RELIABILITY OF DYNAMIC CONTRAST ENHANCED MRI IN DIAGNOSIS OF HEPATIC TUMORS WITH PATHOLOGIC CORRELATION

By

Mostafa M. Shakweer*, Abdelshafy A. Awadalla*, Ahmad M. Mostafa*, and Abdel-Nasser A. Ghareeb**

Radiology Departments, Al-Azhar Faculty of Medicine  (Damietta* and Cairo**)

ABSTRACT

Background: In recent decades, it is clear than any previous era that the morbidity and mortality is increasing in patients with liver tumors. This is either due to liver cirrhosis or metastatic secondaries from distant malignant tumors.  With advanced development of diagnostic tools, it becomes more available to detect early the prevalence of such tumors and distinguish benign from malignant to modify the wellbeing and healthcare of many patients. 

Objective: The study was carried out to evaluate reliability of dynamic contrast enhanced MRI in diagnosis of hepatic tumorswith pathologic correlation to achieve high accuracy diagnostic confirmation prior to the targeted treatment.

Patients and Methods: Forty patients were included, 25 males and 15 females. The age of the patients ranged from 26 to 77 years old and one patient was five months, with mean age ± S.D. “50.86 ± 15.46”. They were referred from Tropical, General Medicine, General Surgery Departments as well as out-patient clinics to Radiology Department, New Damietta, Al-Azhar University Hospital during the period from August, 2011 to December, 2012.   The patients underwent laboratory investigations and ultrasonography as a screening survey and then subjected to full MRI study including pre contrast imaging (T1 and T2 sequence),  dual-echo (in and out) phase, diffusion-weighted images, heavy T2 and dynamic study.    Histopathological study was done as confirmatory for the diagnosis.  All examinations were done after obtaining  informed consent, and parent’s consent was obtained for the infant aged 5 months.

Results: The study showedthat the common hepatic tumor was HCC { solitary lesions (8 cases, 20%) and multicentric HCCs (6 cases, 15%)} followed by metastasis (8 cases, 20%), hemangioma (7 cases, 17.5%), dysplastic nodules (3 cases, 7.5%), cholangiocarcinoma and focal nodular hyperplasia (2 cases for each, 5%), hepatoblastoma, adenoma,  regenerative  nodule and benign lesions (1 case for each, 2.5%).   Thirty three cases (82.5%) were prepared for histopathological correlation; The remaining 7 cases (17.5%) were excluded for biopsy which were diagnosed by MRI as hemangiomas and underwent follow up.

Conclusion: Magnetic resonance imaging has the advantage of achieving high resolution images of the liver without the use of ionizing radiation in diagnosis of liver tumors, and pathological study considered a mandatory final diagnostic tool for the conflict cases.

INTRODUCTION

     The greatest range of benign and  malignant disease affects the liver.  As a result, much researches and many tools developed for a variety of imaging modalities geared toward improvement of visualization of liver diseases (Richard et al., 2011).   Benign liver masses may be found in more than 20% of the general population (Amital et al., 2008).   Benign hepatic lesions include hepatic cysts, hemangiomas, adenomas, and focal nodular hyperplasias (Devaki et al., 2012).

    Malignant hepatic lesions are either primary or secondary. Primary lesions include regenerative, dysplastic nodules, hepatocellular carcinoma, fibrolamellar hepatocellular carcinoma and cholangio-carcinoma, while secondary lesions include hypovascular, hypervascular and hemorrhagic lesions  (Vilgrain et al., 2005).    Hepatocellular carcinoma (HCC) is the most common primary malignant tumor of the liver due to excessive alcohol intake, chronic hepatitis or primary biliary cirrhosis.   Imaging  plays a central role in the management of hepatocellular carcinoma including screening popula-tions at risk, confirming the diagnosis, planning treatment, guiding therapy, and follow up after treatment (Peterson and Baron, 2007).

      MR imaging is establishing a role as a primary diagnostic technique with increasing evidence showing MR imaging to have advantages over CT regarding diagnostic sensitivity and specificity for many pathologies of solid organs, bile and pancreatic ducts, bowel, peritoneum, and retroperitoneum. In addition, there are increasing concerns regarding the risks of radiation and iodinated contrast associated with CT imaging of the abdomen (Diego et al., 2010 and Saima et al.,2011).   MRI has many advantages, e.g. high contrast resolution, multiplaner images, lack of ionizing radiation, and the safety of using particulate contrast media rather than those containing iodine. Lesions morphology, signal intensity, and contrast enhancement pattern are taken into consideration when characterizing masses with MRI (Demir et al., 2009). It leads to significantly better detection of hepatic focal lesions following improvements in technology and techniques (Arguedas, 2007).

     The aim of this study was to evaluate the reliability of dynamic contrast enhanced MRI in diagnosis of hepatic tumors with histopathologic correlation.  

PATIENTS AND METHODS

Patients: This research was performed at Radiology Department, (New Damietta), Al-Azhar University Hospital during the period from August, 2011 to December, 2012. The study included 40 patients; 25 males and 15 females. The age of the patients ranged from 26 to 77 years old and one patient was five months, with mean age ± S.D. "50.86 ±15.46".  They were referred from Tropical, General Medicine, General Surgery Departments as well as out-patient clinics after obtaining informed consent, and parent’s consent was obtained for the infant aged 5 months.

Methods: All patients were subjected to the following:

A) History taking including personal history, age, sex, occupation and special habits of medical importance as well as viral hepatitis, alcoholism and use of oral contraceptives in female patient.

B) Patient complaints:  Right hypochon-drial pain, fever and jaundice.

C) Malignant symptoms: Weight loss, anorexia, cachexia, fever and gastrointestinal symptoms.

Clinical examination: General and local examinations.

Laboratory investigations:

A) CBC, serum creatinine, HBSAg and HCV.

B) Liver function tests.

C) Serum alkaline phosphatase.

 D)Tumor markers: Alpha fetoprotein and carcino-embryonic antigen.

• Ultrasound examination for the abdo-men using Ultrasonix, SP, with convex probe 2-5 MHz as a screening examination.

• Dynamic MRI of the liver using Philips, achiva-1.5 T-XR-Netherlands, 2010.

Technical Considerations for dynamic MRI of the liver scanning protocol:

A) Pre contrast imaging included:

- T₁ weighted (T₁W) gradient echo sequence (GRE) with and without fat suppression (FS): TR=100-200ms, TE ≤8ms, number of excitations (NEX) 1-4, matrix 128-256x256, slice thickness 5mm and slice gap 0-2 mm, flip angle =90.

- T₂ weighted (T₂W) images (fast spin echo sequence) with and without fat suppression (FS): repetition time (TR) 2000ms, echo time (TE) = 90-120 ms, number of excitations (NEX) 1-4, matrix 192-256x256 with a field of view as small as possible, slice thickness 5mm, slice gap 0-2mm, flip angle =90.

B) Dual-echo ( in and out) phase:  Using the shortest possible out-of-phase and in-phase echo times assured the best quality images, with better signal and fewer susceptibility artifacts.

- In  and  out phase: TR=99ms  in both in and out  phases, TE =4.6ms in in-phase WI and 2.3ms in out-phase WI, flip angle = 80 in both in and out phases.

C) Diffusion weighted image: We used breath-hold single-shot SE echo-planar imaging of the liver. 

D)Heavy T2:  MRI offered the additional benefit of T2-weighted imaging, in which heavy T2-weighted imaging with an echo time >112 ms was generally used to differentiate hepatic hemangiomas from malignant lesions, as the former retained their higher signal on this sequence (TR: 8000 msec, TE: 200 msec, flip angle: 90) 

E) Dynamic study:  Dynamic study was done  using  T1W GRE sequence by administration  of bolus  injection of  0.1mmol/kg of  gadolinium chelates at a rate of 2ml/s  {arterial phase (16-20 sec.), portovenous phase (45-60 sec.) and delayed equilibrium phase (3-5 min.)} followed immediately by  administration of 20ml of sterile 0.9% saline solution from the antecubital vein using pump injector.

     MRI pulse sequence parameters were collected in table (1).

Table (1): MRI   pulse sequence parameters.

     Histopathological study was done as confirmatory for the diagnosis.

Statistical analysis of data: The collected data were organized, tabulated and statistically analyzed using statistical package for social science (SPSS) version 19 (SPSS Inc, Chicago, USA), running on IBM compatible computer with Microsoft  Windows 7 operating System. Mean, standard deviation, frequency and percen-tage were used as descriptive. Chi square test (X²) was used for testing significance of observed differences between studied patients. The level of significance was adopted at p< 0.05. Sensitivity, speci-ficity, positive predictive value, negative predictive value and accuracy were used as measurements of validity for MRI regarding to histopathology.

RESULTS

      Forty patients were included in this study; 25 patients (62.5%) were males  and 15 patients (37.5% ) were females .

     The most common age incidence for hepatic tumors was in the seventh decade of life (15 cases; 37.5%), followed by the fifth decade (11cases; 27.5%), sixth decade (6 cases; 15%), forth decade (5 cases, 12.5 %), third decade (2 cases each; 5%) and one patient was 5 months (2.5%) (Table 2).

Table (2): Distribution of the studied cases regarding to age.

      Regarding MRI  presentations, there were 24 cases presented by liver cirrhosis, 4 cases presented by portal vein  throm-bosis, 24 cases presented by hepatitis, 2 cases  presented by intra hepatic biliary radicles dilatation, 16 cases presented by enlarged spleen, 4 cases presented by ascites (Table 3).

Table (3):  Distribution of the studied cases regarding to MRI presentation.

      In our study, sixteen cases (40%) were presented by multiple hepatic focal  lesions,  24 cases (60%) presented by solitary hepatic focal lesions, three cases (7.5%) presented by lymph node enlargement, two cases (5%)  presented by lung metastases and one case (2.5%) presented by bone metastases  (Table 4).

Table (4): Distribution of the studied cases regarding to focal, multiplicity, lymph nodes and metastases.

     Regarding MRI findings in diagnosis of hepatic focal lesions, the common hepatic tumor was HCC } solitary lesions (8 cases 20%) and  multicentric HCCs (6 cases 15%){ followed by metastasis (8 cases 20%), hemangioma (7 cases 17.5%), dysplastic nodules (3cases 7.5%), cholan-giocarcinoma and focal nodular hyper-plasia (2 cases for each 5%), followed by hepatoblastoma, adenoma,  regenerative  nodule and  benign lesions (1 case for each 2.5%) (Table 5).

Table (5): Distribution of the studied cases regarding to MRI diagnosis.

Distribution of the studied cases regarding biopsy, 33 cases (82.5%) were prepared for histopathological correlation. The remaining 7 cases (17.5%) were excluded for biopsy which were diagnosed by MRI as hemangiomas and underwent follow up.

     Regarding  histopathological  results, the common  hepatic tumor was HCC }solitary lesions (7 cases 17.5%) and  multicentric HCCs (4 cases, 10%){ followed by metastasis (10 cases 25%), followed by dysplastic nodules (3 cases, 7.5%), followed by cholangiocarcinoma and  focal nodular  hyperplasia and benign lesions (2 cases for each 5%) followed by hepatoblastoma, regenerative nodule and lymphoma  (1 case for each 2.5%) (Table 6).

Table (6):  Distribution of the studied cases regarding histopathological results.

    In our study, 24 cases (60 %) were presented by liver cirrhosis; 20 cases (50 %) showed malignant lesions, and 4 cases (10%) showed benign lesions. On the other hand, 16 cases (40%) were  presented by non cirrhotic liver; 11 cases (27.5%) showed benign  lesions and 5 cases (12.5%) showed  malignant lesions (Table 7).

Table (7):   Relation of  liver cirrhosis in the studied cases to the tumor types.

S= significant  X²*=chi square value.

     Regarding  the correlation of MRI in the diagnosis of hepatic tumors to the histopathological study were tabulated  in table (8).

Table (8): Relation of MRI in diagnosis of hepatic tumors to the histopathological results.

MRI findings of Fig. (1): Liver showed multifocal variable sized lesions, appeared low signal in T1 (a), in-phase (d) and out-phase (e), and high signal in T2 (b) and DWIS (c).  In dynamic study (f-h), the lesions revealed near to homogenous enhancement in the arterial phase with rapidly wash out in delayed phase.  I) TIC showed rapid wash in phase of hypervascularity of HCC, and rapidly wash-out denoting pathologically microcirculation in the tumor.    Diagnosis: Hepatocellular carcinoma.    

Histopathological diagnosis (j): Highly differentiated hepatocellular carcinoma.

MRI findings of Fig. (2): Liver showed multiple variable sized focal lesions with slight low signal in T2 (b) and DWIs (c).  In dynamic study (g-i), the lesions revealed non homogenous early faint enhancement in arterial phase with washout in delayed phases. Right thoracic rib metastasis was noted with enhancement was seen in axial post contrast T1 WI (f).   TIC  showed slow wash in phase of slight increased blood flow until mild part of curve followed by rapid washout phase in later portion, while at mid part curve showed higher peak of pathologically microcirculation of metastases (j). 

MRI diagnosis:   Metastatic  liver and chest wall lesions  versus multicenteric hepatocellular carci-noma with chest wall metastasis. Histopathological diagnosis (k):  Metastatic deposits from colorectal carcinoma.

DISCUSSION

     Hepatocellular carcinoma is the most common primary malignant tumor of the liver and often develops in patients with underlying liver cirrhosis due to excessive alcohol intake, chronic hepatitis or primary biliary cirrhosis (Janine et al., 2012).

    Eleven cases (27.5%) presented by HCC representing the most common primary malignant hepatic tumor. Four of them (10%) presented by multicentric HCC and remaining seven cases (17.5%) presented by solitary HCC.  With dynamic gadolinium-enhanced imaging, the lesion enhances in the arterial phase then becomes isointense in the portal phase then becomes hypointense in the delayed phase  (Asmaa et al., 2009).

     The signal intensity of HCC on DW-MRI is variable and appears to depend on its degree of differentiation. About 90% of moderately to poorly differentiated HCCs show intermediate-high signal intensity. Furthermore,  hypovascular well-differentiated HCCs appear isointense to the liver parenchyma, but the moderately to poorly differentiated hypovascular HCCs display high signal intensity (Muhi and  Ichikawa, 2011).  In our study, all HCCs (eleven cases) appeared as low signal intensity on T1-weighted images and high signal on T2-weighted images as well as DWI. Dynamic study revealed early contrast enhancement in arterial phase with rapidly wash out in delayed phases.

     The most common malignant tumors of the liver are metastases (about 30% of hepatic tumors). Liver metastases usually appear as solitary or multiple lesions. (Shahid et al., 2009).  In our study, we have reported eight cases (20%) presented by metastatic lesions (5 hypervascular and 3 hypovascular) representing the most common secondary malignant hepatic tumor.

     The imaging appearance of metastases depends on the degree of underlying hepatic arterial supply. Hypovascular  metastases show decreased enhancement on portal venous phase images. In addition, hypervascular metastases enhance earlier on arterial phase and show washout on delayed images. On diffusion-weighted images, these lesions are hyperintense (Alvin et al., 2012).  On DW-MR images in both hypervascular and hypovascular metastasis, the non-necrotic component of metastases display high signal intensity reflecting the cellularity of the solid tumor (Muhi and  Ichikawa,  2011).  In our study, all metastatic lesions showed variable degrees of low signal intensity on T1-weighted images and appeared hyperintense on T2 and DWIs. Dynamic study revealed five of these lesions presented by hypervascular metastases displayed early contrast enhancement on arterial phase and rapidly washout in delayed phases. The remaining three cases that presented by hypovascular metastases displayed slight low signal  in T1 as well as in phase and out phase WIs and  heterogenous high signal  in T2 and DWIs.

    Hemangiomas display high signal intensity on DW-MR images, but usually retain their high signal intensity at high b-value (b = 1,000 s/mm2) DW-MRI  (Kele and Jagt,  2012).  In our study,  all lesions appear as low signal intensity in T1WI, high signal intensity in T2 and DWIs and markedly hyperintense in heavy T2WI. Dynamic study revealed homogenous uniform contrast enhancement  on arterial phase in only one lesion and remaining six lesions revealed peripheral nodular enhancement with centripetal progression in delayed phases giving closed iris sign. Such cases biopsy was not recommended but follow up of these lesions 6 months later confirmed the diagnosis and revealed stationary course of the disease.

     Cholangiocarcinomas represent 10 to 20% of all primary liver malignancies, with an overall incidence rate of 0.95 cases per 100,000 adults. As a result of differences in local risk factors, genetics as well as classification issues, the rate varies markedly around the world from 2 cases per 100,000 to 96  cases per 100,000 (Christine et al., 2012).  In our study, we have only two cases out of forty cases (5%) presented by cholangiocarcinoma.    Cholangiocarcinoma is characterized on MRI as a hypointense lesion  on T1-weighted images and a hyperintense lesion on T2-weighted images. In dynamic contrast enhanced MRI, cholangiocarci-noma is usually recognized by delayed  moderate  peripheral  enhancement. Involved  bile ducts are identified by irregular ductal narrowing with proximal dilatation  (Boris and Gregory., 2010).

       In this study,  lesions appeared as slight high signal in T1-WI and near isointense in T2-WI as well as in phase WI and out phase WI and hyperintense  in DWI.  In dynamic study, the lesion revealed homogenous  enhancement in arterial phase with delayed washout in portal and delayed  phases. MRCP revealed diffuse intrahepatic biliary dilation in both cases.

     Typically, focal nodular hyperplasia "FNH" is iso or hypointense on T1- (94–100%), slightly hyper or isointense on T2- (94–100%) and with a bright central scar on T2-weighted images (84%). FNH shows intense homogeneous enhancement in the arterial phase with enhanced central scar and septa in the later phases of the gadolinium-enhanced images. The central scar shows high in signal intensity on T2-weighted images and shows enhancement on delayed contrast-enhanced images (Shahid et al., 2009).     In our study, the lesions appeared as near isointense with central hypointensity in T1, in phase  and out phase WIs and isointense with central high signal in T2WI and high  signal  in DWI.    In dynamic study, the lesions  revealed homogeneous vivid enhancement in the arterial phase with non enhanced central scarring and rapidly wash out in the delayed phases with enhanced central scar.

    Hepatoblastoma appears primarily in children younger than 5 years of age, while hepatocellular carcinoma occurs primarily after 10 years of age, The incidence of hepatoblastoma in children over 10 years appear only 0.001 % per million (Khunton et al., 2010).  In our study, we have only one case out of forty cases (2.5 %)  presented by hepatoblastoma.  This case was 5 months.     Accuracy of magnetic resonance imaging (MRI) compared with spiral computed (CT) in the diagnosis of liver mass was reported as following:  Sensitivity, specificity, positive predictive value, negative predictive value and  diagnostic accuracy of MRI  were 85%, 71%, 92%, 56%  and 82% respectively,  and that of spiral computed (CT) were 70%, 86%, 95%, 43% and 74% respectively (Nimer et al., 2012).

      In this study,  accuracy of magnetic resonance imaging (MRI) in the diagnosis of liver masses was  reported as achieving sensitivity; 86.7%,  specificity; 66.7%, PPV; 96.3%, NPV; 33.3% and accuracy; 84.8%.

CONCLUSION

      MRI has been used to improve detection and characterization of hepatic malignant lesions.   It was found to be more accurate than CT or US in detecting HCC and dysplastic nodules in patients with cirrhotic liver.   MRI also is a very good modality in delineating the internal architecture of the tumor, tumoral margins and intrahepatic vascular invasion.  It has the advantage of achieving high-resolution images of the liver without the use of ionizing radiation and has a better sensitivity and specificity compared with CT and ultrasound in cirrhotic patients in whom it can be difficult to distinguish HCC from other lesions.  The histopatho-logical study confirmed the diagnostic accuracy of the targeted lesions and tumors.

REFERENCES

1. Alvin CS.,  James ME.,  Ann EM. and Mashal AJ.  (2012):  MR imaging of hypervascular liver masses.   Radiographics,  29(2): 385-402.

2. Amital MM, Haider MA and Rappaport DC. (2008): Multi-detector row helical CT in preoperative assessment of small liver metastases: is thinner collimation better.  Radiology,  225(1): 137-142.

3. Arguedas MR (2007): Screening for hepatocellular carcinoma why, when, how?. Journal Current Gastroenterol Reports,   5(1): 57-62. Publisher; Current Medicine Group., London –UK.

4. Asmaa IG, Shahid AK, Edward LS and Imam W.  (2009): Diagnosis of hepatocellular carcinoma.  World J Gastroenterol.,   15(11): 1281-1288.

 5. Boris R.A. and Gregory JG. (2010):  Cholangiocarcinoma. Clin Liver Dis.,  12: 131-150.

6.  Christine S, Ghalib J, Stephen C.  and Cathy F. (2012): Intrahepatic cholangiocarcinoma: new insights in pathology: incidence, risk factors, and pathogenesis of intrahepatic cholangiocarcinoma. Seminars in Liver Disease,  31(1): 49-60.  

7. Demir OI, Obuz F, Sagol O. and Dicle O. (2009): Contribution of diffusion weighted MRI to the differential diagnosis of hepatic masses. Diagn. Interv. Radiol., 13: 81-86.

8. Devaki S, Sacks K, Patrick J. and Luke C. (2012): Value of PET/CT in the Management of Primary Hepatobiliary Tumors.   AJR;   197: 256-259.

9.  Diego RM, Raman D and Shahid MH (2010): MR imaging of the liver in children. Journal of Magnetic Resonance Imaging,  31(5): 637-656.

10.  Janine R, Ernst-Michael J, Andreas G  and Patrick H. ( 2012): MR arterioportography: A new technical approach for detection of liver lesions. World Journal of Gastroenterology, 17(13): 1739-1745.

11.  Kele PG and Jagt EJ. (2012): Diffusion weighted imaging in the liver. World J Gastroenterol.,  16(13): 1567­1576.

12.  Khunton W, Pope K and Surapon W. (2010): Incidence of hepatocellular carcinoma in children in Khon Kaen (Thailand) before and after national hepatitis B vaccine program. Asian Pacific Journal of Cancer Prevention, 5(3): 25-34.   

13. Muhi A  and Ichikawa T. (2011): DW-MRI for Disease Characterization in the Abdomen in: Diffusion-Weighted MR Imaging Applications in the Body,  chapter 8, P: 117-142.  Publisher; Springer USA. 

14. Nimer A, Gattas N, Agness D and Zaza B. (2012): Approach to Solid Liver Masses in the Cirrhotic Patients. Clin. Gastroenterology, 40(26); 774-782.      

15. Peterson M and Baron R. (2007): Screening the Cirrhotic Liver for Hepatocellular Carcinoma with CT and MR Imaging: Opportunities and Pitfalls. RadioGraphics,  21: 117-132.  

16. Richard CS, Cance WG, Marcos HB and Mauro MA (2011): Hepatic imaging.   Radiol. Clin. N Am., 43: 54-64.

17. Saima H, Cristina BM and Filipe CA. (2011): Liver lesions, Triphasic-CT scan; Radiology, 61(6): 571-75.

18. Shahid MH, John LG, and Richard CS (2009): Liver MRI correlation with other imaging modalities and histopathology., Clin. liver Dis. Ch 17, page 36.   Publisher;  Springer USA. 

19. Vilgrain V,  Mathieu D, and Mahfouz A. (2005): Benign liver tumors. Magn. Reson. Imaging. Clin. N. Am.,  25(5): 255-288.