The concept of staging cancer originated in the early twentieth century as a way of understanding the extent and severity of disease and thereby estimating prognosis. These early efforts at classifying cancer were anatomically based and as such this has provided the foundation for modern staging. The tumor, node, metastasis classification system of staging was first introduced in the 1940s as a standardized method of anatomic staging and used assessment of primary tumor extent, lymph node involvement, and distant disease. Through an objective assessment of these three parameters, patients were able to be grouped in a rough estimation of the severity of the disease. This approach was adopted by the American Joint Committee on Cancer (AJCC) and provides the underpinning of the modern staging system for solid tumors. Refinements have been made for individual malignancies whereby other pertinent factors such as serum markers and histologic grade have been included in staging.

As therapeutic strategies have improved and become more complex, stage has played an important role in treatment decisions, in comparisons of results across institutions and across time, and in assessment of prognosis. The decision to include elements of multimodal care is often based on stage, such as in adjuvant therapy for malignant melanoma or neoadjuvant approaches to rectal cancer. Additionally, standardized staging is important in classifying patients for clinical research so that findings about treatment and outcome can be applied to the appropriate population of patients. Finally, staging plays an important role in cancer epidemiology. Measuring and defining the rate of early and late stage presentation can help to better direct prevention, screening, and diagnosis, as well as the application of the correct therapy at the correct point in time.

Staging of soft tissue sarcoma presents some unique difficulties, which Maki et al. illustrate.1 Sarcoma is a relatively rare disease and is both anatomically and histologically diverse in origin. The authors harness the power of the Memorial Sloan-Kettering Cancer Center sarcoma database to illustrate some of the pitfalls of the current staging system, as well as opportunities for improvement. The data presented emphasize the importance of histology, size, and location in classifying patients. The relative insensitivity of some elements of the current system are demonstrated, such as the use of a binary system to classify tumor size (<5 cm vs. larger), which in reality is a continuous rather than discontinuous variable.

The essential role of histology in treatment and prognosis is illustrated by the recognition of gastrointestinal stromal tumors as a group of tumors which behave and are treated differently than other forms of sarcoma. In the current (seventh) version of the AJCC manual, these tumors have been given their own staging schema. This is important in light of the recognition of their unique behavior, as well as the availability of targeted therapeutics for this group of tumors. It is likely that with increased understanding of the clinical outcomes and molecular underpinnings of other histologies, additional tumors will be split off to create specific staging that will aid clinicians in providing prognostic information and making therapeutic decisions for patients with soft tissue sarcoma.

Recent advances in the laboratory have led to a greater appreciation of the importance of genomics in determining prognosis and making treatment decisions for many forms of cancer. It is clear that in breast cancer, receptor status and gene expression play an important role in treatment and outcome, even beyond the well-established difference in survival, stage for stage, between hormone receptor-positive and -negative patients.2 A recent study described a strong association between specific gene signatures in colorectal cancer and response to certain therapies and overall prognosis.3 Although these advances have yet to be formally incorporated into staging systems, it is clear that the information provided through such analysis compliments traditional anatomic staging in an effort to provide true precision therapy for cancer.

Although the relative rarity of sarcoma has limited the availability of the large data sets that enable this sort of clinical correlation with genomic analysis, strides have nonetheless been made in providing patients with more accurate information about their disease. One approach to the heterogeneity of soft tissue sarcoma has been the development of nomograms for prognosis. Over the last decade, multiple versions of these tools have been created that assign points on the basis of various clinical and pathologic characteristics of patients with sarcoma. These points can then be translated into a value for expected chance of survival or freedom from relapse. The earliest nomograms grouped together patients with tumors originating from a variety of anatomic locations, despite the important differences between extremity and retroperitoneal or visceral sarcoma. More recent nomograms have better stratified patients by both the location of the primary tumor as well as histologies. Although most nomograms have been derived from single-institution databases, a recent publication has created a nomogram that used data from multiple institutions that were been validated with a fourth institution data set with good predictive power for overall and disease-free survival for patients with resected retroperitoneal sarcomas.4 Improvement of nomograms and the development of other sophisticated instruments such as Bayesian belief networks will further aid patients and caregivers alike in making treatment decisions and planning for the course of the disease.

As our understanding of soft tissue sarcoma increases, both from the standpoint of clinical outcomes and its molecular basis, further refinements to the classification of patients can be anticipated. New tools will become available that will allow caregivers to counsel patients regarding disease treatment and prognosis in a more precise manner. However, a balance must be drawn between the grouping of a heterogeneous patient population (“lumping”) and the recognition of unique clinical, pathologic, and genomic characteristics of patients and tumors (“splitting”). Techniques of genomic analysis that allow individualized approaches to tumors must be combined with clinical experience, especially as manifested in decision tools. Understanding both the commonalities and differences of patients with this disease will allow for better clinical care and facilitate academic endeavors.