3-T Magnetic Resonance Diffusion-Weighted Imaging (DWI) for Characterization of Hepatic Masses

Objective: To evaluate the role of DWI and ADC in differentiating benign from malignant hepaticmasses Materials and methods: Forty-one patients with fifty-two focal hepatic masses were included in our study. MRI was done using coronal T2-weighted single-shot turbo spin-echo, breath hold axial 3-D gradient-echo, breath hold 2D gradient-echo in and out-of-phase, respiratory-triggered axial turbo spin-echo T2 sequence with fat saturation, followed by free breathing Diffusion-weighted MR imaging using a single-shot spin-echo echo planar imaging sequence and finally triphasic -MRI. Results: Forty-one patients (Fifty-two lesions) were included in our study. Twenty-three lesions were benign; eight of them were cysts (mean ADC values of 3.15 ± 0.34 × 10-3 mm2/s) and fifteen lesions were hemangiomas (mean ADC values of 2.10 ± 0.25 × 10-3 mm2/s). Twenty-nine lesions were malignant; twelve HCC lesions (mean ADC values of 1.10 ± 0.32 × 10-3 mm2/s) and seventeen masses were metastasis (mean ADC values of 0.96 ± 0.23 × 10-3 mm2/s). Sensitivity and specificity of DWI in differentiating malignant from benign hepatic masses were 96.6% and 95.7% respectively. Conclusion: DWI is an easy technique to obtain and to be evaluated. ADC values can differentiate benign from malignant liver masses with high sensitivity and specificity.


Introduction
The progress of imaging modalities increases the sensitivity of hepatic focal lesions detection during routine radiological examination. Benign hepatic focal lesions arises usually on top of noncirrhotic liver. Hemangioma is the most common benign hepatic lesions followed by focal nodular hyperplasia and adenomas [1,2]. Metastasis is the commonest malignant hepatic lesion arising on top of non-cirrhotic liver. On the other hand, hepatocellular carcinoma and intra-hepatic cholangiocarcinoma arise more on top of chronic liver diseases. The progression of imaging modalities including functional imaging increased accuracy characterization of hepatic lesions, therefore decreased the rate of unnecessary biopsies for benign lesions which is related to high incidence of morbidity (2.0% to 4.8%) and mortality (about 0.05%) [3,4]. Ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI) are the most widely used hepatic imaging modalities. Most of the published studies concluded their similar diagnostic accuracy in the assessment of focal hepatic lesions with no significant difference in their specificities ranging from 82%-89% [4,5].
Over the last few years, several published series have studied the role of DWI in the evaluation of different hepatic lesions. DWI is more sensitive than T2-weighted sequences [6][7][8][9] and than different super paramagnetic iron oxide, enhanced MR studies in the detection and accurate characterization of liver lesions [10]. DWIs has a great role in detection and proper characterization of small focal hepatic lesions which makes DWI superior than other modalities [6,8,10]. High SNRs and high lesion-to-liver signal intensity ratios are seen better at low b values alleviating the depiction of focal liver lesions. Also, black blood effect of diffusion-weighted images help in differentiating small focal lesions from intra-hepatic vessels. DWI is helpful in early detection of small hepatocellular carcinomas (HCCs) in patients with liver cirrhosis due to less clarity of parenchymal heterogenecity and less signal intensity of the cirrhotic liver related to regeneration nodules and fibrosis on DWI in comparison to other T2-weighted images [11]. The differentiation between malignant and benign focal hepatic lesions is a diagnostic dilemma. Therefore, in a trial to improve the characterization of hepatic lesions, new evolving techniques of preexisting modalities, such as MRI, computed tomography (CT) and ultrasonography are being increasingly developed. Recently, some researches have noted that the apparent diffusion coefficient (ADC), which is one of calculated parameters of DWI, is a useful new technique in differentiating benign and malignant lesions in the liver [12].
The diffusion coefficient is measured by the degree of molecular mobility of water particles therefore, it varies according to tissue properties including size of the extracellular space (which reflects the rate of unhindered moving water protons), viscosity and cellularity [13]. Measuring diffusion coefficients has been shown to be helpful for the characterization of focal and diffuse diseases of the liver [14,15].

Objective
The aim of our study was to assess the role of DWI and the ADC in differentiating between benign and malignant hepatic masses.

Patient populations
Between February 2011 and June 2013, Forty-one patients (25 males and 16 females with age range, 21-67 years; mean age, 49 years) with focal hepatic masses were included in our study. We apply our study on all patients referred to our department for MRI of the liver at the period of the study. Patients were imaged using conventional MRI, DWI and triphasic-MRI before biopsy of liver masses. Written informed consent was obtained from all patients before MRI.
Diffusion-weighted MR imaging was performed before dynamic imaging, using a single-shot spin-echo Echo Planar Imaging (EPI) sequence with the b factors of 50, 400, and 800 s/mm 2 along the three orthogonal directions. To improve the signal-to-noise ratio and for patient convenience, DWI under free-breathing was performed. The sequence was obtained free breathing using the following technical parameters TR=6200 ms, TE=95 ms, matrix=125 × 192, slices thickness=5 mm with inter-slice gap=20%, FOV=40 cm, and the average=4.
Acquisition time for the entire liver using the three different b factors was about 3.0 min. The parallel imaging algorithms (GRAPPA), with an acceleration factor of 2, were added to reduce the acquisition time. Spectral fat saturation was employed systematically to suppress the chemical-shift artifacts. ADC maps regarding isotropic images were automatically acquired and all mean ADCs of the lesions were measured on those maps.
After determination of the optimal timing for the arterial phase imaging, using the timing bolus technique, dynamic contrastenhanced imaging was obtained after the administration of a bolus injection of the gadopentetatedimeglumine (0.1 mmol/kg of Magnevist), at a rate of 2 mL/s. We used a 3D gradient echo sequence (VIBE) with ultrafast image reconstruction, using the parallel imaging algorithms ( A dynamic series consisted of one pre-contrast series, followed by three successive post-contrast series, including an early arterial, late arterial, and portal phase imaging, at 31 s intervals (17 s for image acquisition with breath-holding and 14 s for re-breathing), at the start of each phase imaging followed by a 5-min delayed phase imaging. Triphasic MRI was done in 38 patients and was contraindicated in 3 patients with renal failure on haemodialysis and GFR less than 30 ml/ min. Benign appearing focal lesions were subjected to follow up for 2 years to assess the stability of the size and appearance of the lesions (after 6 months then after 1 years for cysts and more setting after 2 years follow up for haemangioma).
Biopsy was done for malignant appearing lesions and correlated with the MRI findings.

Results
Fifty-two lesions were detected in the 41 patients included in our study. Twenty-nine lesions were malignant, 12 lesions were HCC and 17 lesions were metastasis according to the final histopathological study and Twenty-three lesions were benign: 8 hepatic cysts and 15 hemangioms according to US, CT, MRI findings and follow-up.  According to the ADC values (Table 1) twenty three lesions were benign with 8 lesions were hepatic cysts ( Figure 1) with mean ADC values of 3.15 ± 0.34 × 10 -3 mm 2 /s and ADC range of 2.61-3.40 × 10 -3 mm 2 /s and fifteen lesions were hemangioms (Figure 2) with mean ADC values of 2.10 ± 0.25 × 10 -3 mm 2 /s and ADC range of 1.60-2.50 × 10 -3 mm 2 /s. In our study twenty nine lesions were malignant: twelve lesions were HCCs (Figures 3 and 4) and showed mean ADC values of 1.10 ± 0.32 × 10 -3 mm 2 /s and ADC range of 0.91-1.76 × 10 -3 mm 2 /s and seventeen lesions were metastasis ( Figure 5) with mean ADC values of 0.96 ± 0.23 × 10 -3 mm 2 /s and ADC range of 0.75-1.25 × 10 -3 mm 2 /s.

Discussion
Diffusion describes the random (Brownian) motion of water molecules. With a very strong bipolar gradient pulse inserted into either a spin-echo pulse sequence (i.e., Stejskal-Tanner technique) or a gradient echo pulse sequence, MR imaging can be sensitive to the diffusion of water molecules in the tissue [16]. Diffusion restriction increases in highly cellular tissues and decreases in low cellular tissues with large extracellular space or with broken-down cellular membranes [17].
In our study, there is a significant difference in the ADC values between malignant (HCC and metastasis) hepatic masses and benign (cysts and hemangiomas) hepatic masses with low ADC in malignant masses and high ADC value in benign masses ( Figure 6). The result of our study is in agreement with other studies had been published concerning the diffusion properties of focal hepatic lesions. Most of the studies revealed that ADC values of benign lesions (cysts and hemangiomas) were significantly higher than those of malignant lesions attributed to high cellularity of malignant masses [14,18,19].
In our study, the benign hepatic focal lesions had a mean ADC values of about 2.46 ± 0.28 × 10 -3 mm 2 /s and ADC range of about 1.60-3.40 × 10 -3 mm 2 /s and the malignant hepatic focal lesions showed a mean ADC values of about 1.02 ± 0.26 × 10 -3 mm 2 /s and ADC range of about 0.75-1.76 × 10 -3 mm 2 /s. Several studies have identified significantly lower ADC values in malignant compared to benign focal liver lesions which coincide with our results [9,14,20].
ADCs tend to be larger when using small b-values, because the signal attenuation due to diffusion plays only a minor role in that case, and ADC values are contaminated by micro-perfusion. When higher b-values are used, ADCs tend to decrease, in relation with less perfusion contamination. Also, the ADC measurements of benign and malignant hepatic masses were significantly different, which supports similar previous findings with cysts and hemangiomas had the highest ADC values while malignant masses had the lowest ADC values [21][22][23][24]. In the liver, b values of 0 and 500-600 s/mm 2 are typically used. Although at least two b values are required for diffusionweighted imaging analysis, the application of a greater number of b values will improve the accuracy of the calculated ADC. The disadvantage of using multiple high b values is an associated increase in scanning time [25,26].
Quantitative measurement of ADC has been shown to be an indicator of malignancy in focal liver lesions, with a reduction in mean ADC (low signal intensity on an ADC map) of malignant lesions compared with benign lesions [9]. Bruegel et al. [9] reported that an ADC threshold of 1.63 × 10 -3 mm 2 /sec could be used to correctly characterize 88% of lesions as either benign or malignant. Other study using a threshold ADC value of 1.5 × 10 -3 mm 2 /s was able to differentiate benign from malignant lesions with 84% sensitivity and 89% specificity. Potential limitations will include necrotic and cystic metastases, where ADC might be elevated, and the diagnosis will then rely on post-contrast images [14].
In our study, we avoid to measure ADC in necrotic parts of the mixed tumor and depending only in measurement of solid component of mixed hepatic masses. A cut-off point of ADC value of 1.65 × 10 -3 mm 2 /s was able to differentiate between malignant and benign hepatic masses with 96.6 % sensitivity, 95.7% specificity, and total accuracy of about 96.2 %.

Conclusion
In conclusion, DWI is easy to obtain and easy to evaluate, and ADC values can differentiate between benign and malignant liver masses with high sensitivity and specificity 96.6% and 95.7%, respectively. DWI is problem solving sequence in patients with contraindications to contrast media. We recommend addition of DWI sequence to the standard MRI examination of the liver before contrast study and according to the result of our study it is a promising sequence. However, more studies with more variants of hepatic focal lesions are needed for more evaluation of DWI in characterization of hepatic masses.