Pegylated Liposomal Doxorubicin Combined with Ifosfamide for Treating Advanced or Metastatic Soft-tissue Sarcoma: A Prospective, Single-arm Phase II Study

Soft-tissue sarcoma (STS) represents a heterogeneous group of tumors that includes approximately 70 subtypes (1). Advanced and/or metastatic disease is commonly observed in patients with STS, and the prognosis of these patients remains poor, with a median overall survival (OS) of 14–17 months (2–4). Doxorubicin (brand name: Adriamycin; ADM) is frequently used for the treatment of metastatic STS; however, the efficacy of ADM alone remains unsatisfactory. Compared with ADM alone, the combination of ADM and ifosfamide (IFO) as the first-line treatment can significantly improve progression-free survival (PFS; 7.4 vs. 4.6 months). However, the combination treatment does not significantly improve OS. Furthermore, increased occurrence of toxicities (grades 3–4) caused by the combination therapy could affect its efficacy and limit its clinical use. Therefore, it is important to reduce the toxicity of this regimen to improve its efficacy.

Pegylated liposomal doxorubicin (PLD) is a liposomal formulation of ADM that demonstrates reduced uptake by the reticuloendothelial system and has an extended circulation time and enhanced concentration in the tumor. These features may make PLD more advantageous than conventional ADM in terms of safety issues, especially cardiotoxicity (5). A previous study comparing the antitumor activity and toxicity of PLD with those of ADM in patients with advanced or metastatic STS reported comparable activity and an improved toxicity profile for PLD (6). Considering the indispensable role of anthracycline in the treatment of STS and the dose-limiting cardiotoxicity of ADM, PLD can be a potential path to expand treatment options for STS and optimize patient survival. A previous phase I study evaluated the toxicity of combined PLD and IFO as a first-line treatment for patients with advanced or metastatic STS. Five different doses of the combination therapy were evaluated, and PLD at 30 mg/m² on day 1 and IFO at 3 g/m² on days 1–3 every 3 weeks were determined to be the optimal doses (7).

In a previous study, we retrospectively analyzed the efficacy and safety of PLD combined with IFO as the first-line treatment for patients with advanced or metastatic STS. In total, 25 patients received PLD at 30 mg/m² on day 1 plus IFO at 1.8 g/m² on days 1–5 every 21 days. Their objective response rate (ORR) and disease control rate (DCR) were 36% and 84%, respectively, and the median PFS was 7.3 [95% confidence interval (95% CI), 4.6–10.0] months. The observed grade 3/4 toxicities included leukopenia (20%), neutropenia (28%), anemia (4%), and vomiting (4%; ref. 8).

Sensitivity to chemotherapy widely varies among patients with STS, and potential predictors of clinical outcomes remain to be identified. Neutrophils can reflect the state of host inflammation, whereas lymphocytes are associated with the immune response against cancer. Certain pretreatment inflammatory indices, such as the neutrophil-to-lymphocyte ratio (NLR) and lymphocyte-to-monocyte ratio (LMR), have emerged as biomarkers exhibiting prognostic value (9, 10). Cheng and colleagues reported that NLR and LMR were independent prognostic factors for PFS and OS, respectively, in patients with synovial sarcoma (11). Recently, DNA homologous recombination deficiency (HRD)—known as BRCAness—has been observed in patients with STS, and it may affect the sensitivity of such patients to chemotherapeutic drugs. BRCAness is used to describe HRD tumors that lack BRCA1/2 germline mutations. A previous study conducted whole exon sequencing on 22 STS samples to identify the possible genomic and molecular features. The results of that study demonstrated that 54.55% (12/22) of the STS samples exhibited the features of BRCAness (12). Similar genomic and molecular features were also detected in 224 STS samples from The Cancer Genome Atlas database and in vitro (12). This suggests that BRCAness is a common genomic and molecular feature in some cases of STS (12). Adams and colleagues explored the effect of BRCA mutations on the efficacy of PLD in patients with recurrent ovarian cancer, and the results of their study demonstrated that patients with BRCA mutations exhibited a significantly higher response rate (RR) (56.5% vs. 19.5%; P = 0.004) to PLD in addition to improved PFS and OS (13).

Therefore, we performed this prospective, single-arm phase II clinical trial to further evaluate the efficacy and safety of the combination of PLD and IFO in de novo patients with advanced or metastatic STS. Furthermore, we attempted to identify the potential biomarkers of inflammation and genetic abnormalities.

Patients aged 18–75 years were considered eligible for inclusion in the study. Selected patients were those who were diagnosed with advanced or metastatic STS (not amenable to curative surgery or radiotherapy) and had not undergone chemotherapy. Other inclusion criteria included at least one measurable lesion according to the RECIST 1.1, an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–2, and a predicted life expectancy of ≥3 months. In addition, patients with adequate organ functioning were included in the study, including those with a left ventricular ejection fraction (LVEF) of ≥50%, normal electrocardiogram, an absolute neutrophil count of ≥2.0 × 10⁹ cells/L, a platelet count of ≥100 × 10⁹ cells/L, a hemoglobin level of ≥100 g/L, a total bilirubin level of ≤30 μmol/L, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels of ≤2.5 × the upper limit of normal (ULN; if liver metastases were present, ALT and AST ≤ 5 × ULN), a serum creatinine level of ≤ULN, or a creatinine clearance of ≥60 mL/minute (calculated using the Cockcroft and Gault equation).

Patients were excluded if they had osteosarcoma, Ewing sarcoma, primitive neuroectodermal tumors, gastrointestinal stromal tumors (GIST), rhabdomyosarcoma, or dermatofibrosarcoma protuberans. Other key exclusion criteria included clinically active involvement of the central nervous system; the use of other study drugs within 4 weeks before the administration of the first dose in the current study; known allergy to PLD, IFO, or any of its excipients; systemic infection or a body temperature of >38°C; and a ≥grade 2 chronic heart failure that meets the criteria of the New York Heart Association.

The study protocol was approved by Institutional Review Boards and independent ethics committees. Moreover, the study was designed and conducted in accordance with the Good Clinical Practice and the Declaration of Helsinki. All patients provided their written informed consent before enrolment in the study. This study was registered with ClinicalTrials.gov (NCT03268772).

Eligible patients were intravenously administered PLD at 30 mg/m² on day 1 and IFO at 1.8 g/m² plus Mesna (at 0.36 g/m² as an intravenous bolus, with IFO at 0, 4 and 8 hours since the initiation of IFO infusion) on days 1–5 every 3 weeks. Prophylactic GCSF was not administered during the first cycle of chemotherapy. The treatment was continued until disease progression, unacceptable toxicities, or patient death; alternatively, it was continued for up to six cycles. Dose modification was permitted, and dose delays were allowed for up to 3 weeks for patients with ≥grade 3 treatment-related adverse events (AE) as defined by the protocol.

The treatment response was assessed using contrast-enhanced CT and/or MRI after every two treatment cycles and every 3 months thereafter until disease progression and/or patient death. The response was assessed by investigators as per RECIST, version 1.1. In addition, pretreatment inflammatory indices were collected. The NLR was calculated as the absolute neutrophil count divided by the absolute lymphocyte count (ALC). The platelet-to-lymphocyte ratio (PLR) was defined as the number of platelets divided by the ALC, and the LMR was defined as the ALC divided by the number of monocytes. Using echocardiography, LVEF was evaluated at baseline, after two and four cycles, and at the end of treatment. AEs were assessed in accordance with NCI Common Terminology Criteria for Adverse Events 4.0. Follow-up was conducted for 3–4 weeks after the end of treatment and every 3 months thereafter.

Formalin-fixed and paraffin-embedded (FFPE) sections were stained with hematoxylin and eosin to evaluate the content of the tumor cells. Notably, only samples with a tumor content of ≥20% were selected for further analysis. The FFPE tissue sections were dewaxed in mineral oil and then incubated with a lysis buffer and proteinase K at 56°C overnight until the tissues were completely digested. Subsequently, the lysate was incubated for 4 hours at 80°C to reverse the formaldehyde-induced cross-linking. Genomic DNA was isolated from tissue samples using the ReliaPrep FFPE gDNA Miniprep System (Promega) and quantified using the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific) according to the manufacturer's instructions.

The captured libraries were loaded onto a NovaSeq 6000 platform (Illumina) for 100 bp paired-end sequencing. Raw data obtained from paired samples (an FFPE sample and a normal tissue control) were mapped to the reference human genome hg19 using the Burrows–Wheeler Aligner (v0.7.12; ref. 14). In addition, using Picard (v1.130), the PCR duplicates in alignment files were removed. Variant calling was performed only in the targeted regions. Somatic single-nucleotide variants (SNV) were detected using an in-house developed R package to execute a variant detection model based on a binomial test. Local realignment was performed to detect insertions/deletions (INDELs), and variants were filtered by their unique supporting read depth, strand bias, and base quality as described previously. Subsequently, all variants were filtered using an automated false-positive filtering pipeline to ensure sensitivity and specificity at an allele frequency of ≥1%. SNPs and indels were annotated using ANNOVAR against the following databases: dbSNP (v138), 1000Genome, and ESP6500 (population frequency > 0.015; ref. 15). Only missense, stopgain, frameshift, and nonframeshift indel mutations were maintained. Copy-number variations (CNV) and gene rearrangements were detected as described previously (16).

We developed an HRD algorithm—named the 3DMed-HRD algorithm—to assess genomic scars (17–19). Over 10,000 SNPs in the human genome were included in the analysis and incorporated in the 733-gene next-generation sequencing assays. The raw depth of SNP loci and probes used for HRD scoring were calculated using Sambamba (20). Moreover, the sum of heterozygosity (LOH), telomeric allelic imbalance (TAI), and large-scale state transitions (LST) were used to calculate HRD. Briefly, LOH was defined as nB = 0 and nA > 0, with a segment length of >15 mega base (Mb) and shorter than the entire chromosome; TAI was defined as nA! = nB for the segments that extended to the telomeric end of a chromosome but did not cross the centromere; and LST was defined as the breakpoint between two adjacent fragments of >10 Mb (21–23). If a tumor had an HRD score of >30, it was classified as high HRD score, whereas tumors with an HRD score of ≤30 were classified as low HRD score.

The primary endpoint of the study was PFS, and the secondary endpoints included OS, ORR, toxicities, and biomarker analyses. PFS was defined as the interval from the time of enrolment to disease progression or death due to any cause. OS was calculated from the time of enrolment to death due to any cause. ORR was calculated as the proportion of patients with complete response (CR) or partial remission (PR). The DCR represented the percentage of patients who achieved CR, PR, or stable disease (SD).

Descriptive statistical analysis data were presented using frequency with percentage for categorical variables and median with interquartile range (IQR) for continuous variables. The optimal cutoff for the continuous variables was determined using the survminer R package. The Kruskal–Wallis rank-sum test and the Wilcoxon signed-rank test for paired data were used for analyzing LVEF at baseline, during treatment, and at the end of treatment. All patients who underwent treatments were included in the full analysis set, and the efficacy and safety of treatment were analyzed. Response and survival outcomes were calculated using exact binominal CIs and the Kaplan–Meier method, respectively. All analyses were conducted using the R software version 3.3.3 (http://www.R-project.org) and SPSS 22.0 (SPSS Inc).

The sample size was calculated using the method of CI for One Proportion using the PASS software. For achieving the hypothesized 1-year PFS rate of 23.8%, 68 patients were required to be included to achieve a two-sided 95% confidence interval with a width of 0.2.

The study received funding from the Shanghai Anticancer Association Program (grant number: SACA-AX202107).

All data of this study are available per the corresponding author's approval.

The current study was designed and conducted in accordance with Good Clinical Practice and the Declaration of Helsinki. All patients provided written informed consent before enrolment in the study. This study was registered with ClinicalTrials.gov, number NCT03268772.

In total, 69 patients with chemotherapy-naïve advanced or metastatic STS were included in this study between May 2015 and November 2019. Of these, 46 (66.7%) patients had relapsed after prior surgical resection and 23 (33.3%) presented with advanced disease at the time of diagnosis. The most commonly observed histologic subtypes were synovial sarcoma (n = 12, 17.4%), leiomyosarcoma (n = 11, 15.9%), undifferentiated pleomorphic sarcoma (n = 9, 13.0%), sarcoma, not other specified (NOS; n = 9, 13.0%), liposarcoma (n = 8, 11.6%), epithelioid sarcoma (n = 5, 7.2%), and desmoplastic small round cell tumor (n = 5, 7.2%). Most patients (n = 59, 85.5%) had distant metastasis at baseline, and the majority of the patients (n = 67, 97.1%) had an ECOG PS of 0–1. The most common locations of primary cancers were the trunk and extremities (n = 33, 47.8%), followed by the abdomen and thoracic visceral organs (n = 20, 29.0%), retroperitoneal sites (n = 13, 18.8%), and head and neck (n = 3, 4.3%). The most common site of metastasis was the lung (55.1%), and 15 patients had lung metastasis alone. Patients’ baseline characteristics are listed in Table 1.

Patients underwent one to six treatment cycles, with a median of six cycles. Overall, 5 patients withdrew their consent of participation; of these, 2 patients withdrew consent after one cycle of treatment, 1 withdrew after two cycles, and 1 withdrew after three cycles of treatment (not related to toxic effects). The remaining one patient was initially misdiagnosed with fibrosarcoma at another center and was therefore enrolled in the current study; however, after a pathologic review at our center, the patient was diagnosed with dermatofibrosarcoma protuberans. Despite the SD status achieved by the patient after two cycles of treatment, we suggested the patient to withdraw from the study, and imatinib therapy was subsequently started for the patient. Three (4.3%) patients withdrew consent before the first radiological assessment. Till November 4, 2021, the median follow-up time was 47.2 (95% CI: 31.1–63.3) months. At the time of data cutoff, 58 (84.06%) patients had PFS events and 49 (71.01%) died. The median PFS, 1-year PFS rate, and the median OS were 7.3 (95% CI: 5.7–8.9) months, 21.4% (95% CI: 11.2%–31.6%; Fig. 1A), and 20.6 (95% CI: 16.3–25.0) months (Fig. 1B), respectively, for the entire cohort. Among patients with synovial sarcoma, the median PFS and OS were 10.2 (95% CI: 7.3–NA) and 27.7 (95% CI: 18.4–NA) months, respectively.

Of the 69 participants, 18 (26.1%) responded to the treatment, 1 (1.4%) showed CR, and 17 (24.6%) showed PR. SD was observed in 38 (55.1%) patients, resulting in an ORR of 26.1% and a DCR of 81.2%. The variable responses for the different histologic subtypes are listed in Supplementary Table S1. High RRs were observed in patients with synovial sarcoma (50%, 6/12), angiosarcoma (50%, 1/2), undifferentiated pleomorphic sarcoma (44.4%, 4/9), and sarcoma, NOS (44.4%, 4/9).

All 69 patients were included in the safety analysis (Table 2). The most commonly reported AEs were neutropenia (n = 63, 91.3%), leucopenia (n = 58, 84.1%), anemia (n = 47, 68.1%), nausea (n = 39, 56.5%), increased ALT levels (n = 22, 31.9%), vomiting (n = 17, 24.6%), thrombocytopenia (n = 16, 23.2%), and increased AST levels (n = 14, 20.3%). The majorly reported grade 3–4 toxicities included neutropenia (n = 43, 62.3%), leucopenia (n = 39, 56.5%), anemia (n = 12, 17.4%), febrile neutropenia (n = 8, 11.6%), thrombocytopenia (n = 5, 7.2%), and vomiting (n = 2, 2.9%). Grade 1–2 toxicities included allergic reactions (n = 10, 14.5%), mucositis (n = 8, 11.6%), and palmar–plantar erythrodysesthesia syndrome (n = 4, 5.8%). No grade 3–4 cardiotoxicities were observed. Moreover, no significant changes were observed between LVEF at baseline and after six cycles (P = 0.669, Kruskal–Wallis rank-sum test; Supplementary Fig. S1). In patients with LVEF after six cycles, there was no significant change between LVEF at baseline and after six cycles (P = 0.519, Wilcoxon signed-rank test for paired data). No treatment-related deaths occurred. The chemotherapy doses of 14 (20.3%) patients were adjusted for AEs.

The median NLR, PLR, and LMR were found to be 2.5 (IQR, 1.9–3.9), 157.5 (IQR, 124.0–217.2), and 3.8 (IQR, 2.6–5.3), respectively. We subsequently investigated the prognostic values of baseline NLR, PLR, and LMR. The preferred cutoffs for NLR, PLR, and LMR were 2.6, 111.6, and 2.6, respectively. A higher NLR was associated with shorter PFS (P = 0.037; Fig. 2A) and OS (P < 0.001; Fig. 2B). Moreover, as no significant differences were observed in the PFS (P = 0.058; Fig. 2C), a higher OS was observed in patients with a high LMR than in those with a low LMR (P = 0.0027; Fig. 2D) at baseline. No difference was observed in the PFS (P = 0.13) and OS (P = 0.29) of patients with high and low PLR at baseline. After adjusting for the disease stage, the NLR and LMR were found to be uncorrelated with the patient's PFS (Supplementary Table S2), whereas the NLR was inversely correlated with the patient's OS (P = 0.047).

Pretreatment tumor samples were available for 23 patients; of these, 22 samples were taken from primary tumors, whereas only one sample was taken from a metastasis site. Somatic SNVs were detected in all 23 patients (Supplementary Table S3), with each patient having 2–26 somatic SNVs. Moreover, somatic CNVs were identified in 17 of the 23 patients, with each patient having 1–12 somatic CNVs. Alterations including SNV and CNV were detected in 230 genes. CNV gain events occurred in 19 genes, whereas CNV loss events occurred in 41 genes (Fig. 3). TP53 was the most commonly found mutant gene (9/23, 39.1%). Notably, all 23 patients demonstrated microsatellite stability.

We attempted to identify genomic alterations associated with a clinical outcome in patients treated with PLD plus IFO. No patient harboring NF1 mutations responded to PLD plus IFO treatment, and two of three patients experienced disease progression at the first response evaluation. Despite the limited sample size, all 3 patients with NF1 SNV exhibited a shorter OS (P < 0.0001) and PFS (P = 0.044) than those without NF1 SNV. In addition, 3 patients with LIG1 SNV had a shorter OS (P = 0.014) than those without LIG1 SNV. Furthermore, the association between CNV and patient survival was analyzed, and the results indicated that patients with BRCA2 CNV had poor survival rates. Overall, 2 patients with BRCA2 loss demonstrated no response to PLD plus IFO first-line treatment, and they died within 10 months.

Tumor mutation burden (TMB) is defined as the total number of nonsynonymous somatic mutations per Mb. TMB was significantly associated with PFS but not with OS. Patients with a TMB higher than the mean value (n = 13; Mut/Mb) exhibited a longer PFS (8.5 vs. 5.0 months; P = 0.0095; Fig. 4A) than those (n = 10) with a TMB lower than the mean value (n = 10).

The association of HRD scores with patient outcomes was analyzed for 21 patients with evaluated HRD scores. The IQR of the HRD scores was 3–30. The mean HRD score was 18.1 (95% CI: 10.5–25.7), and the median HRD was 11 (95% CI: 5–28). No significant difference was identified in terms of PFS between patients with high HRD score and low HRD score (P = 0.81). A better OS was observed in patients with low HRD score than in those with high HRD score (44.2 vs. 17.3 months; HR = 0.173; 95% CI = 0.043–0.702; P = 0.0056; Fig. 4B).

Our study demonstrated that combination therapy with PLD and IFO was effective and well tolerated as the first-line treatment for patients with advanced or metastatic STS. The median PFS and OS were found to be 7.3 and 20.6 months, respectively, and the 1-year PFS rate was 21.4% (95% CI: 11.2–31.6). The ORR was 26.1%, and the DCR was 81.2%. Safety was manageable overall.

In patients with advanced or metastatic STS, the most commonly used first-line systemic treatment is based on a doxorubicin regimen. Previously, an European Organisation for Research and Treatment of Cancer (EORTC) randomized phase II trial compared the efficacy and safety of PLD and ADM (6). This previous study included some specific subtypes of STS, such as rhabdomyosarcoma and GIST. Notably, after excluding patients with GIST, the RRs were l4% and 12%, and the median times to progression were 65 and 82 days for PLD and ADM, respectively. Although this study suggested an equivalent activity and improved toxicity profile for PLD compared with ADM (grade 3/4 neutropenia: 6% vs. 77% and febrile neutropenia: 2% vs. 16%), the efficacy of anthracycline monotherapy is limited.

The combination therapy of ADM and IFO had been investigated in several trials. In an EORTC phase II trial, ADM and IFO (ADM, 50 mg/m², day 1; IFO, 5 g/m² 24 hours infusions repeated every 3 weeks) as the first-line treatment for advanced STS resulted in a RR of 35%, a median time to progression of 6.7 months, and a median OS of 13.5 months. Myelosuppression was reported as dose-limiting toxicity, and 73% of patients had grade 3–4 leukopenia (24). A previous randomized controlled phase III study conducted by Edmonson and colleagues demonstrated that in first-line treatment of advanced STS, the RR for ADM alone (80 mg/m², day 1, repeated every 3 weeks) was 20% and that for ADM combined with IFO (ADM 30 mg/m², days 1–2, IFO 3 750 mg/m², days 1–2, repeated every 3 weeks) was 34% (P = 0.03). In the ADM combined with IFO group, the incidence of ≥grade 3 myelosuppression was 80%, and the 3 patients (3.4%) experienced treatment-related death because of gastrointestinal bleeding, respiratory arrest, and cardiopulmonary issues, respectively. The incidence of severe nausea and vomiting was 18% (25). The phase III EORTC 62012 trial explored whether the addition of IFO to intensified ADM could improve patient outcomes. It reported that ADM-IFO improved PFS (7.4 vs. 4.6 months; P = 0.003) and ORR (26% vs. 14%; P = 0.006) compared with ADM alone (2). However, no advantage in terms of OS was observed with the use of this combination regimen (12.8 months in the ADM group vs. 14.3 months in the ADM-IFO group; P = 0.076); thus, the combination regimen is commonly reserved for patients with good performance status and rapid tumor progression. Notably, high toxicity has been reported with combination treatments, resulting in an increased number of patients being unable to complete the planned six cycles of treatment compared with the number of patients on ADM alone. The results of the current study revealed favorable activity with the use of the combination of PLD plus IFO. Similar to previous observations, the current study presented a median PFS of 7.3 months. Importantly, the incidence of grade 3–4 AEs, such as nausea (1.4%), vomiting (2.9%), and neutropenia (62.3%), with PLD plus IFO combination treatment was lower than that with the ADM-IFO regimen reported in previous studies. Notably, in the EORTC 62012 trial, even with the pegfilgrastim prophylactic treatment, the incidence of febrile neutropenia reached 46% (2). Furthermore, in the current study, prophylactic GCSF was not used during the first cycle of chemotherapy, and the incidence of febrile neutropenia was much lower (11.6%). Notably, the EORTC 62012 study only included patients aged ≤60 years, and 18% of the patients in the ADM plus IFO arm discontinued the treatment because of toxic effects. Our study included 9 (13.0%) patients ages ≥60 years, and no patients discontinued the treatment because of toxic effects. Moreover, palmar–plantar erythrodysesthesia and mucositis observed in our patients were of grades 1–2. On the basis of these results, the overall safety profile of combination treatment with PLD as a replacement for ADM was acceptable and potentially offers more clinical benefits for de novo patients with advanced or metastatic STS.

ADM can cause cumulative irreversible cardiotoxicity, thereby resulting in damage to myocardial fibers, which is manifested as decreased LVEF and congestive heart failure. This AE can occur after several years of chemotherapy cessation. The cumulative dose of ADM should not exceed 450–550 mg/m² owing to its dose-limiting cardiotoxicity (26, 27). In fact, cardiac AEs were reported in 3.5% of patients with STS who underwent treatment with ADM in an EORTC phase III trial, and these AEs were more frequently observed in the ADM-IFO arm, with grade 3 abnormal functions in 2% of the patients (28). Efforts to lower the occurrences of ADM-related cardiotoxicity are ongoing, and these include optimization of the administration and schedule (29). Previous evidence demonstrated that with dexrazoxane coadministration, ADM can be given at high cumulative doses (>450 mg/m²), with a low rate of cardiotoxicity. However, 50% of the patients exhibited a decrease in LVEF of >10% from baseline after eight cycles of ADM at 75 mg/m² despite dexrazoxane coadministration (30). The cumulative dose of PLD was reportedly as high as 2,000 mg/m² (31, 32), implying that it could be used in more cycles. In the current study, no significant change in LVEF was observed after treatment. The reduced toxicity of PLD-IFO reported in this study compared with that of the ADM-IFO regimen reported previously encourages further exploration of the continued use of PLD in responding patients and in patients exposed to anthracyclines. Furthermore, the acceptable safety profile of PLD-IFO supports further investigation with the addition of immunotherapy or targeted agents to this combination. Optimal selection of patients for this combination treatment could enhance the efficacy and avoid additional toxicity. Histology was found to be correlated with patient response and hence should be taken into consideration when making treatment decisions. Our study demonstrated favorable RRs in patients with synovial sarcoma (50%), undifferentiated pleomorphic sarcoma (44.4%), and sarcoma, NOS (44.4%, 4/9). The subgroup analysis of a randomized controlled phase III study by Edmonson and colleagues also reported that the RR of patients with synovial sarcoma was 88% in the ADM-IFO treatment group and only 20% in the ADM monotherapy group (P = 0.02). These results suggest that the combination regimen is effective for the treatment of these high-grade STS subtypes. Our study results indicated that high NLR and low LMR values at baseline were associated with poor survival. In multivariate analyses, only the NLR was found to be inversely correlated with OS. Similarly, Cheng and colleagues revealed that NLR was independently correlated with poor survival in patients with synovial sarcoma and that the LMR was an independent positive prognostic indicator (11). These results indicated that the pretreatment inflammatory index is a prognostic factor rather than a regimen predictor and that future studies incorporating inflammatory indices, such as LMR and NLR, in prognosis are warranted to better stratify patients’ risk.

In the current study, a biomarker assessment suggested that BRCA2 loss and LIG1 SNV were associated with poor response to first-line PLD plus IFO treatment, and these findings were correlated with poor survival. Indeed, a total of 3 patients [sarcoma, NOS (n = 1), leiomyosarcoma (n = 1), fibrosarcoma (n = 1)] harbored the NF1 mutation. Of these, one patient showed the best SD response, whereas the other two showed progressive disease. The median PFS and OS were found to be 1.4 and 9.3 months, respectively. Two patients exhibited loss of BRCA2, including one with fibrosarcoma and one with leiomyosarcoma. Notably, 2 patients had concurrent NF1 SNV and BRCA2 CNV. Accordingly, as this study included a limited number of patients, studies with a larger sample size are needed to further investigate this correlation.

The correlation between HRD scores and patient outcomes has rarely been addressed among patients with STS. To distinguish HRD, LOH, LST, and TAI were evaluated, and the HRD scores included all three parameters (33–35). In the current study, the median HRD was 18.1, with an IQR of 3–30, indicating a variation of HRD in STS. A higher HRD score was associated with poor OS in our trial. Notably, a high HRD score was previously reported to be associated with poor outcomes in patients with head and neck squamous cell carcinoma (36). Further studies are needed to investigate the association between HRD and survival in patients with STS and to investigate the potential role of PARP inhibitors in patients with STS who have a high HRD score.

This study had some limitations. First, this was a single-arm trial with a small sample size; therefore, the efficacy and safety of PLD-IFO treatment could only be compared with previous studies on ADM-IFO treatment. Further exploration in large cohort studies is warranted. Second, STS represents a heterogeneous group with various subtypes. However, the current study included 14 different subtypes, with limited numbers of patients in each subtype. Histology-specific treatments for STS are lacking since 2015. Patients with indolent subtypes, such as alveolar soft part sarcoma (n = 1), chondrosarcoma (n = 1), and paraganglioma (n = 1), were also included in the current study. Finally, biomarker assessment was conducted for only 23 patients, limiting the power of the statistical analysis and warranting further investigation.

In summary, PLD in combination with IFO is an effective and well-tolerated first-line treatment for patients with advanced or metastatic STS. Further investigations are warranted to evaluate the efficacy of this combination therapy in larger studies using a histological or molecular approach.

X. Liu reports grants from Shanghai Anticancer Association Program during the conduct of the study. No disclosures were reported by the other authors.

X. Liu: Conceptualization, data curation, supervision, funding acquisition, investigation, methodology, writing–original draft, writing–review and editing. S. Jiang: Data curation, investigation, methodology, writing–original draft, writing–review and editing. H. Wang: Data curation, investigation, methodology. X. Wu: Data curation, investigation, methodology. W. Yan: Data curation, investigation, methodology. Y. Chen: Data curation, investigation, methodology. Y. Xu: Data curation, investigation, methodology. C. Wang: Data curation, investigation, methodology. W. Yao: Data curation, investigation, methodology. J. Wang: Data curation, investigation, methodology. L. Yu: Data curation, investigation, methodology. J. Miao: Data curation, investigation, methodology. H. Chen: Data curation, investigation, methodology. J. Xia: Data curation, investigation, methodology. M. Huang: Data curation, investigation, methodology. X. Zhang: Conceptualization, data curation, supervision, investigation, methodology, writing–review and editing. Z. Luo: Conceptualization, resources, data curation, funding acquisition, project administration, writing–review and editing.

We thank all the investigators as well as patients and their families who participated in this study.

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.