European Journal of Cardiovascular Medicine

Because of the widespread clinical use of imaging modalities such as ultrasonography (US), computed tomography (CT), and magnetic resonance imaging (MRI), previously unsuspected liver lesions are increasingly being detected in otherwise asymptomatic patients. The frequency of detection of these lesions varies from modality to modality, with CT scan detection rate is 7.2% - 33%, with MRI it is 10.2% - 34.5%, and with USG, it is 2.3% - 6.2%.1

MRI has many advantages like high soft tissue contrast resolution, multiplanar capability, lack of ionizing radiation and the safety of using particulate contrast media rather than those containing iodine.

Diffusion-weighted imaging (DWI) is a magnetic resonance imaging technique which provides tissue contrast by measurement of diffusion properties of water molecules within tissue. Stejskal and Tanner were the first to describe MR experiment that could be used to observe and measure water diffusion. It uses a pulse sequence                  (T2-weighted spin echo sequence) and two strong motion probing gradients on either side of the 180º refocusing pulse, known as the Stejskal-Tanner sequence.2 Diffusion is expressed in an apparent diffusion coefficient (ADC) which reflects diffusion property unique to each type of tissue.

Diffusion-weighted MRI was first used for the early diagnosis of stroke in neuroradiology.3 Currently, DWI is increasingly used for evaluation of extracranial organs like breast, liver, kidneys, pelvic organs (prostate) as well as whole body imaging in cancer patients. The use of DWI is very promising for the detection and differentiation of benign and malignant lesions, imaging in oncological patients before treatment and for follow-up after treatment.4,5 DWI measurements are quick to perform and do not require administration of exogenous contrast medium.6

As a non-contrast-enhanced technique, DW-MRI is crucial in patients with severe renal dysfunction at risk of developing nephrogenic systemic fibrosis (NSF) after administration of gadolinium-based contrast medium. Although dynamic contrast enhanced examinations have become a routine component of abdominal imaging, high cost/benefit ratio and risk of contrast media side effects remain an issue. Moreover, sometimes it is not possible to distinguish between highly vascular metastases and haemangiomas, even using dynamic examinations.3,7

With recent advances in technology, DW MR imaging is reaching a potential for clinical use in the liver imaging. Restricted diffusion is seen in tissues with high cellularity, e.g. tumors, abscesses, fibrosis and cytotoxic edema whereas unimpeded diffusion is seen in tissues with low cellularity or tissues with disrupted cell membranes, e.g. cysts and necrotic tissues. 

To our knowledge, most of the studies are retrospective in design where selection bias is inevitable. Till date very few Indian studies are available and results are not in consensus. Till date no Indian study is available on usefulness of DWI in differentiation of liver metastasis from primary liver lesions. Some studies are not followed up by histopathological correlation.

This project is a prospective type of observational study, designed to evaluate the contribution of diffusion-weighted imaging (DWI) towards the confident diagnosis and evaluation of liver lesions reducing the need of contrast agents and obviate the use of invasive diagnostic modalities like FNAC or biopsy.

SAMPLE SIZE AND METHODS 

Total 64 patients were evaluated with diffusion-weighted MR imaging at tertiary hospital in Central India. Informed consent was obtained from patients. The final diagnosis was confirmed by histopathological examination or follow up. Multiple liver lesions of similar morphology in a single patient was considered as a single lesion in the study. Clinical history, diagnostic modalities like biopsy, laboratory findings (AFP level etc),  and  other imaging modalities like USG, CT Scan; wherever applicable were  taken into  consideration.

INCLUSION CRITERIA

• Patients referred for MRI with clinically suspected liver lesions.
• Patients referred for MRI with focal liver lesions detected on USG or CT.

EXCLUSION CRITERIA

• Patients having cardiac pacemakers, MRI incompatible prosthetic heart valves, cochlear implants or any incompatible metallic implants.
• Claustrophobic patients.
• Patients who cannot lie down- dyspnoeic patients/ Un-co-operative patients.
• Postoperative cases and recently biopsied cases and recurrent cases are excluded from the study.
• Patients who do not give informed consent to take part in this study.

TECHNIQUE AND TOOLS

Machine: 1.5 Tesla Philips Achieva MRI machine.       Coil: Body Surface coil

Scan parameters:

Table 1: Scan parameters

RECOMMENDED SEQUENCES

Respiratory triggered fat suppresses single shot echo-planar diffusion-weighted imaging (DWI) was performed in axial plane with tri-directional diffusion gradients by using three b values (b=0 s/mm, b=500 s/mm and b=1000 s/mm). The ADC value of each FLL was calculated within a region of interest (ROI) placed in the centre of the assessed FLL, covering more than 50% of its surface. In cases of necrotic FLLs, measurements were taken only in the solid part, trying to avoid inclusion within the ROI of any necrotic part, covering area   > 1 cm.^9,12  T2 weighted  images  were obtained. 

DATA ANALYSIS 

Qualitative analysis of hepatic lesions was done by visual assessment of their signal intensity on diffusion-weighted images at three b values & comparing them with ADC map. Signal intensities of T2WI were also observed for comparison with DWI.  For quantitative analysis, ADC values were recorded for liver lesions. 

Imaging analysis: a lesion was considered¹¹

• Benign: Hyperintense on T2WI and DWI at b = 0 sec/mm2 with a strong signal intensity decreases at b=500 sec/mm² and ADC subjectively higher than liver parenchyma.
• Malignant: mildly to moderately hyperintense on T2WI and on DW images at b=0 sec/mm² and remained hyperintense compared with liver parenchyma at b=500 sec/mm², with an ADC qualitatively lower than that of the surrounding liver.

For statistical calculations,

For statistical calculations, independent sample t-test and independent sample Welch-test for unequal variances were used. ROC curve analysis was done to evaluate the utility of ADC measures for the prediction of malignancy. 

p-value of 0.05 or less was considered statistically significant

AGE-WISE DISTRIBUTION OF PATIENTS

In the present study, out of total 64 patients the mean age amongst the study subjects was 51.42±13.85 years; 41 to 50 years (26.56%) was the most common age group followed by                              51 to 60 years (23.44%). Table 2 below tabulates age wise count of patients followed by histogram chart of the same.

Table 2: Age-wise distribution of patients

Graph 1: Distribution of the study subjects based on the age group

SEX -WISE DISTRIBUTION OF PATIENTS

Table 3: Sex-wise distribution of patients

Graph 2: Sex-wise distribution of patients

The study population had 54.69% female patients and 45.31% male patients. Graphical representation is shown via pie chart below

DISTRIBUTION OF PATIENTS ACCORDING TO DIAGNOSIS

Table 4: Distribution of patients according to diagnosis

Graph 3: Distribution of patients according to diagnosis

In the current study, metastasis (40.63%) was the most common lesion followed by hepatocellular carcinoma (17.19%) and simple cyst(14.06%).Graphical representation for the same is shown below in chart 3.

DISTRIBUTION OF PATIENTS ACCORDING TO BENIGN OR MALIGNANT LESIONS

Table 5: Distribution of patients according to benign or malignant lesions

A pie chart below shows the graphical representation for same.

Graph 4: Distribution of patients according to benign and malignant lesion

In the present study 40 (67.5%) patients had malignant lesions and 24 (32.5%) participants had benign lesions.

DISTRIBUTION OF LESIONS ACCORDING TO LIVER LOBE INVOLVEMENT

Table 6: Distribution of liver lesions according to liver lobe involvement

Graph 5: Distribution of lesions according to liver lobe involvement

In the study, 53.13% lesions were observed in right hepatic lobe, 17.19% in left lobe and 29.69% lesions in both the lobes. A pie chart below shows the graphical representation for same.

AGE-WISE DISTRIBUTION OF DIAGNOSIS OF PATIENTS

Table 7: Age-wise distribution of diagnosis of patients

Graph 6: Age-wise distribution of diagnosis of patients

In the sample size of 64 patients, 40 patients had malignant lesions and  24 patients had benign lesions. Histogram chart for table 9 is shown below.

SEX-WISE DISTRIBUTION OF DIAGNOSIS OF PATIENTS

Table 8: Sex-wise distribution of diagnosis of patients

Graph 7: Distribution of the study subjects based on the diagnosis and gender

In the present study, among total population of 64 patients, 61.54% metastasis and 88.89% simple cyst were found in female population. 54.55% of HCCs were seen in male population. In the study, haemangioma and hydatid cyst were equally seen.

DISTRIBUTION OF LESIONS ACCORDING TO SIZE

Table 9: Distribution of FLLs according to size

Graph 8: Size wise distribution of lesions

In the current study 46 (71.88%) lesions were less than 5.0 cm in size. Average size of lesion was 4.6 cm.

DISTRIBUTION OF LESIONS BY SIZE ACCORDING TO DIAGNOSIS

Table 10: Distribution of FLLs by size according to diagnosis

Graph 9: Distribution of lesion by size according to diagnosis

Graph 9: Distribution of lesion by size according to diagnosis

Table 13 shows that most of the lesions (46 lesions) were less than 5 cm in size. 80.77% metastases were less than 5 cm size. 54.55% HCCs were greater than 5 cm in size. Among total 64 lesions, 9 lesions were greater than 10cm in size. A  bar chart below represents lesion distribution by size.  

MALIGNANT LESION CHARACTERIZATION DEPENDING ON DWI AND ADC SIGNAL INTENSITY

Table 11: Malignant lesion characterization based on DWI and ADC signal intensity

In the present study, all cholangiocarcinoma (3 lesions) showed hyperintense or heterogeneous hyperintense signal intensity on DWI at all ‘b’ values and hypointense/heterogeneous hypointense signal intensity on ADC map.

In the present study, all HCCs (11 lesions) showed hyperintense or heterogeneous hyperintense signal intensity on DWI at all ‘b’ values and hypointense/heterogeneous hypointense signal intensity on ADC map.

None of HCC lesions disappeared at higher ‘b’ value (b=1000sec/mm²) or showed hyperintense signal intensity on ADC map.

In the present study, all metastases (26 lesions) showed hyperintense / heterogeneous hyperintense signal intensity on DWI at all ‘b’ values and hypointense/heterogeneous hypointense signal intensity on ADC map.

None of metastatic lesions disappeared at higher ‘b’ value (b=1000sec/mm²) or showed hyperintense signal intensity on ADC map.

Thus, all cholangiocarcinoma, HCCs and metastatic lesions in the study showed restricted diffusion.

Heterogeneous hyperintense (DWI)/heterogeneous hyperintense (ADC map) signal intensity was seen predominantly in necrotic lesions.

BENIGN LESION CHARACTERIZATION DEPENDING ON DWI AND ADC SIGNAL INTENSITY

Table 12: Benign lesion characterization based on DWI and ADC signal intensity

In the present study, all abscesses (3 lesions) showed heterogeneous hyperintense signal intensity at b=500 sec/mm² and b=1000 sec/mm² and showed heterogeneous hypointense signal intensity on ADC map (restricted diffusion).

All hemangiomas showed hyperintense/heterogeneous hyperintense signal intensity on DWI at all ‘b’ values. However these lesions showed hyperintense or heterogeneous hyperintense signal intensity on ADC map (no restricted diffusion). Persistence of hyperintense signal on higher ‘b’ values could be due to T2 shine through effect (validated by signal intensity on ADC map).

All the hydatid cysts (4 lesions) disappeared at b=1000 sec/mm² and showed hyperintense signal intensity on ADC map (no restricted diffusion).

Out of total 9 simple cysts, 4(44.44%) disappeared at b=500 sec/mm² and all lesions disappeared at b=1000sec/mm² on DWI. All lesions showed hyperintense signal intensity on ADC map (no restricted diffusion/showed facilitated diffusion).

MALIGNANT vs BENIGN LESION CHARACTERIZATION DEPENDING ON DWI AND ADC SIGNAL INTENSITY

Table 13: Malignant vs benign lesion characterization depending on

DWI and ADC signal intensity

In current study population, out of total 40 malignant lesions, all lesions showed hyperintense/heterogeneous hyperintense signal intensity on DWI at higher ‘b’ values and showed hypo/heterogeneous hypointense signal intensity on ADC map which is suggestive of restricted diffusion. 

In case of total 24 benign lesions, 12(50%) lesions showed hyperintense or heterogeneous hyperintense signal intensity at b=1000sec/mm² on DWI. 12(50%) lesions disappeared at higher b values (b=500sec/mm² and/or b=1000sec/mm²) on DWI. 3(12.5%) lesions showed heterogeneous hypointense signal intensity on ADC map which correspond to 3 abscesses showing restricted diffusion. Rest of benign lesions (~87.5%) showed hyperintense/heterogeneous hyperintense signal intensity on ADC map; suggestive of facilitated diffusion.

CASE OF CHOLANGIOCARCINOMA

Figure 1: A) Hyperintense lesion on b=0 sec/mm² B) Htrg-hyperintense lesion on b=500 sec/mm² C) Htrg-hyperintense lesion on b=1000 sec/mm² D) Htrg-hypointense lesion on ADC map, ADC=0.916x 10^-3 mm²/sec E) Mild Hyperintense on T2WI. Final diagnosis was confirmed by FNAC, IHC

CASE OF HEPATIC ABSCESS

Figure 2: A) Hyperintense lesion on b=0 sec/mm² B) Htrg-hyperintense lesion on b=500 sec/mm²  C) Htrg-hyperintense lesion on b=1000 sec/mm² D) Htrg-hypointense lesion on ADC map, ADC=0.1049 x 10^-3 mm²/sec (average of ADCs in centre & periphery of lesion)  E) Hyperintense on T2WI.

CASE OF HEPATIC HYDATID CYST

Figure 3: A) Hyperintense lesion on b=0 sec/mm²B) Htrg-hyperintense lesion on b=500 sec/mm² C) Iso to hypointense lesion on b=1000 sec/mm²                     D) Hyperintense lesion on ADC map, ADC=1.962 x 10^-3 mm²/sec                                                         E) Hyperintense on T2WI. Intralesional daughter cysts seen.

CASE OF HEPATIC HEMANGIOMA

Figure 4:A) Hyperintense lesion on b=0 sec/mm²                                 B) Htrg-hyperintense lesion on b=500 sec/mm² C) Htrg-hyperintense lesion on b=1000 sec/mm² D) Htrg-hyperintense lesion on ADC map, ADC=2.226x 10^-3 mm²/sec     E) mild Hyperintense on T2WI.

CASE OF HEPATIC SIMPLE CYST

Figure 5: A) Hyperintense lesion on b=0 sec/mm²                               B) Hyperintense lesion on b=500 sec/mm² C) Hypointense lesion on b=1000 sec/mm²                         D) Hyperintense lesion on ADC map, ADC=1.973 x 10^-3 mm²/sec                                                        E) Hyperintense on T2WI.

For the current study total of 64 patients were investigated. Final diagnosis was made by biopsy/FNAC and follow up. Clinical history, laboratory findings (AFP level etc), and other imaging modalities like USG, CT Scan; wherever applicable were taken into consideration for final diagnosis of liver lesions. Multiple liver lesions of similar morphology in a single patient was considered as a single lesion in the study.

In current study, majority of patients were in age group of 41-60 years (50%). Mean age of patients was 51.42 years. In the study population, 29% were male patients and 35% were females..

Table 14: Sex and age distribution comparison between  current study and similar studies

Table 14 shows that current study is comparable in terms of age and sex distribution with similar studies in literature referred.

LESION DETECTION ON DWI

In the current study, all focal liver lesions showed higher signal intensity compared to surrounding liver parenchyma, predominantly on lower b value DWI images, enabling lesion detection.

Out of 64 patients, 40 patients (62.5%) had malignant lesions and 24 patients (37.5%) population had benign lesion/s. In the study most common diagnosis of patients was metastases 26(40.63%) followed by HCC 11 (17.19%) and simple cyst 9(14%). 

In the study 53.13% lesions were observed in right hepatic lobe, 17.19% in left lobe and 29.69% lesions in both the lobes.

In the current study 54.5% of HCCs, 100% Cholangiocarcinoma and 100% abscesses were seen in male patients.61.5% of metastases and 88.8% of simple cysts were seen in female patients. Haemangioma and hydatid cysts were equally seen in males and females. Among 64 lesions, 26 lesions were metastases followed by HCC (11 lesions) and simple cyst (9 lesions).

All liver lesions (64) were classified into three groups based on their size (in cm). 46 lesions (71.8%) were <5cm, among which 21 were metastases. 9 lesions were greater than10cm in size. Average size of lesions was 4.6cm.

Table 15: Comparison of number & size of focal liver lesions between current study & other similar studies

Table 15 depicts that current study closely matches with similar studies in literature studied.

MALIGNANT LESION CHARACTERIZATION ON DWI & ADC MAP: QUALITATIVE ANALYSIS

In the     present study, all malignant lesions showed hyperintense/heterogeneous hyperintense signal intensity on DWI at all ‘b’ values and hypointense/heterogeneous hypointense signal intensity on ADC map. Thus, all HCCs, cholangiocarcinoma and metastatic lesions in the study showed restricted diffusion due to high cellularity.

These results are comparable with Hosny et al⁷ who also demonstrated persistent hyperintense signal of solid tumors on high b value on DWI. Haradome et¹⁰ al and Parikh et al¹¹ demonstrated that lesion is considered as malignant if it showed hyperintense signal at b=500sec/mm² & an ADC qualitatively lower than that of surrounding liver. Current study results concur with the same.

BENIGN LESION CHARACTERIZATION ON DWI AND ADC MAP: QUALITATIVE ANALYSIS

In this study, all 8 hemangiomas showed hyperintense/heterogeneous hyperintense signal intensity on ADC map (no restricted diffusion). Persistence of hyperintense signal on higher ‘b’ values could be due to T2 shine through effect (validated by signal intensity on ADC map).^6,14,15,16

All simple cysts disappeared at b=1000sec/mm² on DWI and showed hyperintense signal intensity on ADC map. Inan et al¹⁵ found that, most simple cysts (93%) were isointense with liver at b=1000sec/mm²and showed hyperintense signal on ADC map. Current study results are comparable with Inan et al¹⁵ & Hosny.⁷                            

All the hydatid cysts (2 lesions) disappeared at b=1000 sec/mm² and showed hyperintense signal intensity on ADC map (no restricted diffusion).

All abscesses (3 lesions) showed heterogeneous hyperintense signal intensity at b=500 sec/mm² and b=1000 sec/mm² and showed heterogeneous hypointense signal intensity on corresponding ADC map represented restricted diffusion. This is due to high viscosity in the abscess core. Chan et al¹⁴ showed that all abscesses showed hyperintensity on DWI and hypointensity on ADC maps. Current study results concur with the same.

Limitations:
Sample size and duration of study was small. Being oncology setup, number of malignant lesions was more than benign lesions. Solid benign FLLs (hemangioma) not encountered in this study. Paediatric population was excluded from study.

Qualitative and quantitative assessment of liver lesions was done based on only diffusion-weighted MR imaging and T2WI. Comparison with contrast enhanced MR imaging was not done

1. Chiorean L, Cantisani V, Jenssen C, Sidhu PS, Baum U, Dietrich CF. Focal masses in a non-cirrhotic liver: The additional benefit of CEUS over baseline imaging. Eur J Radiol 2015.
2. Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J Chem Phys 1965;42:288–292.
3. Demir Öİ, Obuz F, Sağol Ö, Dicle O. Contribution of diffusion weighted MRI to the differential diagnosis of hepatic masses. Diagn Interv Radiol 2007;13:81-86.
4. Qayyum A. Diffusion-weighted imaging in the abdomen and pelvis: Concepts and applications. RadioGraphics 2009;29:1797-1810.
5. Malayeri AA, El Khouli RH, Zaheer A, Jacobs MA, CoronaVillalobos CP, Kamel IR, Katarzyna JM. Principles and applications of diffusion-weighted imaging in cancer detection, staging, and treatment follow-up. RadioGraphics 2011;31:17731791.
6. Koh D-M, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR Am J Roentgenol 2007;188:1622-1635.
7. Hosny IA. Diffusion MRI of focal liver lesions. PJR 2010; 20:01-07.

8. Rao UMM, Rao LC. Evaluation of focal liver lesions by magnetic resonance imaging and correlation with pathology. Int J Res Med Sci 2016;4:4659-68.
9. Caraiani CN, Marian D, Militaru C, Calin A, Badea R. The role of the diffusion sequence in magnetic resonance imaging for the differential diagnosis between hepatocellular carcinoma and benign liver lesions. Clujul Med 2016;89:241-249 [PMID: 27152076 DOI: 10.15386/cjmed-567]
10. Haradome H, Grazioli L, Morone M, Gambarini S, Kwee TC, Takahara T, Colagrande S, T2-weighted and diffusion-weighted MRI for discriminating benign from malignant focal liver lesions: Diagnostic abilities of single versus combined interpretations. J. Magn. Reson. Imaging 2012;35:1388-1396.
11. Parikh T, Drew SJ, Lee VS, Wong S, Hecht EM, Babb JS, Taouli Focal liver lesion detection and characterization with diffusion weighted MR imaging: comparison with standard breath-hold T2weighted imaging. Radiology        2008;246 (3):812– 822.
12. Filipe JP, Curvo-Semedo L, Casalta-Lopes J et al. Magn Reson Mater Phy 2013; 26:303. https://doi.org/10.1007/s10334-012-0348-1.
13. Kim T, Murakami T, Takahashi S, Hori M, Tsuda K, Nakamura H. Diffusion-weighted single-shot echoplanar MR imaging for liver disease. AJR Am J Roentgenol 1999 Aug;173:393–398.
14. Chan JHM, Tsui EYK, Luk SH, Fung ASL, Yuen MK, Szeto ML, Cheung YK, Wong KPC. Diffusion-weighted MR imaging of the liver: distinguishing hepatic abscess from cystic or necrotic tumor. Abdom Imaging 2001;26:161–165.
15. Inan N, Arslan A, Akansel G, Anik Y, Sarisoy HT, Ciftci E, Demirci A. Diffusion-weighted imaging in differential diagnosis of simple and hydatid cysts of liver. AJR Am J Roentgenol2007;189:1031-1036.
16. Inan N, Kilinc F, Sarisoy T, Gumustas S, Akansel G, Demirci A. Diffusion-weighted MR imaging in the differential diagnosis of hemangiomas and metastases of the liver. Radiol Oncol 2010;44(1):24-29.