Introduction

Ewing sarcoma is a malignant tumor that can occur in either the bone or soft tissue and is the third most common primary malignant bone tumor after osteosarcoma and chondrosarcoma [1]. The typical age of onset is teenage years, and about 60–80% of Ewing sarcoma cases occur in children, adolescents, and young adults under 20 years of age [2]. Ewing sarcoma is characterized by the proliferation of small round cells and specific fusion genes, including EWSR1-FLI1 and EWSR1-ERG [3].

Localized Ewing sarcoma can be treated with a combination of chemotherapy and surgery. Radiation therapy can be used to achieve durable local control in unresectable Ewing tumors. Notably, some reports have shown local control differences between radiation and surgery [4, 5]; decisions regarding the optimal local control method should be based on patient characteristics and prognosis-related factors.

Currently, multi-agent regimens for Ewing sarcoma, primarily anthracyclines, alkylating agents, and topoisomerase inhibitors, are common; those posing a high risk of myelosuppression are also prevalent. The typical regimen for curative treatment is vincristine, doxorubicin, and cyclophosphamide, alternating with ifosfamide and etoposide (VDC/IE) [6]. VDC/IE was administered repeatedly at 3-week intervals according to the original protocol. However, in recent years, an enhanced regimen, repeated at 2-week intervals, has been developed based on the premise of primary prophylactic administration of granulocyte colony-stimulating factor (G-CSF) [7].

The use of G-CSF as a supportive care medication aims to prevent chemotherapy-induced and febrile neutropenia (FN). When used appropriately, G-CSF can improve chemotherapy outcomes, leading to increased benefits such as a reduced incidence of complications such as FN and associated death, improved quality of life (QOL), and prolonged survival. However, G-CSF also has negative effects, such as pain, the burden of hospital visits, and drug costs. Therefore, its use should be considered carefully if its benefits remain unclear.

The American Society of Clinical Oncology (ASCO), European Society for Medical Oncology (ESMO) 2010 guidelines, and National Comprehensive Cancer Network (NCCN) guidelines, while not specific to Ewing sarcoma, recommend primary prophylaxis when FN incidence exceeds 20%. Yet, definitive evidence for this threshold remains unestablished. The guidelines do not detail the benefits of intensified chemotherapy for primary prophylaxis with G-CSF in adult Ewing sarcoma patients [8,9,10,11,12].

In Japan, the primary prophylactic administration of G-CSF is not covered by insurance, except for some cancers, and the therapeutic administration of G-CSF for FN is the mainstream treatment. A total of 42 members of The Working Group for the Revision of Guidelines for the Proper Use of G-CSF evaluated each question (Clinical Practice Guidelines for the Use of G-CSF, 2022, Japan Society of Clinical Oncology). Even for cancer types and regimens with scant evidence and undefined recommendations, we have prepared a reference to assist in evaluating the risk–benefit balance in each scenario.

This systematic review assessed the effectiveness and safety of primary prophylaxis with G-CSF, in addition to chemotherapy, in patients with Ewing sarcoma.

Patients and methods

Data searching and screening

A systematic review, adhering to the “Medical Information Network Distribution Service (Minds) Handbook for Clinical Practice Guideline Development 2014” [13] and “Minds Clinical Practice Guideline Development Guide 2017” [14, 15], was conducted utilizing PubMed, Ichushi web (Japan Medical Abstracts Society database), and the Cochran Library databases. The search, conducted in April 2020, utilized MeSH keywords including “Ewing sarcoma” and “G-CSF,” covering publications dated from 1990/01/01 to 2019/12/31 in English and Japanese (Supplementary Tables S1, S2). Relevant literature from other databases was incorporated based on the judgment of a systematic review team. According to titles and abstracts, an initial screening was independently conducted by two systematic review team reviewers (TH and MI) to identify ineligible reports (Fig. 1a, b). The second screening entailed a detailed evaluation of chosen articles against inclusion and exclusion criteria, entailing full-text reading, recording exclusion reasons, and omitting duplicates. Disagreements were resolved by consensus among the co-authors. Articles that met the selection criteria were further scrutinized for quality reporting data as described below.

Fig. 1
figure 1

Flow diagram of literature search. a Clinical question #1. b Clinical Question #2

Full size image

Selection criteria

The inclusion criteria encompassed randomized controlled trials (RCT,) non-RCTs, cohort, and case–control trials with adults diagnosed with Ewing sarcoma, requiring the treatment group to have undergone standard intensive induction therapy. Exclusion criteria consisted of guidelines, reviews, letters, abstract-only publications, laboratory studies, systematic reviews, meta-analyses, and grey literature.

Quality of evidence

Following the second screening, a systematic review team member independently reassessed the articles and extracted data using standardized data abstraction forms. The evidence from individual studies on critical outcomes within the clinical questions posed by the guideline creation was categorized based on study design and quality. Critical outcomes included duration of neutropenia or thrombocytopenia, infection-related mortality, disease progression/recurrence, overall survival (OS), and adverse events such as musculoskeletal pain. The authors determined the outcomes under the Population, Intervention, Comparator, and Outcome (PICO) frameworks for the benefits and harms of prophylactic G-CSF. Conflicts and questions were resolved by the team leader (M.E.). Evidence level was evaluated per outcome across studies by design, considering bias risk, inconsistency, imprecision, indirectness, and publication bias. Literature quality and evidence body were assessed via the GRADE approach and categorized into “strong,” “medium,” “weak,” or “very weak” [16].

PICO setting

Our study addressed these clinical questions (CQs): #1: “Does primary prophylaxis with G-CSF benefit chemotherapy for Ewing sarcoma?” #2: “Does G-CSF-based intensified chemotherapy improve Ewing sarcoma treatment outcomes?”.

Regarding CQ #1, PICO consisted of the following: (P) patients, which included patients with Ewing sarcoma receiving chemotherapy; (I) intervention, which included the use of G-CSF as primary prophylaxis; (C) comparison, which did not use G-CSF as primary prophylaxis; (O) outcomes, which included (i) OS; (ii) incidence of FN; (iii) mortality rate from infection; (iv) QOL; and (v) pain. As for CQ #2, PICO was composed of (P) patients, which include patients with Ewing sarcoma receiving chemotherapy, (I) intervention, which includes the use of intensified chemotherapy based on the premise of G-CSF as primary prophylaxis, (C) comparison, which is conventional and not intensified chemotherapy, and (O) outcomes, which include (i) OS, (ii) incidence of FN, (iii) mortality rate for infection, (iv) QOL, and (v) pain.

Results

In CQ #1, 25 articles in English and five in Japanese were listed using PubMed and Ichushi, respectively. After initial screening, nine out of 30 papers were selected based on titles and abstracts. Following this, a full-text review for eligibility narrowed the selection to one cohort study for further analysis. The working group deliberated on converting CQ into future research questions (FQ) because of insufficient data obtained. In this guideline shows, FQ is defined as those for which the systematic review could not be completed due to lack of evidence or other reasons, and for which evidence-based recommendations could not be provided among the foreground questions based on important clinical issues. All questions were designated as CQs at the beginning of the revision work, but CQs and FQs were classified according to the status of subsequent systematic reviews.

A total of 851 patients who underwent VIDE chemotherapy were included (Table 1) [17]. Safety data from 4746 courses of VIDE in 851 patients were collected and compared in terms of the interval between VIDE courses, dose deviations, supportive therapy (G-CSF), incidence of adverse reactions per course, sex, and age. The incidences of FN and infection were 60.8% and 54.7%, respectively, when G-CSF was administered. In contrast, when G-CSF was not administered, the incidence rates of FN and infection were 65.8% and 61.0%, respectively. No statistical analyses were performed. In the cohort of 851 patients, nine fatalities occurred, with five potentially linked to VIDE therapy (three from sepsis, two unspecified). The absence of data on OS, pain, and QOL precluded direct evaluation of CQ #1 outcomes from this review. Consequently, CQ #1 was reclassified as an FQ, restating it as “Primary prophylaxis with G-CSF’s efficacy in Ewing sarcoma remains uncertain outside pediatric cases, though it is commonly administered during curative treatment.”

Table 1 Studies selected for each clinical question
Full size table

CQ #2 was investigated using two databases, PubMed and Ichushi, which yielded 20 articles in English and five in Japanese. Following the initial screening based on title and abstract, six of the 25 papers were considered relevant and subjected to full-text evaluation. Only two studies met the eligibility criteria after the second screening, both of which were RCTs. However, one study compared a dose-intensification regimen that was not commonly used in clinical practice and was deemed inappropriate for inclusion in a systematic review [18]. Another study was a high-quality RCT that examined the use of G-CSF to increase treatment intensity and was considered the basis of the current standard of care for Ewing sarcoma [7]. Following a thorough literature review, the working group responsible for CQ #2 conducted a systematic review.

An RCT comparing every 3-week vs. every 2-week VDC/IE chemotherapy cycle enrolled 568 eligible patients. Filgrastim was used as primary prophylaxis in both groups. The study’s primary endpoint was event-free survival (EFS), and the HR for EFS was 0.74 (95% CI 0.54–0.99, p = 0.048), indicating a statistically significant increase in EFS with every 2-week treatment. Regarding OS outcomes, the hazard ratio (HR) for biweekly versus triweekly cycles was 0.69 (95% CI 0.47–1.0, p = 0.056), indicating a trend toward reduced mortality risk, although the difference was statistically insignificant. OS was not the primary endpoint, and the study was not designed to test its superiority. The findings of this study have established biweekly chemotherapy cycles as the standard in clinical practice. However, the evidence strength is constrained, being derived from a single RCT, resulting in a “weak” rating for OS. Notably, FN incidence appeared elevated in biweekly-treated patients; 7.3% and 6.2% in biweekly and triweekly respectively. Although no statistical validation was provided in the literature, we calculated the odds ratio, which was 1.19 (95% CI 0.99–1.43, p = 0.067 [Fisher’s exact test]). In assessing a single RCT, evidence strength for FN was rated as weak. Mortality included one infection-related death in the biweekly group, with none in the triweekly group. A larger number of events would be needed for a more accurate assessment.

Since this was a single RCT, the strength of the evidence was judged to be very weak based on mortality due to infection. Regarding QOL and pain, studies on this topic were not identified, making it unassessable.

Although no evidence-based assessment of patient values and preferences has been made, there may be variations in the perception of desirable and undesirable effects such as pain. An evidence-based cost and resource evaluation was unfeasible; therefore, we estimated G-CSF costs and assessed if the benefits justified the expenditures and resources involved. Although only one RCT was included in the systematic review, the benefits outweighed the harms of intensified chemotherapy based on primary prophylaxis with G-CSF in treating Ewing sarcoma.

Finally, a draft recommendation, “Weakly recommend intensified chemotherapy based on primary prophylaxis with G-CSF in treating Ewing sarcoma,” was presented to the recommendation decision meeting based on the systematic review report. After discussion and voting by a working group of 25 members (23 physicians, one nurse, and one pharmacist), a consensus was reached, with 24 of the 25 members agreeing to participate in the draft.

Discussion

Currently, multiple-drug regimens centered on anthracyclines, alkylating agents, and topoisomerase inhibitors are frequently used for the treatment of Ewing sarcoma, and treatments with a high risk of FN are widely administered. The typical regimen selected for curative treatment is VDC/IE, and the incidence of FN without primary G-CSF prophylaxis ranges from 49.0% to 72.5% [19, 20]. In clinical practice, G-CSF is widely administered as primary prophylaxis during VDC/IE therapy, considering the incidence of FN described above. Additionally, as explained above for CQ#2, the VDC/IE treatment interval was shortened from 3 to 2 weeks based on the assumption of primary prophylactic administration of G-CSF, which is the standard of care for localized Ewing sarcoma.

However, to date, ASCO guidelines, ESMO guidelines 2010, and NCCN guidelines have not explicitly addressed Ewing sarcoma, and primary prophylactic administration is recommended when the incidence of FN is 20% or higher [8,9,10]. Therefore, to our knowledge, this is the first systematic review to summarize the efficacy of G-CSF administration in Ewing sarcoma.

As a result, only one systematic review of the literature on the primary prophylactic administration of G-CSF is available. A cohort study of 851 VIDE-treated patients with Ewing sarcoma showed that FN and infection incidence rates with and without G-CSF were 60.8%, 54.7%, 65.8%, and 61.0%, respectively, with no significant differences.

Next, we examine the content of the extracted literature for CQ#2. While it comprised a single RCT, its high quality warranted a systematic review, thus defining CQ#2. The HRs for OS and EFS were 0.69 and 0.74, respectively, for every 2 weeks compared with every 3 weeks, showing a trend toward lower mortality for OS and a statistically significant increase in EFS.

Many guidelines worldwide recommend using G-CSF based on the assumption of a “20% of FN incidence rate” cutoff, which lacks scientific evidence. The working group for the revision of the guideline has been discussing the recommendation of using G-CSF and has decided to abandon the assumption of a “20% FN incidence rate” and to evaluate the evidence for each question scientifically. This was the first attempt to create guidelines based on scientific evidence. During the development process, there was a lack of evidence comparing individual cancer types and regimens with or without G-CSF, and there is room for improvement, including the setting of questions, in future revisions of the guidelines.

This systematic review had some limitations. First, only a few reports on these CQs are available. Second, more detailed information on QOL, musculoskeletal pain, and other symptoms must be provided. Sufficient evidence on QOL and bone pain in most cancer subtypes does not exist in the Clinical Practice Guidelines for the Use of G-CSF 2022. Therefore, this FQ needs to be addressed. Third, different regimens for Ewing sarcoma were used in clinical situations, even between CQ #1 and CQ #2.

In conclusion, this study examined primary prophylaxis with G-CSF in Ewing sarcoma. In our systematic review, we assessed patient outcomes with CQ #1 and CQ #2. The benefit of primary prophylactic G-CSF in Ewing sarcoma, apart from pediatric cases, remains uncertain, yet it is commonly used in curative therapies. “Intensified cancer drug therapy with G-CSF as primary prophylaxis in Ewing sarcoma receives a weak recommendation.”