Equilibrium phase images of the liver using a contrast-enhancement boost instead of the portal vein phase

Abstract

BACKGROUND

Three-phase dynamic computed tomography imaging is particularly useful in the liver region. However, dynamic imaging with contrast media has the disadvantage of increased radiation exposure due to multiple imaging sessions. We hypothesized that the contrast enhancement boost (CE-boost) technique could be used to enhance the contrast in equilibrium phase (EP) images and produce enhancement similar to that of portal vein phase (PVP) images, and if this is possible, EP imaging could play the same role as PVP imaging. We also speculated that this might allow the conversion of three-phase dynamic imaging to biphasic dynamic imaging, reducing patients’ radiation exposure.

AIM

To determine if the CE-boost of EP, CE-boost (EP) is useful compared to a conventional image.

METHODS

We retrospectively analyzed the cases of 52 patients who were diagnosed with liver cancer between January 2016 and October 2022 at our institution. From these computed tomography images, CE-boost images were generated from the EP and plane images. We compared the PVP, EP, and CE-boost (EP) for blood vessels and hepatic parenchyma based on the contrast-to-noise ratio (CNR), signal-to-noise ratio, and figure-of-merit (FOM). Visual assessments were also performed for vessel visualization, lesion conspicuity, and image noise.

RESULTS

The CE-boost (EP) images showed significant superiority compared to the PVP images in the CNR, signal-to-noise ratio, and FOM except regarding the hepatic parenchyma. No significant differences were detected in CNR or FOM comparisons within the hepatic parenchyma (P = 0.62, 0.67). The comparison of the EP and CE-boost (EP) images consistently favored CE-boost (EP). Regarding the visual assessment, the CE-boost (EP) images were significantly superior to the PVP images in lesion conspicuity, and the PVP in image noise. The CE-boost (EP) images were significantly better than the EP images in the vessel visualization of segmental branches of the portal vein and lesion conspicuity, and the EP in image noise.

CONCLUSION

The image quality of CE-boost (EP) images was comparable or superior to that of conventional PVP and EP. CE-boost (EP) images might provide information comparable to the conventional PVP.

Key Words: Liver; Computed tomography; Contrast enhancement boost; Image quality; Lesion conspicuity

Core Tip: The aim of this study was to evaluate the contrast enhancement boost method of enhancing equilibrium phase (EP) computed tomography images in liver imaging compared to portal venous phase and EP images, and to reduce radiation dose by switching from triphasic to biphasic imaging. The results show that contrast enhancement boost of EP images have the same or better image quality as portal venous phase images, suggesting that a reduction in radiation dose is possible.

INTRODUCTION

The use of computed tomography (CT) scans is increasing worldwide, and this imaging modality is useful because its application of contrast media (CM) can provide information about the hepatic parenchyma and blood vessels[1-5]. Three-phase dynamic CT imaging, in which contrast is injected followed by three staggered imaging sessions, is particularly useful in the liver region[6,7]. However, dynamic imaging with CM has the disadvantage of increased radiation exposure due to multiple imaging sessions.

The arterial phase (AP) of three-phase dynamic CT imaging is useful for understanding the anatomy of arteries and diagnosing early enhancing tumors[8,9]. The portal vein phase (PVP) is useful for understanding the portal vein's anatomy, enhancing the hepatic parenchyma, and diagnosing ring enhancement of tumors[7,10]. The equilibrium phase (EP), similar to the PVP, is useful for diagnosing tumor ring enhancement and delineating the portal vein, and for the washout of early enhancing tumors[7-9]. These various characteristics are used to help diagnose liver tumors. Hepatocellular carcinoma (HCC) can be diagnosed by the presence of early enhancement in the AP and washout in the EP[7,8]. The diagnosis of hilar cholangiocarcinoma can be made from the enhancement of the tumor over the PVP and EP, Oligometastatic tumors (e.g., metastatic liver cancer) can be diagnosed by hypoattenuation within the hepatic parenchyma in the PVP or EP[11-13].

A technique called the contrast enhancement boost (CE-boost; Canon Medical Systems, Tochigi, Japan) was recently developed to further improve the contrast enhancement of contrast-enhanced CT images[14]. In this technique, a subtraction image is acquired using a new subtraction algorithm, processed for noise reduction, and then added to the original image. Subtracting the non-contrast image from the contrast image to create a CE-boost image can enhance the contrast effect[15].

We hypothesized that the CE-boost technique could be used to enhance the contrast in EP images and produce enhancement similar to that of PVP images, and if this is possible, EP imaging could play the same role as PVP imaging. We also speculated that this might allow the conversion of three-phase dynamic imaging to biphasic dynamic imaging, thereby reducing the number of imaging procedures and reducing patients' radiation exposure. We conducted the present study to compare CE-boost of EP [CE-boost (EP)] images with conventional PVP images and EP images, respectively, and to determine whether CE-boost (EP) images are useful.

MATERIALS AND METHODS

Patients

This was a retrospective study approved by the institutional review board of Oita University Hospital, approval No. 2643). Informed consent was not required. We analyzed the cases of the patients who underwent liver resection between January 2016 and October 2022 at our institution and were diagnosed with cancer based on the pathology of the resected liver. Among these patients, those who had not undergone liver procedure (liver resection, arterial embolization, and radiofrequency ablation) prior to liver resection were selected, and among these patients, we selected the patients with plain CT images and three-phase dynamic CT images taken before liver resection. Patients with imaging artifacts or insufficient patient data were excluded. Finally, only the cases of the patients imaged with a single device were included, using a sequential hybrid-type iterative reconstruction method to eliminate the influence of differences in CT equipment on the images.

CT protocols

All patients underwent imaging with a multidetector CT system at Oita University Hospital. A 320-slice multidetector CT system (Aquilion ONE TSX-301A/2A; Canon Medical Systems, Tochigi, Japan) was used for the CT examinations. The following imaging parameters were used: 120 kVp, 216–541 mAs, rotation time of 0.50 seconds, pitch of 0.60, and slice thickness of 1 mm with a 1-mm reconstruction interval. The noise index [one standard deviation (SD) of the local CT radiation intensity value] for automatic exposure control was set to 10 Hounsfield units. Images were reconstructed using the hybrid-type iterative reconstruction method (adaptive iterative dose reduction 3D enhanced; Canon Medical Systems, Tochigi, Japan) with a mild processing intensity and an FC14 kernel. CT images were acquired with a slice width and spacing of 1 mm. Image data were stored using image archiving and communication systems. A predetermined amount of CM (Iodine per body weight: 360 mgI/kg) was injected over a 30-second period using a power injector. Three different CM agents were used: Iopamidol (Iopamiron; Bayer Schering Pharma, Leverkusen, Germany), iohexol (Omnipaque; General Electric, Boston, MA, United States), and iomeprol (Iomeron; Bracco, Milan, Italy). Scan delays of 40 seconds, 70 seconds, and 180 seconds were used to define the AP, PVP, and EP, respectively.

Image postprocessing

We used both plain and the EP images were considered to generate the CE-boost images (^SURESubtraction Iodine map, Canon Medical Systems, Tochigi, Japan). The CE-boost technique was performed by nonrigid registration and image subtraction. In this study, we subtracted the plain image from the EP image and merged the subtracted image with the EP image to enhance the contrast areas. We used the images with a cross-sectional width and spacing of 1 mm to generate the subtraction images.

Quantitative image analysis

For a quantitative assessment of image quality, the PVP images, EP images, and CE-boost (EP) images were quantitatively analyzed and compared. Comparisons were made between the PVP and CE-boost (EP) images and between the EP and CE-boost (EP) images, respectively. A radiological technologist (Yuji Tachibana with 16 years of experience) performed the quantitative image analysis of whole images. First, the CT radiodensities (in Hounsfield units) of the main portal vein, the right and left branches of the portal vein, the segmental branches of the portal vein, and the inferior vena cava were measured on the PVP, EP, and CE-boost (EP) images respectively. During the measurements, the region of interest was positioned according to the vessel size and was unaffected by surrounding structures.

We calculated the contrast-to-noise ratio (CNR) and the signal-to-noise ratio (SNR) using the following formulas: The CNR = (CT number₀ - CT number_b)/SD_b; the SNR = CT number₀/SD_b, in which the CT number₀ is the mean CT number, the CT number_b is the background CT number, and the SD_b is the background SD. The CT numbers of the erector spinae were used for the background CT number, and the SD values of the erector spinae were used for background noise[16,17].

For a further assessment of the relationship between the dose and the image quality, we calculated figure-of-merit (FOM) values as follows: FOM = CNR²/CT dose index (volume computed tomography dose index). Hepatic parenchyma radiodensities of the PVP, EP and CE-boost (EP) images were then measured. For blood vessels, corresponding CT values were measured and the CNR, SNR, and FOM were derived from these values. For radiodensity measurements, three 150-mm² region of interests were placed on the hepatic parenchyma and their averages were calculated. Images with a slice width and slice spacing of 1 mm were used for these measurements.

Qualitative image analysis

For a qualitative assessment of image quality, we analyzed the PVP, EP and CE-boost (EP) images and compared their values. Two board-certificated radiologists (Kenichiro O and Yoshiki A, with 9 years and 30 years of experience, respectively) independently graded the quality of the PVP, EP, and CE-boost (EP) images for the portal vein and segmental branch vessel visualization, lesion conspicuity, and image noise. The CT data were randomized, and the readers were blinded to the imaging protocol. The analysis was performed using a four-point subjective scale. Vessel visualization was graded as follows: 1 means no blood vessels visible; 2 means blood vessels visible; 3 means good depiction of blood vessels; and 4 means excellent depiction of blood vessels. Lesion conspicuity was graded as follows: 1 means no lesion visible; 2 means lesion visible; 3 means good depiction of lesion; and 4 means excellent depiction of lesion. The image noise in each image was graded on a scale of 1 means unacceptable; 2 means acceptable; 3 means good; and 4 means excellent. Images with a slice width and slice spacing of 1 mm were used to evaluate the blood vessels, lesion conspicuity, and image noise.

Statistical analysis

All values are expressed as the mean ± SD. All of the values used for the quantitative and qualitative image analysis comparisons were performed using the Mann-Whitney U-test. We used the Cohen kappa test to assess the qualitative comparisons of the image analysis and interobserver agreement. The interobserver agreement is expressed as kappa values: Κ values of 0-0.20, slight interobserver agreement; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1.00, almost perfect[18]. Analysis results were considered significantly different at P values < 0.05. All statistical analysis were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface R (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R Commander designed to add functions that are frequently used in biostatistics.

RESULTS

Figure 1 is the flow chart of the study population selection. Of the patients who underwent liver resection, 258 were diagnosed with liver cancer based on the pathology of the resected liver. Of these, we analyzed the cases of the 52 patients who had undergone dynamic imaging of the liver and had no previous surgical history prior to liver resection. The patients’ baseline characteristics are summarized in Table 1.

Figure 1 The patient selection flow chart. CT: Computed tomography; HBIR: Hybrid-type iterative reconstruction; RFA: Radiofrequency ablation; TACE: Transcatheter arterial chemoembolization.

Table 1 The patients’ baseline characteristics.

Characteristics	mean ± SD
Males:Females	37:16
Age (years)	71.63 ± 11.99
Body height (cm)	160.57 ± 9.55
Body weight (kg)	60.65 ± 10.5
BMI (kg/m2)	23.45 ± 3.21
CTDIvol (mGy)	17.38 ± 4.42
SSDE (mGy)	23.92 ± 4.79

BMI: Body mass index; CTDI_vol: Volume computed tomography dose inde; SSDE: Size-specific dose estimates.

Quantitative image analysis

Table 2 provides the results of the comparison of the quantitative image analysis of blood vessels and hepatic parenchyma on the PVP, EP, and CE-boost (EP) images. There were no significant differences in CT values between the PVP and CE-boost (EP) images, with the exception of the inferior vena cava comparison, in which the values were significantly higher for the CE-boost (EP) images. The CE-boost (EP) images showed significant superiority in all comparisons of image noise, CNR, SNR, and FOM, with the exception of the hepatic parenchyma. No significant differences were detected in the CNR or FOM comparisons within the hepatic parenchyma. The comparison of CT values, image noise, CNR, SNR, and FOM between the EP and CE-boost (EP) images consistently favored the CE-boost (EP) images.

Table 2 Comparison of the quantitative image analysis of blood vessels and hepatic parenchyma, mean ± SD.

Characteristics	PVP	EP	CE-boost (EP)	aP value	bP value
CT value (HU)
Main portal vein	185.48 ± 25.51	135.97 ± 18.00	182.53 ± 27.00	0.50	< 0.01
Right portal vein	183.65 ± 24.06	133.65 ± 16.45	179.18 ± 22.47	0.36	< 0.01
Left portal vein	186.21 ± 24.54	134.74 ± 16.69	181.32 ± 23.89	0.36	< 0.01
Segmental branch of the portal vein	177.23 ± 20.31	132.20 ± 15.70	171.61 ± 19.99	0.17	< 0.01
Inferior vena cava	137.11 ± 23.09	122.85 ± 16.61	164.10 ± 24.40	< 0.01	< 0.01
Hepatic parenchyma	120.25 ± 15.67	100.60 ± 12.93	120.36 ± 16.54	0.96	< 0.01
Image noise	18.04 ± 3.33	17.68 ± 3.62	14.15 ± 5.21	< 0.01	< 0.01
Contrast-to-noise ratio
Main portal vein	3.72 ± 1.44	2.08 ± 1.09	4.94 ± 2.66	< 0.01	< 0.01
Right portal vein	3.61 ± 1.40	1.95 ± 1.00	4.69 ± 2.42	< 0.01	< 0.01
Left portal vein	3.77± 1.40	2.04 ± 1.05	4.94 ± 2.70	0.02	< 0.01
Segmental branch of the portal vein	3.26 ± 1.15	1.85 ± 0.83	4.08 ± 1.94	0.03	< 0.01
Inferior vena cava	0.95 ± 1.63	1.30 ± 0.95	3.49 ± 2.30	< 0.01	< 0.01
Hepatic parenchyma	3.07 ± 1.09	1.90 ± 0.83	3.34 ± 1.74	0.62	< 0.01
Signal-to-noise ratio
Main portal vein	10.61 ± 2.30	7.97 ± 1.90	14.48 ± 5.42	< 0.01	< 0.01
Right portal vein	10.50 ± 2.21	7.85 ± 1.84	14.23 ± 5.21	< 0.01	< 0.01
Left portal vein	10.66 ± 2.31	7.94 ± 1.97	14.48 ± 5.53	< 0.01	< 0.01
Segmental branch of the portal vein	10.15 ± 2.13	7.75 ± 1.74	13.62 ± 4.80	< 0.01	< 0.01
Inferior vena cava	7.84 ± 1.98	7.20 ± 1.70	13.03 ± 4.97	< 0.01	< 0.01
Hepatic parenchyma	6.89 ± 1.55	5.90 ± 1.35	9.54 ± 3.44	< 0.01	< 0.01
Figure of merit
Main portal vein	0.94 ± 0.75	0.33 ± 0.49	1.97 ± 2.83	0.01	< 0.01
Right portal vein	0.89 ± 0.75	0.28 ± 0.37	1.72 ± 2.40	< 0.01	< 0.01
Left portal vein	0.98 ± 0.79	0.32 ± 0.40	1.99 ± 2.84	< 0.01	< 0.01
Segmental branch of the portal vein	0.71 ± 0.51	0.24 ± 0.24	1.24 ± 1.32	< 0.01	< 0.01
Inferior vena cava	0.20 ± 0.47	0.15 ± 0.25	1.08 ± 1.87	< 0.01	< 0.01
Hepatic parenchyma	0.68 ± 0.50	0.28 ± 0.23	0.96 ± 1.03	0.67	< 0.01

^aP value of the portal vein phase vs contrast enhancement boost of the equilibrium phase.

^bP value of the equilibrium phase vs ontrast enhancement boost of the equilibrium phase.

PVP: Portal vein phase; EP: Equilibrium phase; CE-boost (EP): Contrast enhancement boost of the equilibrium phase; HU: Hounsfield units.

Qualitative image analysis

Table 3 summarizes the results of the comparison of the qualitative image analysis of blood vessels, lesion conspicuity, and image noise on the PVP, EP, and CE-boost (EP) images. The distribution of subjective image scoring for different protocols by the two radiologists is given in Table 4. Regarding the comparison between the PVP and CE-boost (EP) images, we observed no significant differences in the vessel visualization of the portal vein or segmental branches of the portal vein. However, the CE-boost (EP) images were significantly better in the lesion conspicuity comparisons, and the PVP images were significantly better in the image noise comparisons. In the comparison of the EP images and CE-boost (EP) images, the CE-boost (EP) images were significantly better in the segmental branches of the portal vein and lesion conspicuity, and the EP images were significantly better in image noise. In the portal vein, there were no significant differences.

Table 3 Comparison by qualitative image analysis, mean ± SD.

Characteristics	PVP	EP	CE-boost (EP)	aP value	bP value
Vascular visualization
Portal vein	4 ± 0.00	3.97 ± 0.15	4 ± 0.00	NA	0.16
Segmental branch of the portal vein	3.97 ± 0.15	3.12 ± 0.55	3.88 ± 0.42	0.14	< 0.01
Lesion conspicuity	3.29 ± 0.60	3.21 ± 0.69	3.63 ± 0.51	< 0.01	< 0.01
Image noise	3.85 ± 0.34	3.85 ± 0.34	3.14 ± 0.44	< 0.01	< 0.01

^aP value of the portal vein phase vs contrast enhancement boost of the equilibrium phase.

^bP value of the equilibrium phase vs contrast enhancement boost of the equilibrium phase.

PVP: Portal vein phase; EP: Equilibrium phase; CE-boost (EP): Contrast enhancement boost of the equilibrium phase; NA: Not available.

Table 4 Distribution of subjective image scoring for different protocols by the two radiologists.

Blood vessel visualization (1/2/3/4)	Reviewer 1	Reviewer 2	Agreement (κ)
Portal vein
PVP	0/0/0/52	0/0/0/52	NA
EP	0/0/2/50	0/0/1/51	0.66
CE boost (EP)	0/0/0/52	0/0/0/52	NA
Segmental branch of the portal vein
PVP	0/0/2/50	0/0/1/51	0.66
EP	1/2/39/10	1/4/35/12	0.66
CE boost (EP)	0/2/2/48	0/2/3/47	0.64
Lesion conspicuity (1/2/3/4)
PVP	0/6/21/25	0/6/29/17	0.43
EP	1/6/26/19	1/5/28/18	0.81
CE boost (EP)	0/3/15/34	0/2/14/36	0.38
Image noise (1/2/3/4)
PVP	0/0/10/42	0/0/6/46	0.71
EP	0/0/10/42	0/0/6/46	0.71
CE boost (EP)	0/2/41/9	0/2/40/10	0.84

PVP: Portal vein phase; EP: Equilibrium phase; CE-boost (EP): Contrast enhancement boost of the equilibrium phase; NA: Not available.

With respect to the distribution of subjective image scores by the two radiologists, the values were > 0.60 for all items except lesion conspicuity, which showed substantial or almost perfect agreement. Concerning lesion conspicuity, there was moderate agreement for the PVP images (κ = 0.43), almost perfect agreement for the EP images (κ = 0.81), and fair agreement for the CE-boost images (κ = 0.38). Examples of the clinical PVP, EP, and CE-boost (EP) images used in the analysis are shown in Figure 2.

Figure 2 An example of a clinical image used in the evaluation. A 77-year-old woman with a diagnosis of hepatocellular carcinoma (arrows) based on a pathological diagnosis. The qualitative evaluation results showed that the vessel visualization of the portal vein provided by the contrast enhancement boost of equilibrium phase [CE-boost (EP)] images and that provided by the portal vein phase (PVP) images was comparable. CE-boost (EP) was significantly superior in lesion conspicuity. In this patient’s case, the vessel visualization of the portal vein was rated with an average of 4.00 points on PVP and 4.00 points on CE-boost (EP). The lesion conspicuity was rated with an average of 3.50 points on PVP and 4.00 points on CE-boost (EP). A and B: Portal vein phase images; C and D: Equilibrium phase images; E and F: Contrast enhancement boost of equilibrium phase images.

DISCUSSION

We quantitatively and qualitatively compared PVP images, EP images, and CE-boost (EP) images and observed that in the quantitative comparison of the PVP and CE-boost (EP) images, all of the results were not significantly different or better for CE-boost (EP). The qualitative evaluation showed that the PVP images were superior with respect to image noise, but there was no significant difference with respect to vessel delineation. In addition, the CE-boost (EP) images were qualitatively superior in lesion conspicuity. The results of our comparison of the EP images and CE-boost (EP) images demonstrated that the overall image quality was improved by using the CE-boost technique, although the qualitative image noise increased.

The CE-boost technique is a simple post-processing method that can be applied to images taken at any timing. Several studies have described its usefulness. Yabe et al[19] reported that the combination of the AP and PVP CE-boost (EP) images increases the ability to delineate HCCs < 20 mm. A study by Hou et al[20] revealed that the application of CE-boost technology to PVP can increase the delineation of the portal vein[20]. In addition to the abdominal region, there have been reports of pulmonary vascularization and head and neck vascularization[21-23]. In those reports, the CNR of the targeted vessels and tumors was increased by a factor of 1.48-2.05 by using CE-boost technology. In our present investigation, the CE-boost (EP) images had a CNR that was 2.38 times greater in the portal vein and 1.78 times greater in the hepatic parenchyma compared to the normal EP images, and these data are comparable to those obtained in earlier studies. In addition, the CE-boost (EP) images had a 1.33-times higher CNR for the portal vein and 1.09-times higher CNR for hepatic parenchyma compared to the PVP images. These results indicate that CE-boosted images, in addition to serving the purpose of increasing the contrast of a particular image, may be substituted for images obtained at other time points. This advantage may be useful if applied to other images and body parts, and further studies are expected.

We also observed noticeably low levels of noise in some of the present images (Figure 3). The CE-boost process subtracts images and adds them together to enhance the contrast component[15]. Thus, contrasted blood vessels, tumors, etc. will be enhanced, but if there is some small signal in the image before subtraction, that signal will also be enhanced. We speculate that signals that could not be subtracted for some reason, such as artifacts or noise, remained in the CE-boost images as noise, which affected our qualitative analysis of the image noise. However, the CE-boost performs smoothing on the image during processing, and as a result the overall image noise is reduced and the image quality is improved. In our opinion, CE-Boost is thus superior for qualitative image analyses of lesion conspicuity.

Figure 3 Examples of clinical images of two patients used in the evaluation. A: Images of a 51-year-old woman, equilibrium phase (EP) image; B: Images of a 51-year-old woman, contrast enhancement boost (CE-boost) of EP image (the image noise was reduced by smoothing during the CE-boost process); C: Images of a 78-year-old-man, EP image; D: Images of a 78-year-old-man, CE-boost of EP image, however, after the CE-boost processing, the high-signal noise seen in the EP is emphasized and more prominent in image.

This study has several limitations. It had a retrospective design, and bias may have occurred in the selected subjects. The data were collected from a single center; further validation would require samples from multiple institutions. The sample size was only 52 cases, which limits the generalizability of the study’s results. A more detailed investigation would have been possible had more cases been collected. In addition, there may have been cases in which misregistration artifacts occurred during the processing of the CE-boost method. The CE-boost technique uses subtraction and addition to increase the contrast, but even slight misalignments during automatic alignment may not be detected in the resulting image[22]. There may have been cases in which this misalignment affected the image evaluation.

Moreover, we used only one reconstruction method. If a study similar to the present one is conducted using other reconstruction methods such as deep-learning reconstruction, different results may be obtained. Additionally, the limited enhancement of contrast in the liver parenchyma may limit the effectiveness of the CE-boost in certain diagnostic situations. To improve the robustness of the present findings, more detailed results could be obtained with: (1) Prospective validation in a large and diverse cohort; (2) Additional sensitivity and specificity measures to more comprehensively assess diagnostic accuracy; and (3) Consideration of the application of CE-boost images in other anatomical regions. Finally, there is a lack of a detailed classification of the liver cancers examined. The present patients’ cases included HCC, metastatic liver cancer, and intrahepatic cholangiocarcinoma, which may result in differences in diagnostic performance depending on the tumor. In addition, we were not able to examine the differences in tumor sizes. Our results for small tumors may differ from those obtained with larger tumors[24,25].

CONCLUSION

The image quality of CE-boost (EP) images was comparable to that of conventional PVP images and superior to that of EP images. When images using CE-boost technology are applied to diagnoses, CE-boost (EP) images might provide information comparable to conventional PVP.