Refined high-content imaging-based phenotypic drug screening in zebrafish xenografts

Zebrafish xenotransplantation models are increasingly applied for phenotypic drug screening to identify small compounds for precision oncology. Larval zebrafish xenografts offer the opportunity to perform drug screens at high-throughput in a complex in vivo environment. However, the full potential of the larval zebrafish xenograft model has not yet been realized and several steps of the drug screening workflow still await automation to increase throughput. Here, we present a robust workflow for drug screening in zebrafish xenografts using high-content imaging. We established embedding methods for high-content imaging of xenografts in 96-well format over consecutive days. In addition, we provide strategies for automated imaging and analysis of zebrafish xenografts including automated tumor cell detection and tumor size analysis over time. We also compared commonly used injection sites and cell labeling dyes and show specific site requirements for tumor cells from different entities. We demonstrate that our setup allows us to investigate proliferation and response to small compounds in several zebrafish xenografts ranging from pediatric sarcomas and neuroblastoma to glioblastoma and leukemia. This fast and cost-efficient assay enables the quantification of anti-tumor efficacy of small compounds in large cohorts of a vertebrate model system in vivo. Our assay may aid in prioritizing compounds or compound combinations for further preclinical and clinical investigations.

P11.18.A Localizing value of EEG recordings in patients with glioblastoma

Background

Glioblastoma is associated with a high risk of epileptic seizures ranging from 40% to 60%. Before the advent of modern imaging techniques, electroencephalography (EEG) was a critical component in evaluating patients with space-occupying lesions. In this retrospective single-center study, we aimed (1) to characterize a cohort of patients with glioblastoma with regards to EEG monitoring, seizure frequency and the frequency of prescribed anti-seizure medications (ASM) and (2) to assess the value of EEG as a localizing technique in patients with glioblastoma.

Material and Methods

We reviewed the charts of 179 patients with glioblastomas between January 1st, 2020 and January 1st, 2022, treated at the Medical University of Vienna. The diagnosis was based on MRI and/or confirmed by biopsy according to the 2016 World Health Organization Classification of Tumors of the Central Nervous System. Patients who received an in-house EEG as part of their diagnostic work-up were included if an MRI/CT scan was available (within an average time of +/-60 days). For localization, focal slowing (theta/delta activity) and/or epileptiform discharges were considered. EEG rating was performed by a board-certified electrophysiologist blinded for the diagnosis and MRI/biopsy findings.

Results

We included 52 patients (29.05% of screened cohort) with at least one EEG and MRI or CT scan performed before or after EEG, following inclusion criteria (median: 2 days; mean: 6 days; range: -29 to 52), in the final analysis. Clinical seizure activity and/or epileptiform discharges on EEG were detected in 46 patients (88.46%), and 48 patients (92.31%) were on ASM.

An IDH-wildtype glioblastoma was diagnosed in 45 patients (86.54%), 4 had an IDH-mutant glioblastoma (7.69%), and in 3 patients, IDH-status was unknown (5.77%). In 22 patients (42.31%), biopsy revealed a positive MGMT promoter methylation status, while 28 were unmethylated (53.84%), and two patients had an unknown MGMT promoter methylation status (3.85%). Intermittent and/or continuous focal slow-wave activity was registered in 45 patients (86.54%). In comparison, epileptiform discharges could only be found in 13 patients (25%). When compared to MRI/CT scans, the hemispheric tumor localization could be determined in 42 cases (80.77%). Moreover, the affected brain lobe was accurately predicted in 35 patients (67.31%). Three patients had diffuse EEG changes (5.77%), and EEG was unremarkable in 7 patients (13.46%).

Conclusion

Overall, our presented data indicate that the hemispheric localization of glioblastoma can be reliably predicted by EEG recordings, while a precise (brain lobe-specific) localization was only possible in around two-thirds of cases.

P07.04.B Heme biosynthesis factors and 5-ALA induced fluorescence: analysis of mRNA and protein expression in fluorescing and non-fluorescing gliomas

Background

The intraoperative visualization of adult-type diffuse gliomas with 5-aminolevulinic acid (5-ALA) induced fluorescence is widely used in the neurosurgical field. While visible 5-ALA induced fluorescence is found in the majority of high-grade gliomas, most low-grade gliomas lack visible fluorescence during surgery. Recently, the heme biosynthesis pathway was identified as crucial influencing factor for presence of visible fluorescence since it metabolizes 5-ALA to fluorescing Protoporphyrin IX (PpIX). However, the exact alterations within the heme biosynthesis pathway resulting in visible 5-ALA induced fluorescence in gliomas are still unclear. The aim of the present study was thus to compare the mRNA and protein expression of promising intramitochondrial heme biosynthesis enzymes/transporters in glioma tissue samples of different fluorescence behavior.

Material and Methods

A total of 19 strongly fluorescing and 21 non-fluorescing tissue samples from neurosurgical adult-type diffuse gliomas (WHO grades II-IV) were included in the current analysis. In these samples, we investigated the mRNA expression by quantitative real time PCR and protein expression using immunohistochemistry of the intramitochondrial heme biosynthesis enzymes Coproporphyrinogen Oxidase (CPOX), Protoporphyrinogen Oxidase (PPOX), Ferrochelatase (FECH) and the transporter ATP-binding Cassette Subfamily B Member 2 (ABCG2).

Results

Regarding mRNA expression analysis, we found a significantly decreased ABCG2 expression in fluorescing specimens compared to non-fluorescing samples (p=0.001), whereas no difference in CPOX, PPOX and FECH was present. With respect to protein expression, significantly higher levels of CPOX (p=0.005), PPOX (p<0.01) and FECH (p=0.003) were detected in fluorescing samples. Similar to mRNA expression analysis, the protein expression of ABCG2 (p=0.001) was significantly lower in fluorescing samples.

Conclusion

Distinct alterations of the analyzed heme biosynthesis factors were found primarily on protein level. Our data indicate that heme biosynthesis pathway activity in general is enhanced in fluorescing gliomas with upregulation of PpIX generating enzymes and decreased ABCG2 mediated PpIX efflux outweighing the also increased further metabolization of PpIX to heme. Intramitochondrial heme biosynthesis factors thus constitute promising pharmacological targets to optimize intraoperative 5-ALA fluorescence visualization of usually non-fluorescing tumors such as low-grade gliomas.

P10.15.B Next generation drug screening platform: generation of patient-derived cells forex vivo drug response studies of gliomas

Background

Precision oncology aims to provide individual treatment options for each patient. In this regard, ex vivo drug screening systems have the potential to improve clinical outcomes. Traditionally, cancer drugs are tested on long-term cultured cancer cell line models, but cell lines cannot represent an individual patient from a clinic and are biologically too distinct to be informative for drug screening purposes. Drug screening systems of tumor cells usually rely on viability assays and correlations to genomic alterations. Beside genomic alterations, the cellular metabolism is significantly altered during tumor growth, tumor cell proliferation and tumor cell resistance development. Here we aim to establish a drug screening platform using tumor cells derived directly from the individual patient glial tumor, create patient derived tumor cells (PDCs) and combine the outcomes from standardized viability- and genetic-assays with a new developed metabolomics platform.

Material and Methods

Fresh native tissue from both low- and high-grade glioma are collected. Tumor tissue is used for NMR-based metabolomic analyses and targeted sequencing based genomic analyses as well as PDC isolation using mechanical and enzymatic tissue dissociation. To preserve the original tumor similarity, tissue is short term cultured for two weeks, and PDCs are seeded and treated with a panel of clinical- and preclinical drugs followed by viability assessment, sequencing and metabolomic profiling.

Results

Culturing of PDCs is successful in ≥85% of patient cases, provided that at least 2 g of tumor are available. The automatized high throughput ex vivo drug response helps to identify potential drug candidates which might become relevant for therapeutic approaches in future. Further, it is possible to distinguish between IDH1-wild type and IDH1-mutant glial tumors based on the metabolomic profile, which is confirmed by immunohistochemical staining and molecular analysis of IDH1 R132H-mutation. Strong metabolomic variations have been identified, including GABA, lactate, and myo-inositol levels between tumor and healthy tissue. Data retrieved by the systematic evaluation is retrospectively associated with the clinical course of the patients.

Conclusion

Entangling drug screening and genetic assays with metabolomic profiling of glial tumors enriches the information about cellular drug response and paves the way for future clinical studies and better understanding of underlying drug resistance mechanisms in gliomas.

Disclosure

Funding: K1 COMET Competence Centre CBmed, funded by the Federal Ministry of Transport, Innovation and Technology; the Federal Ministry of Science, Research and Economy, Land Steiermark (Dep. 12, Business and Innovation), the Styrian Business Promotion Agency (SFG), and the Vienna Business Agency. COMET is executed by the Austrian Research Promotion Agency (FFG).

P11.23.B Neurological symptoms independently associate with overall survival in patients with melanoma brain metastases

Background

Based on the the rising detection of asymptomatic brain metastases (BM) and the remarkable intracranial responses of targeted- and immunotherapy in specific subgroups of asymptomatic BM patients, we aimed to evaluate the clinical characteristics and prognostic value of neurological symptoms in BM from melanoma.

Material and Methods

670 patients with newly diagnosed melanoma BM were identified from the Vienna Brain Metastasis Registry. We performed retrospective chart review and statistical outcome analyses.

Results

370/670 (55.2%) patients presented with neurological symptoms including neurological deficits (316/370; 85.4%), signs of increased intracranial pressure (126/370; 34.1%) and epileptic seizures (53/370; 14.3%). 57.6% (80/139) of patients with BRAF mutated melanoma presented with asymptomatic BM at diagnosis. Absence of neurological symptoms at BM diagnosis was associated with a longer median overall survival after time of BM diagnosis (7 vs. 5 months; p< 0.001; log-rank test). In multivariate analysis with the diagnosis-specific graded prognostic assessment (DS-GPA: HR:1.4; 95% CI, 1.25-1.48; p<0.001) and the graded prognostic assessment for melanoma using molecular markers (Melanoma-molGPA: HR:0.8; 95% CI, 0.66 -0.92; p<0.001), the presence of neurological symptoms (HR 0.8; 95% CI, 0.69 -0.96; p=0.003/ HR:0.6; 95% CI, 0.59-0.72; p<0.001) was found to be independently associated with survival prognosis from time of BM diagnosis respectively.

Conclusion:

Neurological symptoms at the time of BM diagnosis are an independent and strong prognostic factor in melanoma BM patients. Our results argue for the integration of neurological symptoms into the prognostic assessment of patients with BM from melanoma.

OS03.7.A Clinical characteristics, treatment and long-term outcome of patients with brain metastases from thyroid cancer - an analysis of the Vienna Brain Metastasis Registry

Background

The incidence of brain metastases (BM) in patients with thyroid cancer (TC) depends on the histological subtype. About 1% of patients with differentiated (DTC), 3% with medullary (MTC) and up to 10% with anaplastic thyroid cancer (ATC) develop BM. The diagnosis of BM drastically worsens the prognosis of TC. Given the rare incidence, little is known about the presentation and outcome of this cohort.

Material and Methods

Patients with a histologically verified diagnosis of TC and BM were identified from the Vienna Brain Metastasis Registry, a comprehensive database managed by the Division of Oncology, Medical University of Vienna, including patients with cerebral metastasis since 1990. Data were obtained from medical records comprising clinicopathological features, treatment, BM-specific characteristics and outcome.

Results

20/6074 patients included in the registry had a diagnosis of TC and radiologically verified BM. 13/20 (65%) were female and the median age at diagnosis of TC and manifestation of BM was 56 years (range 21-75) and 68 years (range 45-75), respectively. In terms of histology, 18/20 (90%) had DTC, one MTC and one ATC. Interestingly, 10/18 DTC presented with follicular histology which underlines the more aggressive course of this rare subtype. 6/20 (30%) had BM at primary diagnosis (DTC n=5, ATC n=1), while the remaining developed BM during follow-up. The median time to diagnosis of BM was 2.6 years for DTC (range 0-42), 22 years in the MTC patient and 2 months for ATC. Regarding BM-specific characteristics, all but one patient had symptoms due to BM (neurological deficits n=11, increased intracranial pressure n=5, seizures n=3). Most patients (13/20) had a singular BM, commonly located in the left hemisphere (8/13), and only one had more than three BMs. Upfront treatment for BM was local therapy (resection n=9, stereotactic radiosurgery n=7, whole brain radiotherapy n=3); one patient received supportive care only. The median overall survival (mOS) from diagnosis of TC was 6 years for DTC (range 2.5 months-42 years), 33 years for the MTC and 9 months for the ATC. The mOS from diagnosis of BM was 14 months for DTC (1.8 months-16 years), 22 years for the MTC and 3 months for the ATC.

Conclusion

Few patients with TC develop BM, which present commonly as single lesion. While in general BM constitute a poor prognostic factor, individual patients experience long-term survival following local therapy. More information about the optimal management of BM in TC is needed to enable guideline recommendations.

P11.44.A The impact of heme biosynthesis regulation on glioma aggressiveness: correlations with most recent diagnostic molecular markers

Background

The prognosis of patients with diffusely infiltrating gliomas is dismal but varies greatly between individuals. While characterization of gliomas was primarily relied on typical histopathological features, specific molecular markers increasingly gained importance and play a key role in the recently published 5th edition of the World Health Organization (WHO) classification. Heme biosynthesis represents a crucial pathway due to its key role in oxygen transport, energy production or drug metabolism. Recently, we described a “heme biosynthesis mRNA expression signature” that correlates with histopathological glioma grades and patient survival. The aim of the current study was to correlate the heme biosynthesis mRNA expression signature with the most recent diagnostic molecular markers for glioma stratification.

Material and Methods

In this study, patient data were derived from the “The Cancer Genome Atlas” (TCGA) lower-grade glioma and glioblastoma cohorts. We identified diffusely infiltrating gliomas correlating molecular tumor diagnosis according to the most recent WHO classification with heme biosynthesis mRNA expression. The following molecular markers were analyzed: EGFR amplification, TERT promoter mutation, CDKN2A/B homozygous loss, concurrent chromosome 7 gain/10 loss, MGMT methylation, IDH mutation, ATRX loss, p53 mutation and 1p19q co-deletion. Subsequently, we calculated the heme biosynthesis mRNA expression signature and correlated this signature with distinct molecular glioma markers as well as the resulting molecular subgroups.

Results

A total of 649 patients with available data on up-to-date molecular markers and heme biosynthesis mRNA expression were included. According to analysis of individual molecular markers, we found a significantly higher heme biosynthesis mRNA expression signature in gliomas with IDH wildtype (p<0.0005), without 1p19q co-deletion (p<0.0005), with homozygous CDKN2A/B loss (p<0.0005) and with EGFR amplification (p=0.001). Furthermore, we observed that the heme biosynthesis mRNA expression signature increased with the aggressiveness of the molecular subgroups (p<0.0005), being lowest in WHO grade 2 oligodendrogliomas and highest in WHO grade 4 glioblastomas.

Conclusion

Our data demonstrate a significant correlation between diagnostic molecular markers and heme biosynthesis regulation in diffusely infiltrating gliomas. Consequently, heme biosynthesis expression is a promising biomarker for glioma aggressiveness and might constitute a potential target for novel therapeutic approaches.

P11.55.B Postoperative MRI is able to detect an unexpected residual tumor after surgery of brain metastases: experience from 5 specialized centers

Background

Brain metastases (BM) constitute the most common central nervous system tumors. The treatment options of BM consist of surgery, radiotherapy, radiosurgery, chemotherapy, and immunotherapy. Regarding surgery in BM, the extent of resection (EOR) represents a crucial factor for patient prognosis. However, first studies using postoperative MRI demonstrated that an unexpected residual tumor after surgery of BM is not uncommon despite these tumors were considered to be well-demarcated. The aim of this study was thus to investigate in a large cohort including multiple neurosurgical centers the EOR following BM resection, potential risk factors for incomplete resection and postsurgical follow-up data.

Material and Methods

In the current retrospective study conducted at 5 specialized neurosurgical centers in Austria, we included patients with BM resection and available postoperative MRI. The EOR following BM resection was determined by postoperative MRI (complete vs incomplete resection). Additionally, the data on the intraoperative judgement of the EOR of the performing neurosurgeon were collected. Moreover, potential factors for incomplete resection including tumor localization, tumor volume, primary tumor, pattern of contrast media enhancement on imaging and tumor eloquence were investigated. Finally, the rate of local progression of BM after initial surgery was analyzed in the follow-up period and overall survival data were collected.

Results

Altogether, 548 patients with 649 surgically treated BM were included. According to postoperative MRI, complete resection was achieved in 407 (66%) of 649 BM and incomplete resection in 176 (29%) of 649 BM. Misjudgment of the EOR by the neurosurgeon was found in 25% of cases and resulted in an unexpected residual tumor which was evident on postoperative MRI in 122 (22%) BM. Preoperative tumor volume was significantly larger in incompletely resected BM. Moreover, local progression was significantly more common in cases with incompletely resected BM and was also associated with shorter overall survival.

Conclusion

Our data of this study including multiple centers indicate that postoperative MRI is capable to detect a relatively high rate of unexpected residual tumors following resection of BM. Since local progression was more common in BM with residual tumors and this was associated with shorter survival, special attention should be paid to achieve a complete tumor resection.

JS07.4.A Correspondence of glutamine and glycine imaging based on 7T MRSI to amino acid PET

Background

New approaches for 7 Tesla magnetic resonance spectroscopic imaging (MRSI) allow the simultaneous imaging of multiple neuro-oncological biomarkers with 3-4 mm resolution in clinically feasible measurement times. Specifically, the amino acids glutamine (Gln) and Glycine (Gly), were previously limited to single voxel detection at lower fields. Both could add to our capabilities to resolve heterogeneous tumour metabolism.

Purpose

To progress the validation of Gln and Gly as neuro-oncological markers by conducting the first comparison to amino acid PET in a cohort of glioma patients.

Material and Methods

In 24 glioma patients (WHO 2021 classification), we quantitatively compared 7T MRSI (3D, 3.4 mm isotropic resolution, 15 min scan time) and routine PET (FET or MET). Within manual tumour segmentations, we defined hotspot volumes of interest (VOI) for the ratios of total choline (tCho, clinical standard reference), Gln, Gly to total N-acetylaspartate (tNAA) and PET tumour-to-brain ratios (TBR), all with a cut-off threshold of 1.6. For these VOIs, we calculated VOI volumes and median ratios as well as Dice similarity coefficients (DSC) and centre of intensity distance (CoI), between MRSI and PET ratios.

Results

We found that Gln and Gly ratios to tNAA had a higher correspondence to PET-based amino acid metabolism than tCho. Our resulting median VOI volumes were 19.08±23.10 cm³ for tCho/tNAA, 33.68±24.60 cm³ for Gln/tNAA, and 22.38±18.49 cm³ for Gly/tNAA compared to 24.33±30.46 cm³ for PET, with correlation coefficients >0.5 for all MRSI hotspot values in relation to PET volumes. Median ratios were 0.52±0.13 for tCho/tNAA, 0.61±0.25 for Gln/tNAA, 0.33±0.15 for Gly/tNAA and 2.11±0.42 for PET. The median DSCs to PET amounted to 0.53±0.36 for tCho/tNAA, 0.66±0.40 for Gln/tNAA, and 0.57±0.36 for Gly/tNAA, while the median CoI distances were 0.56±0.43 cm for tCho/tNAA, 0.39±0.22 cm for Gln/tNAA, and 0.45±0.48 cm for Gly/tNAA.

Conclusion

With this first study that compared high-resolution 3D-MRSI at 7 Tesla to amino acid PET and a quantitative evaluation, we demonstrated that Gln and Gly corresponded better to PET than tCho, which is the main marker used in clinical MR, both within the study and compared to previous literature. Future research is needed to clearly define the benefits of 7T MRSI for neuro-oncology such as the identification of tumour microenvironments or non-invasive determination of molecular-pathologic markers. Gln could be further explored by the application of Gln-based PET tracers to MR-PET. We still see further developments of MRSI methods, such as motion correction or absolute quantification of concentrations instead of ratios, as necessary to obtain such goals.

P11.37.B When to resect or biopsy for patients with supratentorial glioblastoma: a multivariable prediction model

Background

The prospects of a patient with suspected glioblastoma may rely heavily on the indication for surgical resection versus biopsy only. Biopsy percentages vary considerably across hospitals and guidelines for treatment of glioblastoma lack criteria for surgical decision-making. To identify patient and tumor characteristics associated with the decision to resect or biopsy a glioblastoma and to develop and validate a prediction model for decision support.

Material and Methods

Clinical data and pre-operative MRI scans were collected for adults who underwent first-time surgery for supratentorial glioblastoma from a registry-based cohort study of 12 hospitals from the Netherlands, Germany, France, Italy, and the United States between 1st of January 2007 and 31st of December 2011. The main outcome was the type of surgical procedure: surgical resection or biopsy only. Predictors were patient- and tumor-related characteristics. Radiological factors were extracted from MRI using an automated tumor segmentation method. A prediction model was constructed using multivariable logistic regression analysis. The model was cross-validated and externally validated with a leave-one-hospital-out approach.

Results

Out of 1053 patients treated for glioblastoma, 28% underwent biopsy only. Biopsy rates varied from 15-40% across hospitals. The prediction model showed excellent discrimination with an average area under the curve of 0.86. Of the patient-related characteristics, younger age was associated more with resection and Karnofsky Performance Score of 60 or less with biopsy. Of the tumor-related characteristics, a location in the right hemisphere, unifocality, no tumor midline crossing, and no involvement of the cortical spinal tract, were associated with resection, as well as a high expected resectability index, a location in the right occipital lobe, and a higher percentage of tumor in Schaefer’s dorsal or ventral attention, limbic, and default networks. External validation proved acceptable to outstanding discrimination with areas under the curve ranging between 0.79 and 0.92 for hospitals.

Conclusion

A prediction model is presented and validated to support the decision to resect or to biopsy a patient with a suspected supratentorial glioblastoma. In this prediction model, tumor-related characteristics were more informative than patient-related factors. This may support surgical decision-making for individual patients, or facilitate comparisons of patient cohorts between surgeons or institutions.

PL01.5.A Towards modernizing intraoperative histopathological assessment in brain and spinal tumors - Comparison of the novel Stimulated Raman Histology with conventional H&E staining

Background

By intraoperative analysis of fresh frozen sections, neuropathologists provide important information of different brain and spinal tumors to the neurosurgeon during surgery. This facilitates characterization of these tumors intraoperatively to optimize the surgical strategy and patient management. However, preparation and staining are time consuming using conventional techniques of intraoperative fresh frozen section. Stimulated Raman Histology (SRH) was introduced as novel technique providing high-resolution digital images of unprocessed tissue samples directly in the operating room comparable to conventional histopathological images. Additionally, SRH images are fast and easily accessible by neuropathologists. Recently, first data showed promising results on the accuracy and feasibility of SRH in comparison to conventional H&E staining.

Material and Methods

In a time period of 4 months, patients with different brain or spinal tumors who underwent neurosurgical resection or open/stereotactic biopsy at the Dept. of Neurosurgery, Medical University Vienna were included in this study. Tumor tissue samples were collected intraoperatively whenever safely possible for analysis with SRH. Subsequently, unprocessed tissue samples were scanned by SRH, and intraoperative histopathological images were created directly in the operating room within a few minutes. All collected tissue samples were then sent for routine neuropathological workup. In an overall analysis, SRH images and H&E staining of all patients were analyzed separately by two board certified neuropathologists. Information on age, localization and suspected diagnosis was provided in each case in order to simulate the situation of intraoperative fresh frozen section. In a next step the technical feasibility and diagnostic accuracy of SRH was calculated.

Results

In this study, tissue samples of 95 patients who underwent neurosurgical resection or open/stereotactic biopsy of different brain and spinal tumors were collected intraoperatively and analyzed by SRH. In total, 31 gliomas, 30 meningiomas, 19 metastases, 7 neurinomas and 8 rare tumors were analyzed. In the present study the use of SRH was technically feasible in all cases and could be easily integrated in the neurosurgical workflow to provide rapid digital histopathological images for the analyzing neuropathologists. According to our data, SRH provided high diagnostic accuracy (>95%) in the investigated different brain and spinal tumors.

Conclusion

Based on our preliminary data the technical use of SRH is feasible and showed a high rate of diagnostic accuracy in a large series of different brain and spinal tumors. By using this promising technique, we intend to modernize intraoperative histopathological assessment by providing rapid digital images of brain and spinal tumors to optimize the management of these patients.

OS01.5.A Neuron-specific enolase (NSE) and S100 serum levels in patients with active brain metastases from HER2-positive breast cancer treated with trastuzumab-deruxtecan (T-DXd): A biomarker analysis from the TUXEDO-1 trial

Background

T-DXd is a novel antibody-drug conjugate with high activity in HER2-positive metastatic breast cancer. The prospective, single-arm, single-centre phase II TUXEDO-1 trial showed clinically relevant activity of T-DXd in HER2-positive BC pts with active BM with an intracranial response rate of 73.3%. This biomarker sub-study of TUXEDO-1 aimed to investigate changes in the extent of metastases-induced brain damage in patients with and without response to therapy by measuring the serum levels of two proteins constitutively expressed in the human brain and measurable upon brain damage in the blood serum: NSE and S100.

Material and Methods

We assessed serum NSE (sNSE) and serum S100 (sS100) levels in a total of 37 blood samples drawn at cycles 1, 4 and end of treatment (EOT) in all patients of the intent-to-treat population of the TUXEDO-1 trial using commercially available ELISA assays. Intracranial radiological response was centrally assessed by a board-certified neuro-radiologist using RANO criteria. sNSE and sS100 levels were compared between responders and non-responders using the Mann Whitney U test and a significance level of 0.05.

Results

At baseline, we detected no significant difference of sNSE or sS100 levels between T-DXd responders and non-responders, respectively. Baseline median sNSE level was 10.6 ng/ml (interquartile range (IQR) 8.6-12.2) in T-DXd responders as compared with median 12.5 ng/ml (IQR 12.2-12.9) in the non-responder group (n.s.). At cycle 4, corresponding numbers were 8.1 ng/ml in responding patients (IQR 7-11.2) and 12.7 ng/ml (IQR 12.2-12.9) in non-responders, respectively (p=0.009). No differences in sS100 levels were observed between the groups at any time point.

Conclusion

In patients showing intracranial objective response to T-Dxd, sNSE levels were significantly lower at cycle 4 as compared with non-responders, suggesting a reduction in metastases-induced brain damage as a direct treatment effect. sNSE may be a clinically useful biomarker for longitudinal assessment of brain metastasis burden.

P13.04.B Dissecting high-grade glioma immune infiltration in samples from fluorescence-guided surgery: digital pathology with automated image analysis

Background

Fluorescence-guided surgery with 5-aminolevulinic acid (5-ALA) is a widely used technique to conduct maximum safe resection of high-grade gliomas (HGG). 5-ALA accumulates in malignant tumor tissue where it is metabolized to Protoporphyrin IX (PpIX), an agent with fluorescence properties. It helps neurosurgeons to distinguish between tumor-infiltrated tissue and healthy brain parenchyma. Even though fluorescence-guided surgery is clinically well established, the biological properties of different fluorescence levels are not comprehensively studied yet. A relevant current gap in that respect is the pattern of immune cell infiltration in fluorescent versus non-fluorescent tissue samples. In light of future research, reliable, standardized histopathology methods that allow high-throughput analysis are desirable and digital pathology with automated image analysis is an attractive option to explore.

Material and Methods

128 formalin-fixed paraffin-embedded (FFPE) tissue blocks of 39 patients who underwent fluorescence-guided surgery of a HGG were included. Samples were selected based on their documented 5-ALA fluorescence intensity status (strong, vague, negative). Microtome-cut sections of the tissue were stained with antibodies against CD8, CD68, CD163 and FOX P3, representing immune cell populations of specific interest (cytotoxic T cells, glioma-associated macrophages, regulatory T cells). A total of 512 stained slides were then available for assessment. In addition to a classical, semi-quantitative analysis by two independent human reviewers, the immune infiltration intensity was measured via automated image analysis with the digital pathology software QuPath Version 0.3.2.

Results

Across all stained FFPE samples CD68 showed the overall highest intensity, followed by CD163. CD8 and FoxP3 showed generally lower average intensities. In 5-ALA negative and positive high-grade glioma samples the immune cell infiltration pattern was the same. Quantitative automatic digital pathology correlated well with the classical human histopathological analysis for the majority of markers evaluated.

Conclusion

We successfully explored and established novel digital pathology technologies for the study of immune cell infiltration patterns in neurooncology, specifically in the context of fluorescence-guided resection. Leveraging this platform could allow parallelized and high-throughput analysis of immune cell infiltration also in other contexts.

OS03.5.A Characterization of the inflammatory tumor microenvironment composition in solid cancer patients with brain metastases after progression to immune checkpoint inhibitor therapy

Background

Immunotherapy (IO) has changed the treatment landscape of metastatic cancer patients, however, treatment resistance is frequent. We aimed to characterize the inflammatory tumor microenvironment in brain metastases (BM) after IO to gain a deeper understanding of immunologic escape mechanisms.

Material and Methods

Solid cancer patients who had BM resection after IO progression (IO cohort) were retrospectively identified. We analyzed tumor-infiltrating immune cell subsets (CD3, CD8, CD45RO, FOXP3) and expression of immune checkpoint molecules (PD-L1, PD-1, LAG-3) by immunohistochemistry. A control cohort of BM tissue samples without prior IO served for comparison (no immunotherapy cohort, NIO).

Results

Twenty-eight IO patients (12/28, 42.9% females; 16/28, 57.1% males; median 61 years; 14/28, 50% lung cancer; 5/28, 17.9% melanoma; 4/28, 14.3% renal cell carcinoma; 1/28, 3.6% breast cancer; 4/28, 14.3% other cancer entities) and 57 NIO patients (28/57, 49.1% females; 29/57, 50.9% males; median 58 years; 35/57, 61.4% lung cancer; 9/57, 15.8% breast cancer; 4/57, 7.0% melanoma; 3/57, 5.3% renal cell carcinoma; 6/57, 10.5% other cancer entities) were included. IO patients had a median of one (range 0-4) systemic therapy line prior to IO. Median time from last IO application until BM resection was 5.6 months (range 0.2-49.8 months). Patients received a median number of 7 (range 1-56) IO applications (14/28, 50% PD-1-targeting IO; 8/28, 28.6% PD-L1; 2/28, 7.1% CTLA4; 4/28, 14.3% CTLA4+PD-1; 3/28, 10.7% IO+chemotherapy). No statistically significant differences in the densities of investigated TILs or PD-L1 expression between the IO and the NIO cohort were observed. Patients of the IO cohort showed higher PD-L1 expression compared to the NIO cohort (57.1 vs. 42.1%, Chi-square, p>0.05). Overall survival (OS) was similar in both cohorts, with a median OS of 11.0 months (range 5.0-17.0) in the IO cohort and 11.0 months (range 5.5-16.5) in the NIO cohort.

Conclusion

Our findings show an upregulation of PD-L1 in BM occurring after prior IO therapy in the absence of other overt changes in the inflammatory microenvironment. Ongoing analyses in this cohort are investigating possible molecular driver of resistance by analyzing DNA methylation profiles of pre-and post-IO tissue samples of the IO cohort to potentially gain insights on inflammatory IO resistance mechanisms in BM patients.

P11.27.B Whole genome DNA methylation as predictive biomarker in CNS WHO grade 2 and 3 oligodendroglioma patients receiving early postoperative treatment

Background

Oligodendrogliomas are glial tumors with a relatively favorable survival prognosis of >10 years. While immediate postoperative treatment prolongs survival, long-term toxicities of adjuvant radio-chemotherapy remain a concern. Predictive biomarkers guiding postoperative treatment decisions are limited.

Material and Methods

In this retrospective study, we included patients treated for a newly diagnosed oligodendroglioma (isocitrate dehydrogenase (IDH)-mutated, 1p/19q-codeleted, CNS WHO grades 2 and 3) in 1992 - 2019 at the Medical University of Vienna or the Kepler University Hospital Linz (Austria). Early treatment was defined as radiotherapy, chemotherapy, or both within 6 months after resection, whereas benefit from early treatment was defined as progression-free survival (PFS) above the median in the overall cohort. DNA methylation analysis was performed using Illumina MethylationEPIC 850k microarrays.

Results

Of all 201 eligible patients, sufficient tumor tissue for DNA methylation analysis was available in 46 patients. Of these, 25/46 (54.3%) were diagnosed with CNS WHO grade 2 and 21/46 (45.6%) with grade 3 oligodendroglioma. Median age at diagnosis was 41 years (range: 23-70). In total, 21/46 (45.6%) patients received early treatment, of whom 13/21 (61.9%) received radio-chemotherapy, 6/21 (28.6%) radiotherapy only and 2/21 (9.5%) chemotherapy only. Median PFS was 134.0 months (95%CI: 78.3 - not reached) in patients receiving early treatment versus 87.2 months (95%CI: 66.8 - 150) in patients who did not. In patients receiving early treatment, differences in DNA methylation profiles could be detected between patients who drew benefit from postoperative treatment (group 1) versus those who did not (group 2). Based on the top 1000 differentially methylated CpG sites between both groups, two clusters were detected which comprised either patients of group 1 or 2. Clustering was independent from gender, WHO grade, extent of resection, type of postoperative treatment, treating center, and O6-methylguanine-methyltransferease (MGMT) promoter methylation status. Gene set enrichment analysis of the top 1000 differentially methylated gene sites mapped to 694 genes showed differential methylation in genes involved in fibroblast growth receptor 1 (FGFR1) signaling, Wnt signaling, integrin signaling, and actin cytoskeleton regulation. Conversely, in patients not receiving early treatment, PFS did neither correlate with DNA methylation clustering nor with MGMT promoter methylation.

Conclusion

In our cohort, whole genome DNA methylation was associated with PFS in patients who received early postoperative treatment, suggesting a predictive but not prognostic role. As the predictive value of MGMT promoter methylation is limited in oligodendroglioma, whole genome DNA methylation should be considered in future clinical trials.

P15.03.A The value of 7 Tesla MR spectroscopic imaging for improved preoperative determination of the tumor grade and IDH status in gliomas: preliminary data

Introduction

A new generation of MR spectroscopic imaging (MRSI) methods using 7T scanners have demonstrated the capability to resolve more neuro- and oncometabolites at higher resolutions than clinical routine MRSI. In a cohort of glioma patients, we explored the automated preoperative and noninvasive classification of IDH-mutation status and tumor grade based on 7T MRSI.

Methods

This retrospective study included 36 patients (15 female) with histologically confirmed diffusely infiltrating glioma WHO grade 2-4 (9 grade 2, 9 grade 3 and 18 grade 4) and known IDH status (21 IDH1-mut, 15 IDH-wt) with an available 7T MRSI scan of sufficient data quality. The 3D MRSI scan had a 3.4 mm isotropic resolution and 15 minutes acquisition time. 12 spectral components were classified voxel-wise, including choline, glutamine and glycine. Within a tumor segmentation based on routine 3T imaging, we used a random forest algorithm for the voxel-wise classification of IDH mutation and grade (into low or high grade). Training used the leave-one-out cross validation method (i.e., for every patient data set, the other 35 datasets were used as training set) and feature selection out of the available combinations for metabolite ratios (e.g., glutamine to choline). The resulting voxel classifications were aggregated into a mean probability per patient that was the base for receiver-operator characteristic (ROC) curves both for grade and IDH status.

Results

The classification algorithm obtained an area under the curve (AUC) for IDH determination of 0.85 (e.g., 75% sensitivity and 95% specificity). For grade determination, the AUC was 0.88 (e.g., 87% sensitivity and 89% specificity). In comparison, the AUC per voxel would have resulted in an AUC of 0.66 for both. Further, classification by individual metabolite ratios resulted in lower AUCs in all cases.

Conclusions

According to our preliminary data, preoperative 7T MRSI is capable to determine the correct glioma grade and IDH status with high sensitivity and specificity by leveraging the extended metabolic panel width and voxel amount. By increasing this cohort in future, we intend to confirm our initial results and we also plan to extend classification to more molecular-pathological features (e.g., TERT). Thus, even a voxel-wise classification of tumor microenvironments could be attempted. Further improvements in 7T MRSI methodology such as absolute instead of relative quantification would also aid these attempts.

In summary, 7T MRSI has shown its potential for improved preoperative characterization of gliomas.

P14.03 Shifting trends and entity-specific aspects in patients with brain metastasis: real-life analysis from 6031 individuals over an observation period of 30 years

BACKGROUND

We aimed to investigate the changing clinical characteristics of patients with brain metastases (BM) over the last three decades as the foundation for modern BM specific clinical trial planning.

MATERIAL AND METHODS

6031 patients with newly diagnosed BM from different solid tumors treated between 1986–2020 were identified from the Vienna Brain Metastasis Registry.

RESULTS

The fraction of BM originating from the most common BM causing primary tumors (lung cancer, breast cancer and melanoma) was stable over the observation period from 1986–2020. BM from renal cell carcinoma, colorectal cancer and cancer of unknown primary (CUP) decreased over time (p<0.001). Synchronous diagnosis of BM and primary tumor was more frequently observed in lung cancer and CUP patients compared to breast cancer patients (p<0.001). An increasing fraction of patients presented with asymptomatic BM (1986–1999: 20.2% vs. 2010–2020: 30.4%; p<0.001), specifically in lung cancer (p<0.001), melanoma (p<0.001) and renal cell cancer (p=0.004). A decrease of neurosurgical procedures (1986–1999: 39.3% vs. 2010–2020: 20.4%) and an increase of radiation treatments (1986–1999: 56.5% vs. 2010–2020: 73.0%) and systemic therapies (1986–1999: 0.6% vs. 2010–2020: 2.4%; p<0.001) was observed. Furthermore, median overall survival significantly increased across entities (1986–1999: 5 months vs. 2010–2020: 7 months; p=0.001). Intracranial progression as the cause of death increased across entities (p< 0.001). The prognostic DS-GPA (Hazard ratio [HR] 1.42; p< 0.001) and the Lung-molGPA (HR 1.67; p<0.001) could be validated.

CONCLUSION

We observed changes of BM presentation and clinical parameters during the observation period depending on primary tumor origins. Future BM studies should follow an entity-specific approach and address the characteristics of modern BM cohorts.

P14.14 Adjuvant treatment versus initial observation in newly diagnosed WHO grade II and grade III oligodendroglioma: real-life data from two academic, tertiary care centers in Austria

BACKGROUND

Oligodendrogliomas are rare, slow-growing brain tumors with a survival prognosis of >10 years. Although adjuvant radio-chemotherapy has been shown to prolong survival, aggressive treatment comes at the cost of increased toxicity. Systematic data on the optimal timing of adjuvant treatment in oligodendroglioma are lacking.

MATERIAL AND METHODS

Patients treated for a newly diagnosed IDH-mutated, 1p/19q-codeleted oligodendroglioma (WHO grades II/III) in 2000 - 2018 at the Medical University of Vienna or the Kepler University Hospital Linz (Austria) were included in this retrospective study. Adjuvant treatment was defined as radiotherapy (RT), chemotherapy (CHT) or radio-chemotherapy (R-CHT) within 6 months after resection in the absence of progression. “Wait and see” was defined as regular follow up with magnetic resonance imaging and treatment at progression.

RESULTS

185 patients were identified, comprising 123/185 (66.5%) WHO grade II and 62/185 (33.5%) WHO grade III oligodendrogliomas. Median age at diagnosis was 42 years (range: 20–82). Gross total resection (GTR) could be achieved in 77/178 (42.3%) evaluable patients. Adjuvant treatment was applied in 63/185 (38.2%) patients, of whom 43/63 (68.3%) underwent R-CHT, 9/63 (14.3%) CHT only and 11/63 (17.5%) RT only. 43/52 (82.7%) received temozolomide-based treatment, 1/52 (1.9%) procarbazine, lomustine and vincristine (PCV), 1/52 dacarbazine/fotemustine and in 7/52 (13.5%) patients, no data on used regimens was available. Adjuvant treatment was more frequently applied in WHO grade 3 tumors (p<0.001), while there was no association of adjuvant treatment with extent of resection (p=0.24). Patients after GTR who underwent adjuvant therapy presented with longer progression-free survival (PFS) compared to patients initially managed with observation (median: 150 months, 95%CI: 100 - not reached (n.r.) vs. median: 101 months, 95%CI: 73.2–115; p=0.053). In non-GTR tumors, patients with adjuvant therapy presented with a significantly longer median PFS of 107.5 months (95%CI: 62.8-n.r.) as compared to patients initially managed with observation (45.3 months, 95%CI: 41.2–78.8; p=0.025).

CONCLUSION

The application of adjuvant therapy was associated with favorable PFS in patients who underwent resection of newly diagnosed oligodendroglioma in this retrospective study. Prospective clinical trials should investigate the risks and benefits of adjuvant treatment versus initial observation in patients with oligodendroglioma.

Primary tumor side is associated with prognosis of colorectal cancer patients with brain metastases

Background

Brain metastases (BM) are a rare complication in colorectal cancer (CRC) patients and associated with an unfavorable survival prognosis. Primary tumor side (PTS) was shown to act as a prognostic and predictive biomarker in several trials including metastatic CRC (mCRC) patients. Here, we aim to investigate whether PTS is also associated with the outcome of CRC patients with BM.

Methods

Patients treated for CRC BM between 1988 and 2017 at an academic care center were included. Right-sided CRC was defined as located in the appendix, cecum and ascending colon and left-sided CRC was defined as located in the descending colon, sigma and rectum.

Results

Two hundred and eighty-one CRC BM patients were available for this analysis with 239/281 patients (85.1%) presenting with a left-sided and 42/281 patients (14.9%) with a right-sided primary CRC. BM-free survival (BMFS) was significantly longer in left-sided compared with right-sided CRC patients (33 versus 20 months, P = 0.009). Overall survival from CRC diagnosis as well as from diagnosis of BM was significantly longer in patients with a left-sided primary (42 versus 25 months, P = 0.002 and 5 versus 4 months, P = 0.005, respectively). In a multivariate analysis including graded prognostic assessment, PTS remained significantly associated with prognosis after BM (hazard ratio 0.65; 95% confidence interval: 0.46-0.92 months, P = 0.0016).

Conclusions

PTS was associated with survival times after the rare event of BM development in CRC patients. Therefore, its prognostic value remains significant even thereafter.

•

Primary tumor side is a relevant and independent prognostic factor in mCRC.

•

Left-sided CRC was associated with a significantly longer BMFS compared with right-sided CRC.

•

OS from initial diagnosis of CRC as well as from BM was significantly longer in patients with left-sided primaries.

Tumor mutational burden and immune infiltrates in renal cell carcinoma and matched brain metastases

BACKGROUND

Tumor mutational burden (TMB) and density of tumor-infiltrating lymphocytes (TIL) have been postulated as predictive biomarkers for immunotherapy. Therefore, we investigated the concordance of TMB and TIL of primary/extracranial renal cell carcinoma (RCC) specimens and matched brain metastases (BM).

PATIENTS AND METHODS

Twenty specimens from 10 patients were retrieved from the Vienna Brain Metastasis Registry (6/10 primary tumor, 4/10 lung metastasis, 10/10 matched BM). TMB was assessed using the TruSight Oncology 500 gene panel with libraries sequenced on a NextSeq instrument. TIL subsets (CD3+, CD8+, CD45RO+, FOXP3+, PD-L1+) were investigated using immunohistochemistry (Ventana Benchmark Ultra system) and automated tissue analysis (Definiens software).

RESULTS

No significant difference in TMB, CD3+, CD8+, CD45RO+, FOXP3+ or PD-L1+ expression was observed between extracranial and matched intracranial specimens (P > 0.05). Higher CD8+ TIL (P = 0.053) and CD45RO+ TIL (P = 0.030) densities in the primary tumor compared with the intracranial samples were observed in specimens collected after exposure to systemic treatment. Neither extracranial sample origin (lung metastasis versus primary RCC) nor extracranial disease status at BM diagnosis (progressive versus stable disease) were significantly associated with TMB or TIL densities in extracranial and intracranial samples (P > 0.05). No significant correlation was found between the median differences of TMB or TIL densities from extracranial to intracranial samples and BM-free survival.

CONCLUSION

The comparable immunological microenvironment of extra- and intracranial tumor samples in our study underscores the immunological activation also in BM from RCC, and therefore, supports the development of immune modulatory treatments also in patients with brain metastatic RCC.

P14.97 High risk of venous thromboembolism in patients with brain metastases from non-small cell lung cancer

BACKGROUND

Venous thromboembolism (VTE) is a common complication in patients with cancer. In general, patients with metastatic disease are at highest risk. Lung cancer belong to those tumor entities with a particularly high risk of VTE, ranging between 3–13.8%. However, little is known about the VTE rate in lung cancer patients with brain metastases.

MATERIAL AND METHODS

Our study was conducted in the framework of the Vienna Brain Metastasis Registry. Clinical data and VTE events during the course of the disease were recorded via retrospective chart review. In this analysis, non-small cell lung cancer (NSCLC) patients with a resection of brain metastases at the Medical University of Vienna between 2006 and 2010 were included.

RESULTS

In total, 69 NSCLC patients with brain metastases were analyzed. Overall, 69.6% (48/69) patients had an adenocarcinoma, 13% (9/69) a squamous cell carcinoma, 8.7% (6/69) a large cell carcinoma and 8.7% (6/69) other primary tumor histologies. After cancer diagnosis, 20.3% (14/69) patients developed VTE during the course of the disease. Of those, 85.7% (12/14) thromboembolic events occurred after the diagnosis of brain metastases.

CONCLUSION

Based on our data, patients with brain metastases from NSCLC have a very high VTE risk. Further investigations are needed in order to identify patients with distinct VTE risk profiles. Patients at high risk might potentially benefit from primary thromboprophylaxis over the high risk of intracerebral bleeding.

Brain metastases as first manifestation of advanced cancer: exploratory analysis of 459 patients at a tertiary care center

Symptomatic brain metastases (BM) are a frequent and late complication in cancer patients. However, a subgroup of cancer patients presents with BM as the first symptom of metastatic cancer. Here we aimed to analyze the clinical course and prognostic factors of this particular BM patient population. Patients presenting with newly diagnosed BM without a history of metastatic cancer were identified from the Vienna Brain Metastasis Registry. Clinical characteristics and overall survival were retrieved by chart review. 459/2419 (19.0%) BM patients presented with BM as first symptom of advanced cancer. In 374/459 (81.5%) patients, an extracranial primary tumor, most commonly lung cancer, could be identified within 3 months after BM diagnosis. In 85/459 (18.5%) patients no extracranial primary tumor could be identified despite comprehensive diagnostic workup within the first 3 months after diagnosis of BM. Survival of patients with identified extracranial tumor differed only numerically from patients with cancer of unknown primary (CUP), however patients receiving targeted therapy after molecular workup showed significantly enhanced survival (20 months vs. 7 months; p = 0.003; log rank test). The GPA score showed a statistically significant association with median overall survival times in the CUP BM patients (class I: 46 months; class II: 7 months; class III: 4 months; class IV: 2 months; p < 0.001; log rank test). The GPA score has a strong prognostic value in patients with CUP BM and may be useful for patient stratification in the clinical setting. Comprehensive diagnostic workup including advanced imaging techniques and molecular tissue analyses appears to benefit patients by directing specific molecular targeted therapies.

Decreased body mass index is associated with impaired survival in lung cancer patients with brain metastases: A retrospective analysis of 624 patients

Body mass index (BMI) is a prognostic factor in several cancer types. We investigated the prognostic role of BMI in a large patient cohort with newly diagnosed lung cancer brain metastases (BM) between 1990 and 2013. BMI at diagnosis of BM and graded prognostic assessment (GPA) were calculated. Definitions were underweight (BMI <18.50), weight within normal range (BMI 18.50-24.99) and overweight (BMI ≥ 25.00). A total of 624 patients (men 401/624 [64.3%]; women 223/624 [35.7%]; median age of 61 [range 33-88]) were analysed. Histology was non-small cell lung cancer in 417/622 (66.8%), small cell lung cancer (SCLC) in 205/624 (32.9%) and not otherwise specified in 2/624 (0.3%) patients. About 313/624 (50.2%) had normal BMI, 272/624 (43.5%) were overweight and 39/624 (6.3%) were underweight. Underweight patients had shorter median overall survival (3 months) compared to patients with normal BMI (7 months) and overweight (8 months; p < .001; log rank test). At multivariate analysis, higher GPA class (HR 1.430; 95% cumulative incidence, CI 1.279-1.598; p < .001; Cox regression model), SCLC histology (HR 1.310; 95% CI 1.101-1.558) and presence of underweight (HR 1.845; 95% CI 1.317-2.585; p = .014; Cox regression model) were independent prognostic factors. Underweight at diagnosis of BM in lung cancer is associated with an unfavourable prognosis.