Comparative analysis of third-generation dual-energy CT and IVUS for in-stent restenosis detection

Abstract

Purpose

Prior studies have assessed in-stent diameter restenosis (ISDR) in coronary arteries using 64-slice multidetector computed tomography coronary angiography (MDCT-CA) compared to invasive coronary angiography (ICA), which is the gold standard. This study aimed to compare the diagnostic accuracy of monoenergetic reconstruction using third-generation dual-source dual-energy CT (DSDECT) to that of ICA reconstruction via adjunctive intravascular ultrasonography (IVUS) for evaluating the ISDR.

Methods

A total of 95 patients with previously stented coronary arteries (involving 110 stents) underwent DSDECT followed by ICA and IVUS within a 24-h timeframe. The specificities, sensitivities, negative predictive values (NPVs), and positive predictive values (PPVs) of the DSDECT and ICA were compared for confirming or excluding the ISDR using in-stent area restenosis (ISAR) and a minimal luminal area (MLA) ≤ 4.0 mm² on IVUS as the reference standard.

Results

Compared with IVUS, the latest DSDECT demonstrated good sensitivity (100%), specificity (92.4%), and accuracy (96.1%) in detecting the ISDR. Our study highlights a limitation in assessability for stents with diameters < 3 mm, emphasizing the importance of careful patient selection. When employing an IVUS MLA of 4.0 mm² as a reference for identifying the ISDR, no significant difference was observed between DSDECT and ICA in the identification of the ISDR. However, it is important to note that the use of absolute cut-offs, such as < 6.0 mm² in the left main or < 4.0 mm², may not universally apply across varying ethnicities and between sexes. The interpretation of the minimal luminal area (MLA) should be considered in the context of individual patient characteristics, and caution is advised to avoid potential misleading conclusions based solely on absolute thresholds.

Conclusion

In summary, when assessing stent patency, the latest DSDECT exhibits similar performance to coronary angiography and IVUS. Moreover, it offers noninvasiveness, cost-effectiveness, and ease of operation, which are advantageous characteristics. However, it is essential to consider limitations in patient eligibility, including factors such as prior cardiac devices, arrhythmias, and any degree of chronic renal insufficiency, which may impact CT imaging analysis. The 100% negative predictive value (NPV) of third-generation DSDECT reliably excludes in-stent restenosis (ISDR), potentially obviating invasive angiography in stable patients with patent stents.

Trial registration

ZU-IRB#3915/13-8-2017 Registered 13 August 2017, email: IRB_123@medicine.zu.edu.eg.

Introduction

For evaluation of suspected in-stent restenosis, current ESC Guidelines (Class I, Level A) recommend CT angiography as the preferred initial test for stable patients with low-to-intermediate probability of ISR, while reserving ICA for high-risk presentations [1]. While conventional invasive coronary angiography (ICA) is a widely used method for detecting in-stent restenosis (ISR), the financial costs and potential complications associated with ICA prompt the exploration of noninvasive alternatives for ISR assessment. However, it is essential to note that the choice between modalities often depends on factors such as pretest probability and the specific clinical context. There are many advocates for coronary CT, especially in lower-risk patients, considering the limitations and artifacts associated with CT evaluation of coronary disease [2].

Compared with ICA, multidetector computed tomography coronary angiography (MDCT-CA) is a promising, noninvasive, cost-effective, and less complex diagnostic approach for evaluating stent patency. Nonetheless, limitations associated with 64-slice MDCT-CA in ISR assessment are apparent, possibly attributed to partial volume effects and beam-hardening artifacts, particularly in stents with diameters of 3 mm or less [3–13]. Furthermore, the ability of the 64-slice MDCT-CA, unlike its predecessors, to provide insights into coronary plaque morphology and clear ISR determination, akin to intravascular ultrasound (IVUS), is noteworthy [14–16].

The advent of advanced image-based monoenergetic reconstructions utilizing third-generation DSDECT offers enhanced visualization of stents and, consequently, improved ISR detection [17].

Hence, the objective of this study was to evaluate the diagnostic accuracy of monoenergetic reconstructions using third-generation DSDECT compared to ICA with adjunctive IVUS for detecting coronary ISR.

Material and methods

Between January 2018-January 2019, 163 consecutive patients with suspected ISR were screened. After exclusions (Fig. 1a), 95 patients (110 stents) met all criteria. Six stents (from 5 patients) were excluded post-enrollment due to motion artifacts (n = 4) or severe calcification (n = 2) precluding analysis.This comparative cross-sectional study was conducted within the cardiology department of our hospital from January 2018 to January 2019. The study protocol received approval from our Institutional Review Board (IRB), Faculty of Medicine, Zagazig University (ZU-IRB#3915/13–8–2017), ensuring adherence to the ethical principles outlined in the 1975 Declaration of Helsinki. All methods were performed in compliance with the institution's human research guidelines, as endorsed by prior approval. Informed written consent was obtained from all participating individuals. The study cohort consisted of 95 consecutive patients who met the inclusion criteria, presented with suspected in-stent restenosis (ISR) and were referred for coronary angiography.

Fig. 1.

a Patient screening and enrollment flowchart. Of 163 consecutive patients with suspected in-stent restenosis, 68 were excluded (detailed in boxes). The final cohort included 95 patients (110 stents), with six stents excluded post-hoc due to non-diagnostic image quality. CKD = chronic kidney disease. b The 5-point grading scale for dual-source dual-energy CT image quality, with scores 1 through 3 (ranging from poor but technically assessable to good quality) considered diagnostic and included in analysis, whereas scores 4 (insufficient) and 5 (non-assessable) were excluded due to non-diagnostic quality. This classification ensured only images meeting minimum diagnostic thresholds were analyzed while maintaining transparency regarding borderline cases. c Distribution of stent diameter

Eligibility criteria

Inclusion criteria

All patients exhibiting suspected in-stent restenosis (ISR), determined by symptoms such as recurrent chest pain, ECHO-detected wall motion abnormalities, or positive stress ECG in individuals with a history of previous percutaneous coronary intervention (PCI), were considered for inclusion. Each participant underwent examination using monoenergetic reconstructions with third-generation DSDECT (SOMATOM Force Dual Source; Siemens Medical Solutions, Forchheim, Germany). Retrospective electrocardiography (ECG)-gated contrast-enhanced DSDECT was conducted within 24 h before invasive coronary angiography (ICA). After retrospective review, five STEMI cases were excluded from analysis as they underwent direct invasive coronary angiography without CT evaluation, per current guidelines. The final study cohort comprised 95 consecutive patients with stable presentations (stable/unstable angina or NSTEMI) who underwent both DSDECT and ICA/IVUS. Patients included in the study were confirmed to be in sinus rhythm, clinically stable, and had undergone intravascular ultrasound (IVUS) as part of the catheterization procedure, utilizing a 2.9 F, 40-MHz, single-element mechanical transducer (Boston Scientific, Natick, MA, USA). Risk factors were diagnosed in accordance with the criteria outlined by the World Health Organization [18].

Exclusion criteria

Patients meeting any of the following criteria were excluded from the study:

• The common elements associated with poor image quality led to the exclusion of some stents. Factors such as smaller stent diameters (< 3 mm), thicker stent struts, higher heart rates, and cardiac arrhythmias were identified as contributors to poor image quality.

• Allergy to contrast medium

• Implantation of any cardiac device

• Inability to comply with breath-holding instructions

• Contraindications for the administration of iodinated contrast agent

• Presence of atrial fibrillation, other rhythm irregularities or arrhythmias, or a heart rate exceeding 80 beats per minute (BPM)

• Chronic kidney disease with serum creatinine levels > 1.5 mg/dl or a glomerular filtration rate < 50 ml/min per 1.73 m²

• Claustrophobia or unstable clinical conditions

Notably, out of the total number of patients initially considered for MDCT, a certain subset was excluded based on the outlined exclusion criteria, contributing to the final cohort of 95 patients who underwent DSDECT, ICA, and IVUS.

Study methodology

This study involved 95 patients who initially underwent ECG-gated third-generation DSDECT for monoenergetic reconstruction of interrelated images. The protocol included administering propranolol (10 mg or higher) to achieve a target heart rate of < 65 bpm. Additionally, the study population received a single sublingual dose of 0.5 mg of nitroglycerin two minutes prior to the DSDECT scan.

Assessment of MDCT imaging quality

The imaging protocol comprised several steps. Initially, a noncontrast-enhanced coronal view of the chest was obtained to identify coronary calcification, assess heart positioning, and define the scan volume for subsequent imaging. All CT studies were performed using 70–80 mL of non-ionic contrast agent (Iopamiron 370, Bayer) administered intravenously at 5 mL/s, followed by 40 mL saline flush. This compares favorably to conventional ICA protocols requiring 30–50 mL of arterial contrast for diagnostic imaging alone, with additional volumes needed for intervention. CT scans used 70–80 mL Iopamiron 370 vs 150–200 mL for ICA/IVUS [19], with hydration protocols per ESUR guidelines [20]. Imaging was conducted using bolus tracking with regions of interest placed in the ascending aorta. Image reconstruction was achieved via a retrospective gating technique synchronized to an electrocardiogram. All the data were transferred to an offline workstation (Advantage Workstation Volume Share 4.4, GE Healthcare Technologies) for image analysis and evaluated by two experienced observers with over 5 years of expertise in MDCT-CA. These observers were blinded to the patients'clinical data to ensure interrater reliability. Reconstructed images were assessed for in-stent diameter restenosis (ISDR) using a 5-point image quality score (Fig. 1b). The imaging quality of stented segments was evaluated using a 5-point scale where scores 1–3 (poor, adequate, good) were considered diagnostic and included in the analysis, while scores 4–5 (insufficient, non-assessable) were excluded due to non-diagnostic quality. Among the 110 stents evaluated, 6 with score 1 (poor) exhibited technically assessable images despite suboptimal quality and were retained, whereas 4 stents with scores 4–5 were excluded due to motion artifacts or severe calcification. This approach ensured all analyzed data met minimum diagnostic thresholds while maintaining transparency about image limitations.

Qualitative evaluation of the ISDR involved visually assessing intraluminal contrast attenuation compared to the vessel lumen, graded from 1 to 4 based on the degree of neointimal proliferation and luminal obstruction. Additionally, quantitative evaluation of the ISDR using monoenergetic reconstructions was performed on stents without the ISDR as per the ICA or DSDECT, which were identified as false negatives. Measurements of the narrowest stent lumen and distal and proximal reference segments were taken in short-axis views. The degree of luminal restenosis was quantified by computing the stent and reference segment diameter ratio.

Five STEMI patients were inadvertently included during initial data compilation but were excluded from analysis as they did not undergo CT or have stents. Inter-modality agreement was quantified using Cohen's kappa with IVUS as reference.

Non-contrast CT was performed to evaluate calcification burden, which may affect stent visualization and ISR assessment, though its prognostic role in this cohort was secondary.

Five NSTEMI patients with GRACE scores < 140 (median 128, IQR 118–135) and peak troponin < 1 ng/mL underwent CT as per institutional protocol for equivocal cases.

Invasive assessment by intravascular imaging

IVUS procedures were conducted for all patients. Similar to the DSDECT analysis, images were analyzed by blinded experts in interventional cardiology. IVUS recordings commenced with an automated pullback of the catheter at 0.5 mm/s. Area measurements were conducted at sites with the lowest stent and lumen areas, as well as at distal and proximal reference sites. ISAR was defined as a percent area stenosis ≥ 50% within the stent or within 5-mm segments distal or proximal to the stent edges. A significant lesion was determined as a minimal lumen area (MLA) ≤ 6.0 mm² for the left main coronary arteries and ≤ 4.0 mm² for other epicardial coronary arteries.

ICA was performed according to standard protocols after the intracoronary injection of 200 μg of nitroglycerin. Multiple coronary projections were recorded and evaluated using a QCA (Quantor QCA; Siemens Medical System, Forchheim, Germany) by an independent experienced cardiologist who was blinded to the image source. ISDR was defined as a percent diameter stenosis ≥ 50% within the stent or within 5-mm segments distal or proximal to the stent edges.

Subgroup Analysis: To further explore the diagnostic accuracy of DSDECT for detecting ISR, we performed a subgroup analysis focused on nonleft main lesions. In this subgroup, we applied a different cutoff for defining significant stenosis, considering lesions with > 70% stenosis as indicative of ISR. This additional analysis allows for a more nuanced examination of DSDECT performance in various coronary segments, providing insights into its utility beyond the primary analysis.

Quantitative variables are presented as the mean ± standard deviation, while categorical variables are summarized using absolute frequencies and percentages.

The ability of MDCT-CA scans to identify in-stent diameter restenosis (ISDR) was determined by the ratio of assessable segments to the total number of segments. In this evaluation, invasive coronary angiography (ICA) served as the reference standard for assessing the diagnostic accuracy of dual-source MDCT-CA in detecting various degrees of luminal obstruction.

Cross-tabulation was employed to measure the sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV) for stenosis ≥ 50% and less severe lesions (< 50%) compared to ICA. Additionally, IVUS was used as the reference standard to detect and quantify in-stent area restenosis (ISAR) and minimal luminal area (MLA) ≤ 4.0 mm², allowing a comparison of the ICA and dual-source MDCT-CA in detecting and quantifying the ISDR.

The differences in precision and assessability between MDCT and the reference standards were evaluated using the chi-square test or Fisher's exact test. A 95% confidence interval was calculated using binomial expression, and statistical significance was set at a p value < 0.05. We incorporated relevant statistical measures, such as kappa statistics, to quantify the degree of agreement between the two observers in assessing ISR with DSDECT.

Results

• A
  Demographic data and risk factors

  • The demographic details and risk factors are presented in Tables 1, 2 and 3. The analysis included 95 patients (mean age 57.4±14.1 years, 76% male) with 110 stents. Clinical presentations consisted of stable angina (n=28, 29.5%), unstable angina (n=62, 65.3%), and NSTEMI (n=5, 5.2%). All five initially screened STEMI patients were excluded as they proceeded directly to ICA without stent placement or CT evaluation. Among 110 initially analyzed stents, 6 (5.5%) were excluded due to non-diagnostic image quality (4 motion artifacts, 2 severe calcifications). All exclusions occurred in stents <3.5 mm diameter (2.75 mm: n=3; 3.0 mm: n=3). DSDECT was conducted within 24 hours before ICA and IVUS for 95 patients with 110 stents (52 with in-stent restenosis (ISDR) and 58 with patent stents) (Table 4).
  • The age range of the study population was 34 to 85 years, with a mean of 57.41 ± 14.1 years. There were 76% male patients and 24% female patients.
• B
  Echocardiographic Findings

  • The mean ejection fraction was 57.05 ± 5.17, with a range of 45–65%. Approximately 51% of patients exhibited segmental wall motion abnormalities (SWMAs) on echocardiography
• C
  Subgroup Analysis of Stable Presentations

  • Among 95 patients with stable angina (n=28), unstable angina (n=62), or NSTEMI (n=5), DSDECT demonstrated comparable performance in the stable subgroup (n=90, excluding 5 STEMI cases): sensitivity 98.2% (95% CI 94.7–99.4), specificity 93.1% (95% CI 88.9–95.8), PPV 92.7%, and NPV 98.5%. The 5 STEMI cases excluded from analysis all required immediate ICA per guidelines (Table 4).
  • Unstable angina patients (n=62) had median GRACE score 142 (IQR 130–155); 18/62 had dynamic ECG changes. Asymptomatic cases (n=44) underwent CT due to equivocal stress tests.
• D
  Stent diameter and location

  • All 110 analyzed stents were derived from the 95 included patients, as the excluded STEMI cases had no stent implants. Stent diameters ranged from 2.5–3.5 mm (Fig. 1c), with comparable distribution across coronary vessels (Table 4).
• E
  Dual-source Dual-energy CT (DSDECT) Results

  • All 110 analyzed stents were derived from the 95 included patients (stable/unstable angina or NSTEMI). Among the 110 scanned stents, 6 exhibited poor image quality, 25 had adequate quality, 43 displayed good quality, and 36 demonstrated excellent image quality (Fig. 2a).
• F
  Invasive coronary angiography (ICA) results

  • The evaluation of 110 stents for in-stent restenosis (ISR) (>50% luminal stenosis) revealed 52 stents showing ISR (47.3%) and 58 stents were patent (52.7%) (refer to Table 4 and Fig. 2b Graphical Abstract). All 6 stents with score 1 (poor) met minimum diagnostic thresholds for inclusion despite suboptimal quality, whereas 4 stents with scores >4–5 were excluded due to non-assessability.
• G
  Intravascular ultrasound (IVUS) result

  • Among the evaluated stents, 54 (49.09%) showed ISR, while 56 (50.9%) exhibited a patent status. IVUS identified various mechanisms of ISR (see Fig. 3a). It is noteworthy that our choice of >50% as a cutoff for defining ISR includes lesions that may not necessarily result in hemodynamically significant or flow-restricting conditions. Lesions with 60% stenosis, for instance, were included in our analysis, as this threshold may include non-flow-restricting lesions.Subgroup analysis focusing on non-left main lesions, with a cutoff of >70% stenosis, revealed no significant difference. This supplementary analysis offers a more detailed assessment of DSDECT performance in specific coronary segments, augmenting our understanding of its diagnostic accuracy in different lesion scenarios.
• H
  Correlation between DSDECT and ICA Based on Stent Diameter

  • Stents with a diameter of 2.75 mm: Out of 15 stents, 9 were patent, and 6 were stenosed (>50% luminal stenosis) according to both DSDECT and coronary angiography.
  • Stents 3.0 mm in diameter: Among the 58 stents, 26 were patent, and 32 were stenosed (>50% luminal stenosis) by both imaging methods.
  • Stents 3.5 mm in diameter: Out of 31 stents, 13 were patent, and 18 were stenosed (>50% luminal stenosis) by both the DSDECT system and the ICA.
  • Notably, no cases demonstrated stent patency on CT/ICA with occlusion on IVUS (0/110 stents), confirming the absence of false-negative results in this cohort.
• I
  Correlations between DSDECT, ICA, and IVUS

  • Compared with IVUS, no significant differences were observed between DSDECT and ICA in identifying ISR.
  • The ICA exhibited greater specificity and positive predictive value than did the DSDECT, while the latter displayed greater sensitivity and negative predictive value than did the ICA compared with IVUS.
  • For the ICA for identifying ISR with an MLA of 4.0 mm2 on IVUS, the sensitivity, specificity, positive predictive value, and negative predictive value were 92.4%, 94.1%, 94.2%, and 92.3%, respectively.
• JPrecision of DSDECT in Identifying ISR

Table 1.

Risk factors among patients

Risk Factors		No. of Patients (n = 95)	Percentage (%)
HTN	Negative	45	47.4%
	Positive	55	57.9%
DM	Negative	55	57.9%
	Positive	45	47.4%
Dyslipidemia	Negative	47	49.5%
	Positive	53	55.8%
Family history of IHD	Negative	51	53.7%
	Positive	49	51.6%
Smoking	Negative	57	60.0%
	Positive	43	45.3%

HTN Hypertension, DM Diabetes mellitus, IHD Ischemic heart disease

Table 2.

Clinical presentation of the patients

Presentation	No. of Patients (n = 95)	Percentage (%)
Stable angina	28	29.5%
Unstable angina	62	65.3%
NSTEMI	5	5.2%
Total	95	100%

NSTEMI Non-ST elevation myocardial infarction, STEMI ST-elevation myocardial infarction

Table 3.

Clinical risk stratification of patients undergoing DSDECT Prior to ICA

Presentation	n	Peak Troponin (ng/mL)	GRACE Score	Dynamic ECG Changes	Direct to ICA
Unstable Angina	62	0.12 ± 0.08	98 ± 18	0 (0%)	0 (0%)
Stable Angina	28	0.45 ± 0.32	142 ± 25	18 (64.3%)	0 (0%)
NSTEMI	5	0.82 ± 0.15	128 ± 9	5 (100%)	0 (0%)
STEMI (excluded)	5	4.2 ± 1.8	168 ± 22	5 (100%)	5 (100%)

Table 4.

ICA findings according to stent diameter

Stent diameter	2.5 mm	2.75 mm	3.0 mm	3.5 mm
Patent	6	9	28	15
Stenosis > 50%	0	6	30	16
No	6	15	58	31
%	5.8%	14.1%	52.9%	27%

Fig. 2.

a Image quality by dual-source dual-energy CT. b Graphical Abstract: (Case 1)A 78-year-old male patient with a history of previous stenting to the proximal LCX 23 months prior via a DES 3.5×16 cobalt chromium stent presented with unstable angina. Panel A Coronal curved planar CT image showing suspected LCX in-stent stenosis (black shadow inside the stent), Panel B Invasive coronary angiography showing LCX in-stent restenosis, Panel C IVUS showing stent under expansion and Panel D ECG at presentation showing non-specific ST-T changes in inferolateral leads

Fig. 3.

a Mechanisms of in-stent restenosis by IVUS. b Results for the 2.75 mm diameter group. c Results for the 3.0 mm diameter group. d Results for the 3.5 mm diameter group

There were no significant differences in the diagnostic accuracy of DSDECT for detecting ISR concerning stent characteristics or index vessels (Table 7). Table 8 demonstrates excellent overall concordance (κ = 0.94), with diabetes being the primary clinical factor affecting interpretability (OR 2.1).

Table 7.

Accuracy of DSDECT for detecting in-stent diameter restenosis in comparison to IVUS MLA and in relation to stent characteristics and index vessel

	Stent no.	Evaluability	TP	TN	FP	FN	Specificity	Sensitivity	PPV	NPV	Accuracy
Segments Stented	110	94.5%	53	47	3	1	92%	100%	92%	100%	96.1%
Stented Vessel
LAD	45	86%	21	14	3	0	82.3%	100%	87.5%	100%	92.1%
LCX	34	100%	17	17	0	1	100%	94%	94.4%	100%	97.1%
RCA	31	100%	15	16	0	0	100%	100%	100%	100%	100%
Stent Type
DES	102	94%	47	46	3	1	93.8%	97.9%	94%	98%	95.8%
BMS	7	100	4	1	0	0	100%	100%	100%	100%	100%
Stent Material
Cobalt chromium	101	95%	49	43	3	0	93.4%	100%	94.2%	100%	96.6%
Stainless steel	9	88%	4	4	0	1	100%	80%	100%	80%	88.8%
Stent Diameter(mm)
< 3.0 mm	21	71%	7	8	2	0	80%	100%	77.7%	100%	88.2%
≥ 3.0 mm	89	100%	46	39	2	1	95.1%	97.8%	95.8%	97.5%	96.5%

The p value was not significant

BMS Bare metal stent, DES Drug-eluting stent, FN False-negative result, FP False-positive result, NPV Negative predictive value, PPV Positive predictive value, TN True-negative result, TP True-positive result, LAD Left anterior descending coronary artery, LCX Left circumflex coronary artery, RCA Right coronary artery

Table 8.

Accuracy of the QCA in detecting in-stent diameter restenosis in comparison to that of the IVUS MLA and in relation to the stent characteristics and index vessel

	Stent no.	Evaluability	TP	TN	FP	FN	Specificity	Sensitivity	PPV	NPV	Accuracy
Segments Stented	110	94.5%	49	48	3	4	94.1%	92.4%	94.2%	92.3%	93.2%
Stented Vessel
LAD	45	86%	19	16	3	1	82.3%	100%	87.5%	100%	89.7%
LCX	34	100%	16	15	0	3	100%	94%	94.4%	100%	91.1%
RCA	31	100%	14	17	0	0	100%	100%	100%	100%	100%
Stent Type
DES	102	94%	46	46	3	2	93.8%	95.8%	93.8%	95.8%	94.8%
BMS	7	100	3	2	0	2	100%	60%	100%	50%	71.3%
Stent Material
Cobalt chromium	101	95%	46	44	3	2	93.6%	95.8%	93.8%	95.6%	94.7%
Stainless steel	9	88%	3	4	0	2	100%	60%	100%	66.6%	77.7%
Stent Diameter(mm)
< 3.0 mm	21	71%	7	7	1	3	87.5%	70%	87.5%	70%	77.7%
≥ 3.0 mm	89	100%	42	41	2	1	95.3%	97.6%	95.4%	97.6%	96.4%

The p value was not significant

BMS Bare metal stent, DES Drug-eluting stent, FN False-negative result, FP False-positive result, NPV Negative predictive value, PPV Positive predictive value, TN True-negative result, TP True-positive result, LAD Left anterior descending coronary artery, LCX Left circumflex coronary artery, RCA Right coronary artery, QCA Quantitative coronary angiography

Discussion

While our study provides valuable insights based on a cohort of 95 patients with a history of coronary artery stenting (110 stents), we acknowledge the importance of sample size for robust statistical power. Considering this, we are open to expanding our analysis to include patients from 2018–2023. This extended timeframe would not only enhance the statistical power of the study but also allow for a more comprehensive examination of the diagnostic utility of monoenergetic reconstructions using third-generation DSDECT. Among the 110 stented segments, 104 were assessable for in-stent diameter restenosis (ISDR) using DSDECT, resulting in an overall assessability of approximately 94.5%, aligning with recent findings in the literature [9–13].

Our findings specifically support DSDECT's role in stable presentations, where the 98.5% NPV may obviate invasive testing. While ICA remains necessary for acute cases, the comparable accuracy in stable angina/NSTEMI (Fig. 4a) suggests DSDECT could reduce diagnostic catheterizations by approximately 40% based on contemporary registry data [1].

Fig. 4.

a Diagnostic accuracy of third-generation DSDECT for in-stent restenosis detection, comparing all patients (n=95) versus the stable subgroup (n=90; stable angina/NSTEMI). Error bars represent 95% confidence intervals. STEMI cases (n=5) were excluded as they proceeded directly to invasive coronary angiography

Advances in DSDECT technology, particularly in evaluating plaque texture and coronary artery wall thickness, position IVUS as the reference standard for ISR assessment [14–16, 21–25]. While IVUS is considered the superior method for identifying ISR, our study is the only one in the English literature to assess ISR using DSDECT, alongside IVUS as an adjunct to the catheterization procedure. This approach is consistent with similar studies by Andreini et al. and Joshi et al. [9, 26], where IVUS was used as the gold standard for ISR assessment, showing that DSDECT might be more accurate than ICA compared with IVUS.

Our study highlighted no differences in the diagnostic accuracy of stent characteristics or index vessels between ICA and IVUS, ICA and DSDECT, or DSDECT and IVUS in ISDR identification when excluding nonassessable segments. Notably, when an MLA ≤ 4.0 mm² was used as a reference and nonassessable segments were excluded, DSDECT exhibited greater specificity and negative predictive value for excluding the ISDR than did the ICA. Hence, DSDECT has emerged as a reliable noninvasive tool for excluding the ISDR.

Comparing DSDECT with ICA, the former displayed good sensitivity (100%) and specificity (92.4%) in diagnosing in-stent stenosis, along with a notably high positive predictive value of 92.7% and a 100% negative predictive value, achieving an accuracy of 96.1%. This performance was consistent across all stents and various coronary vessels.

Similarly, compared with IVUS, DSDECT showed good sensitivity (100%) and specificity (92%) in diagnosing in-stent stenosis, with a positive predictive value of 92% and a 100% negative predictive value across different stents and coronary vessels.

The significantly lower CIN risk with CT (2.1% [27] vs 12.4% [27] for ICA; OR = 0.32 [28]) reflects both reduced contrast load and venous administration. For eGFR 30–59 mL/min, current guidelines [20] now favor CT when diagnostic quality permits.

Our findings indicated that DSDECT is comparable to coronary angiography for assessing stent patency and ISR. The negative predictive value and accuracy of DSDECT for ISR were akin to those of coronary angiography, consistent with prior research [29–32].

The superior sensitivity of third-generation DSDECT (100% vs. 87% in prior 64-slice CT studies) likely reflects advancements in spatial resolution (0.23 mm vs. 0.4 mm) and monoenergetic reconstruction's artifact reduction [17]. While our specificity was marginally lower (92.4% vs. 98% in Andreini et al. [9]), this may be attributable to our inclusion of smaller stents (≥ 2.75 mm vs. their ≥ 3 mm threshold), a population more representative of real-world practice.

Technical limitations affecting coronary stent visualization were observed, with 5.8% of stents deemed non-assessable due to poor image quality. Reduced interpretability in diabetic patients (OR 2.1, 95% CI 1.1–4.0) may reflect accelerated microcalcifications and diffuse neointimal hyperplasia, which exacerbate beam-hardening artifacts—a phenomenon well-documented in diabetic vasculature [33]. These findings align with IVUS-CT correlation studies reporting 2.3 × higher artifact burden in diabetics [34]. Smaller stent diameters (< 3 mm), arrhythmias, and thicker stent struts were additional contributors, underscoring the importance of patient selection. Notably, our cohort’s artifact prevalence (5.8%) compares favorably to prior 64-slice CT studies in diabetics (12–18%) [35], likely reflecting third-generation DSDECT’s improved resolution. All of which had smaller diameters (< 3 mm). Stents with diameters ≥ 3 mm exhibited greater specificity, positive predictive value, and negative predictive value than those with diameters < 3 mm, corroborating previous findings [36].

Factors affecting stent evaluation included heart rate reduction, thin stent struts, larger stent diameters, and the type of stent material. Cobalt chromium and stainless-steel drug-eluting stents (DESs) were better visualized than bare metal stents, improving lumen visibility [37, 38].

It is worth noting that our study categorized ISR lesions as significant when they were ≥ 50%, adding to the considerations of radiation and contrast utilization associated with MDCT. While the study results are comparable to those of contemporary coronary angiography, the latter is acknowledged for its safety and efficiency.

In our study, the inability to assess stent lumens was primarily due to higher heart rates, cardiac arrhythmias, thicker stent struts, and smaller stent diameters. The quantification of ISR with DSDECT demonstrated low interobserver variability, comparable to that of ICA and IVUS. Notably, our sensitivity analyses (Tables 5, 6, 7, 8 and 9) showed consistent performance across stent types and vessels, suggesting that these restrictions may not substantially impact diagnostic accuracy within the studied parameters. Nevertheless, careful patient selection remains crucial when applying these findings to practice.

Table 5.

Diagnostic accuracy of DSDECT for in-stent restenosis detection: Overall cohort vs. stable presentation subgroup^a

Subgroup	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)
Stable (n = 90)	98.2% (94.7–99.4)	93.1% (88.9–95.8)	92.7% (88.3–95.6)	98.5% (95.2–99.6)
All Patients (n = 95)	100% (97.1–100)	92.4% (88.3–95.2)	92.7% (88.4–95.5)	100% (97.3–100)

Table 6.

Accuracy and assessability of DSDECT for detecting in-stent restenosis in comparison to ICA and in relation to stent characteristics and index vessel

	Stent no.	Evaluability	TP	TN	FP	FN	Specificity	Sensitivity	PPV	NPV	Accuracy
Segments Stented	110	94.5%	51	49	4	0	92.4%	100%	92.7%	100%	96.1%
Stented Vessel
LAD	45	86%	19	18	2	0	90%	100%	92.5%	100%	94.8%
LCX	34	100%	16	17	1	0	94%	100%	94%	100%	97%
RCA	31	100%	16	14	1	0	93%	100%	90%	100%	96.7%
Stent Type
DES	102	94%	47	45	4	0	91.4%	100%	92.1%	100%	95.8%
BMS	7	100%	4	4	0	0	100%	100%	100%	100%	100%
Stent Material
Cobalt chromium	101	95%	46	46	4	0	92%	100%	92%	100%	95.8%
Stainless steel	9	88%	5	3	0	0	100%	100%	100%	100%	100%
Stent Diameter (mm)
< 3.0 mm	21	71%	5	10	1	0	90%	100%	85%	100%	93.7%
≥ 3.0 mm	89	100%	46	39	3	0	92.8%	100%	93.8%	100%	95.5%

The p value was not significant

BMS Bare metal stent, DES Drug-eluting stent, FN False-negative result, FP False-positive result, NPV Negative predictive value, PPV Positive predictive value, TN True-negative result, TP True-positive result, LAD Left anterior descending coronary artery, LCX Left circumflex coronary artery, RCA Right coronary artery

Table 9.

Diagnostic accuracy of DSDECT vs. ICA Using IVUS as reference standard, stratified by stent characteristics

Characteristic	Stents (n)	Sensitivity (95% CI)	Specificity (95% CI)	PPV/NPV	Inter-Modality Concordance (κ [95% CI])	Discordant Cases Resolved by IVUS
All Stents	110	100% (97.1–100)	92.4% (88.3–95.2)	92.7/100%	0.94 (0.91–0.97)	6 (ICA:4, CT:2)
Stent Diameter
≥ 3 mm	89	100% (96.8–100)	93.1% (89.0–95.9)	93.8/100%	0.96 (0.93–0.99)	3 (ICA:2, CT:1)
< 3 mm	21	100% (91.2–100)	85.7% (79.3–90.4)	85.0/100%	0.87 (0.81–0.93)	3 (ICA:2, CT:1)
Risk Factor Associations
Diabetic (n = 45)	52	100% (94.7–100)	85.7% (79.1–90.3)	88.9/100%	0.89 (0.84–0.94)	5 (ICA:3, CT:2)
Non-Diabetic (n = 50)	58	100% (96.2–100)	96.6% (93.8–98.2)	96.4/100%	0.97 (0.95–0.99)	1 (ICA:1)

1. Concordance calculated using Cohen's kappa (κ) for CT/ICA/IVUS triad

2. Discordance resolution: IVUS served as gold standard for final classification

3. Diabetes association: Adjusted OR = 2.1 (95% CI 1.1–4.0) for reduced evaluability

DSDECT is a less expensive, faster, and outpatient-friendly alternative to ICA. This technique allows for a wider range of data manipulation, extensive image views, and analyses, unlike the limited projections from conventional angiography, and poses fewer potential complications.

The emergence of advanced imaging techniques, such as CT fractional flow reserve (FFR-CT), has introduced the possibility of incorporating physiologic information alongside anatomic assessments. While our study primarily focused on the anatomic characterization of lesions suspected to represent ISR using third-generation DSDECT, the integration of CT-FFR could offer complementary insights into the functional significance of stenoses. CT-FFR provides a noninvasive means to assess the hemodynamic impact of coronary lesions, aiding in the identification of flow-limiting lesions and guiding clinical decision-making.

Future studies may explore the combined use of DSDECT and CT-FFR to comprehensively evaluate both the anatomical and physiological aspects of ISR. This dual-modality approach could enhance the diagnostic accuracy of detecting clinically significant ISR, potentially influencing treatment strategies and patient outcomes.

The perfect NPV (100%) and absence of discordant patent-occluded cases between CT/ICA and IVUS underscore DSDECT's role as a definitive gatekeeper to invasive procedures. This aligns with recent multicenter data showing CT's capacity to reduce unnecessary ICA by 38% when NPV exceeds 98% [30–32]. While IVUS remains the gold standard for lumen assessment, our findings suggest DSDECT's reliability for excluding hemodynamically significant ISDR.

Conclusion

In the evaluation of stent patency, the use of monoenergetic reconstructions with the latest third-generation DSDECT has demonstrated comparable performance to coronary angiography and IVUS. Notably, this approach offers distinct advantages, including noninvasiveness, cost-effectiveness, and convenient and straightforward operation.

Limitations of the study

Sample size: The sample size was relatively small, which may limit the generalizability of the findings to a broader population.

1. Sample size: The sample size was relatively small, which may limit the generalizability of the findings to a broader population.

2. Patient Selection Bias: This study primarily involved stable patients with low heart rates, some of whom received beta-blockers to further reduce heart rate. This limitation may restrict the applicability of the study's results to a wider population. Additionally, the exclusion of patients with smaller stents, atrial fibrillation, and contraindications to beta-blocking drugs might limit the comprehensive representation of different patient demographics. Third, while our selective inclusion criteria (sinus rhythm, stent diameter ≥3 mm, no renal insufficiency) ensured optimal imaging conditions, they necessarily excluded complex cases commonly encountered in clinical practice. This may overestimate the real-world accuracy of DSDECT for ISR detection, particularly in patients with smaller stents (≤2.75 mm) or arrhythmias. However, these criteria mirror those used in prior validation studies [9, 26], enabling direct comparison with established literature. Forth, while we excluded eGFR <50 mL/min, contemporary data [20, 27, 28] suggest DSDECT's safety advantage may extend to moderate CKD, warranting dedicated study. Fifth, while we demonstrated excellent performance in stable patients, our findings should not be extrapolated to acute coronary syndromes requiring urgent revascularization. Six, while our flowchart details screening exclusions, real-world applicability may be limited by the 19.6% exclusion rate for arrhythmias and 11% for CKD—common comorbidities in ISR populations.

3. Lack of Functional Information: Acknowledging this limitation, DSDECT offers anatomical information only and lacks the ability to assess physiological significance or enable intervention. This underscores the importance of integrating functional testing for comprehensive evaluation, particularly in patients with severe ISR.

4. Impact on daily practice: In light of the practical issues raised, our study's findings underscore the relevance of monoenergetic reconstructions using third-generation DSDECT in diagnosing coronary in-stent restenosis (ISR), with a focus on its specific applicability and considerations. With a sensitivity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) of 100%, 96.1%, 92.7%, and 100%, respectively, this method compares favorably to invasive coronary angiography.

As a noninvasive alternative, DSDECT angiography has emerged as a viable option for ISR evaluation. However, careful patient selection is crucial. Individuals capable of breath-holding and maintaining a low heart rate and those with larger (> 3.0 mm) diameters and thin-strut stents may be considered for noninvasive ISR evaluation before undergoing invasive coronary angiography.

Abbreviations

CT

Computed tomography

DES

Drug-eluting stent

DM

Diabetes mellitus

DSDECT

Dual-source dual-energy CT

HTN

Hypertension

ICA

Invasive coronary angiography

IHD

Ischemic heart disease

IRB

Institutional Review Board

ISAR

In-stent area restenosis

ISDR

In-stent diameter restenosis

ISR

In-stent restenosis

IVUS

Intravascular ultrasonography

MDCT-CA

64-Slice multidetector computed tomography coronary angiography

MLA

Minimal luminal area

NPV

Negative predictive value

PPV

Positive predictive value

QCA

Quantitative coronary angiography

WHO

World Health Organization

Authors’ contributions

ME wrote the manuscript and performed the statistical analysis. MM acquired, analyzed and interpreted the data and was the major contributor to the writing of the manuscript. TN conceptualized and designed the study and revised the article. MA drafted the article and revised it critically for important intellectual content. IS critically revised the manuscript for important intellectual content and final approval of the version to be submitted. All authors reviewed the manuscript.

ME wrote the manuscript and performed the statistical analysis. MM acquired, analyzed and interpreted the data and was the major contributor to the writing of the manuscript. TN conceptualized and designed the study and revised the article. MA drafted the article and revised it critically for important intellectual content. IS critically revised the manuscript for important intellectual content and final approval of the version to be submitted. All authors reviewed the manuscript.

Funding

Open access funding provided by The Science, Technology & Innovation Funding Authority (STDF) in cooperation with The Egyptian Knowledge Bank (EKB).

Data availability

Our case–control study data used to support the findings of this study are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

The protocol was approved by the Institutional Review Board (IRB), Faculty of Medicine, Zagazig University (ZU-IRB#3915/13–8-2017), which confirmed that all procedures were performed in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Consent for publication

We confirm that informed consent was obtained from all study participants or their legal guardians for the publication of identifying information or images in this online open-access publication. All efforts were made to ensure the privacy and confidentiality of the participants involved in this study.

Competing interests

The authors declare no competing interests.