The Liver Imaging–Reporting and Data System (LI-RADS) is a comprehensive system for standardized interpretation and reporting of computed tomography and magnetic resonance examinations performed in patients at risk for hepatocellular carcinoma. LI-RADS includes a diagnostic algorithm, lexicon, and atlas as well as suggestions for reporting, management, and imaging techniques. This primer provides an introduction to LI-RADS for radiologists including an explanation of the diagnostic algorithm, descriptions of the categories, and definitions of the major imaging features used to categorize observations with case examples.
Key points
- •
The Liver Imaging–Reporting and Data System (LI-RADS) provides a diagnostic algorithm that uses imaging features to categorize the observations seen in patients at risk for hepatocellular carcinoma (HCC) along a spectrum from benign to malignant.
- •
The major features used for categorizing observations with LI-RADS include diameter, arterial-phase hyperenhancement, washout appearance, capsule appearance, and threshold growth.
- •
The major features should be considered present only if they are unequivocally present so as to preserve high specificity for LI-RADS 5 (definite HCC) observations.
- •
In addition to a diagnostic algorithm, LI-RADS includes several free online resources to aid radiologists, including an atlas and lexicon, technical recommendations, reporting guidelines and templates, and management suggestions.
Introduction
The American College of Radiology (ACR)–supported Liver Imaging–Reporting and Data System (LI-RADS) is a comprehensive system for standardized interpretation and reporting of computed tomography (CT) and magnetic resonance (MR) examinations performed in patients at risk for hepatocellular carcinoma (HCC). The system has been developed by a committee composed of diagnostic and interventional radiologists, hepatologists, liver transplant and hepatobiliary surgeons, informatics and lexicon experts, and ACR staff members. LI-RADS development has been based on literature review, expert opinion, rounds of testing and iteration, feedback from users, and the need for congruence with HCC diagnostic criteria in the American Association for the Study of Liver Disease’s (AASLD) guidelines and in the United Network for Organ Sharing–Organ Procurement and Transplantation Network’s (UNOS-OPTN) policy. The first version of LI-RADS was launched in March 2011, and an update was released in January 2013. The release of a second update is planned in 2014, and the authors anticipate annual updates for the foreseeable future.
In the last decade, several clinical practice guidelines have been published worldwide by organizations in an attempt to standardize the surveillance, diagnosis, and management of HCC. However, the imaging-based criteria for HCC diagnosis in these guidelines have several limitations, which LI-RADS addresses. The imaging features used to assess lesions in other systems have not been precisely defined or illustrated, which leads to ambiguity in implementation and limits reproducibility in clinical care and for research. LI-RADS provides a free online atlas ( http://www.acr.org/Quality-Safety/Resources/LIRADS ), lexicon, and diagnostic algorithm as a reference for radiologists. Prior systems have generally classified lesions as HCC, indeterminate, or malignant; the suggested management for indeterminate lesions is often biopsy. In these prior systems, however, the indeterminate category can be very broad and often includes lesions that are likely to be benign and could safely be followed with serial imaging rather than biopsy. By making the indeterminate category more granular, LI-RADS allows for more nuanced and personalized management. Other systems apply to specific populations, such as those undergoing ultrasound surveillance or awaiting liver transplantation. LI-RADS applies to observations seen in all patients at risk for HCC, regardless of their prior imaging or transplant eligibility. All other systems neglect macrovascular invasive HCC. Because macrovascular invasion is part of TNM staging and greatly influences treatment options (eg, precluding liver transplantation as a therapeutic option), consistent and accurate diagnosis and reporting of macrovascular invasion is critical. LI-RADS provides a separate category for observations with macrovascular invasion and provides criteria for making the diagnosis. Other systems do not address non-HCC malignancies, such as intrahepatic cholangiocarcinomas (ICC), that may arise in patients at risk for HCC. It is now known that patients at risk for HCC also have an elevated risk for developing ICC. The differentiation between HCC and ICC is important as the prognosis and management differ considerably. LI-RADS provides guidance to radiologists for distinguishing these malignancies and a separate category for reporting.
In addition to addressing the shortcomings of other systems described earlier, the LI-RADS committee has also developed a structured reporting template, guidance on optimal imaging techniques, and management suggestions. This primer demonstrates the practical utility of LI-RADS in clinical and research practice with an emphasis on MR imaging.
Introduction
The American College of Radiology (ACR)–supported Liver Imaging–Reporting and Data System (LI-RADS) is a comprehensive system for standardized interpretation and reporting of computed tomography (CT) and magnetic resonance (MR) examinations performed in patients at risk for hepatocellular carcinoma (HCC). The system has been developed by a committee composed of diagnostic and interventional radiologists, hepatologists, liver transplant and hepatobiliary surgeons, informatics and lexicon experts, and ACR staff members. LI-RADS development has been based on literature review, expert opinion, rounds of testing and iteration, feedback from users, and the need for congruence with HCC diagnostic criteria in the American Association for the Study of Liver Disease’s (AASLD) guidelines and in the United Network for Organ Sharing–Organ Procurement and Transplantation Network’s (UNOS-OPTN) policy. The first version of LI-RADS was launched in March 2011, and an update was released in January 2013. The release of a second update is planned in 2014, and the authors anticipate annual updates for the foreseeable future.
In the last decade, several clinical practice guidelines have been published worldwide by organizations in an attempt to standardize the surveillance, diagnosis, and management of HCC. However, the imaging-based criteria for HCC diagnosis in these guidelines have several limitations, which LI-RADS addresses. The imaging features used to assess lesions in other systems have not been precisely defined or illustrated, which leads to ambiguity in implementation and limits reproducibility in clinical care and for research. LI-RADS provides a free online atlas ( http://www.acr.org/Quality-Safety/Resources/LIRADS ), lexicon, and diagnostic algorithm as a reference for radiologists. Prior systems have generally classified lesions as HCC, indeterminate, or malignant; the suggested management for indeterminate lesions is often biopsy. In these prior systems, however, the indeterminate category can be very broad and often includes lesions that are likely to be benign and could safely be followed with serial imaging rather than biopsy. By making the indeterminate category more granular, LI-RADS allows for more nuanced and personalized management. Other systems apply to specific populations, such as those undergoing ultrasound surveillance or awaiting liver transplantation. LI-RADS applies to observations seen in all patients at risk for HCC, regardless of their prior imaging or transplant eligibility. All other systems neglect macrovascular invasive HCC. Because macrovascular invasion is part of TNM staging and greatly influences treatment options (eg, precluding liver transplantation as a therapeutic option), consistent and accurate diagnosis and reporting of macrovascular invasion is critical. LI-RADS provides a separate category for observations with macrovascular invasion and provides criteria for making the diagnosis. Other systems do not address non-HCC malignancies, such as intrahepatic cholangiocarcinomas (ICC), that may arise in patients at risk for HCC. It is now known that patients at risk for HCC also have an elevated risk for developing ICC. The differentiation between HCC and ICC is important as the prognosis and management differ considerably. LI-RADS provides guidance to radiologists for distinguishing these malignancies and a separate category for reporting.
In addition to addressing the shortcomings of other systems described earlier, the LI-RADS committee has also developed a structured reporting template, guidance on optimal imaging techniques, and management suggestions. This primer demonstrates the practical utility of LI-RADS in clinical and research practice with an emphasis on MR imaging.
Observations
LI-RADS is used to assign a category to each observation in the liver. An observation is an area with imaging features that differ from those of adjacent liver parenchyma. The term observation is preferred over the terms lesion or nodule because some areas categorized by LI-RADS may represent pseudolesions, such as artifacts or perfusion alterations. LI-RADS does not assign a category to the entire liver, and observations with different categories may be identified throughout the liver.
Categories
LI-RADS categories are ordinal scores assigned to observations. The categories span the spectrum from benign to malignant observations often encountered in patients at risk for HCC. LI-RADS categories provide more granularity than prior systems for the degree of suspicion for malignancy in those observations that are not diagnostic for HCC.
LI-RADS 1 (LR-1) category observations are those observations that demonstrate imaging features diagnostic of a benign entity or that are no longer present on follow-up imaging in the absence of treatment. The radiologist should have absolute certainty that the observation is benign to categorize it as LR-1. Examples of definite benign entities that can be considered LR-1 include hemangiomas, cysts, and perfusion alterations ( Table 1 ).
| Definite Benign Entities | Probable Benign Entities |
|---|---|
| Cyst | Cyst |
| Hemangioma | Hemangioma |
| Vascular anomaly | Vascular anomaly |
| Perfusion alteration | Perfusion alteration |
| Hepatic fat deposition or sparing | Hepatic fat deposition or sparing |
| Hypertrophic pseudomass | Hypertrophic pseudomass |
| Confluent fibrosis | Confluent fibrosis |
| Focal scar | Focal scar |
| Observation that spontaneously disappears at follow-up | LR-2 cirrhosis-associated nodule |
LI-RADS 2 (LR-2) category observations are those that are probably benign; however, the imaging features are not diagnostic of a benign entity, such as a hemangioma with an atypical enhancement pattern. With few exceptions, the lists of probable benign entities (LR-2) and definite benign entities (LR-1) are identical (see Table 1 ). One entity that is not present on the list of definite benign entities but can, in some circumstances, be categorized probably benign is the cirrhosis-associated nodule. Cirrhosis-associated nodules that radiologists should consider reporting and categorizing with LI-RADS are those nodules that are distinctive from background nodules because of size, heterogeneity, enhancement, or signal intensity difference. Although a cirrhosis-associated nodule may not appear suspicious for malignancy with MR imaging, neoplastic histologic or molecular features may be present. Because neoplastic histologic or molecular features cannot be excluded with certainty at noninvasive imaging, cirrhosis-associated nodules that are distinctive from background liver cannot be considered definitely benign and, hence, should be categorized LR-2 or greater. In order to be categorized LR-2, a cirrhosis-associated nodule should be homogeneous, less than 20 mm in diameter, and iso-enhancing relative to background nodules. If the nodule does not meet these criteria, it should be categorized LI-RADS 3 (LR-3) or greater. LR-3 observations are those that have a moderate probability of representing either HCC or a benign entity. LR-3 observations do not meet the criteria for other LI-RADS categories.
LI-RADS 4 (LR-4) observations have imaging features suggestive, but not diagnostic, of HCC. Although there is a high probability that an LR-4 observation represents HCC, there is not 100% certainty in the diagnosis.
LI-RADS 5 (LR-5) observations are those that have imaging features diagnostic of HCC or are proven to be HCC at histology. The criteria for LR-5 have been constructed to be very specific for the diagnosis so as to prevent a false-positive diagnosis of HCC and subsequent treatment or transplantation in the absence of HCC. Because patients at risk for HCC may also develop other malignancies, such as ICC, high specificity for HCC is necessary to prevent inappropriate treatment or transplantation for non-HCC malignancies. As a consequence of the high specificity for HCC, some small or early HCCs may be categorized as LR-4 or LR-3. Clinicians should be aware that an LI-RADS category less than LR-5 does not exclude HCC. LR-5 observations are essentially equivalent to OPTN 5, as described in the UNOS-OPTN policy.
LI-RADS 5V (LR-5V) is a specific category for observations with imaging features diagnostic of HCC with macrovascular invasion ( Fig. 1 ). HCC with vascular invasion requires a separate category in LI-RADS because vascular invasion greatly alters patient management and is considered a contraindication to liver transplantation or surgical resection. LR-5V applies to cases with definite enhancing soft tissue in a vein, even if an intraparenchymal HCC is not visible.
LI-RADS M (LR-M) is used for observations that are probably malignant but not specific for HCC, such as observations with features suggestive of ICC or metastasis. The LI-RADS Treated (LR-Treated) category is for observations that have undergone locoregional treatment. Because therapy can alter the appearance of HCC, the evaluation of these lesions following treatment may require different criteria than lesions that have not been treated. This category is under development, and criteria for assessment of residual or recurrent disease based on the type of therapy will be released in future versions of LI-RADS.
Algorithm
The LI-RADS diagnostic algorithm is intended to mimic the natural thought process of a radiologist evaluating an observation in patients at risk for HCC, yet provide additional structure, particularly for indeterminate observations, that allows more reproducible reporting between radiologists. The algorithm for LI-RADS version 2014 is shown in Fig. 2 .
The authors briefly walk the reader through the algorithm. The authors recognize, however, that radiologists may skip decision nodes in actual practice based on pattern recognition and experience.
As emphasized in the algorithm, LI-RADS applies only to observations in patients at risk for HCC; the populations at risk consist mainly of patients with cirrhosis or chronic viral hepatitis B, as discussed in detail in the AASLD’s practice guidelines. The algorithm should not be applied to patients not at risk for HCC development. On identifying an observation on a CT or MR examination in patients at risk for HCC, the radiologist should first determine whether the observation has undergone prior treatment. The designation as LR-Treated should be based on patient history or characteristic imaging features. If an observation is considered untreated, then the radiologist should assess for imaging features that are diagnostic of a benign entity (LR-1). If the observation is slightly atypical for a benign entity, but is still thought to have a high probability for a benign entity, it can be categorized LR-2. If the radiologist decides that the observation is neither definitely nor probably benign, then the algorithm guides the radiologist to assess whether there are features suggestive of a malignancy other than HCC, such as ICC. If an observation is thought to be probably malignant but not specific for HCC, the observation is categorized LR-M. If malignancy other than HCC is not thought to be likely, the radiologist then looks for enhancing tissue in portal or hepatic veins indicating vascular invasion (category LR-5V). If the radiologist does not see definite tumor in vein, he or she is guided to evaluate for major features as shown in the table near the bottom of the algorithm to assess the degree of suspicion for HCC and assign the appropriate category (LR-3, LR-4, or LR-5).
Most 10- to 19-mm observations with arterial-phase hyperenhancement and a single additional major feature are categorized as LR-4. However, some observations with arterial-phase hyperenhancement and a single additional major feature are categorized as LR-5us (the us designation indicates that ultrasound visibility was incorporated into the categorization) or LR-5g (the g designation indicates that diameter increase was used in the categorization). This practice is done to maintain congruence with the AASLD’s and the UNOS-OPTN’s diagnostic systems, respectively. For example, 10- to 19-mm observations detected as discrete nodules at antecedent surveillance ultrasound and showing arterial-phase hyperenhancement and washout appearance (but not capsule appearance or threshold growth) meet the HCC criteria according to the AASLD’s guidelines. Moreover, prospective studies have shown that for 10- to 19-mm observations, this combination of features (visibility as discrete nodule at antecedent surveillance ultrasound, arterial-phase hyperenhancement, and washout appearance) provides approximately 100% specificity for HCC, which is consistent with the conceptual definition of LR-5. For this reason, such observations are categorized as LR-5us by LI-RADS version 2014. Similarly, 10- to 19-mm observations with arterial-phase hyperenhancement and ≥50% diameter increase in ≤6 months (but not washout appearance or capsule appearance) are categorized as HCC and class 5A-g according to the UNOS-OPTN system. For this reason, such observations are categorized as LR-5g by LI-RADS version 2014.
Note that a radiologist experienced in liver imaging will often move quickly through the algorithm and its decision nodes, and LI-RADS categorization is usually a rapid process in clinical practice.
Major features
Major features are imaging features used to categorize LR-3, LR-4, and LR-5 observations and include arterial-phase enhancement, diameter, washout appearance, capsule appearance, and threshold growth. Major features should be considered present only if they are unequivocally present. Absolute confidence in the presence of major features is necessary to maintain high specificity for HCC. Many of the major features are used by other guidelines and are widely recognized, yet have not previously been defined in a standardized manner. Lack of standardized definitions may lead to inconsistent application and interpretation of the features both in clinical care and in research, leading to inconsistency in patient care, educational materials, and published work. One of the key goals of LI-RADS is to provide standardized definitions for these features, which are discussed briefly next.
Arterial-phase hyperenhancement is defined in LI-RADS as enhancement in the arterial phase that is unequivocally greater than that of background liver ( Fig. 3 ). This feature reflects the angiogenesis that accompanies hepatocarcinogenesis in progressed HCC. According to the WHO Classification of Tumors of the Digestive System , HCCs can be divided into early HCCs and progressed HCCs. Early HCCs are an incipient stage of HCC development, analogous to carcinoma in situ or microinvasive carcinoma in other organs, with limited ability to invade vessels or metastasize. These early lesions are typically vaguely nodular in morphology, lack tumor capsules, are well differentiated histologically, and have incomplete neo-arterialization (and, hence, do not hyperenhance in the arterial phase). By comparison, progressed HCCs are advanced lesions with the capacity to invade vessels and metastasize. These progressed lesions are usually distinctly nodular in morphology with frequent capsule and septum formation, although very aggressive forms may show extracapsular extension and/or infiltrative growth. They are moderately or poorly differentiated and have advanced neo-arterialization (and, hence, usually hyperenhance in the arterial phase). Although arterial-phase hyperenhancement is a characteristic feature of progressed HCC, it is nonspecific as it can be seen in benign observations, such as perfusion alterations, atypical cases of focal or confluent fibrosis, hemangiomas, pseudoaneurysms, focal nodular hyperplasias; in nonmalignant cirrhosis-associated nodules, such as some dysplastic nodules; and in non-HCC malignancies, such as ICCs. As opposed to progressed HCCs, most early HCCs and most precursor nodules (cirrhotic nodules, low-grade dysplastic nodules, and high-grade dysplastic nodules) tend to be hypoenhancing or isoenhancing in the arterial phase.
