Maintenance therapy and drug holiday in sarcoma patients: systematic review

Loïc Lebelleca,b, Anne-Sophie Defachellesc, Pierre-Yves Crenb and Nicolas Penelb,d

ABSTRACT

Purpose: Overall prognosis of advanced sarcoma remains poor, optimization of systemic treatment is urgently needed in this setting.
Materials and methods: We systematically reviewed fully published English-speaking literature about maintenance therapy and drug holiday in sarcoma patients management.
Results: We found that switch maintenance therapy with cyclophosphamide/vinorelbine improves the outcome of localized high-risk rhabdomyosarcoma. There is no other maintenance therapy recom- mended in sarcoma patients. After classical chemotherapy, maintenance therapy with immune-stimu- lating agents for localized osteosarcoma, bevacizumab for advanced angiosarcoma or pediatric advanced sarcoma, or mTOR inhibitors for metastatic sarcoma does not improve the outcome. Drug holiday has been assessed for metastatic gastrointestinal stromal tumor treated with imatinib as the first-line therapy or for metastatic soft-tissue sarcoma treated with trabectedin. Drug holiday has been found to lead to rapid disease progression and should be avoided.
Conclusions: Data about both maintenance and drug holiday are spare in sarcoma management.

Abbreviations: 95% CI: 95% confidence interval; CTX: cyclophosphamide; DFS: disease-free survival; EFS: event-free survival; GIST: gastrointestinal stromal tumor; IFNa: interferon-alpha; ITT: intention to treat (analysis); IVA: ifosfamide/vincristine/actinomycin; IVADo: ifosfamide/vincristine/actinomycin/doxo- rubicin; MAI: methotrexate/doxorubicin/ifosfamide; MAP: methotrexate/doxorubicin/cisplatin; mTOR: mammalian target of rapamycin; MTP: muramyl tripeptide phosphatidylethanolamine; OS: overall sur- vival; n: number of patients; p: p-value; PFS: progression-free survival; PP: per protocol (analysis); RIDA: ridaforolimus; RMS: rhabdomyosarcoma; VIN: vinorelbine

Introduction

Sarcomas include a large number of clinico-pathological enti- ties with variable aggressiveness. Management of some local- ized sarcomas could require adjuvant or neoadjuvant treatments. Most patients with metastatic/advanced disease are best treated with systemic and palliative therapies that primarily aim to alleviate symptoms and improve quality of life. The toxicity/efficacy ratio of systemic treatments needs to be cautiously weighted.
Maintenance therapy gathers different approaches depending on both the clinical setting (curative-intent versus palliative treatment) and the nature of maintenance therapy (continuation maintenance therapy versus switch maintenance).
Traditionally, at advanced stage, not amenable to cura- tive-intent strategy, systemic treatment could be adminis- tered until severe toxicity, disease progression, or cumulative dose have been reached (e.g., as in the case of doxorubicin). After the administration of classical chemotherapy and in the absence of disease progression (including complete or partial response or stable disease), maintenance therapy could be discussed. Maintenance therapy aims to maintain response, prevent or delay progressive disease, and improve quality of life. In this setting, there is no consensus on maintenance therapy duration. In this setting, the key-difference between maintenance therapy and further line is that maintenance therapy is given in patients experiencing non-progressive dis- ease and further line is given in those experiencing progres- sive disease.
At the localized stage, maintenance therapy could be used to maintain complete response and delay or reduce the risk of relapse. Maintenance therapy is then part of (neo)ad- juvant strategies. In contrast, at the metastatic/advanced stage, drug holiday, which is planned discontinuation of sys- temic treatment providing stable disease or response, is con- sidered with the aim to reduce treatment-related toxicity and improve the quality of life. Usually in this setting the duration of maintenance therapy is clearly defined.
At least two different kinds of maintenance therapy have been established: (i) same regimen at a lower dose or a part of the ‘induction’ treatment (‘continuation maintenance ther- apy’), or (ii) use of another drug after administration of ‘induction’ treatment (‘switch maintenance therapy’).
In the present review, we summarize available fully English-written published trials assessing maintenance ther- apy and drug holiday in sarcoma patients. This systematic review was conducted using Medline Database (last searched on 24 November 2019) and according to PRISMA guidelines [1]. Only English-written prospective clinical trials assessing ‘drug-holiday’ or ‘maintenance therapy’ in ‘sarcoma’ patients have been retained. The following filters (‘trials’ and ‘humans’) and terms have been used for searching articles ‘treatment interruption þ Sarcoma’ (30 articles), ‘maintenance therapy þ Sarcoma’ (58 articles). Articles have been selected after abstract reading. All articles have been checked by all authors. Additional references have been added after reading of the selected articles.

Results

Maintenance therapy in curative-intent settings

Use of immune-stimulating agents in osteosarcoma patients

Osteosarcoma is treated by preoperative chemotherapy, fol- lowed by curative-intent surgery and then adjuvant chemo- therapy. Adjuvant chemotherapy could be adapted according to the rate of chemotherapy-induced tumor necro- sis. Surgery is the best treatment in cases of limited relapse (e.g., limited lung metastasis). In the case of diffuse metasta- sis, there is no consensus on treatment and no recognized active salvage treatment. Therefore, the efficacy of (neo)adju- vant chemotherapy is critical for improving outcomes, and the addition of maintenance therapy is an appealing approach in this context. Two immune-stimulating agents have been assessed as maintenance therapy in patients with localized high-grade osteosarcoma receiving adjuvant chemotherapy in two large phase 3 trials (Table 1).
The rationale for using interferon-alpha (IFNa) as mainten- ance therapy is based on (i) the fact that IFNa acts as an antiproliferative, differentiation-inducing, pro-apoptotic, and antiangiogenic agent and (ii) the fact that some osteosarcomas could express IFN receptors [2,3]. Mu€ller et al. reported their experience of 89 localized high-grade osteosarcomas treated between 1971 and 1990 with IFNa without chemotherapy. With a median follow-up period of 12 years, the observed 10-year metastasis-free and sarcoma-specific survival rates were 39% and 43%, respectively. The data of this retrospective series were regarded as promising [2]. Therefore, the EURAMOS group launched a large phase 3 trial assessing the role of IFNa after (neo)-adjuvant therapy for localized high-grade osteosarcoma (n ¼ 716) [3]. The patients received two cycles of methotrexate/doxorubicin/cis- platin as preoperative treatment and then underwent cura- tive-intent surgery. Good responders were randomly assigned to receive four additional cycles of chemotherapy or four additional cycles of chemotherapy followed by weekly subcutaneous IFNa-2b for 74 weeks. The primary objective of the trial was to improve event-free survival (EFS) (Table 1). Approximately 18% (64/357) of the patients assigned to the experimental arm refused IFNa-2b. IFNa-2b was prematurely stopped due to toxicity in approximately 13% of the patients assigned to the experimental arm. Of the 271 patients actually treated with IFNa-2b, 132 (49%) required dose reduction or delays. There was no improve- ment in both the 3-year EFS and overall survival (OS) (Table 1).
Muramyl tripeptide phosphatidylethanolamine (MTP), a nonspecific immune modulator, is a synthetic analog of a bacterial membrane. Studies with preclinical models suggest that MTP has antitumoral effects. An exploratory phase II trial was conducted in 16 patients with metastatic osteosarcoma; pharmacodynamic endpoints were reported but clinical end- points were not reported in this trial [4]. Meyers et al. reported a large phase 3 trial with a factorial design. The patients in this trial were first randomly assigned to receive two cycles of preoperative chemotherapy with methotrexate/ doxorubicin/cisplatin (MAP) versus two cycles of methotrex- ate/doxorubicin/ifosfamide (MAI) [5,6]. The patients under- went curative-intent surgery. After surgery, the patients were divided into four arms: MAP (as an adjuvant regimen) versus MAP þ MTP followed by MTP alone versus MAI versus MAI þ MTP followed by MTP alone. MTP was administered twice weekly for 12 weeks beginning at week 12 and weekly for an additional 24 weeks beginning at week 24. Using the factorial design, the authors determined that 585 patients were needed to observe an EFS improvement with an expected hazard ratio of 0.64, a two-sided a risk of 5%, and a b risk of 20%. The key findings are summarized in Table 1. The interpretation of this trial is difficult because of the use of the factorial design, with inherent double randomization constituting the four arms. It is necessary to test first the assumption of equal relative risks for one randomization (ifosfamide: yes, versus no) compared across strata defined by the second randomization (MTP-PE: yes, versus no). In the first analysis, that is the primary analysis, the authors reported an interaction between the ifosfamide-based regi-therapeutic role of the MTP added could not be appropriately determined [6]. In the second analysis published in 2008, the authors did not find an interaction between the ifosfamide-based regimen and MTP; however, although MTP addition failed to significantly improve EFS, it improved OS. Metronomic chemotherapy in high-risk rhabdomyosar- coma patients Rhabdomyosarcoma (RMS) is the most common soft-tissue sarcoma in children and young adults. It has three histo- logical subtypes: embryonal RMS (accounts for approximately 80% of pediatric RMS cases), alveolar RMS (accounts for approximately 15% pediatric RMS cases), and pleiomorphic RMS (the rarest form that develops in aged adults). Alveolar RMS is particularly aggressive. The survival of patients with non-metastatic RMS has been reported to be approximately 70% with the risk-adapted multimodal treatment strategy.
Metronomic chemotherapy consists of continuous admin- istration of low-dose chemotherapy [7]. Cyclophosphamide (CTX) and vinorelbine (VIN) act as cytotoxic, antiangiogenic, and immune-stimulating agents when administered continu- ously at low doses [7,8]. Several exploratory trials have dem- onstrated the feasibility and the clinical activity of this combination in RMS patients with advanced and refractory disease [9,10].
Bisogno et al. reported the results of the RMS 2005 trial (NCT00339118). The trial included patients non-metastatic high-risk RMS with ages ranging from 6 months to 21 years [11]. High-risk RMS patients were defined as follows: patients with incompletely resected non-alveolar RMS (IRSII or III) at unfavorable sites with an unfavorable age (≥10 years) or tumor size (>5 cm), or both; non-metastatic non-alveolar RMS with nodal involvement; and non-metastatic alveolar RMS but without nodal involvement. Patients in remission after standard treatment nine cycles of ifosfamide, vincristine, and dactinomycin with or without doxorubicin, and local treatment (radiotherapy with or without surgery) were ran- domly assigned (1:1) to follow-up or switch maintenance therapy to chemotherapy (six cycles of intravenous VIN 25 mg/m2 on days 1, 8, and 15, and daily oral CTX 25 mg/m2, on days 1–28). The primary endpoint was disease-free sur- vival (DFS). The sample size was calculated based on the fol- lowing assumptions: expected hazard ratio of 0.67, two-sided a risk of 5%, and b risk of 20%. Finally, 371 patients were enrolled (186 in the observation arm and 185 in the switch maintenance therapy arm). The key results are summarized in Table 2. The improvement in DFS did not meet the pre- specified level of significance; nevertheless, the difference between both arms was clinically relevant. This trial is undoubtedly a trial that would change clinical practice since switch maintenance chemotherapy significantly improved the OS in this patient population.

Maintenance therapy in advanced stages

In advanced stages of different clinico-pathological entities, different maintenance therapies (e.g., metronomic chemo- therapy, antiangiogenic agents, and molecularly targeted therapies) have been explored. Most of the related trials have been phase 1 trials, with a few being phase 3 trials.
Switch maintenance therapy with metronomic chemotherapy Klingebiel et al. reported a randomized phase 2 trial in which the activity and safety of two approaches after first-line treat- ment for metastatic soft-tissue sarcoma in children and ado- lescents were assessed: high-dose chemotherapy versus oral metronomic chemotherapy [12]. The first-line treatment con- sisted of six cycles of alternatively administered ifosfamide- vincristine-actinomycin and carboplatin-epirubicin-vincristine. High-dose chemotherapy consisted of one cycle of CTX and thiotepa, a second cycle of melphalan and etoposide fol- lowed by stem cell infusion. In the other arm, metronomic chemotherapy was administered. It consisted of eight cycles in which four cycles of trofosfamide-etoposide (trofosfamide 2 75 mg/m2/day during 10 days plus etoposide 2 25 mg/ m2/day during 10 days) and four cycles of trofosfamide-idaru- bicin (trofosfamide 2 75 mg/m2/day during 10 days plus idarubicin 5 mg/m2 on days 1, 4, 7, and 10) were alternatively administered. The key findings are summarized in Table 3. The interpretation of this trial is difficult since we are not able to distinguish the potential benefit of switch mainten- ance therapy from that of metronomic chemotherapy and the potential deleterious effect of high-dose chemotherapy. Furthermore, the induction treatment and the switch main- tenance therapy are complex, the sample size calculation has not been explained, and compliance with the full protocol is low since the presented data described the outcomes of only 96 patients who completed the treatment, although 295 were registered in the trial.
Bevacizumab as continuation maintenance therapy Angiogenesis plays a key role in sarcoma biology. The hypothesis that antiangiogenic agents could delay disease progression in advanced sarcoma patients is appealing. To the best of our knowledge, three trials have assessed the role of bevacizumab, a monoclonal antibody inhibiting vas- cular endothelial growth factor receptor, as continuation maintenance therapy after combination with chemotherapy.
Two trials focused on adult patients with soft-tissue sar- coma. First, D’Amado et al. reported a non-randomized two- stage phase 2 trial closed to recruitment after analysis of the first stage. Seventeen soft-tissue sarcoma patients (including 11 leiomyosarcoma patients) were enrolled to receive both doxorubicin and bevacizumab, followed by bevacizumab as maintenance therapy. Only 5 out of the 17 patients (29%) received bevacizumab as maintenance therapy. The overall response rate was 12%. There were two instances of partial responses in two patients with leiomyosarcoma. The median time to progression was 8 months [13]. Ray-Coquard et al. reported a non-comparative randomized phase 2 trial assess- ing the addition of bevacizumab to weekly paclitaxel in patients with advanced angiosarcoma [14]. Angiosarcoma accounts for approximately 1% of all soft-tissue sarcomas. In the experimental arm, bevacizumab was administered during the chemotherapy and as maintenance therapy in patients with non-progressive disease. The primary aim was to esti- mate the activity of both regimens. The primary objective was the 6-month progression-free survival (PFS). Fifty patients were included. Bevacizumab was administered as maintenance therapy in eight (32.0%) of the 25 patients in the experimental arm. In both arms, 14 patients showed non-progression at 6 months. Other activity endpoints are summarized in Table 4. Grade 3 or 4 toxicities seemed more frequent in the combination arm. In both studies, bevacizu- mab was administered during chemotherapy, and a limited number of patients received bevacizumab as maintenance therapy because of rapid disease progression during combin- ation treatment. These two trials do not suggest a major value of bevacizumab addition as maintenance therapy in soft-tissue sarcoma.
Chisholm et al. reported a large comparative randomized phase 2 trial assessing the addition of bevacizumab to chemotherapy in children and adolescent patients with metastatic sarcomas. The chemotherapy consisted of four cycles of ifosfamide/vincristine/actinomycin/doxorubicin (IVADo), followed by five cycles of ifosfamide/vincristine/acti- nomycin (IVA). Chemotherapy was followed by maintenance chemotherapy with metronomic CTX and vinorelbine. The patients were randomly assigned to receive either chemo- therapy or chemotherapy plus bevacizumab. The primary objective was to improve EFS. The statistical hypothesis was as follows: hazard ratio of 0.57, two-sided a risk of 5%, and b risk of 20%. One hundred fifty-four patients were randomized to receive chemotherapy (n ¼ 80) or bevacizumab plus chemotherapy (n ¼ 74). The median EFS was 14.9 months with chemotherapy versus 20.6 months with chemotherapy plus bevacizumab; this difference in EFS did not reach the level of significance (HR ¼ 0.93; 95% CI: 0.61–1.41; p ¼ .72). There were no treatment-related deaths and no increase in the incidence of grade 3 or 4 toxicity in the combination arm (Table 4).
Switch maintenance therapy with an mTOR inhibitor Intracellular signaling through the mammalian target of rapa- mycin (mTOR) and associated upstream signaling pathways are dysregulated in most sarcoma subtypes. Ridaforolimus (RIDA) is a non-prodrug analog of rapamycin (sirolimus) with high affinity for mTOR and optimized solubility, stability, and bioavailability. Owing to these pharmacological properties, RIDA was first evaluated as a treatment for advanced sarco- mas in a non-randomized phase 2 trial [16]. A total of 212 heavily pretreated patients were enrolled; 61 patients (28.8%) achieved non-progressive disease at the best response. The best objective response rate (ORR) was 2%. The reported median PFS was 15.3 weeks and the median OS was 40 weeks.
In 2013, Demetri et al. reported a large double-blind phase 3 trial assessing the clinical benefit of switch mainten- ance therapy with oral RIDA in sarcoma patients showing stable disease or response after first-, second-, or third-line treatment for advanced disease [17]. The study population was heterogeneous, including patients aged 13 years or higher and soft-tissue (90%) and bone sarcoma (10%) patients regardless of the histological subtype. The primary objective was median PFS by a blinded central review. The sample size calculation was based on the following assump- tions: expected hazard ratio of 0.75, b risk of 10%, and a two-sided a risk of 5%. Overall, 711 patients were randomly assigned to receive placebo (364) or RIDA (347). The key findings are summarized in Table 5. Technically, it was a positive superiority phase 3 trial meeting the primary object- ive since RIDA significantly improved the centrally-reviewed PFS. However, this trial failed to support the approval of the study drug since there was no improvement in OS and because the patients benefiting from RIDA were not clearly identified. The missing element was a validated predictive biomarker that could clearly target a subpopulation of patients benefiting from treatment with mTOR inhibitors as switch maintenance therapy.

Drug holiday in sarcoma patients

A drug holiday is a planned hold of active treatment to reduce toxicity in a palliative setting. This concept requires some prerequisites: the active drug is held before the devel- opment of unacceptable toxicity, the active drug does not cause cumulative toxicity that limits its administration, and the active drug provides a ‘long-term’ clinical benefit (at least stable disease). In daily practice, both referring physician and patient mutually agree to holding the treatment because the patient would like to rest some time, or make a summer break to meet with relatives, etc. To the best of our know- ledge, the aggressiveness of sarcomas has been explored in a limited number of clinical trials, including two major randomized trials. We have excluded a randomized phase 2 trial with a discontinuation design that assessed the activity of regorafenib in soft-tissue sarcoma, since only eight patients were randomly assigned to continue or discontinue regorafenib after 12 weeks of administration [18].
Soft-tissue sarcoma and trabectedin drug holiday Trabectedin is an alkylating agent prescribed for soft-tissue sarcoma patients. Some patients could receive a considerable number of cycles (e.g., 71 cycles over 6 years) before cumula- tive toxicity develops [19]. Le Cesne et al. reported the results of the T-Dis trial, a non-comparative randomized phase 2 trial with a discontinuation design (NCT01303094). In the T-Dis trial, patients first received up to six cycles of tra- bectedin and the patients who showed stable disease or response were randomly assigned to continue trabectedin until intolerance and disease progression or discontinue tra- bectedin (drug holiday) and then rechallenge at disease pro- gression. The primary objective was to determine the 6- month PFS after randomization. In this exploratory phase 2 trial, there was no sample calculation, the aim was to ran- domize at least 50 patients. One hundred seventy-eight patients were enrolled in the first part of the trial; 53 patients were randomly assigned to continue (n ¼ 27) or dis- continue trabectedin (n ¼ 26). Among 26 patients assigned to drug holidays, rechallenge with trabectedin was per- formed in 22 cases (85%). The data are summarized in Table 6 [20,21]. The sample size did not allow a formal comparison between the two arms; further, there was no quality-of-life analysis in this study. Nevertheless, quality-adjusted time without symptom and tumor progression (Q-TwisT) was cal- culated; Q-Twist was slightly better in the continuation arm, but this Q-Twist difference did not reach the level of signifi- cance [21]. Furthermore, there was a trend suggesting a benefit of continuation for OS; however, this clinically rele- vant difference did not reach the classical level of significance.
Gastrointestinal stromal tumor and imatinib drug holiday Gastrointestinal stromal tumors (GISTs) are the most fre- quently developing sarcomas in adult humans. Approximately 85% of GISTs show activating mutation of c- Kit or platelet-derived growth factor receptor-a (PDGFR-a). Most of these activating mutations are sensitive to imatinib, a tyrosine kinase inhibitor. Imatinib is used as an adjuvant treatment, as preoperative treatment to obtain tumor shrink- age or as a first-line treatment at the metastatic stage. The median OS of metastatic GIST patients is approximately 6 years with imatinib therapy. Whether imatinib should be continued or discontinued after partial or complete response has been achieved remains debatable. Blay et al. reported the results of a randomized trial with a discontinuation design in imatinib-treated GIST patients after 1 year [22] and then after 3 years (2010). In the initial design, the primary objective was the determination of the 1-year PFS. The statis- tical hypothesis was as follows: 1-year PFS of 90% with con- tinuation, 1-year PFS of 75% with discontinuation, two-sided a and b of 10%. In the first part of the trial, 182 patients were enrolled and 58 were randomly assigned to continue (26) or discontinue (32) imatinib. The median PFS was 18 months in the continuation arm versus 6.1 months in the discontinuation arm (p < .0001). The OS was similar in both arms. Quality-of-life data were not analyzed because of excessive missing data [22]. Randomization was planned after 3 years of treatment; 434 patients were enrolled, 50 of whom were free of progression at 3 years and were randomly assigned to continue or discontinue treatment. After ran- domization, the 2-year PFS was 80% (95% CI: 58–91) in the continuation group and 16% (5–33) in the interruption group (p < .0001) [23].

Discussion

Overall, data about the clinical benefit of maintenance ther- apy or drug holiday in sarcoma management are limited. The use of immune-stimulating agents as maintenance ther- apy with IFNa in localized osteosarcoma failed to improve outcomes in a large phase 3 trial. The results of MTP adju- vant treatment in localized osteosarcoma are difficult to interpret. CTX and VIN as switch maintenance therapy for 6 months after chemotherapy for localized high-risk RMS must be considered as part of the standard of care since this metronomic chemotherapy significantly improves OS. The role of metronomic chemotherapy at an advanced stage has not been properly assessed. Bevacizumab continuation main- tenance therapy failed to demonstrate a clinical benefit in advanced angiosarcoma, a paradigm of tumoral angiogen- esis. Addition of bevacizumab to first-line treatment and switch maintenance therapy in children with advanced soft- tissue sarcoma failed to improve outcomes. Lastly, the mTOR inhibitor RIDA significantly improved the PFS in patients with advanced soft-tissue or bone sarcoma but without impacting the OS. Furthermore, predictive biomarkers to better target the population benefiting from switch maintenance therapy with an mTOR inhibitor are yet to be established. To con- clude, the sole clinical setting in which maintenance therapy must be considered is high-risk localized RMS. Further clinical studies with a strong biological rationale and/or convincing data from phase 2 trials are necessary before maintenance therapy could be recommended in other settings.
We stress that CTX and VIN is an active regimen providing an objective response in RMS: ORR of 35–38% in two phase 2 trials [9,10]. On the contrary, RIDA provides a limited ORR of 2% in sarcoma patients [16]. Bevacizumab provided a short-term partial response in only 9% of angiosarcoma patients [24]. There are no published data demonstrating the activity of IFNa or MTP in osteosarcoma. These facts suggest that maintenance therapy could be effective if the selected regimen provides clear evidence of activity at the metastatic stage. Furthermore, the toxicity profile must be carefully con- sidered. In case of advanced disease, the maintenance ther- apy administered to delay disease progression must be well tolerated and must not alter the quality of life. On the con- trary, at the localized stage, maintenance therapy, which reinforces the therapeutic role of adjuvant or neoadjuvant classical chemotherapy, could be associated with more intense toxicity since the objective is to improve the curabil- ity of the disease.
Our systematic had some limitations, firstly this is not a meta-analysis focusing on one particular outcome, with the inherent bias analysis or sensitivity analysis. The limited num- ber of published trials allows only a narrative description of the trial and a study-by-study discussion of the key-findings.
In the present review, we have summarized evidence based on fully published data. However, experts would be well aware that during the 2019 ASCO annual congress, the results of a large phase 3 trial assessing maintenance therapy with olaratumab were presented [25]. The primary aim of the ANNOUNCE Trial (NCT02451943) is to compare OS in an ITT population of advanced sarcoma patients. In this trial, patients have been randomly assigned to receive doxorubi- cin alone versus doxorubicin plus olaratumab followed by continuation maintenance therapy with olaratumab alone; 509 patients have been enrolled in the trial. The results are disappointing since the median OS was 20.4 versus 19.8 months (HR ¼ 1.05, 95% CI: 0.84–1.30; p ¼ .69) and the PFS was lower in the investigational arm in the ITT popula- tion (5.4 versus 6.8 m; HR ¼ 1.23, 95% CI: 1.01–1.50; p ¼ .04).
These disappointing results did not confirm the results of the previously published randomized phase 2 trial that sug- gested a major benefit of olaratumab addition and mainten- ance therapy [25]. In the randomized phase 2 trial, 133 patients were enrolled, and the addition of and maintenance therapy with olaratumab was associated with a clinically rele- vant improvement in PFS (6.6 versus 4.1 months, p ¼ .0615) and a statistically significant improvement in OS (26.5 versus 14.7 months, p ¼ .0003) [26]. The difference in OS observed in the phase 2 trial could be explained by the imbalance in histological subtypes between both arms: tumor growth in the experimental arms was much slower. Nevertheless, we wish to stress that olaratumab, a PDGFR-a-blocking antibody, failed to demonstrate clinical activity in 14 GIST cases with PDGFR-a mutation [27].
Drug holiday or drug discontinuation has been explored in two settings: trabectedin treatment for metastatic soft-tis- sue sarcoma and imatinib treatment for GIST. In both set- tings, drug discontinuation led to rapid disease progression; however, there was no deleterious impact on OS. The main objective of a drug holiday is to reduce treatment-related toxicity and improve the quality of life. None of these 2 trials formally reported a quality-of-life analysis. Drug holidays could be assessed in other clinical settings; however, advanced sarcomas with a long-term stable disease status under treatment are rather rare. Therefore, at the time, regarding the aggressiveness of sarcomas, there were no strong data to support drug holidays in these clin- ical settings.
In conclusion, in this review, we summarized the findings of ongoing trials assessing strategies of maintenance therapy (See Supplementary Appendix). Most of the trials assessed the role of maintenance therapy with a multikinase inhibitor (e.g., pazopanib and regorafenib) or a chemotherapy agent without cumulative toxicity (e.g., trabectedin).

References

[1] Moher D, Liberati A, Tetzlaff J, et al. preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLOS Med. 2009;6(7):e1000097.
[2] Mu€ller CR, Smeland S, Bauer HC, et al. Interferon-alpha as the only adjuvant treatment in high-grade osteosarcoma: long term results of the Karolinska Hospital series. Acta Oncol. 2005;44(5): 475–480.
[3] Bielack SS, Smeland S, Whelan JS, et al. Methotrexate, doxorubi- cin, and cisplatin (MAP) plus maintenance pegylated interferon alfa-2b versus MAP alone in patients with resectable high-grade osteosarcoma and good histologic response to preoperative MAP: first results of the EURAMOS-1 good response randomized controlled trial. JCO. 2015;33(20):2279–2287.
[4] Kleinerman ES, Jia SF, Griffin J, et al. Phase II study of liposomal muramyl tripeptide in osteosarcoma: the cytokine cascade and monocyte activation following administration. JCO. 1992;10(8): 1310–1316.
[5] Meyers PA, Schwartz CL, Krailo MD, et al. Osteosarcoma: the add- ition of muramyl tripeptide to chemotherapy improves overall survival – a report from the Children’s Oncology Group. JCO. 2008;26(4):633–638.
[6] Meyers PA, Schwartz CL, Krailo M, et al. Osteosarcoma: a random- ized, prospective trial of the addition of ifosfamide and/or mur- amyl tripeptide to cisplatin, doxorubicin, and high-dose methotrexate. JCO. 2005;23(9):2004–2011.
[7] Penel N, Adenis A, Bocci G. Cyclophosphamide-based metro- nomic chemotherapy: after 10 years of experience, where do we stand and where are we going. Crit Rev Oncol Hematol. 2012; 82(1):40–50.
[8] Hill SA, Lonergan SJ, Denekamp J, et al. Vinca alkaloids: anti-vas- cular effects in a murine tumour. Eur J Cancer. 1993;29(9): 1320–1324.
[9] Casanova M, Ferrari A, Bisogno G, et al. Vinorelbine and low-dose cyclophosphamide in the treatment of pediatric sarcomas: pilot study for the upcoming European Rhabdomyosarcoma Protocol. Cancer. 2004;101(7):1664–1671.
[10] Minard-Colin V, Ichante J-L, Nguyen L, et al. Phase II study of vinorelbine and continuous low doses cyclophosphamide in chil- dren and young adults with a relapsed or refractory malignant solid tumour: good tolerance profile and efficacy in rhabdomyo- sarcoma—a report from the Societe Francaise des Cancers et Leucemies de l’Enfant et de l’Adolescent (SFCE). Eur J Cancer. 2012;48(15):2409–2416.
[11] Bisogno G, De Salvo GL, Bergeron C, et al. Vinorelbine and con- tinuous low-dose cyclophosphamide as maintenance chemother- apy in patients with high-risk rhabdomyosarcoma (RMS 2005): a multicentre, open-label, randomised, phase 3 trial. Lancet Oncol. 2019;20(11):1566–1575.
[12] Klingebiel T, Boos J, Beske F, et al. Treatment of children with metastatic soft tissue sarcoma with oral maintenance compared to high dose chemotherapy: report of the HD CWS-96 trial. Pediatr Blood Cancer. 2008;50(4):739–745.
[13] D’Adamo DR, Anderson SE, Albritton K, et al. Phase II study of doxorubicin and bevacizumab for patients with metastatic soft- tissue sarcomas. JCO. 2005;23(28):7135–7142.
[14] Ray-Coquard IL, Domont J, Tresch-Bruneel E, et al. Paclitaxel given once per week with or without bevacizumab in patients with advanced angiosarcoma: a randomized phase II trial. JCO. 2015;33(25):2797–2802.
[15] Chisholm JC, Merks JHM, Casanova M, et al. Open-label, multi- centre, randomised, phase II study of the EpSSG and the ITCC evaluating the addition of bevacizumab to chemotherapy in childhood and adolescent patients with metastatic soft tissue sar- coma (the BERNIE study). Eur J Cancer. 2017;83:177–184.
[16] Chawla SP, Staddon AP, Baker LH, al. Phase II study of the mammalian target of rapamycin inhibitor ridaforolimus in patients with advanced bone and soft tissue sarcomas. JCO. 2012;30(1): 78–84.
[17] Demetri GD, Chawla SP, Ray-Coquard I, et al. Results of an inter- national randomized phase III trial of the mammalian target of rapamycin inhibitor ridaforolimus versus placebo to control meta- static sarcomas in patients after benefit from prior chemotherapy. JCO. 2013;31(19):2485–2492.
[18] Pacey S, Ratain MJ, Flaherty KT, et al. Efficacy and safety of sorafenib in a subset of patients with advanced soft tissue sarcoma from a Phase II randomized discontinuation trial. Invest New Drugs. 2011;29(3):481–488.
[19] Penel N, Decoupigny E, Tresch-Bruneel E, et al. Exceptional long- lasting clinical benefit of trabectedin in a patient with metastatic undifferentiated pleomorphic sarcoma: postscriptum from the T- Dis trial. Am J Clin Oncol. 2019;42(9):737.
[20] Le Cesne A, Blay JY, Domont J, et al. Interruption versus continu- ation of trabectedin in patients with soft-tissue sarcoma (T-DIS): a randomised phase 2 trial. Lancet Oncol. 2015;16(3):312–319.
[21] Kotecki N, Le Cesne A, Tresch-Bruneel E, et al. Impact of trabecte- din interruption and subsequent rechallenge on progression in patients with advanced soft tissue sarcoma: Long-term follow-up of the T-DIS trial. Am J Clin Oncol. 2018;41(11):1094–1100.
[22] Blay JY, Le Cesne A, Ray-Coquard I, et al. Prospective multicentric randomized phase III study of imatinib in patients with advanced gastrointestinal stromal tumors comparing interruption versus continuation of treatment beyond 1 year: the French Sarcoma Group. JCO. 2007;25(9):1107–1113.
[23] Le Cesne A, Ray-Coquard I, Bui BN, et al. Discotinuation of imatinib in patients with advanced gastrointestinal stromal tumours after 3 years of treatment: an open-label multicentre randomised phase 3 trial. Lancet Oncol. 2010;11(10):942–949.
[24] Agulnik M, Yarber JL, Okuno SH, et al. Open-label, multicenter, phase II study of bevacizumab for the treatment of angiosarcoma and epithelioid hemangioendotheliomas. Ann Oncol. 2013;24(1): 257–263.
[25] Tap WD, Wagner AJ, Papai Z, et al. ANNOUNCE: a randomized, placebo (PBO)-controlled, double-blind, phase (Ph) III trial of doxorubicin (dox) þ olaratumab versus dox þ PBO in patients (pts) with advanced soft tissue sarcomas (STS). JCO. 2019; 37(18_suppl):LBA3–LBA3.
[26] Tap WD, Jones RL, Van Tine BA, et al. Olaratumab and doxorubi- cin versus doxorubicin alone for treatment of soft-tissue sarcoma: an open-label phase 1b and randomised phase 2 trial. Lancet. 2016;388(10043):488–497.
[27] Wagner AJ, Kindler H, Gelderblom H, et al. A phase II study of a human anti-PDGFRa monoclonal antibody (olaratumab, IMC-3G3) in previously treated patients with metastatic gastrointestinal stromal tumors. Ann Oncol. 2017;28(3):541–546.