Safety and efficacy of toripalimab plus concurrent chemoradiotherapy for locally advanced cervical cancer: a single-arm, phase Ib trial

Abstract

Background

Although two pivotal phase III trials have evaluated the efficacy of adding immune-checkpoint inhibitors (ICIs) to concurrent chemoradiotherapy (CCRT) in locally advanced cervical cancer (LACC), discrepancies between the two studies highlight the need for further exploration of ICIs combined with CCRT in this setting.

Methods

Totally 30 patients with FIGO 2018 stage IB3-IVA cervical cancer were enrolled in this study. All patients underwent pelvic external beam radiotherapy (50.4 Gy in 28 fractions) followed by brachytherapy (30–36 Gy in 5–6 fractions), concurrent with cisplatin (40 mg/m² weekly) and toripalimab (240 mg every two weeks). The primary endpoint was adverse events, and secondary endpoints included the objective response rate (ORR) and progression-free survival (PFS). This trial is registered at ClinicalTrials.gov (NCT04368273) on April 25, 2020.

Results

Twenty-four patients (80%) experienced grade 3 or higher treatment-related adverse events (TRAEs). The most frequent ≥ grade 3 TRAEs were leukopenia (63%), lymphopenia (37%), and anemia (27%). One patient discontinued treatment due to immune-related hemophagocytic lymphohistiocytosis (HLH). Twenty patients (66.7%) experienced late radiation-related toxicities, including rectal hemorrhage (53.3%), hematuria (10%), and peripheral sensory neuropathy (13.3%). No treatment-related death occurred. All patients showed an ORR of 100% (95% [CI]: 0.85.9-1) as their best response. The 2-year PFS was 90% (95% CI 0.72–0.96). The median PFS was not reached with a 27.3-month follow-up.

Conclusions

Combining toripalimab with CCRT had promising activity in LACC patients but might aggravate radiation-related adverse events and even cause serious immune-related adverse events. Therefore, ICIs therapy with CCRT needs further investigation.

Introduction

Cervical cancer is the second leading cause of cancer-related deaths in women aged 20 to 39 years worldwide [1, 2]. For locally advanced cervical cancer (LACC), the standard treatment is concurrent cisplatin-based chemotherapy and pelvic external beam radiotherapy (EBRT) followed by intracavitary high-dose-rate brachytherapy (HDR-BT) [3]. A meta-analysis of 15 trials involving 3,452 women, however, indicated a 6% improvement in 5-year survival by adding chemotherapy to radiotherapy (hazard ratio [HR] 0.81, P 0.001) [3]. Despite the best available therapeutic regimens, the risk of developing recurrent and/or metastatic disease in LACC patients remains as high as 40% [4].

Cervical cancer is almost always associated with human papillomavirus (HPV) infection, which in turn drives the expression of immune-reactive viral neo-antigens and overexpression of programmed death-ligand 1 (PD-L1) in cervical carcinoma tissues. Accordingly, cervical cancer might be particularly amenable to immunotherapy [5]. The KEYNOTE-826 trial established the efficacy and safety of adding immune checkpoint inhibitors (ICIs) to chemotherapy as first line therapy for recurrent or metastatic cervical cancer [6, 7]. Recently, the KEYNOTE-A18 trial suggested that pembrolizumab concurrent with CCRT led to a significant improvement in 36-month PFS rate (69.3% vs. 56.9%) and 36-month OS rate (82.6% vs. 74.8%) versus CCRT alone in LACC patients [8]. In the CALLA trial, however, adding duvalumab to CCRT failed to improve PFS in patients with LACC [9]. The discrepant findings between KEYNOTE-A18 and CALLA trials call for further investigation into ICIs therapy in LACC in the future.

Toripalimab, a humanized IgG4K monoclonal antibody specific for human PD-1, engages a differential binding domain on PD-1 than nivolumab and pembrolizumab by a crystal structure analysis [10], and has proven efficacious as a single agent or in combination with chemotherapy in a variety of solid tumors [11,12,13]. However, the safety and efficacy of toripalimab for cervical cancer remains unclear. Therefore, we conducted a phase Ib trial of toripalimab added to CCRT in treatment-naïve LACC patients.

Methods-patients

Study design and participants

This single-arm, phase Ib trial enrolled treatment-naïve, HPV positive, adult patients (18–75 years of age) with histologically-confirmed stage IB3-IVA (FIGO2018) cervical cancer. Additional inclusion criteria were: (1) HPV infection; (2) at least one measurable lesion per based on MRI; (3) Eastern Cooperative Oncology Group (ECOG) performance status score of 0–2 [14]; (4) an estimated life expectancy of at least 3 months; and (5) adequate organ and bone marrow function. Exclusion criteria included: (1) active or suspected autoimmune disease, serious uncontrolled diseases, human immunodeficiency virus, hepatitis B or C virus infections, or severe infection within 4 weeks; (2) prior history of allogeneic organ transplantation, except corneal transplantation, or allogeneic hematopoietic stem cell transplantation; (3) participation in clinical trials of anti-cancer therapy within one month. This trial was approved by the Institutional Review Board of Peking University Third Hospital (M2019482), and conducted in compliance to the Declaration of Helsinki and the International Conference on Harmonisation Guidelines for Good Clinical Practice. All patients provided written informed consent. Trial reporting conformed to the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. This trial is registered with ClinicalTrials.gov (NCT04368273).

Treatments

Pelvic EBRT (50.4 Gy in 28 fractions) was delivered with volumetric-modulated arc therapy (VMAT). Patients with positive lymph nodes received simultaneous integrated boost (SIB) of 10 Gy in 5 fractions. HDR-BT (total dose: 30–36 Gy) was delivered in 5 or 6 fractions (twice per week), starting one month after the beginning of EBRT. A pre-HDR-BT pelvic MRI was required to assess tumor response and guide subsequent therapy [15]. During pelvic EBRT, cisplatin (40 mg/m²) was administered weekly for 4–6 weeks; toripalimab (240 mg) was given biweekly for a total of four doses. CCRT was followed by consolidated chemotherapy with paclitaxel (260 mg/m²) and cisplatin (40 mg/m²) for 2–4 cycles as needed. Consolidation chemotherapy was administered if pelvic MRI at one month post-treatment showed residual local disease. Prior to brachytherapy, patients with a CR on MRI received five fractions of intracavitary brachytherapy. Those with residual disease received six fractions of intracavitary brachytherapy. The addition of interstitial brachytherapy was based on the location of the residual lesion. Treatment was discontinued in cases of unacceptable toxicity, pregnancy, disease progression, or withdrawal of consent. No dose modifications were permitted.

Assessments and outcomes

Safety was assessed weekly throughout the treatment period and up to 90 days post the final dose of study drugs. Adverse events (AEs) were defined and graded according to the National Cancer Institute-Common Terminology Criteria for Adverse Events (CTCAE, version 5.0) [16]. Late radiation-related AEs (AEs) occurring 1 year or more after the last date of radiotherapy, were graded according to the Radiotherapy and Oncology Group (RTOG) scale [17]. Tumor response was evaluated via pelvic MR by two radiation oncologists (PJ and JW) per RECIST version 1.1 every 3 months during the follow-up. The primary endpoint was the incidence and severity of AEs throughout the treatment period and up to 3 months after the completion of all interventions and late rAEs. The secondary endpoints included the ORR at 3 months post the final dose and PFS.

Statistical analysis

The 1-year PFS rate was 70% after concurrent chemoradiotherapy based on previous literature [9, 18]. We assumed that the addition of toripalimab to concurrent chemoradiotherapy would increase the 1-year PFS rate to 90%. Assuming a power of 80% and a two-sided type 1 error of 0.1, a sample size of 26 patients was required. Allowing for 10% attrition, we planned to enroll 29 patients.

The Kaplan–Meier (KM) method was used to estimate PFS and the corresponding 95% CI. χ2 test was used to compare recurrence rate in patients with different features. Statistical analyses were performed using R v4.1.3 software and GraphPad Prism v9 software was used to generate graphs.

Data availability

The data generated in this study are available upon request from the corresponding author Junjie Wang.

Results

Patient characteristics

Between May 8, 2020 and August 30, 2022, 153 patients with FIGO 2018 stage IB3-IVA cervical cancer were screened; 30 (median age: 58.5 years) were enrolled (Supplementary Fig. 1 and Table 1). Of the enrolled patients, twenty-eight patients (93%) had squamous cell carcinoma, and 2 (7%) had adenocarcinoma. Fourteen patients (47%) had IIIB/C disease, and 4 (13%) had IVA disease (Table 1). The median treatment duration was 51 days (range, 39 to 65). Twenty-nine patients (97%) completed a prescription dose of 50.4 Gy of pelvic EBRT. One patient discontinued EBRT after receiving 43.2 Gy in 24 fractions, four cycles of cisplatin chemotherapy, and five fractions of HDR brachytherapy, due to hemophagocytic lymphohistiocytosis (HLH). All patients received a total HDR-BT dose of 30 Gy (12 [40%]) or 36 Gy (18 [60%]). Seven patients exceeded the 2-cc OARs dose limitations as per published guideline [19], with median 2-cc doses of 72.72 Gy (range, 49.82–76.61) for the rectum, 65.91 Gy (range, 49.27–71.15) for the sigmoid colon, and 78.33 Gy (range, 23.08–90.95) for the bladder (Supplementary Fig. 2 and Supplementary Table 2). Sixteen patients (53%) received SIB. The median number of cisplatin cycles was 6 (range, 4–6), with 25 patients (83%) completing all 6 cycles. Twenty-nine patients (97%) received 4 doses of toripalimab, while one (3%) received 3 doses. Post-CCRT, 5 patients (16.7%) received consolidated paclitaxel and cisplatin (Supplementary Table 3).

Table 1 Demographic and baseline characteristics of the study population

Toxicity and safety

All 30 patients were included in the safety analysis. All patients experienced at least one treatment-related adverse event (TRAE), with grade 3 or higher TRAEs occurring in 24 patients (80%) (Table 2). The most frequent TRAEs were leukopenia (100%, 30/30), anemia (100%, 30/30), and nausea (90%, 27/30). Most TRAEs were mild or moderate in severity. The most frequent grade 3 or higher TRAEs were leukopenia (63%, 19/30), lymphopenia (37%, 11/30), and anemia (27%, 8/30). No TRAEs caused treatment interruptions or discontinuation, study termination, or death, except for one patient who had immune-related HLH and discontinued treatment. Radiation-related AEs (rAEs) occurred in 11 patients (37%; all grade 1 or 2). The most common early rAEs were proctitis (23%, 7/30), cystitis (17%, 5/30), vagina hemorrhage (17%, 5/30), dermatitis (13.3%, 4/30), and mucous membrane reaction (10%, 3/30) (Table 2). Late rAEs occurred in 20 (66.7%) patients, including rectal haemorrhage (16 [53.3%] of grade 2), hematuria (3 [10%] of grade 2), and peripheral sensory neuropathy (4 [13.3%] of grade 1) (Supplementary Table 4). Hematuria improved gradually following hemostasis under cystoscope; peripheral sensory neuropathy persisted until the final follow-up; rectal hemorrhage recurred intermittently despite various medicated enema treatments. For patients with late rectal hemorrhage, pharmacological enemas were administered under the guidance of gastroenterology and traditional Chinese medicine specialists. The enema regimens included aminosalicylates, glucocorticoids, mucosal protectants and astringents, Yunnan Baiyao, recombinant human epidermal growth factor, and traditional Chinese herbal preparations. Grade 1–2 immune-related adverse events (irAEs), were reported in 9 patients (30%) including chromatosis in 5 patients (17%) and rash, pruritus, hypothyroidism, and adrenal insufficiency in 2 patients (7%) each. One patient (3%) had grade 3 or higher irAEs, who developed immune-related HLH during the treatment.

Table 2 Treatment-related adverse events in the study population

Efficacy

All 30 patients were included in the efficacy analysis. At 3 months post-intervention, 27 patients (90%, 95% CI 72.3%−97.4%) achieved a CR and 3 (10%, 95% CI 2.6%−27.7%) achieved a PR, resulting in a 3-month ORR of 100% (30/30, 95% CI 85.9%−100%). Overall, 29 patients (96%, 95% CI 80.9%−99.8%) achieved CR and 1 (3%, 95% CI 0.17%−19%) attained PR as their best response, leading to a best ORR of 100% (30/30, 95% CI 85.9%−100%) (Table 3). The patient with PR experienced parametrial recurrence at 9.9 months. The median time to response was 0.9 months (range: 0.8–3.4), with a median response duration of 25.7 months (range: 8.2–45.1) (Fig. 1A). Tumour shrinkage was observed in all patients (Fig. 1B).

Table 3 Tumor response to treatment

Fig. 1

Responses to concurrent chemoradiotherapy plus toripalimab. A Swimmer plots of time to tumor response (months) of individual subjects with locally advanced cervical cancer as independently assessed by two experienced radiation oncologists according to RECIST version 1.1. Each swim lane represents one subject in the study population. B Waterfall plots of the best percentage changes from baseline for the sum of target lesion diameters are shown for individual subjects. Each bar represents one subject in the study population. Dashed horizontal lines indicate a 30% reduction in target lesion size. Percentage changes greater than 100% are truncated at 100%. CR, complete response; PD, progressive disease; PR, partial response

At the data cut-off date (June 30, 2024), the median follow-up duration was 27.3 months (range: 22.0–46.1). Recurrence was observed in 4 patients (13.3%), including parametrial recurrence in a patient with a best response of PR and distant metastasis in 3 patients who achieved CR. Time to recurrence ranged from 9.2 to 38.8 months (Supplementary Table 5). The median PFS was not reached. Kaplan-Meier estimates indicated a 2-year PFS of 90% (95% CI: 72.2–96.7) and a 2-year local recurrence-free survival (LRFS) of 96.7% (95% CI: 78.6–99.5). No patient deaths were reported during the trial period (Fig. 2).

Fig. 2

Kaplan-Meier estimates of progression-free survival A and local recurrence free survival B of the study population

Discussion

In this trial, all 30 patients responded to a combination regimen that included toripalimab and CCRT. The 2-year PFS rate was 90%; median PFS and OS were not reached. Grade 3 or higher TRAEs were observed in 80% (24/30) of patients. Nevertheless, toripalimab might intensify late rAEs and potentially lead to serious irAEs. These encourage further development of toripalimab combined with CCRT for LACC, but the increase in TRAEs warrants a cautious evaluation of the benefit-risk balance before advancing its clinical use.

Combination of ICIs with CCRT have been examined by large-scale randomized controlled trials for a variety of locally advanced malignancies, however, the majority of these trials failed to demonstrate survival benefits [20, 21]. However, there is conflicting evidence for ICIs plus CCRT in cervical cancer. The CALLA trial reported a negative result for a 2-year PFS benefit, while the KEYNOTE-A18 study was positive for the same endpoint [8, 9]. Speculation for the discrepancy has focused on potential differences in efficacy between PD-1 and PD-L1 inhibitors as well as enrolment of a population at higher risk in KEYNOTE-A18 [22]. Though several pivotal trials have demonstrated remarkable antitumor activities of immunochemotherapy for recurrent or metastatic cervical cancer, data are scant on immunochemotherapy for LACC [23, 24] (Supplementary Table 6). Consistent with the NiCOL trial, in our study, first-line toripalimab plus CCRT achieved an ORR of 100% and a CR rate of 93%, with a median duration of response of 13.9 months [25]. The 2-year PFS in this trial was 90%, compared to 76.0% (1-year PFS) in CALLA trial, and 68% (2-year PFS) in KEYNOTE-A18 study. In CALLA trial, durvalumab plus chemoradiotherapy failed to demonstrate a significant PFS benefit versus CCRT alone [9].Since the initiation of this trial, the efficacy of toripalimab plus CCRT for LACC has been reported by two trials [26, 27]. A phase I/II trial of toripalimab combined with CCRT by Ou et al. reported 2-year PFS of 81.8% and 2-year OS of 90.9% [26]. A phase II trial of consolidation chemotherapy and toripalimab maintenance by Chen et al. reported 2-year PFS of 88.5% and OS of 95.2% [27]. Survival outcomes in our trial seemed to outperform those in these 2 trials (Supplementary Table 6). The absence of a control arm in this single-arm study limits direct comparisons and precludes definitive conclusions about the superiority of our regimen.

The incidence of grade 3 or higher TRAEs in this trial (100% and 80%, respectively) was higher than that reported by Ou et al. (50%) [26] and Chen et a. (20.7%) [27]. The profile of grade 3 or higher TRAEs is also similar, and predominantly involved hematologic and gastrointestinal toxicities. However, chemotherapy-induced toxicity predominated among the most common TRAEs and grade 3 or higher TRAEs. The rate of grade 3 or higher irAEs in this trial was 3%, whereas whereas only grade 1–2 irAEs were reported by the Ou and Chen studies. To investigate potential causes for these differences, we compared the treatment modalities across the three studies, and found higher number and frequency of immunotherapy cycles (four cycles, every two weeks) during CCRT in this trial (two or three cycles, every three weeks) [26, 27]. These findings suggest that the frequency of ICIs administration may be a critical factor in mitigating irAEs and should not be less frequent than once every three weeks. However, further investigation is warranted to confirm this speculation.

The CHOICE-01 study reported higher rate of fatal TRAEs (5.5% vs. 2.6%), serious AEs (44.8% vs. 35.3%) and TRAEs leading to treatment discontinuation (14.3% vs. 3.2%) in the toripalimab arm [13]. Furthermore, the incidence of grade 3 or higher irAEs was higher in the immunochemotherapy arm in the JUPITER-06 trial (7% vs. 1.6%), the CHOICE-01 study (15.6% vs. 3.2%), as well as the JUPITER-02 study (9.6% vs. 1.4%) [11,12,13]. Similarly, two pivotal trials in patients with LACC, the CALLA and KEYNOTE-A18 studies, reported a higher rate of irAEs in the ICI arm (38% vs. 24% in the CALLA study, and 32% vs. 10% in the KEYNOTE-A18 study) [8, 9]. Also, one of the 30 patients in this trial developed immune-related HLH. These findings suggest that toripalimab possibly increased the rate of immune- and chemotherapy-related adverse events, however, any definite conclusions cannot be drawn based on the available data. The rate of late rAEs in this trial was 66.7% was higher than that reported in the CALLA study (10%). In CALLA study, the most common all grade late rAEs reported in durvalumab group were rectal hemorrhage and gastroenteritis, which were consistent with our findings [9]. To rule out the possibility that these late rAEs were due to excessive radiation doses to organs at OARs, we calculated the 2-cc doses to the rectum, sigmoid colon, and bladder. The results showed dose limitation was exceeded in 7 patients (23.3%) for the rectum and in 1 patient (3.3%) for the bladder. Among these 7 patients, only 4 (13.3%) experienced rectal hemorrhage, while the patient with excessive bladder exposure did not develop hematuria. It was not possible to clearly identify the cause for increased radiation toxicity, although the DVH data showed no unacceptable excessive dose to the OARs (Supplementary Table 7). Instead, combining ICIs with CCRT might increase the risk of TRAEs and late rAEs potentially leading to severe irAEs. However, further studies are needed to confirm this conclusion.

This trial has several limitations. This trial was conducted at a single institution with a limited number of patients, lacked a comparator arm, and had a relatively short follow-up period. Additionally, patients were not selected based on positive PD-L1 expression. It should be noted that​ while this study did not analyze prognostic biomarkers, ​we have archived​ paraffin-embedded tumor tissue specimens from all patients. These valuable resources will enable future in-depth exploration through multi-omics approaches. Furthermore, subsequent analyses will incorporate stratification based on PD-L1 expression and HPV subtypes. Failure to collect patient-reported outcome (PRO) data represents a limitation of this study, indicating that future trials should therefore integrate quality of life assessments via EORTC QLQ-CX24, to comprehensively evaluate treatment value. Despite these limitations, our findings encourage further investigation of toripalimab in combination with CCRT as a first-line treatment in LACC patients.

In conclusion, first-line toripalimab plus concurrent chemoradiotherapy demonstrated promising antitumor activity in patients with LACC. However, high rate of irAEs, fatal HLH and late rAEs necessitate careful consideration. Despite remarkable efficacy, the rate of grade ≥ 3 TRAEs was as high as 80%, including a fatal irAE, suggesting that stringent patient selection, optimization of treatment cycles, and the incorporation of a control arm in future studies are imperative to clearly define the risk-benefit ratio.