TGF-β Blockade Plus PD-L1 Inhibition Triggers Hyperprogression in Relapsed Small Cell Lung Cancer

Why This Study Matters

Relapsed small cell lung cancer (SCLC) carries a dismal prognosis, with few treatment options and only modest benefit from PD-1/PD-L1 checkpoint inhibitors alone after platinum-based chemotherapy fails. Combining checkpoint blockade with inhibitors of TGF-β — a stromal, immunosuppressive signaling pathway thought to limit immunotherapy response — has been pursued as a way to overcome this resistance. But this investigator-initiated Phase Ib/II trial from the National Cancer Institute (NCI) shows that the tumor-intrinsic consequences of blocking TGF-β signaling in SCLC can cut both ways: durable responses in some patients, but a markedly elevated rate of hyperprogressive disease (HPD) in others.

Study Design

This was an investigator-initiated, non-randomized, open-label, single-center Phase Ib/II trial (NCI protocol 18-C-0110; NCT03554473) evaluating bintrafusp alfa, a bifunctional fusion protein that simultaneously targets PD-L1 and TGF-β, in patients with relapsed SCLC. The trial ran at the NCI between April 2019 and July 2022 using a Simon two-stage design within each treatment arm. Eligible patients were adults with histologically or cytologically confirmed relapsed SCLC or extrapulmonary small cell carcinoma who had progressed after at least one prior line of platinum-based chemotherapy.

Patients with relapsed SCLC were enrolled across three treatment arms, with a separate cohort for extrapulmonary small cell carcinoma: an arm receiving bintrafusp alfa monotherapy, with the option to cross over to a combination arm at disease progression; an arm combining bintrafusp alfa with topotecan; an arm combining it with temozolomide; and an extrapulmonary cohort treated with bintrafusp alfa plus temozolomide. The primary objective was to evaluate efficacy, measured by objective response rate (ORR).

Patient Population

The trial enrolled 37 patients, of whom 34 were evaluable for efficacy, all treated at a single site. Mean age was 62 years; 78% of patients were White and 62% were male. Most patients (76%) had an ECOG performance status of 1 and had received a median of two prior lines of therapy (range, 1-6). Seventy percent had platinum-resistant disease, defined as progression within 90 days of the last platinum-based chemotherapy, and 54% had received a prior immune checkpoint inhibitor.

Primary Endpoint Results

In the relapsed-SCLC cohort, one partial response occurred in the monotherapy arm, with two additional partial responses among patients who crossed over to the temozolomide-combination arm after progressing. The topotecan-combination arm did not meet its prespecified efficacy threshold in the first stage and was closed early; the temozolomide-combination arm also did not meet its primary endpoint, though one patient in that arm achieved a partial response. In the extrapulmonary small cell carcinoma cohort, two partial responses were observed, including one lasting more than 16 months in a patient with small cell laryngeal cancer.

The Hyperprogressive Disease Signal

The trial's most clinically significant finding was not efficacy but an unexpected safety signal. Among the 21 patients with progressive disease, 13 (38%) (95% CI, 22.9%-55.4%) met pre-specified criteria for hyperprogressive disease (HPD) — a doubling or more of tumor growth rate, time to treatment failure under two months, and at least a 50% increase in tumor burden. That rate is notably higher than has historically been reported with PD-1/PD-L1 monotherapy alone.

Subgroup and Correlative Analyses

Investigators evaluated established clinical risk factors for HPD; two reached statistical significance: elevated baseline LDH (odds ratio 15.0; 95% CI, 1.36-408.0; P=0.04) and a high neutrophil-to-lymphocyte ratio above 4.67 (odds ratio 6.66; 95% CI, 1.50-37.6; P=0.01). Liver metastases and higher metastatic burden were more frequent among HPD patients but did not reach statistical significance, and no significant associations were observed for age, sex, brain metastases, or platinum sensitivity.

Correlative immune profiling comparing patients with clinical benefit (progression-free survival greater than three months, n=10) to those without (n=19) found higher pre-treatment levels of granzyme H (P=0.04), caspase-8 (P=0.02), and pleiotrophin (P=0.02) in the clinical-benefit group, and higher CXCL5 in the no-benefit group (P=0.04). Tumors from patients with clinical benefit also showed higher frequencies of two subsets of CD4+ memory T cells (P=0.0063 and P=0.02) and increased ERK2 phosphorylation in CD4+ T cells (P=0.04), alongside reduced immunosuppressive features such as naive regulatory T cells (P=0.03) and TIM3 and CTLA4 expression on CD8+ T cells (P=0.04 and P=0.03).

The authors also cite hyperprogression rates observed in separate bintrafusp alfa trial datasets shared by collaborators — 39.5% in EBV-positive nasopharyngeal carcinoma, 20.1% in a Phase 2 cervical cancer cohort, and 7.28% in the bintrafusp alfa arm of a Phase 3 non-small cell lung cancer trial versus 0% in the pembrolizumab comparator arm (P=0.0004) — as cross-study, not within-trial, evidence that elevated HPD risk may be specific to bintrafusp alfa rather than confined to SCLC.

Safety Profile

All 37 patients received at least one treatment cycle (median 2, range 1-14). Any grade 3 or higher adverse event occurred in 67% (25/37) of patients overall, and non-hematologic grade 3 or higher events occurred in 19% (7/37). One dose-limiting toxicity occurred during the safety run-in — grade 4 thrombocytopenia in the topotecan-combination arm — with no dose-limiting toxicities in the other two SCLC arms.

Immune-related adverse events included grade 2 hypothyroidism (14%, 5/37), adrenal insufficiency (5%, 2/37), hyperthyroidism (3%, 1/37), and hypophysitis (3%, 1/37). One patient (3%) died of a grade 5 gastrointestinal hemorrhage arising from liver metastases after a single treatment cycle.

These results suggest that tumor microenvironment-directed strategies should be guided by tumor-intrinsic signaling context, with TGFβR2-dependent signaling representing a candidate biomarker for HPD risk.
— Schroeder et al., Discussion, Cancer Discovery (2026)

Interpretation and Broader Context

This trial offers one of the most detailed clinical characterizations to date of hyperprogression following combined PD-L1/TGF-β blockade. While bintrafusp alfa produced durable responses in a subset of heavily pretreated patients — including tumor regression lasting more than a year in some cases — more than a third of patients with progressive disease experienced accelerated, treatment-associated tumor growth linked to significantly worse survival. Integrative multi-omic profiling and functional laboratory studies converged on tumor-intrinsic TGF-β signaling as a key determinant of this divergent outcome: TGF-β appears to act as a growth-restraining signal in a subset of SCLC tumors, and removing that restraint through pathway blockade may unleash hyperproliferation.

The findings argue against unselected use of dual PD-L1/TGF-β blockade in SCLC and instead support a biomarker-driven strategy for patient selection, since tumor-intrinsic TGF-β pathway status appears to determine whether pathway blockade helps or harms an individual patient. Elevated baseline LDH and a high neutrophil-to-lymphocyte ratio emerged as clinical risk factors that could help flag patients at higher risk of HPD before treatment begins.