We present data from long term follow up of 110 patents for our tertiary neuro-surgical centre in the North Midlands, United Kingdom (UK).
Our retrospective cohort includes those completing neurosurgery and chemoradiotherapy. We excluded those with surgical or radiotherapy complications from our cohort as this allowed us to look at the effects of completing adjuvant TMZ on survival outcomes.
The retrospective cohort from 2007 yielded mature OS and PFS outcomes. However the historical time frame led to heterogenous data due to the evolution of practice over time. So, firstly, we excluded those who received reduced fractionation regimes. Our cohort began before the publications by NCBTSG 2012 and Perry et al trials [14–15] which have since led to more standardised hypofractionation radiotherapy practices for patients over 60 years and with performance status 2.
Secondly, MGMT, IDH, 19q16p codeletion and ATRIX biomarkers were not consistently carried out across this time period within the UK so data was not recorded. It is known those who are MGMT methylated have an improved prognosis and response to treatment [11].
Overall we demonstrate median PFS of 11.9 months and median OS of 16 months. PFS data has significant positive skewness 3.21, due to the statistical limitations of long-term retrospective cohorts.
The median OS was 14.6 months within the Stupp trial with analysis taken from the intention to treat population [4]. A more recent retrospective cohort of GBM treated between 2007–2011 found a comparable median OS of 14.9 months in those receiving maximal treatment [1].
The age range within our cohort is 17– 79 years, with a median age of 60 years. 4% of our cohort had a performance status of 2. This gives a real life data of GBM patients and makes direct comparisons to trial populations limited. However within the Stupp trial population (those aged under 18–70 years and PS 2 or less) 85% completed planned chemoradiotherapy and adjuvant TMZ [4]. Our real world non-trial data demonstrated 57% of patients completed 6 cycles of TMZ following chemoradiotherapy. Patients did not complete 6 adjuvant cycles of TMZ for a variety of reasons, including infection, deterioration in performance status, toxicities and progression.
We demonstrate for the first time those patients receiving 4 or more cycles of adjuvant TMZ had an extra 10 months median OS compared to those who received less than 4 cycles of adjuvant TMZ. This is a significant clinical and statistical difference. As clinically expected if patients discontinued adjuvant TMZ it was most likely after the first cycle.
A second key finding was, in our centre Gliadel wafer insertion improved OS by a median of 3.5 months, demonstrating those who do not have complications from Gliadel wafer insertion such as post-operative infection, incur a benefit.
Within our cohort we saw those who have Gliadel wafer insertion and receive 4 or more adjuvant cycles of TMZ both have improved OS. Therefore we went on to hypothesize the combination of both factors would give maximal OS. Indeed, OS was improved by 2 months in this group. Statistical significance was not reached on log-rank test (P-value 0.4) likely due to the small sample size of 12 patients.
95% of our cohort progressed despite primary standard treatment. At progression 20% underwent re-do surgery and 25% had second line chemotherapy with the majority opting for best supportive care (55%). This is comparable to the Stupp trial where 23% underwent surgery and 24% chemotherapy at relapse [4].
Using Cox analysis we demonstrate second-line chemotherapy had a significant impact on OS whereas re-do surgery did not significantly effect OS however there are complex factors to consider. This measure does not account for patients’ quality of life which is an important factor in decision making. However this gives an overview of practices within our centre which are in-line with large European randomized controlled trials.
This review of data gives us meaningful data to explain outcomes in our centre and beyond.