Neuro Oncol (2013) 15 (suppl 2): ii1-ii56. doi: 10.1093/neuonc/not151
Page ii2 of CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States, 2006–2010 currently reads “NPCR provided data on 313,780 primary brain and CNS tumors diagnosed from 2006 to 2010. NPCR cancer registries agree to participate in the CBTRUS Statistical Report and to pass certain data quality standards required by NPCR, thus allowing CBTRUS to receive and analyze the data. An additional 12,991 primary brain and CNS tumor case records for the time period were obtained from SEER. These data were combined into a single data set for analyses. A total of 8,175 records (2.4%) were deleted from the final analytic data set because of one or more of the following reasons: invalid site/histology combinations, duplicate records that included a less accurate reporting source than microscopic confirmation (e.g. radiographic versus microscopic confirmation), duplicate records for bilateral vestibular schwannoma or meningioma, duplicate record for recurrent disease, and errors in time sequence of diagnosis. The final analytic data set included 326,711 records from 50 population-based central cancer registries, including 49 state registries and the District of Columbia.”
The data information provided is incorrect and should read “NPCR provided data on 316,188 primary brain and CNS tumors diagnosed from 2006 to 2010. NPCR cancer registries agree to participate in the CBTRUS Statistical Report and to pass certain data quality standards required by NPCR, thus allowing CBTRUS to receive and analyze the data. An additional 12,931 primary brain and CNS tumor case records for the time period were obtained from SEER. These data were combined into a single data set for analyses. A total of 2,408 records (.73%) were deleted from the final analytic data set because of one or more of the following reasons: invalid site/histology combinations, duplicate records that included a less accurate reporting source than microscopic confirmation (e.g. radiographic versus microscopic confirmation), duplicate records for bilateral vestibular schwannoma or meningioma, duplicate record for recurrent disease, and errors in time sequence of diagnosis. The final analytic data set included 326,711 records from 50 population-based central cancer registries, including 49 state registries and the District of Columbia.”
Why did the numbers change?
NPCR
The report number of 313,780 was incorrectly reported. This number actually represents the total number of observations from NPCR after the edits program was run. The revised number of 316,188 is the total number of observations from NPCR before the edits program was run. Thus, 316,188–313,780 = 2,408 observations that were deleted.
SEER
The report number of 12,991 was a typo. It was supposed to read 12,931, the revised number, from the beginning. It is too difficult to say how many observations were deleted from the SEER data because the initial SEER set contained other tumors besides brain tumors. The SEER data that was initially entered into the edits program contained 80,022 observations, but most of these were not brain tumors. This is why the deleted observations reported should only include those deleted from NPCR.
Total
The total number of observations before the edits program was run is therefore 316,188 + 12,931 = 329,119. The correct total number of observations after the edits program was run is 329,119–2,408 = 326,711, which is the figure used in all of the reports and tables. Thus, the percentage of deleted observations is (2,408 / 329,119) * 100 = 0.73.
The conclusions of the report remain unchanged. The authors apologize for these errors.