LI‐RADS: Future Directions

Abbreviations

AASLD

American Association for the Study of Liver Diseases

ACR

American College of Radiology

AI

artificial intelligence

BI‐RADS

Breast Imaging Reporting and Data System

CEUS

contrast‐enhanced ultrasound

CT

computed tomography

ECA

extracellular contrast agents

FAQ

frequently asked questions

HBA

hepatobiliary contrast agents

HCC

hepatocellular carcinoma

LI‐RADS

Liver Imaging Reporting and Data System

MRI

magnetic resonance imaging

NCCN

National Comprehensive Cancer Network

OPTN

Organ Procurement and Transplant Network

PDFF

proton density fat fraction

RECIST

Response Evaluation Criteria in Solid Tumors

SC

steering committee

UCSD

University of California San Diego

UNOS

United Network for Organ Sharing

US

ultrasound

WG

working group

Background

The Liver Imaging Reporting and Data System (LI‐RADS) standardizes imaging technique, interpretation, and reporting in patients at risk for hepatocellular carcinoma (HCC). The most recent version of LI‐RADS includes four algorithms: ultrasound (US) for surveillance, contrast‐enhanced ultrasound (CEUS) for diagnosis, computed tomography (CT)/magnetic resonance imaging (MRI) for diagnosis and staging, and CT/MRI for treatment response assessment (Fig. 1).

Fig 1.

LI‐RADS algorithms. Reproduced under CC BY‐NC‐ND 4.0.

The development of LI‐RADS is led by a standing steering committee (SC) of expert diagnostic and interventional radiologists, hepatologists, and surgeons. Representatives from other societies, including the American Association for the Study of Liver Diseases (AASLD) and United Network for Organ Sharing (UNOS), hold memberships in the SC, helping to ensure congruence between LI‐RADS and other stakeholders. The SC oversees numerous LI‐RADS working groups (WGs), each responsible for specific charges (Fig. 2). The SC and WGs apply scientific evidence, expert opinion, user feedback, and global perspective to refine, improve, simplify, and expand the algorithms. This review summarizes key goals in the future directions of LI‐RADS.

Fig 2.

LI‐RADS organizational chart.

Continued Validation, Refinement, and Simplification

Endorsed by the American College of Radiology (ACR), LI‐RADS was initially released in 2011 and has undergone several updates since, with the latest release in 2018. Each update introduced new material, addressed ambiguities, and provided enhancements (Fig. 3). The overarching goal to achieve unification between LI‐RADS and other imaging‐based diagnostic HCC systems led to the incorporation of elements from other systems. This included the adoption of capsule appearance, an imaging feature used in the UNOS system, in 2013 and the expansion of the LR‐5 (Definite HCC) criteria in 2018 to achieve consistency with and integration into the AASLD HCC clinical practice guidance. ¹ The need to keep the system up to date with the most recent scientific evidence and technological advancements is balanced against the need for stability, ensuring sufficient time between updates to allow users to build familiarity and researchers to collect and publish data to inform future improvements. As a compromise between these competing needs, future updates to LI‐RADS algorithms are planned to occur every 3 to 5 years, with the next release targeted for late 2021.

Fig 3.

LI‐RADS development timeline.

Universal Lexicon

Clear and accurate communication of imaging findings to the referring provider is the paramount goal of a radiology report. Use of inconstant terminology introduces variability and ambiguity, obscures intended meaning, causes miscommunication, ² and fails to convey the radiologist’s level of confidence. ³ , ⁴ , ⁵ Variability in terminology and inconsistencies in definitions also challenge the synthesis of the published literature, impeding progress. To overcome these problems, LI‐RADS developed a standardized lexicon for liver imaging. The lexicon provides precise terminology and definitions to reduce interpretation variability and errors, enhance communication, enable the development of imaging‐based registries, and facilitate the pooling of scientific data. The current version of the lexicon, vetted and approved by the LI‐RADS Steering Committee, is released and available free of charge at the ACR website (https://www.acr.org/-/media/ACR/Files/RADS/LI-RADS/Lexicon-Table-2020.pdf). Although the lexicon is currently focused on a narrow set of terms relevant specifically to imaging of patients at risk for HCC, the lexicon will be expanded in breadth and scope to all liver imaging contexts, including terms relevant to liver imaging in the general population.

Treatment Response Assessment

Released in 2017, the LI‐RADS Treatment Response algorithm is designed to assess tumor response using CT or MRI after locoregional therapy or surgical resection, and it has been validated for ablation and chemoembolization. ⁶ , ⁷ , ⁸ , ⁹ Unlike other treatment response systems, such as Response Evaluation Criteria in Solid Tumors (RECIST) or Modified RECIST, which were developed for use as research endpoints in clinical trials, the LI‐RADS Treatment Response algorithm is intended to inform treatment decisions in clinical care. The Treatment Response WG is now refining the algorithm to incorporate emerging evidence, expanding the algorithm to address newer treatment options, including radiotherapy and systemic therapy, and collaborating with the CEUS WG to develop treatment response criteria for CEUS.

Clinical Adoption by Community Radiologists

Although LI‐RADS has been widely adopted in academic centers and is an accepted tool for scientific studies, adoption by community and general radiologists has lagged. Perceived system complexity and the presence of competing HCC imaging algorithms are potential barriers to global utilization. ¹⁰ The Outreach and Education WG seeks to promote LI‐RADS adoption across all clinical settings by developing and disseminating educational materials, such as webinars, educational exhibits, review articles, lectures, and hands‐on workshops, to familiarize radiologists and other specialists with LI‐RADS. In addition, users are encouraged to submit questions through the ACR LI‐RADS website (https://www.acr.org/Clinical-Resources/Reporting-and-Data-Systems/LI-RADS). Members of the SC respond to the questions, and the submitted queries are then added to the curated and continuously updated collection of frequently asked questions (FAQ), which is available on the ACR site.

High clinical volume in private practice poses an additional barrier to utilization of LI‐RADS, because standardized reporting is more time consuming than free dictation. Clinical decision support tools are being developed by the ACR to improve interpretation speed. These tools will be integrated into commercial dictation software, guiding the user to describe all relevant LI‐RADS imaging features and automatically generating a structured report, including the appropriate LI‐RADS categories. The LI‐RADS Reporting WG will validate these tools and work with the ACR to keep them updated through future LI‐RADS updates.

LI‐RADS is an imaging‐based system used by radiologists interpreting imaging studies, and one of the purposes of LI‐RADS is improved communication between the radiologists and clinicians (including hepatologists, gastroenterologists, and primary care physicians). As described earlier, the use of standardized reproducible terminology improves clarity of communication between radiologists and clinicians. As LI‐RADS continues to be adapted by the community radiologists, the community clinicians and, most importantly, their patients will benefit from the reproducible and clear reports. Furthermore, standardization of interpretation of these complex studies by the community radiologists may in the long term remove the current requirement for reinterpretation of these studies at the tertiary care centers, thereby potentially reducing any delay in patient care.

ADVANCED Technologies

In addition to the short‐term goals outlined earlier, LI‐RADS plans to leverage emerging technologies to better define the target population and to improve system accuracy and usability. Currently, the LI‐RADS diagnostic population is restricted to patients with cirrhosis, chronic hepatitis B, or personal history of HCC. ¹¹ It is possible, however, that the target population could be expanded by validating and applying laboratory or imaging biomarkers of HCC risk, such as elastography‐determined liver stiffness. A Quantitative Imaging WG will be convened to help guide proper use of quantitative imaging for risk stratification.

Utilization of artificial intelligence (AI) in radiology is rapidly increasing. Potential applications of AI in LI‐RADS include lesion detection, lesion tracking over serial examinations, imaging feature characterization, and LI‐RADS categorization. Natural language processing could be used to extract data elements from free speech dictation and automatically generate structured reports. The creation of large international imaging and clinical data registries will facilitate AI training, inform further algorithm refinement, and ultimately enable granular probability estimation based on the integration of clinical, laboratory, and imaging factors (Fig. 4).

Fig 4.

LI‐RADS vision: future integration of patient and clinical and imaging data to arrive at precise probability of malignancy and hepatocellular origin.

Conclusion

LI‐RADS is a dynamic system that will continue to evolve. Under the oversight of the LI‐RADS SC, various standing WGs apply scientific evidence, address user feedback, and incorporate national and global perspective to improve existing algorithms and develop new algorithms in response to clinical needs.