UICC World Cancer Congress 2006

There are numerous challenges to the continuation of the value of TNM staging. It is critical that the system maintain its integrity in anatomic staging while at the same time providing clinicians and patients with clinically relevant information. In addition, it is key that the process of collecting staging data be streamlined and made as efficient as possible. In the United States, there have historically been three separate staging systems with independent rules and data elements. To satisfy these needs, the AJCC in collaboration with the NCI SEER program, the CDC, NAACCR, NCRA, and NCIC developed and implemented the Collaborative Staging System.

The AJCC / UICC TNM cancer staging system codified in the Cancer Staging Manual of the American Joint Committee on Cancer (AJCC) provides clinicians and population scientists with a validated mechanism to define the extent of cancer at the time of disease presentation. Clinicians use this information to determine prognosis and to guide treatment decisions. Clinical researchers use staging to evaluate the effect of treatment of similar groups of patients, and population scientists use it to study the incidence of cancer and changes in the presentation of cancer over time.

AJCC/UICC TNM staging is based primarily on the anatomic extent of primary cancer, involvement of regional lymph nodes, and presence of distant metastases. One hallmark of the system is that it undergoes periodic modification to incorporate new knowledge about factors that affect cancer prognosis while maintaining sufficient ties to previous versions to allow comparisons across time. In addition to this system, population scientists use two other staging systems that may be more readily applied on a population basis. These are Summary Stage which classifies cancers as localized, regional, or distant, and the Extent of Disease (EOD) system used by the NCI SEER program.

Clinicians and patients increasingly want information that accurately predicts the outcome and treatment response for individual cancer patients. To date, the staging system has been sufficiently robust to accommodate the needs of both clinicians and population scientists. However, the groupings of cases by AJCC/UICC TNM staging, and by Summary Stage and EOD often include patients with sufficiently different prognosis to make these groupings of less clinical utility. Furthermore, newly discovered non-anatomic characteristics of cancers associated with cancer prognosis and response to treatment may provide prognostic information that improves the precision of anatomic staging, or even supplant the need for anatomic information. Such factors include levels of expression of individual markers measured in cancer tissue and body fluids, panels of such markers, and genomic analysis of tumor and germline tissue. In addition, statistical sciences are increasingly sophisticated to deal with available large datasets of information on cancer patients and may allow more precise estimates of individual prognosis than the current stage groupings.

Limitations of the existing staging systems have led other groups to develop alternative prognostic and predictive systems. One example of this is Adjuvant!® for breast and colon cancer (www.adjuvantonline.com). This uses many of the factors that define staging including size and lymph node status plus other factors such as hormone receptor status, tumor grade and co-existing illness to estimate the chance of recurrence and survival at 10 years after diagnosis without treatment, the number that would be expected to recur or die of cancer and other causes, and based on data from clinical trials and meta-analysis, the number of recurrences or lives expected to be spared by use of specific systemic treatments (chemotherapy, and chemotherapy and / or hormonal therapy in the case of breast cancer). Other groups have developed similar systems for defining cancer prognosis in other diseases. One group has developed nomograms based on anatomic extent and other factors for a number of cancer types including prostate cancer, sarcoma, and others.

However useful such systems are for individual patients, the decentralized propagation of prognostic systems for individual diseases jeopardizes other benefits of centralized, standardized data collection on cancer stage, and may lead substantial added expense of data collection for non-compatible data systems. Accepted algorithms applied across the entire community are vital to understanding the incidence of cancer, the changing scope of cancer that is affected by many factors including changing demographics, and improvement in screening and treatment. It will be unfortunate if parallel systems requiring separate data collection schemas are used by clinicians and population scientists. Conversely, if such standardized systems do not meet the needs of clinicians and patients, there will be increased efforts placed on alternative systems.

These challenges make it necessary to reassess the role of purely anatomic staging, evaluate the potential to expand its scope to include predictive factors, and define the best ways to use non-anatomic factors to supplement the information provided by anatomic extent of disease. Further, strong support should be provided for development of statistical methodologies that may provide precise estimates of outcome to individuals. An equal imperative is to assure that any such factors incorporated into staging systems are clearly defined, that the value of their use is fully validated, and that the factors can be reproducibly measured and made widely available.

The Collaborative Staging System:

The AJCC and partner standards setting organizations addressed these issues in 2004 with the implementation of the Collaborative Staging System (see www.cancerstaging.org for a complete list of partner organizations). This is a data standard for collecting and storing the raw data necessary to derive stage, including summary stage, SEER Extent of Disease and AJCC stage. The system also provides the computerized algorithm to derive all three staging systems.

In establishing this system, major rules differences between the TNM, Summary Stage and EOD systems were resolved. The timing rule adopted was to include all information from the time of diagnosis through 4 months. In addition, it was necessary to resolve ambiguities in the terminology used in the staging systems.

Another key decision was to allow combining of pathologic and clinical information into a single best stage. This accommodates the realities of cancer staging in clinical medicine where it is often possible to have complete pathologic data on one element but not another. Consider for example the case of a woman with a 1.5 cm breast cancer that has been resected, but for which there is no axillary staging. Under the prior TNM system, the case would have to be staged by purely clinical means or be unstagable. Using the Collaborative Staging System the cancer is staged as pT1c cN0 cM0.

Key to this system is that the data elements are stored using extensible markup language (XML) tags in an open architecture that makes them easily available for analysis by any user. Data on factors such as size of tumor and number of lymph nodes are stored in their primary form as continuous variables where applicable, and not grouped by the AJCC Stage Group cutoffs. In addition, the system includes multiple data fields for cancer site specific information that can be used to add information on non-anatomic factors. The XML tagging also allows exporting to other data systems to accommodate additional information on other factors or markers.

Finally, the Collaborative Staging system allows the collection of non-anatomic factors for each disease as necessary. These collected in what are termed “site specific factor” fields. SSF include PSA and Gleasons Score in prostate cancer, tumor thickness for melanoma, and hormone receptor status in breast cancer. The current system has space for up to 6 SSF's for each disease. However, it is anticipated that revisions for the next edition of the Cancer Staging Manual will incorporate more SSF's as new non-anatomic factors are incorporated into staging.

The Collaborative Staging System is now in use by all hospital and population registries in the United States. Complete description, specifications and technical information for the Collaborative Staging System are available at the AJCC website at www.cancerstaging.org.