General characteristics
A total of 86 patients with gynecologic cancer who underwent chemotherapy and 30 healthy control subjects were included. The mean age of patients with gynecologic cancers was 50.5 years (IQR 45.0-56.0 years). Among the 86 patients, 48 (55.8%) were diagnosed with ovarian cancer, 14 (16.3%) were diagnosed with fallopian tube cancer, 20 (23.3%) with uterine neoplasm and 4 (4.6%) with trophoblastic tumor. Among the patients, the cardiovascular risk factors were hypertension in 6 patients (7.0%) and diabetes in 3 patients (3.5%); 6 patients (7.0%) had a history of cardioprotective medication use. The median number of chemotherapy cycles was 6.0 (IQR 3.0-11.0). The median number of chemotherapy regimens was 1.0 (1.0, 2.0). 50 patients (58.1%) received only one regimen, among which the taxol plus platinum regimen was the most commonly used (n = 42, 48.83%). Of the 86 patients enrolled, 41 (47.67%) completed the CMR follow-up. During a median interval of 6 months (IQR 3–9 months) between two scans, these patients had undergone a median of 3 cycles (IQR 2–5 cycles) of chemotherapy.
Deformation parameters in healthy controls subjects and patients with gynecologic cancer
No significant difference in LVEF between the healthy control subjects and patients with gynecologic cancer (p = 0.594 ) was observed. The absolute value of LV global radical strain (GRS) (37.30 ± 8.94% vs. 44.32 ± 8.44%; p < 0.001), GCS (-22.12 ± 3.05% vs. -24.08 ± 2.13%; p = 0.002) and GLS (median -15.72% [IQR-17.13 to -13.58%] vs. median-17.60% [IQR -19.05 to -15.84%]; p < 0.001) were all lower in patients compared with controls (Fig.2).
Among 86 patients with gynecologic cancer, 77 patients were classified as PLVEF group, and 9 patients were classified as RLVEF group. We found a decrease in the absolute value of GRS (37.99 ± 8.65% vs. 43.96 ± 8.36%; p = 0.007), GCS (-22.42 ± 2.98% vs. -24.18 ± 2.10%; p = 0.004), GLS (median -15.94% [IQR -17.19 to -13.68%] vs. median -17.55% [IQR -19.08 to -15.80%]; p = 0.001) and several regional strain parameters in PLVEF group compared with control group (Fig.3).
Changes in deformation parameters during chemotherapy
In the 41 patients who completed follow-up, the absolute value of GRS (37.30 ± 8.94% vs. 31.4± 12.11%; p = 0.003), GCS (median -22.67% [IQR -24.25 to -20.03%] vs. median -19.93% [IQR -21.83 to -18.89%]; p = 0.001) and GLS (median -15.72% [IQR -17.13 to -13.58%] vs. median -12.72% [IQR -14.73 to -10.43%]; p< 0.001) were all decreased at follow-up (Fig.4).
Association between LV deformation parameters and general characteristics
Univariate analysis showed that GCS was positively correlated with the number of chemotherapy regimens (r = 0.290, p = 0.007). In the multivariate analysis which was adjusted for demographical confounders, GCS was independently associated with the number of chemotherapy regimens (Standard regression coefficient [β] = 0.397, p < 0.001).
Correlation between LV deformation parameters and LVEF and serum myocardial injury biomarkers
Correlation analysis showed that all global strain parameters and most strain rate parameters were correlated with LVEF (p < 0.05) .
The results of serum myocardial injury biomarkers showed that all patients had normal cTnI (0~0.06 ug/L) and Myo (0~110 ug/L); 85 patients had normal CK-MB (0~5 ug/L) and CK (39~192 U/L). Among the serum biomarkers, CK-MB was correlated with GLS (r = 0.241, p = 0.038) and global longitudinal systolic strain rate (LSSR) (r = 0.297, p = 0.010); and Myo was correlated with GCS (r = 0.298, p = 0.009), global circumferential systolic strain rate (CSSR) (r = 0.342, p = 0.003), LSSR (r = 0.243, p = 0.034) and CDSR (r = -0.258, p = 0.025).