Comparison of pulmonary congestion severity using artificial intelligence‐assisted scoring versus clinical experts: A secondary analysis of BLUSHED‐AHF

Abstract

AimAcute decompensated heart failure (ADHF) is the leading cause of cardiovascular hospitalizations in the United States. Detecting B‐lines through lung ultrasound (LUS) can enhance clinicians' prognostic and diagnostic capabilities. Artificial intelligence/machine learning (AI/ML)‐based automated guidance systems may allow novice users to apply LUS to clinical care. We investigated whether an AI/ML automated LUS congestion score correlates with expert's interpretations of B‐line quantification from an external patient dataset.Methods and resultsThis was a secondary analysis from the BLUSHED‐AHF study which investigated the effect of LUS‐guided therapy on patients with ADHF. In BLUSHED‐AHF, LUS was performed and B‐lines were quantified by ultrasound operators. Two experts then separately quantified the number of B‐lines per ultrasound video clip recorded. Here, an AI/ML‐based lung congestion score (LCS) was calculated for all LUS clips from BLUSHED‐AHF. Spearman correlation was computed between LCS and counts from each of the original three raters. A total of 3858 LUS clips were analysed on 130 patients. The LCS demonstrated good agreement with the two experts' B‐line quantification score (r = 0.894, 0.882). Both experts' B‐line quantification scores had significantly better agreement with the LCS than they did with the ultrasound operator's score (p < 0.005, p < 0.001).ConclusionArtificial intelligence/machine learning‐based LCS correlated with expert‐level B‐line quantification. Future studies are needed to determine whether automated tools may assist novice users in LUS interpretation.

Prognostic significance of Lung ultrasound for Heart Failure Patient Management in Primary Care. A Systematic Review

“…Additionally, LUS-guided treatment was associated with a lower risk of Major Adverse Cardiac Events (MACEs) [59,60]. Nonetheless, while LUS improved fluid status with a significantly greater reduction in the number of B-lines during the first 48 hours, it did not reduce heart failure readmission [53,55,62].…”

“…The LUS showed higher sensitivity ratio 1.2 (95% CI, 1.08-1.34; p < 0.001) compared with CxR [62,63,71], computerized tomography (CT) , and echocardiogram [74] in the diagnosis of HF, and using LUS with the clinical evaluation reduced diagnostic errors as compared to [CxR+Nt-proBNP] combination [64]. However, mortality was associated to significantly lower IVC collapse [65,70,71], and a greater number of lung B-lines; and higher NT-proBNP levels [65,[68][69][70] without differences in the BIA parameters.…”

Prognostic Significance of Lung Ultrasound for Heart Failure Patient Management in Primary Care: A Systematic Review

“…However, no significant differences in death rates were observed [55][56][57]. Additionally, treatment guided by lung ultrasound (LUS) was linked to a reduced risk of Major Adverse Cardiac Events (MACEs) [58,59], and a significantly greater reduction in the number of B-lines during the initial 48 h, but it did not reduce heart failure readmission [57,60,61]. (4) The results of LUS remained independent of NT-proBNP levels [32,43,50,62].…”

“…The LUS showed higher sensitivity ratio 1.2 (95% CI, 1.08-1.34; p < 0.001) compared with CxR, computerized tomography (CT), and echocardiogram [61][62][63][64][65][66] in the diagnosis of HF and using LUS with the clinical evaluation reduced diagnostic errors as compared to [CxR + Nt-proBNP] combination [67]. However, mortality was significantly associated with lower inferior vena cava (IVC) collapse [53,65,68] and a higher number of lung Blines, as well as elevated NT-proBNP levels [54,[68][69][70], with no differences observed in the bioelectrical impedance analysis (BIA) parameters.…”