A Nigerian Phd Dentistry student, John Adeoye, at the Hong Kong University, alongside other doctors, have developed an AI-based web tool for the prediction of patients’ oral cancer risk.
As part of his doctoral research project that focuses on refining the early diagnosis of patients with oral cancer, John Adeoye, alongside other doctors, have developed an AI-based web tool that will help to predict if a patient has oral cancer.
Adeoye is currently pursuing a degree in Oral and Maxillofacial Surgery (OMFS) at Hong Kong University (HKU) with an emphasis on Oral Oncology. He had his primary and secondary education in Oshodi, Lagos State, before proceeding to the College of Medicine, the University of Lagos, to study Dentistry for six years. He had his compulsory one-year internship at the Dental Centre, Lagos University Teaching Hospital (LUTH), before proceeding for youth service in Abuja. He also had one year of general dental practice before proceeding to HKU.
According to Adeoye, many times, before oral cancer occurs, there are white (leukoplakia) or red (erythroplakia) patches in the mouth that may have been present for many years.
Dr Richard Su, Dr Siu-Wai Choi, Dr John Adeoye and Dr Li-Wu Zheng of the Faculty of Dentistry, HKU
“These patches have an inherent predisposition to become oral cancers and are known as premalignant or potentially malignant oral lesions. Not all these lesions become oral cancers, and currently, there are no objective ways for clinicians to determine which lesions may eventually become oral cancer.
“This led us to construct an AI tool as a freely-available web platform that would be able to perform this task for dentists and other health professionals involved in the care of these premalignant diseases.”
Doctors in the research team are Clinical Associate Professor in OMFS, Dr Richard Su; professor of OMFS, Dr Siu-Wai Choi; Clinical Associate Professor in Oral Medicine, Dr Li-Wu Zheng; Research Assistant Professor in Clinical Artificial Intelligence, HKU Faculty of Dentistry, Dr Mohamad Koohi-Moghadam.
Others are Dr Anthony Lo of the Department of Pathology, Queen Mary Hospital, Hong Kong; Dr Raymond Tsang and Dr Velda Chow of the Department of Surgery, Li Ka Shing Faculty of Medicine, HKU and Professor Peter Thomson, College of Medicine and Dentistry, James Cook University, Queensland.
Adeoye added that the AI-based web tool can provide the probability of oral cancer development in patients with white premalignant oral lesions every month after their diagnosis up until 270 months adding that ‘Deepsurv’ is the AI-algorithm on the backend of the web tool. This works in a similar way as the human brain to generate the predictions based on what it has learned from existing data of patients that have had this disease previously.
“However, before a prediction may be generated on the web tool, the user would have to enter 26 pieces of information such as age, medical history, details of the white patch, treatment history, and pathological findings.”
To test the efficacy of the tool, the dentist noted that the team had to perform both internal and external validation of the AI model before deploying it as a web-based tool.
“The internal validation was performed using a subset of the Hong Kong clinical data used to build the AI model while the external validation was done using another patient dataset from Newcastle Upon Tyne, UK.
Adeoye said the AI model performed well in both validation exercises, noting that, however, that is only one phase of validation to be performed. He said in the future, the team will utilise patient data from other countries, adding that they are currently collaborating with researchers in the Department of Oral and Maxillofacial Pathology and Biology, Faculty of Dentistry, the University of Lagos (UNILAG), to enable the team determine how well the model will perform on Nigerian patients with premalignant oral diseases.
The AI web tool, if found to have good performance and impact in the future, would provide knowledge on oral cancer risk to doctors and patients to guide the treatment and monitoring of patients with premalignant oral lesions. It is, however, still undergoing other validations to determine its performance in other dental and general medical centres worldwide. It is at the moment, only a research tool and investigators can test it in their populations to determine and report how well it performs.
Adeoye believes that Nigerian’s attitude toward their health is relatively poor. This, according to him, also affects their behaviour to oral health.
“The neglect shown to taking care of their dentition and mouth is even worse compared to how serious they take care of their general health. Hopefully, this situation will improve in the future with awareness programmes that will educate people on the link between oral health and general well-being as well as the benefit of proper and regular visits to oral health providers.”
He observed that the most common issue people face in terms of oral healthcare is the fact that people do not keep regular (six-monthly) visits with dental health providers. He noted that this may be due to people’s perception of oral health care.
“Many people believe that dental diseases are indolent and not life-threatening which makes them pay less attention to it. However, this is not entirely true as complications from some untreated diseases in the mouth can result in severe morbidity and even death.
“Beyond perception, other persons who may want to visit the dentists may not have the means to finance their dental consultations and treatment. This will cause them to default on routine visitations and divert their meagre funds to other needs.”
As regards the academic system abroad and in Nigeria, Adeoye noted that the obvious gaps in his opinion, pertain to the abundance of resources and opportunities available to students abroad compared to Nigeria.
“I think governments and universities abroad give priority to students and their wellbeing while ensuring that learning takes place in a comfortable environment.”
He said that an improvement in the attitude of stakeholders, as well as budgetary allocation, will be a step in the right direction to bridge the gap between the educational system and the standards available in developed countries.