Table 1.

Examples for the five sentence categories and their corresponding curated sentences

Category	Gene candidate	Example sentence
Relevant	AFP-L3	AFP-L3 is strongly correlated with the HCC patient outcome: a high level of AFP-L3 is indicative of a poor patient prognosis.
Irrelevant	NGF	The serum levels of NGF were measured by enzyme-linked immunosorbent assay.
Negative	Cyclin D1	The expression of Cyclin D1 was not correlated with CDK4 expression, tumor grades, survival rate, and any clinicopathological parameters.
Named entity recognition error	TACE	Aim: To assess the safety and efficacy of TACE in very elderly patients.
Indefinite	AFP-L3, DCP	The specificity of AFP-L3 and DCP in the studied population was 78.5 and 100%, respectively.

The texts highlighted in bold indicate the gene candidates