatory complexities associated with the amalgamation of AI and mHealth demands proactive engagement with regulatory bodies, the establishment of stringent data governance and cybersecurity infrastructures, prioritisation of transparency and accountability in AI algorithms, active involvement in international standardisation efforts, and clarification of clinical responsibilities. The ultimate goal extends beyond mere compliance with regulations, striving instead to champion responsible innovation that genuinely enhances patient care.

7.3 Transversing the ethical landscape: confidentiality and patient privacy

The integration of AI/ML with mHealth platforms holds promise for improved healthcare delivery and patient outcomes. However, these advancements also trigger significant ethical dilemmas, primarily concerning patient privacy and data protection [104]. The mHealth applications, combined with AI’s advanced data processing capabilities, continually accumulate and analyse personal health data, necessitating an immediate focus on these issues.

A major ethical issue arises from the vast quantity of health-related data generated and disseminated by mHealth applications. This collected information is not only extensive but highly personal, covering a wide range of biometric, physiological, and lifestyle data. Without robust data security measures, the pervasiveness of mHealth applications might expose sensitive data to misuse [105]. AI algorithms’ data needs for training and validation further amplify privacy concerns. The limitations of data anonymisation as a privacy preservation strategy have been underscored by advancements in re-identification techniques [106].

The ‘black box’ nature of certain AI algorithms clouds the decision-making process, posing issues of transparency and accountability [107]. This opacity has implications for informed patient consent and intensifies the risk of algorithmic bias, which could lead to potentially harmful healthcare decisions [108].

Current regulatory structures struggle to adapt to these rapidly evolving technologies, with global data flows and multinational tech companies compounding the challenge. Emerging solutions, proposed via interdisciplinary efforts, include implementing robust data governance frameworks with secure data transmission protocols, data minimisation strategies, and dynamic consent models [109]. Innovative cryptography-based methods, such as homomorphic encryption and differential privacy, offer promising pathways for balancing data utility with patient confidentiality [110].

Creating interpretable and explainable AI models can help address transparency and accountability issues. The push for ‘explainable AI’ aims to make AI decision-making processes more understandable and reliable [111]. Addressing algorithmic bias requires the mindful collection of diverse and representative datasets and the inclusion of fairness-centric evaluation metrics within AI systems [68].

The advancement of AI in mHealth necessitates a strong ethical framework, prioritising human dignity, autonomy, and privacy. While these challenges are substantial, they are not insurmountable. Through collective efforts from technologists, healthcare professionals, ethicists, regulators, and patients, the successful integration of AI with mHealth can be achieved. Thus, ethical responsibility, privacy respect, and a steadfast commitment to improving patient outcomes should guide us in navigating this transformative healthcare landscape.

8. Concluding remarks

8.1 Recap and reflections

The interplay between mHealth, AI, and medical imaging delineates a transformative trajectory in healthcare, which this review has sought to explore comprehensively. The merger of mHealth with AI holds the promise of personalisation and efficiency in healthcare service delivery.

AI, specifically machine learning, has exhibited extensive capacities in enhancing clinical decision-making processes, encompassing a spectrum from disease diagnosis to prognosis. This review has analysed the current significance of AI in healthcare and proposed potential future trends.

The evolution of medical imaging from conventional to digital modalities has been a linchpin in clinical diagnosis. The introduction of AI in this field has catalysed transformative changes in image acquisition, interpretation, and management. The synthesis of AI and mHealth in this realm is incrementally improving healthcare quality, with the review providing a detailed overview of innovative applications and case studies in diagnostic imaging, radiology, and pathology imaging. Despite the myriad benefits, these technologies also introduce a set of challenges that demand attention from a technical and regulatory standpoint, underlining the necessity for equilibrium between innovation, patient safety, and privacy.

Looking towards the future, this review highlights the potential ramifications of these advancements on healthcare delivery and patient outcomes. It emphasises the crucial role policymakers and health leaders must assume in steering this transformation while preserving the ethos of patient-centred care.

8.2 Highlighting the transformative impetus

To underscore the transformative potential of AI, mHealth, and medical imaging is to herald a new era in healthcare. The integration of these technologies paves the way for ubiquitous, personalised, and proactive healthcare that transcends theoretical speculation and edges towards actualisation.

AI, primarily through machine learning, bolsters the healthcare ecosystem via its proficiency in real-time, high-throughput data analytics, diagnostic precision, and prognostic accuracy. Concurrently, medical imaging is being reshaped, with AI algorithms fuelling advancements in image acquisition, interpretation, and analysis.

The efficacy of mHealth is underlined by its inherent accessibility, with an ever-increasing prevalence of mobile devices surmounting geographical barriers and enabling patient autonomy. The confluence of AI with mHealth technologies in the sphere of medical imaging enables sophisticated diagnostics at one’s fingertips, marking a significant stride towards equitable healthcare.

Collectively, these advancements herald an era of proactive, predictive, personalised, and participative medicine. As we navigate policy, regulatory, and ethical hurdles, we stand to gain a healthcare system that epitomises patient-centred, data-driven, and equitable care. The potential for these technologies in healthcare is vast and tantalisingly within our grasp.