About six years ago, hepatologists and liver surgeons at the University of California, San Diego (UCSD), were becoming increasingly frustrated with radiologists' inconsistency in interpreting liver lesions in patients with suspected hepatocellular carcinoma (HCC).
"Three different radiologists could look at the same case and use different words to describe the same thing," said Claude Sirlin, M.D., chief of Body Imaging and Abdominal MRI at UCSD. "Even worse, you could show the same case to the same radiologist on two separate occasions and he could use words like 'equivocal' the first time and 'indeterminate' the second time, and maybe even shift his opinion about that lesion from 'probably benign' to 'probably malignant'."
More confusing, a fourth radiologist could look at the same image and say it was "suspicious for cancer" even though specific criteria to determine exactly how suspicious the lesion was didn't exist.
The need for standardization was becoming critical, said Cynthia Santillan, M.D., an assistant professor and vice-chief of the Body Imaging Section and chief of Body CT at UCSD. "If we can't agree among ourselves, it undermines our credibility with clinicians and also impacts patient care," Dr. Santillan said.
"Clinicians told us we needed to create a system that would help radiologists become more consistent," Dr. Sirlin said.
After developing their own system of reporting based on degree of suspicion of HCC—which was received with the praise and appreciation of their fellow hepatologists—Drs. Santillan and Sirlin learned of similar systems being developed at other institutions around the country. The first step was to combine the UCSD system with that of Thomas Jefferson University Hospital in Philadelphia, which had developed an equivalent system. "We took the best of both, starting with an amalgam of the UCSD and the Thomas Jefferson system."
Efforts to create a universal system ultimately resulted in the American College of Radiology-sponsored committee that created Liver Imaging Reporting and Data System (LI-RADS). The committee, chaired by Dr. Sirlin, grew to include not only radiologists but also hepatologists, transplant surgeons, and other physicians from across the U.S. and the world.
Officially released in March 2011, LI-RADS includes five categories: (1) Definitely Benign, (2) Probably Benign, (3) Intermediate, (4) Probably HCC, and (5) Definitely HCC. The criteria for the last category have been constructed to be as specific as possible for HCC, to avoid false positive diagnosis and unnecessary treatment or transplantation, Dr. Santillan said.
"LI-RADS categories aim to reflect the criteria that many expert radiologists use to diagnose HCC," said Dr. Santillan. "Standardized terminology and categories will improve communication between physicians and enable reliable quality control and outcomes analysis."
In presenting, "Liver Imaging Reporting and Data System (LI-RADS): An ACR-supported System for Classification of Hepatic Lesions on CT and MR Imaging in Patients with Cirrhosis," at RSNA 2011, Dr. Santillan stressed the importance of standardization.
"There is a great deal of variability in terms of how people are being screened for hepatocellular carcinoma, using CT or MRI," Dr. Santillan said. "There are vast differences in the types of equipment used, the imaging sequences that are performed, as well as how many post-contrast phases or even pre-contrast phases are obtained in these patients. As you can imagine, the variability in the amount of information available for interpretation leads radiologists to come to different conclusions."
As for reporting, LI-RADS enumerates the elements—including anatomy, number of suspicious lesions and where they're located in the liver—that should be mentioned in every report, Dr. Santillan said. "These elements really need to be mentioned every time you're looking at one of these patients, because they can greatly influence the management options that are available to them," she said.
LI-RADS continues to be a work in progress, Dr. Sirlin said. The committee continues to fine-tune the first version of the project, LI-RADS 1.0, and recruited 100 radiologists from around the world to test the second version, LI-RADS 1.1, in early 2012. LI-RADS 1.1 will be released upon approval, Dr. Sirlin said.
To make LI-RADS more comprehensive, committee members are now developing a lexicon of controlled terminology and an atlas of teaching and illustrative cases, which will "help radiologists in difficult cases arrive at the most appropriate interpretation," Dr. Sirlin said.
While developing criteria for extracellular contrast agents is the focus of LI-RADS 1.0 and 1.1, a working group is being formed to develop criteria for hepatobiliary contrast agents as well, Dr. Santillan said.
Although they may not immediately incorporate hepatobiliary contrast agents, the lexicon and atlas will likely be released sometime in 2012, Dr. Santillan said. "In some respects we have modeled this after Breast Imaging-Reporting and Data System (BI-RADS), which has a very restrained lexicon so that there is a constrained set of terms that can be used to describe findings."
The lexicon will most likely be presented in the form of a digital manual with embedded hyperlinks, said Dr. Santillan, chair of the Lexicon Subcommittee.
The LI-RADS Reporting Subcommittee is developing minimum reporting standards to define what is needed to put into radiological reports "without being overly onerous," said subcommittee chair Benjamin Yeh, M.D., a professor in the UCSD abdominal imaging department.
The reporting committee is also creating a reporting template that radiologists "can drop into a standardized report so a clinician, regardless of institution, will get similar types of reports with similar types of criteria in a similar type of format," Dr. Yeh said.
LI-RADS is an "an active process that will continue to evolve," with a potentially significant long-term impact, Dr. Santillan said. "If we have standardized treatment protocols associated with these standardized liver lesions, then we can compare outcomes across institutions and across the world," she said. "That can really inform the management of these patients."
To view videos of Drs. Claude Sirlin and Cynthia Santillan discussing:
To read the complete LI-RADS criteria, go to www.acr.org and click Quality and Safety Initiatives in the lefthand column.
Daily Bulletin coverage of RSNA 2011 is available at RSNA.org/bulletin.
Join a global community of leaders in the radiologic sciences.
Continue your education with top-quality learning resources.
Witness the world's premier event in medical imaging, Dec 1–6.