Detection and Characterization of Focal Liver Lesions by Split-bolus Multidetector-row CT: Diagnostic Accuracy and Radiation Dose in Oncologic Patients

Patients and Methods

Study design and patients. A retrospective analysis of split-bolus MDCT data of patients consecutively referred from January 2012 to December 2013 was performed in our study. Written informed consent for split-bolus MDCT protocol was obtained from each patient.

Forty-eight oncologic patients (23 males and 25 females; age range=42-80 years, mean age=63.3 years; weight range=60 and 75 kg, mean weight=68 kg), with 296 focal liver lesions detected by an initial triphasic-MDCT examination, underwent split-bolus MDCT scanning of the chest, abdomen and pelvis 6-12 months after their initial CT examination as a part of their follow-up. No patient had cardiovascular insufficiency or liver cirrosis.

All patients had known extrahepatic histologically-confirmed primary cancer: colorectal cancer (n=13), lung cancer (n=15), breast cancer (n=9), pancreatic cancer (n=3), gastric cancer (n=7) and melanoma (n=1).

To determine the benign or malignant nature of a focal liver lesion, we considered the typical features at triphasic-MDCT, at ultrasound or magnetic resonance imaging and their behavior on sequential imaging studies.

Because the patients had chemotherapy after initial triphasic-MDCT, in order to perform a correct comparison of the results of triphasic (representing the patient baseline study) and split-bolus (representing the follow-up evaluation) MDCT techniques, we only included focal liver lesions unchanged in size and appearance (benign, malignant and indeterminated lesions) in our analysis. Malignant lesions that grew, became smaller or disappeared over time at split-bolus MDCT follow-up evaluation were excluded.

Scanning protocols. At initial triphasic and follow-up split-bolus MDCT, all patients were scanned with a Philips Brilliance 64-detector row scanner (Philips Healthcare, Best, the Netherlands). Triphasic-MDCT: This protocol consisted of unenhanced images and a HAP (started 35-40 s after the injection of the contrast medium or 20 s after threshold of 150 HU in the thoraco-abdominal aorta by using bolus tracking technique) of the upper abdomen, a PVP started 40 s after the end of HAP acquisition of the chest, abdomen and pelvis and a DP (5 minutes after the start of injection) of the upper abdomen. Nonionic iodinated contrast medium (Iopamidolo, Iopamiro 370 mgI/ml; Bracco, Milano, Italy and Iopromide, Ultravist 370 mgI/ml Schering AG, Berlin, Germany) was injected in a quantity ranging from 120 ml to 150 ml at a rate of 4 ml/s from an antecubital vein using an 18-gauge needle.

Split-bolus MDCT: This protocol consisted of unenhanced low-milliampere scans of the upper abdomen and a single acquisition of the chest, abdomen and pelvis after i.v. injection of a maximum of 150 ml of contrast medium (as described above), split by an automatic power injector (Stelland CT; Medrad, Indianola, PA, USA) into two boli (Figure 1).

The dose of contrast medium for the first bolus, injection duration and times were established on the basis of literature data (10-12) and broad clinical experience. For our patients (weight ranging from 60 to 75 kg), we obtained an adequate hepatic enhancement during PVP by administering the first bolus of contrast medium (1.2 ml/kg) at a fixed flow rate of 2 ml/s, and with an injection duration ranging from 37 to 45 s. In addition, optimal patient bodyweight-tailored dose of contrast medium for the liver was determined based on a definition of the dose which can give adequate hepatic enhancement (more than 50 HU increase from unenhanced baseline HU of the liver) during PVP (13, 14). HAP occurs about 35 s after the start of i.v. injection of a bolus of 60 ml of contrast medium at 3.5 ml/s.

Using the scout film, a scan range from the pulmonary apex to the pubic symphysis was determined. Then a circular region of interest of the bolus tracking technique (raising the threshold value at 500 HU) was placed in the descending aorta. Approximately at the end of the second bolus injection of contrast medium, the scan started cranio-caudally after a delay of at least 6 s from the arrival of the contrast medium in the aorta. The inherent 6 s delay in the bolus tracking technique is necessary in order to move the scan table to the start of the scan, give breath-hold instructions to the patient, and tune the gantry parameters.

A single contrast-enhanced acquisition was carried out, resulting in a simultaneous contrast enhancement of the arterial and venous systems.

For triphasic and split-bolus MDCT protocol, the following acquisition parameters were used: slice thickness 2.5 mm; gantry rotation speed 0.75 s; reconstruction index 1.25; pitch: 0.935:1; 120 kVp and 200-400 mAs. Automatic tube current modulation was used to minimize radiation exposure. MDCT examinations were completed with sagittal, coronal and curved multiplanar reconstructions.

Images were transferred to an external workstation (Advantage Workstation 4.2 from GE Healthcare, Milwaukee, Wisconsin, USA; and Magic View from Philips Medical Systems, Best, the Netherlands) and stored in image archiving and communication system (PACS; Agfa Impax RIS Healthcare, Mortsel, Belgium, EU).

Image analysis. All CT images were transported from local digital media to a viewing station (Advantage Workstation 4.2; Magicview) and were blindly reviewed by two radiologists (M.S., A.D.A) with more than 10 years of experience in interpreting body CT images, and who independently analyzed triphasic and split-bolus CT.

Attenuation of the aorta, main portal vein and right lobe of the liver were measured by a circular region of interest without major vascular structures, at combined HAP/hepatic enhancement during PVP on split-bolus scans and at HAP and PVP on triphasic-MDCT scans.

The appearance of each lesion at split-bolus MDCT was described on the basis of the attenuation and homogeneity of the lesion in comparison with aorta, relating to the triphasic CT findings (2, 4, 15-18).

After describing the state of each lesion in each phase, the pattern of enhancement over time of each lesion was described as a three-part pattern name that incorporated the appearance of the lesion in each phase (HAP, PVP, DP) of the triphasic CT technique (e.g. hypo-/hypo-/hypo-) and as a two part pattern name that incorporated the appearance of the lesion in the combined HAP/hepatic enhancement during PVP and DP of split-bolus protocol (e.g. hypo-/hypo-).

The comparative analysis between the two protocols was performed only on lesions that did not change in number and size between the initial triphasic and follow-up split-bolus MDCT.

The radiation dose to the patient and the number of images to be analyzed with split-bolus (including DP) and triphasic-MDCT protocol were determined and the results were compared.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Schematic view of split-bolus 64-detector row computed tomographic scanning of the chest and abdomen shows contrast medium administration splitted into two-bolus injections in adult male (weighed 75 kg). First bolus [at the start of bolus injection (or time zero): 90 ml (1.2 ml/kg) of contrast medium at 2.0 ml/s, followed by 20 ml of saline solution at same flow rate, is injected to obtain adequate hepatic enhancement during the portal venous phase; second bolus: 60 ml of contrast medium at 3.5 ml/s followed by 20 ml of saline solution at the same flow rate to obtain hepatic arterial phase. A circular region of interest of the bolus-tracking technique (raising the threshold value at 500 HU) was placed in the descending aorta. At the start of the second bolus contrast medium injection, the scan starts cranio-caudally after a delay of at least 6-seconds from arrival of the contrast medium into the aorta, and a single contrast-enhanced acquisition from the pulmonary apex to the pubic symphysis was carried-out, resulting in a simultaneous contrast enhancement of the arterial and venous systems. Tarr: Time of arrival of contrast medium at the aorta.

Dose radiation calculation. The dosimetric evaluation was carried out by analyzing the triphasic and split-bolus MDCT examinations of the 48 patients included in the study. Radiation dose was based on volume CT dose index (CTDIvol) and dose-length-product (DLP).

Biphasic split bolus and triphasic-MDCT technique were compared by examining the effective dose (ED; in Sv) which was calculated using the following equation:

ED= k × DLP where k=0.015 (conversion coefficient) (19).

Statistical analysis. All the statistical analysis was performed using commercially available software (SPSS, release 17.0 SPSS Inc,; Chicago, IL, USA). Interobserver agreement between the two radiologists on image quality, detection and characterization of focal liver lesions was evaluated using weighted k statistics, with a k value on the following: k value <0.2, poor agreement; 0.2≤k≤0.4, fair agreement; 0.41≤k≤0.60, moderate agreement; 0.61≤k≤0.80, substantial agreement; k≥0.80, excellent agreement (20, 21). For the detection and characterization of focal liver lesions, sensitivity, specificity, positive predictive value and negative predictive value were also assessed. For comparative analysis of the averages and standard deviations (SDs) of attenuation CT values, Student's t-test was used: a p-value less than 0.05 was considered statistically significant.