Autologous haematopoietic stem cell transplantation following high-dose chemotherapy for non-rhabdomyosarcoma soft tissue sarcomas: a Cochrane systematic review*

Abstract

Objectives

We conducted a systematic review to compare the efficacy and adverse events of autologous haematopoietic stem cell transplantation (HSCT) following high-dose chemotherapy (HDCT) versus standard-dose chemotherapy (SDCT) in patients with locally advanced or metastatic non-rhabdomyosarcoma soft tissue sarcomas (NRSTS).

Setting

Patients were observed in hospital units specialised for cancer therapy.

Participants

The review evaluated 294 patients with 19 different subtypes of malignant NRSTS. The patients had a median age between 10 and 46 years (range 2–65) and were mostly men.

Primary and secondary outcome measure

The planned and measured primary outcomes were overall survival and treatment-related mortality. The planned and measured secondary outcomes were progression-free survival, grade 3–4 non-haematological toxicity and secondary neoplasia. Other secondary outcomes including disease-free survival, event-free survival and health-related quality of life were not reported.

Results

We included 62 studies reporting on 294 transplanted patients. We identified 1 randomised controlled trial (RCT) with 38 transplanted and 45 non-transplanted patients and judged a low risk of bias. We further identified 61 single-arm studies with 256 transplanted patients. Overall survival in the RCT was reported not statistically significantly different between autologous HSCT following HDCT versus SDCT. The HR was 1.26 (95% CI 0.70 to 2.29; p=0.44) and the point estimates at 3 years were 32.7% vs 49.4%. Data from single-arm studies were used to extract data on adverse events. Treatment-related mortality was reported in 5.1% (15 of 294) transplanted patients.

Conclusions

Overall survival in patients with locally advanced or metastatic NRSTS was not statistically different after autologous HSCT following HDCT compared with SDCT in a single RCT with a total of 83 patients. No other comparative study was available. The proportion of adverse events among the transplanted patients is not clear.

Strengths and limitations of this study.

• We conducted a comprehensive literature search and strictly adhered to the projected methodology. The WHO classification of soft tissue sarcomas was adopted and modified to define a clear terminology for the study selection process.

• We judged a low risk of bias for the single identified randomised controlled trial, which may serve as the major relevant evidence.

• Single-arm studies provided some estimation about serious adverse events with transplantation.

• Some treatments were performed 10–20 years ago. Thus, the results may not be applicable to patients who are treated today.

• The included studies report various subtypes of non-rhabdomyosarcoma soft tissue sarcomas, and each tumour type may carry an individual risk profile and, therefore, ideally should be evaluated separately.

Introduction

Soft tissue sarcomas (STS) are a highly heterogeneous group of rare malignant solid tumours of non-epithelial extraskeletal body tissue and are classified on a histogenetic basis.¹ The location of the primary tumour can involve any area of the body.² STS can involve any type of tissue and typically affect muscles, tendons, adipose tissue, blood vessels and joints and commonly present as a painless mass.³ In this review, we investigated non-rhabdomyosarcoma STS (NRSTS) provided that they are categorised as malignant according to the WHO 2002 classification.⁴ In Western countries about four new cases of NRSTS are estimated per 100 000 population every year, with the Ewing family of tumours excluded from this statistic.⁵

Surgery is the standard treatment for localised NRSTS and can be curative if distant dissemination is not present.^{6 7} Chemotherapy is regarded mainly as a palliative treatment for high-risk patients who are characterised by inoperable, locally advanced and metastatic disease.⁶ Riedel⁸ provides an overview of current systemic therapies and discusses possible novel therapeutic agents and treatment strategies. High-dose chemotherapy (HDCT) has been evaluated as an alternative treatment option for high-risk patients. The rationale for HDCT is that escalating doses of HDCT may increase survival by capturing putatively remnant malignant cells.⁹ The rationale for autologous haematopoietic stem cell transplantation (HSCT) following HDCT is a planned rescue for HDCT-related severe haematological toxicity.⁹ The primary objective of the present systematic review is to evaluate the effectiveness and adverse events of autologous HSCT following HDCT in patients with advanced or metastatic NRSTS.

Methods

This article is based on a Cochrane systematic review published in The Cochrane Library.¹⁰ Publication of this work is in agreement with the policy of The Cochrane Collaboration.¹¹ While preparing this systematic review, we endorsed the PRISMA statement, adhered to its principles and conformed to its checklist.¹²

Study inclusion criteria

We included patients with NRSTS provided that they are categorised as malignant according to the WHO 2013 classification on STS⁴ as well as malignant haemangiopericytoma and anaplastic sarcoma. We excluded the Ewing family of tumours according to the European Society for Medical Oncology (ESMO) Guidelines Working Group,⁵ chondrosarcomas, osteosarcomas and rhabdomyosarcomas. While writing the Cochrane Review, we referred to the WHO 2002 classification.¹³ For the purpose of the present systematic review, we updated the inclusion criteria and re-evaluated the potentially relevant studies and included the following entities: ‘Gastrointestinal Stromal Tumours’, ‘Malignant peripheral nerve sheath tumour’, ‘Undifferentiated pleomorphic sarcoma not otherwise specified’. Almost all published studies refer to the 2002 classification. Thus, we continued to include the following entities though they were removed and relocated within the 2013 classification: ‘malignant fibrous histiocytoma’ (MFH), ‘undifferentiated sarcoma’, ‘unclassified sarcoma’, and ‘haemangiopericytoma’. Table 1 compares the categories and malignant subtypes of the 2013 versus the 2002 edition of the WHO classification of tumours of soft tissue and indicates which of those are included in the present systematic review. Participants were included regardless of age, severity and clinical stage of disease. Studies were included as long as at least 80% of patients had NRSTS and received the test intervention. The test intervention was autologous HSCT following HDCT containing stem cells from peripheral blood or bone marrow. The comparator was standard-dose chemotherapy (SDCT). The primary outcomes were overall survival and treatment-related mortality (TRM). Secondary outcomes were disease-free survival, progression-free survival, event-free survival, non-haematological toxicity grades 3–4,¹⁴ secondary malignant neoplasia and health-related quality of life.

Table 1.

Inclusion of malignant soft tissue tumours of the WHO classification 2013 vs 2012

Category	Malignant subtypes	2013	2002	Inclusion
Adipocytic tumours		2013	2002	Included
	Dedifferentiated liposarcoma	2013	2002	Included
	Myxoid liposarcoma	2013	2002	Included
	Pleomorphic liposarcoma	2013	2002	Included
	Liposarcoma, not otherwise specified	2013	2002	Included
	Round cell liposarcoma	No	2002	Included
	Mixed-type liposarcoma	No	2002	Included
Fibrobastic/myofibroblastic tumours		2013	2002	Included
	Adult fibrosarcoma	2013	2002	Included
	Myxofibrosarcoma	2013	2002	Included
	Low-grade fibromyxoid sarcoma	2013	2002	Included
	Sclerosing epitheloid fibrosarcoma	2013	2002	Included
	Malignant haemangiopericytoma	No	No	Included
So-called fibrohistiocytic tumours		2013	2002	Included
	Pleomorphic ‘MFH'/(UPS)	No	2002	Included
	Giant cell ‘MFH'/UPS with giant cells	No	2002	Included
	Inflammatory ‘MFH'/UPS with prominent inflammation	No	2002	Included
Smooth muscle tumours		2013	2002	Included
	Leiomyosarcoma (excluding skin)	2013	2002	Included
Pericytic (perivascular) tumours		2013	2002	No
Skeletal muscle tumours		2013	2002	No
	Embryonal rhabdomyosarcoma	2013	2002	No
	Alveolar rhabdomyosarcoma	2013	2002	No
	Pleomorhic rhabdomyosarcoma	2013	2002	No
	Spindle cell/sclerosing rhabdomyosarcoma	2013	No	No
Vascular tumours of soft tissue		2013	2002	Included
	Epithelioid haemangioendothelioma	2013	2002	Included
	Angiosarcoma of soft tissue	2013	2002	Included
Chondro-osseous tumours		2013	2002	No
	Mesenchymal chondrosarcoma	2013	2002	No
	Extraskeletal osteosarcoma	2013	2002	No
Gastrointestinal stromal tumours		2013	No	Included
	Gastrointestinal stromal tumour, malignant	2013	No	Included
Nerve sheath tumours		2013	No	Included
	Malignant peripheral nerve sheath tumour	2013	No	Included
	Epithelioid malignant peripheral nerve sheath tumour	2013	No	Included
	Malignant Triton tumour	2013	No	Included
	Malignant granular cell tumour	2013	No	Included
	Ectomesenchymoma	2013	No	Included
Tumours of uncertain differentiation		2013	2002	Included
	Synovial sarcoma NOS	2013	2002	Included
	Epithelioid sarcoma	2013	2002	Included
	Alveolar soft-part sarcoma	2013	2002	Included
	Clear cell sarcoma of soft tissue	2013	2002	Included
	Extraskeletal myxoid chondrosarcoma	2013	2002	No
	Extraskeletal Ewing sarcoma	2013	2002	No
	Desmoplastic small round cell tumour	2013	2002	Included
	Extrarenal rhabdoid tumour	2013	2002	Included
	Neoplasms with perivascular epithelioid cell differentiation	2013	2002	Included
	Intimal sarcoma	2013	2002	Included
	Malignant Mesenchymoma	No	2002	Included
Undifferentiated/unclassified sarcomas		2013	No	Included
	Undifferentiated spindle cell sarcoma	2013	No	Included
	UPS	2013	No	Included
	Undifferentiated round cell sarcoma	2013	No	Included
	Undifferentiated epithelioid sarcoma	2013	No	Included
	Undifferentiated sarcoma NOS	2013	No	Included

MFH, malignant fibrous histiocytoma; NOS, not otherwise specified; UPS, undifferentiated pleomorphic sarcoma.

Search strategy, selection of studies and data extraction

We conducted an electronic literature database search in MEDLINE (Ovid), EMBASE (Ovid) and Cochrane Library CENTRAL (Wiley) including articles published from inception to an update search on 12 June 2014. The corresponding search strategies have been published in the corresponding Cochrane Review.¹⁰ We retrieved all titles and abstracts by electronic searching and downloaded them to the reference management database EndNote V.X3.¹⁵ We considered studies written in languages other than English. We searched the online registries^{16 17} on 12 June 2014 for additional completed or ongoing studies using the search strategy "sarcoma AND chemotherapy AND transplantation". We searched all retrieved abstracts of annual meetings contained in EMBASE (Ovid). We contacted authors to replenish missing information. All data assessments were performed independently by two independent review authors. We resolved differences by discussion or by appeal to a third review author. We judged whether the autologous HSCT following HDCT could be regarded as a consolidation or a salvage therapy. A consolidation therapy is a treatment that is given after cancer has disappeared following the initial therapy and a salvage therapy is a treatment that is given after the cancer has not responded to other treatments.¹⁸ We considered a consolidation therapy if the status at transplantation was either a complete or a partial response to the preceding therapy and we considered a salvage therapy if the status was less favourable and in case a relapse was described.

Assessment of risk of bias in included studies

We have used four criteria from The Cochrane Collaboration’s tool for assessing risk of bias in randomised controlled trials (RCTs)¹⁹: random sequence generation, allocation concealment, blinding of outcome assessment and selective reporting such as not reporting prespecified outcomes. We extended the Cochrane tool for assessing risk of bias by five criteria that consider non-randomised studies: prospective design, comparable baseline characteristics, assignment of patients to treatment groups, concurrent control, and loss to follow-up. We applied The Cochrane Collaboration's criteria for judging risk of bias.²⁰

Data synthesis

We synthesised aggregate data as narrative because data were too scarce to be pooled. Differently from the Cochrane Review, we did not pool time-to-event data on overall survival from studies with individual data. With respect to survival data, we accepted time of diagnosis and beginning of treatment as starting points. We evaluated all 62 studies to search for reports on TRM and tabulated the identified patient data. We evaluated the seven studies reporting aggregate data to search for reports on grade 3–4 non-haematological toxicity in the autologous HSCT following the HDCT arm and tabulated the identified event data.

Results

Search results

Figure 1 shows the literature search and study flow. We retrieved 1035 records and evaluated 260 full text papers in detail. We included 62 studies with 294 transplanted patients, 1 RCT with 38 transplanted and 45 non-transplanted patients,²¹ 6 single-arm studies reporting aggregate case series data,^22–27 and 55 single-arm studies with individual data. In online registries, we identified 6 studies with a still pending completion and we did not find additional studies in the update search.

Figure 1.

Literature search and study flow.

Baseline data

We provide an overview of the main characteristics of studies and treatment (table 2), of the patients (table 3) and of the frequency of the identified subtypes (table 4). The one RCT was an open, multicenter and randomised phase III study with two parallel treatment groups.²¹ Patients were eligible for randomisation if they had responded to chemotherapy or, for stable disease, if a complete surgical resection of all disease sites could be carried out. The intention-to-treat principle was modified to exclude patients found to be ineligible at a histological review after randomisation. Three of the six single-arm studies reporting aggregate case series data collected the data prospectively^22–24 and three retrospectively.^25–27 Data from the remaining 55 single-arm studies were considered for the description of TRM only.

Table 2.

Characteristics of studies and therapy

Study	Number of centers (country)	Enrolment; years	Prospective design	Autologous HSCT following HDCT
				Drugs	Consolidation vs salvage vs NR; N	PBSCT vs BMT vs NR; N
Aggregate comparative data
Bui-Nguyen et al21	16 (France)	2000–2008	Yes	Ca-Et-If	38 vs 0 vs 0	38 vs 0 vs 0
Aggregate case series data
Bertuzzi et al22	1 (Italy)	1997–2002	Yes	Me-Mi-Th	10 vs 0 vs 0	10 vs 0 vs 0
Bisogno et al23	>1 (Italy)	1999–2008	Yes	Cy-Me-Th	14 vs 0 vs 0	14 vs 0 vs 0
Blay et al24	1 (France)	1988–1994	Yes	Ci-Et-If	0 vs 0 vs 24	0 vs 0 vs 24
Bokemeyer et al25	3 (Germany)	NR	No	Do-If	16 vs 0 vs 0	16 vs 0 vs 0
Cook et al26	29 (USA)	1999–2007	No	Ca-Cy-Et-Me-Th	0 vs 0 vs 36	33 vs 2 vs 1
Philippe-Chomette et al27	>1 (France)	1995–2006	No	Various	14 vs 0 vs 0	0 vs 0 vs 14
Individual cases data
55 studies (142 patients)	Various	Various	No	Various	69 vs 61 vs 12	102 vs 21 vs 19

BMT, bone marrow transplant; Ca, carboplatin; Ci, cisplatin; Cy, cyclophosphamide; Do, Doxorubicin; Et, etoposide=Vepesid=VP 16; HDCT, high-dose chemotherapy; HSCT, autologous haematopoietic stem cell transplantation; If, ifosfamide; Me, melphalan; Mi, mitoxantrone; N, number; NR, information not reported in the article; PBSCT, peripheral blood stem cell transplant; Th, thiotepa.

Table 3.

Characteristics of patients

Study	Patients analysed; N		FU	Subtypes	Age; median years (range)		Gender; % males
	HSCT	SDCT			HSCT	SDCT	HSCT	SDCT
Aggregate comparative data
Bui-Nguyen et al21	38	45	55 (NR)	Various	46 (19 to 65)	43 (18 to 65)	58	50
Aggregate case series data
Bertuzzi et al22	10	NA	35 (14 to 60)	DSRCT	29 (NR)	NA	100	NA
Bisogno et al23	14	NA	27 (NR)	DSRCT	10 (2 to 17)	NA	93	NA
Blay et al24	24	NA	NR	Various	NR	NA	NR	NA
Bokemeyer et al25	16	NA	NR	Various	45 (25 to 57)	NA	NR	NA
Cook et al26	36	NA	44 (4 to 89)	DSRCT	19 (8 to 46)	NA	80	NA
Philippe-Chomette et al27	14	NA	23 (9 to 51)	DSRCT	NR (4 to 29)	NA	86	NA
Individual cases data
55 studies	142	NA	Various	Various	25 (1 to 65)	NA	NR	NA

DSRCT, desmoplastic small-round cell tumour; FU, follow-up of the analysed patients in median months (range); HSCT, autologous haematopoietic stem cell transplantation following high-dose chemotherapy; N, number; NA, not applicable; NR, information not reported in the article; SDCT, standard-dose chemotherapy.

Table 4.

Frequency of subtypes

Subtype	All	Aggregate	Individual
Anaplastic sarcoma	5	0	5
Angiosarcoma	10	4	6
Clear cell sarcoma	2	1	1
Desmoplastic small round cell tumour	109	74	35
Epitheloid sarcoma	2	0	2
Fibrosarcoma	6	1	5
Fibromyosarcoma	1	0	1
Leiomyosarcoma	29	14	15
Liposarcoma	15	8	7
Mesenchymal sarcoma	2	2	0
Malignant fibrous histiocytoma	31	13	18
Malignant haemamgiopericytoma	8	5	3
Malignant peripheral nerve sheath tumour	4	0	4
Rhabdoid tumour, extrarenal, extracerebral	2	0	2
Spindle cell sarcoma	1	0	1
Synovial sarcoma	32	9	23
Unclassified sarcoma	17	12	5
Undetermined sarcoma	13	4	9
Not NRSTS	5	5	0
Total number	294	152	142

NRSTS, non-rhabdomyosarcoma soft tissue sarcomas.

The 62 studies were set in 13 different countries in four different continents. Most of the transplanted patients were studied in France, USA and Germany. We assume that most patients in the studies reporting aggregate case series data received autologous HSCT following HDCT as a consolidation therapy, whereas a considerable number of the individual case data were associated with autologous HSCT following HDCT as a rescue therapy. The majority of all studies used peripheral blood stem cell transplants. Median age varied roughly between 19 and 46 years and there was a male preponderance. Patients had 19 different relevant histological diagnoses. Most patients had desmoplastic small round-cell tumour (N=109 of 294) followed by the new category of undifferentiated pleomorphic sarcomas (N=61), which is composed of MFH (N=31), unclassified sarcoma (N=17) and undetermined sarcoma (N=13).

Primary outcome

Overall survival was not statistically significantly different in the RCT by Bui-Nguyen et al²¹ between autologous HSCT following HDCT versus SDCT regarding the HR of 1.26 (95% CI 0.70 to 2.29; p=0.44; table 5). In this RCT, the point estimates at 3 years were 32.7% vs 49.4% based on 8 vs 17 remaining patients at risk. The patients at risk at baseline were 38 vs 45 patients. With respect to the studies reporting aggregate case series data, overall survival for transplanted patients ranged roughly from 20% to 51% at 2 years and from 32% to 40% at 3 years (table 5). In 10 studies, TRM was associated with 15 of 137 evaluated patients (table 6). Assuming no other TRM in the remaining 157 patients, a risk o procedure-related death might be estimated as 5.1% (15 of 294).

Table 5.

Overall survival in studies reporting aggregate data

Study	Overall survival (95% CI), point estimates			Statistics
	HSCT at 2 years	HSCT at 3 years	SDCT at 3 years
Aggregate comparative data
Bui-Nguyen et al21		32.7%	49.4%	HR 1.26 (0.70 to 2.29), p=0.44
Aggregate case series data
Bertuzzi et al22	20%	NR	NA
Bisogno et al23	48%	38.9%	NA
Blay et al24	NR	NR	NÁ
Bokemeyer et al25	Median 13 months, range 3–19		NA
Cook et al26	NR	40% (24% to 58%)	NA
Philippe-Chomette et al27	51.4% (23.2% to 79.6%)	NR	NA

Some estimates were deduced from Kaplan-Meier plots.

HSCT, autologous haematopoietic stem cell transplantation following high-dose chemotherapy; NA, not applicable; NR, not reported, SDCT, standard-dose chemotherapy.

Table 6.

Treatment-related mortality in the HSCT arm of all included studies

Study	N affected/ N evaluated patients	Specification
Treatment-related mortality
Bui-Nguyen et al21	1/38	Treatment-related leukaemia death 2 years after HSCT
Cook et al26	2/36	NR
Doros et al41	1/1	NR
Engelhardt et al42	3/24	Sepsis (N=2); pneumonia related to lung metastases (N=1)
Kasper et al43	1/14	Cardiac arrest of unknown cause
Matsuzaki et al44	1/1	Multiple organ failure
Navid et al45	1/2	Liver as well as kidney failure
Philippe-Chomette et al27	1/14	Died of treatment toxicity 12 months after HSCT
Saab et al46	2/4	Acute myocardial infarction (N=1); veno-occlusive disease (N=1)
Slease et al47	2/3	Progressive encephalopathy (N=1); sepsis (N=1)
Total	15/137

HSCT, autologous haematopoietic stem cell transplantation following high-dose chemotherapy; N, number; NR, not reported.

Secondary outcomes

Progression-free survival was also not statistically significantly different in the RCT by Bui-Nguyen et al²¹ between autologous HSCT following HDCT versus SDCT regarding the HR of 1.34 (95% CI 0.81 to 2.20; p=0.25). In this RCT, the point estimates at 3 years were 9.3% vs 21.6% based on 3 vs 12 remaining patients at risk. The RCT did not report results on disease-free survival and event-free survival. An overview of the number of events of non-haematological toxicity grade 3–4 is provided in table 7. In the RCT, 11 events were observed in 38 transplanted patients and 1 event (asthenia) was reported regarding the SDCT arm. In 3 of the studies reporting aggregate case series data, 25 events were observed in 54 transplanted patients in the HSCT arm. The other 3 studies did not report toxicity data. We identified one secondary neoplasia in a single case report. Health-related quality of life scales were not addressed in the included studies.

Table 7.

Grade 3–4 NCI-CTCAE non-haematological toxicity in the HSCT arm of studies reporting aggregate case series data

Study	N events/ N evaluated patients	Specification
Aggregate comparative data
Bui-Nguyen et al21	11/38	Digestive (N=8); infection (N=2); pain (N=1)
Aggregate case series data
Bertuzzi et al22	NR	NA
Bisogno et al23	1/14	Mucositis grade 4
Blay et al24	16/24	Neurological grade 4 (N=1); lung grade 3/4 (N=2); renal grade 3/4 (N=5); nausea/vomiting grade 3/4 (N=8)
Bokemeyer et al25	8/16	No grade 4; neurological (N=1); renal (N=2); infection (N=1); mucositis (N=2); nausea/emesis (N=2)
Cook et al26	NR	NA
Philippe-Chomette et al27	NR	NA

HSCT, autologous haematopoietic stem cell transplantation following high-dose chemotherapy; N, number; NA, not applicable; NR, not reported; NCI-CTCAE, National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) grade III–IV.¹⁴

Data quality

Clinical heterogeneity was substantial because tumour subdiagnosis varied considerably between patients. Furthermore, tumour stage and metastasis were not reported for all participants. The RCT by Bui-Nguyen et al²¹ stands out as it is the only study reporting comparative data. We judged a low risk of bias for this trial for random sequence generation and selective reporting. However, the trial does have some drawbacks. We judged an unclear risk for allocation concealment because masking of allocation was not described in full detail. We judged a high risk of bias for blinding of outcome assessment because it was not reported for any outcome. The other 61 of 62 studies are single-arm studies and are therefore not qualified for assessing a treatment effect.

Discussion

Outcomes

We identified one RCT comparing autologous HSCT following HDCT versus SDCT.²¹ The authors reported a difference in overall survival and progression-free survival after the treatment in favour of SDCT but the difference was not statistically significant, respectively. Therefore, there is evidence that patients may not have a better survival after autologous HSCT following HDCT versus SDCT. If at all, this intervention should only be offered after careful consideration and preferably only within a randomised controlled clinical trial. We estimated a TRM of 5.1%, which was somewhat higher than the 2% reported by others.²⁸ Severe toxicity grade 3–4 was sparsely reported. Studies on health-related quality of life were not identified. The frequency of secondary neoplasia in 1 of 294 participants is probably an extreme underestimation of the true frequency due to a relatively short follow-up. The detection of secondary neoplasia depends on a long follow-up and was estimated from 4% to 6.9% by others.^{29 30}

The WHO 2013 classification

The WHO recently published the 2013 classification on STS.⁴ The authors inserted the category ‘Undifferentiated Pleomorphic Sarcoma Not Otherwise Specified’ to lodge those types of STS that are difficult to classify using the current available techniques.^{31 32} The authors integrated the terms ‘MFH’, ‘Undifferentiated Sarcoma’, and ‘Unclassified Sarcoma’ into this newly created category. MFH was characterised by an apparent lack of specific differentiation³³ and it was considered a diagnosis of exclusion.³⁴ MFH was regarded as the most common soft tissue sarcoma of adulthood³³ and accounted for up to 25% of patients in clinical trials on soft tissue sarcoma.³⁴ In 1992, Fletcher³³ reassessed 159 cases with MFH and found 63% (97 of 159) tumours to be specific sarcomas other than MFH. In 2001, Fletcher et al³⁵ confirmed that 84% (84 of 100) tumours of patients with MFH showed sufficient differentiation to assign them to specific subtypes of STS. The techniques to assess cell differentiation have been substantially improved with the effect that the frequency of the tumour within this category has decreased.³⁶ It was supposed that the category of ‘Undifferentiated Sarcoma—Otherwise Not Specified’ may contain liposarcoma, fibrosarcoma, leiomyosarcoma, rhabdomyosarcoma, other sarcomas, and even carcinomas or lymphomas.^{36 37} It was estimated that the pathologist might have difficulties in identifying a specific differentiation in 10–15% of tumours previously called MFH.³⁷ The new edition also removed the term ‘Haemangiopericytoma’.^{31 32} ‘Gastrointestinal Stromal tumours’ and ‘Nerve Sheath tumours’ were relocated from other classifications and appear for the first time in the soft tissue classifications.^{31 32} Consequently, the term ‘Malignant Peripheral Nerve Sheath Tumour’ is newly integrated.

Strengths and limitations

The search strategy had the broad aim of the retrieval of all relevant studies. With respect to historical versions of the Cochrane Review,¹⁰ we applied two different search strategies and retrieved the same studies with aggregate data but different studies with individual cases data. These results show the substantial difficulty associated with the aim of searching for all published cases. This enterprise appears almost impossible. We adopted the new WHO 2013 classification of STS and made minor modifications to define a clear terminology for the study selection process. The group of NRSTS consists of many subtypes that are difficult to diagnose and separate even today. A considerable number of tumours cannot be clearly assigned to a specific histological category. Thus, we may have tumours with a specific label that might not be true. Otherwise, we may have tumours without a specific label that might belong to a specific category. We excluded studies if the proportion of non-eligible participants were greater or equal to 20% of the total population to prevent a significant mixture with disease or interventions that are not included in the present review. Authors were contacted to ask for additional data. We judged a low risk of bias for the one identified RCT, which may serve as the major relevant evidence. All other identified studies were single-arm studies that are not helpful in deciding whether autologous HSCT following HDCT for NRSTS is a meaningful treatment option. Therefore, we removed the survival data of studies reporting individual data. Nevertheless, they provided data for estimation about TRM within all included transplanted patients. We also removed data on non-haematological toxicity of studies reporting individual data because the sparse reporting might have caused a display of non-representative information. The description of consolidation and salvage therapy is based on our judgement and might be judged differently by others. These types of therapy were not precisely reported in most studies. Some treatments were performed 10–20 years ago. Thus, the results may not be applicable to patients who are treated today. All studies report various subtypes of NRSTS and each tumour type may carry an individual risk profile and, therefore, ideally should be evaluated separately. With respect to the individual survival data, follow-up started at different time points, that is, at diagnosis or at start of treatment. The delay between diagnosis and starting HDCT can be considerable.

Other findings and opinions

We wish to point out that some authors have warned against the use of autologous HSCT following HDCT, indicating the possibility of repositioning of malignant cells.³⁸ Others have questioned the use of HDCT with reference to the potential existence of refractory cancer stem cells.⁹ Pedrazzoli et al³⁹ stated that the potential benefit of this treatment option has not been investigated sufficiently in comparative studies. Kasper et al⁴⁰ concluded that the use of HDCT for locally advanced or metastatic adult (soft tissue and bone) sarcomas still remains highly investigational and should not be performed outside clinical trials. The identified RCT by Bui-Nguyen et al²¹ provides meaningful comparative data for the first time and its results questions any benefit of the intervention. Finally, we cannot close the chapter as it can be unsecure to rely on a single trial.

Conclusion

Overall survival in patients with locally advanced or metastatic NRSTS was not statistically different after autologous HSCT following HDCT compared with SDCT in a single RCT with a total of 83 patients. No other comparative study was available. A considerable number of patients were not evaluated concerning adverse events, and its proportion among the transplanted patients remains unclear. If this treatment is offered it should only be after careful consideration and only within an RCT.