Impact of deep learning image reconstructions (DLIR) on coronary artery calcium quantification

Influence of deep learning-based super-resolution reconstruction on Agatston score

OBJECTIVE

To evaluate the impact of deep learning-based super-resolution reconstruction (DLSRR) on image quality and Agatston score.

METHODS

Consecutive patients who underwent cardiac CT, including unenhanced CT for Agatston scoring, were enrolled. Four types of non-contrast CT images were reconstructed using filtered back projection (FBP) and three strengths of DLSRR. Image quality was assessed by measuring image noise, signal-to-noise ratio (SNR) of the aorta, contrast-to-noise ratio (CNR), and edge rise slope (ERS) of coronary artery calcium (CAC). Agatston score and CAC volume were also measured. These results were compared among the four CT datasets. Patients were categorized into four risk levels based on the Coronary Artery Calcium Data and Reporting System (CAC-DRS), and the concordance rate between FBP and DLSRR classifications was evaluated.

RESULTS

For the 111 patients enrolled, DLSRR significantly reduced image noise (p < 0.001) and improved SNR and CNR (p < 0.001), with stronger effects at higher DLSRR strengths (p < 0.01). ERS was significantly enhanced using DLSRR compared with FBP (p < 0.001), whereas there was no significant difference among the three strengths of DLSRR (p = 0.90-0.98). Agatston score and CAC volume were not significantly affected by DLSRR (p = 0.952 and 0.901, respectively). The concordance rate of CAC-DRS classification between FBP and DLSRR was 93%.

CONCLUSION

DLSRR significantly improves image quality by reducing noise and enhancing sharpness without significantly altering Agatston scores or CAC volumes. The concordance rate of CAC-DRS classification with FBP was high, although some reclassifications were observed.

KEY POINTS

Question The utility of deep learning-based super-resolution reconstruction (DLSRR) in coronary CT angiography is well known, but its impact on the Agatston score remains unclear. Findings DLSRR significantly improved image quality without altering the Agatston scores, but some reclassifications of Coronary Artery Calcium Data and Reporting System (CAC-DRS) were observed. Clinical relevance DLSRR should be cautiously used in clinical settings owing to the occurrence of some cases of CAC-DRS reclassification.

Effect of temporal resolution on calcium scoring: insights from photon-counting detector CT

To intra-individually investigate the variation of coronary artery calcium (CAC), aortic valve calcium (AVC), and mitral annular calcium (MAC) scores and the presence of blur artifacts as a function of temporal resolution in patients undergoing non-contrast cardiac CT on a dual-source photon counting detector (PCD) CT. This retrospective, IRB-approved study included 70 patients (30 women, 40 men, mean age 78 ± 9 years) who underwent ECG-gated cardiac non-contrast CT with PCD-CT (gantry rotation time 0.25 s) prior to transcatheter aortic valve replacement. Each scan was reconstructed at a temporal resolution of 66 ms using the dual-source information and at 125 ms using the single-source information. Average heart rate and heart rate variability were calculated from the recorded ECG. CAC, AVC, and MAC were quantified according to the Agatston method on images with both temporal resolutions. Two readers assessed blur artifacts using a 4-point visual grading scale. The influence of average heart rate and heart rate variability on calcium quantification and blur artifacts of the respective structures were analyzed by linear regression analysis. Mean heart rate and heart rate variability during data acquisition were 76 ± 17 beats per minute (bpm) and 4 ± 6 bpm, respectively. CAC scores were smaller on 66 ms (median, 511; interquartile range, 220-978) than on 125 ms reconstructions (538; 203-1050, p < 0.001). Median AVC scores [2809 (2009-3952) versus 3177 (2158-4273)] and median MAC scores [226 (0-1284) versus 251 (0-1574)] were also significantly smaller on 66ms than on 125ms reconstructions (p < 0.001). Reclassification of CAC and AVC risk categories occurred in 4% and 11% of cases, respectively, whereby the risk category was always overestimated on 125ms reconstructions. Image blur artifacts were significantly less on 66ms as opposed to 125 ms reconstructions (p < 0.001). Intra-individual analyses indicate that temporal resolution significantly impacts on calcium scoring with cardiac CT, with CAC, MAC, and AVC being overestimated at lower temporal resolution because of increased motion artifacts eventually leading to an overestimation of patient risk.

Ultralow dose coronary calcium scoring CT at reduced tube voltage and current by using deep learning image reconstruction

OBJECTIVE

To explore the potential of the deep learning reconstruction (DLR) for ultralow dose calcium scoring CT (CSCT) with simultaneously reduced tube voltage and current.

METHODS

In this prospective study, seventy-five patients (group A) undergoing routine dose CSCT (120kVp/30mAs) were followed by a low dose (120kVp/20mAs) scan and another 81 (group B) were followed by an ultralow dose (80kVp/20mAs) scan. The hybrid iterative reconstruction was used for the routine dose data while the DLR for data of reduced dose. The calcium score and risk categorization were compared, where the correlation was evaluated using the intraclass correlation coefficient (ICC). The noise suppression performance of DLR was characterized by the contrast-to-noise ratio (CNR) between coronary arteries and pericoronary fat.

RESULTS

The effective dose was 0.32 ± 0.03 vs. 0.48 ± 0.05 mSv for the two scans in group A and 0.09 ± 0.01 vs. 0.49 ± 0.05 mSv in group B. No significant difference was found on CACSs within either group (A: p = 0.10, ICC=0.99; B: p = 0.14, ICC=0.99), nor was it different on risk categorization (A: p = 0.32, ICC=0.99; B: p = 0.16, ICC=0.99). The DLR images exhibited higher CNR in both groups (A: p < 0.001; B: p = 0.001).

CONCLUSIONS

The DLR allowed reliable calcium scoring in not only low dose CSCT with reduced tube current but ultralow dose CSCT with simultaneously reduced tube voltage and current, showing feasibility to be adopted in routine applications.

Artificial Intelligence in Coronary Artery Calcium Scoring

Cardiovascular disease (CVD), particularly coronary heart disease (CHD), is the leading cause of death in the US, with a high economic impact. Coronary artery calcium (CAC) is a known marker for CHD and a useful tool for estimating the risk of atherosclerotic cardiovascular disease (ASCVD). Although CACS is recommended for informing the decision to initiate statin therapy, the current standard requires a dedicated CT protocol, which is time-intensive and contributes to radiation exposure. Non-dedicated CT protocols can be taken advantage of to visualize calcium and reduce overall cost and radiation exposure; however, they mainly provide visual estimates of coronary calcium and have disadvantages such as motion artifacts. Artificial intelligence is a growing field involving software that independently performs human-level tasks, and is well suited for improving CACS efficiency and repurposing non-dedicated CT for calcium scoring. We present a review of the current studies on automated CACS across various CT protocols and discuss consideration points in clinical application and some barriers to implementation.

Influence of Deep Learning Based Image Reconstruction on Quantitative Results of Coronary Artery Calcium Scoring

RATIONALE AND OBJECTIVES

To assess the impact of deep learning-based imaging reconstruction (DLIR) on quantitative results of coronary artery calcium scoring (CACS) and to evaluate the potential of DLIR for radiation dose reduction in CACS.

METHODS

For a retrospective cohort of 100 consecutive patients (mean age 62 ±10 years, 40% female), CACS scans were reconstructed with filtered back projection (FBP), adaptive statistical iterative reconstruction (ASiR-V in 30%, 60% and 90% strength) and DLIR in low, medium and high strength. CACS was quantified semi-automatically and compared between image reconstructions. In a phantom study, a cardiac calcification insert was scanned inside an anthropomorphic thorax phantom at standard dose, 50% dose and 25% dose. FBP reconstructions at standard dose served as the reference standard.

RESULTS

In the patient study, DLIR led to a mean underestimation of Agatston score by 3.5, 6.4 and 11.6 points at low, medium and high strength, respectively. This underestimation of Agatston score was less pronounced for DLIR than for ASiR-V. In the phantom study, quantitative CACS results increased with reduced radiation dose and decreased with increasing strength of DLIR. Medium strength DLIR reconstruction at 50% dose reduction and high strength DLIR reconstruction at 75% dose reduction resulted in quantitative CACS results that were comparable to FBP reconstructions at standard dose.

CONCLUSION

Compared to FBP as the historical reference standard, DLIR leads to an underestimation of CACS but this underestimation is more moderate than with ASiR-V. DLIR can offset the increase in image noise and calcium score at reduced dose and may thus allow for substantial radiation dose reductions in CACS studies.

Artificial intelligence in coronary artery calcium score: rationale, different approaches, and outcomes

Almost 35 years after its introduction, coronary artery calcium score (CACS) not only survived technological advances but became one of the cornerstones of contemporary cardiovascular imaging. Its simplicity and quantitative nature established it as one of the most robust approaches for atherosclerotic cardiovascular disease risk stratification in primary prevention and a powerful tool to guide therapeutic choices. Groundbreaking advances in computational models and computer power translated into a surge of artificial intelligence (AI)-based approaches directly or indirectly linked to CACS analysis. This review aims to provide essential knowledge on the AI-based techniques currently applied to CACS, setting the stage for a holistic analysis of the use of these techniques in coronary artery calcium imaging. While the focus of the review will be detailing the evidence, strengths, and limitations of end-to-end CACS algorithms in electrocardiography-gated and non-gated scans, the current role of deep-learning image reconstructions, segmentation techniques, and combined applications such as simultaneous coronary artery calcium and pulmonary nodule segmentation, will also be discussed.

Deep learning algorithm applied to plain CT images to identify superior mesenteric artery abnormalities

OBJECTIVES

Atypical presentations, lack of biomarkers, and low sensitivity of plain CT can delay the diagnosis of superior mesenteric artery (SMA) abnormalities, resulting in poor clinical outcomes. Our study aims to develop a deep learning (DL) model for detecting SMA abnormalities in plain CT and evaluate its performance in comparison with a clinical model and radiologist assessment.

MATERIALS AND METHODS

A total of 1048 patients comprised the internal (474 patients with SMA abnormalities, 474 controls) and external testing (50 patients with SMA abnormalities, 50 controls) cohorts. The internal cohort was divided into the training cohort (n = 776), validation cohort (n = 86), and internal testing cohort (n = 86). A total of 5 You Only Look Once version 8 (YOLOv8)-based DL submodels were developed, and the performance of the optimal submodel was compared with that of a clinical model and of experienced radiologists.

RESULTS

Of the submodels, YOLOv8x had the best performance. The area under the curve (AUC) of the YOLOv8x submodel was higher than that of the clinical model (internal test set: 0.990 vs 0.878, P =.002; external test set: 0.967 vs 0.912, P =.140) and that of all radiologists (P <.001). The YOLOv8x submodel, when compared with radiologist assessment, demonstrated higher sensitivity (internal test set: 100.0 % vs 70.7 %, P =.002; external test set: 96.0 % vs 68.8 %, P <.001) and specificity (internal test set: 90.7 % vs 66.0 %, P =.025; external test set: = 88.0 % vs 66.0 %, P <.001).

CONCLUSION

Using plain CT images, YOLOv8x was able to efficiently identify cases of SMA abnormalities. This could potentially improve early diagnosis accuracy and thus improve clinical outcomes.