Retroperitoneal Sarcoma With Infected Necrosis: An Unfavourable Prognostic Factor

Purpose. To report the phenomenon of infected retroperitoneal sarcoma (RPS). Method. Two case reports. Results. Both patients died soon after laparotomy. Discussion. Infected RPS is identified as an entity not clearly documented in the literature. It should probably be added to the list of poor prognostic factors when planning the management of patients with RPS.


Introduction
T he prognosis of patients with retroperitoneal sarcom a (RPS) is poor, with an overall 5-year survival rate reported between 12± 40% 1± 9 with rare exception. 10 T he patients w ith the m ost favou rable prognosis are those with low grade tumours in which com plete excision of the prim ary tum our is achieved. 1,2,10,11,12 Even in this group the cum ulative probability of local relapse has been reported up to 85% at 5 years. 1 There is a role for palliative debulking of RP S and patients w ith low grade tum ours can have repeated surgery over m any years w ith a 5 year survival rate of 80% . 2 M ajor surgery should probably be avoided in patients with concurrent m etastases, low perform ance status and high grade recurrent tum ours unless essential for palliation.
W e were involved in the m anagem ent of two patients w ith aggressive retroperitoneal sarcom as w ho presented with severe pain that was unresponsive or not suitable for other palliative therapy. T hey w ere offered surgery with the hope of im proving their pain. At laparo tomy palliative intracapsular debulking w as undertaken. Both these patients had offensive necrotic tum our and culture con® rm ed contam ination w ith enteric organism s. Infected necrosis in large retroperitoneal or intraperitoneal bowel related tum ours is a recognised but uncom mon clinical ® nding 13 however there is only one report of infected RPS in the literature. 14 Infected RPS has not been related to prognosis previously.

Case reports
C ase 1 T his 51-year-old wom an presented in January 1998 with abdom inal pain and a pelvic m ass. C om puted tomograph y (C T) Scan showed a 9 cm pelvic m ass thought to be arising from the uterus. At the laparotomy by her gynaecologist a retroperitoneal m ass in the lower abdom en was biopsied and found to be a spindle cell sarcom a. Postoperatively she had ongoing pain that required daily morphine but was apyrexial. She was referred to the Royal Marsden H ospital and w as initially advised that there was no other palliative options. T here was a considerable increase in the size of the tum our on repeat C T Scan (see Fig. 1). After repeated representations by the patient and her fam ily a further laparotomy was agreed to. There was a large pelvic m ass of necrotic offensive tum our with adherent loops of sm all bowel. This w as debulked. Postoperatively she developed sepsis and respiratory failure. It w as decided not to reintubate her and she died 48 hours postoperatively. T he bacterial culture from the necrotic tum our grew Streptococcus m illeri, coliform s and anaerobes. Histopathology dem onstrated spindle cell sarcom a (EO RTC grade 2) perhaps representing gastrointestinal strom al tum our (of autonom ic type). T his 35-year-o ld wom an was diagnosed with a retroperitoneal biphasic synovial sarcom a (EO RTC G rade 2) in April 1997. It was incom pletely resected at that time. She received 2 cycles of ifosfam ide and doxorubicin which were poorly tolerated and thus stopped despite a partial response. She had several episodes of unexplained fever with the chemotherapy. In January 1998 she had progressive disease on CT Scan and poorly controlled abdominal pain. D ebulking surgery w as the only option to attempt to im prove her quality of life. Preoperatively she had one tem perature of 39°C and was comm enced on broad spectrum antibiotics. At laparotom y the m ass w as debulked of large volum es of offensive necrotic m aterial. M icrobiological culture grew Streptococcus m illeri and Bacteroides m elaninogenious. Postoperatively she m ade good progress how ever she represented to hospital one m onth after surgery w ith severe abdom inal pain. In the course of investigating the pain, she deteriorated and had an hypovolaem ic cardiac arrest due to presum ed intra-abd om inal bleeding. She was initially resuscitated however under the circum stances it w as considered inappro priate to intervene further. She died that evening.

D iscussion
Retroperitoneal sarcomas have a poor long term prognosis with an overall 5-year survival rate reported between 12 and 40% 1± 9 with one report of 63% 5-year survival. 10 T hey also have an ongoing signi® cant recurrence and mortality rate after that time. 1,5,10,11 Even w hen the tumour is m acroscopically rem oved it has been reported that there is a cum ulative probability of local relapse between 50± 80% at 5 years. 1± 10 D espite this there are m any cases of recurrent low grade sarcomas that are able to be debulked at regular intervals with m edium to long term survival. 2,3 The factors related to poor prognosis include metastatic disease which is often transperitoneal seeding, 1 failure to accomplish complete excision, 1± 3,5,6,10± 12 high grade tumours [1][2][3]5,9,10 and in some series the histopathological type of tumour. 3,15± 17 The failure to accomplish complete excision is a m anifestation of the size of the tumour. This relates to and is in¯uenced by its location in relationship to vital structures and the num ber of organs involved. 1,2,4,5 To the list of poor prognostic features we would add infected RPS. T he ® nding of infected necrosis in advanced abdom inal and retroperitoneal tumours is a recognised but an uncomm on clinical ® nding having been described in lymphoma, renal cell carcinoma and metastatic testicular carcinoma, 13 however, there is only one report of infected RPS involving a duodenal leiomyosarcoma. 14 There are no previous reports suggesting a relationship to prognosis and risks of reoperation for palliation. From our experience, if identi® ed preoperatively infected RPS would deter further palliative surgery in favou r of persisting with conservative therapy including appropriate antibiotics.
In both of our cases the nature of the aggressive disease w as such that bowel was intimately involved and thus the presum ed portal of entry of the infection. T he bacteria cultured are known gut organism s. T he other possible portal of entry is blood born organism s that seed the tumour but this is less likely given both cases were polym icrobial. N o precedent exists for establishing the diagnosis of infected retroperitoneal sarcoma however there are sim ilarities to infected necrosis in severe acute pancreatitis. T herefore CT Scan guided biopsy and culture should be feasible. 18 Sim ilarities to infected pancreatic necrosis and its com plications, such as m ycotic splenic artery aneurysm causing catastrophic bleeding and m ultiorgan failure from sepsis, also bear striking sim ilarity to our cases. 18 In conclusion, retroperitoneal sarcom as have a poor prognosis. Patients should only be offered palliative surgery w hen their sym ptom s cannot be controlled by any other m odality and it is technically feasib le to reduce the bulk of their disease. If infected necrosis is suspected clinically CT guided aspiration for m icrobiological culture could be used for diagnosis. Based on our recent experience infected RPS m ay deter further palliative surgery in favo ur of conservative m anagem ent. Any RPS found to have necrosis at surgery should be sent for m icrobiological culture as well as histopathology.