|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2004.06.025 on September 13 2004 © 2004 American Society of Clinical Oncology.
Molecular Prognostication for Soft Tissue Sarcomas: Are We Ready Yet?
1 Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, and Mayo Clinic and Mayo Foundation, Rochester, MN; Since the discovery more than a decade ago of the EWS-FLI1 fusion gene resulting from the chromosomal translocation(11;22) in the Ewing's sarcoma family of tumors,1,2 major advances in our understanding of the molecular biology of several sarcomas have taken place at an impressively fast pace. Following the example of hematologic malignancies, in which simple reciprocal chromosomal translocations are also common, the new molecular era is opening not only novel investigative pathways with the cloning of multiple fusion genes3 and the identification of gene expression signatures in sarcomas,4-7 but is also stimulating a desire to quickly translate these achievements into better patient care through either enhanced prognostication or identification of new therapeutic targets. At the same time that the appearance of molecular markers promises highly accurate predictive frameworks for sarcoma diagnosis, prognosis, and therapeutic tailoring, some of these data are also generating controversial and, not uncommonly, contradictory findings. The recent debate around the prognostic value of synovial sarcoma fusion genes is an example of many of the issues encountered in studies of molecular prognostic markers in cancer. Synovial sarcoma is an aggressive tumor of uncertain cellular differentiation that predominantly affects young adults; although having a wide anatomic distribution, it shows a predilection for the limbs. Cytogenetically, synovial sarcoma is characterized in almost all cases by the chromosomal translocation t(X;18)(p11;q11). This chromosomal rearrangement typically leads to fusion of the SYT gene on chromosome 18 with either the SSX1 or SSX2 genes on chromosome X.8 In a ground-breaking retrospective pilot study published in 1998, Kawai et al9 found that patients with localized tumors harboring SYT-SSX1 fusion transcripts had decreased metastasis-free survival. Similar results were later obtained in four other retrospective studies10-13 and were further supported by a large multicenter study involving more than 200 patients, published in 2002 by Ladanyi et al.14 In this issue of the Journal of Clinical Oncology, the same question has been readdressed in another multi-institutional study by Guillou et al15 of 165 patients with synovial sarcoma from the French sarcoma group. In contrast to the previous series, these authors found no association between the type of fusion gene and clinical outcome. Moreover, they also report that histologic grade was the strongest predictor for survival in a multivariate model. So what factors can account for these differing results, and what valuable lessons can be taken from these studies? One explanation for the discordant results of this study compared with previous reports is the retrospective design of these studies and their unavoidable shortcomings, including missing data and the possibility of several selection biases.16-18 Despite the fact that known patient and tumor characteristics were apparently similar (such as age, sex, or tumor size), the use of different therapeutic approaches and the presence of other potential confounders (such as comorbidities, access to medical care, or varying size and expertise of treatment centers) should not easily be disregarded. As pointed out by others, unavailability of clinical data in retrospective studies is unlikely to be random in nature16-18 and may have different causes among different studies. Although the various published studies in synovial sarcoma seem to have accrued cases over a similar time period and most likely have used comparable therapeutic approaches, the understandably limited information available, particularly regarding precise details of treatment, make it impossible to exclude with certainty the possibility that therapy, for example in the French study, might have been more successful, thereby obfuscating the prognostic value of fusion gene typing. Different laboratory techniques used for detection of the fusion genes are another well-known cause of variation among studies. Although Guillou et al15 used two distinct analyses in parallel to confirm their results—a polymerase chain reaction (PCR)–based assay and single-strand conformation polymorphism—Ladanyi et al14 evaluated samples in a more heterogeneous fashion in varying reference laboratories, using either PCR or fluorescence in situ hybridization; other studies used variable PCR-based approaches.10-13 Confirmatory DNA sequencing of the PCR products was inconsistently performed in some of these studies, and, more importantly, none of the groups confirmed their results in a second independent laboratory evaluation. These issues are even more troubling when one considers that a recent investigation performed by one of the groups that previously attributed prognostic value to the type of fusion transcript10 found that both transcripts can be expressed in the same tumor.19 Overall, therefore, it seems that the clinical relevance of fusion gene typing in synovial sarcoma remains uncertain at the present time. Statistical analyses in studies of prognostic factors are another rather important, but sometimes disregarded, source of variation.16,17,20 In all of the above studies, data modeling, multivariate models, and the clinical outcomes evaluated were dissimilar, therefore making direct comparisons challenging. Considering that even standard prognostic factors for synovial sarcoma have not been well-defined or validated, it is difficult to be certain of which prognostic factors should be included in any multivariate model. Common sense tells us that important sarcoma variables such as tumor size, histologic grade, or the status of the surgical margins should at least be considered; however, uniform standards, including data stratification or optimized cutoff values for continuous variables,21 should be established a priori. Furthermore, some variables have not been considered in the same way in different centers. Histologic grade is a good example of this.22 Although it has been shown that a well-formulated grading system23 confers superior (albeit not perfect) prognostic information in many types of soft tissue sarcoma, it is often not feasible to generate meaningful histologic grading from resection specimens in a center in which preoperative radiotherapy or neoadjuvant chemotherapy are standard. Not only do these preoperative procedures alter mitotic counts, induce tumor necrosis, or even change the degree of tumor differentiation, but the preoperative biopsies performed in these centers often provide insufficient material for reliable histologic grading assessment, not least because of the inevitable issue of sampling variation.24,25 The study of prognostic molecular markers—in particular, the type of fusion gene for determining prognosis in soft tissue sarcomas—has also been addressed in three other major types of sarcoma. In the Ewing's/peripheral neuroectodermal tumor family of tumors, an initial study performed by Zoubek et al26 suggested that the EWS-FLI1 type I fusion gene was associated with longer relapse-free (either metastasis or local recurrence) survival in patients with localized disease compared with other types of fusion gene by univariate analysis. To confirm these findings, a multicenter study including 112 patients was performed by de Alava et al.27 Despite the fact that these authors found an association between type I EWS-FLI and overall survival by multivariate analysis, relapse-free survival was not evaluated. A third study addressing a closely related question was carried out by Ginsberg et al.28 These authors compared several clinical variables between patients with tumors harboring either EWS-FLI1 or EWS-EGR fusion genes. No prognostic value was attributed to the fusion genes when evaluated for event-free and overall survival by univariate analysis.28 Similar to the synovial sarcoma studies, variations in study design, data modeling, statistical analyses, and evaluated outcomes make definitive conclusions equally difficult. Investigations regarding the prognostic value of the fusion genes PAX3-FKHR and PAX7-FKHR in alveolar rhabdomyosarcoma were initially carried out in a pilot study involving 34 patients by Kelly et al29; these authors observed better clinical outcomes for the PAX7-FKHR group by univariate analysis. These results were later confirmed in a second retrospective study of a large number of patients from the Children's Oncology Group.30 A major merit of this second study was that all laboratory analyses were performed independently in three separate centers for confirmation of the results. Despite these promising findings, the clinical utility of assessing the type of fusion gene in alveolar rhabdomyosarcoma needs to be further confirmed in an independent series and prospectively evaluated using well-defined cohorts. A single study has addressed the prognostic value of the type of fusion gene in myxoid liposarcoma, a common adult soft tissue sarcoma characterized by the presence of the TLS-CHOP fusion gene in 95% of cases. The authors were unable to find any association between the structure of the fusion gene and disease-specific survival but confirmed the value of careful histologic assessment for prognostication.31 As proposed by Altman and Lyman16,17 and later modified by Hall and Going,32 validation of putative prognostic factors should be performed in three major phases: exploratory studies, retrospective confirmatory investigations, and prospective studies. The first phase comprises the study of a plausible prognostic factor in relation to specific outcomes and in comparison with prognostic data already available. These studies are also important for hypothesis formulation. The second phase is designed to confirm the initial observations in larger and well-defined cohorts. In this phase, not only are data stratification, cutoff points, and multivariate models defined, but the results are validated in a second independent series.33 Also, the predictive accuracy of the new prognostic factor with regard to the clinical outcome for which its use is intended should be carefully evaluated.33 This is of greatest importance when considering therapeutic tailoring for individual patients based on prognostic factors—these factors may only explain a small fraction of the observed variation in clinical outcome despite their low P values or high hazard ratios. As an example, a recently proposed multivariate prognostic model for cutaneous melanoma consisting of five variables only explained 12% of the clinical outcomes in a series of more than 1,000 patients.34 In the second phase, the search for more cost-effective surrogate markers must also be considered. For example, if mitotic counts or Ki-67 antigen expression show a high degree of correlation with a certain type of fusion gene—as has been observed in synovial sarcoma,10,11 alveolar rhabdomyosarcoma,35 and Ewing's sarcoma36—it might be much easier (and cheaper) to ensure widespread adoption of these proliferative indices as a prognostic marker, rather than more complex and expensive molecular genetic analyses. The third phase comprises multi-institutional trials in which the effect of the new prognostic factor is prospectively evaluated in relationship to other important factors, including therapeutic choices. Furthermore, a prognostic factor should not only influence clinical decisions, but also these decisions should be relevant for modifying patients' care and clinical outcome. In this context, the value of putative molecular prognostic factors in soft tissue sarcomas is still under intense investigation, therefore making their immediate use in clinical practice both premature and of uncertain utility at the present time. Two additional issues need to be considered when dealing with studies of molecular prognostic factors in sarcomas. One is the possibility of publication bias,37 a well-known phenomenon in which significant results are much more likely to be published than nonsignificant or negative findings, therefore inflating the literature with biased claims. The second is the problem of multiplicity,17,38 a phenomenon that occurs when specific hypotheses are not defined a priori and multiple comparisons are carried out a posteriori to identify associations among different factors. This also leads to an increased likelihood for finding spurious or misleading associations, as has been commonly observed in gene expression profiling studies.39 Despite the many issues involved in the study of molecular prognostic factors in sarcomas and despite the uncertainty that persists concerning the clinical relevance of fusion genes in these tumors, major biologic insights can be gained from this work. Additionally, the identification of therapeutically relevant molecular markers, including oncogenic protein tyrosine kinases susceptible to small molecule inhibitors, has become a paradigm for studies addressing novel molecular prognostic factors in some sarcomas, such as gastrointestinal stromal sarcoma,40,41 and in certain other solid tumors.42,43 In this regard, an era in which molecular prognostication is tightly linked to therapeutic tailoring is unfolding. Authors' Disclosures of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. REFERENCES 1. Delattre O, Zucman J, Plougastel B, et al: Gene fusion with an ETS DNA-binding domain caused by chromosome translocation in human tumours. Nature 359: 162-165, 1992[CrossRef][Medline] 2. Zucman J, Delattre O, Desmaze C, et al: Cloning and characterization of the Ewing's sarcoma and peripheral neuroepithelioma t(11;22) translocation breakpoints. Genes Chromosomes Cancer 5: 271-277, 1992[Medline] 3. Ladanyi M, Bridge JA: Contribution of molecular genetic data to the classification of sarcomas. Hum Pathol 31: 532-538, 2000[CrossRef][Medline] 4. Nielsen TO, West RB, Linn SC, et al: Molecular characterisation of soft tissue tumours: A gene expression study. Lancet 359: 1301-1307, 2002[CrossRef][Medline] 5. Oliveira AM, Fletcher JA: CHIPing soft tissue tumors: Will the paradigms be changed? Adv Anat Pathol 10: 1-7, 2003[Medline] 6. Dirix LY, van Oosterom AT: Gene-expression profiling to classify soft-tissue sarcomas. Lancet 359: 1263-1264, 2002[CrossRef][Medline] 7. Segal NH, Pavlidis P, Antonescu CR, et al: Classification and subtype prediction of adult soft tissue sarcoma by functional genomics. Am J Pathol 163: 691-700, 2003 8. Ladanyi M: Fusions of the SYT and SSX genes in synovial sarcoma. Oncogene 20: 5755-5762, 2001[CrossRef][Medline]
9. Kawai A, Woodruff J, Healey JH, et al: SYT-SSX gene fusion as a determinant of morphology and prognosis in synovial sarcoma. N Engl J Med 338: 153-160, 1998
10. Nilsson G, Skytting B, Xie Y, et al: The SYT-SSX1 variant of synovial sarcoma is associated with a high rate of tumor cell proliferation and poor clinical outcome. Cancer Res 59: 3180-3184, 1999 11. Inagaki H, Nagasaka T, Otsuka T, et al: Association of SYT-SSX fusion types with proliferative activity and prognosis in synovial sarcoma. Mod Pathol 13: 482-488, 2000[CrossRef][Medline] 12. Mezzelani A, Mariani L, Tamborini E, et al: SYT-SSX fusion genes and prognosis in synovial sarcoma. Br J Cancer 85: 1535-1539, 2001[CrossRef][Medline] 13. Panagopoulos I, Mertens F, Isaksson M, et al: Clinical impact of molecular and cytogenetic findings in synovial sarcoma. Genes Chromosomes Cancer 31: 362-372, 2001[CrossRef][Medline]
14. Ladanyi M, Antonescu CR, Leung DH, et al: Impact of SYT-SSX fusion type on the clinical behavior of synovial sarcoma: A multi-institutional retrospective study of 243 patients. Cancer Res 62: 135-140, 2002 15. Guillou L, Coindre JM, Bonichon F, et al: Histologic grade, but not SYT-SSX fusion type, is an important prognostic factor in synovial sarcoma patients. J Clin Oncol 22:4040-4050, 2004 16. Altman DG, Lyman GH: Methodological challenges in the evaluation of prognostic factors in breast cancer. Breast Cancer Res Treat 52: 289-303, 1998[CrossRef][Medline] 17. Simon R, Altman DG: Statistical aspects of prognostic factor studies in oncology. Br J Cancer 69: 979-985, 1994[Medline]
18. McGuire WL: Breast cancer prognostic factors: Evaluation guidelines. J Natl Cancer Inst 83: 154-155, 1991 19. Yang K, Lui WO, Xie Y, et al: Co-existence of SYT-SSX1 and SYT-SSX2 fusions in synovial sarcomas. Oncogene 21: 4181-4190, 2002[CrossRef][Medline]
20. Concato J, Feinstein AR, Holford TR: The risk of determining risk with multivariable models. Ann Intern Med 118: 201-210, 1993
21. Altman DG, Lausen B, Sauerbrei W, et al: Dangers of using "optimal" cutpoints in the evaluation of prognostic factors. J Natl Cancer Inst 86: 829-835, 1994 22. Oliveira AM, Nascimento AG: Grading in soft tissue tumors: Principles and problems. Skeletal Radiol 30: 543-559, 2001[CrossRef][Medline]
23. Guillou L, Coindre JM, Bonichon F, et al: Comparative study of the National Cancer Institute and French Federation of Cancer Centers Sarcoma Group grading systems in a population of 410 adult patients with soft tissue sarcoma. J Clin Oncol 15: 350-362, 1997 24. Brown FM, Fletcher CDM: Problems in grading soft tissue sarcomas. Am J Clin Pathol 114:S82-S89, 2000 (suppl) 25. Association of Directors of Anatomic and Surgical Pathology Recommendations for the reporting of soft tissue sarcomas. Mod Pathol 11: 1257-1261, 1998[Medline]
26. Zoubek A, Dockhorn-Dworniczak B, Delattre O, et al: Does expression of different EWS chimeric transcripts define clinically distinct risk groups of Ewing tumor patients? J Clin Oncol 14: 1245-1251, 1996
27. de Alava E, Kawai A, Healey JH, et al: EWS-FLI1 fusion transcript structure is an independent determinant of prognosis in Ewing's sarcoma. J Clin Oncol 16: 1248-1255, 1998
28. Ginsberg JP, de Alava E, Ladanyi M, et al: EWS-FLI1 and EWS-ERG gene fusions are associated with similar clinical phenotypes in Ewing's sarcoma. J Clin Oncol 17: 1809-1814, 1999
29. Kelly KM, Womer RB, Sorensen PH, et al: Common and variant gene fusions predict distinct clinical phenotypes in rhabdomyosarcoma. J Clin Oncol 15: 1831-1836, 1997
30. Sorensen PH, Lynch JC, Qualman SJ, et al: PAX3-FKHR and PAX7-FKHR gene fusions are prognostic indicators in alveolar rhabdomyosarcoma: A report from the Children's Oncology Group. J Clin Oncol 20: 2672-2679, 2002
31. Antonescu CR, Tschernyavsky SJ, Decuseara R, et al: Prognostic impact of P53 status, TLS-CHOP fusion transcript structure, and histological grade in myxoid liposarcoma: A molecular and clinicopathologic study of 82 cases. Clin Cancer Res 7: 3977-3987, 2001 32. Hall PA, Going JJ: Predicting the future: A critical appraisal of cancer prognosis studies. Histopathology 35: 489-494, 1999[CrossRef][Medline] 33. Altman DG, Royston P: What do we mean by validating a prognostic model? Stat Med 19: 453-473, 2000[CrossRef][Medline]
34. Vollmer RT, Seigler HF: Using a continuous transformation of the Breslow thickness for prognosis in cutaneous melanoma. Am J Clin Pathol 115: 205-212, 2001 35. Collins MH, Zhao H, Womer RB, et al: Proliferative and apoptotic differences between alveolar rhabdomyosarcoma subtypes: A comparative study of tumors containing PAX3-FKHR or PAX7-FKHR gene fusions. Med Pediatr Oncol 37: 83-89, 2001[CrossRef][Medline]
36. de Alava E, Panizo A, Antonescu CR, et al: Association of EWS-FLI1 type 1 fusion with lower proliferative rate in Ewing's sarcoma. Am J Pathol 156: 849-855, 2000 37. Montori VM, Smieja M, Guyatt GH: Publication bias: A brief review for clinicians. Mayo Clin Proc 75: 1284-1288, 2000[Abstract] 38. Tukey JW: Some thoughts on clinical trials, especially problems of multiplicity. Science 198: 679-684, 1977 39. Ransohoff DF: Rules of evidence for cancer molecular-marker discovery and validation. Nat Rev Cancer 4: 309-314, 2004[CrossRef][Medline]
40. Singer S, Rubin BP, Lux ML, et al: Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol 20: 3898-3905, 2002
41. Heinrich MC, Corless CL, Demetri GD, et al: Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin Oncol 21: 4342-4349, 2003
42. Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350: 2129-2139, 2004
43. Paez JG, Janne PA, Lee JC, et al: EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy. Science 304: 1497-1500, 2004
Related Article
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
