Longer versus Shorter Schedules of Vincristine, Irinotecan, and Temozolomide (VIT) for Relapsed or Refractory Ewing Sarcoma: A Randomized Controlled Phase 2 Trial

Ewing sarcoma is a small round blue cell malignant cell tumor and is the second most common primary bone cancer in children, adolescents, and young adults (1, 2). Multiagent chemotherapy is recommended (3, 4), but outcomes are considerably worse for patients with post-treatment disease recurrence or resistance to initial therapy (5, 6). Less than 20% of patients with relapsed or refractory disease show long-term survival (7).

Irinotecan has long been used for treatment of pediatric sarcomas. Preclinical experiments have shown enhanced antitumor activity for longer-duration, lower-dose administration of irinotecan compared with single large doses, an observation consistent with its S-phase specific mechanism of action (8). The first pediatric phase I clinical trial of irinotecan administered a protracted d × 5×2 schedule (daily from days 1 to 5 for two consecutive weeks; ref. 9) as opposed to the single-dose therapy given every 3 weeks for colon cancer. A number of subsequent trials have explored various schedules of vincristine, irinotecan, and temozolomide (VIT) administration in patients with sarcoma. The schedules were as follows: One large single dose (10, 11), 5d consecutively (d × 5; ref. 12), d × 5×2 (13, 14), oral dose for 5d every 3 weeks (15, 16). Each schedule was safe but diverse antitumor activities were reported (17). A single study compared the efficacy of the d × 5 and d × 5×2 schedules in patients with rhabdomyosarcoma (RMS; ref. 18) and found no differences in objective response rate (ORR) or overall grade 3/4 adverse events (AE). The result is that the more convenient d × 5 schedule has been widely adopted. However, whether the results regarding irinotecan schedule in RMS can be extrapolated to Ewing sarcoma is unclear. Conflicting results have been reported in retrospective or prospective single-arm clinical trials for patients with Ewing sarcoma (17, 19).

Because the administration schedule of irinotecan appeared to be an important in vivo determinant of antitumor activity, we conducted a randomized phase II trial to directly compare the d × 5 and d × 5×2 schedules in relapsed or refractory patients with Ewing sarcoma.

A phase 2, randomized controlled trial was conducted at two hospitals in China. Patients were eligible for this trial if age <30 years at the time of diagnosis and Ewing sarcoma was confirmed by FISH or next-generation sequencing (NGS) detection of EWSR1 gene rearrangement; relapse or disease progression occurred during first-line chemotherapy (first-line chemotherapy included doxorubicin, vincristine, cyclophosphamide, ifosfamide, and etoposide, known as VDC-IE or VIDE) or less than 6 months after ceasing any prior systemic therapy; had not received irinotecan before enrollment; and obtained an Eastern Cooperative Oncology Group performance status of 0, 1, or 2 and a life expectancy of at least 3 months was established. Patients were required to have measurable disease by CT or MRI by RECIST 1.1. Adequate organ function was confirmed, as defined by hemoglobin >8 g/dL (transfusion allowed); absolute neutrophil count >1,500/μL; platelet count >80,000/μL; serum creatinine <1.5 × normal for age or creatinine clearance/radioisotope; serum bilirubin < normal for age; serum alanine transaminase <2.5 × normal for age; albumin ≥25 g/L. The following are the exclusion criteria: metastases to the brain and spinal cord; tumor embolus in large vein; persistent clinically significant toxicities caused by previous cancer therapy; pregnant/lactating.

The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guideline. Patients provided written informed consent and the trial was approved by national and institutional research ethics committees (registered with ClinicalTrials.gov, number NCT03359005).

Eligible patients were randomly assigned (1:1) to d × 5 group or d × 5×2 group by random number table method. No further stratification was considered in our study. Neither patients nor investigators were masked to treatment allocation.

d × 5 patients received intravenous 50 mg/m²/d irinotecan on D_1–5 and 1.4 mg/m²/d (maximum 2 mg) vincristine on D₁. d × 5×2 patients received 20 mg/m²/d irinotecan on D_1–5,8–12 and 1.4 mg/m²/d (maximum 2 mg) vincristine on D₁ and D₈. Temozolomide (100 mg/m²/d) was administered intravenously on D_1–5 one hour before irinotecan. The above treatments were repeated every 3 weeks until disease progression or unacceptable toxic effects, up to a maximum of eight cycles. Irinotecan and temozolomide were obtained from Jiangsu Hengrui Medicine, Lianyungang, China. The administration of 10 mg/kg/d twice daily cefixime (maximum 400 mg/d) 3 days before irinotecan and continuing through the cycle is specified by protocol guidelines to reduce abdominal pain/diarrhea. Atropine and intestinal alkalization before irinotecan was also considered to avoid acute diarrhea and loperamide given for late diarrhea.

Pretreatment evaluation included a medical history, physical examination, complete blood cell count, serum biochemical tests, and chest, abdominal, and pelvic CT scans or PET/CT scans. Clinical examination and laboratory tests were conducted before and during every treatment cycle. Radiographic imaging assessments of disease status were performed every 6 weeks according to RECIST 1.1. Patients were followed-up every 12 weeks after treatment progression to assess survival with follow-up survival monitoring still ongoing.

The primary endpoint was the objective response rate at 12 weeks (ORR_12w). ORR_12w was defined as the proportion of patients who achieved complete response (CR) or partial response (PR) at 12 weeks. CR, PR, stable disease (SD), or progressive disease (PD) was assessed according to RECIST 1.1. The secondary endpoints were progression-free survival (PFS), overall survival (OS), and safety. PFS was defined as the time from randomization until documented progression, protocol violation due to investigator's discretion, serious adverse effect, or death from any cause. OS was defined as the time from randomization until death from any cause. No central review of response was completed. Efficacy and safety analyses were performed in the intention-to-treat (ITT) population.

AEs were assessed and graded according to Common Terminology Criteria for AEs version 4.03. Dose modifications for AEs were done according to the protocol.

The study was powered to detect a 30% improvement in the response rate from 20% of patients receiving d × 5 schedule to 50% receiving d × 5×2 (α = 0.1, 1−β = 0.9, one-sided test favors d × 5 schedule because the only difference of clinical importance was an improved response with the more prolonged but inconvenient schedule of d × 5×2). A sample size of 30 patients per group (60 randomly assigned patients) was required to detect a significant improvement in ORR_12w. The χ² analysis was used to compare the differences in baseline patient characteristics, ORR_12w, and safety data between schedules. PFS and OS were estimated using the Kaplan–Meier method, and the curves were compared using the log rank test in IBM SPSS 22.0. Some categorical variables were compared between schedules by PFS and OS using Cox univariate analysis. A P value of <0.05 was considered statistically significant.

This work was funded by Project (RDX2019–09) of Peking University People's Hospital Scientific Research Development Funds.

Data are available on request from the authors and the study protocol is shown in the Supplementary Materials.

Between May 21, 2020 and August 8, 2022, 24 patients were randomly assigned to the d × 5 schedule and 22 to the d × 5×2 schedule (Fig. 1).

Demographic and baseline characteristics of 24 patients randomly assigned to the d × 5 and 22 to the d × 5×2 schedule (Fig. 1) are shown in Table 1. EWSR1 gene rearrangement was confirmed by FISH or NGS and central pathology review was not required. Median follow-up was 10.7 months (IQR, 9.7–13.9) for d × 5 and 8.3 months (IQR, 4.4–15.3) for d × 5×2. No significant difference was found between the two groups in proportions for baseline patient characteristics (P > 0.05), including age, gender, primary tumor location, site of target and non-target lesions, time from first diagnosis to enrollment, and lines of previous chemotherapy. All 46 patients were included in ITT analysis and safety analyses.

Table 2 shows the response at 12 weeks and survival data. ORR_12w was lower in the d × 5 than in the d × 5×2 group [5/24 (20.8%) vs. 12/22 (54.5%); P = 0.019]. One CR, 4 PR, 6 SD, and 13 PD were recorded at week 12 among the d × 5 group and 1 CR, 11 PR, 3 SD, and 6 PD for the d × 5×2. One patient received only one d × 5×2 cycle and was unable to access further cycles due to the COVID-19 isolation policy. He could not be included in the efficacy analysis and received oral anlotinib as a salvage therapy during his home quarantine. PD was recorded for this patient in ITT analyses and he was not included in per-protocol analyses.

There was no significant difference in PFS between groups [median PFS 2.3 months; 95% confidence interval (CI), 0.0–4.7; in d × 5 group vs. 4.3 months (2.7–6.0) in d × 5×2 group; hazard ratio (HR), 0.956; 95% CI, 0.84–1.09; P = 0.434; Fig. 2].

There was also no significant difference in OS (median OS 14.8 months; 95% CI, 12.0–17.6 vs. 12.8 months; 95% CI, 6.9–18.7; HR, 0.957; 95% CI, 0.81–1.12; P = 0.594). Five patients (2 from d × 5 to 3 from d × 5×2) failed to continue allocated treatment after the first 12 weeks due to domestic isolation policy of COVID-19. One d × 5 patient with SD dropped out after two cycles due to inability to travel from his hometown to our hospital and showed subsequent disease progression. The other d × 5 patient continued with SD for five cycles before his isolation and received radiotherapy for lung lesions instead of chemotherapy and had no evidence of disease (NED) at final follow-up. Two d × 5×2 patients had PR after four cycles before being quarantined at home and receiving radiotherapy on release. One progressed after 3 months and the other one remained with NED at final follow-up. The last d × 5×2 patient had PR after first three cycles and switched to anlotinib after home quarantine to avoid disease progression. Three patients received local therapy before all eight cycles. One d × 5 patient with SD received surgical resection, and one d × 5 patient with PR received radiotherapy and one d × 5×2 patient received radiotherapy after PR for four cycles.

Grade 3/4 AEs were summarized in Table 3. No unexpected toxicities or deaths during treatment were observed. d × 5 patients reported more grade 3 and 4 AEs, including 23/97 (23.7%) cases of diarrhea/abdominal pain versus 8/95 (8.4%) for d × 5×2 (P = 0.005) and vomiting/nausea 6 cases (6.2%) versus 1 (1.15%, P = 0.035). Other common grade 3 and 4 AEs (>5%) were similar between the two schedules (P > 0.05), including fatigue with 24 (24.7%) versus 14 (14.7%) cases, leukopenia 6 (6.2%) versus 8 (8.4%) cases, neutropenia 5 (5.2%) versus 8 (8.4%) cases, and anemia 5 (5.2%) versus 7 (7.4%) cases.

Controversial results have been recorded for the antitumor activity of irinotecan administered using different schedules in various tumor xenograft models. Prolonged exposures prompted improved activity in RMS and colon cancer; however, no difference between the protracted and shorter schedules was found in ovarian cancer, soft tissue sarcoma, and brain tumors (8, 20–22). In contrast with the single large dose of irinotecan used in colorectal cancer (FOLFIRI regimen; ref. 23), several regimens with different durations based on irinotecan have been evaluated in patients with sarcoma to improve activity, reduce toxicity, and increase the convenience of therapy. The following schedules have been explored: One large single dose of 600 mg/m² every 3 weeks (10, 11); 125 mg/m²/dose weekly 4x every 6 weeks (24); 180 mg/m²/d D_1–3 every 4 weeks or 50 mg/m²/d d × 5 every 3 weeks (12, 25); 20 mg/m²/d d × 5×2 every 3 weeks (the original protracted schedule; refs. 13, 14), 15 mg/m²/d d × 5×2 every 3 weeks when combined with anlotinib (26), and 90 mg/m²/d D_1–5 every 3 weeks [the most recent oral irinotecan (VOIT) regimen; refs. 15, 16]. Regimens and efficacy results were summarized in Table 4. The ORR of protracted schedule seems somewhat higher than shorter schedule (55.0%–63.2% vs. 34%–54.5%) with not much difference. As all the seven published studies were retrospective, there was no prospective plan for imaging. It may not be very rigorous and scientific if using integrated data from retrospective studies to analyze ORR of a certain schedule. Our study was the first RCT to compare protracted and shorter schedules of irinotecan in patients with Ewing sarcoma.

Only one study—a phase II study conducted by the Children's Oncology Group in patients with RMS (18)—compared the efficacy of different schedules. No difference in ORR and overall grade 3/4 AEs was found between the two schedules, either on a d × 5 or d × 5×2 schedule every 3 weeks. These findings resulted in the recommendation of a shorter and more convenient schedule. However, whether the results regarding irinotecan schedule in RMS can be extrapolated to Ewing sarcoma is unclear.

Recently, Slotkin and colleagues (19) from MSKCC retrospectively reviewed VIT studies in Ewing sarcoma and recommended a longer administration of irinotecan schedule for its better efficacy (ORR of 53% in d × 5×2 vs. 29% in d × 5), consistent with the current findings (ORR_12w of 54.5% vs. 20.8%). The most recent global trial of rEECur also used the d × 5 schedule. The second interim analysis showed that the combination of irinotecan and temozolomide was inferior to two of the other regimens (high-dose of ifosfamide and cyclophosphamide/topotecan). The ORR of d × 5 irinotecan/temozolomide regimen was only 20%, according to the report of the 2020 ASCO meeting, lower than the 53% reported for the d × 5×2 schedule (17, 19). A phase 1 trial of regorafenib, vincristine, and irinotecan is ongoing with high expectations (NCT 02085148) and regorafenib is given sequentially, rather than concomitantly, due to safety issues on a shorter d × 5 rather than d × 5×2 schedule. We have also used anlotinib (VEGFR-TKI, tyrosine kinase inhibitor) and irinotecan on a d × 5×2 schedule (26). The final results from the last two trials may give us more information about ideal irinotecan schedules.

We acknowledge some limitations to the current study. First, vincristine was given only on D₁ in d × 5 schedule and on D_1,8 in d × 5×2 schedule. Vincristine is an active chemotherapeutic agent for Ewing sarcoma and may act synergistically with irinotecan according to preclinical studies (27). The greater ORRs in d × 5×2 schedule may be due to more frequent dosing with vincristine rather than the protracted irinotecan strategy. However, it is generally agreed that vincristine is an adjuvant drug and not the central component of VIT. In addition, the current patients were relapsed or refractory to previous first-line chemotherapy that included vincristine. Second, the PFS and OS secondary endpoints may be affected by many social factors in addition to treatment. Confounding factors included drop-out due to the COVID-19 domestic isolation policy resulting in premature adoption of local treatment. The improved ORR was a robust result but insufficient data were available to distinguish differences in PFS between the two schedules. Treatments following VIT were varied and included chemotherapy, local treatment, target-drug and other clinical trials, making OS difficult to assess with consistency. Small-molecule TKIs targeting VEGFR-TKIs have shown good efficacy in Ewing sarcoma. Examples are anlotinib (26, 28), cabozantinib (29, 30), regorafenib (31), and apatinib (32). Variations in follow-up therapy may deeply impact the final outcomes.

As an open-label RCT, patients and investigators were not blinded and this may lead to some bias in response assessment. However, we have two senior radiologists to review images independently. They were not provided with grouping information or treatment details. We believed that bias brought by the open-label nature of our study would not significantly impact the final outcome.

The number of patients finally enrolled in our study is not enough. As designed in the protocol, at least 30 patients in each group were needed to get a statistically meaningful result of ORR. As for PFS, we would need 118 patients in each arm to find a statistical improvement of 6-month PFS from 8.9% of d × 5 to 20.6% of d × 5×2 according to the result of this premature study (α = 0.1, 1 – β = 0.9, one-sided test). However, the isolation policy of COVID-19 significantly restricted the travel of patients from their hometown to our hospital. Given the relatively slow speed of recruiting, the sponsor stopped following support to our study.

Cefixime, atropine, intestinal alkalization, and loperamide were used to avoid serious diarrhea (see methods section), based on AEs observed in previous irinotecan studies in patients with sarcoma (33–35). Grade 3/4 diarrhea/abdominal pain remained a big problem with a similar risk to the records previously, especially in the first VIT cycle (36). Notably, d × 5×2 patients experienced much less diarrhea/abdominal pain than d × 5 patients and the latter group also had worse vomiting/nausea. The higher single-dose of irinotecan may be blamed for vomiting/nausea. The lower risk of fatigue in d × 5×2 group may be attributed to lower incidence of diarrhea/abdominal pain or vomiting/nausea at the same time.

It is worth noting that doctors should find a balance between antitumor efficacy and quality of life (QoL) before prescription in the setting of patients with refractory or relapsed disease. The inconvenience of protracted regimen kept patients in clinic 2 out of every 3 weeks, and this might impact their social needs. Although less toxicities in protracted regimen may protect them from some AEs such as diarrhea or vomiting, the longer period of in-clinic setting may affect their QoL. Unfortunately, QoL assessment was not involved in our study according to protocol and now we feel it difficult to give a precise answer to these QoL questions. Recently, liposomal irinotecan has been tested in metastatic pancreatic cancer (37, 38) and biliary tract cancer (39) with promising results. This more convenient dosage form of irinotecan may be explored in patients with Ewing sarcoma in the future.

VIT is an effective schedule as a second-line systematic therapy for recurrent and refractory patients with Ewing sarcoma. The current study failed to find a significant survival benefit but the protracted d × 5×2 VIT schedule produced a superior response rate and favorable tolerability compared with the shorter d × 5 VIT schedule. A protracted schedule of regular irinotecan or new dosage of long-effect irinotecan should be considered in the future.

No disclosures were reported.

J. Xu: Conceptualization, writing–original draft. L. Xie: Conceptualization, methodology. X. Sun: Resources, data curation. K. Liu: Data curation, software. X. Liang: Data curation, methodology. Z. Cai: Data curation. X. Tang: Supervision, methodology, project administration. W. Guo: Methodology, project administration.

This work was funded by Project (RDX2019–09) supported by Peking University People's Hospital Scientific Research Development Funds. The authors would like to express their gratitude to EditSprings for the expert linguistic services provided.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.