Amit Singal, MD: We’ve been using these imaging criteria for years, and now the idea of LI-RADS has come into the diagnostic radiology literature. This is a reporting system that can actually standardize the way that imaging is reported by radiologists across the country. At UT Southwestern Medical Center, our radiologist would have reported it one way. A radiologist at Memorial Sloan Kettering Cancer Center would have reported it another way. A radiologist at Georgetown University Medical Center may have reported this completely different. And so, as a clinician, you didn’t know what the radiologist was truly trying to tell you. LI-RADS was proposed as a nomenclature, where we could start talking to each other in a very standardized fashion. LI-RADS basically takes any lesion and looks at criteria including size, the presence of arterial enhancement, delayed washout, and other features and classifies a lesion from LI-RADS 1, which is definitely benign, all the way up to LI-RADS 5, which is definitely HCC. The lesions that we truly get concerned about are LI-RADS 4 lesions (ie, suspicious for HCC) and LI-RADS 5 lesions (definite for HCC).

Mark W. Karwal, MD: Most oncologists are familiar with LI-RADS. They’re called BI-RADS (Breast Imaging Reporting and Data System) in breast imaging. It’s the same thing. In breast imaging, the same number, BI-RADS 5, means that you have breast cancer until proved otherwise. LI-RADS 5—that’s liver cancer until proved otherwise.

Ghassan K. Abou-Alfa, MD: I hear you.

Mark W. Karwal, MD: With a biopsy, there may be a 5% chance of stuff growing in the needle track, right where the biopsy is.

Ghassan K. Abou-Alfa, MD: Oh, no, no, no. Let’s not go there. The data have shown that the risk from the biopsy is 0.0000. Shall I keep going?

Mark W. Karwal, MD: It depends on which study you look at. But you can find 1% or 2%.

Ghassan K. Abou-Alfa, MD: We have to be careful here. Manish, what else do we need the biopsy for?

Manish R. Sharma, MD: From the perspective of a medical oncologist, we’re used to having that tissue. These days, we’re doing molecular profiling with it. We’re all using these panels, whether they’re our own in-house panels or commercial panels, to look for gene mutations that we think may be actionable—that may have some meaning in terms of what we can offer the patient, right? In HCC, that has lagged behind. Nonetheless, the growing trend among oncologists is that they want that information. We will occasionally discover things in a patient with HCC. I’ve had patients in whom a BRCA mutation was discovered. MSI-high status, as we know, is a marker for immunotherapy, right?

So there are reasons why you might gain valuable information from having a tissue diagnosis in addition to what Mark brought up about mixed tumors and things like that, where it looked like HCC but it ended up being more of a cholangiocarcinoma. These are really important pieces for us, but I think molecular profiling is a strong argument for it. We still have a lot of clinical trials that require a tissue diagnosis as well. Being a part of an institution where we do a lot of that, that’s also an important factor.

Ghassan K. Abou-Alfa, MD: I hear you. Ruth, we’re not necessarily trying to defend biopsies, but, interestingly, the MSI story in regard to HCC is not really that important for clarity. We’re giving checkpoint inhibitors to every patient with HCC, regardless of their MSI status. Do you agree?

A. Ruth He, MD, PhD: Yes. The rate of MSI in liver cancer is very low. We do know that there are patients with MSI-stable disease, which is most of the HCC cases that we see. And we know that 15% to 20% of those patients respond to immunotherapy.

Ghassan K. Abou-Alfa, MD: Absolutely. We’re going to talk more about that, but this is an important point in the context of what we’re discussing.

Transcript Edited for Clarity