Improved quantification of coronary artery luminal stenosis in the presence of heavy calcifications using photon-counting detector CT

Abstract

Coronary CT Angiography (cCTA) is commonly used to detect and quantify luminal stenoses in patients with coronary artery disease (CAD). However, its use is limited in patients with heavy coronary calcifications due to calcium blooming, which is caused by insufficient spatial resolution. This study evaluated the ability of a photon-counting-detector (PCD) CT in quantifying luminal stenosis in the presence of heavy calcifications relative to an energy-integrating-detector (EID) CT. Cylindrical rods of 4.5 mm diameter (with 3 mm lumen), which contained calcium hydroxyapatite (CaHA) to emulate calcifications of varying shapes and sizes and an iodine or blood analog to emulate the coronary lumen, were placed within an anthropomorphic thorax phantom and scanned at matched dose on an EID-CT and a PCD-CT scanner. Stenoses were qualitatively evaluated and quantified using commercial software. Measured percent area stenosis was compared to reference values. PCD-CT provided better visualization of calcium plaques and the patent lumen, and more accurate stenosis quantification for all plaques. In one rod (75% occlusion with ring-shaped plaque), only PCD-CT was able to determine that the vessel was not fully obstructed. The phantom results indicate luminal stenoses that were previously considered non-assessable due to the presence of heavily-calcified plaques can be assessed using PCD-CT. Clinical studies to support these conclusions are underway.

1. INTRODUCTION

Coronary artery disease (CAD) and its subsequent complications have the highest mortality rate of any disease in the United States [1]. Many individuals with CAD often do not experience symptoms prior to onset of a coronary event [2]. Contrast-enhanced coronary CTA (cCTA) is commonly performed to provide a noninvasive assessment of the coronary lumen to detect presence of significant CAD [3], which can reduce hospital stay and expenses in comparison to invasive angiography [4]. However, presence of heavy calcium in coronary arteries presents a challenge in many cCTA scans, making them difficult to visualize contrast-filled lumen [5]. Although sensitivity and negative predictive value are high, the specificity of cCTA is relatively low due to the inability to accurately evaluate lumen patency in the presence of heavy calcifications [6]. This is primarily due to the finite spatial resolution of current CT scanners, which results in calcium blooming artifact that compromises clear visualization of the lumen.

Imaging small coronary artery vessels with cCTA is currently limited due to the spatial resolution of clinical scanners using energy integrating detectors [7]. Visualization of stenosis in coronary artery calcified plaques is limited due ot blooming and partial volume artifacts, leading to inaccurate estimation of vessel lumen [8]. PCD-CT systems use direct conversion technology and combat high count rates with small pixel size as compared to EID-CT counterparts leading to improved intrinsic spatial resolution [9]. PCD-CT can be especially advantageous when employed to cardiac imaging due to its high temporal resolution, radiation dose efficiency, reduction in metal artifact, and high spatial resolution of greater than 33 lp/cm [10]. To address the issues above, this phantom study aimed to demonstrate the impact of PCD-CT, which can achieve improved spatial resolution without sacrificing dose efficiency, on the accuracy of luminal stenosis assessment for cCTA.

2. METHODS

2.1. Stenosis phantom preparation

Four coronary artery stenosis phantoms with CaHA plaques (QRM GmbH, Möhrendorf, Germany) were used to emulate coronary arteries containing stenosed vessels. Two calcified plaque geometries were available (those existing on only one side of the artery phantom and those forming concentric rings, Figure 1 A/B). The lumens were filled with materials that emulated non-iodinated and iodinated blood to mimic non-contrast-enhanced and contrast-enhanced scans. The one-sided (OS) phantoms contained plaques with 50%, 25%, and 15% area stenosis, while the ring (R) phantom contained plaques with 50% and 75% area stenosis. Both phantom rods had a diameter of 4.5 mm and a length of 65 mm; the plaques were 10 mm in length (Figure 1 A/B).

Figure 1:

Diagram of the coronary artery stenosis phantoms. (A) Phantom vessel with three one-sided CaHA plaques, referred to as OS-A/B/C, with the degree of stenosis indicated. (B) Phantom vessel with two CaHA plaques shaped like rings, referred to as R-A/B, with the degree of stenosis indicated. These coronary phantoms were available with both blood-mimicking (non-contrast) and iodine-mimicking (contrast-enhanced) materials in the lumen. (C) The stenosis phantoms were placed inside of a water-filled acrylic cylinder at the center of a thoracic phantom. The non-contrast and contrast-enhanced vessels for a given calcification type were scanned simultaneously. Figure 1 A/B were modified from figures in the phantom manual (QRM GmbH).

The stenosis phantoms were placed on a stationary holder in the center of a water-filled cylinder, which was inserted into a 35 cm (lateral width) anthropomorphic thorax phantom (QRM GmbH, Möhrendorf, Germany), as shown in Figure 1-C. For a given calcified plaque configuration, both the non-contrast-enhanced (NC) and contrast-enhanced (C) vessel phantoms were placed in the thorax phantom and scanned. This was repeated for all stenosis phantoms.

2.2. EID-CT and PCD-CT scans and reconstructions

For each coronary phantom configuration, the thorax phantom was scanned on a state-of-the-art dual-source EID-CT scanner (SOMATOM Force, Siemens Healthineers GmbH) and a dual-source PCD-CT scanner (NAEOTOM Alpha, Siemens Healthineers GmbH). The PCD-CT scan was acquired using a high-resolution (HR) mode that uses a 120 x 0.2 mm detector pixel size, which yields an effective acquisition slice width of 200 μm, and one energy threshold (20 keV). EID-CT and PCD-CT data were acquired using the same tube potential (120 kV) and radiation dose (CTDIvol (32 cm) = 9.3 mGy).

Images were reconstructed using each platform’s native iterative reconstruction algorithm. EID-CT images were reconstructed using a routine vascular kernel (Bv49; 0% MTF at 9.37 lp/cm), while the PCD-CT images were reconstructed with a sharper kernel (Br60; 0% MTF at 12.9 lp/cm), both at 50 mm display field-of-view. The sharper kernel is available on the PCD-CT to take advantage of the better spatial resolution of the detector.

2.3. Stenosis assessment

Both qualitative and quantitative assessments were performed for the luminal stenosis. Visual assessment was performed to evaluate the calcification’s shape in the presence of iodine contrast, with the non-contrast phantom used to provide a visual reference. In addition, line profiles through the center of each vessel were calculated and compared. For the ring-shaped plaques, the profiles were taken at multiple orientations between 0° and 180° and averaged.

Percent area stenosis in the contrast-enhanced vessel phantoms was quantified using commercially available software⁵ (Syngo Via CT Vascular; Siemens Healthineers GmbH). The vessel was manually segmented, and both the plaque and two fully patent lumens (one before and one after the calcified plaque) were selected by the operator. Upper and lower CT number thresholds were chosen to encompass the range of CT numbers in the contrast-enhanced phantom lumens. The area stenosis was determined by calculating the ratio of the cross-sectional area at the position of the stenosis relative to at the fully patent portions of the vessel and reported as a percentage (0% indicates a fully patent lumen and 100% indicated a fully blocked lumen). The reported values were obtained by averaging eight individual measurements of the percent area stenosis across the 10 mm length of the calcified plaque, which were then compared to the known values provided by the phantom manufacturer.

3. RESULTS

Figure 2 provides a visual comparison between the PCD-CT and EID-CT images for the three one-sided plaques. For both the EID-CT and PCD-CT images, the presence of a calcified plaque and the iodinated lumen are visualized for all three stenosis levels, indicating a partial blockage. The PCD images more clearly convey the shape and size of the calcified plaques compared to EID-CT.

Figure 2:

Visual assessment of the one-sided plaques (OS), with the degree of stenosis indicated in parentheses. Vessels with iodine in the lumen (contrast-enhanced scan) are on the left side of each image pair, while vessels on the right side of each image pair contained only blood in the lumen (non-contrast scan). Note that the orientation of the plaque segment may vary between EID-CT and PCD-CT due to rotation of the rods that occurred when the phantom was transferred between scanners.

The improvement in plaque/lumen delineation is more obvious for the ring-shaped plaques, shown in Figure 3. For both the 50% and 75% stenosis, the ring-shaped plaque is clearly delineated from the lumen on PCD-CT images, but cannot be appreciated in the EID-CT images. This finding is supported by the line profiles shown in Figure 3. For both the 50% stenosis and 75% stenosis phantoms, the iodine contrast in the PCD-CT images can be clearly distinguished from the CaHA plaque, with the signal amplitude decreasing in the patent portion of the lumen. For the EID-CT images, the center of each vessel shows elevated signal due to calcium blooming, making it difficult to determine the lumen size.

Figure 3:

Visual assessment and line profiles for the ring-shaped plaques with and without iodine in the lumen. The degree of stenosis is indicated in parentheses. Line profiles were measured at multiple angles and averaged, with the standard deviation indicated by the shaded region.

These observations are reinforced by the percent area stenosis measurements, which are presented in Table 1. When compared with the reference values for percent area stenosis provided by the phantom manufacturer, PCD-CT measurements are more accurate than EID-CT measurements for all phantom configurations. Although the improvement for the one-sided plaques is relatively small, PCD-CT measurements are notably more accurate for the two ring-shaped plaques. This is likely due to the fact that lumens of ring-shaped plaques are completely surrounded by the calcifications and hence calcium blooming has a much greater impact. This is especially notable for the 75% stenosis configuration, where only PCD-CT demonstrated that the vessel was not completely blocked.

Table 1.

Percent area stenosis measurements

	Reference	EID-CT	PCD-CT
OS-A	50%	56.9%	54.8%
OS-B	25%	33.1%	25.1%
OS-C	15%	18.5%	13.4%
R-A	50%	82.5%	66.9%
R-B	75%	100.0%	90.8%

^*OS: One-sided calcified plaque; R: ring-shaped calcified plaque

4. CONCLUSIONS

PCD-CT, with its improved spatial resolution, provides a more accurate assessment of coronary luminal stenosis than EID-CT, especially for ring-shaped plaques. For the heaviest calcification evaluated (a ring-shaped plaque with 75% area stenosis), only the PCD-CT images were able to resolve the presence of iodine in the lumen, indicating that the vessel was not fully blocked. These data demonstrate PCD-CT’s potential to assess heavily calcified plaques that were previously considered to be non-assessable due to high percentage of luminal stenosis. Clinical studies to measure the impact of the improved accuracy in quantifying percent area stenosis are underway in our center.