Disease Incidence and Results of Extremity Lesion Treatment: Mersey Region Soft Tissue Sarcomas (1975–1985)

Purpose. The incidence and treatment results of extremity soft tissue sarcoma (STS) in the Mersey Region, in the absence of a Multi-Disciplinary Unit, for the period 1975–1985, have been analysed. Subjects and methods. Data from cases presenting with STS within the Mersey region, from 1 January 1975 until 31 December 1985, were reviewed. Only patients with sarcoma of head and neck, thoracic wall, abdominal wall, retroperitoneum, limb girdle or extremity were included. Extremity lesions were staged according to the MTS system. Pathological data also were assigned a grade according to tumour differentiation, mitosis count and tumour necrosis. Data from patients with a minimum follow-up of 5 years were collated, and patterns of treatment failure were investigated. Finally, time to first occurrence was analysed. Results and Discussion. The incidence of STS in this study was identical to that reported by the US Department of Health in 1976. Five year survival rate for Stage I tumours was only 51.7% which compares very unfavourably with contemporary series from Multi-Disciplinary Units. Five year survival rate following wide local excision ± adjuvant therapy is 52.4%, while that following amputation ± adjuvant therapy is 45.5%. While not attaining the results reported by other centres, limb-sparing surgery does not appear to appreciably prejudice long-term survival. Conclusions. STS are rare in the UK, leading to poor classification and suboptimal treatment of lesions. It is important to establish multidisciplinary teams of surgeons, radiologists, radiotherapists and oncologists to plan and organise multimodality therapy for STS.


Introduction
Soft Tissue Sarcom as (STS) are rare tumours, accounting for less than 1% of m alignant neoplasm s. 1 D espite their rarity, recent advances in lim b-sparin g techniques and the use of adjuvant radiotherapy and chem otherapy 2 in m ultidisciplinary units 3,4 has seen an im provem ent in their m anagement. In the M ersey Region between 1975 and 1985 no such specialist unit existed. In this analysis, the incidence and treatm ent results of extrem ity soft tissue sarcom as occurring in M ersey between 1975 and 1985 w ere investigated.

S ubjects and m ethods
T he M ersey Region Cancer Registry database was utilised for this study. T he case records of patients presenting w ith soft tissue sarcom as within the M ersey Region were review ed from 1 January 1975 to 31 D ecem ber 1985, inclusive. For eligibility, patients required a histologically proven diagnosis of sarcom a arising from the head and neck, thoracic wall, abdom inal wall, retroperitoneum , lim b girdle or extrem ity. V isceral sarcom as and m etastatic sarcom as of unknown primary w ere excluded. T he eligible patient population was then examined to derive age/sex distribution and age-sp eci® c annual incidence. Patients with lim b girdle and extremity STS were further exam ined to calculate their age/ sex distribution data and the relative incidence of histological types of ST S.
Extremity lesions were staged according to the M T S staging system , 1980. 5 T he pathological reports w ere reviewed and assigned a grade (high or low) according to the work of Trojani et al. 6 This m ethod uses tum our differentiation, m itosis count and tum our necrosis to provide an estim ation of grade. The surgical technique used and results of investigation w ere correlated to allow estimation of surgical stage. Patients with a m inim um follow-up of 5 years were identi® ed and m ethods of prim ary m anagem ent were collated. T he 5-year survival rates were analysed by prim ary treatm ent employed, by com m onest histological type and by disease stage (M T S Surgical Staging System , 1980). Patterns of treatment failure were then investigated. Finally, time to ® rst recurrence was analysed.

Results
Incidence of STS (M ersey 1975± 1985) T he above criteria were satis® ed by 544 patients; 309 m ales and 235 fem ales, giving a m ale to fem ale ratio for the study of 1.32:1. The m ean age of the 309 m ales was 53.9 years (range 1 month to 97 years). T he m ean age of the 253 fem ales was 55.5 years (range 9 days to 88 years). The age/sex distribution of ST S in the M ersey Region (1975± 85) is shown in Fig. 1 Fig. 3. The peak incidence of ST S for both sexes occurs in the over 85 years age group, in which STS are twice as com m on in m ales as in females. It should be noted that, for both sexes, ST S becom e progressively m ore comm on after 45 years of age. The sites of occurrence of STS for the 544 patients in the study are shown in Table 1. A group of 274 patients with lim b girdle and extremity ST S were identi® ed. T he m ean num ber of new ST S occurring annually in the low er and upper lim bs was 18.2 (range 12± 34) and 6.7 (range 3± 13), respectively. The m ean num ber of extremity lesions occurring annually was thus 24.9 (range 18± 41). Further age/sex distribution data for extremity lesions were calculated (Table 2) giving a m ale to fem ale ratio of 1.3:1 and a lower lim b to upper lim b ratio of 2.7:1.
The relative incidence of histological types of ST S of the extrem ities is presented in T able 3.    operatively after 2± 3 weeks delay to allow w ound healing. T ypically, radiotherapy w as adm inistered by m egavoltage technique using parallel opposed ® elds, vital structures being protected by wedging the X-ray ® eld. Palliative radiotherapy in one or tw o fractions was used in som e patients, but m ost patients received their therapeutic external beam irradiation to the prim ary site, drain tracks and lym ph nodes; 35± 65 G y over 3± 6 w eeks being the usual regim e. Chem otherapy was used as part of the prim ary m anagem ent in certain situations, which included patients presenting with dissem inated disease and patients with advanced primaries not am enable to surgical treatment and w ho did not respond to radiotherapy. A variety of chem otherapeutic regimes rately oncological surgical m argins (intralesional, m arginal, wide or radical) from exam ination of the surgical records and pathological reports.
Radiotherapy, when utilised, was given post-   (Table 5). T he overall 5-year survival rate for the series w as 40.8%. T he am putation rate was 9.4% . T he 5-year survival rate of patients treated by wide local excision 6 adjuvant w as 52.4% . T his com pares w ith the 5-year survival rate of 45.5% for those patients treated by am putation 6 adjuvant therapy. Lim b-spari ng surgery did not appear to prejudice long-term survival. Excision biopsy alone is to be condem ned as a m ethod of treatment as this is associated w ith only 10.5% 5-year survival.
The 5-year survival rates occurring in the 10 m ost com m on histological types were calculated (Table  6). T he results of treatment by disease stage (M T S Surgical Staging System , 1980) were also investigated ( Table 7).
The patterns of failure of 233 patients were then studied. Lym ph node involvem ent occurred in 18 (6.6% ) patients. Local failure occurred in 120 (43.8% ). D istant m etastases without local recurrence occurred in 31 (11.3%). T reatm ent failure by histological type is shown (Table 8).
Finally, in this study, time to ® rst local recurrence in patients w as analysed. Seventy percent of tum ours that recurred locally did so by the ® rst year, 95% recurred by 3 years, and only two ® rst local recurrences occurred m ore than 5 years from prim ary surgery.

D iscussion
In this study the incidence of ST S in M ersey Region was investigated. At 2.0 per 100 000 (2.36/100 000 for males and 1.68 per 100 000 for fem ales) this rate is identical to the rate reported by the US D epartm ent of H ealth in 1976. 1 T he incidence of ST S increases steadily with age after 34 years, and is higher in m ales, such that in the greater than 85year age group STS are tw ice as comm on am ongst m en. T his increase with age m ay be due to environm ental exposure to carcinogens, cum ulative degenerative genetic dam age and im m unosuppression of old age. Other studies have shown an increased occurrence of those tum ours am ongst m en. 7 In this study m ales out-num bered fem ales by 1.32 to 1. T he reason for this is unknown but m ay be due to occupational exposure to carcinogens.
W hen compared w ith previous studies, 8± 11 the relative incidence of histological types of STS show som e m arked disparities. Unclassi® ed tumours are far m ore com m on in this series (19.7% of cases com pared with 10.5% , a mean of four series). This m ay be a re¯ection of the lack of expertise in pathology, as app ropriate biopsy technique and histological experience can enable m ore tum ours to be categorised. Synovial sarcom as and rhabdom yosarcom as occurred w ith a reduced frequency when com pared to other studies. Fibrosarcom as comprised the largest group of tumours, being 23.3% of the total. M alignant ® brous histiocytom a was diagnosed rather infrequently (6.6%). This m ay again be related to pathological inexperience, as m any tum ours form ally regarded as ® brosarcom as are now classi® ed as m alignant ® brous histiocytom as. Electron microscopy and the use of special stains have enabled greater categorisation of ST S. T his is clearly a reason for the establishm ent of STS Pathology U nits.
The distribution of ST S was largely as reported in previous studies. 8± 11 Retroperitoneal tumours were m ore com m on than expected (14.5% of cases compared w ith a m ean of 11.6% in four other studies), and lower lim b tumours were less com m on than expected (36.8% com pared with a m ean of 41.8% ).
Thus, data from the M ersey study appears consistent with previous investigations, and the few results app earing outside the expected range app ear to be due to inexperience in histological diagnosis.

E xtremity STS
In this study, low er limb lesions outnum bered upper lim b lesions by 2.7 to 1. T his probably re¯ects the greater soft tissue bulk of the lower lim b. Fem ales developing tum ours were older than their m ale counterparts; fem ales with upper lim b tumours w ere, on average, 4.2 years older than their m ale patients. Anatom ically, ST S were distributed such that 44.2% of extremity lesions occurred in the buttock/thigh region, knee and low er leg tumours com prised 20.4% of lesions, and shoulder and arm tum ours com prised 16% of all lesions. Soft tissue sarcom as at other m ore distal sites w ere relatively rare. T his overall distribution com pares well with previous studies.
Examination of histological types in extrem ity lesions con® rm ed the preponderance of ® brosarcom as and unclassi® ed STS together w ith the dim inished incidence of m alignant ® brous histiocytom as. It was noted that synovial sarcom as are increased in incidence (2.8% for all sites, 5.1% for extremity lesions) and leiom yosarcom as are less com mon in the extremities (7.2% of all sites; 4.0% for extremity lesions).
The 274 extrem ity lesions were staged according to the Enniking M TS surgical staging system , 1980. T his form of retrospective cross-correlating between surgical technique, pathological exam ination, preoperative exam ination and investigation is w ell known to be inaccurate. Better surgical techniques along oncogenic lines w ill allow better future evaluation of data; however, within these constraints the distribution of extremity lesions was as follows: stage Ia 15.3%, Ib 21.9% , IIa 19.0% , IIb 33.6%, IIIa 4.0% and IIIb 6.2%. This com pares well to the w ork of Sim on and Enniking w ho exam ined a large series of extrem ity STS. 12 There are several interesting observations to be m ade upon the m anagement of extrem ity STS in the M ersey Region. Firstly, the am putation rate at 9.9% is low and com pares favou rably with contemporary series. 3,13 Excision biopsy alone was perform ed in 22 cases, 23 patients received excision biopsy and radiotherapy. T here is abundant evidence to show that excision biopsy alone is an inadequate form of treatm ent. T here appears to be no logical explanation why so m any patients treated by excision biopsy were not subsequently given radiotherapy. Surgeons may not have referred these patients or, conversely, radiotherapists m ay not have considered these tum ours radio-responsive. Again, in the case of wide local excision, 41 patients had supplem entary radiotherapy whilst 115 patients had surgery alone. If the results are exam ined the 5-year survival rate for wide local excision alone was 57.1% com pared with 48.6% for wide local excision plus radiotherapy. This m ay im ply that radiotherapy was reserved for patients with larger tumours resulting in poorer prognosis. The 5-year survival rate for M ersey extrem ity ST S (40.8% ) appears low when contemporary series of patients treated by lim bsparing surgery are attaining 61% (range, 47.6± 67.0% ) 5-year survival rates. 14± 16 Solutions to explain this poor outcom e cannot be derived from these data due to the lack of com parable groupings.
The present study reveals that 5-year survival rate following wide local excision 6 adjuvant therapy is 52.4% , and that following am putation 6 adjuvant therapy it is 45.5% . T he present study therefore reveals that lim b-spar ing surgery, as perform ed in the M ersey Region, w hilst not attaining the results of other centres, 17 does not appear to com prom ise long-term survival when com pared to am putation.
Five-year survival by histopathological type and stage of disease at presentation was also reviewed. Better differentiated tum ours were noted to have better 5-year survival rates. T he unclassi® ed group, as would be expected, had the poorest prognosis, and the 5-year survival rate was only 17.6%. F iveyear survival by disease stage was as follows: stage I 51.7% , stage II 36.7% , stage III 0% . Contem porary studies show 90± 95% 5-year survival rates for stage I lesions, 45% survival for stage II lesions and 5% survival for stage III lesions. 14,15 M ost of these contemporary series have patients with sim ilar stage disease but differ in that aggressive radiotherapy was used either pre-or post-operatively.
The  18 The m arked rate of local recurrence in this series cannot be linked causally to distant m etastasis. It does, however, suggest inadequate treatm ent in the m anagement of local disease.
Lym ph node m etastasis occurred in 18 patients (7.6% ) treated during the study period; this agrees w ell with the 5.8% lym ph node involvem ent rate reported by Ariel. 19 M etastasis to lym ph nodes occurred with increasing frequency in patients with leiom yosarcom as, synovial cell sarcomas and ® brosarcom as, in com pariso n with other series.
The local recurrence rate for excision biopsy alone w as found to be 81.8%; this com pares to a local recurrence rate of 47.8% when local excision is supplem ented with radiotherapy. T he addition of radiotherapy to wide local excision decreases the local recurrence rate from 35.7 to 29.3% . Local recurrence rate for patients treated by am putationw as 18.5% ; this com pares favo urably w ith reported local failure rates of 15.8% for patients treated by am putation. 12,20,21 Unfortunately there was no way of categorising am putees as being m arginal, wide or radical. D e® nite conclusions about what constitutes adequate local treatment cannot be drawn from this w ork due to the likely heterogeneous nature of both the patient groups and treatm ent protocols. Patients w ho develop local recurrence are clearly at increased risk of developing distant m etastasis and tum ourrelated m ortality. T here was under-use of radiotherapy in the M erseyside Region during the study period, this cannot explain the poor survival rates reported here, but m ay provide an explanation for the very high and undesirable rates of local recurrence. T his strengthens the argum ent for a combined m odality treatment app roach to the m anagement of ST S.
Histological type also appears to in¯uence local recurrence. Spindle cell sarcom as, rhabdom yosarcom as and synovial cell sarcom as had high local recurrence rates at 90.0, 80.0 and 69.2% , respectively. W ide variation in these rates have previously been noted. 22 Local recurrence rates for ® brosarcom as, liposarcom as and leiom yosarcomas were 50.0, 35.7 and 62.5%, respectively. It is well known that certain histological types show a predeliction to local recurrence. However all these rates are high when com pared with other studies. 12,23 Clearly the local control w as inadequate with regard to all histological types. The time to local recurrence was in accord w ith the work of Cantin et al. 24 and revealed that of these tum ours destined to recur 95% had done so by 5 years.
In sum m ary, this investigation has shown the M ersey Region population is at average risk of developing ST S. The rarity of these lesions m ean that in the U K, a given surgeon w ith an average sized practice will only see an ST S every 1± 2 years. T his inexperience has lead to suboptim al treatm ent of extremity lesions in terms of both local recurrence and 5-year survival. This inexperience is also m anifest in pathological assessm ent, as there were clear weaknesses in classi® cation of the lesions during the study period. T his work has further highlighted the im portance of establishing m ultidisciplinary teams involving surgeons, radiologists, radiotherapists and oncologists to plan and organise m ultimodality therapy for ST S.