“Histologic subtype shouldn’t exclude patients from getting checkpoint inhibitor therapy,” says Natalie J. Miller, MD, PhD.
San Francisco-Response to immune checkpoint inhibition does not vary by histologic subtype of urothelial carcinoma (UC). In a multi-institutional retrospective cohort study, the overall response rate (ORR) and overall survival (OS) were similar between patients with pure UC or variant UC, reported Natalie J. Miller, MD, PhD, at the Genitourinary Cancers Symposium in San Francisco.
“There are mixed data as to whether variant UC responds as well as pure UC to standard-of-care therapy, such as chemotherapy, as well as radiotherapy,” said Dr. Miller, of the University of Washington, Seattle.
“There are also some studies that show that variant UC can be more aggressive, present at higher grade or higher stage and be more likely to have quicker recurrences,” added Dr. Miller, working with Petros Grivas, MD, PhD, and co-authors.
Despite these differences, the data demonstrated that the ORR to immune checkpoint inhibition between patients with pure UC and variant UC was comparable, at 28% versus 29%, respectively (p=.90). Median OS was 11.0 months among the patients with pure UC versus 10.1 months for those with variant UC (p=.60).
“We looked at patients who were treated with any one of five checkpoint inhibitors, any line up to fourth line,” she said.
The study included 424 patients from 18 academic institutions with UC who received immune checkpoint inhibitors and met inclusion criteria; 395 were evaluable for ORR, 406 for OS, and 403 for progression-free survival (PFS) analysis.
Among the 406 patients in the OS analysis cohort, there were 286 with pure UC, 120 with variant UC, and nine with neuroendocrine UC. Slightly more than two-thirds in each cohort were male and 65% to 78% were ever smokers. Fourteen to 16% had upper tract tumors. Fifty-one percent in the pure UC cohort had undergone cystectomy or (nephron)ureterectomy compared with 60% in the variant UC cohort and 29 in the neuroendocrine cohort.
When assessed by variant histology, the ORR was 28% in those with patients with squamous histology, 28% with micropapillary histology, 38% with sarcomatoid histology, 17% with plasmacytoid histology, 22% with adenocarcinoma, and 25% with neuroendocrine UC. The ORR for patients with variant UC excluding neuroendocrine was 29%.
Next: No difference in response according to the line of checkpoint inhibitor therapyThere was no difference in response according to the line of checkpoint inhibitor therapy (first line or subsequent lines).
“When you break the variant group down into the particular type of variant, it showed that patients with neuroendocrine carcinoma variant had a significantly shorter OS of about only half as long as the patients with either pure or other variant histologies [4.6 vs. 11.0 vs. 10.1 months; p=.003 vs. pure or variant UC],” Dr. Miller said. “That was significant still on multivariate analysis. Obviously, neuroendocrine, no matter what type of cancer, is going to be a little more aggressive, but even accounting for some of those factors, it appeared to be associated with poorer OS, and we see a similar trend on PFS.”
By histologic subtype, median OS was 10.5 months in those with squamous UC, 9.8 months with micropapillary histology, 11.5 months with sarcomatoid histology, 10.1 months with plasmacytoid histology, not reached in patients with adenocarcinoma, and 4.6 months in those with neuroendocrine UC.
In patients with upper tract tumors, OS was significantly shorter in those with variant versus pure UC (4 vs. 14 months; p=.03).
“One of the surprising factors of the neuroendocrine subset is that they responded but progressed,” Dr. Miller said. PFS after initiating immune checkpoint inhibitor therapy was 5.2 months in those with variant histology excluding neuroendocrine, 4.1 months in patients with pure UC, and 3.7 months in those with neuroendocrine UC (p=.03 vs. variant without neuroendocrine).
“Histologic subtype shouldn’t exclude patients from getting checkpoint inhibitor therapy,” she said. “There’s a lot of evidence that PD-L1 expression in these tumors is fairly similar between pure and variant UCs and that tumor mutational burden is also pretty similar, so that could be why they’re responding similarly.”