To investigate the multimodality imaging features of cardiac myxoma (CM) and explore the multimodality diagnosis model of CM.

One hundred and seventy-eight patients who were initially screened for CM by two-dimensional transthoracic echocardiography at Fuwai Hospital from July 2016 to August 2019 were retrospectively included in our study. Using pathological results as the gold standard, the subjects were divided into either a CM group or a non-CM group.The CM group was further divided into a typical CM subgroup and an atypical CM subgroup according to the characteristics of echocardiography. The clinical characteristics and the imaging features of echocardiography, multi-slice computed tomography (MSCT), and cardiac magnetic resonance imaging (CMR) were compared between groups by t-test, chi-squared test, or Fisher's exact test.

Of the 178 patients with an initial diagnosis of CM, 160 (89.9%) underwent surgical treatment and pathological examination (89.9%); 150 patients were diagnosed with CM, 1 with low-grade fibromyxoid sarcoma, 1 with undifferentiated pleomorphic sarcoma, 1 with angiosarcoma, 2 with hemangioma, 2 with lipoma, and 3 with thrombosis. The diagnostic accuracies of echocardiography, MSCT, and CMR were 93.8 %, 96.3%, and 100%, respectively. The proportions of patients with hypertension and atrial fibrillation in the typical CM subgroup were significantly lower than those of the atypical CM subgroup (25.4% vs 53.1%, χ²=8.978, P=0.003; 2.5% vs 18.8%; χ²=9.027, P=0.003); the left atrial anteroposterior diameter and the early diastolic peak flow velocity of the mitral valve were significantly higher in the typical CM subgroup than in the atypical CM subgroup [(39.21±6.34) mm vs (35.92±5.72) mm, t=2.357, P=0.020; (1.59±1.67) m/s vs (0.90±0.25) m/s; t=2.040; P=0.043]. The average long diameter of atypical CM was significantly smaller than that of typical CM [(34.81±17.43) mm vs (45.99±16.73) mm; t=3.324, P=0.001], and the width of tumor pedicle or base was significantly larger than that of typical CM [(13.02±7.28) mm vs (9.97±4.73) mm; t=-2.506, P=0.014]. The distribution, attachment site, morphology, mobility, presence of tumor pedicle, and presence of atrioventricular valve orifice obstruction in atypical CM were significantly different from those in typical CM (P＜0.05 for all). Low-attenuation was the main MSCT feature of atypical CM, and the proportion of patients with medium or mixed attenuation and enhancement was higher than that of typical CM, but the difference was not statistically significant (P＞0.05). The signal intensity of atypical CM on T1 and T2 weighted images was similar to that of typical CM (both P＞0.05). The average long diameter of the non-CM masses was smaller than that of typical CM, but the difference was not statistically significant (P＞0.05). The average short diameter was similar to the typical CM, and the width of the tumor pedicle or base was significantly larger than that of the typical CM [(13.35±6.80) mm vs (9.97±4.73) mm; t=-2.026, P=0.046]. The distribution, attachment site, morphology, presence of tumor pedicle, and mobility of non-CM masses were significantly different from those of typical CM (P＜0.05 for all). Based on the above analysis, the echocardiographic characteristics of non-CM masses were similar to those of atypical CM. The main MSCT feature of non-CM masses was mixed attenuation, which was different from that of typical CM (P＜0.05). The non-CM masses were iso-hyperintense on T1-weighted images and hyperintense on T2-weighted images, and showed enhancement at about one-half of cases on first-pass perfusion imaging and late gadolinium enhancement imaging, which did not show significant difference from those of typical CM (P＞0.05 for all).

Based on the multimodality imaging features of CM, echocardiography should be used to discriminate typical CM from atypical CM and further cardiac CT and CMR are used to distinguish between malignant and benign tumors and to assess histological types of cardiac masses considered as atypical CM, which can improve the diagnostic efficiency of CM, reduce the misdiagnosis rate, and contribute to preoperative planning.