Advances in biomonitoring technologies for women’s health

Abstract

In global healthcare systems, sex and gender biases have favored cisgender males, which has led women and transgender individuals to be understudied and underrepresented in medical literature. Thus, these populations are largely overlooked in health policy making. Persistent gender inequalities, socioeconomic divides, and racial-ethnic discrimination, particularly in low-resource communities, have exacerbated women’s health concerns, delaying advancements in care and accessibility. However, recent years have seen the emergence of tracking technologies and wearable devices that enable long-term biomonitoring of key health biomarkers which promise to facilitate early disease diagnosis for women from all walks of life. These innovations value education and accessibility, which can break down barriers to health care access and management that has affected generations of women around the world. This review discusses emerging biomonitoring technologies for diagnosing and managing critical women’s health conditions as defined by the World Health Organization, including breast and gynecological cancers, vaginal infections, fertility, pregnancy and post-menopausal osteoporosis. Additionally, we examine the current commercial landscape of women’s health technologies, highlighting barriers to adoption, such as medical insurance access and socioeconomic status, as well as discuss opportunities for future innovation.

In global healthcare systems, sex and gender biases have led to a health gap that affects diverse populations of women. This review highlights emerging wearable, biomonitoring, and diagnostic technologies to address this health gap, examines the commercial landscape and barriers to adoption, and discusses opportunities for future innovation.

Introduction

Despite many achievements in women’s civil rights over the past century, gaps in healthcare, medical research, sociocultural standing and economic power continue to affect women’s lives in the modern day^1,2.

From an economic perspective, women have struggled with attaining equal resources to men because of barriers in accessing education, job opportunities, safe housing, and healthcare services¹. For instance, women’s opportunities in receiving medical insurance coverage and healthcare infrastructure resources are limited and often do not fully encompass women’s health needs³. This leads to nearly 10.8 million women remaining medically uninsured in the USA, and 1.5 billion women and girls worldwide lacking access to healthcare facilities or resources^3,4. The healthcare gap is exacerbated amongst populations of women in middle to low-income communities and transgender people^5,6. In addition, funding and investment for health research and medical innovation has a biological sex bias⁷. Despite $198 billion USD being invested in health research in the USA, only 4% is allocated to women’s cancer and 1% for all other female conditions⁷. Additionally, conditions such as hepatitis B, which affects men 2$\times$ more than women, receive 41$\times$ more attention and funding in clinical or health research compared to conditions predominantly affecting women⁷. This disparity stems from the viewpoint that ‘women’s health’ is niche, resulting in women-prevalent conditions being neglected and understudied by funding agencies, policy makers and researchers⁷.

Other than research funding, biological sex biases affect clinical and medical studies⁸. Historically, biological male participants, animal models and cell lines have been the bases of medical studies, which become generalized to women. These studies do not consider biological sex differences present in genetics, hormones, symptomology and bodily functions⁸. This bias results from the fear of negatively impacting women’s fertility and pregnancy, resulting in only 37% of women representation amongst academic journals⁸. Furthermore, amongst in vitro cell studies, 70% of cell lines used are derived from biological males⁸. Consequently, the lack of sex considerations and limited health research is highly associated with increased incidence of harmful, ineffective or inappropriate treatment strategies implemented for females⁸.

The presence of these gaps leads to women often feeling limited in their understanding of their own bodies and unaware of their health choices². Personalized, cost-effective and reliable technologies can address the concerns that women face regarding the overall limited medical data on women-prevalent conditions⁷. Women themselves have also shown interest in adopting wearable and wireless technologies for tracking their own health. For instance, there is a 77% acceptability rate towards using ovulation tracking via bracelet wearables⁹. Transgender individuals also express interest in leveraging technologies for understanding their health. In qualitative studies conducted amongst transgender individuals, key factors listed for these technologies include the promotion of fluidity and gender identity, ability to support different kinds of transitions, and easy-to-manage user interfaces for data concerning transition and gender identity^6,10.

The World Health Organization (WHO) and various government agencies have reported that the most prevalent diseases and concerns amongst women are cancer, reproductive health, maternal health, and osteoporosis (Fig. 1a–e)^11,12. Moreover, biomonitoring technologies, such as wearable devices and point-of-care (POC) diagnostic tools, have gained attention due their ability to enable early detection and long-term health management of these conditions. This is done by measuring and analyzing personalized vitals, hormones and biomarkers⁵. In this review, we discuss the emerging and accessible technologies for the diagnosis and monitoring of women’s health amongst diverse populations, based on the WHO’s reports. We describe the advantages and limitations of these emerging technologies, identifying gaps and areas in need of more attention or development. Furthermore, we highlight how artificial intelligence (AI) and machine learning (ML) algorithms present an unparalleled opportunity for improving women’s health diagnostics by enhancing the accuracy of current tools in predicting health problems¹³.

Fig. 1. Overview of women’s health conditions.

a Depiction of conditions over-represented in women. b Schematics are shown for the progression of gynecological conditions. c Depiction of breast health conditions, including breast tumors, adenomas, mammary duct abnormalities, and fibrocystic tissues. d Illustration of the monthly female hormonal cycle, highlighting the different phases of menstrual cycle. e The role of female hormones in the progression of postmenopausal osteoporosis and the activity of bone-related cells. Created in BioRender. Hosseinidoust, Z. (2025) https://BioRender.com/t20f565.

Wearable technologies for biomonitoring

Advancements in bioengineered wearable technologies have enabled real-time, long-term self-monitoring, providing women with reliable health data and gaining traction in both research and clinical practice (Fig. 2)¹⁴. This section focuses on wearable technologies for monitoring female prevalent conditions recommended by the WHO (Supplementary Table 1)^11,12. Additionally, critical conditions that remain underrepresented in wearable health monitoring solutions are discussed, such as cardiovascular disease (CVD) and chronic pain, because of the limited experimental data available on these conditions to inform wearable technology design parameters for women^15,16. Literature on these conditions mainly report observational studies and qualitative assessments from small pools of women¹⁶.

Fig. 2. Depiction of wearable biomonitoring technologies.

a Nocturnal BBT monitoring with wrist wearable and intravaginal temperature monitoring with wireless logger. b Components of adhesive patches in hormonal monitoring. c Adhesive patches in microfluidics for oestradiol and cortisol monitoring. d Antenna pair system in healthy breast tissues compared to malignant tumor tissue. e Smart clothing and textiles, such as smart bras for breast cancer monitoring, conductive yarn for sweat cortisol monitoring and in-sole sensors for osteoporosis gait monitoring. f Wireless biosensor using pH sensitive chips embedded in underwear, vaginal rings and tampon surfaces. Created in BioRender. Hosseinidoust, Z. (2025) https://BioRender.com/t20f565.

Biometric monitoring for fertility and pregnancy health

Monitoring fertility and maternal health is crucial for both the mother and child, as 810 and 6700 maternal and fetal deaths, respectively, occur every day^17,18. Additionally, the WHO reports that, annually, 40 million women develop long-term health conditions during pregnancy¹⁹. If left improperly diagnosed or unaddressed, these conditions negatively impact women’s lives for years after giving birth, showing a need for health monitoring solutions^19,20. Women in low-income and low-middle income countries can especially benefit, as 92% of maternal deaths in these countries are preventable by addressing the lack of access to high-quality healthcare resources, well-informed clinical guidelines and reliable data on women’s reproductive health²⁰.

Biometric parameter changes, trackable with wearable devices, are characterized by physiological fluctuations throughout the menstrual cycle and pregnancy²¹. These parameters include basal body temperature (BBT)²², peripheral temperature on the skin²³, sweat rate²⁴, and heart rate²⁵. Some wearables offer a non-invasive alternative for measuring these parameters compared to traditional clinical assessments that require physician facilitation. These include temperature measurements recorded with a thermometer either orally, rectally or intravaginally^26–28, thermocouple or infrared sensor heat mapping conducted in clinics²³, and off-body sweat sample analyses performed in laboratories²⁹.

Wearable devices for BBT monitoring should be capable of detecting fluctuations in the range of 0.5–0.8 °C, as this is the typical variation observed throughout the menstrual cycle^30,31. Generally, estradiol levels increase throughout the follicular phase, peaking at ovulation, which is associated with increased vasodilation and body cooling³². Subsequently, after ovulation, high progesterone levels lead to body heating by affecting the hypothalamus and central nervous system to increase the body’s thermoregulatory setpoint^27,33. The most common BBT monitoring wearables are wrist-worn activity trackers. However, the challenges with measuring daily BBT through smart activity watches and bracelets are that fluctuations throughout the day, depending on level of activity, lead to fluctuating accuracy³¹. These trackers are best suited for detecting biometric changes at night, when the body is at rest³¹.

BBT monitoring via wearable technologies for intravaginal temperature tracking offers a stable and accurate method for continuous reproductive health assessment, compared to manual or wrist-worn recorded temperature measurements at specific timepoints each day (Fig. 2a)³⁴. Intravaginal loggers, once inserted, provide automated real-time temperature measurements over multiple days without the need for removal³⁵. These devices, such as OvuSense, show 69.2% acceptability rate amongst women, with 76.9% reporting convenience and comfort during use over three menstrual cycles³⁶.

Wearable devices that monitor biometrics specific to pregnancy can provide an overall surveillance of the pregnancy’s trajectory³⁷. Pregnancy-specific biometrics include maternal heart rate (interchangeable with pulse rate but focused on mother rather than fetus)³⁸, maternal cardiac function (traditionally monitored via electrocardiogram (ECG) or transthoracic echocardiogram (TTE))³⁹, uterine contractions⁴⁰, blood oxygenation³⁹ and sleep patterns⁴¹. For instance, wearable soft sensor systems leveraging traditional methods, such as ECG, electrohysterogram (EHG), and Doppler ultrasound, that can be worn over the abdomen provide a comprehensive solution for maternal biometrics tracking⁴². Moreover, sensors integrated into conductive clothing, referred to as smart textiles, offer a non-invasive approach to monitoring uterine contractions, maternal heart rate, and respiration^43,44. Although still in early development, large-scale clinical testing is required to refine smart textiles for broader adoption.

Gestational diabetes mellitus (GDM) is a common complication developed during pregnancy that can benefit from biometric-based wearable technologies. GDM affects ~14% of pregnant women worldwide and leads to 10$\times$ increase in the risk of postpartum type 2 diabetes^45–48. GDM is characterized by an intolerance towards glucose that develops during the second or third trimester of pregnancy, with lifestyle being the number one risk factor^45,49. Lifestyle factors impacting the likelihood of developing GDM include activity levels, body mass index (BMI), stress levels, nutrition, and sleep patterns^45,50,51. Wearable technologies are designed to continuously track and record pregnancy and diabetes biometric parameters, such as physical activity indicators (e.g., number of steps, sitting and standing occurrences, actigraphy, heart rate), blood and interstitial glucose levels, renal function, and circadian rhythm^{41,45,50–53}.

Commercial wearable tracking devices worn on the hips and wrists capture specific biometric data over long time frames, sufficient for second and third trimester monitoring and GDM detection⁴⁵. Such devices include the Abbott FreeStyle Libre Pro⁴⁵ (interstitial glucose levels), Medtronic Guardian Connect⁵⁰ (tissue glucose under the skin), CamNTech ActiHeart device⁴⁵ and Firstbeat Bodyguard 2⁵⁰ (focused on heart rate variability to detect abnormalities), GENEActiv⁴⁵ (focused on sleep analysis), Exsed from UKK-Institute⁵⁰ (focused on tracking movement), Vivosmart 3⁵⁰ and Fitbit Charge 3⁵³ (step count and speed). Combining these data onto one wearable device can provide a comprehensive lifestyle profile for pregnant women. For instance, Ava Bracelet trackers capture data on movement during sleep, sleep duration, pulse rate, breathing rate and skin peripheral temperature⁴¹. These biometrics, collectively, provide results on sleep analysis, emotional states and stress levels throughout pregnancy and postpartum⁴¹.

On the research front, miniaturized electrochemical biosensors or immunosensors targeting biomarkers associated with disease progression show promise for GDM monitoring^50,54. For instance, an electrochemical wearable technology, combining miniaturized circuitry, cost-effective material and microneedle sample collection, has been proposed for detecting β-hydroxybutyrate and/or lactate alongside glucose. The former two being key components for the diabetes ketosis pathway seen in GDM^55,56. This technology achieved a LOD of 50 μM with 95% stability over 6 h in artificial interstitial fluid⁵⁶. Cystatin C (Cys-C) is another such biomarker shown to increase severity in the presence of GDM. Cys-C is associated with multiple metabolic and physiological mechanisms as well as adverse immunological, renal and fetal growth effects⁵². In a label-free miniaturized electrochemical immunosensor wearable, leveraging screen-printed electrodes coated with MXene and gold nanoparticles, Cys-C ranges of 50–5000 ng/mL were monitored over 30 days when placed on skin^52,57. Multiplex detection with dopamine and uric acid was also used to provide comprehensive insight into the renal function during GDM⁵⁷.

Another important aspect of GDM management is reducing reported feelings of failure, anxiety, and powerlessness many pregnant women face after receiving a diagnosis of GDM⁵⁰. By tracking both self-reported and wearable monitored biometrics on accessible mobile health (mHealth) applications, pregnant women can remain informed about their own health and lifestyle decisions^46,58. mHealth applications tracking nutritional health and blood glucose levels, inputted manually 3–6 times a day, are amongst the most used worldwide, such as GDm-Health (UK), Dnurse (China), MoTHER (Australia), MyDiabby (France), Diamond (Australia), and Pregnant+ (Norway)⁵⁸.

Despite the notable advancements on this front, the main barrier faced is the need for accurate data interpretation to direct proper management of GDM⁴⁹. Currently, even with proper utilization of wearables and smartphone applications, many women seek healthcare provider support to understand the significance of collected mHealth data. These women report hesitance in making health decisions or behavioral changes without professional feedback⁵⁰. As such, wearable and mHealth technology are implemented as an enhancement to the clinical care for pregnant women rather than as a complete substitute⁴⁶.

Next generation designs of these mHealth technologies are expected to leverage remote professional feedback tools built into user interfaces, which can increase adoption amongst different demographics^50,51,58. One example of such emerging wearable devices includes the Fitbit Flex activity and nutrition tracker, which incorporated a feedback and encouragement text messaging system to promote positive behaviors and lifestyle change goals⁴⁸. Another step for wearable pregnancy monitoring involves launching community outreach and education initiatives focused on the wearables technologies in clinics and healthcare systems, amongst pregnant women and their healthcare providers. Such initiatives will bridge digital literacy gaps and encourage users to adopt these technologies⁵⁹. Wireless and remote communication is also key, as it enables remote monitoring between pregnant women and their clinicians, allowing for telemedicine application in pregnancy care⁶⁰. Telemedicine has the advantage of allowing pregnant mothers to monitor their health alongside a health professional from the comfort of their home, reducing need for clinic visits⁶⁰. However, for these initiatives and communication platforms, diverse populations of women need to be considered, including those from different cultural backgrounds, age groups, socioeconomic standings and those with limited internet access^61,62.

Hormonal monitoring wearables

While general biometric wearables, such as smartwatches and fitness trackers, provide useful data, they fail to capture the complexity of menstrual cycles, hormone therapies, and their physiological implications. Integrating hormone biosensors into existing platforms could enhance diagnostic capabilities and lead to more personalized healthcare solutions.

Hormonal fluctuations play a central role in the menstrual cycle, influencing the fertility window and ovulation⁶³. In addition to progesterone and estradiol, other key hormones of the menstrual cycle include follicle-stimulating hormone (FSH), estrone-3-glucuronide (E1G), cortisol and pregnanediol glucuronide (PdG), which dip and peak at different phase of menstrual cycle⁶⁴. For example, sweat-based nanobiosensors integrated into microfluidic wearable devices show promise in detecting hormones non-invasively (Fig. 2b, c)^63,65. Aptamers functionalized onto gold nanoparticle-MXene (AuNP-MXene) electrodes enable highly sensitive hormone detection at picomolar concentrations, significantly enhancing predictive accuracy⁶⁶.

Cortisol is also a key biomarker of reduced bone health as it influences calcium absorption and bone metabolism leading to osteoporosis^67,68. Electrochemical detection of osteoporosis using sweat-based wearables, including microfluidics, microneedles, and textiles, enable non-invasive cortisol detection (Fig. 2b,c)^66,69,70. Smart textiles incorporating modified carbon yarn fibers coated with iron oxide (Fe₂O₃)-based conductive nanomaterials can selectively detect cortisol via chemically crosslinked monoclonal antibodies in a miniaturized nanobiosensor platform (Fig. 2e)^69,71. In addition, during pregnancy, cortisol monitoring is essential, as cortisol levels are 2–4$\times$ higher when pregnant vs. non-pregnant, which has been correlated with bone loss^72,73. One device that has the potential to address this risk by targeting cortisol monitoring in pregnant women is U-RHYTHM, which can monitor 24-hr ultradian rhythm fluctuations of cortisol and its derivates in the serum of pregnant women with different body weights^25,74. By collecting data on multiple hormonal signals, devices such as U-RHYTHM enable deeper investigations of underlying biological mechanisms associated with changes in these signals^25,74.

Despite established correlations between hormonal changes and women-prevalent conditions, for chronic pain⁷⁵, endometriosis⁷⁶, polycystic ovarian syndrome (PCOS)⁷⁷, osteoarthritis⁷⁸, CVD⁷⁹, and lung cancer⁸⁰, no wearable technologies currently exist target these conditions⁸¹. For instance, endometriosis, characterized by chronic menstrual pain, is associated with abnormal hormone levels⁷⁶. Gonadotropin-releasing hormone (GnRH) agonists, commonly used for treatment, ultimately suppress FSH, LH, and estrogen levels, leading to temporary menopause-like states and reducing bone density⁸². However, no wearable devices exist for real-time hormone tracking before, during, or after endometriosis treatment. Similarly, testosterone is the key biomarker for PCOS diagnosis, yet there are no at-home biomonitoring solutions for its detection⁸³. Estrogen is also particularly relevant for various health conditions, including migraines, fibromyalgia, CVD, and lung cancer^80,84–86.

Transgender considerations in wearables design

The literature on women’s health wearable technologies focuses disproportionately on design parameters, usability and experiences of cisgender women. However, a gender identity gap is also present in medical practice⁸⁷. Little consideration in medical literature has been given to transgender experiences, before, during and after undergoing transition⁸⁸. Fertility is another consideration regarding this health gap, as 76% of transgender individuals consider fertility preservation before transitioning. Yet only 12% successfully preserve their fertility, due to limited health guidelines¹⁷. As well, there is almost no collected and reported medical data for monitoring transgender health conditions, with little information on the potential effects of gender affirming hormone therapy on the body^89,90. Wearables for hormone-based monitoring of gender affirming therapy can allow transgender individuals to feel confident in how their body is transitioning through tangible, accessible data⁹¹.

For transgender women, undergoing male-to-female transition, estrogen therapy requires the reduction of testosterone from 200 ng/dL (biological male levels) to 75 ng/dL (biological female levels), by intaking estrogen promoting drugs, thereby increasing estrogen and progesterone levels in the body^89,90. When monitoring male-to-female transition, estrogen hormone therapy is a key factor for wearables design as it can lead to increased risk of disease by which estrogen is an underlying mechanism⁹². Additionally, one of the main side effects of estrogen promoting drugs, such as cyproterone acetate, includes increased depression and hyperprolactinemia⁹⁰.

Outside of hormone therapy, lifestyle factors play a key role in the development of long-term health conditions amongst the transgender population^88,93. For instance, lifestyle habits such as drinking alcohol, smoking and diet, influences the chances of developing breast cancer for transgender individuals^88,93. Additionally, gender dysphoria, experienced by transgender women and men, often leads to anxiety, depression and substance abuse, which can in turn affect biometrics like heart rate and sleep patterns^94–97. As such, a combined biometric and hormonal wearable device that takes gender identity into account would greatly improve experiences of transition⁹¹. Other challenges to overcome for transgender health technologies are the scarcity of health professionals knowledgeable in transgender health to assist in data interpretation, limited healthcare infrastructure to inform device design and integration into transgender lives, and lack of medical insurance plans supporting gender identity to receive high quality health care^88,91.

Wearables tracking abnormal tissues using imaging modalities

For various female-prevalent diseases, the clinical gold standard for diagnostics is imaging abnormal tissue growth or rapid tissue loss^98–100. For imaging tissue growths amongst women, the breast is an organ of concern as 50% of women develop fibro cysts⁹⁸, and 25% develop benign tumors¹⁰⁰, referred to as fibroadenomas, over their lifetime. Mammography and ultrasound are the international standard of screening and diagnosing patients^101–103. Similarly, osteoporosis in women is correlated with low estrogen levels that occur after menopause, resulting in lack of estrogen binding to osteoblasts and increased secretion of cytokines that promote osteoclast differentiation of progenitor cells. Ultimately, this leads to overactivity of bone resorption by high numbers of active osteoclasts (Fig. 1e)¹⁰⁴. For this objective, women aged over 50 years old, tissue loss particularly in bones, leads to 4$\times$ higher occurrences of osteoporosis compared to men¹⁰⁵. Regarding bone loss, x-ray-based imaging in hospitals, such as dual-energy x-ray or computed tomography, is the most common diagnostic method¹⁰⁶.

While effective, these gold standard diagnostic methods are limited in accessibility, relying on highly expensive equipment and healthcare systems¹⁰⁷. In communities and countries with limited resources and funding for medical equipment, wearable imaging modality technologies, in comparison, provide cost-effective and resource-efficient alternatives^108,109. Also, early onset diagnostics are limited, but having these conditions determined at early stages can prevent worsening disease symptomology¹¹⁰. Additionally, many reports from women indicate feelings of discomfort, fear and anxiety before and after undergoing diagnostic imaging because of prolonged wait times to receive results¹⁰⁷. To address these concerns, emerging literature points to miniaturizing and integrating of imaging modalities into wearable devices for a faster result¹⁰⁷.

Sensing antenna development is a promising component of breast imaging wearables^111–113. Antennas transmit electromagnetic waves (e.g., microwaves, ultrasound, or infrared) into breast tissue, with reflection analyzed to differentiate healthy from malignant tissue based on stiffness and density differences (109–167 kPa for cancerous vs. 10.9–12.4 kPa for normal tissue) (Fig. 2d)^113,114. While sensing antenna systems can provide full breast imaging, high probe counts (16–512 antennas) complicate user adoption. Recent developments in omnidirectional antennas, which allow simultaneous signal transmission and acquisition, have reduced probe requirements, improving user comfort¹¹⁵. Regardless, these systems remain technically complex and require clinical expertise for data interpretation.

Ultrasound breast patches offer better structural flexibility and compatibility with existing ultrasound imaging techniques and have thus been proposed as a preferred alternative for breast diagnostics and screening¹¹⁵. Full breast monitoring and deep tissue imaging of cysts as small as 0.3 cm has been reported with ultrasound breast patches composed of crystalline materials, making them a viable option for early-onset detection¹¹⁵. An additional advantage of crystal-based patches is their ability to detect electromagnetic waves in the terahertz frequency range, which is safer for patients and more compatible with current clinical workflows than microwave or infrared imaging^116,117. However, these patches are typically single-use and require clinical evaluation, limiting their application for continuous at-home monitoring^115,116.

Smart textiles provide a practical solution for everyday breast health monitoring by incorporating wireless sensors into fabrics like cotton, polyester, nylon, and denim (Fig. 2e)^112,118,119. Textile-based antennas, either stand-alone or embedded into smart bras, can be implemented to measure tumor size and detect tissue abnormalities¹²⁰. These antennas leverage scalable textile manufacturing techniques, such as high-throughput weaving, making them viable for commercial production¹²¹. Wireless connectivity allows real-time data transmission to smartphone applications, allowing women to monitor their breast health autonomously¹¹⁹. Currently, most wearable breast health technologies and smart bras focus exclusively on breast cancer detection. However, conditions such as fibroadenomas and fibrocystic breasts, which involve benign tissue growths, could also benefit from similar wearable monitoring approaches. Future iterations of smart textiles and biosensors should be adapted to differentiate between benign and malignant tissue, expanding their diagnostic and preventative monitoring capabilities¹²².

Imaging and tracking wearables can also help in early management of osteoporosis. Osteoporosis is a musculoskeletal condition characterized by bone mineral density (BMD) loss due to heightened osteoclast activity¹²³. The decline in estrogen levels and high potential of developing osteoporosis during menopause accelerates bone resorption, increasing the risk of falls, fracture, sarcopenia and total hip arthroplasty (THA)^124,125. Gait pattern can be assessed through motion sensors-based mapping and accelerometry, in combination with biometrics such as oximetry and cardiac monitoring, incorporated into wearable devices¹²⁶. Optical polymer fibers in textiles and walking devices track gait through heart rate, respiration, and gait cadence¹²⁷. Meanwhile, in-sole gait monitoring sensors utilizing inertial measurement units (IMUs) offer a discreet, real-time assessment (Fig. 2e)¹²⁵. For postmenopausal women, wearable sensors embedded in textiles or shoes are easily adoptable and can enhance compliance for more accurate monitoring. Regardless, these gait sensors primarily focus on regular walking patterns, with limited data on dynamic movements like exercise or abnormal walking patterns.

Wearables measuring BMD provide a more direct strategy for monitoring osteoporosis compared to gait. For instance, emerging wearable armbands use ultrasonic transduction to measure BMD, detecting femoral ultrasonic velocity (3404.46 m s⁻¹) and identifying osteoporosis through T-score testing (T-score ≤ −2.5)¹²⁸. Regardless, BMD abnormalities can also indicate conditions like rheumatoid arthritis, osteoarthritis, and Parkinson’s Disease, requiring BMD monitoring to be paired with electrochemical biosensing to enhance diagnostic precision^129–131. These wearables, while promising, have not been assessed with large participant populations of elderly, menopausal women, and will require preclinical and clinical trials before being brought to market.

Wireless wearable technologies for vaginal infection

Vaginitis, or vaginal infection, one of the most common reasons women seek medical attention, impacts sexual and reproductive health¹³². Symptoms include itching, pain, abnormal discharge, and odor, often leading to feelings of embarrassment, anxiety, and reduced quality of life^133,134. The three most common types of vaginitis are bacterial vaginosis (BV), vulvovaginal candidiasis (VVC), and trichomoniasis^135,136.

BV, the most frequent vaginal infection among reproductive-age women, is characterized by microbiome imbalances, particularly a loss of Lactobacilli and an overgrowth of anaerobic bacteria such as Gardnerella, Prevotella, and Mobiluncus^132,134. Diagnosis traditionally relies on microscopy-based Nugent scoring and Amsel’s criteria, assessing discharge appearance, odor, and vaginal pH (pH > 4.5). However, these methods suffer from poor sensitivity, inter-observer variability, and reliance on skilled personnel¹³². Similarly, VVC and trichomoniasis are diagnosed through microscopy, and in vitro pH assessment, both of which have limited accuracy¹³⁶. Polymerase Chain Reaction (PCR) diagnostic techniques enhance the specificity and sensitivity of detection by leveraging molecular mechanisms through microorganism DNA detection in biological samples¹³⁷. Regardless, current PCR platforms are only used for clinical and laboratory testing as equipment is costly and requires trained personnel, making it difficult for at-home monitoring¹³⁷.

Research on wearable vaginal infection diagnostics remains limited, underscoring the need for further innovation. Several pH-based biosensors have been developed for BV and VVC detection. For example, intravaginal biosensing rings monitor pH levels continuously using a fluorescent polyelectrolyte-coated porous silica sensor on a PDMS ring¹³⁸. Similarly, a startup company, ALMA, reported the development of biosensing underwear that integrates potentiometric and amperometric sensors into the gusset to track vaginal pH and lactate levels¹³⁹. Additionally, waterproof electronic decals (WPEDs) mounted on tampons measure vaginal fluid pH, transmitting data wirelessly to a smartphone (Fig. 2f)¹⁴⁰.

These biosensors provide cost-effective, real-time data, but they also face design challenges related to comfort, durability, safety, and hygiene. Issues with adhesion, connectivity, and interference from biological fluids (e.g., menstrual blood) further complicate their adoption. Additionally, pH monitoring alone is insufficient for definitive diagnosis, as abnormal pH is common to both BV and trichomoniasis, necessitating supplementary diagnostic methods¹⁴¹.

Beyond pH sensing, innovative biosensors targeting specific vaginal pathogens could enhance diagnostic accuracy. For instance, a thiolated aptamer-based AuNP sensor detects glucans on Candida albicans, triggering a colorimetric change from pink to blue upon pathogen detection. Compared to pH-based approaches, this method has increased specificity for VVC¹⁴². However, its usability may be hindered by both the subjective interpretation of results by non-experts and the complexity of the device’s assembly process, such as calibration and user interface navigation¹⁴³.

Despite advances in BV and VVC detection, wearable diagnostics for trichomoniasis remain unexplored. Expanding biosensor technologies to include pathogen-specific markers and multi-analyte detection could facilitate at-home vaginal health monitoring, offering more accurate, accessible, and user-friendly solutions.

Portable diagnostics of women-prevalent conditions

Portable diagnostic devices have notable advantages including POC, real-time results, and minimal infrastructure requirements, making them particularly beneficial for women’s health. However, research and development efforts have largely focused on vaginitis and cancer, leaving many other conditions unexplored. Expanding existing POC technology could substantially improve diagnostic accessibility and quality of life. WHO emphasizes that POC devices should align with ASSURED criteria—affordable, sensitive, specific, user-friendly, rapid, equipment-free, and deliverable¹⁴⁴.

Point-of-care technologies for vaginal infection detection

Timely and accurate detection of vaginitis is crucial, particularly in pregnancy, to prevent complications¹⁴⁵. Recent efforts have focused on low-cost microfluidic, paper-based, and electrochemical POC devices. The two most common approaches for vaginal infection detection are enzyme-activity-based and DNA-based methods.

Enzyme-activity-based technology for POC devices

Bacterial enzymes such as vaginolysin (VLY) and sialidase (SLD) are key virulence factors in BV¹⁴⁶. SLDs, found commonly in vaginal discharge, play a crucial role in cleaving sialic acid from cervicovaginal mucus and epithelial cells, promoting bacterial growth, biofilm formation, and toxin formation (Fig. 3a). Bacterial SLDs also is associated with undesirable pregnancy outcomes and preterm birth^146–149.

Fig. 3. Diagnostic devices for women-prevalent conditions.

a The mechanism of SLD activity, whereby disialylation of sialoglycon improves bacterial adhesion and biofilm formation. b LAMP-based lateral flow strips steps: DNA amplification accrues during the 1 h at isothermal condition; the test line of strip immobilized with anti-FAM antibody; conjugation with biotin/FAM-modified DNA; capturing streptavidin-modified AuNPs; test line signal. Copyright 2022 Frontiers¹⁵⁸. c Biomarker enrichment methods used for breast and gynecological cancer detection. d Microfluidic platform technology such as lateral displacement, µ-sieving, and channel surface modifications can be used for enrichment based on physical properties such as size and surface charges. Created in BioRender. Hosseinidoust, Z. (2025) https://BioRender.com/t20f565.

SLD has been the target of numerous colorimetric and fluorescent biosensors. A reported colorimetric biosensor uses 20-(5-bromo-4-chloroindol-3-yl)-a-D-N-acetyl-neuraminic acid (BCIN), which changes color upon contact with vaginal fluid containing SLD¹⁵⁰. To enhance reaction speed and enzyme absorption, poly(ethyleneimine) is used as a catalyst^150,151. Although colorimetric biosensors are user-friendly, they often lack quantitative precision. To improve sensitivity, fluorescent SLD detection can be leveraged alongside colorimetric methods. Disposable Y-shaped paper-based biosensors or well-plate sensing platforms incorporate cadmium selenide zinc sulfide (CdSe@ZnS) quantum dots (QDs) or fluorescein isothiocyanate (FITC)-modified antibodies can achieve quantifiable results through fluorescence^148,152.

LAMP-based POC devices

In the diagnosis of VVC, chlamydia trachomatis (CT), and neisseria gonorrhoeae (NG), current gold standard tests include cell culture technology, enzyme-linked immunosorbent assays (ELISAs), nucleic acid amplification tests (NAATs), and PCR-based assays. Of these methods, PCR is the most clinically used as it has >90% sensitivity and specificity^153,154. These techniques, while effective in clinics, are time consuming, labor some and difficult to incorporate into at-home POC devices.

More accessible POC-compatible alternatives to PCR are loop-mediated isothermal amplification (LAMP) and recombinase polymerase amplification (RPA). LAMP is a rapid, low-cost, and straightforward method for nucleic acid amplification, capable of amplifying the target gene within one hour under an isothermal cycle (50–60 °C) (Fig. 3b)¹⁵³. In comparison, RPA operates at lower temperatures (37–42 °C) and requires fewer primers. Although RPA is more temperature-flexible, LAMP is more widely used due to its resistance to inhibitors, making it ideal for complex clinical samples such as blood and vaginal fluids^153,155,156.

While lateral flow assays (LFAs) are commonly used in BV POC biosensors, they lack the sensitivity and speed of NAATs^137,157. To overcome this, LAMP-integrated lateral flow assays (LILFAs) have been developed for multiplex detection of NG and CT. LILFAs utilize biotin- and fluorescein-labeled DNA probes, forming a sandwich-like structure upon target binding, enhancing detection accuracy¹⁵⁸. However, LILFAs remain qualitative, and further research is needed to incorporate quantitative fluorescence-based detection for bacterial load analysis and infection staging^157,159.

Bacteriophage-based biosensing technologies

Bacteriophage (phage)-based biosensing is an emerging tool for the detection of bacterial pathogens in various settings^160,161. Phages selectively infect bacteria, injecting genetic material for replication, making them ideal for highly specific bio-probes in electrochemical and optical biosensors¹⁶². While phage-based diagnostics have been applied in various medical fields, no studies have explored their use in vaginal infection detection. Given their specificity, low-cost, rapid detection, and low detection limits, phage biosensors hold promise for BV diagnostics and multi-pathogen detection in the future¹⁶³.

Diagnostic technologies for breast and gynecological cancers

Early cancer detection significantly improves survival rates by enabling timely, less invasive treatment¹⁶⁴. However, 31% of breast cancer and 49% of cervical cancer diagnoses occur at late stages¹⁶⁵. As a result, five-year survival rates drastically lower, from 93% (stage 1) to 26.9% (stage 3) and 13.4% (stage 4)¹⁶⁶. Current detection methods—ultrasound, computed tomography, and magnetic resonance imaging (MRI)—lack sensitivity for early-stage diagnosis, are costly, or involve toxic chemicals¹⁶⁴. POC diagnostics can enhance early biomarker isolation and detection, as summarized in Supplementary Table 2.

Enrichment of early-stage women cancer biomarkers

Serological biomarkers are commonly used in immunosensors, ELISAs, and PCR, with key breast and gynecological cancer biomarkers including breast cancer associated gene 1 and 2 (BRCA1 and BRCA2), cancer antigen-125 (CA-125), human epididymis protein 4 (HE-4), mesothelin alpha fetoprotein, osteopontin, human epidermal growth factor receptor 2 (HER2), and human papillomavirus (HPV) virus. However, these markers may overlap with non-cancerous conditions, leading to false negatives^167,168. For instance, CA-125 is not detectable amongst 50% of women in the first stage of ovarian cancer¹⁶⁹. Moreover, traditional solid-tissue biopsies are invasive and fail to capture tumor heterogeneity in real time¹⁷⁰. In comparison, liquid biopsies offer a timely, non-invasive method for cancer detection of a diverse range of biomarkers, found in body fluids such as blood, urine, and saliva^171,172.

Alternative biomarkers for cancer detection focus on biomolecules secreted by breast and gynecological cancer cells into biological fluids, such as extracellular vehicles (EVs), circulating tumor DNA (ctDNA), microRNA (miRNA), and cancer-derived lipids or proteins. For example, exosomes, ranging from 30 to 200 nm in size, can also be secreted by cancer cells into body fluids at different stages of cancer development depending on the cancer type^141,142. Additionally, circulating tumor cells (CTCs)—shed from tumors during early metastasis—serve as early-stage cancer biomarkers^173,174. As all these biomarkers are found in fluid, future diagnostic methods will shift towards liquid biopsies, minimizing the reliance on more invasive solid tumor biopsy, which can severely alter the structure and function of female tissues¹⁷⁵.

Despite their promise, CTC concentrations in biological fluids are extremely low (1–100 CTCs per million hemocytes), necessitating isolation and enrichment¹⁷⁴. Conventional methods (flow cytometry, ELISA, ultracentrifugation) are time-consuming, costly, and inefficient for detecting exosomes (<200 nm)¹⁷⁶. Two key alternative techniques for biomarker isolation with minimal structural disruption and high size resolution are biochemical and biophysical separation^175,177,178.

Biochemical enrichment relies on biomarker interaction with biorecognition molecules. Magnetic beads (MBs) functionalized with cancer-specific probes enhance biomarker isolation in liquid biopsies. In addition, folate receptors, overexpressed in breast and ovarian cancers, are ideal MB targets for CTC capture (Fig. 3c)^179,180. However, biochemical methods are limited to singleplex detection, requiring multiple probes for multiplexing¹⁸¹. Microfluidic chips, with micrometric channels for single-cell studies, reduce sample volume and costs, making them widely adopted for cancer detection^182,183.

Biophysical separation, a label-free approach, overcomes biochemical limitations by utilizing filtration, inertial forces, ultrasound, or electrical fields¹⁸¹. Recent studies combine size, charge affinity, and surface modifications for enhanced isolation (Fig. 3d)^184,185. For example, CTCs preferentially adhere to positively charged surfaces, such as lipid or polymer-modified surfaces, leveraging their stronger negative charge compared to normal cells^181,186. CTCs can also be enriched through filtration or binding to antibody-functionalized polymeric nanoparticles (e.g., polystyrene), then collected downstream in a microfluidic chip^156,157.

Exosome isolation can also be achieved via surface charge manipulation. Chitosan-coated microfluidic channels, sensitive to pH changes, allow negatively charged exosomes to bind during sample flow. After capture and sample removal, pH modulation releases the exosomes, facilitating collection¹⁸⁶. This approach can be further enhanced by combining size and charge-based separation¹⁸¹.

POC technologies for breast and gynecological cancer

Standard cancer diagnostics require trained personnel and hospital-based equipment, limiting accessibility, particularly for women with restricted healthcare access. At-home POC devices, such as LFAs, provide affordable, rapid screening that meets WHO’s ASSURED criteria¹⁴¹.

LFAs are paper-based POC devices composed of a sample pad that interacts with the sample fluid and transfers it to the conjugate pad for labeling. The labeled sample moves through the strip by capillary forces, towards the detection pad, modified with aptamers, antibodies and other biomolecules¹⁸⁷. To increase sample interaction time and increase sensitivity, promising approaches include using magnetic materials, such as probe-modified magnetic particles, which attach to biomarkers on the detection pad and extend interaction time under a magnetic field¹⁸⁸.

However, conventional LFAs remain primarily qualitative or semi-quantitative, limiting their sensitivity compared to ELISA-based assays. To address sensitivity limitations, luminescent-based dual-mode LFAs incorporate fluorescent microspheres (FMs), QDs, and organic dyes, enabling both visual and quantitative detection^189–191.

While LFAs effectively detect abundant cancer biomarkers (e.g., CA-125, HER2, HE-4), CTCs and exosomes, due to their low concentrations, require preconcentration before detection. Achieving simultaneous preconcentration, signal amplification, and detection of CTCs and exosomes in LFAs remains a challenge, though emerging strategies have been reported. Pen-based LFAs, using a catalyzed hairpin assembly (CHA) technique, are especially promising in their ease of use and all-in-one design¹⁹².

Multiplex detection for women-prevalent cancers

Relying on a single biomarker for cancer detection can lead to high false-positive rates, as some biomarkers for women’s cancers are associated with other women’s conditions. For example, CA-125, a key ovarian cancer biomarker, is also elevated during endometriosis^193–195. Similarly, breast cancer subtypes—HER2-positive, Luminal A/B, and Basal-like—require distinct biomarker panels (HER2, estrogen receptor (ER), and progesterone receptor (PR)) to guide treatment selection^196,197.

Multiplex biosensors enhance diagnostic specificity by detecting multiple cancer biomarkers simultaneously. For instance, miRNA profiles (miR-21, miR-106a, miR-155) vary across breast cancer subtypes, necessitating multiplex detection strategies¹⁹⁸. Electrochemical biosensors are well-suited for the development of flexible sensors due to fast response, inherent miniaturization, convenient operation, portability, and integration of cost-effective electrodes. Unlike optical methods, electrochemical biosensors provide direct electrical readouts, enabling real-time biomarker analysis in complex biological samples¹⁹⁹. Enzyme-based electrochemical (EBEC) and electrochemiluminescence (ECL) biosensors enhance signal amplification and sensitivity, which can be achieved by labeling the probes with redox species such as methylene blue, horseradish peroxidase (HRP), and metal ions (e.g., gold, silver) for EBEC, or with fluorescent labels for ECL^200,201.

Multiplex LFA-based POC devices are particularly relevant for ovarian cancer, one of the most difficult cancers to detect in its early stages. Since HE-4 levels in urine vary depending on hydration status and urine concentration, creatinine levels are used as a normalizer for urinary biomarkers^193,202. It is also essential to measure the HE-4/creatinine ratio for detecting ovarian cancer in urine. The cut-off value for HE-4 is 3.5, with the ratio range being around 2 for healthy individuals and up to 45 for late-stage cancer patients. For this purpose, a dual LFA system has been reported with signals quantified using a smartphone and scanner¹⁹⁵. One strip was designed to detect HE-4 using a sandwich-like system with antibodies and antibody-modified gold shells as reporters, while the other strip was designed for creatinine detection through a series of chemical reactions.

Despite advancements, few LFAs are available for detecting multiple protein biomarkers. LFAs sensitivity, in both singleplex and multiplex detection, is often limited by the hook effect that prevents the formation of the antibody-antigen complex²⁰³. This leads to false-negative results when targeting more than two protein biomarkers at high concentrations. Vertical flow assays overcome this limitation. Surface-enhanced Raman spectroscopy (SERS)-vertical flow assays have shown promise for the multiplex detection of exosome proteins (Mucin short variant S1 (MUC1), HER2, and CEA) which are useful for identifying breast cancer subtypes. This vertical flow assay contains a single aptamer-modified test pad that captures each biomarker on a specific spot, followed by detection using SERS probes that offer ultrasensitive, non-destructive analysis capabilities and minimal photobleaching²⁰⁴.

Commercialization, barriers to entry and future scopes

Recent advancements in women’s health technologies–particularly in wearable biosensors and POC diagnostic devices–have accelerated the transition from conceptualization to clinical trials, driving increased stakeholder investment and introduction of these technologies into international markets (Supplementary Table 3)²⁰⁵.

Barriers to entry of women’s wearable technologies

A recent surge in technologies in the commercial space of women’s health have advanced significantly in the product pipeline for new product development (NPD)²⁰⁵. For instance, a majority of fertility wearable tracking devices are readily available for clinicians, researchers and consumers to purchase and use^206–210. Additionally, companies such as Dexcom and Freestyle Libre have led the development of continuous glucose monitoring wearables in the commercial space^211,212. For highly specialized devices, such as those targeting women-prevalent cancer screening and osteoporosis monitoring, these technologies are not as advanced in the product development pipeline, with many going through clinical trials or the early stages of product scale up^213–220. The success of these commercial devices relies on ease of integration into the currently existing infrastructure, where devices designed to be compatible with current medical best practices and mHealth applications have a higher chance of sucess^221,222. On the other hand, devices that are costly in terms of manufacturing and materials are more likely to fail^221,222. Women’s health is a rapidly growing commercial space, therefore intellectual property in the field is expected to grow and thus patenting new technologies becomes even more important for technologies geared towards commercialization²²³. Bringing women’s health technologies from prototype to commercialization involves multiple challenges. This section highlights key barriers in usability, adoption, and policy-making that impact market success.

Usability, adoption and compliance for diverse women

Women’s health technologies in commercialized spaces are most successful when they are designed for diverse populations of cisgender women and transgender people from different socioeconomic and cultural communities^224,225. Age is also essential to consider when designing user interfaces for monitoring and diagnosing disease associated with menopause. Digital knowledge gaps may restrict usability for older generations of women^226,227. Meeting these considerations can ensure compliance and autonomy when using digital devices²²⁸.

Furthermore, women from low socioeconomic status or enfranchised communities adopt biomonitoring technologies less than women of high socioeconomic status²²⁹. The reason for this is the lack of access to digital health platforms and internet needed for device function, as well as limited opportunities to partake in educational or policy decision making initiatives⁶². Often, the voices of these women are not heard on the world political stage, and these women do not have expectations that wearable devices can be designed for them, creating a market barrier in a similar way as differences in age. This is because of poverty, cultural disparities and income inequality, as well as home or job insecurity^62,230,231. These experiences are exacerbated amongst transgender individuals, who are over twice as likely to experience low socioeconomic conditions compared to the cisgender population in the US^232–234.

Global policy making for women’s health technologies

One challenge in the development of technologies is the lack of global policy making practices for women’s health. Within the past decade, a large positive shift towards the globalization of women’s health and sexual rights has taken place. On an international front, the United Nations (UN) and WHO acknowledged the lack of policy making and women’s sexual rights laws which has limited women’s access to healthcare and medical devices^235,236. Many international governments have taken initiatives and measures to remove gender and sex discrimination, and promote women’s self-autonomy, medical education and access to healthcare resources. This includes the United States (e.g., National Institutes of Health’s (NIH’s) Women Health Initiative), Canada (e.g., Health Canada’s National Women’s Health Research Initiative), Europe (e.g., European Union (EU) Manifesto for Women’s Health), and Australia (e.g., Australia’s National Women’s Health Strategy 2020–2030)^237–241. Additionally, international organizations, such as United Nations (e.g., United Nations Population Fund) and WHO (e.g., Global Strategy for Women’s, Children’s and Adolescents’ Health), are collaborating with countries in Africa, the Middle East, and Asia to initiate women’s health and equality policies by 2030^242,243.

Artificial intelligence in women’s health technologies

From a clinical perspective, AI and ML can be leveraged to significantly improve predictive capabilities and offer dynamic pattern recognition to enhance early-onset diagnostics, automate data collection and streamline care²⁴⁴. Examples of AI and ML algorithms of interest include convolutional neural networks (CNNs)²⁴⁵, artificial neural networks (ANNs)¹⁸, support vector machine (SVM)²⁴⁶, multivariate logistic regression (LR), random forest (RF), extreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM)²⁴⁷. In women’s health, AI and ML algorithms can enhance health outcomes for women facing pregnancy complications (e.g., preterm births, spontaneous abortion, GDM, preeclampsia, neonatal deaths, congenital anomalies)^18,246, vaginal infections^245,248, and cancers^249–256 (Supplementary Table 4).

In diagnostics, new biomarker discovery using AI can introduce unique or more all-encompassing diagnostic strategies by providing more holistic understanding of disease mechanisms for conditions affecting women, thus bridging the knowledge gap²⁴⁶. An example can be identifying new set of biomarkers in urine for determining GDM that affects many organ systems of pregnant women and has multiple disease mechanisms for different symptomologies^246,257. AI and ML algorithms can also collect and compile a wide range of predictive values simultaneously, including EHG, metabolic biomarker levels, ultrasound images of abdomen and cervix, and demographic profiles, to predict complication occurrences in the first trimester²⁴⁶. Additionally, AI-powered mHealth applications offer a valuable tool for assessing early symptoms and recommending timely interventions, an example includes AI-powered smart phone tools using MGH Perinatal Depression Scale (MGHPDS) for postpartum depression, which affects approximately 15% of mothers^258,259. Identifying diseases by comparing hundreds of sample images is another capability, as AI and ML algorithms can quickly pick up discrepancies that can be missed by the human eye. For instance, CNNs can detect the presence of yeast hyphae from microscopy images for detecting VVC, exceeding expert diagnostics by 4.58%²⁴⁵.

Some mHealth applications and wearable technologies that leverage AI and ML have already been commercialized for women’s health applications or are close to commercialization²⁶⁰. For example, the OvuSense platform, which has been incorporated into viO’s OvuFirst device, is a commercial device that obtains temperature data for fertility monitoring and utilizes algorithms for improved fertility predictions²⁶¹. Copan Diagnostics has also developed and released PhenoMatrix to detect multiple vaginal infections by comparing images of vaginal swab plate cultures²⁴⁸. Select studies have shown how AI has been implemented for improving accuracy of cervical and breast cancer diagnosis; these technologies have received Food and Drug Administration (FDA) regulatory approval for commercial use^249–256. One example is General Electric’s FDA approved AI-powered breast cancer prediction tool based on ultrasound data²⁶².

In applying AI and ML for clinical practice, there is a dire need to understand the effects of generalized data and how we can shift away from it for reliable outputs²⁶³. An untapped use of AI is in assessing the possibility of undetermined or undetected biological sex differences amongst diseases that affect both males and females²⁶⁴. One example is the use of AI-powered statistical analyses comparing women with and without menstrual cycles, as well as males, to determine the variance of body temperature to overcome sex biases by including equal data on females and males²⁶⁵. If trained on unbiased data sets, AI and ML have the power to distinguish factors that lead to predominance of diseases in women compared to men, such as CVD and osteoporosis²⁶⁴. These algorithms also have the potential to identify other unknown conditions that predominantly affect women, providing insight and knowledge into unidentified health gaps²⁶⁴. However, the main challenge in algorithm development is that they typically leverage historically reported biased medical data and biomedical research as part of their training processes^264,266. As mentioned, a large biological sex bias and gender identity disparity exists in literature and medical records, which could translate to the model after training^264,267. This barrier to entry for emerging AI and ML models necessitates the need for further biomedical research addressing this biological sex bias. It also highlights the importance of reporting unbiased data that includes equal representation of all sexes, genders and backgrounds²⁶⁸.

Outlook

This review highlights key bioengineering advancements in women’s health diagnostics and biomonitoring. As more prototypes and in vivo studies emerge, the field is progressing toward commercialization, transforming economic, healthcare, and societal perceptions of women’s health monitoring and diagnostics. One outstanding challenge in these emerging technologies is the limited data from underrepresented populations to inform device design and user interface accessibility. To ensure equitable healthcare solutions, future biomedical devices must be inclusive, addressing the needs of cisgender women and transgender individuals across diverse backgrounds.

As more wearable technologies and detection platforms for women’s health are created, integrating wearable technologies with ML and AI will not only enhance detection accuracy but also reduce user error in data interpretation. AI-driven systems can personalize diagnostics, ensuring precise monitoring of cancers, infections, and reproductive health conditions. AI algorithms can predict bacterial imbalances before a full-blown infection develops, allowing for earlier intervention²⁶⁹. Similarly, in cancer detection, AI can analyze biomarker fluctuations, predict cancer progression, metastasis risk, and immune responses (e.g., NK cell activation), leading to earlier and more targeted treatments.

The future of women’s health technologies is shifting toward wireless, AI-integrated wearables that provide real-time predictive feedback, enhanced user interfaces, and seamless remote monitoring. These innovations will assist women by enabling personalized, proactive healthcare management. By bridging bioengineering, AI, and digital health, the next generation of wearable diagnostics will transform women’s health monitoring into a more accessible, data-driven, and patient-centered paradigm.

Author contributions

T.F.D., Z.H., and W.G. conceived the idea of the review paper, designed the outline and scope, and provided funding for this project. Sh.M. and L.N. conducted the literature research, compiled the tables, and wrote the manuscript. Sh.M., L.N., and F.B. designed the figures. T.F.D., Z.H., W.G., Sh.M., L.N., Y.L., F.B., A.P., S.K., and A.B. edited the manuscript, references, figures and tables.

Peer review

Peer review information

Nature Communications thanks Ramses Martinez, and the other, anonymous, reviewers for their contribution to the peer review of this work.

Competing interests

The authors declare the following competing interests: W.G. is co-founder and advisor at Persperity Health. T.F.D. and Z.H. are principal investigators on an NSERC Alliance grant with Women’s Global Health Innovations.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-63501-3.