Burden of Sarcoma May Be Underestimated in the National Cancer Database

Submitted by admin5 on Fri, 11/17/2017 - 16:31

Data from the National Cancer Database, which is heavily dependent on institutional reporting, could be misleading about the burden of sarcomas in the United States, according to a presentation at the Connective Tissue Oncology Society 2017 Annual Meeting (November 8-11, 2017; Maui, Hawaii).

Institutions have their own policies regarding whether difference cancers are coded by disease type or organ site. Due to this inconsistency of World Health Organization disease coding across United States institutions, there is potential for miscalculations of the global burden of many cancer types.

A recent study made efforts to characterize inaccuracies and errors in sarcoma coding practices that have resulted in incorrect sarcoma surgical diagnostic codes and tumor registry data at a single institution. Researchers analyzed data from one of the five surgeons who treated patients with sarcoma. Three of the surgeons were surgical oncologists, while the remaining surgeons were orthopedic oncologists.

Researchers noted that World Health Organization coding was compared with the according pathology report.

A total of 1237 pathologically confirmed sarcoma cases were analyzed from January 2012 through December 2016. Among these cases, only 63% (n = 764) were coded as sarcoma, whereas 16% (n = 180) were coded as non-cancers and 22% (n = 260) were coded by organ site.


Related Content

External Beam Radiotherapy Effective for Skin Kaposi Sarcoma Treatment

SEOM Releases Clinical Management Guidelines for Soft-Tissue Sarcoma


Furthermore, researchers acknowledged that more than 50% of breast angiosarcomas were inputted as breast cancers. Similarly, nearly 50% of gastrointestinal stromal tumors were not coded correctly.

In their concluding remarks, researchers asserted that their findings suggest a trend toward misleading information of sarcoma burden in the United States, based on data from the National Cancer Database. However, they noted that their study only considered data from one institution; if these results are to be generalized across other institutions, further research is needed to determine whether a similar trend occurs in larger datasets.—Zachary Bessette