Fig 1.
Prompt optimization process.
Table 1.
Prompt optimization process and corresponding average readability grade level (aRGL) scores.
Table 2.
Average readability grade level (aRGL) scores of the original radiology report, and outputs from the basic prompt “simplify” and the 5 best prompts. Average represents aRGL. Medians and Quartile 1 – Quartile 3 are depicted for each prompt tested on 255 radiology reports. Prompt A-E are significantly lower than both the Report and Simplify for all scales, p < 0.0001. Simplify is significantly lower than the Report for all modalities, p < 0.0001.
Fig 2.
Stratification between included modalities for the main primary prompt: Prompt D.
*, **, ***, **** correspond to p < 0.05, p < 0.01, p < 0.001, and p < 0.0001, respectively. aRGLs of the the final prompt and report were used.
Fig 3.
Accuracy, completeness, and inclusion of extraneous information of Rads-Lit tool.
Fig 4.
Rads-lit interface.