Photon-counting computed tomography for paediatric congenital heart defects yields images of high diagnostic quality with low radiation doses at both 70 kV and 90 kV

Computed tomography (CT) is used in paediatric patients with congenital heart defects for diagnosis, preoperative planning, and postoperative evaluation [1‐3]. CT is widely available and fast. It can often be used without sedation or with feed-and-sleep methods, with the benefit of avoiding general anaesthesia. It also has high spatial resolution and enables simultaneous assessment of extracardiac structures such as the pulmonary parenchyma, pleura, and skeleton. The relative disadvantage is that the patient is exposed to radiation and receives an iodine contrast agent with potential risk for renal function impairment (although this risk is low when using iso-osmolar contrast agents) [4, 5].

Photon-counting CT (PCCT) was recently made available for clinical use. It yields images with negligible electronic noise and improved contrast between soft tissue and iodinated contrast agents [6]. However, this means that the images look different from conventional CT images, further dependent on the chosen reconstruction algorithm. This new CT method may potentially impact diagnostic accuracy because the observer’s experience is different.

PCCT provides higher spatial resolution than conventional CT, as well as intrinsic spectral information in every scan [7]. Thus, image quality may be improved with a lower radiation dose. A study on small children that used PCCT at 90 kV for diagnosis of congenital heart defects achieved higher image quality than conventional CT but yet reported a similar effective dose (E_eff) [8].

It is unknown if the essentially noise-free PCCT images impact observers impression of diagnostic quality, and it is unknown which PCCT radiation dose is adequate for children with congenital heart defects. It also remains to be investigated to what extent radiation dose is reduced when moving from 90 kV to 70 kV using PCCT.

The aim of this study was therefore to compare the diagnostic quality and radiation dose in paediatric non-gated cardiovascular PCCT for diagnosis of congenital heart defects at 70 kV and 90 kV.

The regional ethics committee approved this retrospective study using pseudonymised data from patients referred for a clinical examination, waiving the need for individual consent. Paediatric non-gated cardiovascular PCCT (Naeotom Alpha, Siemens Healthineers, Erlangen, Germany) examinations with intravenous contrast agent completed between 30th September 2021 and 1st March 2023 in children with suspected or confirmed congenital heart defects were eligible. Initially, only 90 kV protocols were available as this was an early installation. After 70 kV was made available by the vendor, all clinical examinations were performed at 70 kV, which also limited the number of 90 kV examinations in the current study. The aim was to include the same number of examinations using 70 kV as had been completed using 90 kV.

Figure 1 shows the inclusion and exclusion flow chart. Table 2 shows patient characteristics including distribution of cardiovascular malformations. Although there were differences in the distribution of cardiovascular defects between the 70 kV and 90 kV groups, there was no difference between groups for age (P > 0.99), height (P = 0.86), or weight (P = 0.83). However, scan length differed (P = 0.002) between 70 kV and 90 kV groups, and therefore comparisons were made between radiation doses for 70 kV and 90 kV both using all available data and corrected for different scan lengths by matching. As expected, there was a numerical change in the P-value for DLP when adapting for scan length, but not for CTDI_vol, SSDE, or E_eff.

Table 3 shows the quality scores for all observers, with a median of 4 (high diagnostic quality) for all observers for both 70 kV and 90 kV. However, all observers scored some 90 kV cases below 3, whereas for 70 kV there were no scores below 3. Sum-of-ranks was different between 70 kV and 90 kV for observer 2 for overall diagnostic scoring (median 4 vs 4, sum-of-ranks 1,032 vs 1,454; P = 0.001; all other P-values > 0.34), and for observer 3 for specific diagnostic scoring (median 4 vs 4, sum-of-ranks 1,129 vs 1,357; P = 0.03; all other P-values > 0.06).

The average percent rater agreement for 70 kV was 79% (range 66–94%) for overall scores, and 89% (range 80–97%) for specific scores. For 90 kV, the average percent rater agreement for overall scores was 72% (range 60–77%) and for specific scores 86% (range 83–89%).

The intraclass correlation for 70 kV was 0.58 (95% confidence interval [CI] 0.26–0.75) for overall scores, and 0.71 (95% CI 0.51–0.84) for specific scores. For 90 kV, intraclass correlation for overall scores was 0.84 (95% CI 0.72–0.91) and for specific scores 0.85 (95% CI 0.75–0.92).

Figures 2 and 3 show clinical examples of 70 kV and 90 kV examinations. Maximum intensity projections or 0.6 mm images are presented to more clearly show relevant anatomical structures.

Radiation dose parameters are shown in Table 4, and radiation dose parameters versus age and body surface area in Fig. 4 for CTDI_vol, in Fig. 5 for SSDE, and in Fig. 6 for E_eff.

Jonckheere-Terpstra’s tests showed that there was no trend between E_eff and overall scores (70 kV: P = 0.55; 90 kV: P = 0.98) or between E_eff and specific scores (70 kV: P = 0.55; 90 kV: P = 0.29).

This study shows that PCCT yields images of high diagnostic quality for paediatric congenital heart defect examinations with a concomitant low radiation dose. Further, vessels were well-defined with a diluted contrast agent dose at 70 kV, with scoring equal to or higher than the full contrast dose used for 90 kV.

Paediatric cardiovascular PCCT therefore has the potential to become the primary complementary modality to echocardiography in congenital heart defects due to its high diagnostic quality, low radiation dose, and low contrast agent dose. This is particularly important in complex congenital heart defects where repeat examinations are needed in small radiosensitive children.

Although magnetic resonance imaging has the benefit of providing both functional and anatomical information, it has limited spatial resolution. CT will thus remain the gold standard for anatomical assessment of small vessels in conditions such as congenital heart defects with pulmonary atresia, pulmonary vein stenosis, collaterals, or coronary artery anomalies, where precise delineation of anatomy is central to planning and guiding surgical or interventional treatment.

All cases were of high diagnostic quality and could answer the clinical questions. Observers gave scores of 3 and 4 to the 70 kV images, but scores from 1 to 4 for the 90 kV images. This indicates that the essentially noise-free images at 70 kV do not impact observers negatively. Also, agreement between observers was generally high. In regard to image quality, this is similar to a 90 kV PCCT study showing a 97% success rate; however, that study only showed moderate or poor interreader agreement [8].

The current study showed a median E_eff of 0.39 mSv at 90 kV for all patients between birth and 17 years. For those below 1 year of age, median E_eff was 0.30 mSv. Both values are lower than those in the only previous comparable PCCT study, where mean E_eff was 0.50 mSv (± 0.23 mSv) at 90 kV, even though the children in that study (66 [10–161] days) were much younger than in the current study [8]. Further, Dirrichs et al. [8] showed a mean DLP at 90 kV of 14.3 (± 6.6; range 3–42) mGy × cm for their very young children, compared with the current study’s mean DLP of 4 (range 3–8) mGy × cm. It is unclear why the previous study showed such high radiation exposure, but possible explanations include larger scan length or the use of a lead apron which may increase radiation dose unless automated tube current modulation is turned off [13].

For 70 kV, the current study showed a median E_eff of 0.30 mSv with all ages up to 17 years included, and for those below 1 year of age, the median E_eff was 0.22 mSv. In comparison, a recent 70 kV dual-source conventional CT study that only included young children (3.5 [0.2–6.6] months) showed a median E_eff of 0.20 mSv [14].

Different CT methods, dose calculations, and use of weighting factors make direct comparison between studies challenging. There is nevertheless a trend towards reduced radiation doses for non-gated cardiovascular CT, where the current PCCT study shows a further dose reduction with E_eff of 0.15–0.82 mSv for 70 kV and 0.22–1.51 mSv for 90 kV in children between 0 and 17 years. This is lower than previous conventional non-gated CT studies in small children between 0 years and 4.5 years, where E_eff was between 0.20 and 1.95 mSv [15‐17].

Both CTDI_vol and DLP are commonly used to estimate radiation exposure. They are however relatively blunt measures—especially DLP—when comparing radiation doses, which is supported by the current results. For comparisons between studies, SSDE may be preferred as a reliable tool to estimate average radiation dose depending both on CT parameters and size of the specific patient, even though it does not include organ or tissue weighting factors [18]. So far, relatively few studies have presented SSDE, and none of them have used PCCT for congenital heart defects. The current study, by presenting SSDE, provides an opportunity for future comparisons.

This study has limitations. The PCCT software was updated during the study. However, image scores did not change with time or stepwise after software updates, and so the impact of the updates is considered small. Even though diagnostic quality was assessed it would be appropriate for clinical impact to study anatomy by PCCT, and compare it to anatomy during surgery. Surgery was not performed in all patients and there is no other examination more accurate than PCCT available for comparison. Therefore, neither accuracy nor potentially missed diagnoses were assessed. It is however unlikely that accuracy is decreased using PCCT compared to conventional CT either for congenital heart defects as studied here, or for other possible associated cardiovascular abnormalities. Finally, there is no international consensus regarding which QIR level to use for congenital heart defects, but QIR 4 in its present form as used in the current study is now preferred in several centres.

Paediatric cardiovascular PCCT yields images for congenital heart defects of high diagnostic quality with low radiation dose at both 70 kV and 90 kV.