Materials and Methods

Patients

Our retrospective study was approved by the institutional review board.

During June 1, 2012 to December 31, 2014 all MRI studies with focal liver lesions were included. Inclusion criteria in the study were (1) All pathological proven the largest focal liver lesions more than 10 mm in each MRI (2) The focal liver lesions without a history of any treatment and (3) The largest benign focal liver lesions with classical MRI findings in each MRI studies which follow-up examinations using US, CT or MRI at least 6 months duration.  The exclusion criteria were all lesions with low images quality of DWI and ADC map. The lesions were classified into benign or malignant groups.

MRI

MR imaging

MR imaging was performed on a 1.5-T system (MagnetomAera, Siemens Medical Solutions, Erlangen, Germany) with  sixteen channel body phased array coils anterior and two spine clusters (three channels each) posterior.

Coronal T2-weighted half-Fourier single-shot turbo spin-echo (HASTE) sequence and an axial T2-weighted turbo spin-echo sequence were acquired as well as an axial dynamic T1-weighted, three-dimensional spoiled gradient-recalled echo sequence (volumetric interpolated breath-hold examination (VIBE) sequence with spectral fat saturation) following the intravenous administration of Gd-DTPA.

Diffusion-weighted images were acquired using a single-shot echo-planar imaging sequence. Thus, the gradient factors (b values) were 0, 150, and 800 s/mm². The technical parameters were as follows: echo time, 65 ms; EPI factor, 125; echo spacing, 0.77ms; receiver bandwidth, 1736 Hz/pixel; spectral fat saturation; field of view, 292 × 360 mm; matrix, 125 × 192; section thickness, 6 mm. For shortening of the echo train length, integrated parallel imaging techniques (iPAT) by means of generalized autocalibrating partially parallel acquisitions (GRAPPA) with a twofold acceleration factor were used. For respiratory triggering, PACE (prospective acquisition correction) was implemented. Data were acquired during the end-expiratory phase. DWI was performed before the administration of Gd-DTPA.      

Procedure

All MRI images of the patients who were eligible in this study were reviewed. Demographic data included age, gender, diagnosis (benign and malignant), subtypes of diagnosis (HCC, CCA, liver metastasis, cyst, hemangioma, FNH, and adenoma) were collected. The ADC measurement of liver masses was measured from ADC maps.

Image analysis

All MR images were analyzed retrospectively by a radiologist specialized in gastrointestinal imaging with more than 9 year experienced.  All lesions equal or larger than 1 cm were selected to avoid volume errors and underwent 3 measurements per lesion and then averaging them to get the final ADC of each lesion. Malignant lesions with necrotic or fibrous core were measured only solid portion to void heterogeneous ADC value. The largest lesion was selected in patient who has multiple similar characteristic pathology.

Statistical analyses

Demographic data variables which included baseline characteristics were divided into dichotomous or polytomous variables. All variables were summarized using descriptive statistic presentation in percentage, mean and standard deviation. However, if the distribution of this data was not a normal distribution, then median and inter-quartile range were used instead. Wilcoxson signed rank test was used to analysis the differences between the ADC values of benign and malignant liver mass. The ROC curve was used to summarize the overall accuracy of the DWI using ADC measurement for detecting malignant liver mass. Then an optimal cut-off point was determined. The performance of the test was summarized by the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), likelihood and Youden’sindex.

Sample size calculation

Sample size calculation is based on the objective of the study which was the performance of the DWI using ADC measurement in detecting malignant abdominal mass.  ROC curves are used to summarize the accuracy of diagnostic tests. Therefore, calculating sample size was based on the area under the ROC curve (AUC) according to the methodology of Hanley and McNeil (1983) ²⁹.

          This method varies the sample size until a sufficiently small S.E. of the area under the ROC curve is achieved. The sample size of at least 70 cases were enough and feasible to conduct in clinical practice at the SE of p<0.05.

Results

Baseline data

During a studied period, there were 180 consecutive focal liver lesions included. The baseline and characteristics of lesions are demonstrated in Table 1.

The 79 malignant lesions were pathological proven by surgery or biopsy composed with 52 CCAs, 20 HCCs and 7 liver metastases. There were 101 eligible benign liver lesions are included in the study. In 44 hemangiomas, the diagnosis was pathologically confirmed by surgery only 1 lesion and the other 43 lesions were based on typical MRI findings and follow up images. In 39 cysts, the diagnosis was based on typical MRI findings and follow up serial CT or MRI examination at least 6 months (range, 6-23 months). In 11 FNHs, 4 lesions were histopathologically confirmed with surgery and the other 7 lesions base on typical MRI findings criteria using hepatocyte-specific contrast agents. In all 7 adenomas, the diagnosis was based on typical MRI findings criteria using hepatocyte-specific contrast agents.

Table 1 Baseline characteristics of studied populations

Note; N: total number of liver lesions, IQR: inter-quartile range, FNH: focal nodular hyperplasia, HCC: hepatocellular carcinoma, CCA: cholangiocarcinoma

Discussion

          Many preliminary studies have measured the ADCs of the liver lesions with diffusion-weighted echoplanar MR imaging and have demonstrated lower ADC values in malignant than in benign lesions, some of them reported overlap values. Miller et al. ⁸ reported the more common benign liver lesions were hemangiomas and cysts, had significantly higher mean ADC values compared to other lesions. Similar to this study, cyst was the most common liver lesions, followed by hemangiomas. Liver cysts and hemangiomas were also higher in ADC values than other lesions, most likely secondary to their higher fluid content resulting in more free water molecule diffusion.

In contrast, solid lesions show the low ADC values likely due to their high cellularity and the resultant restricted diffusion of water molecules. Hemangiomas are not pure fluid-containing lesions but can contain vascular endothelial tissue, fibrous septa, and blood. These non-fluid elements do restrict water motion and consequently ADC values of hemangiomas are lower than that of cysts as seen in our study. The median ADC value of cyst and hemangioma in our study were 2.91.04x10-3 mm²/s and 1.841.04x10-3 mm2/s, respectively.

          The result of this study suggested the optimal cutoff ADC value for distinction between malignant and benign liver lesions was 1.49x10^-3 mm²/s (when using a maximum b value of 800 s/ mm²), with sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 84.8% , 81.2%, 4.51, and 0.19, respectively).

This study reported lower sensitivity and lower specificity than Chen ZG et al.¹⁰, a meta-analysis stud found DWI was useful for differentiating between malignant and benign hepatic lesions with sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 93% (95% confidence interval (CI): 0.91-0.95), 87% (95% CI: 0.83-0.91), 7.28 (95% CI: 4.51-11.76), and 0.09 (95% CI: 0.05-0.17), respectively. They mentioned the diagnostic capability might be overestimated due to the possibility of selection bias.

          The  optimal cutoff ADC value of this study was closed to the results of Kim et al. [4] (1.6x10^-3 mm²/s when using a maximum b value of 846 s/ mm²), Taouli et al.⁵ (1.5x10^-3 mm²/s, maximum b value of 500 s/ mm²), Miller et al.⁸ (1.5x10^-3 mm²/s, maximum b value of 500 s/ mm²), and Gourtsoyianni et al.² (1.47x10^-3 mm²/s, maximum b value of 1000 s/ mm²), whereas the results of Ichikawa et al.⁶  is 5.5x10^-3 mm²/s when using a maximum b value of 55 s/ mm². The similarity or difference between cutoff ADC values should be due to the magnitude of the maximum b value used for the other studies, as discussed in the preceding text. Ichikawa et al ⁶ mentioned in their report that larger b factor, those greater than 400 s/mm² should be chosen for more precise evaluation of ADCs in the abdomen, but the image quality of a diffusion-weight image with greater b values would be greatly diminished because the T2 relaxation time of abdominal tissues is short. However, Kim et al. ⁴ mentioned in their report that greater b values can be used for characterizing liver lesions because the targets for characterization should be the lesions, which have long T2 relaxation time and show high signal intensity on T2-weighted images. Moreover, ADCs of liver lesions measured with larger b values would be less affected by the magnitude of the maximum b value and would be more similar to a diffusion coefficient less affected by perfusion. In agreement with Kim et al.⁴ and Ichikawa et al.⁶, we selected b 800 s/mm2factors to obtain diffusion-weight images in this study.

          Due to the fact that solid benign liver lesions, including FNH and adenoma, being hypercellular lesions, are expected to present with low ADC values. The prior study of Miller et al.⁸ and Sandrasegaran et al.⁹ reported some overlaps in ADC values between solid benign liver lesions (FNHs and adenomas) and malignant lesions. Our results similar to previous reports which no significant difference between overall ADC value of malignant and solid benign liver lesions. However, 1.04x10-3 mm2/s value of ADC was used to be the optimal cut off ADC value to differentiated liver metastases from solid benign liver lesions, with a sensitivity and a specificity (85.7%, and 83.3%, respectively) that similar sensitivity and specificity of  meta-analysis of Wei et al.¹¹, who reported DWI was useful for differentiation between liver metastases and benign focal liver lesion with sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of 87%, 90%, 8.50, and 0.17, respectively.  The cut off ADC value in our study was lower than study of Testa et al.³⁰, who reported 1.2x10-3 mm²/s was the ADC cut off value to differentiate between solid benign and malignant liver lesion with a sensitivity of 71%, and a specificity of 71%.

We also found no significant difference between ADC value of the primary and metastasis liver malignancy which is similar to reported of Goya C. et al.³¹, who found ADC value are not helpful in differentiation between HCCs (1.06x10-3 mm²/s) and liver metastases (1.102x10-3 mm²/s ).   Among the primary liver tumor CCA is the largest number of primary liver tumor because the study population live in the northeast of Thailand where it has the highest incidence of CCA³². The median value for ADC of CCA was 1.15x10-3 mm²/s which is lower than the study of Xing-Yu Cui et al.³³, who found the mean value for ADC of extrahepatic CCA was 1.31 ± 0.29x10-3 mm²/s.

          Our study has few potential limitations. First, the single-shot echoplanar images contain a distortion artifact caused by the susceptibility effect, which tends to be more severe with a larger b value. This advantage may be reduced in the future if it becomes possible to obtain diffusion-weighted images with other fast imaging sequences that are less sensitive to the effects of static magnetic field inhomogeneities. Second, we did not have histopathologically confirmed diagnosis for all patients or lesions; however, those without surgery or biopsy-proven lesions were only included in the study if they has typical features on MR images as well as adequate follow-up to confirm the diagnosis.

Conclusion

            In conclusion, ADC value is useful for differentiating malignant from benign liver lesions using 1.49x10^-3 mm²/s optimal cutoff ADC value and 1.04x10^-3 mm²/s optimal cutoff ADC value to differentiated metastases from benign solid liver lesions.

Acknowledgment

Funding This study was funded by Faculty of Medicine Khon Kaen University, Thailand (Grant Number IN59346).

References

1. Koh DM, Collins DJ. Diffusion-weight MRI in the body: applications and challenges in oncology. AJR Am j Roentgenol 2007;188:1622-1635.

2. Gourtsoyianni S, Papanikolaou N, Yarmenitis S, Maris T, Karantanas A, Gourtsoyiannis N. Respiratory gated diffusion-weighted imaging of the liver: value of apparent diffusion coefficient measurements in the differentiation between most commonly encountered benign and malignant focal liver lesion. Eur Radiol 2008; 18: 486-492.

3. Vergara ML, Fenandez M, Rodrigo MT. Diffusion-Weighted MRI Characterization of solid liver lesions. Rev Chil Radiol 2010; 16(1): 5-10.

4. Kim T, Murakami T, Takahashi S, Hori M, Tsuda K, Nakamura H. Diffusion-weighted single-shot echoplanar MR imaging for liver disease. AJR Rm I Roentgenol 1999; 173: 393-398.

5. Taouli B, Vilgrain V, Dumont E, Daire JL, Fan B, Menu Y. Evaluation of liver diffusion isotropy and characterization of focal hepatic lesions with two single-shot echo-planar MR imaging sequences: prospective study in 66 patients. Radiology 2003;226:71–78.

6. Ichikawa T, Haradome H, Hachiya J, Nitatori T, Araki T. Diffusion-weighted MR imaging with a single-shot echoplanar sequence: detection and characterization of focal hepatic lesions. Am J Roentgenol 1998;170:397–402.

7.Bruegel M, Holzapfel K, Gaa J, Woertler K, Waldt S, Kiefer B, et al. Characterization of focal liver lesions by ADC measurements using a respiratory triggered diffusion-weighted single-shot echoplanar MR imaging technique. EurRadiol 2008 ;18: 477-485.

8. Miller FH, Hammond N, Siddiqi AJ, Shroff S, Khatri G, Wang Y, et al. Utility of diffusion-weighted MRI in distinguishing benign and malignant hepatic lesions. J MagnReson Imaging 2010;32:138–147.

9. Sandrasegaran K, Akisik FM, Lin C, Tahir B, Rajan J, Aisen AM. The value of diffusion-weighted imaging in characterizing focal liver masses. AcadRadiol 2009;16:1208-1214.

10. Chen ZG, Xu L, Zhang SW, Huang Y, Pan RH. Lesion discrimination with breath-hold hepatic diffusion-weighted imaging: a meta-analysis. World J Gastroenterol 2015; 21(5): 1621-1627.

11. Wei C, Tan J, Xu L, Juan L, Zhang SW, Wang L, et al. Differential diagnosis between hepatic metastases and benign focal lesions using DWI with parallel acquisition technique: a meta-analysis. Tumor Biol 2015;36:983-990.

12. Caraiani C, Chiorean L, Fenesan DI, Lebovici A, Feier D, Gersak M, et al. Diffusion weighted magnetic resonance imaging for the classification of focal liver lesions as benign or malignant. J Gastrointestin Liver Dis 2015; 24:309-317.

13. Hasan NMA, Zaki KF, Alam-Eldeen MH, Hamedi HR. Benign versus malignant focal liver lesions: diagnostic value of qualitative and quantitative diffusion weighted MR imaging. The Egyptian journal of radiology and nuclear medicine 2016;47:1211-1220.

14. Xia D, Jing J, Shen H, Wu J. Value of diffusion-weighted magnetic resonance images for discrimination of focal benign and malignant hepatic lesions: a mata-analysis. J MagnReson Imaging 2010;32:130-137.

15. Li Y, Chen Z, Wang J. Differential diagnosis between malignant and benign hepatic tumors using apparent diffusion coefficient on 1.5-T MR imaging: a meta analysis. Eur J Radio l2012;81:484-490.

16. Cieszanowski A, Anysz-Grodzicka A, Szeszkowski W, Kaczynski B, Maj E, Gornicka B, et al. Characterization of focal liver lesions using quantitative techniques: comparison of apparent diffusion coefficient values and T2 relaxation time. EurRadiol 2012;22:2514-2524.

17. Demir OI, Obuz F, Sagol O, Dicle O. Contribution of diffusion-weighted MRI to the differential diagnosis of hepatic masses. DiagnIntervRadiol 2007;13(2):81-86.

18. Tokgoz O, Unlu E, Unal I, Serifoglu I, Oz I, Aktas E, et al. Diagnostic value of diffusion weighted MRI and ADC in differential diagnosis of cavernous hemangioma of the liver. Afri Health Sci 2016;16(1):227-233.

19. Holzapfel K, Bruegel M, Eiber M, Ganter C, Schuster T, Heinrich P, et al. Characterization of small (≤10 mm) focal liver lesions: Value of respiratory-triggered echoplaner diffusion-weighted MR imaging. Eur J Radiol 2010;76:89-95.

20. Koike N, Cho A, Nasu K, Seto K, Nagaya S, Ohshima Y, et al. Role of diffusion-weighted magnetic imaging in the differential diagnosis of focal hepatic lesions. World J Gastroenterol 2009;15(46):5805-5812.

21. Saito K, Tajima Y, Harada TL. Diffusion-weighted imaging of the liver: current applications. World J Radiol 2016;8(11):857-867.

22. Battal B, Kocaoglu M, Akgun V, Karademir I, Deveci S, Guvenc I, et al. Diffusion-weighted imaging in the characterization of focal liver lesions: efficacy of visual assessment. J comput Assist Tomogr 2011;35(3):326-331.

23. Erturk SM, Ichikawa T, Sano K,  Motosugi U, Sou H, Araki T. Diffusion weighted magnatic resonance imaging for characterization of focal liver masses: Impact of parallel imaging (SENSE) and b value. J comput Assist Tomogr 2008;32:865-871.

24. Lyng H, Harald seth O, Rofstad EK. Measurement of cell density and necrotic fraction in human melanoma xenografts by diffusion weighted magnetic resonance imaging. MagnReson Med 2000;43:828-836.

25. Sugita R, Yamazaki T, Furuta A, Itoh K, Fujita N, Takahashi S. High B value diffusion-weighted MRI for detecting gallbladder carcinoma: preliminary study and results. EurRadiol 2009;19:1794-1798.

26. Yamada Rijswijk CS, Kunz P, Hogendoorn PC, Taminiau AH, Doornbos J, Bloem JL. Diffusion-weighted MRI in the characterization of soft tissue tumors. J MagnReson Imaging 2002;15:302-307.

          27. Yamada I, Aung W, HimenoY,Nakagawa T, Shibuya H.Diffusion coefficients in abdominalorgans and hepatic lesions: evaluationwithintravoxel incoherent motionecho-planar MR imaging. Radiology 1999;210(3):617–623.

          28. Sun XJ, Quan XY, Huang FH, Xu YK. Quantitative evaluation of

diffusion-weighted magnetic resonance imaging of focal hepatic lesions. World

J Gastroenterol 2005;11:6535–6537.

          29. Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology1983;148(3):839-843.

          30. Testa ML, Chojniak R, Sene LS, Damascena AS, Guimaraes MD, Szkiaruk J, et al. Is DWI/ADC a useful tool in the characterization of focal hepatic lesions suspected of malignancy?.PLoS ONE 9(7): e101944. doi:10.1371/journal.pone.0101944

31. Goya C, Hamidi C, Onder H, Inal A, Tekbas G, Firat U, et al. Primary and metastatic liver malignancy: utility low and high B value (1600-2000). Hepatogastroenterology 2015; 62(140):962-5.

32. Green A, Uttaravichien T, Bhudhisawasdi V,Chartbanchachai W, Elkins DB, Marieng EO, et al. Cholangiocarcinoma in north east Thailand. A hospital-based study. Trop Geogr Med 1991; 43(1-2):193-198.

33. Cui XY, Chen HW.Role of diffusion-weighted magnetic resonance imaging inthe diagnosis of extrahepatic cholangiocarcinoma. World J Gastroenterol 2010; 16(25): 3196-3201.