Table 1.

Measured and transformed variables for each of the 21 subjects and 504 lesions.

Variable name	Description
Histologic evaluation	The ground truth of each lesion. Its possible values were: “Adenoma”, “Non-Adenoma”
Human judgment, S1 and S2	The optical diagnosis of the lesion by an endoscopist in each session, mapped as mentioned above. It takes the values: “Adenoma”, ‘Non-adenoma”, and “Uncertain”
Human confidence, S1 and S2	The confidence of the previous judgment expressed by the endoscopist. It takes the values: “Very high”, “High”, “Low”, “Very low”, “Uncertain”. We classified the confidence as “Uncertain” whenever the associated lesion evaluation was “Uncertain”
Algorithmic ai diagnosis	The diagnosis about a given lesion provided by the ai output in S2, as interpreted by an automatic algorithm. It takes the values: “Adenoma”, “Non-Adenoma”, and “Undetermined”
Perceived ai diagnosis	The endoscopists’ interpretation of the ai dynamic output in S2. It takes the following values: “Adenoma”, “Non-Adenoma”, “Uncertain”, and “I am not sure/I did not notice the output of the ai”
Evaluation of ai confidence	The endoscopists’ appreciation of the level of reliability of the ai diagnosis in S2. It takes the following values: “Very high confidence”, “High confidence”, “Low confidence”, and “Very low confidence”
Transformed variable name
Human correct diagnosis, S1 and S2	A binary variable that indicates whether each lesion was correctly diagnosed by each endoscopist, in each session
Accuracy${}_s$ of the perceived ai diagnosis	A binary variable indicating whether each lesion is correctly diagnosed by the ai. We considered the ai diagnosis perceived by the endoscopist. “Uncertain” and “not sure/not notice” were excluded from the computation
Accuracy${}_u$ of the perceived ai diagnosis	A binary variable indicating whether each lesion is correctly diagnosed by the ai. We considered the ai diagnosis judged by the endoscopist. Classifications of the ai output as “Uncertain” and “not sure/not notice” were conservatively considered errors
Confidence score, S1 and S2	A discrete numerical variable ranging 1 to 9 that measures the belief of each endoscopist about each lesion in each session. The score of 9 indicates a strong belief that the lesion is an adenoma. At the other extreme, the score of 1 denotes a strong belief that the lesion is not an adenoma. The score of 5 indicates “Uncertain” diagnoses