Study Participants.
Between August 2017 and February 2021, 34 consecutive patients with CKD were referred for cardiac MRI examinations to identify cardiac function and myocardial tissue changes. They were 25 men and 9 women ranging inage from 31 to 89 years (mean age, 67.1 years). There were 18 patients (52.9%) with hypertension and 14 (41.2%) with diabetes mellitus. Dialysis was given to 5 of 34 patients (14.7%). In addition, 8 (23.5%), 3 (8.8%), and 1 (2.9%) patient had heart failure, coronary artery diseases, and cardiopulmonary arrest, respectively. For comparison, 10 controls (5 men and 5 women; mean age, 41.1 years; age range, 21−78 years) were enrolled: their electrocardiogram or echocardiography was borderline positive, but cardiac MRI revealed no functional abnormalities and their myocardial T1 and T2 values were within the normal range14. LGE MRI showed no myocardial scar in the controls, while LGE MRI was not performed in any CKD patients because of their renal functional impairment. This retrospective, cross-sectional study was approved by our institutional review board, and informed consent was given by all the patients and controls. All methods were performed in accordance with the relevant guidelines and regulation.
Magnetic Resonance Imaging.
All MRI examinations were performed using a 1.5 T system (Ingenia, Philips Healthcare, Best, the Netherlands). Breath-hold black-blood T2-weighted STIR was performed with the typical parameters as follows: repetition time (TR), 2 RR; effective echo time (TE), 80 ms; field of view (FOV), 320 x 320 mm2, imaging matrix, 220 x 138; and slice thickness, 10 mm. Cine SSFP imaging was acquired with the imaging parameters as follows: TR, 3.2 ms; TE, 1.6 ms; FOV, 350 x 350 mm2, imaging matrix, 192 x 170; and slice thickness, 8 mm. T1 mapping was performed with a 5s (3s) 3s modified Look–Locker inversion recovery sequence acquired with a single-shot SSFP readout27. Therefore, 7 to 11 multicontrast images were acquired during 11-s breath-holding in short-axis midventricular slices. The imaging parameters for T1 mapping were as follows: TR, 2.7 ms; TE, 1.3 ms; flip angle, 35°; FOV, 300 ´ 300 mm2; imaging matrix, 152 ´ 150; and slice thickness,10 mm. The delay time was 159.5 ms with an inversion time increment of 350 ms. T2 mapping imaging was performed with multiecho gradient- and spin-echo (GRASE) sequences28. The imaging parameters for GRASE were as follows: TR, 1 RR; effective TE, 7.7–69.2 ms with 9 echoes and an echo train length of 5; flip angle, 35°; FOV, 300 ´ 300 mm2; imaging matrix, 152 ´ 145; and slice thickness, 10 mm.
Image Analysis.
Color-coded T1 and T2 maps were automatically generated on the operator console after T1T2 mapping scanning, and a team of three MR-dedicated radiological technologists measured T1T2 values of the septal myocardium according to the previous study10,29. Texture analysis was performed by a radiologist with 26 years of experience in cardiac MRI using the open-access software (MaZda; Institute of Electronics, Technical University of Lodz, Poland)23,27,30. The intensity range of the image under texture analysis was normalized by the default method: the intensity changes of the image ranged from 1 to 256 and 6 bits per pixel were used for image quantization. The entire left ventricular myocardium was enclosed manually, and a texture parameter calculation was conducted for T2-weighted STIR and cine images. The feature reduction methods, including Fisher coefficients, classification error probability, average correlation coefficients, and mutual information measure, were used to select the most discriminating features between the patients with CKD and controls.
Statistical Analysis.
The differences in the T1T2 values and texture features selected by the feature reduction methods were assessed between the patients with CKD and controls. An unpaired t test was used to evaluate the differences, and P < 0.05 was defined as significant. ROC analysis was used to determine the cutoff value for each analysis. The AUC was compared between the analyses and P < 0.05 was defined as significant.
Ethical approval.
Voluntary informed consent was obtained from all subjects prior to their participation. The protocol was approved by the Nihon University Hospital Institutional Review Board.