|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology
Prognostic Factors for Sarcomas: Hard and SoftMemorial Sloan-Kettering Cancer Center, New York, NY WHY DO ONCOLOGISTS study prognostic factors? There are three reasons. First, we need to counsel our patients and their families about their prognosis so that they can make informed choices about treatment. We typically estimate prognosis for a patient as a probability of survival or event-free survival at a specified interval from diagnosis. We recognize that such estimates of probability have validity for groups of patients with similar presentation. For an individual patient, such estimates may be confusing. For the individual patient, outcome will be success or failure, not some fraction of success. Individuals given the same estimate of success will make radically different choices. Nonetheless, these estimates are of great importance to patients. Second, if our estimates of probability are sufficiently robust, they may allow us to assign patients to different treatments. We may offer therapy of lesser intensity to patients with a more favorable prognosis and therapy of greater intensity with an inferior prognosis. This has the potential to avoid toxicity to one group of patients, and to expose to potentially greater toxicity only those patients whose inferior prognosis justifies such risk. Typically, our paradigm for this strategy requires two steps. Initially, factors of potential prognostic significance are identified by retrospective analysis. These factors are then tested by prospective identification to ensure validity. Third, identification of factors that predict outcome may lead to identification of novel targets for therapy. In addition, the precise delineation of prognostic factors allows us to establish benchmarks to evaluate new therapies with assurance that we are comparing similar cohorts of patients. This issue of the Journal of Clinical Oncology includes two reports of meticulous registries of large groups of patients, which allow the identification of prognostic factors with great reliability. The German-Austrian-Swiss group has studied osteosarcoma (OS) for two decades.1 The authors note that registration of patients to their studies closely parallels national registry data. This analysis is based on a population-based analysis which captured virtually all younger patients with OS and recorded potential prognostic factors prospectively at the time of registration. The Memorial Sloan-Kettering Cancer Center (MSKCC) Multidisciplinary Soft Tissue Sarcoma Group has registered every patient with soft tissue sarcoma (STS) seen since 1982.2 Although MSKCC is a tertiary referral center, the value of this analysis is enhanced by the large size of this group of prospectively registered patients and the uniformity of disease assessment. The Cooperative Osteosarcoma Study (COSS) Group reports an analysis based on 1,702 patients enrolled on sequential studies from 1980 through 1998. Almost all previous analyses of prognostic factors in OS have focused on younger patients with primary localized OS of the extremity. The COSS Group enrolled all patients with a diagnosis of OS, including older patients, patients with metastatic disease, patients with axial primaries, and patients with OS arising after prior treatment for malignancy (secondary OS). This has allowed them to provide new and important information about OS. They have shown that older patients with OS of the extremity do as well as younger patients. The overall inferior prognosis of older patients seems to be related to a higher probability of axial primary sites. They have shown that patients with resectable secondary OS have the same chance for cure as patients with resectable primary OS. The apparently inferior prognosis for OS arising after treatment for another malignancy can be explained by a higher probability of unresectable primary site. The population from which they drew their analysis was large enough to allow multivariate analysis, and they have identified factors that predict outcome with substantial statistical significance. For all patients presenting with OS, the COSS Group has also identified primary metastatic disease and axial primary site as unfavorable prognostic features. These features have been previously recognized as predictors of poor outcome, and reports for outcome, if not strategies for treatment, have long been stratified for these factors.3,4 For patients presenting with osteosarcoma of the extremity, they have identified proximal tumors and primary metastasis as unfavorable features. The statistical reliability of the ascription of unfavorable prognosis to proximal site is powerful, but the odds ratio is only 1.5. In other words, the clinical magnitude of the increased risk for patients presenting with a tumor in a proximal extremity is probably not large enough to justify stratification on the basis of this risk factor alone. The other two factors identified by the COSS Group as having an impact both on all patients in aggregate and on patients with primary tumor of the extremity are histologic necrosis after initial chemotherapy and macroscopic residual tumor after surgical resection. Both of these factors are robust in both settings, with odds ratios from 2.5 to 4. They are both, however, not traditional prognostic factors, and cannot be assessed at initial patient presentation. They are assessed only after an initial period of treatment. The patients who were the subject of this analysis received between 9 and 18 weeks of chemotherapy before they underwent definitive surgical resection of the primary tumor and any metastases with evaluation for necrosis and surgical margins. These factors cannot be used for stratification of patients at the time of initial diagnosis, but only to identify patients at increased risk for failure after an initial period of therapy. Other groups have used the same prognostic factors to identify patients at increased risk for failure and attempted to modify this inferior prognosis by modifying treatment after definitive surgery.3 Almost all of these attempts have failed to demonstrate the ability to improve the outcome for these patients. The MSKCC group has performed a fastidious analysis of all patients seen at that center with STS. They chose to exclude treatment variables from modeling. Was this decision correct? The MSKCC group previously published a prospective randomized trial for STS and was unable to detect a benefit for adjuvant chemotherapy.5 A recent meta-analysis of adjuvant treatment for adult STS identified an increase in probability of event-free survival with adjuvant chemotherapy, but the magnitude of the difference was small, and the same analysis could not identify any improvement in overall survival.6 Because the MSKCC analysis focused on death from sarcoma, it is reasonable to expect that omitting treatment variables from modeling would not have a major impact on the conclusions of the study. The MSKCC group has prospectively analyzed size, depth, site, histology, age, and grade. All of these factors have been reported both by the MSKCC investigators and others. All of these factors can be assessed at the time of initial patient presentation. The important and novel contribution of the present study is the simultaneous analysis of all these factors in a very large cohort of patients, studied prospectively and followed meticulously over many years. This makes it possible to construct a nomogram that can predict the probability for death from sarcoma within 12 years from diagnosis. The authors do not have the space to provide the data regarding the odds ratios for each of the factors they have studied, but they can be inferred from the nomogram. Grade is such a powerful predictor of outcome that they have used two linear axes in parallel for low- and high-grade lesions after allocation of all the other prognostic factors. At some points along these axes, a shift from low to high grade will almost triple the probability of death from sarcoma. Does this study identify prognostic factors that are sufficiently robust to allow stratification of patients for treatment? The answer is emphatically yes. The 40-year-old patient with a large, deep, high-grade synovial sarcoma of the trunk has a greater than 80% probability of 12-year sarcoma-specific death. The 40-year-old patient with a 7-cm, superficial, low-grade fibrosarcoma of the lower extremity has a less than 5% probability of 12-year sarcoma-specific death. The authors have validated their model in several different ways and it is highly reliable. These data will allow us to design prospective trials for new treatments for STS with the confidence that we will subject to toxicity only patients at high risk for death from their primary cancer diagnosis. The two series reported in this issue have been extremely successful in their goal of identifying and validating with precision factors that predict outcome in OS and adult STS. The MSKCC study provides the oncology community with a potentially useful tool with which to select patients for novel therapies and gauge their impact. For OS, the challenge is different. Successful adjuvant therapies already exist, and any new treatment must be compared with the effectiveness of current standard therapy. The COSS Group’s study demonstrates the need for dramatically novel therapies to improve the prognosis for patients identified at high risk after initial therapy. It also suggests a need to identify new factors that can be assessed at the time of diagnosis which predict outcome. Potential factors have been identified by retrospective review, and include P-glycoprotein immunohistochemistry, HER2/erbB-2 expression, and loss of heterozygosity of the Rb gene.7-9 These factors are currently undergoing validation in prospective trials. An alternative strategy would be to develop techniques to identify patients who are not achieving favorable necrosis earlier than definitive surgery by using noninvasive assays. The COSS Group and others are testing dynamic magnetic resonance imaging, positron emission tomography, and thallium radionuclide imaging as potential tools for this purpose.10-12 This could potentially allow modification of therapy at an earlier time point, and this might improve the possibilities for modification of therapy to improve the prognosis for poor initial responders. Ultimately, we need to achieve the third objective of prognostic factor identification. We need to identify novel targets for therapy, both for OS and STS. REFERENCES
1.
Bielack SS, Kempf-Bielack B, Delling G, et al: Prognostic factors in high-grade osteosarcoma of the extremities or trunk: An analysis of 1,702 patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol 20: 776-790, 2002
2.
Kattan MW, Leung DHY, Brennan MF: A postoperative nomogram for 12-year sarcoma-specific death. J Clin Oncol 20: 791-796, 2002 3. Meyers P, Heller G, Healey J, et al: Chemotherapy for non-metastatic osteogenic sarcoma: The Memorial Sloan-Kettering experience. J Clin Oncol 10: 5-15, 1992[Abstract]
4.
Meyers PA, Heller G, Healey JH, et al: Osteogenic sarcoma with clinically detectable metastasis as initial presentation. J Clin Oncol 11: 449-453, 1993 5. Casper ES, Gaynor JJ, Hajdu SI, et al: A prospective randomized trial of adjuvant chemotherapy with bolus versus continuous infusion of doxorubicin in patients with high-grade extremity soft tissue sarcoma and an analysis of prognostic factors. Cancer 68: 1221-1229, 1991[CrossRef][Medline] 6. Sarcoma Meta-analysis Collaboration: Adjuvant chemotherapy for localised resectable soft-tissue sarcoma of adults: Meta-analysis of individual data. Lancet 350: 1647-1654, 1997[CrossRef][Medline]
7.
Baldini N, Scotlandi K, Barbanti-Brodano G, et al: Expression of p-glycoprotein in high-grade osteosarcoma in relation to clinical outcome. N Engl J Med 333: 1380-1385, 1995
8.
Gorlick R, Huvos AG, Heller G, et al: Expression of HER2/ErbB-2 correlates with survival in osteosarcoma. J Clin Oncol 17: 2781-2788, 1999
9.
Feugeas O, Guriec N, Babin-Boilletot A, et al: Loss of heterozygosity of the RB gene is a poor prognostic factor in patients with osteosarcoma. J Clin Oncol 14: 467-472, 1996 10. Reddick WE, Wang S, Xiong X, et al: Dynamic magnetic resonance imaging of regional contrast access as an additional prognostic factor in pediatric osteosarcoma. Cancer 91: 2230-2237, 2001[CrossRef][Medline] 11. Franzius C, Sciuk J, Brinkschmidt C, et al: Evaluation of chemotherapy response in primary bone tumors with F-18 FDG positron emission tomography compared with histologically assessed tumor necrosis. Clin Nucl Med 25: 874-881, 2000[CrossRef][Medline] 12. Imbriaco M, Yeh SD, Yeung H, et al: Thallium-201 scintigraphy for the evaluation of tumor response to preoperative chemotherapy in patients with osteosarcoma. Cancer 80: 1507-1512, 1997[CrossRef][Medline]
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
