Fig. 4: Heatmap overview of the evaluation metrics used in the reviewed studies.
From: Large language model integrations in cancer decision-making: a systematic review and meta-analysis
(Top) The table presents a 3 × 3 matrix defining the data and evaluation types, with parentheses indicating the number of papers in each category. Each paper is assigned to only one category. (Bottom) The heatmap illustrates the 15 evaluation categories identified across 34 salient papers, following the same data type classification as the top table. Unlike the top table, a single paper may employ multiple evaluation metrics and thus fall into multiple categories. The complete breakdown of data types and evaluation metrics for each paper is provided in Table 2.
