Chonthicha Tanking, Nanthasit Samansit, Chanwit Wuttichaipradit
In this small retrospective cohort, early LASr decline was associated with, and tended to precede, GLS change in women who subsequently developed anthracycline-associated CTRCD. These hypothesis-generating findings support LASr as a potential early, high-sensitivity adjunctive marker requiring prospective validation.
PURPOSE: Global longitudinal strain (GLS) is the recommended marker for cancer therapy-related cardiac dysfunction (CTRCD) but may miss the earliest myocardial changes. We examined whether early decline in left atrial strain (LAS) precedes GLS change in patients who subsequently develop CTRCD.
METHODS: In a single-center retrospective cohort of 62 women with breast cancer, speckle-tracking echocardiography measured LAS and GLS at baseline (T0), after the fourth anthracycline cycle (T1), and 3-6 months after treatment (T2). CTRCD (2022 ESC criteria) was classified by onset (T1 or T2). Analyses were exploratory.
RESULTS: CTRCD occurred in 26 patients (41.9%): 8 T1-onset, 18 T2-onset. In T2-onset patients, the early (T0→T1) relative decline in LA reservoir strain (LASr) already exceeded that of no-CTRCD patients (-19.5 ± 13.9% vs -6.4 ± 11.3%; p = 0.001), whereas early GLS change did not differ (p = 0.43). Early LASr change had the highest area under the curve (AUC) for subsequent CTRCD (0.78; 95% CI 0.65-0.90) versus GLS (0.59; DeLong p = 0.10). An exploratory 6.35% LASr decline threshold gave 100% sensitivity (95% CI 82-100%), 56% specificity (95% CI 40-71%), and 100% negative predictive value; bootstrap-corrected AUC was 0.77. LASr and GLS were only weakly-to-moderately correlated (r = -0.29 baseline; r = -0.41 early change).
CONCLUSION: In this small retrospective cohort, early LASr decline was associated with, and tended to precede, GLS change in women who subsequently developed anthracycline-associated CTRCD. These hypothesis-generating findings support LASr as a potential early, high-sensitivity adjunctive marker requiring prospective validation.