Skip to main content
. 2026 Aug 12;17:1907116. doi: 10.3389/fimmu.2026.1907116

Figure 5.

Calibration plot on the left shows observed versus predicted MPR probability, with values closely following the diagonal reference line, indicating good model calibration. Decision-curve analysis on the right compares net benefit across threshold probabilities for four strategies: Clinical plus ENI, Clinical baseline, ENI alone, Treat all, and Treat none, with Clinical plus ENI yielding the highest net benefit across most thresholds.

Calibration and clinical utility of the ENI-Score: (A) calibration plot comparing observed and predicted probabilities of MPR; (B) decision-curve analysis comparing the clinical-plus-ENI model, clinical baseline model, ENI-Score alone, treat-all strategy, and treat-none strategy.