Sarcomas gone bad: What to do about recurrent Ewing sarcoma

R eporting institutional or cooperative group results for survival after recurrence of sarcoma is an academic cottage industry. These retrospective analyses identify features at initial presentation or at recurrence that are associated with outcome. ‘‘Survival following recurrence’’ manuscripts also uniformly lament our communal failure as pediatric oncologists, serving as both a call-to-arms for treatment improvement and a collective exercise in group psychotherapy. As a voluntary disclosure statement, my own curriculum vitae has examples of ‘‘Outcome following recurrence of (fill in the blank) sarcoma’’ papers for the three most common pediatric sarcoma, so I know of what I write. Despite their somewhat predictable conclusions, these analyses of survival after recurrence serve an essential role in directing future therapeutic approaches for patients with newly diagnosed and recurrent sarcomas, particularly when the results are consistent between study populations and treatment eras. In this issue of Pediatric Blood & Cancer, Stahl et al. report the outcome following recurrence of Ewing sarcoma from the combination of three large, front-line, cooperative group clinical trials (CESS-81, CESS-86, and EICESS-92), illustrating the strength and limitations of this type of analysis. Drawing from a study populationwithwell-characterized initial clinical features and treatment, Stahl et al. demonstrate that the timing and pattern of recurrence are the strongest prognostic factors in recurrent Ewing sarcoma. Perhaps most striking is the similarity to the results seen from the Children’s Oncology Group (COG) cohort using the North American intergroup study, INT0091 [1]. Nearly three times the size of the COG cohort, the CESS/EICESS analysis has nearly identical key findings, including the 5-year overall survival of 7% in both series for early recurrence (<2 years after initial diagnosis) and 29– 30% for late recurrence. Also consistent between the North American and European cohorts is the timing (72–79% within 2 years of initial diagnosis) and pattern of recurrence, with isolated distant recurrence seen in 63–73%. The prognostic significance of the pattern of recurrence is also similar, with worse survival seen for combined local and distant recurrence. The remarkable consistency of pattern, timing, and prognostic significance of clinical features at diagnosis and recurrence despite geographically and therapeutically disparate study populations strongly supports the universality of the observations made both the CESS/EICESS and COG teams. Themost intriguing question raised by the analysis of Stahl et al. is what treatment was used in the few patients who were able to survive after recurrence? Neither the European nor the North American analyses included details regarding the therapy after recurrence. It is possible that some received topotecan with cyclophosphamide (TC), which has a 33–35% response rate in recurrent Ewing sarcoma [2,3]. More recently, irinotecan and temozolomide (IT) has become a common treatment option, with response rates of 29–63% [4,5]. Although the response rates to TC and IT are encouraging and suggest potential curative treatment approaches for recurrent Ewing sarcoma, each of these four response reports have limitations [2–5]. By their nature, phase II trials have a relatively small sample size, represent a selected population of recurrent patient, with the possibility to overor under-estimate the probability of long-term survival, and allow additional therapy after initial response evaluation that could also contribute to survival. One particularly controversial additional treatment is high-dose therapy (HDT) for patients with recurrent Ewing sarcoma who respond to second-line therapy. In the absence of a randomized trial, retrospective analyses report positive [6,7], and negative results [8]. Although restricted to newly diagnosed Ewing sarcoma patients with isolated pulmonary metastases rather than after recurrence, EuroEWING99will finally determine the role of HDT in a randomized study [9]; its results will likely be extrapolated to the management of recurrent Ewing sarcoma. The outcome reported by Stahl et al. and prior publications of phase II response to either TC or IT confirm one unavoidable conclusion: the treatment options available to patients with recurrent Ewing sarcoma are inadequate and unlikely to lead to long-term survival after recurrence. Even patients with favorable clinical features (late or isolated local recurrence) have a postrelapse survival rate of less than 30%. Such poor survival clearly justifies consideration of enrollment in a phase I or II trial at first recurrence. In the face of this cold truth, the search for more effective, biologically based therapy appeared unusually promising over the past few years, with the clinical development of antibodies that block the insulin-like growth factor-1 receptor (IGF-1R). To borrow a phrase fromAlanGreenspan, the Ewing sarcomafield rode a wave of ‘‘irrational exuberance’’ on the therapeutic potential for IGF-1R antibodies. Reality replaced enthusiasm when the response rate to IGF-1R antibodies as single agents proved disappointing [10] and as Pharma’s interest in this class of agents withered. Recent experience with IGF-1R antibodies notwithstanding, patients with recurrent Ewing sarcoma represent an excellent population to evaluate other biologically target therapy in combinationwith either TC or IT, both of which have rather modest toxicity profiles amenable to the addition of non-cytotoxic agents including