Patient and study site characteristics

In the current dose survey, a total of 487 patients were enrolled from 17 hospitals, with a median age of 71 (IQR: 63–75) years and 301 (62%) being men. The median patient height was 162 (IQR: 155–168) cm, weight 61 (IQR: 55–65) kg, and HR at the scan was 59 (IQR: 54–65) beats per minute (bpm). Sinus rhythm was present in 445 (91%) of the patients. In comparison, the previous dose survey included 442 eligible patients from 17 hospitals, with a median age of 70 (IQR: 62–75) years and 269 (61%) being men. The median patient height was 162 (IQR: 156–168) cm, weight 61 (IQR: 55–66) kg, with HR 60 (IQR: 54–66) bpm and sinus rhythm present in 413 (93%) of the patients. Nitroglycerin was administered in 98% and 99% of cases in 2021 and 2023, respectively (p = 0.28). β-blockers were used in 63% of patients in 2021 and 61% in 2023 (p = 0.46) (Table 1).

Table 1 Patient characteristics

In the current dose survey, 252 (52%) of scans were performed using area-detector CT scanners, 118 (24%) using dual-source CT scanners with EID, and 8 (2%) using dual-source CT scanners with PCD. In contrast, in the previous dose survey, 187 (42%) of scans were performed using area-detector CT scanners and 108 (25%) using dual-source CT scanners with EID. The characteristics of the CT scanners used in the dose survey 2021 and 2023 are listed in Supplementary Table 1. AEC were used in almost all scans for both the current and previous dose surveys. Regarding image reconstruction techniques, no cases used filtered back projection. All scans were reconstructed using either conventional IR or DLIR. In the 2023 survey, IR and DLIR were used in 346 (71%) and 141 (29%) cases, respectively, whereas in the 2021 survey, the numbers were 366 (83%) and 76 (17%), respectively. Regarding scan techniques, ECG-gated retrospective helical scanning accounted for 18% of all scans in the current survey, down from 46% in the previous survey. In contrast, the use of prospective ECG-triggered scanning increased from 47 to 68%. Low tube potential scanning was employed in 67% of the scans in the current survey, whereas it was used in 33% of scans in the previous survey (Table 2, Fig. 3).

Table 2 Imaging characteristicsFig. 3

Scan techniques and tube potentials used in the previous and current dose surveys. A Proportions of scan techniques in the 2021 and 2023 surveys. B Proportions of tube potentials in the 2021 and 2023 surveys. Abbreviations: CTA/CFA, Computed Tomography Angiography/Cardiac Function Analysis

Association between heart rate distribution and trends in scan techniques

In the previous survey, the HR during scanning was below 60 bpm in 212 patients, between 60 and 64 bpm in 109 patients, and 65 bpm or above in 121 patients. In the current survey, 256 patients had HR below 60 bpm, 116 had HR between 60 and 64 bpm, and 115 had HR 65 bpm or above. The proportion of patients with HR below 65 bpm was 73% in the previous survey, compared to 76% in the current survey (p = 0.31).

Regarding scan techniques, in the previous survey, approximately 50% of the scans were performed by using ECG-gated retrospective scanning, regardless of HR (43% for HR below 60 bpm, 50% for HR 60–64 bpm, and 47% for HR 65 bpm or above). In the current survey, the use of ECG-gated retrospective scanning significantly decreased, particularly in patients with lower HR: 9% for HR below 60 bpm, 21% for HR 60–64 bpm, and 35% for HR 65 bpm or above (Fig. 4).

Fig. 4

Heart rate during scanning and scan techniques by heart rate in the previous and current dose surveys. A Proportions of patients classified by heart rate categories (< 60 bpm, 61–64 bpm, and ≥ 65 bpm) in the 2021 and 2023 surveys. B Proportions of scan techniques used within each heart rate category in the 2021 and 2023 surveys. Abbreviations: bpm, beats per minute; CTA/CFA, Computed Tomography Angiography/Cardiac Function Analysis

Image quality of coronary CT angiography

The image quality scores of CCTA were a median of 3.5 (IQR: 3–4) in the previous survey and a median of 3.75 (IQR: 3.25–4) in the current survey, respectively (p < 0.01). In the current dose survey, according to HR categories, the median scores were 3.75 (IQR: 3.5–4) for HR below 60 bpm, 3.75 (IQR: 3.25–4) for HR 60–64 bpm, and 3.5 (IQR: 3–4) for HR 65 bpm or above (p < 0.01). When categorized by tube potential, the image quality scores were 3.5 (IQR: 3–4) for 120 kVp or above and 3.75 (IQR: 3.5–4) for 100 kVp or below (p < 0.01), respectively. In the current survey, image quality was also significantly higher in the DLIR group than in the IR group, with median scores of 4 (IQR: 3.5–4) and 3.75 (IQR: 3–4), respectively (p < 0.01).

Review of cardiac CT examination protocols of each institution

Sixteen institutions provided valid data for both secondary surveys, having not been excluded based on the criterion of submitting fewer than 20 cases. Among them, two institutions had changed equipment since the previous survey. Cardiac CT protocols were modified 14 (88%) of the institutions. Six institutions adjusted the HR threshold for administering oral or intravenous β-blockers, six decreased tube potential for non-obese patients, and six modified the scanning mode dependent on HR and rhythm. Additionally, four hospitals changed the scanning range, and one adjusted the image noise settings. No complaints, including degradation of image quality or workflow issues, were reported following the protocol modification. However, four institutions continued to use standard tube potential imaging at 120 kVp for over 80% of cases. The reasons for employing 120 kVp included concerns about image degradation at lower tube voltages due to relatively older CT systems in three hospitals, and the need to reduce noise associated with the routine use of sharp kernels which is preferred by the attending radiologist in one.

Total DLP from cardiac CT and CTDIvol of coronary CT angiography

The DLP and CTDIvol for the CCTA portion at each hospital are shown in Fig. 5 and Fig. 6, respectively. In the present study, the median DLP and CTDIvol for CCTA were 266 (176–395) mGy·cm and 19 (12–31) mGy, respectively. Both values were significantly lower than those in the previous survey, where the median DLP was 494 (390–764) mGy·cm and the CTDIvol was 33 (25–48) mGy (p < 0.01 for both), approaching the levels reported in the PROTECTION VI study. Supplementary Table 2 summarizes the DLP and CTDIvol for CCTA in both the previous and current surveys, along with the annual number of cardiac CT examinations performed and the number of hospital beds at each participating facility.

Fig. 5

Dose-length product for coronary computed tomography angiography across sites in the previous and current surveys. Left: Box plots showing the DLP for all CCTA examinations. Right: Variability of median DLP (± IQR) among sites. Reference lines: red, Japan DRLs 2020 (1300 mGy·cm); orange, previous survey (764 mGy·cm); blue, current survey (395 mGy·cm); green, PROTECTION VI study (338 mGy·cm). Abbreviations: DLP, dose-length product; CT, computed tomography; DRL, diagnostic reference level

Fig. 6

Computed tomography dose index volume for coronary computed tomography angiography across sites in the previous and current surveys. Left: Box plots showing the CTDIvol for all CCTA examinations. Right: Variability of median CTDIvol (± IQR) among sites. Reference lines: red, Japan DRLs 2020 (66 mGy); orange, previous survey (48 mGy); blue, current survey (31 mGy); green, PROTECTION VI study (24 mGy). Abbreviations: CTDIvol, computed tomography dose index volume; CT, computed tomography; DRL, diagnostic reference level

The relationship between scan techniques and CTDIvol for CCTA in the 2021 and 2023 surveys is shown in supplementary Fig. 3. The CTDIvol for CCTA using ECG-triggered prospective scanning, CTA/CFA mode, and ECG-gated retrospective scanning were 19.7 mGy (10.8–34.1), 30.7 mGy (22.4–37.7), and 49.4 mGy (37.6–63.4) in 2021, and 14.4 mGy (9.2–22.5), 27.3 mGy (10.8–36.5), and 39.5 mGy (30.2–45.6) in 2023, respectively.