OS07.2.A Evaluation of Intraoperative Surgical Adjuncts and Resection of Glioblastoma (ELISAR GB): A UK and Ireland multicentre, prospective observational cohort study

Background

Despite operative and adjuvant therapies, glioblastoma remains incurable, with the extent of resection being one of few treatments that can improve survival. To improve resection, operative adjuncts are used, with neuronavigation and 5-aminolevulinic acid (5-ALA) recommended as a standard of care in those aimed for maximal safe resection. Despite the standards, meta-analysis concluded that the impact of 5-ALA on the extent of surgical resection is of low quality due to bias in reporting tumour location and additional image guidance used, factors impacting on extent of resection as well as short-term neurological outcomes being uncertain. Therefore we aimed to evaluate the availability and use of 5-ALA and other adjuncts and compare surgical outcomes of 5-ALA-guided versus non-5-ALA-guided resections.

Material and Methods

A multicenter prospective observational cohort study was conducted across 27 out of 31 available centres in the UK and Ireland from 6 January until 19 March 2020. Inclusion criteria included adults with first diagnosis, supratentorial glioblastoma undergoing resection. Primary outcomes included: i) the availability and use of surgical adjuncts and ii) complete resection of enhancing tissue (CRET). Secondary outcomes included adverse events, new onset of postoperative neurological deficit and post-operative neurological function. Descriptive and inferential statistics were used for analysis with a p-value <0.05 deemed significant.

Results

232 consecutive cases were identified. 142/232 cases were aimed for maximal safe resection subsequently divided into 5-ALA-guided (n=92) versus non-5-ALA-guided (n=50) resections. 5-ALA and neuronavigation were available across all centres. Neuronavigation and 5-ALA were used in 91% (n=129/142) and 65% (n=92/142) of cases aimed for maximal safe resection whereas 83% (n=75/90) and 49% (n=44/90) for debulk surgery. 35 unique combinations of surgical adjuncts were used in 232 operations. 5-ALA-guided resection yielded a higher percentage of CRET than without (55% versus 28%, p < 0.01). The two groups showed no difference in adverse events (p=0.98), new onset of neurological deficit (p=0.88) nor neurological function (p=0.7). A logistic regression analysis showed that 5-ALA was an important predictor of CRET regardless of additional adjuncts used (OR 2.4, CI 0.96-5.97, P = 0.05), tumour location and molecular characterisation (OR 3.48, CI 1.61-7.51, P <0.01).

Conclusion

Firstly, we showed that 5-ALA is not always used for glioblastoma aimed for CRET. Secondly, we report a great heterogeneity of adjuncts used for resection, possibly explained by a lack of high-quality evidence and surgeon training. Thirdly we demonstrate that 5-ALA-guided resection leads to higher percentage of CRET regardless of other adjuncts used, tumour location and molecular characterisation.

P13.04 Survival analyses for determining clinical relevance of molecular markers in translational glioblastoma research - a systematic review

BACKGROUND

Pre-clinical glioblastoma studies can assess the relevance of their findings to patient survival using integrated clinical and genomic data. Validity of univariable analyses requires an assumption that molecular markers are randomly distributed across patient characteristics to mitigate the confounding effects of clinical variables. Multivariable survival analyses adjusting for clinical variables do not change the association if this assumption holds. We aimed to assess this by summarising the types of survival analyses and their results in translational glioblastoma research.

MATERIAL AND METHODS

We systematically searched Medline and Embase Jan 2008 to Feb 2021 for glioblastoma cell line or animal studies validating their molecular markers in The Cancer Genome Atlas (TCGA) or the Chinese Glioma Genome Atlas (CGGA) using survival analyses. Studies that exclusively used genomic data without laboratory findings were excluded. Two reviewers independently assessed study eligibility and extracted data. Data items included patient inclusion criteria, characteristics of survival analyses, and whether molecular markers had statistically significant association with overall survival.

RESULTS

Of 1,047 potentially eligible studies, we included 59 pre-clinical glioblastoma studies that tested the association between their molecular markers and survival using TCGA or CGGA data. All studies used TCGA data and 2 also used CGGA data. Sixteen (27%) studies specified their patient inclusion criteria from TCGA for survival analysis. Eight studies exclusively investigated sets of molecular markers, leaving 51 studies reporting 126 molecular markers. Eighteen (31%) studies used multivariable survival analysis in addition to univariable analyses. All molecular markers underwent univariable analyses, of which 12 (10%) molecular markers had additional multivariable survival analyses. In the 13 multivariable analyses on 12 molecular markers, four (31%) markers were associated with survival in the univariable analyses but not in the multivariable analyses.

CONCLUSION

Most pre-clinical studies used univariable survival analyses alone in public genomic repositories to assess the relevance of their results to patient survival. Our findings demonstrated that multivariable analyses are needed to account for confounding effects of clinical variables. Using relevant components from reporting guidelines for observational studies can improve the transparency and quality of translational studies.

P13.09 Inconsistent effect of temozolomide exposure on cell viability in glioblastoma cell line models - a systematic review

BACKGROUND

The alkylating agent temozolomide is part of standard care for patients with glioblastoma. Potential novel therapeutic agents are often first evaluated against temozolomide in glioblastoma cell line models. Despite the importance of this step in compound development, there is no standard concentration or exposure duration of temozolomide in laboratory research, and consistency in the effect of temozolomide on glioblastoma cell lines has not been assessed. This systematic review aimed to summarise the concentration and exposure duration of temozolomide and its effect on cell viability in studies using glioblastoma cell lines.

MATERIAL AND METHODS

We searched Medline and Embase Jan 1994 - Feb 2021 for studies that used at least one glioblastoma cell line and reported a measure of cell viability associated with temozolomide exposure. Studies were excluded if they used modified cell lines or did not report a cell viability measure associated with temozolomide as monotherapy. One reviewer screened all records and two reviewers assessed potentially eligible studies for inclusion. The main data items included the cell lines used, the concentration and exposure duration to temozolomide, and cell viability measures. We summarised findings using descriptive statistics.

RESULTS

Of 1,533 potentially eligible studies we included 213 studies reporting 209 different cell lines. The most common cell lines were U87, U251 and T98G, used in 61%, 41%, and 27% of studies, respectively. Twenty-five (12%) studies used patient-derived cell lines. The concentration of temozolomide used ranged from 0 to 8000μM. The temozolomide exposure duration ranged from <24 hours to >96 hours, with 29% studies using 72 hours. The most common cell viability measure was half maximal inhibitory concentration (IC50), which was reported in 183 (86%) studies. The median IC50 in 32 studies using the U87 cell line was 180μM (interquartile range [IQR]: 52–254μM) at 48-hour temozolomide exposure and 202μM (IQR 52–518μM) at 72-hour exposure. The median IC50 in 31 studies using U251 cell line was 84μM (IQR: 34–324μM) at 48-hour exposure and 102μM (IQR: 35–358μM) at 72-hour exposure.

CONCLUSION

Experimental setup of temozolomide and its effect on cell viability vary widely between studies using similar glioblastoma cell lines. This inconsistency of response to temozolomide questions reproducibility and the translational value of study findings.

Prospective, multicentre study of external ventricular drainage-related infections in the UK and Ireland

OBJECTIVES

External ventricular drain (EVD) insertion is a common neurosurgical procedure. EVD-related infection (ERI) is a major complication that can lead to morbidity and mortality. In this study, we aimed to establish a national ERI rate in the UK and Ireland and determine key factors influencing the infection risk.

METHODS

A prospective multicentre cohort study of EVD insertions in 21 neurosurgical units was performed over 6 months. The primary outcome measure was 30-day ERI. A Cox regression model was used for multivariate analysis to calculate HR.

RESULTS

A total of 495 EVD catheters were inserted into 452 patients with EVDs remaining in situ for 4700 days (median 8 days; IQR 4-13). Of the catheters inserted, 188 (38%) were antibiotic-impregnated, 161 (32.5%) were plain and 146 (29.5%) were silver-bearing. A total of 46 ERIs occurred giving an infection risk of 9.3%. Cox regression analysis demonstrated that factors independently associated with increased infection risk included duration of EVD placement for ≥8 days (HR=2.47 (1.12-5.45); p=0.03), regular sampling (daily sampling (HR=4.73 (1.28-17.42), p=0.02) and alternate day sampling (HR=5.28 (2.25-12.38); p<0.01). There was no association between catheter type or tunnelling distance and ERI.

CONCLUSIONS

In the UK and Ireland, the ERI rate was 9.3% during the study period. The study demonstrated that EVDs left in situ for ≥8 days and those sampled more frequently were associated with a higher risk of infection. Importantly, the study showed no significant difference in ERI risk between different catheter types.