The burden of oral cancer falls disproportionately on people in low resource settings who lack access to oral cancer specialists. In India, oral cancer accounts for 23% of all cancer-related deaths. With this in mind, an international group of researchers tested the use of an app-based screening strategy for early diagnosis of oral cancer.
The team used an mHealth oral cancer surveillance program called Oncogrid for this study. Oncogrid offers “specialist consultation for dentists and health workers in low resource settings” in order to identify high risk patients susceptible to oral cancer.
The Android-based tool was developed to risk stratify patients and aid in the evaluation of their mouth by an oral cancer specialist. It contains oral screening software with a clinical decision algorithm for risk assessment and diagnosis. The mHealth system was developed using the MIT-developed Sana platform for picture & video exchange and the Open MRS open source EMR.
The study involved a tertiary care center and two remote sites with two arms in the study. One cohort was a target group from two different rural villages with people who had a high prevalence of oral cancer risk habits. They were seen in a primary health center.
A second opportunistic cohort was comprised of people who were seen by dentists in a primary dental clinic. The intervention consisted of a risk evaluation questionnaire in the Oncogrid app, taking a picture of any suspicious lesions with the mobile phone app, and sending the risk information and photos to an oral cancer specialist.
The specialist then determined if the image was nonneoplastic, potentially malignant or malignant. Patients with potentially malignant or malignant lesions were referred for a biopsy at the primary health center or dental clinic. Those at low risk received counseling and education focused on decreasing oral cancer risk behaviors.
In this study, they found that:
- In the targeted cohort, 45% of interpretable images were confirmed by specialists as suspicious while, not surprisingly, 100% were confirmed as suspicious in the opportunistic cohort.
- 48% of recommended biopsies were completed. Critically though, in the primary health center group, it was only 4% (1 out of 23) while in the dental clinic group it was 57% (61 out of 106).
These results were an indication that the Oncogrid app could be useful in identifying oral cancer in low resource communities, particularly among patients presenting to dental clinics. However, there is clearly more work that would need to be done to improve the yield among patients presenting to primary health centers & frontline health workers.
The researchers concluded that expansion of the use of Oncogrid would require studying larger populations. They also need to include a control group to determine whether the app is effective or some other aspect of the study is effective. For example, in the present study access to a dentist may be the key factor in identifying oral cancer, not the use of an app. In addition, the researchers need more rigorous statistical analysis of their findings. Descriptive statistics was provided in this study with no inferential statistics being used. Randomization of app use would also be useful in future studies.
Even with these future concerns and challenges, the researchers were able to help individuals in low resource communities in Bangaglore, and provided a step towards earlier identification and treatment of a problem that impacts a substantial number of people.