Skip to main content
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: Abdom Radiol (NY). 2021 Jan 19;46(6):2556–2566. doi: 10.1007/s00261-020-02892-x

Figure 2.

Figure 2.

Decision trees from CART analysis are shown. Decision tree 1 includes the combination of auto-LPDI and assessment of CE as a binary sign. Decision tree 2 includes auto-LPDI and CEvol. Contingency table results and AUCs for the two analyses is presented in Table 3. Decision tree analysis revealed an optimal cut off of ≥ 12% for ruling in MHAI when CE is either absent (decision tree 1), or diminutive at < 0.05 mL (decision tree 2). (CART-classification and regression trees, auto-LPDI- automated liver parenchymal disruption index, CE- contrast extravasation, CEvol- contrast extravasation volume)