Immunotherapy in locally advanced cervical cancer: Integrating KEYNOTE-A18 into management strategies

Summary

In locally advanced cervical cancer (LACC), the benefit of PD-1 blockade was unknown. In KEYNOTE-A18, Lorusso et al. compared the efficacy and safety of adding pembrolizumab to chemoradiation in LACC and demonstrated favorable outcomes. Given multiple approved indications of pembrolizumab in cervical cancer, strategies for optimal integration into management will be needed to maximize overall survival.

Main body

As the 4^th most common malignancy in women, cervical cancer continues to be a major global health issue with an annual incidence and mortality of 660,000 cases and 350,000 deaths worldwide¹. Although screening and prevention strategies such as Papanicolaou smears and Human Papilloma Virus (HPV) vaccination have proven ability to decrease the incidence of this cancer, their implementation is far from uniform and locally advanced cervical cancer (LACC) still accounts for over one-third of newly diagnosed cases². LACC, defined by the International Federation of Gynecology and Obstetrics (FIGO) as FIGO 2018 stage IB3-IVA, is treated with radiation with concurrent chemotherapy (typically weekly cisplatin; referred together as “chemoradiation”). Nearly all cervical cancers are caused by chronic high-risk HPV infection that is unable to be effectively cleared by the immune system. Persistently infected cells subsequently progress to cervical dysplasia and carcinoma over a period of several years. Given this association, there has been great interest in utilizing immunotherapy in the management of cervical cancer. The use of pembrolizumab (anti-PD-1 monoclonal antibody) has demonstrated efficacy as monotherapy and in combination with chemotherapy and bevacizumab for recurrent or metastatic cervical cancer^3,4. However, the feasibility and efficacy of combining PD-1 inhibitors to chemoradiation and whether there may be synergistic anti-tumor effects was unknown⁵. In a phase III, multicenter, randomized control trial (KEYNOTE-A18, NCT04221945), Lorusso et al. sought to evaluate the efficacy and safety of chemoradiation with or without pembrolizumab in LACC⁶. These results were published in Lancet and will be reviewed here⁶.

In KEYNOTE-A18, 1060 participants were randomized 1:1 to the chemoradiation (external beam radiation followed by brachytherapy with weekly cisplatin chemotherapy for 5 – 6 cycles) with pembrolizumab or placebo every 3 weeks for 5 cycles. Following the 5 cycles, maintenance pembrolizumab or placebo were administered every 6 weeks (maximum of 15 cycles). The two primary co-endpoints were progression-free (PFS) and overall survival (OS). Key eligibility criteria included 1) FIGO 2014 stage IB2 – IIB with node-positive disease (corresponding to stage FIGO 2018 stage IIIC) or stage III-IVA regardless of nodal status 2) cervical squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma histologic subtypes 3) no prior systemic therapy, definitive surgery, or radiation. PD-L1 positivity was not required for eligibility.

Median PFS and OS were not reached for either group but median follow-up was only 17.9 months. The 24-month PFS rates were 68% (95% CI 62 – 73) and 57% (95% CI 51 – 63) for the pembrolizumab and placebo groups, respectively. The pembrolizumab group had statistically significant improvement in PFS (HR 0.70, 95% CI 0.55 – 0.89, one-sided p = 0.002) compared to the placebo group. Subgroup analyses demonstrated improved PFS with pembrolizumab among patients with PD-L1 positive tumors (HR 0.72, 95% CI 0.56 – 0.92) but not PD-L1 negative tumors (HR 0.61, 95% CI 0.18 – 2.07). Also, pembrolizumab was favorable for patients with FIGO 2014 stage III-IVA (HR 0.58, 95% CI 0.42 – 0.80) but not those with node-positive FIGO 2014 stage IB2 – IIB (HR 0.91, 95% CI 0.63 – 1.31). For OS, data is still immature, however in the current report, there was no difference between the pembrolizumab and placebo groups (HR 0.73, 95% CI 0.49 – 1.07). The 24-month OS rates were 87% (95% CI 82 – 91) and 81% (95% CI 75 – 86), respectively. Toxicity was higher in the pembrolizumab group with greater risk difference (≥5%) of any adverse event with respect to thyroid disorders, leukopenia, hypokalemia, and aspartate aminotransferase elevations. Grade ≥3 treatment-emergent (75% vs 69%), grade ≥3 treatment-related (67% vs 61%), and serious treatment-related (17% vs 12%) adverse events were higher in the pembrolizumab group compared to the placebo group; the most common toxicities were leukopenia, neutropenia, and anemia.

The KEYNOTE-A18 investigators should be applauded for their excellent work and publication. Because of the KEYNOTE-A18 results, pembrolizumab was granted US FDA approval on January 12, 2024 for use with chemoradiation for untreated FIGO 2014 stage III-IVA regardless of PD-L1 status. This indication translates to patients with FIGO 2018 stage IIIA, IIIB, or IVA regardless of nodal involvement. Given pembrolizumab with or without chemotherapy has approval in the metastatic and recurrent setting, this approval provides another indication for the drug in cervical cancer.

However, this trial also raises further questions on how to best integrate immune checkpoint inhibitors into management. In KEYNOTE-826, chemotherapy with pembrolizumab with or without bevacizumab demonstrated a 12-month OS benefit compared to chemotherapy with or without bevacizumab in the immunotherapy-naïve population (28.6 vs 16.5 months, respectively)⁴. Given the modest benefit in PFS, with as of yet unknown OS benefit, observed in the interim analyses of KEYNOTE-A18, it is unclear whether addition of PD-1 blockade should be employed in the frontline chemoradiation setting (per KEYNOTE A-18) or reserved for patients who experience a recurrence (per KEYNOTE-826). Furthermore, it is unknown whether retreatment with pembrolizumab following recurrence will still provide an OS benefit in patients exposed to this drug during chemoradiation. Lastly, with multiple indications for pembrolizumab in cervical cancer, sequencing and cost-effectiveness studies will be crucial to ascertain optimal management strategies.

Additionally, it should be noted that the proportion of patients with PD-L1 positivity in KEYNOTE-A18 (94.3%), is on the high-end of the reported range of prevalence (range 34.4 – 96%)⁸. Given known benefit of pembrolizumab in patients with PD-L1 positive tumors, current FDA approval for pembrolizumab for PD-L1 positive tumors in the recurrent setting, and lack benefit of in the PD-L1 negative tumor subgroup in KEYNOTE-A18 (albeit small cohort), the use of pembrolizumab with chemoradiation in PD-L1 negative LACC should be investigated in further studies.

It is noteworthy, that in another similarly designed randomized phase III trial (CALLA NCT03830866), chemoradiation with durvalumab (PD-L1 inhibitor) did not demonstrate improvement in PFS compared to chemoradiation with placebo in a similar population of LACC (HR 0.84, 95% CI 0.65 – 1.08, p = 0.17)⁹. The 24-month PFS rates were 65.9% (95% CI 59.8 – 71.4) and 62.1% (56.2 – 67.4%). The reasons for the lack of observed PFS benefit in CALLA but detectable statistical difference in KEYNOTE-A18 are unclear. With very similar (but not identical) study patient populations, the 24-month PFS rates were close in numerical value in the pembrolizumab (68%) and durvalumab (65.9%) groups. However, the placebo arm in CALLA outperformed that on KEYNOTE-A18 (62.1% vs. 57%). Two factors have been described as potential attributions. Although similar immune checkpoints, pembrolizumab and durvalumab may have variable impact on anti-tumor immunity given target differences (PD-1 on cytotoxic T-cells vs PD-L1 on tumor cells). Additionally, the KEYNOTE-A18 authors highlight differences in eligibility criteria regarding size and number of lymph nodes to be considered node-positive (e.g. KEYNOTE-A18 required ≥2 nodes with 1.5 cm in short axis; CALLA required ≥1 nodes with 1 cm in short axis). The authors postulate that KEYNOTE-A18 represented a higher risk population and therefore may be more likely to benefit from immunotherapy. Although that is a possibility, there was no observable difference in the node-positive stage 1B2 – IIB subgroup in KEYNOTE-A18 and it is unclear the proportion of stage III/IVA patients also had nodal positivity since it was not requirement for that population.

Another important ongoing study is the INTERLACE trial¹⁰. In this multicenter phase III study, 500 patients with LACC were randomized to receive induction chemotherapy (weekly carboplatin and paclitaxel for 6 cycles) followed by chemoradiation versus chemoradiation alone. In the preliminary results, the 5-year PFS rates were significantly greater in the induction chemotherapy group compared to the chemoradiation alone group (73% vs 64%; HR 0.65, 95% CI 0.46 – 0.91, p = 0.013). Additionally, there was an improvement in 5-year overall survival in the induction chemotherapy group (80% vs 72%; HR 0.61, 95% CI 0.40 – 0.91, p = 0.04). Whether immune checkpoint inhibitors can augment the potential benefit of induction chemotherapy followed by chemoradiation in LACC remains to be determined.

Lastly, CheckMate-358 demonstrated that dual CTLA-4 and PD-1 blockade in chemotherapy- and immunotherapy-naïve recurrent cervical cancer patients may hold significant therapeutic benefit with sustained anti-tumor responses when compared with patients who had prior systemic treatment⁷. Currently, there is an ongoing trial evaluating the use of dual CTLA-4 and PD-1 blockade first prior to physician’s choice systemic therapy as a sequencing triage strategy to improve OS in chemotherapy-naïve advanced or recurrent cervical cancer [NCT05475171]. It remains to be seen if dual checkpoint blockade may provide additional benefit beyond anti-PD1 blockade in patients with advanced cervical cancer.

The results of KEYNOTE-A18 reaffirm of role of PD-1 inhibitors in the overall management strategy of cervical cancer management. Like CALLA, the updated PFS and OS results of KEYNOTE-A18 are eagerly anticipated. Future studies should focus on how to best utilize immune checkpoint inhibitors in combination and/or sequencing strategies to improve overall survival and minimize toxicity and costs for cervical cancer patients.