Hutson Compares Avelumab for PD-L1 Positive Versus PD-L1 Negative Bladder Cancer

Thomas Hutson, DO, PharmD, director, Urologic Oncology Program and cochair, Urologic Cancer Research and Treatment Center, Baylor University Medical Center, discussed the case of a 66-year-old patient with bladder cancer during a virtual Targeted Oncology Case-Based Roundtable event.

Targeted Oncology^TM: How did data compare between chemotherapies used in bladder cancer prior to the checkpoint inhibitor era?

HUTSON: There was a tail on the [Kaplan-Meier] curve for cisplatin that we don’t see with carboplatin.¹ In lung cancer, they’re used interchangeably. Even for some small cell lung cancers, physicians are using it interchangeably. But like with testicular cancer, platinum chemotherapy in genitourinary cancer and bladder cancer is not equivalent. There’s better activity with cisplatin than there is with carboplatin. Because of this [tail on the curve], there’s been a focus on [whether] we should be giving chemotherapy [in patients with bladder cancer] because there is a small group of patients that get chemotherapy that do well.

I think if you treat enough patients with bladder cancer, you see patients who end up having great results. The numbers are small, so you have to have a large number [of patients] to treat. You don’t get that [tail] with carboplatin or gemcitabine. There is also quite a difference in the overall survival [OS] in patients that can get the gemcitabine/cisplatin combination.

The overall response rates [ORRs] are quite different [between chemotherapy regimens]. Dose-dense methotrexate, vinblastine, doxorubicin, and cisplatin [MVAC] produced the highest ORR and complete response [CR] rate [of 72% and 25%, respectively], and median OS was 15.1 months with this regimen.

There’s some debate as to if it’s worth the toxicity, but these types of data show us that if [a patient] can get cisplatin, they should. Because if we don’t do that, if we negate chemotherapy in general, we’re going to lose any chance for a CR in these patients. We should consider chemotherapy still as the front line, especially if someone’s able to get cisplatin.

We take these data and extrapolate them in the perioperative setting. Right now, we’ll use these data to support neoadjuvant therapy. In that setting, the neoadjuvant therapy should be cisplatin. There’s a big movement, an education movement by the American Society of Clinical Oncology, to remind us that carboplatin is inferior, and you shouldn’t consider carboplatin in the neoadjuvant setting.

At Baylor University Medical Center, urologists are aware of the data of chemotherapy, and they tell me, “If I can’t give cisplatin [as a neoadjuvant therapy], I want to take the person to surgery. I don’t want neoadjuvant carboplatin.”

It’s hard for a community oncologist to know that kind of thing. But really, physicians should consider not giving carboplatin/gemcitabine [as a neoadjuvant therapy], because it’s so inferior. That’s pretty much what our experts tell us: We should be using cisplatin, or you can use split-dose gemcitabine/cisplatin.

A phase 2 trial [investigated] split-dose gemcitabine/cisplatin, every 21 days, that’s supposed to be more renally safe. I’ve [treated with] cisplatin down to creatinine clearances in the 35 mL/min range using a split-dose. That’s pretty much what we’re doing at Baylor University Medical Center—trying to go with gemcitabine/cisplatin one way or the other.

For patients who can’t get cisplatin, or I start them on cisplatin [and they] don’t do well, we send them to surgery because I’m always reminding myself that the curative therapy is not neoadjuvant chemotherapy. That adds at best 10% to their cure rate. The cure is surgery or radiation chemotherapy. That’s the take-home point. If they’re not tolerating chemotherapy, there’s no advantage to pushing it.

The other reminder is that in order to do radiation and chemotherapy as an alternative to surgery, you want to have had maximal transurethral resection of the bladder tumor. Then you still need to consider neoadjuvant chemotherapy.

Which trial backs the use of avelumab (Bavencio) in this setting?

The trial that really changed the standard of care, presented by Thomas Powles, MD, MBBS, MRCP, was a late-breaker abstract, LBA1.² JAVELIN Bladder 100 [NCT02603432] was a study that established avelumab maintenance as a standard of care. They took patients who had first-line chemotherapy with either cisplatin or carboplatin/gemcitabine for 4 to 6 cycles and scanned [them]. Their response was either stable disease [SD], partial response [PR], or CR.

[Patients] were randomized to avelumab, continued every 2 weeks until unacceptable toxicity or progression, or [patients received] best supportive care [BSC] alone. The stratification was based upon the type of chemotherapy and whether they had a CR, PR, or SD.

[Investigators] looked at metastatic sites, visceral versus nonvisceral, and PD-L1 status. The primary end point was OS. They looked at all randomized patients and the PD-L1–positive population. The secondary end points were progression-free survival [PFS], safety, and patient-reported outcomes. There was no crossover allowed. There were 700 patients.

What were the efficacy data in the JAVELIN Bladder 100 trial?

When we look at the Kaplan-Meier curve of OS, the separation [of the curves] was maintained throughout.

The median OS was 21.4 months [with avelumab] versus BSC alone, at 14.3 months. The hazard ratio was 0.69 [95% CI, 0.56-0.86], so that’s a pretty attractive hazard ratio for a cancer trial, and it had a statistically significant P value [P < .001].

It appears that patients who have at least SD with chemotherapy have a survival benefit by starting the checkpoint inhibitor right away and trying to maintain the benefit they got from chemotherapy. The median follow-up was 19.6 months with avelumab and 19.2 months with BSC.

The curves come back together [at a certain point], and there are a lot of theories as to why. One theory is that the patients who were [in the] BSC arm, when they progress, get a checkpoint inhibitor, so the checkpoint inhibitor is rescuing them. That’s why the curves come back together some. We know checkpoint inhibitors affect survival, and it’s the second-line agent of choice. Even though they get rescued with checkpoint inhibitors, it doesn’t catch up. There’s still benefit to early use of checkpoint inhibitors.

Across the board, all subgroups favored avelumab maintenance therapy. A couple of groups crossed the line. Age and performance status favored avelumab. The group with CR or PR seemed to have the greatest benefit [with avelumab], but it’s pretty similar to the SD group. Remember, the numbers are much smaller in the SD group. That may explain why it [crosses] 1.00. Then [with a] PD-L1 status of positive, negative, or unknown, there was still potential benefit or perceived benefit irrespective of PD-L1 status. However, PD-L1 positive status definitely increased likelihood of response.

Does PD-L1 status matter when using switch maintenance?

The OS [data in JAVELIN Bladder 100 looked at] the overall intention-to-treat population, which includes the PD-L1– positive or PD-L1–negative [population] versus the PD-L1–positive population alone. The OS was similar but [the PD-L1–positive group saw a] more enriched curve. They don’t have an OS median in that patient population [yet]. The BSC arm was at 17.1 months median OS. Further updates on these data are going to be useful.

How did patients do with the secondary end points?

PFS doesn’t really tell the story in this situation. For ORR in the PD-L1–positive population, [avelumab] does better than what we see in the overall population.

There were patients [who received prior] chemotherapy. They may not have had a CR [with chemotherapy], then they go on a checkpoint inhibitor and a few developed CRs. There are some people that had chemotherapy and had at least SD. You [may] put them on a checkpoint inhibitor, and they have progression.

Checkpoint inhibitors don’t work for everyone. The reason you’re choosing to [use them] is because of the entire patient population. You don’t know how your patient’s disease is going to behave. The checkpoint inhibitors have these benefit [data], and it will just have to play out on that individual patient. Unfortunately, there’s going to be about a third of your patients that don’t respond to a checkpoint inhibitor.

What were the adverse events (AEs) observed in this study?

When we look at AEs, the high level is what you would expect to see with a checkpoint inhibitor in this population [98.0%]. The AEs at grade 3 or more with the checkpoint inhibitor are low [47.4%]; the bulk of the AEs are going to be grade 1 or 2.

Remember, these patients are coming off chemotherapy and going on a checkpoint inhibitor, so they may already come into the checkpoint inhibitor with fatigue or cytopenias.

What’s interesting is the BSC arm gives you background [on how these patients would do normally]. That gives you the level of AEs that patients have that represents the disease itself and represents the AEs of the chemotherapy arm. Patients finish chemotherapy with some degree of AEs.

The treatment-emergent AEs led to discontinuation in about 12% of patients, which is low. That’s similar to what I’m familiar with avelumab when we’ve used it in kidney cancer. Death was attributed by investigator to study toxicity in 2 patients. One was sepsis, and the other was ischemic stroke.

If we look at immune-related AEs, the grade 3 level is quite low: 7% in total. Grade 1 or 2 was most predominant, and the most common [immune-related AE] was hypothyroidism. We saw skin rash and pneumonitis also.

High-dose corticosteroids, which they defined as greater than 40 mg of total daily prednisone or equivalent, were [used] in just 9% of avelumab-treated patients. That’s lower than what we’ve seen in other things like pembrolizumab [Keytruda] and nivolumab [Opdivo], where there’s a higher percentage [of patients] who need steroids.

Is avelumab approved in this setting?

Based upon these data—this is the first time we’ve ever had anything that’s shown survival benefit in patients after chemotherapy—the FDA approved [avelumab] for urothelial carcinoma maintenance treatment.³

Because of that, it’s been incorporated into the NCCN criteria as a preferred treatment after gemcitabine/cisplatin and the dose-dense MVAC.⁴ For cisplatin ineligibility, the preferred regimens…give you other choices, recognizing that there may be a reason to [treat with] atezolizumab [Tecentriq] or pembrolizumab in that setting too. What they’re looking at is PD-L1 expression. If the patient had high PD-L1 expression, then they were not eligible for platinum chemotherapy. Atezolizumab or pembrolizumab would [then be used] because that was the criteria for those trials, and they checked PD-L1 status.

Avelumab maintenance has really been incorporated into the standard of care and is now considered for all patients who are eligible for chemotherapy.

How do you sequence the therapies in this setting?

What sequence yields the longest median OS? The best OS is with switch maintenance.1 The greatest estimated median OS was 25 months with maintenance when you add up [the different therapies]. There was a phase 2 trial [NCT02500121] of [switch maintenance] with pembrolizumab. Avelumab and pembrolizumab had similar outcomes in the maintenance setting.

_References:

_{1. Plimack E. Discussion of LBA1. Presented at: 2020 American Society of Clinical Oncology Virtual Scientific Program; May 29-31, 2020; virtual.}

_{2. Powles T, Park SH, Voog E, et al. Maintenance avelumab + best supportive care (BSC) versus BSC alone after platinum-based first-line (1L) chemotherapy in advanced urothelial carcinoma (UC): JAVELIN Bladder 100 phase III interim analysis. J Clin Oncol. 2020;38(suppl 18):LBA1. doi:10.1200/JCO.2020.38.18_suppl.LBA1}

_{3. FDA approves avelumab for urothelial carcinoma maintenance treatment. FDA. June 30, 2020. Accessed March 9, 2021. https://bit.ly/3jaforD}

_{4. NCCN. Clinical Practice Guidelines in Oncology. Bladder cancer, version 1.2021. Accessed March 9, 2021. https://bit.ly/39PQGtI}