Construction of a Web-Based Dynamic Nomogram for Predicting the Prognosis in Acute Heart Failure

Figure 1. Selection of key features related to survival by the least absolute shrinkage and selection operator (LASSO) regression. (A) The LASSO regression coefficient trendlines of the 51 candidate features. (B) Determination of the tuning parameter λ by 10‐fold cross‐validation. Two vertical lines were drawn, representing a more concise model within one standard error. The tuning parameter λ = 0.094 was selected under the 1‐SE criteria with 6 non‐zero coefficients included. ACEI, angiotensin‐converting enzyme inhibitor; AF, atrial fibrillation; ALT, alanine aminotransferase; ARB, angiotensin receptor blocker; AST, aspartate transaminase; BUN, blood urea nitrogen; CRT, cardiac resynchronization therapy; DBP, diastolic blood pressure; DCM, dilated cardiomyopathy; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; IHD, ischaemic heart disease; LBBB, left bundle branch block; LVDd, left ventricular end‐diastolic diameter; LVDs, left ventricular end‐systolic diameter; LVEF, left ventricular ejection fraction; MAP, mean arterial pressure; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; NYHA, New York Heart Association; PASP, pulmonary artery systolic pressure; PBW, phase…

Figure 2. The DVCNS [D: DBP; V: VHD; C: CRT; N: NT‐proBNP; S: rest scar burden] heart prognostic nomogram for predicting the risks of 1, 3, and 5 year all‐cause mortality in patients with acute heart failure. The level or category of each variable corresponded to a score on the points scale. After summing the scores of all variables and locating the total score on the total points scale, a line was drawn straight down to 1, 3, and 5 year survival scales and the estimated probabilities of death were obtained. *P < 0.05; ** P < 0.01; *** P < 0.001. CRT, cardiac resynchronization therapy; DBP, diastolic blood pressure; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; VHD, valvular heart disease.

Figure 3. Kaplan–Meier curves of overall survival (OS) stratified by quartiles of risk scores (Q1: <0.27; Q2: 0.27–0.59; Q3: 0.60–1.87; Q4: ≥1.88) in patients with acute heart failure. Multiple comparisons between groups were adjusted by the Benjamini–Hochberg method.

Figure 4. Prognostic performance of the nomogram model to predict the risk of all‐cause mortality in patients with acute heart failure. (A) The mean time‐dependent area under the curves before and after 100 iterations of three‐fold cross‐validation. (B) Distribution of time‐dependent area under the curves after 100 iterations of three‐fold cross‐validation at 1, 3, and 5 years.

Figure 5. The 49‐sample bootstrapped calibration plots of the nomogram model to predict the risk of all‐cause mortality in patients with acute heart failure at (A) 1, (B) 3, and (C) 5 years. Blue line represents the ideal fit; circle represents nomogram‐predicted probability of death; cross represents the bootstrap‐corrected estimate; and error bar represents the corresponding 95% confidence interval.

Figure 6. Decision curves of the nomogram model to predict the risks of (A) 1, (B) 3, and (C) 5 year all‐cause mortality in patients with acute heart failure. Dark red line represents net benefit of the nomogram model; light red line represents net benefit of the hypothesis that all patients were alive; blue line represents net benefit of the hypothesis that all patients were dead; the remaining line represents the respective net benefit of each variable included in the nomogram model. Clinical applicability of the model increases with increasing net benefit. CRT, cardiac resynchronization therapy; DBP, diastolic blood pressure; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; VHD, valvular heart disease.