NBMRF Innovation Grants for 2024
Congratulations to Department of Obstetrics, Gynaecology and Newborn Health researchers who were successfully awarded Norman Beischer Medical Research Foundation (NBMRF) Research Innovation Grants for 2024.
These NBMRF Innovation grants were awarded to the following Dept OGN staff and honoraries in November 2023:
The Endometriosis AI Study: Developing Novel Deep Learning AI Methods to Better Detect Pelvic Endometriosis
Dr Debjyoti Karmakar, A/Prof Fiona Brownfoot, Prof Marimuthu Palaniswami, Mr Yuchong Yao, Dr Kate Stone, Ms Dorothy McGinnes, Dr Lenore Ellett.
The AI_Endo_Study aims to revolutionize endometriosis diagnosis and management using AI for ultrasound image analysis, improving diagnostic accuracy beyond expert sonologists. This responds to the challenge of diagnosing endometriosis, a significant issue in women's health. Utilizing computer vision, Support Vector Machines, and Convolutional Neural Networks, the project offers a non-invasive diagnostic alternative to invasive surgeries. Conducted with Mercy Hospital's endorsement and Norman Beischer Foundation funding, it involves a retrospective observational study using data from 2018-2023, including 500 individuals. The study focuses on refining AI models through deep learning, with laparoscopic and histopathological observations as diagnostic benchmarks. The project, beginning its AI modelling phase in February 2024, aims to improve surgical decision-making and set new standards in endometriosis care by optimizing surgical treatment pathways and enhancing ultrasound diagnostic specificity.
Development of an obstetric growth centile calculator and app
Dr Natasha Pritchard, Prof Sue Walker, Dr Emerson Keenan, Prof Stephen Tong, Dr Anthea Lindquist, and Dr Richard Hiscock.
Identifying babies that are small is a critical objective of pregnancy care. Small babies are at risk of many complications during pregnancy and birth, the most serious of which is stillbirth. A primary way of finding small babies during pregnancy is to estimate their weight using ultrasound. This weight is then plotted on a growth centile chart, which provides a percentile that compares the baby to others of the same gestational age. If the fetus is classified below the 10th centile, this acts as a key trigger to increase monitoring during pregnancy, and to offer interventions such as induction of labour or a caesarean section. The chart that is used to classify the baby’s size is therefore critically important, with substantial impacts on pregnancy care.
Unfortunately, many different growth centile charts are used within Australia. These different growth standards can lead to wildly discrepant classifications for the same infant, for example 5th centile compared to 35th centile. This means that simply performing an ultrasound at two different locations within Australia can change the centile attributed to an infant! This can alter the entire trajectory of pregnancy for the mother and baby. If the wrong chart is used, some pregnancies that are at high risk of complications will not receive the monitoring that they need. Other low-risk pregnancies could experience unnecessary interventions.
Over the past four years, the researchers have performed a series of rigorous investigations to help identify the ‘optimal’ growth standard to use in an Australian obstetric population. They have combined this research with international guidelines to come up with the key elements of an obstetric growth standard that we believe should be used. These include 1) Using a growth standard that is adjusted for the average birthweights of an Australian population (as the average birthweight in other countries can be very different), 2) Using a growth standard that compares to healthy babies still inside the mother’s womb (as babies born preterm are often smaller than average and so are not a good comparator) 3) Providing the option of a sex-specific growth standard (if the sex is known during pregnancy), as we know that male and female babies are different sizes even before birth, 4) Adjusting for each day of pregnancy, rather than each week of pregnancy, as babies can grow a lot throughout each week.
The study aims to:
1) Incorporate the elements discussed above, to create the ‘optimal’ growth standard for use in an Australian population
2) Design and publish an app that can be used on computers and smartphones, so that the growth standard can be adopted into clinical practice
3) Validate the growth centiles calculator, by applying it to a large, statewide dataset of all Victorian births over a 10-year period, and comparing it to other older obstetric growth charts
4) Audit its use in real-life ultrasound data from the Mercy Hospital for Women
Investigating the most effective heart rate detection methods for premature babies in the first minutes of life
Dr Elizabeth Baker, Dr Kari Holte, Prof Claus Klingenberg, Dr Rebecca Szabo, Ms Tatiana Zecher, Prof Peter Davis, A/Prof Marta Thio.
This study, titled ‘A validation study of a new dry-electrode ECG device for preterm infants’ aims to evaluate a new method of monitoring the heart rate of a preterm baby (born between 26- and 31-weeks’ gestation) immediately after birth. The study will compare current gold standard tests with the ‘NeoBeat MiniTM’, a reusable and consumable-free heart rate monitor that simply and quickly wraps around baby’s chest.
Preterm babies often require lifesaving support in the first minutes of life as they adapt to conditions outside the womb. The need for resuscitation and response to these lifesaving interventions are guided by the newborn’s heart rate. A low heart rate at five minutes is associated with a four-fold increase in risk of death for babies born at less than 32 weeks’ gestation.
A rapid and reliable method of measuring the newborn’s heart rate is critical. However, standard heart rate monitoring methods have limitations, including a delay in detecting heart rate, challenges in application during resuscitation, and availability in low resourced settings.
The NeoBeat Mini, a new therapeutic goods administration approved device, adapts existing electrocardiogram (ECG) technology into a small portable device that is quick and easy to apply around a preterm baby’s abdomen. Though approved for use, the device has not been validated in clinical settings. The Royal Women’s Hospital, and its international collaborators, aim to test if the NeoBeat Mini, compared with existing methods, accurately and quickly measures a preterm baby’s heart rate.
If accurate, the NeoBeat Mini may provide many advantages over current methods of heart rate monitoring, and aid clinicians in those first minutes after a baby is born to provide critical life-saving care.
Adenomyosis in Pregnancy: Can we see it, and does it change?
Dr Samantha Mooney, Dr Vanessa Ross, A/Prof Martin Healey, A/Prof Ricardo Palma-Dias, Dr Debbie Nisbet, Dr Sofie Piessens, Dr Kate Stone, Dr Tristan McCaughey, Dr Clair Shadbolt, Dr Natalie Yang, Dr Stephen Esler, Dr Marsali Newman, Dr David O’Keefe, Prof Peter Rogers
Adenomyosis is a condition where the lining cells of the uterus (the ‘endometrium’) are present within the muscular layer. This results in thickening and potential scarring of the muscular layer of the uterus. In recent studies, adenomyosis has been associated with an increased risk of troubles becoming pregnant, and complications during a pregnancy. However, there is very little evidence explaining why adenomyosis may be associated with these complications, including miscarriage, preterm birth, and high blood pressure conditions in pregnancy. A further complication reported to be associated with uterine adenomyosis is placenta accreta spectrum (PAS), where the placenta attaches too deeply to the uterine wall.
As a foundation for investigation of the links between adenomyosis and pregnancy outcomes we plan to perform a suite of descriptive studies.
1. A prospective observational study will be performed where high quality ultrasound techniques will be used to observe the appearance of the uterine muscle in patients with adenomyosis who become pregnant. Patients will be invited to participate if they attend for an early pregnancy ultrasound and have previously had a non-pregnant ultrasound demonstrating features that are strongly suggestive of adenomyosis. Participants will then undergo an additional ultrasound 6months postpartum, to again assess the uterine muscle for features of adenomyosis.
2. A retrospective cohort study across two major obstetric hospitals in Melbourne will be conducted. A common indication for MRI in pregnancy is for the investigation of abnormal placenta, eg PAS. It is hypothesized that patients with PAS have an increased likelihood of adenomyosis on their pregnancy MRI. Patients who have previously undergone MRI in pregnancy and been diagnosed with PAS will have their MRIs reviewed, with specific mention of features that suggest adenomyosis. This will be an observational study reporting the ability to see features of adenomyosis in pregnant patients undergoing MRI in pregnancy for investigation of PAS. In this same cohort, some patients will have undergone caesarean section with hysterectomy (as is often required treatment for PAS). The histology of this subgroup will be reviewed to describe the prevalence of adenomyosis in this cohort. Similarly, many of this cohort will have had a pre-pregnancy and/or early pregnancy ultrasound at one of the two study centres. The images of this subgroup will be retrospectively reviewed by one of the study experts, and categorized according to the presence of certain features of adenomyosis.
Tumor Necrosis Factor Inhibition: A New Frontier in Preeclampsia Therapeutics
Dr Natasha de Alwis, Professor Natalie Hannan, Dr Natalie Binder, Dr Alina Roman
Preeclampsia is a serious and deadly complication of pregnancy, taking the lives of over half a million babies and 70,000 mothers every single year. In pregnancies complicated by preeclampsia, poor oxygen supply causes the placenta to be in a state of stress, which initiates the prolonged release of toxic inflammatory and damaging factors into the maternal circulation. These toxins spread throughout the body causing major injury to the mother’s blood vessels and major organs, especially the cardiovascular system. As such, preeclampsia is a leading cause of maternal and neonatal death and long-term disability worldwide, with disease burden being highest in developing nations where obstetric and neonatal care is not readily accessible. Unfortunately, there is no effective treatment for preeclampsia; current clinical practice is to prematurely deliver the placenta and baby to save the mother. Urgent development of therapies for preeclampsia is desperately needed.
In this study, the researchers will investigate the potential of the anti-inflammatory medication, infliximab to treat preeclampsia. Infliximab works by inhibiting the action of a key toxic inflammatory mediator, tumor necrosis factor (TNF). It is commonly used in the clinical treatment of chronic inflammatory conditions such as rheumatoid arthritis and inflammatory bowel disease. TNF is elevated in preeclampsia and directly contributes to the inflammatory environment of this insidious disease, making TNF an excellent target for treatment.
Specifically, they investigate the potential of repurposing TNF inhibitor infliximab to treat preeclampsia in their laboratory models of human pregnancy. They will assess whether infliximab can reduce injury to the placenta and blood vessels, and closely examine the passage of infliximab through the placenta to the fetus. This study will uncover whether blocking these toxic factors released from the stressed placenta might be an effective way to treat preeclampsia.
Importantly, infliximab is currently used clinically in pregnant women with underlying inflammatory conditions, and is generally considered safe for use in pregnancy. Thus, they envisage expedited translation of our findings into human clinical trials in pregnant women. Concurrent to this project, the team and clinical collaborators will be performing small clinical cohort studies to investigate the action of infliximab in cohorts of pregnant women at risk of/or diagnosed with preeclampsia.
Together, this pre-clinical laboratory and clinical trials data will provide key information regarding the safety and efficacy of infliximab to treat preeclampsia. The team would then initiate a large-scale international clinical trial to validate the effectiveness of these drugs in different pregnant cohorts. If successfully translated, infliximab could be an exceptional first step towards treating a pregnancy complication that has existed for millennia but has never had an effective treatment – potentially saving hundreds of thousands of lives every year and preventing long-term injury in many more.
Predicting unplanned mother and newborn’s hospital admissions in the first year of life: a big data and machine learning partnership with Safer Care Victoria
Melvin Marzan, A/Prof Lisa Hui, Prof Andrew Wilson, Karrie Long, Prof Mark Umstad, Dr Jake Valentine.
Childbirth is a profound life experience for families. Yet, it’s not without its challenges from both personal and societal perspectives. Our project is dedicated to making this journey safer for mothers and babies and cost-effective for the Victorian Health Care System. One of the most dramatic and sustained impacts of the COVID-19 pandemic is the reduction in hospital length of stay for both mothers and newborns. No Victorian study has yet explored the potential health and economic tradeoffs from the shortened length of stay. Our recent work with Safer Care Victoria has exposed unintended consequences of shorter length of stay, including an increase in unplanned readmissions of newborns for feeding problems and suspected infections. An unplanned hospital admission is hugely disruptive for young families as well as costly for the health system.
This project proposes to harness the power of big data, predictive analytics, and machine learning, to transform the way care for new mothers and babies by providing personalized risk assessment of unplanned readmissions. This will enable to identify those who might benefit from additional support, for example, through inpatient services, domiciliary midwife care, or community health services.
To make this happen, the team will tap into the richness of healthcare information contained in the statewide maternity data collections. These records will be linked to hospital and emergency admission datasets, medical benefits schedule (MBS) and pharmaceutical benefit schemes (PBS) datasets. The data will be teamed up with advanced epidemiological and statistical modelling and machine learning. By supplying the models with data from thousands of birth records, hospital readmission records, healthcare utilization through the PBS and MBS schedules, computers can be trained to predict if a mother or baby might need extra medical attention in the year after birth.
Machine learning can apply its computational power to routinely collected datasets to achieve better care for mothers, babies, and families. If the team can predict unplanned readmissions, they can step in with the right care, at the right time, for the right patient. This could prevent unnecessary trips to the hospital, save time, money, and most importantly, ensure better health outcomes for young families. The project’s envisions a Victorian healthcare system that doesn’t just wait for problems to happen, but actively prevents them. This means better healthcare decisions, resources used wisely, and happier families as emphasized in the “Targeting zero program” of the State Government of Victoria. It’s about giving families peace of mind, knowing that the best care is there when they need it.
Asphyxial injury and the Delivery of Oxygen after REsuscitation (the ADORE study)
Dr Shiraz Badurdeen, Prof Peter Davis, A/Prof Susan Donath, A/Prof Margarita Moreno-Betancur, Ms Kate Francis, Prof Jeanie Cheong, A/Prof Sue Jacobs, A/Prof Hamish Graham, Prof Stuart Hooper, Prof Graeme Polglas
Collaborators contributing data- Prof Helen Liley (Aus), Dr Rakesh Rao (USA)
Birth asphyxia is the result of a critical shortage of oxygen supplying the brain before and during birth. Despite improvements in obstetric care, birth asphyxia affects 3 in every 1000 births in Australia and over 1 million births per year worldwide. Damage to the baby’s brain from birth asphyxia leads to a condition called hypoxic-ischaemic encephalopathy (HIE). HIE is a leading cause of childhood death and lifelong disability, including cerebral palsy, epilepsy and intellectual difficulties. Despite the enormous health and economic burden, there have been no new treatments for HIE over the last decade.
Oxygen is the most commonly used drug during newborn resuscitation. Too little oxygen is harmful but the latest research in the laboratory found that following resuscitation, the baby’s brain is extremely vulnerable to receiving too much oxygen. Excessive oxygen damages the brain’s mitochondria, which are the cell’s powerhouse. This triggers the key pathways responsible for additional brain injury that typically follows the initial injury at birth from asphyxia. Based on these laboratory discoveries, the team looked at their own historical data of babies with HIE. The preliminary findings showed that high oxygen levels within 2 hours of birth increased the chances of death or disability at 2 years of age.
The aim to confirm the findings in a world-first study, creating a large, combined dataset of babies with HIE. Latest statistical methods will be applied to analyse data from babies who participated in 4 international research studies over the last 10 years. In line with ethical recommendations, the team will be making best use of existing high-quality, consistently measured data that research participants have already provided.
Specifically, the aim is to examine whether early exposure to high oxygen levels (measured using a blood test within 2 hours after birth) increases the risk of death or disability (measured at 2 years of age) in babies born at ≥35 weeks’ gestation with moderate-to-severe HIE. To achieve this, the researchers will combine the expertise of our diverse team of statisticians, clinicians, and scientists. Individual data from approximately 1300 babies from research studies in the USA and Australia will be collated. The combined dataset will be analysed using cutting-edge “causal inference” statistical methods. These methods are essential to accurately answer the question because they adjust for biases that distort the true cause-and-effect relationship between high oxygen exposure and death/disability. Therefore, the team will able to find out whether avoiding excess oxygen in babies with HIE could reduce the risk of death and disability.
This research may open a new avenue of treatment for babies with HIE, for whom the chances of a poor outcome remain unacceptably high. Unlike new drugs or devices, the implementation of our findings to clinical care will be quick. Getting oxygen right will become a priority in the care of these vulnerably babies, both in Victoria and around the world.
The Norman Beischer Medical Research Foundation Innovation Grants provide seed funding to support highly innovative research projects with potential for significant clinical impact. They may lead to major grant applications with such institutions as the NHMRC and the Medical Futures Fund.
These grants are open to any researcher at any career stage and are generally for one year, but in certain circumstances may be for longer.
More information about these grants and others can be found on the NBMRF website.