KIM-1 Demonstrates Potential Utility as a Biomarker With Worse DFS Outcomes in KIM-1–High RCC

With the presentation of an exploratory analysis of KIM-1 biomarker data from the phase 3 IMmotion010 trial (NCT03024996) at the 2024 ASCO Annual Meeting, a third prospective trial has shown that higher KIM-1 levels at baseline result in worse disease-free survival (DFS) outcomes for patients with resected high-risk renal cell carcinoma (RCC), according to Brian Rini, MD. However, treatment with immunotherapy may improve DFS vs placebo among patients with high KIM-1 levels at baseline.

Data from the analysis revealed that patients with high KIM-1–high status at baseline (n = 300) experienced a median DFS of 35.88 months vs 57.23 months among those with KIM-1–low status at baseline (n = 452; HR, 1.75; 95% CI, 1.40-2.17). Further, patients with high baseline KIM-1 serum levels treated with atezolizumab (Tecentriq; n = 151) achieved a median DFS of not estimable compared with 21.16 months for those given placebo (n = 149; HR, 0.72; 95% CI, 0.52-0.99). A baseline KIM-1 level of 86 pg/mL was determined to be the optimized threshold for distinguishing between KIM-1–high vs KIM-1–low subgroups in the study.¹

“[KIM-1] needs to be validated prospectively in another data set [and] there are other adjuvant trials, [however,] the phase 3 KEYNOTE-564 study of pembrolizumab [Keytruda; NCT03142334] is the only positive trial out,” Rini said in an interview with OncLive®. “It would be great to see data in that setting [and] to understand whether [KIM-1] can be complementary to our risk stratification models. To estimate risk of recurrence for an individual patient, you look at T stage and N stage, grade, and perhaps necrosis. But those models are 2 to 3 decades old, and they’re not very good. You can put the same patient data into 3 different models and get 3 very different estimates of recurrence, which tells you none of them are very good [or] accurate.”

In the interview, Rini, who is chief of Clinical Trials, Ingram Professor of Cancer Research, and a professor of medicine in Hematology/Oncology at Vanderbilt-Ingram Cancer Center in Nashville, Tennessee, detailed the findings to date regarding KIM-1 in RCC and next steps with the potential biomarker.

OncLive: How could KIM-1 potentially aid clinicians in determining the risk status of patients with HCC?

Rini: I see a lot of patients in clinic who have T3 disease, but it’s low grade and barely invasive into the perinephric fat, where I’m thinking ‘I don’t think this patient has high-risk [disease], but they’re high-risk by criteria.’ Conversely, a patient who has only T2 disease, but has [poor] histologic features, [I would] think ‘this patient is probably higher risk than the nomograms would estimate.’ KIM-1 measurement could assist in those patients were our traditional models fall short. Whether those patients should get adjuvant therapy [is a] big question that will require big trials. But it’s at least a starting point and we haven’t had a starting point until now.

There’s still not uniform consensus on who the patients are [with RCC who are] most suitable for adjuvant therapy. Despite the DFS and overall survival benefit with pembrolizumab, there’s still some debate in the field and different doctors who know a lot about this, do things differently, which tells you it’s not uniform. What if you have a patient who’s technically high risk, but only has a 30% risk of recurrence, [and] you’re exposing them to potential lifelong or life-threatening toxicities [with adjuvant therapy]?

I tend to give adjuvant pembrolizumab [to patients at] the higher end of the KEYNOTE-564 risk [criteria], but other doctors give it to everybody, and other doctors don’t give it anybody. If [KIM-1 expression] could at least help refine that, remove some of those lower-risk patients or include them appropriately, much like ctDNA [does], the treating physician would feel a lot better about exposing that patient to the risk [of adverse effects].

What data have been presented surrounding KIM-1 as a potential biomarker for kidney cancer?

The most relevant and recent [data on KIM-1] are from 3 prospective studies—the adjuvant phase 3 ASSURE [NCT00326898] study with sorafenib [Nexavar] and sunitinib [Sutent] vs placebo, the phase 3 CheckMate 914 [NCT03138512] trial with ipilimumab [Yervoy]/nivolumab [Opdivo] or nivolumab vs placebo, and more recently IMmotion010 with atezolizumab vs placebo. Each of those looked at KIM-1, which is a soluble protein, at baseline, at a couple of post-nephrectomy time points, and at recurrence, roughly—not all the studies did it exactly the same way.

The assays for KIM-1 are not identical, although similar, and the cut points used were not identical either [in the trials]. When you have 3 different groups looking at the same biomarker, they’re not all going to do it identically. That’s a relatively minor point, I don’t believe it detracts, but it’s important to set that limitation and nobody’s prospectively defined the level or the cut point of KIM-1 that’s more or less prognostic or predictive. There’s a lot of work to do in standardizing the assay and refining the cut points.

Having said that, all those studies showed that patients with high baseline KIM-1 levels had a worse outcome receiving placebo, so it’s prognostic—independent of receiving any treatment or [placebo] patients who had baseline high KIM-1 levels have a worse outcome in terms of DFS. To me, [these findings are] analogous to ctDNA in other diseases. If a patient has an elevated ctDNA after getting a presumably localized tumor cut out, maybe it wasn’t localized, and the patient is at higher risk for recurrence. I believe at least by analogy [KIM-1] has some parallels there.

[KIM-1 is] also shown to be predictive looking at the immune-based therapy studies. CheckMate 914 used quartiles of expression [and] IMmotion010 used a cut point that was not quite at the median—it was approximately at the 60% mark in terms of higher KIM-1 levels predicting response to immune therapy—[and] those patients did better [with immunotherapy]. Higher KIM-1 levels are adversely prognostic but predicted a greater benefit to immunotherapy in both those studies. Despite some of the differences, there are relatively consistent results.

What did the biomarker analysis of the IMmotion010 study presented at ASCO 2024 demonstrate?

Data presented at first wasn’t just focused on KIM-1—the initial [biomarker identification portion] looked at [approximately 3000] circulating proteins and KIM-1 stood out in terms of its predictive and prognostic ability. There was then work to look at the optimal cut point and it’s important to note the limitations; a certain cut point wasn’t defined in one dataset and then validated in another, [investigators] looked for predictive potential and found that the 60% cutoff vs the 40% was optimal in terms of differentiating the arms.

[KIM-1] was both predictive and prognostic, independent of arm, high KIM-1 [levels] were associated with worse survival. Patients with KIM-1–high [disease] had a greater effect from receiving adjuvant atezolizumab which was not seen in the intention-to-treat population [of IMmotion010 where] there was no benefit overall. But the subset of patients with KIM-1–high disease as defined [experienced] a benefit [from treatment with atezolizumab]. It’s exploratory [data and] hypothesis generating, so there are a lot of limitations but the strength of the data both within that set and now across 3 major phase 3 trials caught people’s attention to say, ‘this is a viable biomarker’.

Also, ctDNA detection has not been as robust in kidney cancer for various reasons. There are new methods being developed, it will have a home, but it’s generally been low. Unlike bladder cancer, and other non-genitourinary cancers, we haven’t had ctDNA to guide therapy or to design prospective trials. By analogy, [KIM-1] may be a biomarker to do that.

What do you want your colleagues to know about KIM-1 as it stands today?

KIM-1 is a circulating protein that is expressed in patients with resected, high-risk kidney cancer, and higher expression appears to be associated with worse DFS independent of therapy [received] and [patients also experience] a better outcome with immunotherapy. We need to do a lot more work to make this applicable in clinical practice. This is not something you can measure if you want to in the clinic, and even if you could I’m not sure what I would do with the results. Unlike [how] there are commercial ctDNA companies you can send your patient’s blood to; I’m often not quite sure what to do with those results, I don’t act on it clinically, but many do. This is even further behind that. We’re not even at the point of commercial assays or quite knowing what to do with it, but I believe those days are coming.

Reference

Albiges L, Bex A, Suárez C, et al. Circulating kidney injury molecule-1 (KIM-1) biomarker analysis in IMmotion010: a randomized phase 3 study of adjuvant (adj) atezolizumab (atezo) vs placebo (pbo) in patients (pts) with renal cell carcinoma (RCC) at increased risk of recurrence after resection. J Clin Oncol. 2024;42(suppl 16):4506. doi:10.1200/JCO.2024.42.16_suppl.4506