Incidence, distribution of histological subtypes and primary sites of soft tissue sarcoma in China

Abstract

Objective

Soft tissue sarcomas (STSs) are rare malignancies deriving from mesenchyme. In this study, we reported the epidemiology of STS in China using population-based cancer registry data.

Methods

In 2017, qualified data from 339 cancer registries were included in the national database. All STS cases were retrieved based on the morphological and topographical codes of International Classification of Diseases for Oncology, and were categorized into different histological subtypes and primary sites accordingly. Nationwide new STS cases were estimated using incidence rate of STS and the national population, and were reported for gastrointestinal stromal tumor (GIST) and STSs other than GIST separately by sex and region. Distribution of histological subtypes and primary sites of STS were calculated, as well as primary sites of GIST.

Results

Approximately 39,900 new STS cases occurred nationwide in China in 2014, accounting for 1.05% of overall cancer incidence. The crude incidence rate was 2.91/100,000 and generally increased with age. An overall female predilection was found. GIST was the most common histological subtype, followed by nerve sheath tumor and malignant peripheral nerve sheath tumor, leiomyosarcoma, liposarcoma, and fibrosarcoma. About 67.5% of GIST occurred in stomach while 1.4% were recorded outside the gastrointestinal tract. Connective, subcutaneous and other soft tissues were the most common primary site, of which extremities were the major subsite.

Conclusions

The burden of STS is not serious in China relatively. However, due to their histological and topographical complexity, STSs should not be unnoticed, and more basic and clinical studies should focus on STSs.

Introduction

Sarcomas, consisted of soft tissue sarcomas (STSs) and bone sarcomas, are a heterogeneous group of mesenchymal malignancies that can develop at any age, comprising approximately 1% of all adult malignancies and 15% of pediatric malignancies¹. Compared with other common cancer types, STSs can develop at almost any anatomical sites², and are more prone to occur at childhood³. Although the etiology of STSs is still unclear, known risk factors include inherited syndromes, chemical and radiation exposures, viral infections and genetic mutations^4,5.

Due to their rarity, STSs are always outnumbered by carcinomas in many primary sites, therefore not reported in most researches categorizing cancers with the International Classification of Diseases⁶. In United States, it is estimated that 13,040 new cases and 5,150 deaths were caused by STSs in 2018, representing 0.75% of overall cancer incidence and 0.84% of overall cancer mortality⁷. In Europe, nearly 23,600 new STS cases rose annually and the crude incidence rate was 4.7 per 100,000⁶. In UK alone 3,300 new cases were diagnosed each year, with about 90 cases in children under 15 years oldd⁴. So far, most national and subnational epidemiological researches on the burden of STSs were carried out in developed countries^3,8-15.

Of all histological subtypes of STSs, gastrointestinal stromal tumor (GIST) was categorized as STS instead of digestive system cancer for the first time in the 2013 World Health Organization (WHO) Classification^5,16. Compared with other common cancer types in gastrointestinal tract, GIST starts in different types of cells and needs different types of treatment. Therefore, GIST was either treated as an entity equal to STSs and bone sarcomas^6,17, or reported separately from other STS subtypes^18-20 in some researches. The crude incidence rate of GIST was 1.1–1.4 per 100,000 according to French and Spanish researches^20,21. Stomach is the most common site for GIST²¹ while there are some extremely rare tumors that arise outside the gastrointestinal tract but show the features of GIST. These tumors were therefore named extra-GIST²². Similar to the condition of overall STSs, few researches on the epidemiology of GIST and extra-GIST were carried out in developing countries.

With world’s largest population, 23.7% of the global new cancer cases were estimated to occur in China in 2018²³. However, little is known about the incidence of STS in this country. In order to have a comprehensive understanding on the cancer burden of STS in China, in this study we described the distribution of histological subtypes and primary sites of STS using updated data from 339 population-based cancer registries. Estimation of the incidence of STS in the whole Chinese population was made as well and was reported for GIST and STSs other than GIST (non-GIST STSs) separately. We also described the distribution of GIST and explored the proportion of extra-GIST.

Materials and methods

Cancer registry data source

By 30th August 2017, 449 cancer registries from 31 provinces in China submitted cancer registry data for 2014 to National Central Cancer Registry of China (NCCRC). Local population data were provided along with the cancer data. After quality control based on the criteria of Chinese Guideline for Cancer Registration, data from 339 registries met the criteria and were finally included in this analysis. Among them, 129 were located in urban areas (covering a population of 144,061,915) and 210 were in rural areas (covering a population of 144,181,432). The overall population covered by these 339 cancer registries was 288,243,347 (146,203,891 men and 142,039,456 women), accounting for 21.07% of the national population that year.

All cancer cases were coded according to the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) and the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10). Of all STS cases, the proportion of morphology verified cases (MV%) was 84.97%. More information on the national cancer registry program and data quality control procedure could be found elsewhere^24,25.

Data extraction, classification and statistical analysis

STS cases in the 339 registries were retrieved and categorized into different histological groups according to their ICD-O-3 morphological codes as shown in Table 1 based on a previous study¹². The primary site of a case was defined by the ICD-O-3 topographical code. Cases with topographical codes being C40-C41 were deleted to exclude sarcomas in bones, joints and articular cartilages.

1.

Histological subtypes by ICD-O-3 morphological codes

Histological group	ICD-O-3 codes
NOS: not otherwise specified
Sarcoma NOS	M8800–8806, M8000–8004 located in C48.0–49.9
Gastrointestinal stromal tumor	M8936
Leiomyosarcoma	M8890–8896
Endometrial stromal sarcoma	M8930–8935
Liposarcoma	M8850–8858
Malignant fibrous histiocytoma	M8830
Angiosarcoma	M9120–9133, M9150 and M9170
Rhabdomyosarcoma	M8900–8920 and M8991
Fibrosarcoma	M8810–8815
Nerve sheath tumor and malignant peripheral nerve sheath tumor	M9540–9571
Dermatofibrosarcoma	M8832–8833
Other specified soft tissue sarcoma
Carcinosarcoma, NOS	M8980
Synovial sarcoma	M9040–9043
Mixed tumor, malignant NOS	M8940
Primitive neuroectodermal tumor, NOS	M9364 and M9473
Granular cell tumors and alveolar soft part sarcoma	M9580–9581
Paragangliomas and glomus tumors	M8680–8711
Malignant mesenchymoma	M8990
Malignant myoepithelioma	M8982
Clear cell sarcoma	M9044
Kaposi sarcoma	M9140
Rhabdoid tumor	M8963
Extraskeletal osteosarcoma and chondrosarcoma	M9180–9243
Myxosarcoma	M8840
Malignant giant cell tumors	M9251

Crude incidence rates were calculated by area (urban/rural), sex (male/female) and age group (0, 1–4, 5 to 80 by 5 years, 85+) based on the retrieved cases. To estimate the number of new STS cases in the whole country, incidence rates stratified by area, sex and age group were multiplied by corresponding nationwide population in each stratum. National population data were obtained from Nation Bureau of Statistics. Chinese population in 2000 and World Segi’s population were used for age-standardization.

The estimated nationwide incidence was reported for overall STSs, GIST, and non-GIST STSs separately. The distribution of histological subtypes was reported by sex and age group (0–19, 20–64, and over 65 years old) in order to reveal the sex and age predilection. Primary sites of all STS cases were classified into 14 groups and reported by sex.

Results

Incidence of soft tissue sarcoma

Approximately 39,900 new STS cases (19,200 men and 20,700 women) occurred nationwide China in 2014, as shown in Table 2. The crude incidence rate was 2.91/100,000 (2.72/100,000 in men and 3.11/100,000 in women), while the age-standardized rate by Chinese population (ASRcn) and world population (ASRwld) were 2.21 and 2.13 per 100,000, respectively.

2.

Estimated number of nationwide new cases of soft tissue sarcomas in China, 2014

	Areas	Gender	Number of new cases	Incidence rate (1/100,000)	ASRcn (1/100,000)	ASRwld (1/100,000)
STS, soft tissue sarcoma; GIST, gastrointestinal stromal tumor; ASRcn, Age –standardized incidence rate by China Population; ASRwld, Age –standardized incidence rate by Segi’s Population
Overall STS	All areas	Total	39,900	2.91	2.21	2.13
		Male	19,200	2.72	2.10	2.05
		Female	20,700	3.11	2.32	2.23
	Urban areas	Total	26,600	3.55	2.55	2.47
		Male	12,900	3.38	2.47	2.40
		Female	13,700	3.71	2.65	2.54
	Rural areas	Total	13,300	2.14	1.74	1.69
		Male	6,300	1.94	1.61	1.58
		Female	7,000	2.36	1.89	1.81
GIST	All areas	Total	5,700	0.42	0.29	0.29
		Male	3,200	0.45	0.32	0.32
		Female	2,500	0.38	0.26	0.25
	Urban areas	Total	3,900	0.52	0.34	0.34
		Male	2,200	0.57	0.38	0.38
		Female	1,700	0.47	0.31	0.30
	Rural areas	Total	1,800	0.29	0.22	0.21
		Male	1,000	0.30	0.24	0.23
		Female	800	0.28	0.20	0.19
Non-GIST STS	All areas	Total	34,200	2.50	1.92	1.85
		Male	16,000	2.27	1.78	1.73
		Female	18,200	2.73	2.06	1.97
	Urban areas	Total	22,700	3.03	2.21	2.13
		Male	10,700	2.81	2.09	2.02
		Female	12,000	3.25	2.34	2.25
	Rural areas	Total	11,500	1.85	1.52	1.48
		Male	5,300	1.64	1.37	1.35
		Female	6,200	2.09	1.69	1.61

There were 5,700 GIST cases (3,200 men and 2,500 women) nationwide in 2014, accounting for about 14.3% of the overall STSs. The crude incidence rate was 0.42/100,000 (0.45/100,000 in men and 0.38/100,000 in women). On the other hand, approximately 34,200 non-GIST cases (16,000 men and 18,200 women) were diagnosed, accounting for about 85.7% of the overall STSs. The crude incidence rate was 2.50/100,000 (2.27/100,000 in men and 2.73/100,000 in women). The incidence of overall STSs and non-GIST STSs were higher in women than in men in urban and rural areas, whereas the incidence of GIST was higher in men than in women in both areas.

As shown in Figure 1, the age-specific incidence rates of overall STSs and non-GIST STSs were the lowest in children aged 5 to 9 years, increased dramatically after 35 years old and peaked at 75–79 years old in women and 80–84 years old in men for overall STSs, and 80-84 years old in both sexes for non-GIST STSs. As for GIST, the age-specific incidence rate was unstable due to the scarcity in individuals under 25 years old. The rate increased dramatically after 45 years old and peaked at 75–79 years old in women and 70–74 years old in men.

1.

Age-specific incidence for soft tissue sarcoma by sex in China, 2014. (A) All soft tissue sarcoma. (B) Gastrointestinal stromal tumor. (C) All soft tissue sarcoma excluding gastrointestinal stromal tumor.

Distribution of histological subtypes

As shown in Table 3, altogether 8,475 new STS cases (4,033 men and 4,442 women) were reported from the 339 cancer registries. The most common histological subtype was GIST (14.38%), followed by nerve sheath tumor and malignant peripheral nerve sheath tumor (MPNST) (12.48%), leiomyosarcoma (5.97%), liposarcoma (5.75%), and fibrosarcoma (5.73%).

3.

Histological distribution of soft tissue sarcomas by sex and age in China, 2014

Histological subtypes	Proportion (%)	Gender			Male/female ratio	0-19 years old (%)	20-64 years old (%)	65+ years old (%)
		Total, n	Male, n	Female, n
NOS, not otherwise specified
Sarcoma NOS	28.65	2,428	1,198	1,230	1.0	3.50	54.08	42.42
Gastrointestinal stromal tumor	14.38	1,219	669	550	1.2	0.41	56.52	43.07
Leiomyosarcoma	5.97	506	95	411	0.2	0.40	79.05	20.55
Endometrial stromal sarcoma	4.46	378	30	348	0.1	0.53	84.13	15.34
Liposarcoma	5.75	487	281	206	1.4	2.26	66.74	31.00
Malignant fibrous histiocytoma	5.12	434	269	165	1.6	0.92	51.15	47.93
Angiosarcoma	4.59	389	212	177	1.2	3.34	59.64	37.02
Rhabdomyosarcoma	1.43	121	64	57	1.1	44.63	47.93	7.44
Fibrosarcoma	5.73	486	245	241	1.0	2.67	72.22	25.10
Nerve sheath tumor and malignant peripheral nerve sheath tumor	12.48	1,058	523	535	1.0	2.74	80.72	16.54
Dermatofibrosarcoma	2.28	193	114	79	1.4	6.74	78.24	15.03
Other specified soft tissue sarcoma
Carcinosarcoma, NOS	1.58	134	38	96	0.4	0.00	61.19	38.81
Synovial sarcoma	1.18	100	51	49	1.0	4.00	84.00	12.00
Mixed tumor, malignant NOS	0.86	73	31	42	0.7	4.11	72.60	23.29
Primitive neuroectodermal tumor, NOS	0.55	47	29	18	1.6	14.89	76.60	8.51
Granular cell tumors and alveolar soft 　part sarcoma	1.16	98	15	83	0.2	6.12	84.69	9.18
Paragangliomas and glomus tumors	1.13	96	45	51	0.9	3.13	79.17	17.71
Malignant mesenchymoma	0.88	75	37	38	1.0	2.67	68.00	29.33
Malignant myoepithelioma	0.17	14	10	4	2.5	0.00	64.29	35.71
Clear cell sarcoma	0.34	29	18	11	1.6	6.90	72.41	20.69
Kaposi sarcoma	0.59	50	29	21	1.4	4.00	58.00	38.00
Rhabdoid tumor	0.08	7	4	3	1.3	28.57	57.14	14.29
Extraskeletal osteosarcoma and 　chondrosarcoma	0.26	22	13	9	1.4	0.00	90.91	9.09
Myxosarcoma	0.22	19	7	12	0.6	0.00	57.89	42.11
Malignant giant cell tumors	0.14	12	6	6	1.0	16.67	58.33	25.00
Total	100.00	8,475	4,033	4,442	0.9	3.12	64.65	32.24

Overall male/female ratio was 0.9, indicating a female predilection. Highest ratios were found in malignant myoepithelioma, clear cell sarcoma, malignant fibrous histiocytoma and primitive neuroectodermal tumor NOS, whereas lowest ratios were found in endometrial stromal sarcoma, granular cell tumors and alveolar soft part sarcoma, leiomyosarcoma, carcinosarcoma NOS, and myxosarcoma.

Children and adolescents under 20 years old accounted for 3.12% of all STS cases, while adults aged 20-64 years and over 65 years accounted for 64.64% and 32.24%, respectively. In major histological subtypes, children and adolescents represented 44.63% of all rhabdomyosarcoma cases, whereas endometrial stromal sarcoma (84.13%) and nerve sheath tumor and MPNST (80.72%) occurred mainly in adults aged 20 to 64 years, and proportions of malignant fibrous histiocytoma (47.93%) and GIST (43.07%) were relatively high in elders above 65 years old.

Distribution of primary sites

As shown in Table 4, more than 22% of STSs were located in the connective, subcutaneous and other soft tissues, of which 574 were in lower limb and hip (lower extremities), 234 were in upper limb and shoulder (upper extremities). About one-fifth of STSs occurred in digestive organs, of which 952 were in stomach, 294 were in small intestine, and 153 were in liver. Retroperitoneum and peritoneum and eye, brain and other parts of the central nervous system both accounted for nearly 14% of all STSs.

4.

Distribution of primary sites of soft tissue sarcomas by gender in China, 2014

ICD-O-3 code	Primary sites	Gender			Proportion (%)
		Male, n	Female, n	Total, n
NOS, not otherwise specified
C00-C14	Lip, oral cavity and pharynx	82	57	139	1.64
C15-C26	Digestive organs	982	721	1,703	20.09
C16	Stomach	528	424	952
C17	Small intestine	164	130	294
C22.0	Liver	100	53	153
C30-C39	Respiratory system and intrathoracic organs	214	112	326	3.85
C44	Skin	203	146	349	4.12
C47	Peripheral nerve and autonomic nerve system	44	48	92	1.09
C48	Retroperitoneum and peritoneum	567	660	1,227	14.48
C49	Connective, subcutaneous and other soft tissues	1,061	855	1,916	22.61
C49.0	Head, face and neck	112	84	196
C49.1	Upper limb and shoulder	142	92	234
C49.2	Lower limb and hip	325	249	574
C49.3	Thorax	76	66	142
C49.4	Abdomen	88	105	193
C49.5	Pelvis	81	80	161
C49.6	Trunk	85	47	132
C49.8	Overlapping lesion	4	4	8
C49.9	NOS	148	128	276
C50	Breast	2	87	89	1.05
C51-C57	Female genital organs	0	924	924	10.90
C60-C63	Male genital organs	51	0	51	0.60
C64-C68	Urinary tract	71	43	114	1.35
C69-C72	Eye, brain and other parts of the central nervous system	550	582	1,132	13.36
C71	Brain	117	146	263
C72	Spinal cord, cranial nerves and other parts of central nervous system	396	362	758
C73-C75	Thyroid and other endocrine glands	39	46	85	1.00
C76, C77, C80	Lymph nodes and other sites	167	161	328	3.87
Total		4,033	4,442	8,475	100.00

As shown in Table 5, of all GIST cases, 98.6% arose in digestive organs. Stomach was the most common site (67.51%), followed by small intestine (20.92%), other digestive organs (4.59%) and rectum (3.12%). Extra-GIST was found in retroperitoneum and peritoneum, bronchus and lung, soft tissues, thyroid gland, and other unknown sites, representing 1.4% of all GIST cases.

5.

Distribution of primary sites of gastrointestinal stromal tumor by sex in China, 2014

ICD-O-3 code	Primary sites	Gender			Proportion (%)
		Male, n	Female, n	Total, n
C15	Esophagus	6	3	9	0.74
C16	Stomach	444	379	823	67.51
C17	Small intestine	141	114	255	20.92
C18	Colon	10	9	19	1.56
C20	Rectum	26	12	38	3.12
C21	Anus and anal canal	1	0	1	0.08
C25	Pancreas	1	0	1	0.08
C26	Other digestive organs	29	27	56	4.59
C34	Bronchus and lung	0	1	1	0.08
C48	Retroperitoneum and peritoneum	5	1	6	0.49
C49	Connective, subcutaneous and other soft tissues	2	0	2	0.16
C73	Thyroid gland	1	0	1	0.08
C76	Other and ill-defined sites	3	3	6	0.49
C80	Unknown primary site	0	1	1	0.08
Total		669	550	1,219	100.00

Discussion

In this study, using population-based cancer registry data, we found that nearly 39,900 new STS cases (5,700 GIST and 34,200 non-GIST) were diagnosed nationwide in China in 2014, representing 1.05% (0.90% in male, 1.23% in female) of all cancer incidence²⁴, larger than the proportion (0.75%) in United States⁷. Crude incidence rates of STSs and GIST were 2.91 and 0.42 per 100,000 respectively, lower than the corresponding rates in western countries (4.7 and 0.78 per 100,000)^6,19.

Incidence of STS in China generally increased with age. The rate in Austria showed similar trend but peaked and dropped at younger ages¹⁴. Unlike carcinomas, STSs comprise a large proportion of pediatric malignancies and are an important cause of death in adolescents and young adults^3,4. In our study, 3.1% of all STSs occurred in individuals aged 0 to 19 years, whereas in United States and Japan, corresponding proportions were 5.6% and 6.7%, respectively^3,12. Rhabdomyosarcoma was the only entity with a median age < 20 years in United States. It was also the most common subtype in people under 20 years old in these three studies.

Unlike most carcinomas, to which men are prone to develop, contradictions remained as for the major gender of STS patients. In Europe, there were more female patients^6,14, while in Korea and Taiwan China, results showed the opposite^13,26. Similar to these Asian results, men outnumbered women in most primary sites and histological subtypes in our study. However, there were more female patients due to the numerous cases in breast and female genital organs, and subtypes like leiomyosarcoma, which occurs mainly in genital organs, bladder and uterus.

GIST was the most common histological subtype in Chinese population, as was found in Europe¹⁵. In France, 16.7% of the identified sarcoma cases were GIST. French distribution of GIST was similar to our study, with 61% and 27% arose in stomach and small intestine, respectively. The proportion of extra-GIST nevertheless was greater because 5% of GIST arose in peritoneum. This extravagant percentage can be partly explained by the small sample size²¹. In contrast to our and American results¹⁹, GIST showed a female predilection in France. In United States, there were relatively less GIST in stomach (55.3%) and more in small intestine (28.7%), while the proportion of extra-GIST was 3.0%, also larger than our result¹⁹.

Kaposi sarcoma, accounting for less than 0.6% of all STSs, was only a minor subtype in China and even less common (0.2%) in Japan¹², whereas it had very high incidence in African countries²⁷. Kaposi sarcoma is also the most common subtype (31.7%) in United States³. This could be explained by the relatively low prevalence of HIV in East Asia because Kaposi sarcoma is strongly related with immune suppression²⁸.

Other than Kaposi sarcoma, common subtypes in United States were fibrohistiocytic tumors (30.4%) and leiomyosarcoma (27.4%)³. Even after the inclusion of GIST into the leiomyosarcoma category, corresponding proportions in our study, which were 5.12% and 20.35% respectively, were still smaller than these results. In Japan, leiomyosarcoma (28.1%) was the most prevalent subtype, followed by liposarcoma (9.6%), and malignant fibrous histiocytoma (9.0%)¹². In contrast, nerve sheath tumor and MPNST is far more prevalent in China, representing more than 12% of all STSs in our study, compared with 3.5% in Japan and 4.6% in United States.

In terms of primary sites, according to American Cancer Society, STSs can develop in soft tissues, internal organs as well as retroperitoneum²⁹. However, the WHO classification did not include internal organs as primary sites for STSs and listed only the extremities, trunk wall, and retroperitoneum⁵. Therefore, STSs were recognized as an entity with a variety of primary sites in some studies^10-12,14, while in others, STSs were more confined and sarcomas in internal organs were categorized as visceral sarcoma separately^8,15.

In this study, we used the classification of American Cancer Society, and found that connective, subcutaneous and other soft tissues was the most common site. Delving into more specific topographic categories, lower extremity was the most common subsite, and the number of cases in lower extremity was 2–3 times of that in upper extremity. Results in Korea¹³, Japan^12,30 and Europe¹⁵ further proved ours. Digestive organs, which was recognized as visceral sarcomas in some studies, were the second most common site for STSs, of which the majority was GIST. We found 17 extremely rare extra-GIST cases based on this large population-based study. This will provide valuable epidemiological evidence for further researches.

STSs have more than 70 histological subtypes¹⁵. Together with the various primary sites, they make up a large number of combinations that vary tremendously in terms of survival and clinical treatment, which means that same histological subtype in different sites may need completely different therapies^6,31. The rarity of STSs and the even smaller number of cases in each combination further add to the difficulty faced by researchers and physicians. Although the national survival of STSs in China is still unknown, overall five-year survival rate of STSs in developed countries ranged from 55% to 78%^4,8,30, indicating a relatively optimistic prognosis. Nevertheless, survival may vary tremendously (48.0%-99.5%) between different subtypes³⁰, and would worsen along with ageing³. This, together with its scarcity, emphasize that more attention should be paid to this less noticed disease and its patients.

Some limitations of this study should be noticed. First, probable methodology bias remains. We retrieved data based on ICD-O-3 morphological codes in this study because STSs can arise in almost every part in the human body, making topographical codes less useful than in the analysis of carcinomas. Nevertheless, according to a European study, nearly 42% of the first histological diagnoses of STSs were inaccurate. This heterogeneity and discrepancy can lead to misreport on subtype incidence^4,32 and consequently bias the results on the epidemiology of STSs. Second, proportion of sarcoma NOS was too high compared with Japanese results¹², and MV% was lower than that in Europe⁶, indicating a necessity to strengthen the pathological diagnosis especially in rural China.

Conclusions

This study is the first in China mainland to report the nationwide incidence of STS as well as the distribution of histological subtypes and primary sites. As a group of rare malignancies in China, STSs are more common in children than carcinomas. With numerous histological subtypes and primary sites, tremendous variations remain in STSs, which impose an arduous challenge to both medical researchers and doctors and emphasize more attention.

Acknowledgements

We sincerely appreciate the effort of every cancer registry for the data collection, data sorting, data verification and database creation. This study was supported by CAMS Innovation Fund for Medical Sciences (CIFMS) (Grant No. 2016-12M-2-004) and Ministry of Science and Technology (Grant No. 2014FY121100).

Conflict of interest statement

No potential conflicts of interest are disclosed.