Catching CKD Early: Utilizing AI and its Clinical Significance

EP: 1.Overview of Chronic Kidney Disease (CKD) and Impact on Patients

EP: 2.Identifying CKD: Screening and Guideline Implementation

EP: 3.Filling the Gap: Testing to Identify Early-Stage CKD

EP: 4.Identifying and Managing CKD With Other Comorbid Conditions

EP: 5.Emerging Treatments and Preventative Care for CKD

Now Viewing

EP: 6.Catching CKD Early: Utilizing AI and its Clinical Significance

EP: 7.AI CKD Test Clinical Implementation and Provider Pearls

EP: 8.Final Pearls and Takeaways for Testing and Managing CKD

Transcript

Dhiren Patel, PharmD: So you mentioned the KidneyIntelX tool, and it’s basically, I would say, an AI tool that kind of integrates type 2 diabetes and CKD data for earlier staging and management, which fills a big void in kind of a gap that we have, even though we have both of these tests, and in many cases they may not get ordered, but it gives you a very complete picture, especially I think in what you said, the earlier places where the uACR [urine albumin-creatinine ratio] or the estimated glomerular filtration rate might otherwise show that it’s normal, and then you’re able to pick this up. So could you talk to us a little bit about fundamentally what this KidneyIntelX test is? And then next we’ll go into the clinical utility, but for our viewers maybe just a high-level overview of what this is going to be able to help with.

Robert Busch, MD: So the viewers are probably familiar with what a CRP [C-reactive protein] is, looking for an inflammatory marker. There are inflammatory markers within the kidney that when inflammation is in the tubules, there are these markers called TNF1 and 2 [tumor necrosis factor 1 and 2], which are inflammatory or fibrosis markers going on in the tubules of the kidney. Normally TNF2 is not present unless there’s inflammation, and TNF1 can be inflammation in the glomerulus or in the tubule. So those markers might be very elevated, and another test called kidney injury marker, or KIM. So what this [Renalytix KidneyIntelX] did is give different weightings for these markers, TNF1 [and] 2, and kidney injury markers, as well as look at some clinical things from the patient’s history, their GFR [glomerular filtration rate], their blood pressure, the uACR, plus some other factors. And they weighted each of these based on what you said with artificial intelligence. And they came up with a point score, a very simple test. It looks like a speedometer where it’s low, medium, or high. And if you’re high risk, 90 or over, and it’s read on the test—the test comes back very easily interpretable—it means do something. It’s like an electric shock to you as the provider to…do something. Same as if you had a high cardiac calcium score in a patient with diabetes, you would be much more aggressive with lowering their LDL [low-density lipoprotein] and [give a] big statin…vs just, OK, they’re on a statin, they’re fine. So it puts that patient at higher risk for prognosis of getting end-stage kidney disease, and it tells you to do something for that patient, do something more to benefit their kidneys other than the RAS—maximize the RAS, add an SGLT2 [sodium-glucose cotransporter 2], add finerenone, do what you can to benefit the kidneys.

Dhiren Patel, PharmD: Since this past ADA [American Diabetes Association] conference in June—and the tool and the company has done a tremendous job of kind of creating a body of evidence. There [are] some great real-world evidence studies, and there [are] also publications that we saw at the ADA in which we saw reductions in the estimated GFR slope and A1c reductions over a year. How do you interpret the clinical significance of something like this?

Robert Busch, MD: So when they incorporated it in the Mount Sinai health system, and the patients had the point score done, those patients were at high risk, their clinicians were more aggressive, maximizing the RAS, maximizing blood pressure control, adding an SGLT2, maybe referring to a nephrologist early on—vs, if the score wasn’t that high, they didn’t have to throw the book at the patient and do the works on everyone. So they did this in a health system already, and even over 6 months they saw changes and the changes you pointed out, that your GFR slope, instead of going down 4 or 5 points per year, stabilized somewhat, or these patients’ A1C improved because they were more aggressive controlling the diabetes…to avoid kidney disease. So it’s a marker that’s showing you your patients on the high-speed slope to dialysis. Do something. Don’t just leave them alone on what they’re on because they’re getting worse despite that.

Dhiren Patel, PharmD: And they also have a study that they’ve looked at in the [veterans] population, is that correct?

Robert Busch, MD: They did a large study looking at thousands of patients to get this different weighting of the different factors, and that’s why these proprietary tests, KIM and TNF1 and 2 receptors…were incorporated with the clinical picture vs just looking at KDIGO [Kidney Disease Improving Global Outcomes] guidelines, these enhanced KDIGO guidelines to point out who’s much more likely to go on to end-stage renal disease, and it’s an actionable test to do something about it.

Dhiren Patel, PharmD: I think what you just said is a great way to look at it. It’s an actionable test. It immediately tells you that you need to do more, like you said.

Transcript was AI-generated and edited for clarity.