Sir--In their report, Brian O'Sullivan and colleagues (June 29, p 2235)
1 compare the role of preoperative radiation therapy with that of postoperative radiation therapy for soft-tissue sarcomas of the limb.
An important point that was not entertained in this study is whether chemotherapy increases the rate of wound complications. This is an important question since chemotherapy--especially doxorubicin--is a major treatment option for this disorder. One of the main complications of radiation therapy associated with doxorubicin is "radiation recall". Skin reactions can be as mild as warmth and erythema,
2 but at times can involve desquamation and ulceration,
3 which may delay the healing process. For some limb sarcomas, chemotherapy is given before radiation and surgery to decrease the size of the tumour and surgical morbidity.
4 Whether recall reactions with doxorubicin occur with concurrent radiotherapy is unclear.
Nevertheless, the article is important because the prognosis for patients with sarcomas larger than 5 cm is inferior to those with tumours of less than 5 cm.
5 This finding is vital, because if the goal is palliation, then preoperative radiotherapy may not be the best approach.
*Christian Schultheis, Bradley Sachs, Syed Haider
Ellis Fischel Cancer Hospitals and Clinics, University of Missouri, Columbia, MO 65203, USA (e-mail:schultheisc@health.missouri.edu)
1 O'Sullivan B, Davis AM, Turcotte R, et al. Preoperative versus postoperative radiotherapy in soft-tissue sarcoma of the limbs: a randomised trial.
Lancet 2002;
359: 2235-41. [
Text]
2 Riggs CE. Antitumor antibiotics and related compounds. In: Perry MC, ed. The chemotherapy sourcebook. Philadelphia: Lippincott Williams and Wilkins, 2001: 234.
3 Donaldson SS, Glick JM, Wilbur JR. Adriamycin activating a recall phenomenon after radiation therapy.
Ann Intern Med 1974;
81: 407-08. [
PubMed]
4 Edmonson JH, Petersen IA, Shives TC, et al. Chemotherapy, irradiation, and surgery for function-preserving therapy of primary extremity soft tissue sarcomas: initial treatment with ifosfamide, mitomycin, doxorubicin, and cisplatin plus granulocyte macrophage-colony-stimulating factor.
Cancer 2002;
94: 786-92. [
PubMed]
5 Brennan MF, Alektiar KM, Maki RG. Sarcomas of the soft tissue and bone. In: Devita VT, ed. Cancer: principles and practice of oncology, 6th edn. Philadelphia: Lippincott Williams and Wilkins, 2001: 1874-75.
Sir--The randomised trial by Brian O'Sullivan and colleagues
1 would have been adequately powered to detect a difference in wound complications--its primary endpoint. Therefore, we cannot understand why the difference between the two groups in the incidence of serious wound complications (35%
vs 17%) is regarded as significant, whereas an almost identical difference in overall mortality (15%
vs 28%), or even in sarcoma-related deaths (7%
vs 19%), qualifies as only "slightly better survival (p=0·0481)". I am sure the authors have not just inversely related the significance of these results to the p value. The lower overall mortality is also internally consistent: the curves of preoperative radiotherapy group are always above those of the postoperative radiotherapy for local recurrence-free survival, distant relapse-free survival, or progression-free survival, although of course not having the power to lower the p value.
Preoperative radiotherapy could exercise survival benefit through several mechanisms including inhibition of angiogenesis at distant metastatic sites.
2 Before we dismiss it as a statistical quirk, it is essential that we await the long-term results of this trial and hopefully those of a much larger trial.
*Jayant S Vaidya, F Fausto Palazzo
Department of Surgery, Whittington Hospital, Royal Free and University College London Medical School, London W1W 7EJ, UK (e-mail:j.vaidya@ucl.ac.uk)
1 O'Sullivan B, Davis AM, Turcotte R, et al. Preoperative versus postoperative radiotherapy in soft-tissue sarcoma of the limbs: a randomised trial.
Lancet 2002;
359: 2235-41. [
Text]
2 Hartford AC, Gohongi T, Fukumura D, Jain RK. Irradiation of a primary tumor, unlike surgical removal, enhances angiogenesis suppression at a distal site: potential role of host-tumor interaction.
Cancer Res 2000;
60: 2128-31. [
PubMed]
Authors' reply
Sir--We acknowledge that other adjuvant treatments for sarcoma, such as chemotherapy, may have their own side-effects and complications, but their effect is often confounded by the combined use of several of these therapies. In our study, chemotherapy was not used and we cannot comment specifically from these results. Meric and colleagues
1 compared morbidity associated with radical surgery in soft-tissue sarcoma in 104 patients who received induction chemotherapy before surgery and 204 patients who had surgery first. The incidence of surgical complications was not different for patients undergoing preoperative chemotherapy from patients who had surgery alone, in those with sarcomas of the limbs (34%
vs 41%) and in those with retroperitoneal or visceral sarcomas (29%
vs 34%).
With respect to "radiation recall", we appreciate that this could cause difficulties for groups that use chemotherapy frequently and where scheduling issues of the type mentioned by Christian Schultheis and colleagues render it more likely. On the other hand, as they point out, the occurrence of such reactions is controversial and the data shown above indicate that much of the risk for surgical morbidity is unlikely to be due to chemotherapy on its own.
Finally we accept that a cut-point at 5 cm definitely declares patients at different risk of adverse outcome. Nevertheless, we do not consider that patients with lesions of greater than 5 cm should generally receive palliative approaches since a large proportion of such patients can be cured of their sarcoma.
Jayant Vaidya and Fausto Palazzo correctly note that our trial was powered and designed to detect a difference in major wound complications in soft-tissue sarcomas of the limbs undergoing preoperative or postoperative radiotherapy. Indeed, the trial was designed to be stopped early if the primary question was answered at an interim analysis after the midpoint of accrual (which is what took place). However, for practical reasons, the trial was not designed with sufficient power to reliably assess secondary endpoints such as survival and cancer-specific outcome.
Our reason for not emphasising the secondary endpoint was that our sample-size estimation was based on wound complication not overall survival. The observed significance should be adjusted for early comparison to avoid the possibility of potential bias due to small numbers of events, and the p value for survival should not be interpreted at its face value and must be interpreted with caution. Additionally, the p value is for time to event, based on the survival curve presentation for the secondary endpoints referred to, and not on the proportion of events which are different analytic approaches. Also, as we noted in the paper, the timing of the survival analysis was not specified before the trial started, whereas the primary endpoint analysis was clearly built into the trial design.
We respect Vaidya and Palazzo's comments concerning survival benefit. The biology of cancer and metastasis is complex and mediated by numerous factors and pathways
2,3 that could be influenced at many levels. They mention angiosuppression by radiation,
4 but other mechanisms could also apply. The point is that we do not know. Future research is needed and, as Vaidya and Palazzo point out, longer follow-up from our trial and other trials. In the meantime we should not overinterpret the results that are available.
*Brian O'Sullivan, Robert Bell, Aileen Davis, Robert Turcotte, Benny Zee
Department of Radiation Oncology, Princess Margaret Hospital, University of Toronto, Toronto, Ontario M5G 2M9, Canada (e-mail:Brian.OSullivan@rmp.uhn.on.ca)
1 Meric F, Milas M, Hunt KK, et al. Impact of neoadjuvant chemotherapy on postoperative morbidity in soft tissue sarcomas.
J Clin Oncol 2000;
18: 3378-83. [
PubMed]
2 Blobe GC, Schiemann WP, Lodish HF. Role of transforming growth factor beta in human disease.
N Engl J Med 2000;
342: 1350-58. [
PubMed]
3 Hanahan D, Weinberg RA. The hallmarks of cancer.
Cell 2000;
100: 57-70. [
PubMed]
4 Hartford AC, Gohongi T, Fukumura D, Jain RK. Irradiation of a primary tumor, unlike surgical removal, enhances angiogenesis suppression at a distal site: potential role of host-tumor interaction.
Cancer Res 2000;
60: 2128-31. [
PubMed]
