Separation of liver focal nodular hyperplasia (FNH) and liver malignant tumors by a combination of T2-weighted imaging signal and three diffusion magnetic resonance metrics of diffusion-derived vessel density, slow diffusion coefficient, and apparent diffusion coefficient

Abstract

Background

This study investigates the differentiation of liver focal nodular hyperplasia (FNH) from liver malignant tumor (MT) by a combination of T2-weighted imaging (T2WI), diffusion-derived vessel density (DDVD), slow diffusion coefficient (SDC), and apparent diffusion coefficient (ADC). Based on the odds ratio (OR) for a sign to suggest the possibility of a lesion being FNH, we propose a liver mass sum score (LiverMss-FNH) scheme to facilitate the diagnosis.

Methods

Liver diffusion-weighted magnetic resonance imaging included 13 cases of FNH and 82 cases of MT. DDVD was calculated from b=0 and b=10 s/mm² images, SDC was calculated from b=500 and b=800 s/mm² images, and ADC was calculated from b=0 and b=800 s/mm² images. For liver semi-quantitative analysis, relative to the adjacent liver signal, a liver lesion’s signal was assigned to five categories: low signal, iso-signal, slightly high signal, high signal, and markedly high signal. The lesion on T2WI being not high signal was assigned a sub-score “1” (otherwise scored 0); the lesion being iso-signal on DDVD was assigned a sub-score “1.5” (otherwise scored 0); the lesion on SDC being not high signal was assigned a sub-score “1” (otherwise scored 0); the lesion on ADC being not low signal was assigned a sub-score “0.5” (otherwise scored 0); the existence of stellate scar was assigned a sub-score “0.5” (otherwise scored 0). The sum of these five sub-scores was termed LiverMss-FNH.

Results

A total of 26 MT cases had large (median 8.1 cm, standard deviation: 4.2 cm) and very heterogeneous masses which were very unlikely to be FNH. The remaining 13 FNH cases (median 3.8 cm, standard deviation: 1.7 cm) and 56 MT cases (median 4.9 cm, standard deviation: 4.3 cm; hepatocellular carcinoma, n=40; metastasis, n=12; intrahepatic cholangiocarcinoma, n=4) were evaluated with LiverMss. Liver lass lesion being not high signal on T2WI, being iso-signal on DDVD, being not high signal on SDC, being not low signal on ADC, and the existence of stellate scar had ORs of 49.1, 45.8, 30, 8.5, and 13.3, respectively, favoring the diagnosis of FNH. A total of 69.2% (9/13) of the FNH had LiverMss-FNH ≥4.0, while the remaining 4 cases (30.8%) all had a LiverMss-FNH of 3.0. A total of 89.3% (50/56) of the MT had LiverMss-FNH ≤1.5.

Conclusions

Liver FNH tend to have lower DDVD signal and lower SDC signal than liver MT. A LiverMss ≥4 can strongly suggest the diagnosis for a liver mass being FNH, and while a LiverMss-FNH ≤1.5 can strongly suggest the diagnosis for a liver mass being MT.

Highlight box.

Key findings

• A combination of magnetic resonance T2-weighted imaging (T2WI) and diffusion imaging can differentiate liver focal nodular hyperplasia (FNH) from liver malignant tumors.

What is known and what is new?

• Liver FNH tends to have lower T2 signal and higher apparent diffusion coefficient than liver malignant tumors.

• Liver FNH tends to have lower diffusion-derived vessel density signal and lower slow diffusion coefficient signal than liver malignant tumors.

What is the implication, and what should change now?

• A substantial proportion of liver lesions, as long as they are of reasonable size, can be classified based on a combination of T2WI and diffusion metrics, without the need for a contrast agent injection.

Introduction

Liver focal nodular hyperplasia (FNH) is a benign lesion of hepatocytic hyperplasia that commonly arises in a background of normal or nearly normal liver. It is the second most common benign liver lesion after hemangioma, representing 8% of primary hepatic lesions. It has an incidence of 0.9%, is much more common in women than in men (8:1). Histologically, FNH is hyperplastic growth of morphologically normal hepatocytes, without normal development of the portal tract. A central fibrovascular scar and radiating fibrous septa contain large malformed feeder arteries and branches (1). Most cases of FNH are asymptomatic and incidental. FNH lesions have no malignant potential, and hemorrhage and rupture are rare. Asymptomatic FNH requires no treatment or imaging follow-up in patients with no known malignancy or underlying liver disease (2,3).

Magnetic resonance imaging (MRI) has higher sensitivity and specificity for FNH than do ultrasonography or computed tomography. Typically, FNH is iso- or hypointense on T1-weighted images, is slightly hyper- or isointense on T2-weighted images (T2WIs), and has a hyperintense central stellate scar on T2WIs. FNH demonstrates intense homogeneous enhancement during the arterial phase of gadolinium-enhanced imaging and enhancement of the central scar during later phases. Currently, the standard of diagnosis for FNH is based on the application of hepatobiliary contrast agents including gadoxetic acid (Gd-EOB-DTPA, Primovist; Bayer, Berlin, Germany) and gadobenate dimeglumine (Gd-BOPTA, MultiHance; Bracco, Milan, Italy). FNH, as a hyperplastic growth of hepatocytes, almost always shows hepatobiliary phase uptake, whereas lesions of nonhepatocellular origin, such as hemangiomas or metastases, do not take up hepatobiliary contrast agents and appear hypointense during the hepatobiliary phase. Other lesions of hepatocellular origin, such as hepatocellular carcinoma (HCC) and hepatic adenoma, are also usually hypointense during the hepatobiliary phase (1,4-7).

Recently, we introduced two new diffusion weighted imaging (DWI) metrics. Liver micro-vessels, including sub-pixel vessels, show high signal when there is no motion probing gradient (b=0 s/mm²) and low signal when even very low b-values (such as b=2 s/mm²) are applied (8). Thus, the signal difference between images when the motion probing gradient is “off” and “on” reflects the extent of tissue functional vessel density (in the physiological sense), and we term this as diffusion-derived vessel density (DDVD) (8-10). The clinical usefulness of DDVD as a straightforward diffusion imaging biomarker for liver imaging has been recently demonstrated (11-16). DDVD is a useful parameter for distinguishing livers with and without fibrosis, and livers with severe fibrosis tend to have even lower DDVD measurements than those with milder liver fibrosis (8,10). Li et al. applied DDVD to assess the perfusion of HCC, showing higher perfusion of HCC relative to adjacent liver parenchyma (13). Hu et al. (14) described that liver hemangiomas, which show very high DDVD values, can be mostly differentiated from liver mass-forming lesions (HCCs and FNH) based on the DDVD map. Slow diffusion coefficient (SDC) was proposed to measure tissue slow diffusion (17). In its basic form, SDC is derived from a high b-value DWI image and a higher b-value DWI image. With the conventional apparent diffusion coefficient (ADC) approach, the spleen has been reported to have a much lower ADC than liver, HCCs have a lower ADC than liver parenchyma, and simple liver cysts have a higher ADC than liver hemangiomas. On the other hand, with SDC analysis, Xu et al. (17) reported that the spleen has a faster diffusion than liver, HCCs have a faster diffusion than liver parenchyma, and liver hemangiomas have a faster diffusion than simple liver cysts. The liver and spleen have a similar amount of blood perfusion, the spleen is waterier than the liver, and the spleen tissue has a higher contrast-enhanced computed tomography (CT) extracellular volume fraction than the liver (15,18). HCCs are mostly associated with increased blood supply and increased proportion of arterial blood supply and with edema. It is more reasonable with SDC results that spleen and HCC have a faster diffusion than liver parenchyma. Due to the “flushing” of blood flow inside the hemangioma, it is also more reasonable with SDC results that the diffusion of hemangioma liquid is faster than the more “static” liquid of the cysts. Hu et al. reported that a combination of DDVD and SDC offers an accuracy of >95% in separating liver hemangioma and liver mass lesions (19).

In our recent attempt to evaluate FNH with DDVD (20), FNH was shown to have a lower DDVD value than malignant lesions, with DDVD qualification showing an area under the receiver operating characteristic curve of around 0.9 for separating FNH and liver malignant tumors (MT). Building on the initial promising results, the current study further investigates the differentiation of FNH from liver MT by a combination of T2WI, DDVD, SDC, and ADC. Based on the odds ratio (OR) for a sign to suggest the possibility of a lesion being FNH, we propose a scoring system scheme, liver mass sum score for FNH (LiverMss-FNH), to facilitate the diagnostic decision.

Methods

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All imaging data were acquired with institutional ethical approval from the ethics committee of Sun Yat-sen Memorial Hospital (approval ID, S1709442250839) and with informed consent obtained from individual participants. The liver imaging data were initially prospectively acquired for an intravoxel incoherent motion (IVIM) study of liver lesions, and the data were then retrospectively re-analysed in the current study. Liver imaging was performed with a 3.0-T magnet (Vida Magnetom, Siemens Healthineers, Erlangen, Germany). The diffusion imaging was based on a single-shot spin-echo type echo-planar sequence, with respiratory gating. The default spectral pre-saturation technique was used for fat suppression. DWI images with b-values of 0, 10, 500, and 800 s/mm² were utilized in this study. The repetition time (TR) was 2,500 ms, and the echo time (TE) was 84 ms. Other parameters included slice thickness =5 mm and inter-slice gap =1 mm, matrix =128×128, field-of-view (FOV) =350 mm × 350 mm, number of excitation (NEX) =1 for b=0, 10 s/mm² images, and NEX =3 for b=500, 800 s/mm² images.

In this study, we initially included 14 cases of liver FNH and 86 cases of liver MT including HCC, metastasis (Mets), and intrahepatic cholangiocarcinoma (ICC). All HCC and ICC cases had histopathological diagnosis. Eight FNH lesions had histopathological diagnosis, and the remaining 6 lesions were diagnosed with typical findings of hepatobiliary contrast agent enhanced MRI. The diagnosis of Mets was based on histopathology or a combination of complete patient history and typical imaging features.

DDVD weighted image was calculated from b=0 and b=10 s/mm² images, and derived from the equation (8-10):

$$\text{DDVD}=\frac{Sb0}{{\text{ROI}}_{\text{area0}}}-\frac{Sb10}{{\text{ROI}}_{\text{area10}}}\left[\text{unit:\hspace{0.17em}arbitrary\hspace{0.17em}unit}\left(\text{au}\right)\text{/pixel}\right]$$ [1]

where ROI_area0 and ROI_area10 refer to the number of pixels in the selected region-of-interest (ROI) on b=0 and b=10 s/mm² DWI, respectively. S(b₀) refers to the measured sum signal intensity within the ROI when b=0, and S(b₁₀) refers to the measured sum signal intensity within the ROI when b=10 s/mm². If we consider a pixel to be an individual ROI, DDVD pixelwise map (DDVDm) can be constructed pixel-by-pixel with this same principle (11).

SDC was calculated from b=500 and b=800 s/mm² images, and derived from the equation (17):

$$SDC=\frac{S\left(b_1\right)-S\left(b_2\right)}{b_2-b_1}\left[\text{unit:\hspace{0.17em}au/s}\right]$$ [2]

where b₁ and b₂ refer to a high b-value (i.e., 500 mm²/s in this study) and a higher b-value, respectively (i.e., 800 mm²/s in this study), S(b₁) and S(b₂) denote the image signal intensity acquired at the high b-value and the higher b-value, respectively.

ADC was calculated according to:

$$ADC=\frac{ln\left(S\left(b_1\right)/S\left(b_2\right)\right)}{b_2-b_1}\left[\text{unit:\hspace{0.17em}}{\text{mm}}^{\text{2}}\text{/s}\right]$$ [3]

where b₂ and b₁ refer to the high b-value (i.e., 800 mm²/s in this study) and low b-value, respectively (i.e., 0 mm²/s in this study), where S(b₂) and S(b₁) denote the image signal intensity acquired at the high b-value and low b-value, respectively.

The image analysis and scoring were conducted by two readers in consensus [a specialist radiologist (Y.X.J.W.) and a senior radiology trainee (C.Y.L.)]. A semi-quantitative (SQ) analysis was conducted on: (I) T2WI, (II) DDVD map, (III) SDC map, and (IV) ADC map, respectively. Relative to the adjacent liver signal, a liver lesion signal was assigned to five SQ categories: low signal (scored as “0”), iso-signal (scored as “1”), slightly high signal (scored as “1.5”), high signal (scored as “2”), and markedly high signal (scored as “3”). On T2WI and SDC, “high signal” was generally the expected spleen signal; “markedly high signal” was usually close to the liquid or blood vessel signal. When it was difficult to sign a score of “2” or “3”, an additional score of “2.5” (“higher signal”) was assigned (20). On T2WI, an absolute iso-signal of a lesion rarely exists; thus, being consistent with literature, a faintly higher signal was assigned a score of “1”.

As explained in Figure 1, we further used a “sub-scoring scheme”: the lesion on T2WI being not high signal was assigned a score “1” favoring the FNH diagnosis (i.e., signal score ≤1.5; otherwise scored 0); the lesion being homogenously iso-signal on DDVD map, except the high signal due to stellate scar or potential artefacts, was assigned a sub-score “1.5” favoring the FNH diagnosis (otherwise scored 0); the lesion on SDC map being not high signal was assigned a sub-score “1” favoring the FNH diagnosis (i.e., signal score ≤1.5; otherwise scored 0); the lesion on ADC map being not low signal was assigned a sub-score “0.5” favoring the diagnosis (i.e., signal score ≥1; otherwise scored 0); the existence of stellate scar was assigned a sub-score “0.5” (otherwise scored 0). The sum of these five sub-scores was termed LiverMss-FNH (further shortened as LiverMss in this article), and the maximum value of LiverMss was 4.5. The ORs for these 5 items were calculated for each sign in each image (see the “Results” section), and the “sub-scoring scheme” was proposed considering these ORs. The OR results were consistent with existing literature. On T2WI, FNH is typically faintly high signal (sometimes described as iso-signal in literature) (1,3,21), thus, T2WI high signal favors the diagnosis of MT. Typically, liver MT has a lower ADC relative to adjacent liver parenchyma; ADC being iso-signal or high signal favors the diagnosis of FNH (21,22). Dohan et al. described that ADC of FNH was greater than that of the spleen, and slightly lower than adjacent hepatic parenchyma (21), while Donati et al. found no difference in ADC values between FNH and the hepatic parenchyma (23). A stellate scar favors the diagnosis of FNH; however, it can be “sometimes” difficult to differentiate between FNH stellate scar and MT central necrosis, i.e., this “stellate scar pattern” was not considered pathognomonic for FNH.

Figure 1.

LiverMss-FNH scoring scheme. T2WI not high signal, DDVD iso-signal, SDC not high signal, ADC not low signal, and the existence of stellate scar favor the diagnosis of FNH. The maximum LiverMss-FNH is 4.5. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; FNH, focal nodular hyperplasia; MT, malignant tumor; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Statistical analysis for OR was performed using GraphPad Prism (GraphPad Software, San Diego, CA, USA).

Results

Out of the initially included 100 cases, 5 were excluded due to that reliable DDVD and/or SDC maps could not be obtained (caused by respiration motion). Out of the included 95 cases, 26 MT cases (13 HCC cases and 13 Mets) had a “large (median size: 8.1 cm) and very heterogeneous mass” which were very unlikely to be FNH (Figure 2). The remaining 69 cases included 13 cases of FNH, 40 cases of HCC, and 12 cases of liver Mets (6 cases from colon, 2 cases from breast, 2 cases from pancreas, 1 case from biliary duct, and 1 case from adrenal gland), and 4 cases of ICC, totaling 56 cases of MT (Table 1).

Figure 2.

Examples of 4 cases of HCC. (A-D) The lesions are too large and too heterogeneous signal to be FNH. No scoring was conducted. DDVD, diffusion-derived vessel density; DWI, diffusion weighted imaging; FNH, focal nodular hyperplasia; HCC, hepatocellular carcinoma; SDC, slow diffusion coefficient.

Table 1. Patients’ demographic information and tumor size.

Types	Gender		Age (years)	Size (cm)
	Male	Female
Excluded FNH (n=1)	0	1	30	0.8
Excluded MT (n=4)	3	1	63 [59–65]	1.4 (0.31; 0.9–1.6)
FNH (n=13)	6	7	38 [22–46]	3.8 (1.7; 1.5–7.3)
MT without scoring (n=26)	18	8	53 [32–80]	8.1 (4.2; 2.4–16.6)
MT with scoring (n=56)	44	12	55.5 [23–76]	4.9 (4.3; 1.7–22.6)

Data are presented as n, median [range], or median (SD; range) unless otherwise indicated. FNH, focal nodular hyperplasia; MT, malignant tumor; SD, standard deviation.

With the 13 FNH and 56 MT evaluated with the SQ scoring scheme, the OR results for each SQ sign are shown in Table 2. Lesion being not high signal on T2WI, being iso-signal on DDVD, being not high signal on SDC, being not low signal on ADC, and the existence of stellate scar had an OR of 49.1, 45.8, 30, 8.5, and 13.3, respectively, favoring the diagnosis of FNH. The distribution of SQ score of FNH and MT for each sign is shown in Figure 3.

Table 2. ORs of a FNH or MT lesion being not high signal on T2WI, being iso-signal on DDVD, being not high signal on SDC, being not low signal on ADC, and the existence of stellate scar.

SQ score criterion	Yes for FNH or MT (n)		OR (95% CI)	P
	Yes for FNH	Yes for MT
T2WI signal score				<0.001
0–1.5	12	11	49.1 (5.8–418.9)
≥2	1	45
DDVD				<0.001
DDVD iso signal (score =1)	11	6	45.8 (8.1–258.1)
Not DDVD iso signal	2	50
SDC signal score				0.002
0–1.5	12	16	30 (3.6–250.1)
≥2	1	40
ADC score				0.009
≥1	11	22	8.5 (1.7–42.1)
<1	2	34
Stellate scar				<0.001
Stellate central scar exists	8	6	13.3 (3.3–54.2)
No stellate central scar	5	50

ADC, apparent diffusion coefficient; CI, confidence interval; DDVD, diffusion-derived vessel density; FNH, focal nodular hyperplasia; MT, malignant tumor; OR, odds ratio; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 3.

Percentage distribution of lesion SQ scores on T2WI (A), on SDC map (C), and ADC map (D), lesion being iso-intensity or not iso-intensity on DDVD (B), and existence (+) or nonexistence (−) of stellate scar appearance (E). ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

LiverMss results for FNH and for MT are shown in Figure 4. A total of 69.2% (9/13) of the FNH had LiverMss ≥4.0. Three FNH lesions scored 3.0 had only one signal sign being untypical of FNH, and one FNH lesion scored 3.0 had respiration motion artifact. A total of 89.3% (50/56) of the MT had LiverMss ≤1.5. Six MT scored between 2.5 and 3.5 would not be firmly diagnosed as FNH as explained in Figure 4.

Figure 4.

Distribution of LiverMss for MT and FNH. A total of 69.2% (9/13) of the FNH had LiverMss ≥4.0. Three FNH lesion scored 3.0 had only one signal sign being untypical of FNH, and one FNH lesion scored 3.0 had respiration motion artifact. A total of 89.3% (50/56) of the MT had LiverMss ≤1.5. Six MT scored between 2.5 and 3.5 would not be firmly diagnosed as FNH. Blue dots denote the cases with liver cirrhosis signs visible on MRI or with a combination of ascites and apparent splenomegaly seen on MRI (all had confirmed cirrhosis or advanced liver fibrosis). Case 2 and Case 4 had histological grade 2 liver fibrosis, but no liver fibrosis/cirrhosis signal on morphological MRI. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; F, female; FNH, focal nodular hyperplasia; HCC, hepatocellular carcinoma; ICC, intrahepatic cholangiocarcinoma; M, male; MRI, magnetic resonance imaging; MT, malignant tumor; SDC, slow diffusion coefficient; T2WI, T2-weighted imaging.

Visualizations of the difference between FNH and MT, and the SQ score results, are shown in Figures 5-12. T2WI of the 5 excluded cases is shown in Figure 13.

Figure 5.

Examples of 3 cases of HCC [orange arrows in case A (A1-A4), case B (B1-B4), and case C (C1-C4)], and the corresponding SQ score on each image. (A1) The high signal inside the lesion is determined to be necrotic changes rather than central scar. (A2) Though the average SQ score is assigned to be 1, the lesion has very heterogenous signal (denoted with an asterisk). (B2) The lesion has high DDVD signal. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; HCC, hepatocellular carcinoma; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 6.

Examples of 4 cases of HCC [orange arrows in case A (A1-A4), case B (B1-B4), case C (C1-C4), and case D (D1-D4)] and the corresponding SQ score on each image. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; HCC, hepatocellular carcinoma; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 7.

Examples of 4 cases of Mets [orange arrows in case A (A1-A4), case B (B1-B4), case C (C1-C4), and case D (D1-D4)], and the corresponding SQ score on each image. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; Mets, metastasis; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 8.

Examples of 2 cases of ICC [orange arrows in case A (A1-A4) and case B (B1-B4)], and the corresponding SQ score on each image. Case B is Case 6 in Figure 4, with biliary duct dilatation noted. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; ICC, intrahepatic cholangiocarcinoma; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 9.

Examples of a case of HCC (orange arrows) and the corresponding SQ score on each image (A1-A5). (B1-B3,C1-C3,D1-D3) Adjacent slice images of b=0 and b=10 s/mm² DWI and the corresponding DDVD map. Pseudo signal on DDVD map is noted for A3, B3 and D3. The lesion appears pseudo low signal on A3 and D3, and unreliable signal on C3. The darker ring [yellow thin arrow (A3,D3)] suggests respiratory motion artifact. The shape of the lesion on C3 is inconsistent with its morphology on C1. In this case, the central necrosis also mimics stellate scar. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; DWI, diffusion weighted imaging; HCC, hepatocellular carcinoma; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 10.

Examples of 4 cases of FNH with typical signal [orange arrows in case A (A1-A4), case B (B1-B4), case C (C1-C4), and case D (D1-D4)] and the corresponding SQ score on each image. Note that, for Case D (D1-D4), central scar signal is shown both on DDVD map (D2) and ADC map (D4). ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; FNH, focal nodular hyperplasia; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 11.

Examples of two cases of FNH [orange arrows in case A (A1-A4) and case B (B1-B4)] and the corresponding SQ score on each image. The Case A lesion is the Case 8 in Figure 4. The lesion has a non-homogeneous slightly high signal on DDVD map (SQ score: 1.5, A2), however, it has stellate scar, and signal on other images are typical of FNH. Case B lesion is typical of FNH. ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; FNH, focal nodular hyperplasia; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 12.

Example of 3 atypical cases of FNH [orange arrows in case A (A1-A4), case B (B1-B4), and case C (C1-C4)] and the corresponding SQ score on each image. The Case A lesion is the Case 9 in Figure 4, with SDC high signal, no central scar, otherwise signals typical of FNH. The Case B lesion is the Case 7 in Figure 4, with T2W high signal, no central scar, otherwise signals typical of FNH. The Case C lesion is the Case 10 in Figure 4. Respiratory motion between different diffusion weighted images was noted for the lesion, and the high DDVD signal and high ADC signal are likely due to the respiratory motion (note that respiratory motion can be particularly a problem at this location of anterior abdomen). ADC, apparent diffusion coefficient; DDVD, diffusion-derived vessel density; FNH, focal nodular hyperplasia; SDC, slow diffusion coefficient; SQ, semi-quantitative; T2WI, T2-weighted imaging.

Figure 13.

T2WI of 5 cases excluded from the analysis (lesion denoted with arrows). Due to respiratory motion, DDVD and SDC maps could not be reliably assessed. (A-C) Lesions are located close to the diaphragm, thus more susceptible to respiratory motion. (D,E) Lesions are quite small. DDVD, diffusion-derived vessel density; SDC, slow diffusion coefficient; T2WI, T2-weighted imaging.

Discussion

The imaging appearance of FNH has been well described (1,3,24,25). Because FNH is mainly composed of hepatocytes, it appears similar to the background liver on unenhanced images. FNH lesions may be isointense or mildly T1-hypointense or slightly T2-hyperintense on MR images. Although FNH lesions generally have higher ADC than those of malignant lesions and hepatocellular adenomas, substantial overlap exists, and ADC is considered not a reliable differentiating feature in isolation (22). After administration of standard extracellular agents such as gadolinium-diethylenetriamine-penta-acetic acid (Gd-DTPA), FNH typically shows an intense enhancement during the arterial phase, followed by homogenous wash-out during the portal-venous and equilibrium phases which is similar to the kinetics of liver parenchyma (23-25). Hepatocyte-specific contrast agents provide additional diagnostic information and have dual functions, acting as extracellular contrast agents up to the venous phase and transitioning to hepatocyte-specific contrast agent uptake over time. Hepatobiliary phase imaging is typically performed at 20 minutes after administration of (Gd-ethoxybenzyl-DTPA) Gd-EOB-DTPA. The application of hepatocyte-specific contrast agents incur additional costs and additional MRI scan time. It will be highly relevant if majority of FNH cases, even not all cases, can be diagnosed with non-contrast MRI without the need for a contrast agent injection.

It has been recently shown that FNH had a lower DDVD value than those of HCC and Mets (20). Compared with our recent study (20), the current study further utilized earlier dataset-1 and expanded the data with additional 5 cases of FNH and 11 cases of Mets. The current study also made use of additional non-contrast scan data, including T2W, SDC, and ADC. The dataset-2 in our recent study was not utilized in the current study as there were no images for constructing SDC and ADC maps (20). In the current study, we proposed a scoring scheme, LiverMss, to evaluate lesion signal intensity as well as the existence of stellate scar of the lesion. The weighting for each sign was based on the OR of each sign (Table 2). Certain signs, including the existence of stellate scar and T2WIs being iso-signal or slightly higher signal favoring the diagnosis of FNH and lower ADC favoring the diagnosis of MT, have been well described in literature and have been further confirmed in the current study (1,3,21-23). New attention was paid to FNH being iso-signal on DDVD and not high signal on SDC maps. Higher SDC for MT than for FNH is consistent with that MT tends to have higher T2 signal and thus “waterier” than FNH. The results of the current study suggest that about 70% FNH can be confidently separated from liver MT by a combination of three diffusion metrics and T2WI, without the need for a contrast agent injection. Based on the data from the current study, if we take a conservative approach only assign a liver mass lesion to be FNH when LiverMss is ≥4.0, then unlikely a mistake will be made. Based on the data from the current study, LiverMs of ≤1.5 strongly suggest a liver mass lesion to be malignant. In this study, 6 MT cases had a LiverMss ≥2.5, while 89.3% (50/56) of the MT had LiverMss-FNH ≤1.5. Of these 6 MT cases, one was ICC and it was noted that some ICC may appear iso-signal on DDVD (26), thus ICC may score a higher LiverMss than HCC and MT though not as high as a typical FNH. Three cases had MRI-visible liver cirrhosis signs. Liver cirrhosis is associated with a longer T2 relaxation time and lower ADC of the liver parenchyma (27-31). A longer T2 and brighter background T2WI liver and spleen may lead to the lesion appearing relatively “less bright”. We suggest that LiverMss cannot be applied to diagnose FNH on apparently cirrhotic liver background, but may be applied to diagnose MT [see a further study (26)].

It should be noted that the TE for DWI dataset in this study (i.e., TE =84 ms) was relatively long as a shorter TE of around 60 ms is commonly used. In our experience till now, such a long TE is useful for the separation of FNH and MT, as a shorter TE (such as 60 ms) would lower the DDVD value for MT which has a T2 of around 60 ms (32). The liver has a T2 around 40 ms, and FNH is expected to have an only slightly longer T2 than liver as shown FNH is only slightly hyperintense on T2WI. Another point is that, besides stellate scar, we mainly considered the imaging signal without fully utilizing other morphology information. The differentiation of stellate scar from cancerous necrosis may depend on the skill and experience of a radiologist. Other morphological features, such as tumor pseudo-capsule, biliary duct dilatation (in the case of ICC), would favor MT diagnosis. In real practice, the readings based on the experience of the radiologist should always carry weight. Clinical history also offers pre-test probability favoring FNH or MT. A lesion detected in healthy young women would also favor the diagnosis of FNH. It is our expectation that, if we emphasize the diagnosis specificity, it is unlikely that an MT would be misdiagnosed as FNH. If a lesion demonstrates typical MRI characteristics of FNH during all sequences, then a confident diagnosis can be made. However, if not all classic features are present or one or more atypical features are identified, then diagnosis based on hepatobiliary contrast agent, follow-up imaging or tissue diagnosis should be pursued.

This study has a number of limitations. This is a single center study and the number of FNH was limited (n=13). Data from different scanners and different centers will be acquired to validate the findings of the current study. The data acquisition protocol can be further improved in future studies. For example, the NEX was only 1 for images of b-value =0, 10 s/mm². Misalignment between b=0 and b=10 s/mm² images, or between b=500 and b=800 s/mm² images, or between b=0 and b=800 s/mm² images can lead to erroneous DDVD/SDC/ADC signal (14,20), while the imaging data in this study were not acquired with a breathhold. As this study uses previously acquired liver IVIM data, the selection of the four b-values was out of convenience. However, these four b-values are probable of reasonable choice. For many standard clinical scanners, the lowest non-zero b-value is 10 s/mm², and liver DWI image with b>800 s/mm² likely contains a high level of noises (8). For SDC mapping, our initial testing also showed that, by further increasing the interval between the two high b-value images (500–800 s/mm² in this study, with an interval of 300 s/mm²), the signal-to-noise ratio of the SDC map will further increase, the lesion to (adjacent) tissue contrast may decrease. Lesions located near the diaphragm or near the anterior portion of the liver are more susceptible to respiration motion artifact. Single breathhold imaging for DDVD or SDC data is feasible in principle (8,13), and it is anticipated that artificial intelligence powered accelerated data reconstruction will enable single breathhold imaging of DDVD and SDC with sufficient signal-to-noise ratio (33,34). As noted in our recent study, a drop from b=0 s/mm² DWI high signal or slightly high signal to DDVD iso-signal suggests the diagnosis of FNH (20). This feature was not applied in the current study. As noted in the method section, the definition of a stellate scar within the lesion is associated with a degree of subjectivity and this depends on the experience and skill of the reading radiologist. According to this study, the existence of it offers an OR of 13.3 [95% confidence interval (CI): 3.3–54.2] favoring the diagnosis of FNH and it was allocated a LiverMss sub-score of 0.5. We have not studied FNH-like lesions within an abnormal liver background (such as the cases of Fontan-associated liver disease, Budd-Chiari syndrome), and this will be topic for further study. This study did not differentiate between three types of MTs (i.e., HCC, Mets, and ICC). We did not compare the differential diagnosis performance of DDVD with those of contrast-enhanced CT/MRI/ultrasound, as our goal is to provide a more cost-effective method without the need for contrast agent administration. Finally, future studies of less common pathologies, such as hepatic adenoma, etc., will be of interest. Hepatic adenoma is a rare neoplasm that is most common in young women and carries substantial risks for malignant transformation and hemorrhage (35,36).

Conclusions

In conclusion, LiverMss, which reflects a liver mass lesion’s signal features with T2WI and three diffusion metrics of DDVD, SDC, and ADC, offers practical diagnostic separation of liver FNH and MT for the majority of patients. More than 2/3 of liver FNH, as long as they are of reasonable size and the images are not degraded by respiratory motion, can be diagnosed by LiverMss without the need for contrast agent administration. It is anticipated that diagnostic confidence can be further enhanced when other MRI morphological signs (such as the existence of a pseudo-capsule) and patient history are considered.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All imaging data were acquired with institutional ethical approval from the ethics committee of Sun Yat-Sen Memorial Hospital (approval ID, S1709442250839) and with informed consent obtained from individual participants.

Data Sharing Statement

Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-aw-860/dss

jgo-17-01-20-dss.pdf ^{(138.2KB, pdf)}

DOI: 10.21037/jgo-2025-aw-860