Association of chemotherapy dose intensity and age with outcomes in patients with Ewing's family sarcoma

2.1 Patients

Patients were a subset of adult patients from the Australian Comprehensive Cancer Outcomes and Research Database (ACCORD), and pediatric patients identified from the medical records of one pediatric hospital. ACCORD is a national electronic resource capturing clinical data and treatment outcomes of patients with bone and soft tissue sarcoma at major centers in Australia. Eligible patients for this study analysis were aged 10 years and above, diagnosed with EFS between January 2010 and September 2020, and had completed first-line treatment (chemotherapy, radiotherapy, and/ or surgery) by September 2020. All eligible patients were treated at an Australian specialty sarcoma center (one pediatric, three adult centers). This study was approved by the South Eastern Sydney Local Health District Human Research Ethics Committee and supported by the Australian and New Zealand Sarcoma Association.

2.2 Data collection

Clinical characteristics, treatment modalities, and survival data were collected for all patients from the ACCORD database, and/or individual medical records review. Treatment information included the duration of chemotherapy cycle (14- or 21-day schedule), received total dose of individual drugs adjusted for body surface area (including dose-escalation), and reasons for dose reduction, treatment delay, or discontinuation. A typical VDC/IE regimen comprised seven cycles each of VDC and IE: vincristine 1.5 mg/m², doxorubicin 37.5 mg/m² (2 days), cyclophosphamide 1200 mg/m², alternating with ifosfamide 1800 mg/m² (5 days), etoposide 100 mg/m² (5 days) with pegfilgrastim support.²² A full course of doxorubicin was considered as 375 mg/m² (usually five cycles) due to the risk of cardiac toxicity, and omission for the final two cycles of VDC was not recorded as a dose reduction if the cumulative maximum dose had been received.

2.3 Statistical analysis

Descriptive statistics were used to summarize patients’ demographics, clinical characteristics, and patterns of treatment, both for the entire cohort of EFS patients and the subgroup who received VDC/IE.

RDI was calculated using a previously described method, as the ratio of proportional dose ($$\delta $$) to proportional time delay ($$\tau $$), $${\mathrm{RDI}}\ \ ( \% ) = \ \frac{\delta }{\tau } \times 100$$.¹⁴ Proportional dose was equal to the received chemotherapy dose divided by the anticipated chemotherapy dose (100% of protocol), $$\delta \ = \ \frac{{{\mathrm{received\ dose}}\ ( \% )}}{{{\mathrm{anticipated\ dose}}\ ( \% )}}$$. Proportional time delay was equal to the number of days between the first and last day of treatment, divided by the planned time as per the published treatment schedule or clinical trial protocol (either 14 or 21 days), $$\tau \ = \ \frac{{{\mathrm{actual\ duration\ of\ chemotherapy\ (days)}}}}{{{\mathrm{anticipated\ duration\ of\ chemotherapy\ (days)}}}}$$.²² For those who underwent surgery, there was an allowance for 18 days hematological recovery before surgery, and 14 days surgical recovery prior to resuming chemotherapy. A patient who received treatment without dose reduction or delay would have $$\delta $$ = 1, $$\tau $$ = 1, and RDI = 100%. Acceptable RDI of chemotherapy was defined as ≥85% based on previous studies in solid tumors demonstrating a statistically significant benefit above this threshold.²³

All statistical analyses were performed using R Statistical Software Version 4.2.2 (R Foundation for Statistical Computing). PFS and OS for those receiving VDC/IE were described using the Kaplan−Meier method. Univariable and multivariable Cox proportional hazards regression models were used to compare outcomes by RDI, age, gender, Ewing's type, primary site, and stage. Clinical predictors of achieving an acceptable RDI were explored using univariable and multivariable logistic regression.

3.1 Patient characteristics

A total of 155 consecutive patients with histologically confirmed EFS were identified from ACCORD (n = 145, 94%) and pediatric medical records (n = 10, 6%) between 2010 and 2020. Nine patients were excluded due to incomplete baseline record (transfer of care to another center with no ongoing follow-up), leaving 146 patients eligible. Of the 146 patients, 135 (93%) received chemotherapy as part of primary treatment; VDC/IE (n = 104, 77%), VIDE with or without autologous stem cell transplantation (n = 30, 22%), and oral etoposide (n = 1). All patients over 60 years of age (n = 8) did not receive chemotherapy as primary treatment, and a further three patients received surgery with or without radiotherapy. Of 104 patients who received VDC/IE, 76 (73%) had complete chemotherapy data for RDI analysis, the remainder 28 (27%) were excluded due to insufficient records (n = 21), chemotherapy delivered at an external facility (n = 5), or previous treatment for non-EFS cancer (n = 2).

The baseline characteristics of all patients with EFS (146) and the subgroup who received VDC/IE and had complete record for RDI analysis (n = 76) were similar, as shown in Table 1. For the VDC/IE cohort, the majority were male (59%), with a median age of 25.4 years (range 10.3–59). The proportion with skeletal and extra-skeletal histology were 43% and 57%, respectively. Over half had localized disease at diagnosis (stage II, n = 28 [37%]; stage III, n = 21 [28%]).

3.2 VDC/IE cohort: Chemotherapy completion rate, dose reductions, and interruptions

Of the 76 patients who received VDC/IE, 54 (71%) patients completed scheduled chemotherapy without dose reduction. VDC/IE was delivered on a 14-day cycle in 24 patients (32%) and a 21-day cycle in 49 patients (65%), with three patients (4%) switching from a 14- to 21-day cycle. Fifteen patients (20%) at one site underwent chemotherapy dose-escalation from a starting dose of 100% of protocol due to good tolerance, as part of institutional guidelines, with 10 of these 15 (67%) having doxorubicin and cyclophosphamide dose escalated (range 110%–130% of protocol dose) and 14 (93%) dose escalated ifosfamide and etoposide (range 110%–120%).

Table 2 shows number of instances of dose reduction, cycle delay or omission, and treatment cessation, by age group. Twenty (26%) had a cumulative dose reduction of 1%–20%, two (3%) required a dose reduction by 21%–50%, and none required a dose reduction greater than 50%. For those whose chemotherapy was dose-escalated, they were only considered to have a dose reduction if their cumulative dose fell below 100% of the original anticipated dose.

There were 43 instances of drug dose reduction in 22 patients (29%). The most frequently reduced agent was doxorubicin (49% of cycles), followed by cyclophosphamide (35%), ifosfamide (35%), etoposide (33%), and vincristine (23%). Doxorubicin was more commonly reduced in the 40- 59-year-age group (43%), compared to 20–39 years (38%) and 10–19 years (19%). The reasons for dose reduction are detailed in Table 3. Myelosuppression was the most common reason for dose reduction, with at least 40% of total reductions due to febrile neutropenia.

There were 95 occasions of cycle delay greater than 7 days, among 49 patients (65%). The most common reason for delay was myelosuppression (neutropenia in 36%, thrombocytopenia in 31%, and anemia in 21%). Other documented reasons included patient preference (10%), mental health problems (17%), other infection (39%), and unrelated surgery (17%).

Treatment was stopped early in 19 patients (25%), which included six patients (32%) with disease progression. Six patients were assessed as ceasing treatment due to toxicity (32%), three due to infection (16%), and two due to patient preference (11%). Two patient deaths occurred, both over 20 years of age, one due to disease progression and one unknown. Cycle omission was rare, with only eight cycles omitted among three patients due to neutropenia, nephrotoxicity, fatigue, and patient preference.

3.3 RDI and survival outcomes

Fifty-seven percent of patients achieved an acceptable RDI of ≥ 85% (67% aged 10–19, 61% aged 20–39, and 29% aged 40–59). Median RDI was 87% (98% aged 10–19, 89% aged 20–39, and 80% aged 40–59). Compared to those aged 10–19 years, the odds ratio of an acceptable RDI for patients aged 20–39 was 0.79 (95% CI 0.24−2.46, p = 0.70) and for patients aged 40–59 was 0.20 (95% CI 0.04−0.86, p = 0.04).

The median follow-up for patients receiving VDC/IE was 37.3 months. For all patients, the median PFS was 22.4 months (2-year PFS 45%) and the median OS was 113 months (2-year OS 72%). RDI ≥85% conferred a significantly higher PFS than RDI < 85% (2-year PFS 56% vs. 30%; adjusted HR 0.39, 95% CI 0.18−0.82; p = 0.01; Figure 1A). OS was also higher in patients achieving RDI ≥85% (2-year OS 88% vs. 49%; adjusted HR .25, 95% CI 0.10−0.63; p = 0.004; Figure 1B). Regarding the effect of age, PFS and OS varied according to age group in univariable analyses (log-rank p = 0.06 and 0.05, respectively), but not when limited to those who received RDI ≥85% (log-rank p = 0.9 and 0.6, respectively), thus confirming the prognostic importance of RDI.

RDI was a prognostic factor for OS in both univariable and multivariable models, after adjustment for age, gender, Ewing's subtype, primary site, and stage (Table 4). The only other prognostic factor for OS on multivariate analysis was stage at diagnosis, with metastatic disease conferring a statistically significant inferior outcome.