Entry for:2020 Queensland Women in STEM Prize
Knee arthroscopy is a minimally invasive surgery to diagnose and treat knee pathologies. Over 4 million surgeries are performed each year, costing the global healthcare system 15 billion $. This procedure is complex for the surgeons to perform and can frequently cause unintended damage to the patient. To reduce these shortcomings, our research group is developing robotic systems to make this surgical procedure more safe, precise and eventually autonomous. In my research, I am creating a sort of GPS to guide the robot during the surgery using ultrasound imaging. Ultrasound is non-invasive and allows to capture 3D volumes in real-time. I am investigating its feasibility for knee arthroscopy through experiments on knee phantoms, volunteers and cadaver models. To automatically interpret ultrasound images, I am implementing advanced machine learning algorithms, to identify and track the internal knee structures, which is vital for the robot to navigate inside the knee joint.
Nearly all of us experience knee pain of some sort in our lifetime. If knee pain is not treated, patients are at risk of falling into a downward spiral where they become less active and gain weight, worsening the knee pain and eventually leading to knee replacement. The first line of minimally invasive diagnostic and treatment procedure is known as knee arthroscopy. Recent studies show that this procedure not only requires a long learning curve, but in many cases, it can also cause unintended to the patient and/or post-operative complications.
Ninety-three orthopaedic surgeons all over Australia responded to a questionnaire to determine their perceptions about the level of difficulties involved in knee arthroscopy and their predisposition towards adopting robotic platforms to improve the surgical procedure. About half of the respondents involved reported the occurrence of cartilage damage in 1 out of 10 cases, and one third in 1 out of 5 cases. A smaller number (about 7.5%) declared that the damage occurred in every procedure performed. Around 99% of the participants reported difficulties in visualizing some sections of the knee while operating and around 86% would use robotic systems if they ensured decreased cartilage damage.
Driven by these issues, in our transdisciplinary research team, we are developing robotic knee arthroscopy systems to improve clinical outcomes for the patients and to support the surgeons during this procedure, limiting their physical and mental stress. The development of this technology holds potential to reduce the cost of arthroscopy procedures, promoting sustainable healthcare.
As part of this research team, my contribution is to introduce ultrasound imaging in this surgical procedure, in order to overcome the difficulties associated with the knee tissues visualization and identification during surgery. Through my research project, I am also building the tools for automatic interpretation of ultrasound images that could bring benefits to many other clinical applications.
Ultrasound is broadly used to scan many body regions due to its capability to visualize both bony surfaces and soft tissues, e.g. abdominal, vascular, muscular-skeletal sonography. Moreover, it has many advantages over other imaging modalities such as cost effectiveness, portability, non-invasiveness and volumetric real-time capability, making this imaging modality appealing for many applications. Despite these facts, US imaging is currently not exploited at its full potential due to challenging aspects in the scanning process and image interpretation. Sonographers (i.e. professional specialized in the use of ultrasound imaging) are in fact trained for several years to reach baseline competency in ultrasound-based applications.
The automation of ultrasound image interpretation could be adopted in clinics, to fasten (and thus potentially increase the number of) the diagnostic procedures, possibly improving their quality by increasing the consistency in image acquisition, and quantitative analysis of the pathology. These applications could be beneficial both to developed health systems and to third world countries and remote territories lacking sonographers and surgeons to perform the medical examinations.
In my research path, I attended many conferences and workshops in Queensland and around the world to have the chance to present my research and work together with people from different nationalities and research fields.
My PhD project is part of the Australia-India Strategic research Fund that promotes innovative research between the two countries. For this reason, in the past years, I had the chance to work with our Indian partner team and to participate in a joint research project at the Indian Institute of Technology-Madras in Chennai (India), where each research team presented and discussed their research progress.
With our research team, I also travelled to Munich (Germany) and to Udine (Italy) to explore possible collaborations with a German company (Imfusion) and an Italian University (University of Udine).
In 2018, I was part of an exhibition made in collaboration with a robotic industry (KUKA) at the IEEE International Conference on Robotics and Automation in Brisbane, where we showed a robotic arm performing autonomously an ultrasound scan of a knee phantom. Being this conference one of the most important in the world for robotics, I had the chance to meet and discuss with the top experts in this field.
In 2018, I also participated to the 13th IEEE-EMBS International Summer School on Biomedical Imaging in Saint Jacut de la Mer (France). During the school, I attended many courses taught by the key opinion leaders in my research field, presented my ongoing project and met colleagues from more than 20 different countries.
I loved this journey. And I hope through this competition I will have the chance to disseminate even more my research findings!