How do you use technology/policy to improve citizens’ lives? Tell us about your role or organisation.
I am a breast surgeon and head of the Breast Service at TTSH. I am also a clinician-scientist, and one of my main goals is to discover and develop new ways of improving the management and outcomes of women with breast cancer and other breast diseases. My work with patients gives me insight into gaps and unmet needs, and my role as the head of service provides me with a better understanding of the hospital system. One of the most exciting advances involves the adoption of Artificial Intelligence (AI) into medical imaging.
Breast imaging, mammography, and ultrasound constitute a major part of breast assessment. It is increasingly clear that AI will become a part of breast imaging. Interpretation of breast imaging requires specialised training and is dependent on the expertise of the radiologist. The most intriguing aspect of AI technology is the accumulation of such knowledge. The potential for AI technology in the interpretation of screening mammography is already appreciated, and AI software to serve as a second reader is already introduced to reduce mammography miss rates.
What was the most impactful project you worked on this year?
We are working on developing AI solutions for breast ultrasound, though in its infancy, it has great potential. Ultrasound is used to complement mammography in Asian women who more often have dense breasts, reducing mammography sensitivity and younger women. The use of ultrasound has contributed to increased biopsy rates, many of which return as non-malignant. AI algorithm can better characterise a lesion as benign or malignant, increasing accuracy of ultrasound assessment and reducing false-positive while ensuring that miss rates are minimal. Given the number of ultrasound assessments performed yearly, AI algorithms will have a wide impact.
What is one unexpected learning from 2021?
Social restrictions and the need to minimise in-person contact brought many aspects of everyday life to a halt. As Covid became a long-drawn-out fight, we found new ways to do things, using remote platforms such as Zoom for meetings and working from home. While telemedicine is widely used, it cannot replace in-person encounters. It is less personal, more challenging to pick up subtle body language, and it is not a good substitute for physical examination. But at the same time, there are advantages like reducing time spent on travelling and more access to seminars, workshops and meetings. It was a struggle initially, especially for non-tech savvy people like me, but I learnt that people can adapt. If we take the right attitude, change can make things better.
What’s your favourite memory from the past year?
Family time! Lockdowns and school closures meant that the kids were home a lot. I have to admit that although we drove each other crazy at times, it was a precious moment that I will always cherish. I learnt so much more about them when we went through homework, tried to figure out why we could not log in or submit assignments on Student Learning Space, (tried to) pack their room, baking and cooking, online shopping and binge-watched shows. For me, Covid made me take a break from my routine and allowed me to “re-calibrate” my priorities.
What’s a tool or technique you’re excited to explore in 2022?
I hope to make headway with developing AI solutions for breast ultrasound assessment. It is an area of unmet need, particularly in our local setting. Interpretation of breast ultrasound is dependent on the radiologist’s experience. AI algorithms can potentially improve the accuracy of lesion characterisation and reduce the need for histological confirmation through biopsy.
The other project is the development of a chatbot to help with post-operative wound care. It maintains an open real-time communication channel between the team and patients upon discharge. It can assure us that the post-operative recovery is uneventful or alert us to potential issues that need prompt attention. When optimised, it would reduce the number of post-operative clinic visits. It could also give patients more confidence for same-day discharge.
What are your priorities for 2022?
To have a prototype for the AI-assisted breast ultrasound assessment ready for validation and testing in patients. Ultrasound assessments will be more accurate. False-positive biopsy rates will decrease, leading to more efficient utilisation of resources. It could speed up reporting, and patients may be able to assess their reports within hours. It could also allow us to triage cases to ensure that more complex cases receive optimal care, while others may be suitably transferred to primary care providers.
Who are the mentors and heroes that inspire you?
My parents. They are not famous people. They did not have fancy jobs, nor did they have hobbies or great achievements. They always found a way to make the best of any situation and believed that change can be for the better. There were times when they had to take a step back and reconsider whether some things were ‘worth the effort’. They stopped if needed, but pushed ahead if they believed it would make a difference.
It is this spirit and attitude that inspires me. Research is less predictable compared to clinical work. Promising ideas and data are not always feasible in routine practice, and they do not always produce the expected clinical benefit. Funding is always an issue, especially for less established teams. Regulations and documentation can seem daunting. But we have to make the most of the given circumstances. We have to continually re-look and re-assess the data at hand. Look at alternatives, press on and find a way if we believe in our work’s potential.
What gets you up in the morning?
The practical answer is I have to get the kids to school. I am not the sort who jumps out of bed excited to experience the challenges and offerings of a new day. I get up in the morning because I have a job to do. Don’t get me wrong. I love what I do. I like operating; I like seeing my patients in the clinic.
I like exploring new research ideas, looking at data, and brainstorming. But not every day goes well. Some days are great, some days are not so good, and some days are downright awful. But, I am lucky to have more good days than bad ones. So I go to bed most days happy and content that the day was well spent, satisfied that I did my job well. I suppose you could say that this is what gets me up the next morning.