Sustainable healthcare solutions aim to address global health challenges while minimizing environmental and economic impacts. They focus on creating resilient, equitable, and environmentally responsible systems, promoting resource-efficient technologies, reducing costs, and making essential services accessible to underserved populations. With the global population projected to exceed 9 billion by 2050, sustainable healthcare ensures efficient resource use, eco-friendly practices, and mitigation of environmental pollutants. Sustainable healthcare innovations aim to balance financial and environmental costs with the demand for high-quality medical care. Improvements include telemedicine, decentralized care delivery, and renewable energy-powered facilities. Point-of-care diagnostics and portable devices provide early disease identification in limited resources, minimizing reliance on centralized labs and total expenses [[1], [2], [3]]. High sensitivity and specificity, rapid response time, and the ability to provide quantitative and qualitative data in real-time are some of the major characteristics of point-of-care (POC) sensing [4]. POC sensing enhances diagnostic accessibility in remote or resource-limited settings, reduces delays, and lowers operational costs, making it more resource-efficient than traditional diagnostic tools, and its field is rapidly expanding [5]. In optical sensing, these characteristics are achieved through the interaction of light with the target analyte, allowing for non-invasive and real-time detection [6]. Optical sensors, using light-based techniques like fluorescence, surface plasmon resonance, and Raman spectroscopy, identify biomarkers or infections, providing real-time data for quick medical decisions. They are suitable for clinics, distant locations, and limited resources, improving patient outcomes and healthcare procedures [7]. Lab-on-a-chip (LoC) technologies reduce waste and reagent use by enabling multiplexed diagnostics with minimal sample volumes. Biodegradable instruments identify illnesses and pollutants, while wearable sensors reduce clinical visits and invasive tests. Major optical sensing techniques, including fluorescence, UV-Visbile, and Raman spectroscopy, have been explored for low analyte detection. Surface-enhanced Raman scattering (SERS) has achieved ultra-high sensitivity, enabling single-molecule detection of target analytes in complex samples [8,9]. This makes SERS particularly sensitive in POC systems, where quick, accurate diagnostics are crucial for healthcare applications [10].

Over the years, SERS has emerged as a powerful optical-based diagnostic tool for personalized health management [[11], [12], [13], [14], [15]]. SERS technology offers an environmentally friendly alternative to traditional diagnostic methods, offering multiplexing capability, ultra-high sensitivity, label-free detection, and single laser application. Advances in technology have led to reusable substrates and can be integrated into wearable devices for continuous, non-invasive health monitoring [[16], [17], [18], [19], [20], [21], [22]]. SERS technology offers personalized treatment plans, reduced trial-and-error, and high precision, reducing healthcare costs. Its early detection capabilities prevent disease progression and minimize treatment costs. As technology advances, implementation costs decrease, making it viable for resource-limited settings. Integrating SERS with 3D printed micro and nanosensors could create customized, on-demand diagnostic tools [[23], [24], [25], [26], [27]]. SERS offers unique fingerprint spectra of molecular binding events with low concentration measurement, particularly sensitive for liquid samples due to water's low scattering cross-section, gaining wide biological and medical applications [28]. Research demonstrates SERS's ultra-sensitivity for disease diagnosis in various samples, including urine, blood, serum, plasma, and saliva. Its versatility in measuring samples in various forms allows for early disease detection and monitoring, while its specificity enables the identification of biomarkers in clinical samples [29]. These features are essential for applications such as cancer diagnostics, heart disease, wound monitoring, toxic sensing, where early and precise detection can significantly improve disease diagnosis [[30], [31], [32], [33]]. The sophistication of SERS-based sensors has relied on manufacturing modalities. The weak Raman signal can be enhanced by 104–1010 using a photolithographically produced metallic nanostructure known as SERS providing single molecule sensitivity [17]. It is noted that biomedical research produces huge data and medical records which require high-end computer assistance for quick analysis to remove high variability and distinguish between similar spectra patterns.

Artificial intelligence (AI) is a computer science approach that develops programs and algorithms to make devices intelligent and efficient for tasks requiring human intelligence. It includes subsets like machine learning (ML), deep learning (DL), and neural networks (NN) improving medical sciences by simplifying human intervention in clinical diagnosis and decision-making [[34], [35], [36], [37]]. The integration of AI and SERS is crucial for unlocking the full potential of SERS in healthcare applications, enhancing precision, scalability, and usability. ML-enhanced SERS is revolutionizing the screening, diagnosis, and early detection of various conditions, including cancer, infectious diseases, neurodegenerative disorders, cardiovascular and metabolic diseases, rare and genetic disorders, and environmental and occupational health. Early detection of breast, lung, prostate, oral, and cervical cancers can be achieved through biomarker detection in blood/saliva, label-free tissue analysis, and real-time SERS-ML during surgery. SERS-ML can also identify drug-resistant bacteria, viral infections, neurodegenerative disorders, cardiovascular and metabolic diseases, rare and genetic disorders, and environmental and occupational health risks.

Large and complicated datasets produced by SERS may be processed by AI, especially ML techniques [38]. This makes it possible to find patterns and minute variations in spectral data that could point to certain biomarkers or stages of illness [39,40]. Some of the well-known examples are early-stage cancer identification using medical images, personalized medicine for genetic disorders, remote patient monitoring from wearable devices, predictive analytics tools for disease outbreaks, etc. [[41], [42], [43]]. To detect low-concentration analytes more accurately—which is essential for identifying biomarkers at the earliest stages of disease—AI can reduce noise and improve the signal quality of SERS data. By training ML models to identify particular spectral patterns linked to particular illnesses or substances [44], the specificity of SERS as a diagnostic tool may be increased [[45], [46], [47]]. For time-sensitive outcomes along with predictive analysis, AI can recognize these different patterns and remove ambiguity to minimize data variations for better clarity of the result [[48], [49], [50]]. Over recent years, both non-neural and neural algorithms have been deployed in biosensors to process biomedical data, distinguishing between healthy and unhealthy users and quantifying biomarker levels [51].

AI integration allows for fully automated workflows, reducing manual intervention and improving reproducibility. AI models can identify and enhance subtle features in SERS spectra, reducing false positives and negatives. AI-driven algorithms make SERS accessible for non-specialist users by simplifying data interpretation. AI can deconvolute overlapping signals, enabling simultaneous detection of multiple analytes in complex samples. AI models account for substrate-specific variations, facilitating standardization and patient-specific analysis [45]. AI-SERS platforms enable longitudinal studies and continuous monitoring of disease progression or therapy response. AI-SERS synergy is expanding into fields like nanomedicine, environmental sensing, and food safety. Combining AI-SERS with other technologies enhances multi-functional capabilities [52] (Fig. 1).

Also, other difficulties faced in the SERS system are preprocessing steps which are noise removal, smoothing, baseline correction, background subtraction, and cosmic ray removal. ML can greatly automate these spectral processes and improve the efficiency of the raw data (data collection, preprocessing, and downstream analyses) for high accuracy [38,53]. ML proved to be a versatile algorithmic approach capable of effectively managing vast, intricate, and diverse datasets which mainly consist of Support Vector Machine (SVM), NN, Principal Component Analysis (PCA), Partial Least Squares (PLS), convolutional neural network (CNN), distributed arithmetic (DA), artificial neural network (ANN), random forest (RF), Quadratic Discriminant Analysis (QDA), linear discriminant analysis (LDA), partial least squares (PLS-DA), etc. [54]. These algorithms find application across various healthcare domains, spanning from diagnostic procedures and treatment strategies to personalized drug selection for individuals [55]. Therefore, AI and ML which could provide numerous algorithms in data analysis with interpretation ability to screen, monitor, and diagnose diseases for personalized treatment has been integrated with SERS [56,57]. Further, the incorporation of the Internet of Medical Things (IoMT) would provide a real-time solution for better health management. We will also discuss the merits and limitations of different types of AI-ML-IoMT integrated SERS systems, including nanomaterial-based sensors, microfluidic devices, wearables, etc. [58,59].

This article will provide a comprehensive analysis of the existing state of research, highlighting the design, advantages, and application of SERS-based diagnostic assay in conjunction with emerging computer-based tools. Further, this review article will explore the integration of optical sensing incorporation with AI, ML, and IoMT technologies to develop next-generation SERS systems to improve the personalized healthcare system. This futuristic system can revolutionize the modern healthcare system to improve medical diagnostics through a customized medicine approach. Finally, in the concluding section, the article will highlight the opportunities, challenges, and prospects of utilizing SERS, AI, ML, and quantum sensing in health management applications.