Peer-reviewed by 2 reviewers with median rating of 12/20. Review process was triple-blinded.
Round 1 (12/20)
Round 2 (12/20)
Conceptual advance and Impact5
It was very difficult to understand the technology and the data points it generated, yet the device itself. Methodology in reference to collection, there is no standardization. The data points to do analysis is not clear, no mention of controls, what is being measured, what is the sensitivity, specificity, accuracy, etc.
The title is too general, low resource meaning what? and oral lesion as all oral lesions? and what type of screening tool?
There should be more detailed information related to the study, how many patients? there are no detailed summary of how it is being detected.
Please describe in the legend the different chemistries. I would normalize the scales of intensities to be in the same range.
Results & Discussion
The results are not informative, they are just description of the method being used. Lack of detail in reference to what is being generated at to the data points and what type of training set was used, parameters, etc for each type of machine learning algorithm.
Poor or lack of date to justify that this is even a potential alternative, lack of samples (there is not inclusion/exclusion criteria), standardization of sample collection, no validation, etc.
More details of the number of data points, the detail of the technology, and the results (output) that is generated.
Please cite work from David Wong, UCLA.
There is no description as to the device and the detection method, the volume of saliva needed, etc.
There was a lack of a clear study design to grasp the take home message.
They are making too many assumptions without justification. No controls, or standardization of methods, what is the dyamic range, sensitivity, specificity of the device and/or approach. Poorly explained.