We read the article by Lawrence et al1 with interest. The authors found an improvement in the overall survival of patients with glioblastoma between 2001 and 2007; however, no survival improvement was seen among patients aged ≥70 years.
We think that a difficult detection and collection of cases, particularly among older patients, can influence analysis on survival and treatment of glioblastoma patients.
Data from the SEER-17 Registries2 showed that the proportion of microscopically confirmed cases of glioblastoma (International Classification of Diseases, Oncology [ICD-O] 9440) was higher among patients with age <70 years than among older patients. In the period from 2001 to 2007, 96% of patients younger than 70 years and 81% of patients aged 70 years and older had microscopic confirmation.
Glioblastoma is basically a histological diagnosis: the exclusion of cases with clinical or radiographic diagnosis “Glioma, Not Otherwise Specified“ without histologic definition could hide the real incidence of tumors in older patients, where microscopic examination is generally less frequent than in young patients. Sometimes the term “glioma” includes glioblastoma multiforme and anaplastic astrocytoma,3 and this may cause some conflicting results. The analysis by cell type is difficult because of the lack of details; thus, many studies reported information on brain tumors rather than on specific histologic type of brain tumors. Recently, diagnostic methods and tumor classification have improved histologic diagnosis of specific brain tumors4; during the 1990s, changes in brain tumor classification and in ICD-O coding occurred, but how these changes influenced histology-specific rates is unclear.4 The variation in diagnosis and coding could be influenced by the rates of glioblastoma over the time.
Some studies suggested that the use of magnetic resonance imaging could have influenced the increase in incidence of brain tumor, particularly in patients with older age3 and that in elderly patients the histologic confirmation of lesion was more available recently than in the past. A possible explanation may be the technological improvement in diagnosis, which allows the characterization of glioblastoma that was previously not easily detectable.
Difficulty in detection of cases, low proportion of microscopic characterization, tumor misclassification, or unspecific coding could influence survival analysis in older glioblastoma patients.