ª Springer Science+Business Media New York 2015

Abdominal Imaging

Abdom Imaging (2015) DOI: 10.1007/s00261-015-0507-8

Role of imaging for patients with colorectal hepatic metastases: what the radiologist needs to know Michael J. Reiter,1 Nathan P. Hannemann,2 Ryan B. Schwope,2,3 Christopher J. Lisanti,2,3 Peter A. Learn4 1

Department of Radiology, Stony Brook University Medical Center, HSC Level 4, Room 120 East Loop Road, Stony Brook, NY 11794, USA 2 Department of Radiology, Brooke Army Medical Center, San Antonio, TX, USA 3 Uniformed Services University of the Health Sciences, Bethesda, MD, USA 4 Department of Surgery, Brooke Army Medical Center, San Antonio, TX, USA

Abstract Surgical resection of colorectal metastatic disease has increased as surgeons have adopted a more aggressive ideology. Current exclusion criteria are patients for whom a negative resection margin is not feasible or a future liver remnant (FLR) of greater than 20% is not achievable. The goal of preoperative imaging is to identify the number and distribution of liver metastases, in addition to establishing their relation to relevant intrahepatic structures. FLR can be calculated utilizing crosssectional imaging to select out patients at risk for hepatic dysfunction after resection. MRI, specifically with gadoxetic acid contrast, is currently the preferred modality for assessment of hepatic involvement for patients with newly diagnosed colorectal cancer, to include those who have undergone neoadjuvant chemotherapy. Employment of liver-directed therapies has recently expanded and they may provide an alternative to hepatectomy in order to obtain locoregional control in poor surgical candidates or convert patients with initially unresectable disease into surgical candidates. Key words: Colorectal cancer—Hepatic metastases—Portal vein embolization—CT—MRI

Colorectal cancer is the third most common cancer in both men and women in the United States as well as the

Presented as an educational exhibit at RSNA 2014. Correspondence to: Michael J. Reiter; email: [email protected]

rest of the world [1]. Up to 65% of patients develop metastases either at the time of diagnosis or over the course of their disease and the liver represents the most common anatomic site, seen in 33–50% cases [2–5]. Approximately, 20% of patients with hepatic metastases will have liver-only disease and the extent of this disease is one of the main determinants of survival. The median survival is 5–12 months for patients who are left untreated while 5-year survival following palliative chemotherapy alone is exceedingly rare [3]. Marginnegative surgical resection, also referred to as R0 resection, is known to represent the only curative therapy and, as a result, has been the standard of care for colorectal hepatic metastases [6]. Resection has resulted in 5-year survival rates reported between 30–60% [2, 7]. The significant improvement in 5-year survival rates for patients with colorectal hepatic metastases over the past two decades is in part due to a paradigm shift in the philosophy of treatment, in addition to improved chemotherapeutic agents [8]. Rather than focus on which metastatic lesions can be removed, the current emphasis instead is on what should be left behind [9]. This, along with improved surgical techniques, has permitted treating physicians to become more aggressive and further expand the pool of patients who are considered candidates for resection. Nonoperative liver-directed therapies can also now be administered with curative intent for those patients who are not surgical candidates. Coinciding advances in medical imaging have seen the radiologist play an increasing role in the staging and management of colorectal cancer patients. However, the landscape of treatment is continually changing and radiologists need to stay up to date in order to provide

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

relevant information. This article will review the indications, both current and historic, for surgical resection of colorectal hepatic metastases, examine the accuracy of various imaging modalities employed today, and highlight the pertinent radiologic features which critically impact the implementation of an optimal treatment plan.

Indications for surgical resection Historical perspective The fundamental theory which served as the basis for the fervent surgical treatment of colorectal hepatic metastases was the belief that malignant colorectal cancer cells spread to the liver first. Removing metastatic lesions thereby prevented disease dissemination from the liver to other anatomic sites. Although it is now realized that the biology of metastasis is more nuanced, it remains clear that patients with liver-only metastatic disease have the potential to be cured by resection. The goal of surgical resection is therefore removal of all metastatic disease from the liver while preserving sufficient functional parenchyma. Restrictive criteria for eligibility of resection were introduced in the late 1980s and were as follows: four or fewer metastatic lesions, maximum tumor diameter less than 5 cm, the absence of extrahepatic disease, and a negative surgical margin of at least 1 cm (Fig. 1) [10, 11]. However, all of these criteria have been challenged over the past decade as they were derived from antiquated studies and the definition of resectability has been modified [7].

Current criteria There has been incentive to expand the criteria for resectability to include a larger number of patients since it has been realized that the only true constraints are the ability to achieve negative margins and maintain adequate hepatic reserve. More precisely, resectability is no longer defined by what is being removed but now focuses on what will remain after excision and these criteria continue to evolve over time [4]. A recent expert consensus concluded that a minimum of two contiguous hepatic segments with adequate vascular inflow and outflow as well as biliary drainage comprising at least 20% of the initial liver volume should remain following surgery [6]. Therefore, up to six anatomical segments can be safely resected in patients with normal liver function without inducing postoperative hepatic failure. The amount of liver parenchyma that remains after resection is commonly referred to as the future liver remnant (FLR). The definition of an adequate FLR is based on the patient’s baseline preoperative liver function. Typical designations include an FLR of at least 20% for normal patients, 30–40% in the setting of steatosis, or 40% or greater in the presence of cirrhosis [4].

Fig. 1. Two patients with colon cancer and hepatic metastases with differing treatment strategies based on historical criteria for resection. A 61-year-old man with isolated 4.5 cm mass in segment 4a (white arrow) on axial CT, which would be an ideal lesion for surgical resection as it is less than 5 cm in size and it is more than 1 cm from the left portal vein (white arrowhead). B 45-year-old man with bilobar masses in segments 2 and 7 (black arrows) on axial CT. This patient would have been classified as nonresectable based on historical criteria.

Hepatic metastases are considered nonresectable if there is invasion of both main portal pedicles, invasion of one portal pedicle and the contralateral hepatic vein, invasion of all three major hepatic veins, or if complete removal of all metastatic deposits results in an inade-

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

quate FLR [12]. Inability to obtain an R0 resection is also a contraindication to surgery and the most common scenario in which a positive margin will occur is close proximity of the tumor to the portal vein, hepatic vein, or vena cava [9]. Local extension of hepatic metastases into adjacent structures and intraluminal biliary tumor extension do not qualify as true extrahepatic disease and, as such, should not serve as a contraindication to surgery when present in isolation. Much more controversy exists regarding the eligibility for hepatic resection of patients with more diffuse extrahepatic metastatic disease. Pulmonary metastases tend to be treated as a unique subset since survival rates are remarkably high for patients undergoing combined liver and lung lesion resection in the setting of no other known sites of disease [4]. The segmental anatomy of the liver was first described by Couinaud et al in 1957 and it comprised eight segments, each with a portal vein, hepatic artery, bile duct, and hepatic venous branches. Consequently, each segment can be resected separately without affecting the remaining liver segments [13]. Hepatectomy is classified as minor if one or two segments are removed, major when three or more segments are resected and extended when 5 or more segments are excised. However, hepatectomy is not required to be segmental or lobar in approach and non-anatomic resections are possible as well. The type of resection, wedge (non-anatomic) vs. segmentectomy (anatomic), has no impact on patient outcome. Wedge resections have become more common of late as a means to preserve hepatic parenchyma whilst treating a greater number and broader distribution of metastatic lesions (Fig. 2) [14]. Several prognostic scoring systems have been developed which incorporate clinical factors such as patient age, serum carcinoembryonic antigen level, and diseasefree interval in an attempt to aid surgeons with the choice of whether or not to operate based on a prediction of patient outcome after resection. Fong’s clinical risk score is the most widely utilized system today. However, none of these scoring systems are relied upon for absolute determination of resectability due to inherent inaccuracies and as such, they function solely as a guideline [7]. Preoperative radiologic imaging, therefore, plays the primary role in staging of colorectal cancer, specifically with regard to the feasibility of hepatic resection.

Role of preoperative imaging for staging and surgical planning Detection of colorectal hepatic metastases Various imaging modalities are available for the detection of colorectal hepatic metastases. Ultrasonography (US) was widely employed in the past but currently has

Fig. 2. Examples of the two types of hepatic resection. A Axial CT of a 46-year-old man who had several right hepatic lobe colon cancer metastases and underwent right lobectomy. Note the surgical clips along the margin of the resection plane (white arrow). The remaining left hepatic lobe is hypertrophied due to preoperative portal vein embolization. B 62-year-old woman with several right hepatic lobe metastases from rectal cancer. Wedge resection of lesions in segments 5 and 6 was performed as evidenced by surgical clips (black arrow) on axial CT. Note the difference in the amount of residual parenchyma following this non-anatomic resection as compared to right lobectomy in A.

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

fallen out of favor and is rarely utilized due to its inferior performance compared to computed tomography (CT) and magnetic resonance imaging (MRI). However, the introduction of contrast-enhanced US may change this trend as it permits the identification of more hepatic metastases compared to routine US and also improves the specificity for characterizing focal liver lesions. The advantages of CT include its speed of image acquisition, relative inexpensive cost, and propensity to detect extrahepatic sites of metastatic disease as well. The portal venous phase remains the optimal imaging time for detection of hepatic metastases on CT. The choice of slice selection is important for visualization of small metastatic lesions, and 2.5 or 3.75-mm-thick slices have been shown to be superior to 5-mm-thick slices. However, when reduced to 1-mm thickness, no further improvement in lesion detection is appreciated. As a result, a slice thickness of 2–3 mm is ideal [15]. Drawbacks of CT are radiation exposure to the patient and difficulty in characterizing lesions less than 1 cm in size. Due to the high metabolic activity of the normal liver, Fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT may not be helpful in imaging intrahepatic lesions, particularly smaller masses, although it can identify radiologically occult extrahepatic disease, altering treatment plans in up 24% of cases [16– 19]. However, the value of FDG PET/CT for staging has been questioned of late as it may not result in as frequent changes in surgical management as initially reported [20]. The primary limitations of FDG PET/CT include its reduced sensitivity in detecting subcentimeter lesions, lesions with high mucin content, and lesions which have undergone neoadjuvant chemotherapy treatment, the latter of which has a reported drop in sensitivity for lesion detection to 60–70% (compared to sensitivity of 90% in chemonaive patients) [9]. In theory, the combination of PET with MRI is preferable to PET/CT as there is less radiation exposure to the patient as well as better soft tissue contrast resolution. The utilization of PET/MRI in the realm of identification of colorectal hepatic metastases is an exciting potential future application. MRI is the preferred method of evaluating the liver in patients with newly diagnosed colorectal cancer based on a meta-analysis revealing it to be the most sensitive modality (mean sensitivity of 88.2% and mean specificity of 92.5% on per patient basis), particularly for lesions less than 10 mm in size [21]. However, a large number of institutions still employ CT as the primary imaging modality and reserve MRI for problem-solving situations only or for those patients set to undergo resection [13, 17, 22]. The authors presume this is due to greater familiarity of surgeons with CT as compared to MRI and FDG PET/CT in addition to the more widespread availability of CT. CT still flaunts a high performance in lesion detection, most notably amongst radiologists with substantial experience with this modality [23]. In fact, the

diagnostic yield of MRI for hepatic metastases is low for newly diagnosed colorectal cancer patients with either negative CT results or the presence of too small to characterize hypoattenuating lesions [24]. MRI utilizing hepatocyte-specific agents such as gadoxetic acid (Gd-EOB-DTPA, Eovist; Bayer Healthcare Pharmaceuticals) boasts even higher sensitivities (range, 92–98%) compared to MRI with conventional extracellular contrast agents [25–27]. It is a useful technique for assessing equivocal hepatic lesions identified on preoperative staging CT for patients with newly diagnosed colorectal cancer (Fig. 3) [28]. The addition of diffusion weighted imaging (DWI) permits detection of small lesions adjacent to hepatic vessels that otherwise would be obscured on postcontrast images due to the suppression of intrahepatic vascular signal on DWI [29]. Despite having the highest upfront cost amongst most imaging modalities, MRI with gadoxetic acid actually shows the best cost savings when reimaging and unnecessary surgical procedures are factored into the equation [30]. Therefore, MRI with gadoxetic acid is the recommended first-line imaging study.

Typical imaging features of colorectal hepatic metastases Untreated colorectal hepatic metastases are variable in appearance, although they typically manifest as hypoattenuating masses as compared with normal surrounding parenchyma on noncontrast CT, either related to a lack of glycogen within metastases or a relative increase in the amount of water in the extracellular spaces of malignancies. In a background of hepatic steatosis, however, metastases may be isodense or even hyperdense due to the abnormal low attenuation of the fatty liver and, thus, may be overlooked [31]. Heterogeneity is standard appearance for colorectal hepatic metastases on CT, the degree of which is dictated by the relative proportions of mucin, fibrosis, and calcification [32]. Up to 11% of colorectal liver metastases are calcified at initial presentation, which are often best visualized on unenhanced CT [15]. Characteristic features include ill-defined margins and attenuation higher than that of normal fluid, such as bile or cerebrospinal fluid (CSF). This is in contradistinction to hepatic cysts, which are sharply marginated and isodense with fluid (Fig. 4) [33]. The ring-like enhancement surrounding hepatic metastases during the arterial phase corresponds to nontumoral tissue with inflammatory infiltration or vascular proliferation [31]. Metastatic lesions may centripetally fill in with contrast over time, mimicking hepatic hemangiomas. However, complete filling is more suggestive of hemangiomas as is a discontinuous and nodular ring of enhancement in the earlier phases [34]. On MRI, colorectal hepatic metastases are typically ill-defined with mild to moderate hyperintense signal

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Fig. 3. 52-year-old man with colon cancer and indeterminate subcentimeter hypoattenuating hepatic lesion on initial staging CT (not shown). A Axial postcontrast T1-weighted image with fat saturation obtained during the arterial phase shows an 8-mm lesion in segment 6 (arrow) with incomplete peripheral hyperenhancement and central low signal intensity. B Axial postcontrast T1-weighted image with fat saturation obtained during the portal venous phase reveals the lesion to homogeneously

fill-in and has slightly hyperintense signal relative to the adjacent normal hepatic parenchyma (arrow). This lesion could simulate a hemangioma based on its enhancement characteristics. However, the lesion is only mildly hyperintense (arrow) on axial HASTE image, C, and also demonstrated restricted diffusion (not shown) which is suspicious for metastasis. Surgical resection confirmed the lesion to be metastatic from patient’s primary colon malignancy.

relative to the normal liver on T2-weighted images and corresponding low signal on T1-weighted images. Continuous rim enhancement with central progression over time can be seen after contrast administration [33]. For necrotic lesions, however, there will be a lack contrast filling centrally. With use of hepatocyte-specific contrast

agents, both rim enhancement and a ‘‘target sign’’ (central hyperenhancement and a hypointense rim) have been described for metastases on the hepatobiliary phase, although homogeneous hypointense lesions are most common (Fig. 5) [30]. Lesions can be classified as benign if they demonstrate homogeneously high signal intensity

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Fig. 4. 50-year-old man with colon cancer metastasis in hepatic segment 6 on noncontrast axial CT. The lesion is not sharply marginated nor isoattenuating with fluid, which distinguishes it from a hepatic cyst. The hyperattenuating portion of the lesion represents partial calcification.

on T2-weighted images similar to that of CSF and absent enhancement (cyst) or discontinuous nodular enhancement with delayed fill-in (hemangioma) [33].

Critical imaging findings to report For patients with colorectal hepatic metastases, several goals of preoperative imaging exist. First, the number and size of the metastatic lesions, in addition to their lobar distribution, needs to be described. The Couinaud segmental anatomy is the most universally accepted system and should therefore be used to illustrate the location of all masses. The number of lesions provides prognostic information and may affect the decision to perform a combination procedure including both resection and local ablative therapy if there are too numerous small masses to excise (Fig. 6). Size also is used to determine prognosis and, additionally, can impact the feasibility of ablative therapies, particularly radiofrequency ablation (RFA), as they are inadequate for masses greater than 3 cm [16]. It is important to recognize that not all hepatic lesions detected during staging will represent metastases. Small lesions, defined as those 15 mm or less in diameter, have a high probability of being benign, even in patients with a known extrahepatic malignancy [35]. A previous study of hypodense liver lesions found in women with breast cancer and which were ‘‘too small to characterize’’ (less

Fig. 5. 62-year-old woman with colon cancer hepatic metastases. A Axial gadoxetic acid postcontrast T1-weighted image with fat saturation obtained during hepatobiliary phase (20 min after contrast injection) demonstrates one lesion in segment 6 (arrow) with mild central enhancement (although lesion remains hypointense relative to the normal adjacent hepatic parenchyma). There is a peripheral rim of low signal intensity and this appearance has been described as the ‘‘target sign.’’ B Axial T2-weighted image with fat saturation shows the same lesion (arrowhead) to have only mild hyperintense signal relative to the normal hepatic tissue, distinguishing it from cysts and hemangiomas which tend to be isointense with fluid.

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Fig. 6. 44-year-old man with bilobar hepatic metastases from colon cancer on coronal CT. This patient would ideally undergo chemotherapy in an attempt to decrease the tumor burden and a combination of wedge resection and ablative therapy would then be a viable treatment option if several masses persisted throughout the liver which were too small to excise in total.

than 15 mm in diameter) demonstrated that 93% were benign [36]. A similar study assessed hepatic lesions less than 10 mm in diameter in patients with various primary malignancies and found that 14% were malignant in the subset of patients with colorectal cancer [37]. Comparison with prior imaging studies is the most helpful tactic in these cases, as demonstration of stability for 12 months or greater supports benignity while a new hepatic lesion raises suspicion for malignancy. Close attention to small indeterminate lesions on follow-up surveillance imaging studies is also critical. Second, a determination of whether or not surgical resection is technically feasible is undertaken. This is further subdivided into an assessment of whether the metastatic disease burden can tangibly be removed in its entirety as well as if such a resection will leave behind an adequate FLR to prevent hepatic failure. To judge for the possibility of resection, the primary focus is on the proximity of lesions to vital structures, such as the vascular and biliary tree. If there is metastatic disease abutting the right or left portal vein on CT or MRI then the presumption is that a negative margin will not be obtained without taking the vessel as well and, therefore, the entire lobe will need to be removed. Again, involvement of the bifurcation of the main portal vein or all three major hepatic veins is a contraindication to resection. If a lesion abuts a vascular structure that must be preserved and cannot be removed then these patients are

best treated with chemotherapy prior to resection in order to shrink the mass and decrease the chance of a positive margin (Fig. 7) [38]. Intrabiliary invasion and portal venous thrombosis, while not absolute contraindications to resection, can significantly impact the surgeon’s decision. Biliary duct invasion by colorectal liver metastases is an uncommon pattern of dissemination but is important to recognize preoperatively as it can result in partial tumor resection due to underestimation of disease spread. Anatomic right or left lobectomies are preferred in this setting in lieu of wedge resection. Similarly, the phenomenon of insufficient margins with biliary invasion should be considered for local therapies such as RFA, as it puts the patient at risk for a higher likelihood of local recurrence. A parenchymal metastatic lesion associated with dilated biliary ducts, either upstream or downstream, suggests ductal invasion [39]. In order to ensure that an adequate amount of functioning hepatic parenchyma will be left behind following resection, the FLR can be calculated using preoperative cross-sectional imaging, either CT or MRI. Both modalities can provide accurate and reproducible measurements with the various post-processing software packages available today (Fig. 8) [40]. While the FLR is a reliable predictor of postsurgical hepatic dysfunction, an absolute FLR volume is not exact in isolation given that larger patients require a larger amount of parenchyma compared to smaller patients. The solution is to standardize the FLR to patient size and this is expressed as a percentage of the total liver volume (TLV). The TLV is measured first and then the volume of metastatic disease is subtracted from it. Alternatively, TLV can be calculated using a formula based on its association with body surface area (BSA). This formula has been shown to correlate with clinical outcomes and is as follows: TLV = -794.41 + 1267.28 9 BSA (m2) [41]. MRI with gadoxetic acid may help in determining the risk of liver failure after major hepatic resection based on the preoperative relative enhancement of the hepatic parenchyma during the hepatobiliary phase. Previous research has shown that the lower the degree of enhancement, the greater the likelihood of postoperative failure [42]. For those patients in whom the FLR is too small to prevent hepatic failure after resection, portal vein embolization can be used to increase the FLR and expand the pool of surgical candidates. Portal vein embolization has long been recognized to induce compensatory hypertrophy of the contralateral lobe and, as such, has become an integral part of liver surgery [17]. Polyvinyl alcohol particles, platinum coils, cyanoacrylate, and gelatin are commonly used materials for embolization [43]. Contraindications to portal vein embolization include portal hypertension, uncorrectable coagulopathy, renal failure, ipsilateral portal vein tumor

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Fig. 7. 63-year-old man with colon cancer. A Axial CT image performed for staging reveals innumerable bilobar hepatic metastases (white arrows). Neoadjuvant chemotherapy was initiated with the goal of decreasing his tumor burden to allow for surgical resection. Two axial CT images performed 5 months later, one more cranial (B) and the other more caudal (C), show interval decrease in size of many of the previously demonstrated masses (white arrows). However, there is new left portal vein thrombosis (white arrowhead).

There is also a new metastatic lesion present in segment 6/7 (black arrow) which abuts the right portal vein (black arrowhead). Despite the positive partial response to therapy, this patient is a nonsurgical candidate as the left portal vein thrombosis would necessitate left lobectomy for treatment of left sided disease while right lobectomy would be required to extirpate all right sided disease since wedge resection of the new segment 6/7 lesion could yield a positive margin given its proximity to the portal vein.

thrombus, and the absence of a normal portal vein bifurcation. Recognition of variations in the ‘‘standard’’ portal venous anatomy on preoperative imaging is important to avoid injury to the FLR secondary to

undesirable embolization. While only representing a relative contraindication, biliary obstruction is important to note since it will require biliary drainage prior to embolization [44].

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

avoid misinterpretation. The peripheral regions of the treated liver segments can demonstrate increased enhancement during the arterial phase because of preferential flow, potentially mimicking a hypervascular lesion (Fig. 9). Over time, thrombosis of distal portal veins

Fig. 8. 61-year-old man with rectal cancer. A CT 3D volumetry reformat image shows single left hepatic lobe metastasis in yellow with an estimated volume of 158 mL. B CT 3D volumetry reformat image for planned resection of this metastasis highlights the left hepatic lobe in purple and the right lobe in green, with estimated volumes of 425 and 868 mL, respectively. The total liver volume is 1293 mL and the calculated FLR would be 76% of the total liver volume [868/(1293 - 158)], rendering left hepatectomy a safe treatment option.

Repeat imaging, typically with CT given that the susceptibility artifact from coils hampers image quality on MRI, is performed within 4 weeks after portal vein embolization to assess the degree of hepatic hypertrophy [45]. A 30–50% increase in the FLR occurs at 6 weeks in most series, although this is variable and patient dependent. While hypertrophy of the FLR may continue for up to 6 months, a plateau is often reached around 6–8 weeks and this serves as the typical timeframe of planned resection [17]. Awareness of the altered hemodynamics that occur after embolization is necessary to

Fig. 9. 46-year-old man with right hepatic colon metastases requiring lobectomy. Predicted FLR was 18% based on preoperative CT volumetry and right portal vein embolization was pursued instead. Axial CT images obtained 4 weeks later, one more cranial (A) and the other more caudal (B), reveal increased enhancement in the right lobe during the arterial phase due to altered hemodynamics after embolization. Metastatic lesion present within segment 8 (arrow) and coils are present within the right lobe (arrowhead), resulting in beamhardening artifact. FLR increased to 27% which was sufficient to permit right lobectomy.

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

in the treated segments may become apparent, a finding which indicates successful coil placement [44]. The final consideration on preoperative imaging is the identification of any extrahepatic metastatic disease. This does not represent an absolute contraindication to hepatic resection but this disease would need to be amenable to complete extirpation as well [16]. Assessment of regional lymph nodes is of paramount importance because in the setting of perihepatic lymph node involvement, the majority of studies reveal poorer survival as compared to patients without lymph node disease [3]. Lymph node metastases located in the porta hepatis near the hilum should not be considered an absolute contraindication to resection. However, if lymph nodes about the celiac trunk are involved, there is no survival benefit and surgery is typically avoided [46, 47]. The most widely accepted method of discriminating between normal and pathologic lymph nodes is based on short-axis size [48]. Normal porta hepatis and retrocrural lymph nodes should not exceed 6 mm, whereas celiac axis and retroperitoneal lymph nodes greater than 10 mm are considered abnormal [49]. Irregular margins and central necrosis are additional morphologic features suggestive of metastatic nodes. Lymph nodes which are metabolically active on FDG PET/CT are highly suspicious, presuming that the primary colorectal tumor was FDG avid.

Imaging findings following neoadjuvant treatment Unfortunately, only 15–25% of patients with metastatic colorectal liver metastases are deemed candidates for resection at the time of initial presentation. With the advent of newer and more efficacious chemotherapeutic agents, such as oxaliplatin and irinotecan, an additional 12% -30% of patients will be rendered eligible for surgery following neoadjuvant treatment due to a decrease in the bulk of metastatic lesions [4, 14, 50]. Irinotecan and oxaliplatin both can induce hepatic toxicity. The former is associated with chemotherapy-associated steatohepatitis (CASH), while the latter has been shown to cause vascular changes such as sinusoidal dilatation with erythrocytic congestion, perisinusoidal fibrosis, and fibrotic veno-occlusive disease. CT or MRI may demonstrate these changes as a nodular hepatic contour, patchy enhancement of the hepatic parenchyma due to venoocclusive disease, capsular retraction secondary to fibrosis, or fatty infiltration. Reticular hypointensity may be encountered on MR with gadoxetic acid hepatobiliary phase images if sinusoidal obstruction is present as a result of surrounding fibrosis. Peliosis hepatis has been associated with certain drugs to include oxaliplatin and it is important to not mistake this benign lesion as a new metastasis on post-chemotherapy imaging. In patients with a normal appearance of the liver parenchyma who

then present with evidence of liver damage on imaging over a short time interval, the injury is highly likely to be secondary to CASH. This highlights the importance of obtaining pretreatment imaging studies. Drug-induced hepatic injury is important to recognize as it may increase the risk of perioperative morbidity and a larger FLR will be required for resection if steatohepatitis is present [12]. Chemotherapy-related changes are reversible although they can persist for up to 4 months [6]. MRI is the preferred imaging modality for preoperative detection of colorectal hepatic metastases in patients treated with neoadjuvant chemotherapy, with a pooled sensitivity of 85.7% [51]. Specifically, MRI with gadoxetic acid has been shown to be superior to CT in the setting of steatosis, most notably for metastases less 1 cm in size [52]. CT is the best alternate to MRI, and FDG PET/CT is the worst performer of the three modalities. Despite being highly effective in chemonaive patients, FDG PET/CT is presumed to have decreased sensitivity and specificity after neoadjuvant treatment due to reduced size of the metastatic lesions which are below the detection threshold, decreased metabolic activity of the tumor cells, or therapy-induced necrosis. For CT, lower attenuation and less prominent contrast enhancement of the liver due to steatosis induced by the chemotherapy results in decreased liver to lesion contrast [51]. One caveat to be mindful of is the identification of new hypoattenuating or hyperattenuating lesions on post-treatment CT which may represent lesions which were imperceptible on pretreatment CT due to isoattenuation with the normal hepatic parenchyma. This phenomenon may mimic disease progression in spite of chemotherapy but is actually a result of alteration in the attenuation of the liver due to steatosis [53]. Radiologic evaluation is the current cornerstone for assessing disease response and changes in tumor size using Response Evaluation Criteria in Solid Tumors (RECIST) are sufficient following neoadjuvant treatment with conventional chemotherapy. RECIST utilizes unidirectional measurements and a partial response is defined as a 30% decrease in size [54]. However, for those patients who have received antiangiogenic agents, such as bevacizumab, a monoclonal antibody against vascular endothelial growth factor resulting in a cytostatic effect, the measured size differences over time are inadequate [55]. Morphologic criteria correlate better with pathologic response as those lesions which have optimally responded will become homogeneously low in attenuation and sharply marginated without enhancement, the socalled ‘‘pseudocyst’’ appearance (Fig. 10) [32]. While these morphologic findings have been validated for CT, the same has not yet been done for MRI [16]. Recently, measurements utilizing volumetric methods, such as slice-based segmentation or threshold-based segmentation, have gained favor over the traditional RECIST technique of calculating lesion volume based on its lar-

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

neoadjuvant therapy is warranted here as its optimal management remains unclear. ‘‘Disappearing’’ metastases, occurring in 5–48% of patients, describe those lesions which undergo a complete response and are no longer visualized on repeat imaging after treatment [16, 46]. In up to 80% of cases of complete response, persistent microscopic pathologic disease has been encountered following surgery [9]. This serves as the rationale for attempted resection of all initially detected sites of hepatic metastatic disease in spite of the challenge the surgeon may face without the ability to identify the precise anatomic location of ‘‘disappearing’’ metastases via either palpation or imaging [12]. Radiologists interpreting imaging following neoadjuvant chemotherapy should be cognizant that, in general, all of the original sites of hepatic metastatic disease noted on pretherapy imaging need to be resected or ablated, not just the residual disease identified after therapy. This, therefore, represents a departure from the typical method of radiologic reporting, in which nonexistent findings on the current study would not receive significant mention (Fig. 11).

Alternative therapies to surgery

Fig. 10. 46-year-old man with colon cancer treated with neoadjuvant chemotherapy (fluorouracil, leucovorin, oxaliplatin, and bevacizumab) for unresectable hepatic metastases. Axial CT images before (A) and after (B) therapy demonstrate a lesion in segment 8 (arrows) which has had a partial response to treatment by RECIST as there has been a greater than 30% decrease in lesion diameter. The lesion also has ‘‘pseudocyst’’ appearance as it is sharply marginated and nonenhancing on B.

gest axial diameter. This is a result of higher intraobserver reproducibility and increased accuracy of the volumetric measurements compared to RECIST [56]. Deliberation on a specific entity, which is commonly referred to as ‘‘disappearing’’ metastases, following

Liver-directed locoregional therapies, to include RFA, hepatic arterial infusional chemotherapy, transarterial chemoembolization (TACE), radioembolization with the beta-emitter yttrium-90 (Y-90), and cryotherapy, are playing an increasing role in the management of colorectal hepatic metastases. While previously reserved for the palliative setting, these therapies may now be used to convert patients with initially unresectable disease into surgical candidates or for definitive treatment in patients who have either unresectable disease, would have an insufficient FLR, or who are poor surgical candidates. Analogous to the rationale that surgical resection can be curative for isolated hepatic metastases, liver-directed therapies should be pursued as alternative to hepatectomy in order to obtain locoregional control and prevent further dissemination of disease [5]. The exact role of these treatments, however, remains unclear and still needs to be elucidated. RFA has risen to the forefront amongst available local ablative therapies due to advances in technology which maximize effectiveness, while minimizing associated morbidity. However, it has yet to be proven equivalent to surgical resection in randomized control trials since studies comparing the two treatments for resectable disease have yet to be performed [57]. Patients who undergo RFA alone for metastatic lesions immediately adjacent to a hepatic vein have higher incidences of intrahepatic recurrence due to the well-documented ‘‘heat-sink’’ effect. Caution is required for lesions which are centered in the hilar region as RFA may result in

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Fig. 11. 53-year-old man with hepatic metastases from rectal cancer. A Axial CT image for staging shows a lesion in segment 6 (arrow). Several other hepatic masses consistent with metastases were also present (not shown). B Axial CT image obtained 3 months later after neoadjuvant chemotherapy shows complete response of the segment 6 lesion (arrow), consistent with ‘‘disappearing’’ metastasis. However, this lesion needs to be specifically documented on the post-therapy CT since this site of disease will still require resection or ablation. Note the new nodular contour of the liver (arrowhead) representing chemotherapy-induced steatohepatitis.

biliary stricture [9]. Assessment of the effectiveness of RFA with CT or MRI is the current standard. If a 5-mm margin of normal appearing hepatic parenchyma is present within the ablation zone during the portal venous phase then the ablation is considered successful. Evaluation of the margin is critical as it has been recognized as an independent risk factor for tumor progression [58]. Transarterial therapies allow for higher response rates on average as compared with systemic therapies for the treatment of colorectal hepatic metastases. This is presumably because a larger proportion of the treatment dose will be concentrated to the tumor cells instead of the normal hepatocytes given that metastases are preferentially fed by the arterial system. However, few studies exist investigating TACE as a treatment for colorectal hepatic metastases because it is best suited for hypervascular liver lesions, such as hepatocellular carcinoma, whereas metastatic deposits usually have less arterialization [59]. The efficacy of TACE is through a combination of ischemia of the tumor cells due to arterial occlusion and antineoplastic effect from the chemotherapy [60]. Radioembolization delivers therapeutic radiation in a targeted fashion to metastatic lesions via the injection of microspheres with bound Y-90 which are released into the arterial supply of the liver [59]. Relative to TACE, there exists more evidence within the literature to support the use of Y-90 radioembolization to treat patients with unresectable colorectal hepatic metastases. Selection criteria for Y-90 include unresectable metastatic disease isolated to the liver, bilirubin less than 2 mg/dL, and the absence of uncorrectable coagulopathy [61]. Evaluating the response of metastases following radioembolization is a specific role best suited for FDG PET/CT. Two months after treatment, good correlation has been shown between the response on FDG PET/CT and the relative uptake of macroaggregated albumin on the pretreatment simulation exam. FDG PET/CT is superior to both CT and MRI for this indication as the latter modalities cannot readily differentiate between post-treatment necrosis and fibrosis vs. disease progression [5]. Recognition of tumor-supplying collateral vessels parasitized from extrahepatic arteries on CT or MRI is essential prior to any transarterial treatment. Commonly parasitized vessels include the right inferior phrenic and right T8–T11 intercostal arteries. If present, these vessels will need to be embolized before TACE or Y-90 radioembolization can be performed. This step is undertaken not to induce tumor ischemia but instead to re-establish the normal intrahepatic arterial flow so that the treatment will reach all of the desired target lesions [61].

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

Conclusion What qualifies as resectable disease for patients with colorectal hepatic metastases has evolved over the past few decades. Hepatic metastases are currently considered resectable as long as sufficient liver volume will remain to include at least two contiguous liver segments with adequate inflow, outflow, and biliary drainage. Cross-sectional imaging is vital in order to determine technical resectability and locate extrahepatic disease. MRI, specifically with gadoxetic acid, is the most sensitive modality for the detection of colorectal hepatic metastases, particularly those less than 1 cm in size as well as for patients with steatosis or other hepatotoxicity-related changes secondary to chemotherapy. Conflict of interest All of the authors have no competing interests and have nothing to disclose.

References 1. Siegel R, Naishadham D, Jemal A (2014) Cancer statistics, 2014. CA Cancer J Clin 64:9–29 2. Haddad AJ, Bani Hani M, Pawlik TM, Cunningham SC (2011) Colorectal liver metastases. Int J Surg Oncol 2011:285840 3. Chun YS, Vauthey JN (2007) Extending the frontiers of resectability in advanced colorectal cancer. Eur J Surg Oncol 33(Suppl 2):S52–S58 4. Pawlik TM, Schulick RD, Choti MA (2008) Expanding criteria for resectability of colorectal liver metastases. Oncologist 13:51–64 5. Seront E, Van den Eynde M (2012) Liver-directed therapies: does it make sense in the current therapeutic strategy for patients with confined liver colorectal metastases? Clin Colorectal Cancer 11:177–184 6. Zdenkowski N, Chen S, Van der Westhuizen A, Ackland S (2012) Curative strategies for liver metastases from colorectal cancer: a review. Oncologist 17:201–211 7. Guye ML, Schoellhammer HF, Chiu LW, et al. (2013) Designing liver resections and pushing the envelope with resections for hepatic colorectal metastases. Indian J Surg Oncol 4:349–355 8. Kopetz S, Chang GJ, Overman MJ, et al. (2009) Improved survival in metastatic colorectal cancer is associated with adoption of hepatic resection and improved chemotherapy. J Clin Oncol 27:3677–3683 9. Sharma S, Camci C, Jabbour N (2008) Management of hepatic metastasis from colorectal cancer: an update. J Hepatobiliary Pancreat Surg 15:570–580 10. Ekberg H, Tranberg KG, Andersson R, et al. (1986) Determinants of survival in liver resection for colorectal secondaries. Br J Surg 73:727–731 11. Meriggi F, Bertocchi P, Zaniboni A (2013) Management of potentially resectable colorectal cancer liver metastases. World J Gastrointest Surg 5:138–145 12. Lupinacci R, Penna C, Nordlinger B (2007) Hepatectomy for resectable colorectal cancer metastases—indicators of prognosis, definition of resectability, techniques, and outcomes. Surg Oncol Clin N Am 16:493–506 13. Primrose JN (2010) Surgery for colorectal liver metastases. Br J Cancer 102:1313–1318 14. Grundmann RT (2011) Current state of surgical treatment of liver metastases from colorectal cancer. World J Gastrointest Surg 3:183–196 15. Weg N, Scheer MR, Gabor MP (1998) Liver lesions: improved detection with dual-detector-array CT and routine 2.5 mm thin collimation. Radiology 209:417–426 16. Adams RB, Aloia TA, Loyer E, et al. (2013) Selection for hepatic resection of colorectal liver metastases: expert consensus statement. HPB 15:91–103 17. Frankel TL, D’Angelica MI (2014) Hepatic resection for colorectal metastases. J Surg Oncol 109:2–7

18. Wiering B, Krabbe PF, Jager GJ, Oyen WJ, Ruers TJ (2005) The impact of fluorine-18-deoxyglucose-positron emission tomography in the management of colorectal liver metastases. Cancer 104: 2658–2670 19. Zhuang H, Sinha P, Pourdehnad M, et al. (2000) The role of positron emission tomography with fluorine-18-deoxyglucose in identifying colorectal cancer metastases to liver. Nucl Med Commun 21:793–798 20. Moulton CA, Gu CS, Law CH, et al. (2014) Effect of PET before liver resection on surgical management for colorectal adenocarcinoma metastases: a randomized clinical trial. JAMA 311:1863– 1869 21. Niekel MC, Bipat S, Stoker J (2010) Diagnostic imaging of colorectal liver metastases with CT, MR imaging, FDG PET, and/or FDG PET/CT: a meta-analysis of prospective studies including patients who have not previously undergone treatment. Radiology 257:674–684 22. Mayo SC, Heckman JE, Shore AD, et al. (2011) Shifting trends in liver-directed management of patients with colorectal liver metastasis: a population-based analysis. Surgery 150:204–216 23. Scharitzer M, Ba-Ssalamah A, Ringl H, et al. (2013) Preoperative evaluation of colorectal liver metastases: comparison between gadoxetic acid-enhanced 3.0-T MRI and contrast-enhanced MDCT with histopathological correlation. Eur Radiol 23:2187–2196 24. Han K, Park SH, Kim KW, et al. (2015) Use of liver magnetic resonance imaging after standard staging abdominopelvic computed tomography to evaluate newly diagnosed colorectal cancer patients. Ann Surg 261:480–486 25. Patel S, Cheek S, Osman H, Jeyarajah DH (2014) MRI with gadoxetate disodium for colorectal liver metastasis: is it the new ‘‘imaging modality of choice’’? J Gastrointest Surg 18:2130–2135 26. Zech CJ, Korpraphong P, Huppertz A, et al. (2014) Randomized multicentre trial of gadoxetic acid-enhanced MRI vs. conventional MRI or CT in the staging of colorectal cancer liver metastases. Br J Surg 101:613–621 27. Kim HJ, Lee SS, Byun JH, et al. (2015) Incremental value of liver MR imaging in patients with potentially curable colorectal hepatic metastasis detected at CT: a prospective comparison of diffusionweighted imaging, gadoxetic acid-enhanced MR imaging, and a combination of both MR techniques. Radiology 274:712–722 28. Cho JY, Lee YJ, Han HS, et al. (2015) Role of gadoxetic acidenhanced magnetic resonance imaging in the preoperative evaluation of small hepatic lesions in patients with colorectal cancer. World J Surg 39:1161–1166 29. Donati OF, Fischer MA, Chuck N, et al. (2013) Accuracy and confidence of Gd-EOB-DTPA enhanced MRI and diffusionweighted imaging alone and in combination for the diagnosis of liver metastases. Eur J Radiol 82:822–828 30. Lafaro KJ, Roumanis P, Demirjian A, Lall C, Imagawa DK (2013) Gd-EOB-DTPA-enhanced MRI for detection of liver metastases from colorectal cancer: a surgeon’s perspective!. Int J Hepatol 2013:572307 31. Kanematsu M, Kondo H, Goshima S, et al. (2006) Imaging liver metastases: review and update. Eur J Radiol 58:217–228 32. Boonsirikamchai P, Asran MA, Maru DM, et al. (2011) CT findings of response and recurrence, independent of change in tumor size, in colorectal liver metastasis treated with bevacizumab. AJR Am J Roentgenol 197:W1060–W1066 33. Ward J, Robinson PJ, Guthrie JA, et al. (2005) Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced MR imaging. Radiology 237: 170–180 34. Sica GT, Ji H, Ros PR (2000) CT and MR imaging of hepatic metastases. AJR Am J Roentgenol 174:691–698 35. Jones EC, Chezmar JL, Nelson RC, Bernardino ME (1992) The frequency and significance of small (<15 mm) hepatic lesions detected by CT. AJR Am J Roentgenol 158:535–539 36. Khalil HI, Patterson SA, Panicek DM (2005) Hepatic lesions deemed too small to characterize at CT: prevalence and importance in women with breast cancer. Radiology 235:872–878 37. Schwartz LH, Gandras EJ, Colangelo SM, Ercolani MC, Panicek DM (1999) Prevalence and importance of small hepatic lesions found at CT in patients with cancer. Radiology 210:71–74

M. J. Reiter et al.: Role of imaging for patients with colorectal hepatic metastases

38. Weiss MJ, D’Angelica MI (2012) Patient selection for hepatic resection for metastatic colorectal cancer. J Gastrointest Oncol 3:3–10 39. Peungjesada S, Aloia TA, Kaur H, et al. (2013) Intrabiliary growth of colorectal liver metastasis: spectrum of imaging findings and implications for surgical management. AJR Am J Roentgenol 201:W582–W589 40. Shin DS, Ingraham CR, Dighe MK, et al. (2014) Surgical resection of a malignant liver lesion: what the surgeon wants the radiologist to know. AJR Am J Roentgenol 203:W21–W33 41. May BJ, Madoff DC (2012) Portal vein embolization: rationale, technique, and current application. Semin Intervent Radiol 29: 81–89 42. Wibmer A, Prusa AM, Nolz R, et al. (2013) Liver failure after major liver resection: risk assessment by using preoperative gadoxetic acid-enhanced 3-T MR imaging. Radiology 269:777–786 43. Mahnken AH, Pereira PL, de Baere T (2013) Interventional oncologic approaches to liver metastases. Radiology 266:407–430 44. Ganeshan DM, Szklaruk J (2012) Portal vein embolization: crosssectional imaging of normal features and complications. AJR Am J Roentgenol 199:1275–1282 45. Geisel D, Gebauer B, Malinowski M, Stockmann M, Denecke T (2014) Comparison of CT and MRI artefacts from coils and vascular plugs used for portal vein embolization. Eur J Radiol 83: 692–695 46. Jaeck D (2003) Significance of hepatic pedicle lymph nodes metastases in surgical management of colorectal liver metastases and of other liver malignancies. Ann Surg Oncol 10:1007–1011 47. Gallinger S, Biagi JJ, Fletcher GG, et al. (2013) Liver resection for colorectal cancer metastases. Curr Oncol 20:255–265 48. Koh DM, Hughes M, Husband JE (2006) Cross-sectional imaging of nodal metastases in the abdomen and pelvis. Abdom Imaging 31:632–643 49. Einstein DM, Singer AA, Chilcote WA, Desai RK (1991) Abdominal lymphadenopathy: spectrum of CT findings. Radiographics 11:457–472

50. Adam R, Wicherts DA, de Haas RJ, et al. (2009) Patients with initially unresectable colorectal liver metastases: is there a possibility of cure? J Clin Oncol 27:1829–1835 51. van Kessel CS, Buckens CF, van den Bosch MA, et al. (2012) Preoperative imaging of colorectal liver metastases after neoadjuvant chemotherapy: a meta-analysis. Ann Surg Oncol 19:2805–2813 52. Kulemann V, Schima W, Tamandl D, et al. (2011) Preoperative detection of colorectal liver metastases in fatty liver: MDCT or MRI? Eur J Radiol 79:e1–e6 53. Tirumani SH, Kim W, Nishino M, et al. (2014) Update on role of imaging in management of metastatic colorectal cancer. Radiographics 34:1908–1928 54. Eisenhauer EA, Therasse P, Bogaerts J (2009) New response evaluation criteria in solid tumors: revised RECIST guideline (version 1.1). Eur J Cancer 45:228–247 55. Adam R, De Gramont A, Figueras J, et al. (2012) The oncosurgery approach to managing liver metastases from colorectal cancer: a multidisciplinary international consensus. Oncologist 17:1225–1239 56. Rothe JH, Grieser C, Lehmkuhl L, et al. (2013) Size determination and response assessment of liver metastases with computed tomography—comparison of RECIST and volumetric algorithms. Eur J Radiol 82:1831–1839 57. Shah DR, Green S, Elliot A, McGahan JP, Khatri VP (2013) Current oncologic applications of radiofrequency ablation therapies. World J Gastrointest Oncol 5:71–80 58. Minami Y, Kudo M (2013) Radiofrequency ablation of liver metastases from colorectal cancer: a literature review. Gut Liver 7:1–6 59. de Baere T, Deschamps F (2011) Arterial therapies of colorectal cancer metastases to the liver. Abdom Imaging 36:661–670 60. Johnston FM, Mavros MN, Herman JM, Pawlik TM (2013) Local therapies for hepatic metastases. J Natl Compr Cancer Netw 11:153–160 61. Wang DS, Louie JD, Sze DY (2013) Intra-arterial therapies for metastatic colorectal cancer. Semin Intervent Radiol 30:12–20