Patients with No Known Cancer or Chronic Liver Disease

In the absence of risk factors such as cancer or chronic liver disease, the likelihood that a small, incidental, indeterminate liver lesion is malignant is extremely low, regardless of the age of the patient or number of lesions. Therefore, we do not automatically recommend additional evaluation of incidental nonspecific lesions smaller than 1 cm in diameter in patients without risk factors for malignant disease, provided no other compelling indications exist. Examples of lesions that might require additional evaluation include multiple lesions in an immunocompromised patient or a lesion that shows evidence of growth.

Large lesions detected incidentally in asymptomatic healthy individuals have a greater chance for being malignant than small lesions. In a study of 107 asymptomatic patients with hepatic lesions incidentally discovered with a variety of imaging techniques, Liu and colleagues found that 58% of lesions were malignant. On multivariate analysis, male sex, age older than 50 years, and tumor size greater than 4 cm were predictive of malignant disease.

Ultrasound may be a reasonable initial step to evaluate a low-attenuation liver lesion in an accessible location in an average-size patient. Lesions less than 5 mm are unlikely to be detected with ultrasound. However, up to two thirds of lesions larger than 5 mm detected with CT can be located by ultrasound, provided the CT scan is used to direct the search. Targeted ultrasound is much more likely to succeed for a lesion in a favorable location (e.g., not the dome of the liver) and in a person of average body habitus. If ultrasound reveals a simple cyst, no further evaluation is necessary. Targeted ultrasound of an indeterminate hepatic lesion may also help to determine whether sonographically guided biopsy is an option.

Magnetic resonance imaging (MRI) has the highest sensitivity and specificity of available hepatic imaging techniques but is typically reserved for the evaluation of patients with risk factors for malignant tumor or infection. MRI is capable of distinguishing between benign and malignant lesions in the majority of cases referred for a small indeterminate lesion discovered with CT. However, MRI is costly and should be reserved for cases when further characterization of a lesion is likely to affect patient management. Such situations include oncology patients for whom hepatic resection or an alternate chemotherapy regimen is contemplated.

Nuclear medicine techniques are rarely helpful in the evaluation of small, incidentally discovered hepatic lesions because of limited spatial resolution.

Patients with Known Extrahepatic Cancer

Even in patients with known primary malignancy, a small incidental lesion discovered with CT is more likely benign than malignant (Fig. 7-1). In patients with known primary cancer, a lesion considered “too small to characterize” by CT has a greater than 80% chance of being benign according to most studies. The actual number can be expected to vary with the population studied, the criteria used to define small or “too small to characterize,” the imaging technique, and the interpreting individual. Patients with breast cancer with one or more small (≤15 mm) hypoattenuating lesions on a baseline contrast-enhanced CT scan without other evidence of hepatic metastases are no more likely to experience development of subsequent hepatic metastases than patients with no such lesions. In another study of patients with breast cancer, too small to characterize lesions represented benign findings in more than 90% of women. Patterson and researchers studied the MRI evaluation of “too small to characterize” liver lesions discovered with CT in patients with breast cancer and found that only 5% of such lesions were shown to represent metastases. In a study of hepatic lesions 15 mm or smaller in patients with gastric and colorectal cancer, almost 80% were benign. In that study, if patients with larger coexistent liver metastases were excluded, small hypoattenuating lesions were metastases in only approximately 2% of patients.

Figure 7-1 Enhanced axial computed tomographic image through the abdomen of a patient with colorectal carcinoma and multiple low-attenuation liver lesions. Given the small size and number of lesions, a diagnosis of biliary hamartomas was suggested and subsequently confirmed at the time of colectomy.

In some respects, it is not surprising that small incidental lesions discovered with CT are likely benign. After all, benign hepatic lesions are extremely common, and tiny, low-attenuation, simple cysts are likely to be more conspicuous against a background of enhancing liver than equally sized enhancing metastatic lesions.

A brief analysis of lesion features can sometimes help identify worrisome lesions, although no imaging features are entirely specific. Benign lesions tend to be lower in attenuation and have more discrete margins than malignant lesions, whereas target enhancement (low-attenuation border surrounding a lower attenuation center) is suggestive of metastatic disease. For patients with extrahepatic tumors, one may also take into account the stage and imaging characteristics of the primary tumor when analyzing a small focal hepatic lesion. For example, patients with known extrahepatic metastases are more likely to have liver metastases than patients with carcinoma in situ.

Patients with Chronic Liver Disease

Patients with chronic viral hepatitis or cirrhosis are at increased risk for development of hepatocellular carcinoma. Unfortunately, patients with chronic liver disease are also likely to be plagued by small, incidental, arterially enhancing lesions on dynamic, multiphase, contrast-enhanced imaging. The cause of such lesions varies and may include arterial-portal shunt, regenerative nodule, or benign neoplasm (e.g., hemangioma or focal nodular hyperplasia–like lesions). Fortunately, small (<2 cm), arterially enhancing nodules not visible on other phases of enhancement are more likely to be benign than malignant in the setting of chronic liver disease, although it may be difficult to distinguish benign nodules from hepatocellular carcinoma based on imaging criteria alone. Therefore, patients with chronic liver disease who have small arterially enhancing nodules on dynamic, multiphase, contrast-enhanced CT (or MRI) are usually followed with CT or MRI at approximately 6-month intervals. Small, arterially enhancing lesions that demonstrate intralesional washout of contrast material during the portal venous phase, a nodule-within-nodule appearance, rim enhancement, or evidence of a pseudocapsule on portal phase images should be considered malignant until proved otherwise.