2.2. 2010
The year 2010 represented a turning point regarding the issue of the non-invasive diagnosing of hepatic metastases. In fact, in their large meta-analysis, Niekel et al. confirmed the highest sensitivity on a per-patient basis of FDG-PET vs. MRI and CT (94.1%, 88.2%, and 83.6%, respectively)
[7]. However, MRI reached the best sensitivity on a per-lesion basis, especially for lesions smaller than 10 mm (80.1%). For these reasons, the authors concluded that the preferred first-line modality for evaluating CRLM should be represented by MRI, while FDG-PET should be used as a second-line modality.
In the same year, in accordance with the meta-analysis of Niekel et al.
[7], Floriani et al. analyzed the per-lesion sensitivity of all the above-mentioned imaging techniques, such as FDG-PET, MRI, and CT in their meta-analysis
[8]. In agreement with Niekel et al., Floriani et al. confirmed the highest sensitivity on a per-patient basis of FDG-PET vs. MRI and CT (93.8%, 81.1%, and 74.8%, respectively) and the longstanding belief that MRI showed the same sensibility as FDG-PET (86.3% and 86%, respectively) and a better sensitivity than CT (82.6%;
p < 0.0001) in per-lesion analysis
[8]. Moreover, MRI sensitivity was significantly different when liver-specific contrast agents were administered (86.3%) and the preferential use of MRI for the detection of liver metastases was confirmed.
However, the two meta-analyses published in 2010 presented some differences. The first was the different time period analyzed by the two groups. Niekel et al.
[7] carried out a comprehensive search for articles published from January 1990 to January 2010, a very long period in which many technical developments were introduced, probably causing bias. On the other hand, Floriani et al.
[8] carried out a search for articles published from January 2000 to August 2008, a very short period. The second difference was the choice of only evaluating papers which used MRI performed with hepatobiliary contrast agents. In fact, Niekel et al.
[7] stated that data regarding liver-specific contrast materials, such as gadobenate dimeglumine (Gd-BOPTA, Multihance; Bracco, Princeton, NJ, USA) or gadoxetic acid (Gd-EOB-DTPA, Primovist; Bayer Schering, Berlin, Germany), were unavailable or limited to other focal liver diseases such as hepatocellular carcinoma. Instead, in their analysis of 25 articles, Floriani et al.
[8] included eight articles which utilized liver-specific contrast media, such as mangafodipir trisodium (MnDPDP) or resovist (SPIO). Therefore, in the same year (2010), MRI was considered to be the first-line technique in detecting liver metastases according to two different meta-analyses, and, moreover, MRI with hepatobiliary contrast agents was demonstrated to have the best sensitivity in this field according to one meta-analysis. However, MRI did not appear in the majority of the most widely used international oncological guidelines.
The year 2010 and those immediately thereafter represented decisive moments in liver imaging. In fact, on the one hand, robust data demonstrated that MRI with hepatospecific contrast agents was the most sensitive imaging technique in evaluating liver metastases. On the other hand, some hepatospecific contrast media commercially available at that time, such as MnDPDP or SPIO, have since been withdrawn from the market. Fortunately, in the same years, the “second generation” of hepatobiliary contrast agents was diffusely utilized in current clinical practice, continuously changing MRI evaluation of the liver. These new contrast media were Gd-EOB-DTPA and Gd-BOPTA. In particular, as compared with the previous “old” liver-specific contrast agents (i.e., MnDPDP or SPIO), the new hepatobiliary contrast agents furnished both the dynamic phases (arterial, portal, and delayed phases) and the hepatobiliary phase in a single injection, reducing examination time and complications. In 2010, in one of the first published series concerning Gd-EOB-DTPA MRI, Choi et al. demonstrated that this technique had a sensitivity of 90.3% in detecting liver metastases ≤1 cm
[9]. As a consequence, MRI performed with the “second generation” of hepatobiliary contrast agents became the modality of choice in clinical practice for the detection of hepatic metastases, surpassing the other dynamic imaging techniques. In 2014, an important randomized multicenter trial (VALUE) compared the diagnostic performance of three different imaging techniques such as Gd-EOB-DTPA MRI, MRI with extracellular contrast medium, and contrast-enhanced CT in patients with suspected CRLM
[12]. The study showed the diagnostic superiority of Gd-EOB-DTPA MRI as compared to the other imaging modalities. In fact, additional imaging was required in 0 of 118 patients initially evaluated with Gd-EOB-DTPA MRI, in 19 (17.0%) of 112 patients initially assessed using MRI with extracellular contrast agent, and in 44 (39.3%) of 112 patients primarily evaluated with contrast-enhanced CT (
p < 0.001). Moreover, the authors demonstrated that confidence in diagnosis and therapeutic decision were high or very high in 98.3% of patients when the initial imaging modality was Gd-EOB-DTPA MRI, significantly superior (
p < 0.001) to MRI with extracellular contrast medium (85.7%) and contrast-enhanced CT (65.2%)
[12]. Furthermore, the information provided by Gd-EOB-DTPA MRI resulted in a decreased number of patients with intraoperative modifications of the established surgical plan in patients undergoing liver resection
[12]. The observation that a higher percentage of patients, in the group where Gd-EOB-DTPA MRI was the initial imaging procedure, reached a higher successful treatment was particularly interesting. Perhaps the higher confidence in the preoperative imaging stimulated more audacious treatment decisions, especially regarding more aggressive surgical approaches.
In the same period in which the vast majority of scientific studies was focused on the use of the new second-generation of hepatospecific contrast media, a new MRI technique was also taking hold, namely diffusion-weighted imaging (DWI). Diffusion-weighted imaging measures the degree of diffusion of water molecules in biological tissues in vivo; diffusion represents the random motion of water molecules, also known as Brownian motion; during its motion, each particle collides with those nearby and moves, following an erratic course
[13]. The diffusion of water molecules is a biophysical parameter which correlates with the structural characteristics of tissues under both physiological and pathological conditions
[13]. The rapid evolution and therefore diffusion of DWI worldwide has led to an important milestone in abdominal radiology and, in particular, in liver imaging. In fact, many papers have evaluated the value of DWI in the detection and characterization of focal liver lesions. In particular, many authors compared the diagnostic performances of DWI-MRI, Gd-EOB-DTPA MRI and the combination of both methods in evaluating liver metastases
[14][15][16][17][18]. In a prospective study in 2015, Kim et al. clearly demonstrated that Gd-EOB-DTPA MRI combined with DWI was much more accurate than contrast-enhanced CT (98% vs. 85%, respectively), having important effects in patients with potentially resectable disease
[19]. Subsequently, in a prospective study, Schulz et al. confirmed the significantly highest per-lesion sensitivity of combined Gd-EOB-DTPA MRI with DWI with respect to all the other imaging modalities, such as CT and FDG-PET (90%, 68%, and 61%, respectively), especially for lesions <10 mm (74%, 16%, and 9% respectively)
[20]. Therefore, these authors
[19][20] concluded that the detection of liver metastases should be based on MRI as a first-line imaging technique in order to avoid underestimation of liver involvement and ensure correct management. It is important to underline that the differential uptake of FDG in liver metastases above the liver background update is mainly due to the upregulation of different glucose transporters: GLUT2 in liver parenchyma and different transporters in metastases depending on primary tumors (i.e., metastases from lung and breast cancers are GLUT5 positive while those from colorectal cancer are GLUT3 positive). This aspect will affect the sensitivity of this technique.
2.3. 2016
After all these experiences, the year 2016 represents the third and last fundamental step in the imaging of liver metastases. In fact, in 2016, a systematic review and meta-analysis, involving more than 1200 liver lesions evaluated over a 15-year period, confirmed the higher per-lesion sensitivity of Gd-EOB-DTPA MRI as compared to contrast-enhanced CT (median 94.9% vs. 74.2%, respectively;
p < 0.001) without statistical differences in terms of specificity (median 86.6% vs. 94.1%;
p = 0.44)
[21]. The superiority of Gd-EOB-DTPA MRI over contrast-enhanced CT was also demonstrated for lesions <1 cm (per-lesion median sensitivity of 85.7% and 50%, respectively;
p < 0.001).
Moreover, in the same year (2016), Vilgraine et al. published another comprehensive meta-analysis in which they included 1989 patients with 3854 hepatic metastases
[10]. The authors, assuming the superiority of MRI over other imaging techniques such as CT, analyzed the sensitivity of DWI alone, Gd-EOB-DTPA MRI alone, and a combination of the above techniques (DWI and Gd-EOB-DTPA MRI) for detecting liver metastases on a per-lesion basis. When considered independently, DWI-MRI was less sensitive than Gd-EOB-DTPA MRI for detecting liver metastases (87.1% vs. 90.6%); however, the combination of both techniques demonstrated the highest value of per-lesion sensitivity (95%;
p < 0.0001). Moreover, the authors found that similar results were observed in articles which compared the three techniques simultaneously, with only CRLM and liver metastases smaller than 1 cm. The conclusion was that in metastatic patients, the combination of DWI and Gd-EOB-DTPA MRI ensured the highest sensitivity for detecting liver metastases on a per-lesion basis.
The economic implications of the initial imaging with MRI in patients with suspected liver metastases have always been a matter of debate in the scientific community. In the majority of cases, this is an argument in favor of the use of CT instead of MRI as the first-line technique, together with the scarce spread of MRI in the world. However, there were no robust scientific evidences on these issues until 2016.
In fact, in 2016, Zech et al. assessed the costs of diagnostic workup and surgery for patients with liver metastases from colorectal cancer, comparing three different imaging strategies: Gd-EOB-DTPA MRI, MRI with extracellular contrast media, and contrast-enhanced CT
[22]. Surprisingly, the cost of a diagnostic workup in the majority of countries analyzed, including Austria, Germany, Italy, Sweden, Switzerland, and Thailand, was lower when Gd-EOB-DTPA MRI was used as the initial imaging technique as compared with the other strategies. In fact, although Gd-EOB-DTPA MRI was the most expensive imaging modality, no patient in the Gd-EOB-DTPA MRI group required additional imaging examinations in order to reach a decision regarding treatment as compared to 18.1% and 39.7% of the patients in the extracellular contrast media-enhanced MRI and contrast-enhanced CT groups, respectively. Specifically, in the majority of European countries, the diagnostic workup costs were higher due to the need for additional MRI procedures, and the cost of surgery was higher in the Gd-EOB-DTPA MRI group since significantly more patients underwent surgery for a curative approach. Therefore, the superior sensitivity of Gd-EOB-DTPA MRI in detecting liver metastases, the benefits in avoiding additional imaging examinations, and similar diagnostic workup costs suggest that Gd-EOB-DTPA MRI should be the preferred initial imaging modality for evaluating liver resectability in patients with hepatic metastases (). In another field of research, such as hepatocellular carcinoma, a study was conducted regarding the economic impact on the management of these patients, using the three different techniques as starting imaging, namely Gd-EOB-DTPA MRI, MRI with extracellular contrast medium, and contrast-enhanced CT
[11]. Using a mathematical model, the authors demonstrated that, starting from a similar budget for the three different methods, Gd-EOB-DTPA MRI promoted cost savings, reducing treatments for false positive patients and reducing the requirement for other examinations or biopsies
[11]. The main advantage of this approach was that the major part of the money could be used to treat the true positive patients. The authors believe that these results can also be demonstrated in the field of patients with liver metastases in which the use of Gd-EOB-DTPA MRI could reduce unnecessary treatment for false positive patients and could reduce the use of other imaging techniques or biopsies, thus allowing the use of increasingly limited economic resources for the treatment of those patients who could really benefit from treatment, especially for aggressive ones.
Figure 1. A 61-year-old male with a clinical history of gastric cancer. (A) 18-fluorideoxyglucose positron emission tomography (18FDG-PET) image shows two focal lesions in the segment VIII of the liver with increased uptake (SUVmax 4.7) (arrowhead and arrow). (B) In the axial computed tomography (CT) scan image, no suspicious lesions were visible. Diffusion-weighted imaging (DWI)-Gd-EOB-DTPA magnetic resonance imaging (MRI) examination showed only one slight hyperintense focal lesion on T2-weighted images (arrow in (C)) with strong hyperintensity in diffusion-weighted image (arrow in (D)) and hypointensity during the hepatobiliary phase (arrow in (E)), suspected for a liver metastasis. Moreover, a very small (2 mm) focal lesion with hyperintensity on diffusion-weighted image (arrow in (F)) and hypointensity during the hepatobiliary phase (arrow in (G)) was also detectable cranially in the same liver segment. An axial contrast-enhanced CT scan performed after six months confirmed this hepatic metastasis (arrow in (H)).