Abstract
Background: Detection of asymptomatic focal liver lesions (FLL) is increasing because of a widespread use of modern radiologic imaging. Most of these lesions are benign, though malignancy often has to be ruled out, which is posing a diagnostic challenge. Aim: To critically evaluate our treatment strategy in the context of recently published American College of Gastroenterology (ACG) guidelines. Patients and Methods: The medical records of patients who underwent surgery for asymptomatic, incidentally detected FLL from 2005-2012 were reviewed. Primary end-points were the congruence of suspected diagnosis and final pathology, as well as the identification of predictors of malignancy. A systematic review was undertaken to help define a standardized management. Results: Eighty patients, 37 male and 43 female with a mean age of 57 years (range=16-83) were included, harboring 39 (49%) malignant and 41 (51%) benign, asymptomatic hepatic lesions. Hepatocellular carcinoma (HCC) (n=24) represented most of malignant FLL, followed by cholangiocellular carcinoma (CCC) (n=10), whereas focal nodular hyperplasia (FNH) (n=19) and liver hemangioma (n=7) were the predominant benign entities. Fifty-one patients (64%) had a correct preoperative diagnosis, while patients with FNH were most commonly misdiagnosed (53%). We identified age (p<0.001) and male sex (p=0.013) as risk factors for malignancy in an asymptomatic FLL. Conclusion: Despite recent technical advances of the modern radiology setting a correct preoperative diagnosis in an asymptomatic FLL remains challenging. Male gender and old age seem to correlate with malignancy. In the absence of biomarkers and evidence-based guidelines, a multidisciplinary approach in an experienced tertiary referral center is recommended for an optimized individual management.
Increasing numbers of incidentally detected focal liver lesions (FLL or incidentaloma) are based on widespread use and a cheaper access to high resolution radiologic imaging (1-3). In order to allow an individual and cost-effective treatment of every FLL detected, it is inevitable to know the biologic behavior of possible tumor entities. Furthermore, knowledge of the characteristic radiological appearance of respective hepatic lesions may contribute to an adequate management (Figure 1). Nevertheless, clinicians are oftentimes confronted with the dilemma of choosing adequate diagnostic means in the clinical workup. Even after comprehensive efforts for diagnosis in specialized centers, there is a remaining risk for incorrect preoperative diagnoses and associated over- or undertreatment with possible severe clinical consequences for the patient. This is mainly owed to the absence of definitive diagnostic criteria, reliable biomarkers and evidence-based guidelines that may sufficiently guide treatment decisions (4). Thus, whenever diagnosis remains unclear, an indication for surgery has, thus far, been based on suspected malignancy, growth in size and symptoms (5). In 2014, the American College of Gastroenterology (ACG) published a guideline to allow a more standardized diagnostic approach and adequate treatment for any incidental liver lesion (3). In the absence of information from reasonably powered prospective trials, this guideline merely reaches a low evidence level leaving a relevant potential to further improve this clinical challenge.
The aim of the present study was to retrospectively evaluate the treatment and outcome of patients with FLL at our surgical Department in the context of recently published ACG guidelines (3). We additionally performed a systematic review of relevant recent literature on the management of FLL. Our long-term goal is to help resolve the diagnostic dilemma and significant economic burden associated with incidental FLL.
Demographic data.
Patients and Methods
Patients. A retrospective analysis of all patients with resected, incidentally detected FLL treated at the Department of General, Visceral, Thoracic and Vascular Surgery, University Hospital of Bonn, from January 2005 till December 2012, was conducted. Ethical standards of the University of Bonn were fully acknowledged. Exclusion criteria were symptoms at the time of presentation, known hepatic cirrhosis, hepatitis B/C infection, known primary malignancy and a follow-up of less than 12 months. Demographic, clinical, pathological and operative data were collected and analyzed.
FLL was defined according to the ACG guideline as a “solid or cystic mass or area of tissue that is identified as an abnormal part of the liver” (3). Morbidity was defined as a clinical adverse event within 30 days postoperatively and surgical mortality was determined as death within the same period of time. Complications were categorized by the Clavien-Dindo classification (6) and further subdivided into major and minor complications (major ≥grade III).
Statistics. Continuous variables were expressed as median (range) and compared using the Mann-Whitney U-test. Categorical variables were compared using the χ2 test or Fisher exact test, where appropriate. Values are expressed as mean or median and range unless otherwise stated. Discriminant analysis was performed on continuous variables to identify the best numeric values to distinguish patients with malignant diseases from those with benign pathologies. Binary logistic Cox regression model was used to identify prognostic factors for malignant disease. Differences of p<0.05 were considered statistically significant. Analyses were performed using SPSS computer software (SPSS version 20; IBM Corp., New York, NY, USA).
Systematic literature review. A systematic MEDLINE query was performed using the words “focal liver lesion” OR “liver incidentaloma”. The search was limited by the following criteria: “English” (language category), “humans” (participants' category), “published in the last 10 years” (dates' category). Eligible for our study were analysis on handling, diagnostic approach and diagnostic criteria on patients with FLL. The reference part of each article was scrutinized for further matching publications. We excluded studies focusing only on imaging characteristics or single case reports. The search for studies was performed independently by several authors (AS, PL and HM) and revealed 35 relevant publications.
Results
Over a period of seven years (2005-2012), liver resection for asymptomatic FLL was performed on 80 matching patients at our Department. We included 43 women (54%) and 37 men (46%) with a mean age of 57 years (range=17-83) (Table I).
Thirty-nine FLL (49%) proved to be malignant and 41 tumors (51%) were benign. Hepatocellular carcinoma (HCC) was the most prevalent diagnosis in malignant lesions (n=24, 30%), followed by cholangiocellular carcinoma (CCC) (n=10, 13%). The most prevalent benign lesions were focal nodular hyperplasia (FNH) (n=19, 24%) and hemangiomas (n=7, 8%) (Table II). A majority of 88% of the tumors (n=70) appeared as a solitary hepatic mass with a mean maximum diameter of 6.3 cm.
Tumor entities of FLL in descending frequency.
Male gender (67% vs. 33%, p=0.013, odds ratio (OR)=3.543) and age at presentation (68 years vs. 47 years, p<0.001, OR=35.7) was associated with a higher probability of malignancy (Table III). Additionally, γ-glutamyltransferase levels (γ-GT) differed between the groups (61.4 vs. 108.6 U/l, p=0.02). All other parameters including tumor markers and liver enzymes showed no effect on the probability of malignancy (Table I).
In 51 cases (64%), the preoperative suspected diagnosis was matching the final histology, whereas, in 17 patients (21%), they did not match and in 12 cases (15%) there was no documented unequivocal preoperative diagnosis. Preoperatively, malignant lesions were more frequently diagnosed correctly compared to benign tumors (88% vs. 62%, p=0.012). The most commonly misdiagnosed entity was FNH (8 of 15 cases, 53%). In comparison, HCC was diagnosed correctly in 86% (19 of 22 patients) and CCC in 88% (7 of 8 patients).
The main indication for surgical resection was a suspected or proven malignancy of the tumor (n=60, 75%), a significant growth in size (n=13, 16%) and a strong intention by the patient to rule out malignancy by tumor resection (n=7, 9%). Mostly, an atypical resection was performed (n=32; 40 %), followed by segmental resection (n=24, 30%), hemihepatectomy (n=19, 24%) and enucleation (n=5, 6%). In total, 29 patients suffered from complications (36%, 15 minor and 14 major). Patients with malignant tumors of the liver had a longer operation time (256 min vs. 199 min, p<0.01), an extended hospital stay (14.8 vs. 8.9 days, p<0.001) and a higher rate of postoperative complications (49 vs. 24%, p=0.036).
Review of literature. Generally, FLL may originate from hepatocytes, biliary epithelium, mesenchymal tissue and connective tissue or represent metastasis from distant sites (3, 7).
Binary logistic regression model of factors predicting malignancy.
Malignant FLLs
Hepatocellular carcinoma (HCC). HCC is the second leading cause of cancer-related death and the fifth most common tumor worldwide (8). Despite all therapeutic efforts these tumors still carry a poor prognosis with a 5-year survival rate of ~16%(8). The most important risk factor for HCC is cirrhosis, which is more prevalent in hepatitis B and C patients (9). This implicates that a FLL, in an elderly patient with known cirrhosis, is particularly suspicious for HCC (10). In 85% cross-sectional imaging studies, mainly represented by computed tomography (CT) and magnetic resonance imaging (MRI), show a characteristic, strong contrast enhancement in the arterial phase and a wash out in the portal venous or the delayed phase (11). HCC-typical contrast pattern in one contrast-enhanced imaging study may, nowadays, be regarded as “radiologic diagnosis” of HCC in case of liver cirrhosis. A second, contrast-enhanced cross-sectional study is recommended if the lesion is under 1 cm or the quality of imaging is low (7). In case of diagnostic uncertainty or absence of cirrhosis, a biopsy is necessary to confirm HCC preoperatively (12). Elevated α-fetoprotein (AFP) levels may additionally be suggestive for HCC but cannot be used as single diagnostic tool (3, 7). New biomarkers, e.g. microRNA panels or exosome-derived proteins, show promising results but, thus far, no single parameter reaches sufficient sensitivity and specificity (13, 14). It is important to emphasize that an early diagnosis and resection of HCC significantly improves postoperative outcome (45.3 vs. 15 months of median disease-free survival) (15). In HCC limited to the liver, a curative resection or transplantation may be possible (16) (For more information see (7)).
Cholangiocellular Carcinoma (CCC). CCC are highly aggressive tumors with a devastating prognosis and an increasing incidence (17). Currently, these tumors represent ~20% of primary hepatic malignancies and may arise intrahepatically but may also grow in a perihilar and extrahepatic locations with sometimes secondary hepatic infiltration (18). CCC lack specific early symptoms or laboratory changes and are, therefore, often diagnosed in an advanced stage (19). Besides primary sclerosing cholangitis, several other risk factors have been identified for cholangiocarcinogenesis, including cirrhosis, liver fluke, choledochal cysts and cholelithiasis (20). Presentation of CCC in modern imaging is often inconclusive (3), although dilatation of the bile duct system or an abrupt change in bile duct diameter, together with a progressive enhancement in the arterial and venous phase of contrast studies, may be suggestive of CCC. Despite the conventional biomarkers, carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (Ca 19-9), that show a moderate sensitivity and specificity, new serum-based biomarkers have extensively been examined. Of these, mucines, epidermal growth factor receptor (EGFR) and several microRNAs currently show the highest potential in diagnosis and prognosis prediction (21, 22). Radical surgery remains the only curative option for CCC because only few patients meet the strict criteria justifying transplantation (18). Even subsequent to successful surgery with free margins, recurrence rates are high due to the aggressive biology with early metastatic spread (23) (For more information see (17, 24)).
Metastasis. The liver is the most common site for metastatic spread (25). In fact, liver metastases are diagnosed 20-fold more often than primary malignant hepatic tumors. Metastasis of colorectal cancer and neuroendocrine tumors are the most frequent ones (26). In patients with known malignancy, small incidentalomas are particularly suspicious for metastasis, although the diverging capabilities for hepatic metastasis formation of different cancer entities have to be kept in mind (27). Since their appearance is strongly related to the associated primary tumor, these lesions appear heterogeneous in cross-sectional imaging. In general, metastases show irregular margins, necrotic areas and – depending on the entity – an arterial contrast enhancement (28). Although contrast enhancement pattern differs, attention has to be paid as highly vascularized metastasis may mimic hepatic hemangiomas in native CT studies (29). Treatment varies depending on the stage and entity of the patients' malignancy and comprises of palliative to aggressive treatment. Especially in colorectal cancer, outcome is improved by focal liver resection (30).
Benign FLLs
Hepatic hemangioma (HH). HH are the most frequently encountered benign liver lesions (31). Their prevalence is reported to be as high as 20 % in specific populations (32) with a strong female predominance (5 women: 1 man) (33). Despite their frequency, the underlying pathogenesis is still not fully elucidated (31). HH typically arises in middle-aged women of 30-50 years who normally show no or few symptoms (31). Imaging techniques have a high specificity and sensitivity for diagnosis (31). CT and MRI studies reveal in ~70% a well-demarked lesion exhibiting peripheral nodular enhancement with a centripetal progression and lack of a wash out in later contrast phases (3, 32). Combined with sonographic correlation, this typical radiological appearance serves as an important criterion for the distinction to HCC (34). Based on their minor potential for complications and predominantly benign and subclinical course, no treatment for HH is indicated in asymptomatic patients (3). With respect to small HH suspicious lesions under 3 cm in size with inconclusive imaging, follow-up studies should be performed after six months (3, 31). Only in case of symptoms, such as a growth tendency over 3 cm per year or a diameter over 10 cm, an intervention should be considered. Well-established treatment methods are parenchyma-saving surgical enucleation, intra-arterial embolization or irradiation (1, 2, 31, 35).
Simple hepatic cyst (SHC). SHC develop as congenital exclusions of hyperplastic bile duct remnants without connection to the biliary ductal system (36). Their prevalence is reported to be up to 18% (37). These common, round to oval tumors are predominantly found in women above the age of 40 years and show virtually no malignant transformation (3). Usually, SHC diameter does not exceed one centimeter (38). Ultrasound, as the leading diagnostic tool, reveals an anechoic, smoothly-walled and fluid-filled lesion that shows no contrast enhancement on MRI and CT (3, 29, 37). Because of their natural history, most patients with SHC do not need any treatment or follow-up (3). Indications for treatment are symptomatic cysts, proof of septations, calcification or suspicion of biliary cystadenoma and cystadenocarcinoma (39). Surgical intervention comprises of marsupialisation, fenestration, enucleation, aspiration and sclerotherapy (40, 41).
Focal nodular hyperplasia (FNH). FNH are common benign liver lesions with a prevalence of 0.3-3%, which are mainly found in middle-aged women (3, 42). In contrast to hepatocellular adenoma (HCA), up to 40% of patients develop symptoms (1); however, FNH do hardly ever show a malignant transformation (1, 26). These tumors rarely exceed 5 cm in maximum diameter and show no elevation of liver enzymes (29). A central scar with a spoke-wheel pattern is a characteristic imaging feature of FNH in contrast-enhanced cross-sectional imaging techniques (43). This phenomenon describes an early and homogeneous enhancement of the lesion, followed by a decrease in density or intensity down to the same level of the surrounding liver parenchyma. Furthermore, a late phase enhancement of a central scar (44) may be observed when extracellular contrast agents have been applied. Calcifications or necrosis are extremely rare (29). These radiologic characteristics allow an accurate diagnosis with a sensitivity of 70% and a specificity of 98% (3, 45, 46). Albeit many pathognomonic imaging features for each entity exist, differentiation between FNH and hepatocellular adenoma remains arduous in up to one third of patients. Treatment recommendations for asymptomatic FNH include conservative options because of their natural stability and low complication rate (47). Nevertheless, FNH should be followed up every 6-12 months for evaluation of growth (2, 48). In case of a strong growth tendency or relevant symptoms, atypical resection (29, 32) or embolization (49) has to be considered.
Hepatocellular adenoma (HCA). HCA are rare tumors that show a strong association with hormonal intake, e.g. oral contraceptives or androgen intake (31, 50). Additionally, HCAs are highly prevalent in patients with hereditary glycogen storage disorders, comprising a subgroup with unique clinical presentation and treatment (51). Despite their benign nature, HCA often cause symptoms and are associated with a relevant risk for complications. The most prevalent complication is spontaneous hemorrhage in 11-29% of patients (52). Nearly five percent of these lesions also show a malignant transformation into HCC, especially in elderly male patients and patients with large HCAs (53, 54). Radiologic diagnosis of HCAs may be extremely challenging because of necrotic and hemorrhagic areas within the tumor, mimicking features of well-differentiated HCC (26, 29, 55). In this context, MRI seems to be superior to CT, especially when using liver-specific contrast agents (56). Based on the probability of complications and the risk of a malignant transformation, HCA exceeding 5cm in diameter should be considered for surgical resection (3, 31, 48). Subsequently, a strict follow-up every 6-12 months is mandatory for the detection of recurrence (2). Asymptomatic HCA, smaller that 5 cm, may be followed closely for 2 years (3, 57). In women, discontinuance of oral contraceptives should be recommended to reduce tumor volume (3).
Cystic echinococcosis (CEC). CEC of the liver is caused by infection with the tapeworm Echinococcus granulosus or multilocularis, while humans act as an intermediate host. The cysts mainly arise in the liver and appear quite similar to simple cysts when small (3). In contrast, lesions over five centimeters develop a thickened, calcified wall with daughter and satellite cysts best seen with CT and MRI imaging (29, 58). The so called “water lily sign” is pathognomonic, resembling a central germinal membrane floating in the cysts' fluid (29). There is only a low evidence level for treatment modalities in CEC but, in general, treatment decisions should be made taking size, location and symptoms into consideration (3, 59). For small, asymptomatic and inactive cysts a watchful waiting seems appropriate. In contrast, larger cysts carry a risk for rupture and should, therefore, be evaluated for resection. It is important to stress that seeding of cyst content has to be prevented during operation to avoid an anaphylactic shock (60). Perioperative treatment with antihelmintic agents should be started 1 month prior to surgery and continued 6 months after surgery (61). An alternative, less aggressive treatment is the “PAIR technique” (puncture, aspiration, injection, re-aspiration) that has a comparable outcome to surgery and is recommended in multimorbid patients (62).
Discussion
Within the last decades, a major increase in incidentally detected FLL has been noted. These lesions are found in up to one third of individuals screened with cross-sectional imaging in the absence of any liver related diseases or symptoms (27, 63). This virtual epidemic of hepatic incidentalomas is mainly based on the constant improvement and widespread availability of radiologic technologies and their inflationary use in routine clinical work. This progress holds both opportunities and risks: On the one hand, it may result in the detection of a life-threatening lesion at an early, curable stage. On the other hand, FLL detection may result in a harmful over-treatment of patients in whom merely follow-up or no treatment at all is necessary (64). We critically analyzed management and outcome of asymptomatic FLL at our Department. In this study, surgically-resected tumors were included for comparability of preoperative suspected and postoperative proven diagnosis. Detailed descriptions of the radiological imaging features may be found in respective publications (4, 26, 32, 65) and are briefly summarized in Table IV.
According to our own experience and with respect to recent literature on this topic, a stepwise and standardized diagnostic approach to FLL at a multidisciplinary expert center seems inevitable to cope with the variety of underlying causes for FLL and to keep the rapidly rising costs in a justifiable range. Initially, a detailed history, including established risk factors for various liver tumors (e.g. hormonal intake, presence of infectious disease, presence of gastrointestinal cancer) is crucial. Routine blood work and the few available tumor markers (e.g. AFP, CEA, CA19-9) are additionally important standard tasks in all patients with unclear FLL. There has been a huge effort to develop other biomarkers for non-invasive disease detection, such as e.g. exosome-based assays or circulatory miRNAs, though none of these have been established in clinical routine thus far. In the near future, liquid biopsy techniques may hold a key for safe and precise FLL diagnosis (14, 66, 67).
Radiological characteristics of selected focal liver lesions.
To rule out malignancy, in most cases it remains crucial to perform a triple phase CT/MRI with an arterial, portal venous, as well as a delayed venous phase or contrast-enhanced ultrasound, in order to most accurately display the tumor's contrast pattern. An exception constitutes asymptomatic cystic liver lesions, where a contrast-enhanced ultrasound reveals almost every relevant information. Thus far, contrast-enhanced imaging is very helpful and provides key information in the non-invasive diagnostic workup (68). In accordance to Marrero et al., we further suggest to sub-divide the incidentally detected FLL into solid, cystic and vascularized lesions, which should be further investigated as indicated in Figure 2.
It is important to underline that a vast majority of FLLs are benign (3). Our and other studies revealed a malignancy rate of 40% or more in FLL (15, 69). This high number is essentially owed to an inclusion bias measured at a tertiary high-volume center (2, 15, 48). Almost all patients in our study were referred to our hepatobiliary outpatient center with previous imaging studies, implying that referral was triggered by suspected or proven malignancy. This bias casts light on a common problem of patients with FLL. Prior to their referral, patients' diagnostic workup has often been initiated by general practitioners with the risk for inappropriate, delayed or simply unnecessary imaging, posing both a clinical and financial burden. Despite extensive complete diagnostics, the precise diagnosis of a FLL remains unclear in up to 50% (68, 70, 71) and we similarly report a rate of merely 64% correctly diagnosed lesions. In our collective, FNH showed the highest diagnostic error-proneness, while HCC and CCC presented the lowest error-proneness. This may be explained by the number of large FNH and their rapid increase in size before surgery in our collective. Four patients also presented with very rare hepatic lesions, including angiomyolipoma, mesothelioma and lymphangioma. Despite preoperative biopsies in three of these patients, all but one were not correctly diagnosed, highlighting the difficulties with rare entities.
The importance of a preoperative biopsy for FLL is still a matter of controversy. Some experts favor this approach because of an additional diagnostic tool to verify the diagnosis. Furthermore, histology of normal liver parenchyma may be helpful to estimate postoperative outcome especially if major hepatectomies are planned (72). Others criticize biopsies for their high rate of false negative results (73), their low consistency between biopsy's histology and final histology (74), as well as their potential risk for tumor cell seeding (64, 75). In general, the complication rate of a percutaneous biopsy seems low (76). If imaging studies fail to reveal a precise diagnosis, biopsy may be needed, especially in cirrhotic livers and multimorbid patients to confirm the diagnosis preoperatively and to sufficiently plan the operation (12). In healthy patients with suspected malignancy, we rather tend to perform an open or laparoscopic biopsy (1, 17). This approach has two major advantages: First, larger tumor samples can be obtained allowing a more precise pathology and, second, tumor resection may be performed within the same operation (29, 47, 70). In our opinion, a multidisciplinary approach, e.g. as an equivalent to acknowledged tumor boards, has to be set up for every patient presenting with a FLL.
Radiologic and macroscopic appearance of mucinous cystic neoplasms in an asymptomatic patient. A 40-year-old, asymptomatic and healthy woman with long-term contraceptive intake presented with a progressing FLL in the left liver lobe first diagnosed 6 months prior to admission. Extensive workup, including 3-phase-CT, MRI with magnetic resonance cholangiopancreatography (MRCP), colonoscopy and gastroscopy could not fully exclude malignancy. Cross-sectional imaging described several tumors, including a multi-chambered cystic mass with a nodular lesion at its ventral wall (segments 2, 3) and a maximum diameter of 10.5 cm suspicious for biliary cystadenoma, hamartoma or a cystic echinococcosis. Several other lesions were diagnosed as an uncomplicated hemangioma (segment 7), a FNH (segment 4a), a fatty lesion and two artery-venous malformations (segments 5 and 6). Routine blood work, tumor markers and serologic testing for cystic echinococcosis were normal. An extended bisegmentectomy (segments 2, 3 and atypic resection of segment 4a) was performed. Final pathology revealed a mucinous cystic neoplasms (MCN) without evidence of malignancy, as well as a FNH in segment 4a. The postoperative course was complication-free. a: Portal-venous phase transversal image, b: T1 weight contrast-enhanced coronar image, c: T2 weight transversal MRI image with Turbo Spin Echo and d: macrospopic view of the resected specimen.
Approach to FLL (from (3)).
Because sensitive and specific biomarkers for FLL are still missing, demographic factors may additionally help guiding the treatment decision. We and others were able to show that age over 55 years and male gender may act as a risk factor for malignancy (15, 69). Other risk factors that have been described previously, e.g. size and multilocular appearance, were not confirmed in our collective (15, 69) (Table III). The present study has several significant limitations mainly because of the retrospective study design and the small collective compromising statistic power and evidence level of our data. Large prospective studies are needed to investigate the best approach for the rapidly increasing population of asymptomatic FLL patients and eventually allow evidence-based diagnostic and therapeutic treatment.
Conclusion
A standardized and evidence-based approach in individuals presenting with asymptomatic FLL is still missing. This, however, will be necessary to prevent unnecessary under-treatment and potentially harmful (surgical) over-treatment in the future and, also, decrease the financial burden associated with FLL.
Acknowledgements
This work was supported by a BONFOR (“Bonner Forschung”) Research Grant (O-1120051) provided to HM by the University Medical Center, University of Bonn, Bonn, Germany.
Footnotes
- Received March 16, 2016.
- Revision received April 19, 2016.
- Accepted April 20, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.