Contrast-enhancing Tumor Research Articles

CTNI-48. NOVEL TREATMENT OF END STAGE RECURRENT GLIOBLASTOMA TREATED WITH A NONINVASIVE ONCOMAGNETIC DEVICE USING OSCILLATING MAGNETIC FIELDS – A NEW AND POWERFUL NONINVASIVE THERAPY

Abstract Noninvasive treatment with alternating electric fields at 200 kHz (Optune therapy) has been approved for recurrent and newly diagnosed glioblastoma (GBM). We have recently tested and confirmed in vitro, and in patient derived xenografted mouse models, the oncolytic potential of powerful newly developed device generating oscillating magnetic fields (OMF) using rotating magnets. This device is completely different and has different mechanisms of action than Optune therapy. We report the successful treatment of an endstage recurrent GBM in a 53-year-old man. The patient had undergone radical surgical excision and standard of care chemoradiotherapy, as well as experimental gene therapy for a left frontal tumor. 18 months after this treatment, there was recurrence with progressive enlargement and leptomeningeal spread, making the patient’s condition not suitable for further treatment. We conducted a 5-week treatment (followed by a pause of one week) consisting of daily exposure (on weekdays) with intermittent OMF. The patient tolerated the treatment well. At 4 weeks there was a progressive reduction of the contrast-enhanced tumor (CET) volume to less than 60% of the pre-treatment volume on magnetic resonance imaging scans. Caregiver report indicated mild subjective improvement in speech and cognitive function. The daily decrease in volume was correlated with the amount of daily treatment, with longer treatments showing an increase in shrinkage of the CET volume. After the treatment pause in the 6th week, there was an apparent increase in CET volume. However, a marked reduction in the choline peak and an increase in the lipid peak on a proton magnetic resonance spectrogram of the tumor at this timepoint documented that this represented necrotic tissue. We are excited about this new and powerful treatment for glioblastoma and will present the details of treatment and imaging studies during therapy.

STEM-05. SINGLE CELL SEQUENCING CHARACTERIZES SPATIAL AND TEMPORAL RELATIONSHIPS OF ENHANCING AND NON-ENHANCING REGIONS IN HIGH-GRADE GLIOMA

Abstract INTRODUCTION Treatments of high grade glioma focus on the contrast enhancing (CE) portion of the tumor. However, given the invasive nature of glioma, residual tumor cells responsible for recurrence likely exist in the non-enhancing (NE) region. METHODS In 3 patients undergoing surgery for malignant glioma, we used pre-operative magnetic resonance images to prospectively identify biopsy targets in the CE region and locations 0.5-2.0cm beyond the CE edge. We carried out single cell sequencing on image guided biopsy specimens to generate 12,528 RNA profiles. RESULTS In all samples, tumor cells clustered into three predominant groups: tumor cells with astrocyte markers (ASC-like), tumor cells with oligodendrocyte markers (ODC-like), and tumor cells with neither marker. This last group consisted of a small proportion of tumor cells and expressed putative stem cell markers (PROM1, NES, SLC1A3, A2B5, ID1). A trajectory analysis consistently positioned this group as branching off into either ASC-like or ODC-like cells. CE regions had different cellular compositions than NE regions, with higher proportions of tumor, endothelial cells, T-cells, and NK cells. Using location and cell density data, we modeled expected tumor burden, predicting tumor cells up to 1.5cm beyond the CE region of the tumor. There were significant differences in gene expression between CE and NE tumor cells, with increased inflammation and hypoxia in the CE region versus increased proliferative markers in the NE region. Tumor cells in the NE region were characterized by increased proliferation, markers of invasion, and markers of self-renewal. CONCLUSIONS Single cell sequencing illustrates multiple glioma cell types and suggests a hierarchical relationship between tumor cell types. CE and NE regions exhibit different tumor and non-tumor cell populations as well as different gene expression profiles within individual cell types. There remains a significant tumor cell burden in the NE portion of tumor, including actively proliferating cells.

SURG-22. PATTERN OF RECURRENCE ALONG EXTENT OF RESECTION IN PATIENTS WITH GLIOBLASTOMA, IDH-WILDTYPE: A RETROSPECTIVE STUDY

Abstract OBJECTIVE The concept for maximal resection of glioblastoma(GBM) is changing to resect not only contrast-enhanced (CE) tumor but non-contrast-enhanced (NCE) tumor. Although maximal resection is associated with longer survival, the pattern of recurrence in glioblastoma, isocitrate dehydrogenase (IDH)-wildtype patient is poorly understood. METHODS This retrospective, single-center study included 358 eligible patients through November 1, 2005 to December 31, 2018. Those patients were grouped along extent of resection (subtotal resection (STR, n=123), gross-total resection (GTR, n=194), and supratotal resection (SupTR, n=41)). The pattern of recurrence was classified based on the range of radiation treatment, into infield gross target volume (GTV), infield clinical target volume (CTV), marginal, outfield and CSF seeding. RESULTS Tumor recurrence was observed for each group as follows: STR group, 80.5%, GTR group, 75.3%, and SupTR group, 63.4%. Statistical differences in pattern of recurrence among each group were observed in infield GTV (STR group:75.8%, GTR group: 71.9%, and SupTR group: 42.3%, p=0.003) and Outfield (STR group:11.0%, GTR group: 21.2%, SupTR group: 34.6%, p=0.012) recurrence. The survival for SupTR group was significantly longer than that for the STR and GTR in terms of both overall survival (OS) and progression free survival (PFS) (OS, STR: 15.7m, GTR: 21.6m, SupTR: 44.7m, PFS, STR: 10.4m, GTR: 14.3m, SupTR: 35.5m) CONCLUSIONS As the extent of resection increased, local recurrence decreased and distant recurrence increased. During treatment of glioblastoma, expectation

NIMG-23. COMPARISON OF A RADIO-PATHOMIC MODEL VERSUS A RADIOLOGY-ONLY TUMOR SEGMENTATION MODEL FOR THE DETECTION OF INFILTRATIVE TUMOR IN GLIOMA PATIENTS

Abstract This study used large format autopsy tissue samples to compare radio-pathomic maps of brain cancer to a current tumor segmentation algorithm. We hypothesized that an MRI-based machine learning model trained with actual histology rather than radiologist annotations cellularity would 1) improve delineation between tumor and treatment effect, and 2) detect abnormal pathology beyond the contrast-enhancing tumor region. Seventeen patients with pathologically confirmed glioma were included in this study. At autopsy, 43 tissue samples were collected from 17 subjects from whole brain slices sectioned to align with the last axial MRI prior to death. Cellularity was calculated using a color deconvolution on 40X digitized H&E stained slides from the tissue samples. In-house custom software was used to align tissue samples and cellularity information to the FLAIR image using manually defined control points. The DeepMedic algorithm was trained to segment tumors using the BraTs 2017 dataset, and then applied to our patients in order to create automated tumor probability maps. An MRI-based ensemble algorithm using a 5x5 voxel searchlight (input: T1, T1C, FLAIR, ADC) was used to predict cellularity at each voxel, using tissue samples from 14 subjects as ground truth. Both models were applied to 3 withheld test subjects in order to compare tumor probability and cellularity predictions to the pathological ground truth. The mutual information between tumor probability and actual cellularity was 1X10-15, relatively low compared to the rad-path predicted cellularity (=0.16), despite the tumor prediction model accurately highlighting regions of contrast enhancement. Additionally, the radio-pathomic ensemble model correctly identify regions of hypercellularity beyond the tumor segmentation model as well as regions of within the segmented tumor area. This study demonstrates the utility of training machine learning models with pathological ground truth rather than radiologist annotations for predicting localized tumor information, particularly in the post-treatment stage.

NIMG-04. MRI EVALUATION OF THE EFFECT OF WHOLE BRAIN RADIOTHERAPY ON BRAIN METASTASIS

Abstract Poor prognosis of breast cancer brain metastasis partially results from the inability of therapeutics to penetrate the Blood-Brain-Barrier (BBB). Whole Brain Radiotherapy (WBRT) is standard-of-care for brain metastasis, however the early effects of fractionated WBRT on brain metastases have not been evaluated. Here, we applied longitudinal MRI to study WBRT’s effect on tumor volume, BBB permeability, and tumor-cell killing using anatomic, Dynamic Contrast Enhanced (DCE), and Diffusion-weighted (DW) MRI, evaluating tumor volume, permeability, and cytotoxic effect, respectively. The BCBM mouse model was established with left ventricle MDA-MB231-Br cell injection. MRI occurred before and 24h after 3 daily fractionated 4Gy WBRT (2.4Gy/min) or sham irradiation. In addition to anatomic MRI, DW MRI measured tissue ADC to study cytotoxic effect, and DCE MRI measured BBB permeability, quantified as the permeability variable Ktrans. Before treatment, T2-weighted images revealed hyperintense multifocal lesions in brains, and many had intact BBB indicated by no T1-w contrast-enhancement. WBRT tumors were smaller than sham irradiated tumors after 3 days (p< 0.05), while no difference in the number of new lesions was observed. DW MRI demonstrated increased ADC in WBRT tumors (p< 0.05), correlating well with elevated Caspase-3 staining of apoptotic cells. T1-w contrast enhanced images of tumors after treatment revealed no differences in the population of contrast-enhanced tumors between the WBRT and sham radiation groups. However, quantitative DCE MRI showed higher Ktrans in WBRT treated tumors relative to sham treated lesions, and an increase in Ktrans of WBRT tumors (p< 0.05), despite high permeability heterogeneity both inter- and intra-tumorally, indicating remaining impermeable BBB. In addition to monitoring tumor volume, longitudinal MRI provided non-invasive assessments of temporal and spatial changes in cellularity and vascular permeability of brain metastasis in response to WBRT, which may prove to be useful for defining an ideal treatment window for potential combination BCBM treatments.

SURG-21. A CROWDSOURCED CONSENSUS ON SUPRATOTAL RESECTION VERSUS GROSS TOTAL RESECTION FOR ANATOMICALLY DISTINCT PRIMARY GLIOBLASTOMA

Abstract Gross total resection (GTR) of contrast-enhancing tumor is associated with significantly increased overall survival in primary glioblastoma (GBM). Even when achieved, recurrence is likely, in part due to malignant cells infiltrating outside enhanced regions. Subsequently, there has been increasing interest in performing supratotal resections (SpTRs) for GBM. Published results have varied in part due to a lack of consensus on the definition of SpTR in GBM and its appropriate use. A crowdsourcing approach was used to survey 21 academic neurosurgical oncologists representing 13 health systems nationwide. Participants’ demographics including fellowship training status, years of experience, and operative volume with various techniques was collected. Participants were presented with 11 definitions of SpTR from published, peer-reviewed studies and asked to rate the appropriateness of each definition. Subsequently, participants reviewed T1-weighed post-contrast and FLAIR MR imaging videos in the axial, coronal, and sagittal planes for 22 GBMs. Participants were asked to assess eloquence of the tumor’s location, perceived equipoise of enrolling patients in a randomized clinical trial comparing GTR to SpTR, and their own personal surgical treatment plans. Most neurosurgeons surveyed (n=18, 85.7%) agree or strongly agree that GTR plus resection of some non-contrast enhancement is an appropriate definition for SpTR. Overall, there was only moderate inter-rater agreement, measured using sample variance and the index of qualitative variation, regarding eloquence, equipoise, and personal treatment plans. Neurosurgeons who performed more than 10 SpTRs for GBMs in the past year were more likely than counterparts to recommend it as their personal treatment plan (p< 0.005). Anterior temporal and right frontal GBMs were considered the best randomization candidates. We established a consensus definition for SpTR of GBM and identified anatomically distinct locations deemed most amenable to SpTR. These results will be used to plan prospective trials further investigating the potential clinical utility of SpTR for GBMs.

Prediction of pseudoprogression and long-term outcome of vestibular schwannoma after Gamma Knife radiosurgery based on preradiosurgical MR radiomics

Background and purposeGamma Knife radiosurgery (GKRS) is a safe and effective treatment modality with a long-term tumor control rate over 90% for vestibular schwannoma (VS). However, numerous tumors may undergo a transient pseudoprogression during 6–18 months after GKRS followed by a long-term volume reduction. The aim of this study is to determine whether the radiomics analysis based on preradiosurgical MRI data could predict the pseudoprogression and long-term outcome of VS after GKRS. Materials and methodsA longitudinal dataset of patients with VS treated by single GKRS were retrospectively collected. Overall 336 patients with no previous craniotomy for tumor removal and a median of 65-month follow-up period after radiosurgery were finally included in this study. In total 1763 radiomic features were extracted from the multiparameteric MRI data before GKRS followed by the machine-learning classification. ResultsWe constructed a two-level machine-learning model to predict the long-term outcome and the occurrence of transient pseudoprogression after GKRS separately. The prediction of long-term outcome achieved an accuracy of 88.4% based on five radiomic features describing the variation of T2-weighted intensity and inhomogeneity of contrast enhancement in tumor. The prediction of transient pseudoprogression achieved an accuracy of 85.0% based on another five radiomic features associated with the inhomogeneous hypointensity pattern of contrast enhancement and the variation of T2-weighted intensity. ConclusionThe proposed machine-learning model based on the preradiosurgical MR radiomics provides a potential to predict the pseudoprogression and long-term outcome of VS after GKRS, which can benefit the treatment strategy in clinical practice.

Differentiating Nonenhancing Grade II Gliomas from Grade III Gliomas Using Diffusion Tensor Imaging and Dynamic Susceptibility Contrast MRI

Contrast enhancement in a brain tumor on magnetic resonance imaging is typically indicative of a high-grade glioma. However, a significant proportion of nonenhancing gliomas can be either grade II or III. While gross total resection remains the primary goal, imaging biomarkers may guide management when surgery is not possible, especially for nonenhancing gliomas. The utility of diffusion tensor imaging and dynamic susceptibility contrast magnetic resonance imaging was evaluated in differentiating nonenhancing gliomas. Retrospective analysis was performed on imaging data from 72 nonenhancing gliomas, including grade II (n= 49) and III (n= 23) gliomas. Diffusion tensor imaging and dynamic susceptibility contrast data were used to generate fractional anisotropy, mean diffusivity, axial diffusivity, and radial diffusivity as well as cerebral blood volume, cerebral blood flow, and mean transit time maps. Univariate and multivariate logistic regression and area under the curve analyses were used to measure sensitivity and specificity of imaging parameters. A subanalysis was performed to evaluate the utility of imaging parameters in differentiating between different histologic groups. Logistic regression analysis indicated that tumor volume and relative mean transit time could differentiate between grade II and III nonenhancing gliomas. At a cutoff value of 0.33, this combination provided an area under the curve of 0.71, 70.6% sensitivity, and 64.3% specificity. Logistic regression analyses demonstrated much higher sensitivity and specificity in the differentiation of astrocytomas from oligodendrogliomas or identification of grades within these histologic subtypes. Diffusion tensor imaging and dynamic susceptibility contrast imaging can aid in differentiation of nonenhancing grade II and III gliomas and between histologic subtypes.

Partial resection offers an overall survival benefit over biopsy in MGMT-unmethylated IDH-wildtype glioblastoma patients

BackgroundIsocitrate dehydrogenase (IDH)-wildtype glioblastoma patients with O6-methylguanine-DNA-methyltransferase (MGMT)-unmethylated tumors have the worst outcome of all glioblastoma patients. The overall survival (OS) benefit of partial resection of glioblastoma compared to biopsy only remains controversial specifically in relation to molecular factors. In this report, we analyzed the effect of incomplete resection on OS compared to biopsy only in a cohort of IDH-wildtype glioblastoma patients who were uniformly treated with temozolomide-based chemoradiotherapy (TMZ-CR) after surgery.Material & MethodsA retrospective study was conducted including only glioblastoma patients who were treated with TMZ-CR after surgery from two centers. Surgical groups were defined as biopsy only, partial resection (PR) or gross total resection depending on the presence of contrast-enhancing tumor on postoperative imaging. IDH-mutation was determined using next generation sequencing technique and MGMT-methylation was analyzed with semi-quantitative methylation-specific polymerase chain reaction. Next to descriptive statistics, univariate and multivariate survival analyses were performed using Kaplan-Meier estimates and Cox regression models.ResultsIn total, 159 patients were included. 37 patients underwent biopsy only and 73 partial resections. 99 patients (62.3%) harbored unmethylated tumors. Median OS for the whole patient group was 13.4 months. In the subgroup of patients with unmethylated tumors, PR yielded a median OS of 12.2 months vs 7.6 months for biopsy patients (P = 0.003). PR proved an independent beneficial prognostic factor in multivariate Cox regression model, together with age, Karnofsky Performance Score and MGMT-methylation.ConclusionIn IDH-wildtype glioblastoma patients with MGMT-unmethylated tumors, treated with chemoradiotherapy after surgery, PR yields a significant OS benefit compared to biopsy.

The role of radiation therapy in treatment of adults with newly diagnosed glioblastoma multiforme: a systematic review and evidence-based clinical practice guideline update.

The role of radiation therapy in treatment of adults with newly diagnosed glioblastoma multiforme: a systematic review and evidence-based clinical practice guideline update.

Magnetic Resonance Molecular Imaging of Extradomain B Fibronectin Improves Imaging of Pancreatic Cancer Tumor Xenografts.

The survival of pancreatic cancer patients can be greatly improved if their disease is detected at an early, potentially curable stage. Magnetic resonance molecular imaging (MRMI) of oncoproteins is a promising strategy for accurate, early detection of the disease. Here, we test the hypothesis that MRMI of extradomain-B fibronectin (EDB-FN), an abundant oncoprotein in the tumor extracellular matrix, can overcome the stromal barriers of pancreatic cancer to facilitate effective molecular imaging and detection of small tumors. Specimens of normal, premalignant, and malignant human pancreatic tissues were stained with a peptide-fluorophore conjugate (ZD2-Cy5.5) to assess EDB-FN binding and expression. MRMI with ZD2-N3-Gd(HP-DO3A) (MT218) specific to EDB-FN and MRI with Gd(HP-DO3A) were performed in three murine models bearing human pancreatic cancer xenografts, including a Capan-1 flank model, a BxPC3-GFP-Luc and a PANC-1-GFP-Luc intrapancreatic xenograft model. Tumor enhancement of the contrast agents was analyzed and compared. Staining of human tissue samples with ZD2-Cy5.5 revealed high EDB-FN expression in pancreatic tumors, moderate expression in premalignant tissue, and little expression in normal tissue. MRMI with MT218 generated robust intratumoral contrast, clearly detected and delineated small tumors (smallest average size: 6.1 mm2), and out-performed conventional contrast enhanced MRI with Gd(HP-DO3A). Quantitative analysis of signal enhancement revealed that MT218 produced 2.7, 2.1, and 1.6 times greater contrast-to-noise ratio (CNR) than the clinical agent in the Capan-1 flank, BxPC3-GFP-Luc and PANC-1-GFP-Luc intrapancreatic models, respectively (p < 0.05). MRMI of the ECM oncoprotein EDB-FN with MT218 is able to generate superior contrast enhancement in small pancreatic tumors and provide accurate tumor delineation in animal models. Early, accurate detection and delineation of pancreatic cancer with high-resolution MRMI has the potential to guide timely treatment and significantly improve the long-term survival of pancreatic cancer patients.

The value of intraoperative MRI in recurrent intracranial tumor surgery

OBJECTIVE Identifying tumor remnants in previously operated tumor lesions remains a challenge. Intraoperative MRI (ioMRI) helps the neurosurgeon to reorient and update image guidance during surgery. The purpose of this study was to analyze whether ioMRI is more efficient in detecting tumor remnants in the surgery of recurrent lesions compared with primary surgery. METHODS All consecutive patients undergoing elective intracranial tumor surgery between 2013 and 2018 at the authors’ institution were included in this retrospective cohort study. The cohort was divided into two groups: re-craniotomy and primary craniotomy. In contrast-enhancing tumors, tumor suspicion in ioMRI was defined as contrast enhancement in T1-weighted imaging. In non–contrast-enhancing tumors, tumor suspicion was defined as hypointensity in T1-weighted imaging and hyperintensity in T2-weighted imaging and FLAIR. In cases in which the ioMRI tumor suspicion was a false positive and not confirmed during in situ inspection by the neurosurgeon, the signal was defined as a tumor-imitating ioMRI signal (TIM). Descriptive statistics were performed. RESULTS A total of 214 tumor surgeries met the inclusion criteria. The re-craniotomy group included 89 surgeries, and the primary craniotomy group included 123 surgeries. Initial complete resection after ioMRI was less frequent in the re-craniotomy group than in the primary craniotomy group, but this was not a statistically significant difference. Radiological suspicion of tumor remnants in ioMRI was present in 78% of re-craniotomy surgeries and 69% of primary craniotomy surgeries. The incidence of false-positive TIMs was significantly higher in the re-craniotomy group (n = 11, 12%) compared with the primary craniotomy group (n = 5, 4%; p = 0.015), and in contrast-enhancing tumors was related to hemorrhages in situ (n = 9). CONCLUSIONS A history of previous surgery in contrast-enhancing tumors made correct identification of tumor remnants in ioMRI more difficult, with a higher rate of false-positive ioMRI signals in the re-craniotomy group. The majority of TIMs were associated with the inability to distinguish contrast enhancement from hyperacute hemorrhage. The addition of a specific sequence in ioMRI to further differentiate both should be investigated in future studies.

Diagnostic utility of Restriction Spectrum Imaging in the characterization of the peritumoral brain zone in glioblastoma: Analysis of overall and progression-free survival

PurposeWe studied the ability of Restriction Spectrum Imaging (RSI), a novel advanced diffusion imaging technique, to estimate levels of cellularity in different glioblastoma regions, evaluated their prognostic value compared with established clinical diffusion metrics such as fractional anisotropy (FA) and mean diffusivity (MD). MethodsForty-two patients with untreated glioblastoma, IDH-wildtype, were examined with an advanced MRI tumor protocol. The region of interest (ROI) was obtained from the contrast-enhancing part of tumor and the peritumoral brain zones and then co-registered with RSI-cellularity index, FA and MD maps. Histogram parameters of diffusion metrics were assessed for all ROI locations and compared to MGMT promoter methylation status and survival. The ability of RSI-cellularity index, FA, and MD to stratify survival and were assessed by Cox proportional hazard regression, adjusted for significant clinical predictors. ResultsThe highest RSI-cellularity index was measured in contrast-enhancing tumor core with a negative gradient from tumor core to the periphery of peritumoral zone with predictive accuracy 81 % (P < 0.001). Shorter overall survival was significant associated with higher RSI-cellularity index (hazard ratio (HR) 3.6, 95 % confidence interval (CI) 1.3−9.5, P = 0.002) with synchronal decrease in MD (HR 0.31, 95 %CI 0.1−0.8, P = 0.008) in the contrast-enhanced tumor core. This association was also consistent for RSI-cellularity index value measured in the peri-enhancing zone (HR 3.6, 95 % CI 1.0−12.3, P = 0.041). No statistically significant differences were noted between RSI-cellularity index, FA, nor MD and MGMT promoter methylation. ConclusionRSI-cellularity index may be used as prognostic biomarker to improve risk stratification in patients with glioblastoma.

Extent of resection affects prognosis for patients with glioblastoma in non-eloquent regions

Glioblastoma (GBM) is a malignant cerebral neoplasm carrying poor prognosis. The importance of extent of resection (EoR) in GBM patient outcomes has been argued in the literature. Previous studies included tumors in eloquent regions of the brain. This confounds the role of EoR by including patients with intrinsically worse outcomes but will be over-represented in the reduced EoR category. In a homogenous group of patients in whom GTR was considered achievable, we investigated the effect of increasing EoR on survival. A retrospective review of 51 patients was undertaken. Quantitative, volumetric analysis of pre-operative and post-operative magnetic resonance image was compared with corresponding clinical details. The primary outcome measured was post-operative overall survival. Median overall survival was 18.3 months for GTR patients compared to 11.6 months for non-GTR (p = 0.025). Median pre-operative contrast-enhancing tumor volume for GTR patients was 54.7 cm3 and 24.9 cm3 for non-GTR. Post-operative median residual tumor volume was 1.1 cm3 in the non-GTR cohort. In multivariate analyses, GTR (HR [95% CI] = 0.973 [0.954–0.994], p = 0.00559) and increasing EoR (HR [95% CI] = 0.964 [0.944–0.985], p = 0.000665) remained predictors of survival. Centile dichotomization of EoR revealed 74% (HR [95% CI] = 0.351 [0.128–0.958], p = 0.0409) as the lowest threshold conferring statistically significant survival benefit. Where technically feasible, both GTR and EoR remained as independent prognostic factors for survival. GTR remains the gold standard for surgical treatment of GBM in patients, 74% being the minimum EoR required to confer survival benefit.

Radiomics model predicts granulation pattern in growth hormone-secreting pituitary adenomas.

To investigate whether radiomic features from magnetic resonance image (MRI) can predict the granulation pattern of growth hormone (GH)-secreting pituitary adenoma patients. Sixty-nine pathologically proven acromegaly patients (densely granulated [DG] = 50, sparsely granulated [SG] = 19) were included. Radiomic features (n = 214) were extracted from contrast-enhancing and total tumor portions from T2-weighted (T2) MRIs. Imaging features were selected using a least absolute shrinkage and selection operator (LASSO) logistic regression model with fivefold cross-validation. Diagnostic performance for predicting granulation pattern was compared with that for qualitative T2 signal intensity assessment and T2 relative signal intensity (rSI) using the area under the receiver operating characteristics curve (AUC). Four significant radiomic features from the contrast-enhancing tumor (1 from shape, 1 from first order feature, and 2 from second order features) were selected by LASSO for model construction. The radiomics model showed an AUC, accuracy, sensitivity, and specificity of 0.834 (95% confidence interval [CI] 0.738-0.930), 73.7%, 74.0%, and 73.9%, respectively. The radiomics model showed significantly better performance than the model using qualitative T2 signal intensity assessment (AUC 0.597 [95% CI 0.447-0.747], P = 0.009) and T2 rSI (AUC 0.647 [95% CI 0.523-0.759], P = 0.037). Radiomic features may be useful biomarkers to differentiate granulation pattern of GH-secreting pituitary adenoma patients, and showed better performance than qualitative assessment or rSI evaluation.

Validation of combined use of DWI and percentage signal recovery-optimized protocol of DSC-MRI in differentiation of high-grade glioma, metastasis, and lymphoma.

With conventional MRI, it is often difficult to effectively differentiate between contrast-enhancing brain tumors, including primary central nervous system lymphoma (PCNSL), high-grade glioma (HGG), and metastasis. This study aimed to assess the discrimination ability of the parameters obtained from DWI and the percentage signal recovery- (PSR-) optimized protocol of DSC-MRI between these three tumor types at an initial step. DSC-MRI using a PSR-optimized protocol (TR/TE = 1500/30 ms, flip angle = 90°, no preload) and DWI of 99 solitary enhancing tumors (60 HGGs, 24 metastases, 15 PCNSLs) were retrospectively assessed before treatment. rCBV, PSR, ADC in the tumor core and rCBV, and ADC in peritumoral edema were measured. The differences were evaluated using one-way ANOVA, and the diagnostic performance was evaluated using ROC curve analysis. PSR in the tumor core showed the best discriminating performance in differentiating these three tumor types with AUC values of 0.979 for PCNSL vs. others and 0.947 for HGG vs. metastasis. The ADC was only helpful in the tumor core and distinguishing PCNSLs from others (AUC = 0.897). Different from CBV-optimized protocols (preload, intermediate FA), PSR derived from the PSR-optimized protocol seems to be the most important parameter in the differentiation of HGGs, metastases, and PCNSLs at initial diagnosis. This property makes PSR remarkable and carries the need for comprehensive DSC-MRI protocols, which provides PSR sensitivity and CBV accuracy together, such as the preload use of the PSR-optimized protocol before the CBV-optimized protocol.

Abstract 866: Radiomic prediction of survival in recurrent high-grade glioma patients treated with CAR T-cell therapy

Abstract Introduction: High-grade glioma (HGG) is the most common subtype of primary brain tumors with high recurrence rate and poor survival. The emergence of targeted molecular and cellular therapies (e.g., pembrolizumab, Chimeric Antigen Receptor [CAR] T-cell therapy) are potentially promising in improving overall survival. Due to increased intratumor heterogeneity and inhomogeneous treatment response, there is an unmet need of imaging biomarkers predictive of treatment response and survival. Radiomics using machine learning methods have shown promise in predicting treatment response in various solid tumors, including HGG. In this study, we compare the survival prediction performance using machine learning models with different radiomic features individually derived from T1- and T2-weighted MR images in patients suffering from HGG treated with CAR-T cell therapy. Methods: In this IRB-approved phase 1 clinical trial, 61 patients (39 males, median age = 49) suffering from recurrent HGG underwent surgical resection and CAR T-cell therapy1. All patients underwent baseline MRI scans prior to both surgical resection and CAR T-cell administration in the resection cavity. For patients with a complete set of T1- and T2-weighted MRIs (n = 50), we generated segmentations in a semi-automated manner, labeling with each tumorous voxel as either contrast-enhanced tumor (ET), non-enhancing tumor (NET) and edema. From each tumor label, we extracted shape-based, texture-based and image-filtered radiomic features2. We utilized gradient-boosted tree models (lightGBM) to classify whether survival is above or below group median (188 days) by using two nested loops of 10-fold cross validations each. For the inner validation loop, we determined the optimal model from hyper-parameters including regularization. For the outer validation loop, we tested this model on the hold-out data and the predictions were used as radiomic risk scores. Results: For each of the ET, NET, and edema tumor ROIs, we extracted 1313 radiomic features for predictive modeling. The outer validation loop Area Under the Receiver Operating Characteristic Curve (AUC) for ET, NET and edema were 0.55, 0.70, and 0.46, respectively, suggesting that radiomic features calculated from NET voxels are the most predictive of survival compared to features from ET and edema voxels. We also stratified the patients into two distinct prognostic sub-groups (25 patients each group) using the NET radiomic risk scores obtained from the outer validation loop, with a log-rank test p-value of 0.01. Conclusions: In patients suffering from recurrent HGG who were treated with CAR T-cell therapy, we found that radiomic features derived from NET voxels are predictive of survival while the other two tumorous voxel types (ET and edema) are not. Further work is needed to incorporate clinical and molecular features that also may be predictive of survival. 1N Engl J Med 2016; 375:2561-2569. 2Cancer Research, 77(21), e104-e107 Citation Format: Chi Wah Wong, Sohaib Naim, Vincent La, Seth Michael Hilliard, Eemon Tizpa, Rashi Ranjan, Hannah Jade Young, Kimberly Jane Bonjoc, Aleksandr Filippov, Saman Tabassum Khan, Christine Brown, Behnam Badie, Ammar Ahmed Chaudhry. Radiomic prediction of survival in recurrent high-grade glioma patients treated with CAR T-cell therapy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 866.

Primary Central Nervous System Lymphoma: Clinical Evaluation of Automated Segmentation on Multiparametric MRI Using Deep Learning

Precise volumetric assessment of brain tumors is relevant for treatment planning and monitoring. However, manual segmentations are time-consuming and impeded by intra- and interrater variabilities. To investigate the performance of a deep-learning model (DLM) to automatically detect and segment primary central nervous system lymphoma (PCNSL) on clinical MRI. Retrospective. Sixty-nine scans (at initial and/or follow-up imaging) from 43 patients with PCNSL referred for clinical MRI tumor assessment. T1 -/T2 -weighted, T1 -weighted contrast-enhanced (T1 CE), and FLAIR at 1.0, 1.5, and 3.0T from different vendors and study centers. Fully automated voxelwise segmentation of tumor components was performed using a 3D convolutional neural network (DeepMedic) trained on gliomas (n = 220). DLM segmentations were compared to manual segmentations performed in a 3D voxelwise manner by two readers (radiologist and neurosurgeon; consensus reading) from T1 CE and FLAIR, which served as the reference standard. Dice similarity coefficient (DSC) for comparison of spatial overlap with the reference standard, Pearson's correlation coefficient (r) to assess the relationship between volumetric measurements of segmentations, and Wilcoxon rank-sum test for comparison of DSCs obtained in initial and follow-up imaging. The DLM detected 66 of 69 PCNSL, representing a sensitivity of 95.7%. Compared to the reference standard, DLM achieved good spatial overlap for total tumor volume (TTV, union of tumor volume in T1 CE and FLAIR; average size 77.16 ± 62.4 cm3 , median DSC: 0.76) and tumor core (contrast enhancing tumor in T1 CE; average size: 11.67 ± 13.88 cm3 , median DSC: 0.73). High volumetric correlation between automated and manual segmentations was observed (TTV: r = 0.88, P < 0.0001; core: r = 0.86, P < 0.0001). Performance of automated segmentations was comparable between pretreatment and follow-up scans without significant differences (TTV: P = 0.242, core: P = 0.177). In clinical MRI scans, a DLM initially trained on gliomas provides segmentation of PCNSL comparable to manual segmentation, despite its complex and multifaceted appearance. Segmentation performance was high in both initial and follow-up scans, suggesting its potential for application in longitudinal tumor imaging. 3 TECHNICAL EFFICACY STAGE: 2.

The Association Between the Extent of Glioblastoma Resection and Survival in Light of MGMT Promoter Methylation in 326 Patients With Newly Diagnosed IDH-Wildtype Glioblastoma.

Background: The association between contrast enhanced (CE) and non-contrast enhanced (NCE) tumor resection and survival in patients with glioblastoma in relation to molecular subtypes is poorly understood. The aim of this study was to assess the association between CE and NCE tumor resection and survival in light of MGMT promoter methylation in newly diagnosed IDH-wildtype glioblastoma.Materials and methods: Patients with newly diagnosed IDH-wildtype glioblastoma who underwent surgery were eligible. CE and NCE tumor volumes were assessed on pre- and post-operative MRI scans and extent of resection was calculated. The association between CE and NCE tumor resection and survival was evaluated using multivariable Cox proportional hazards models and Kaplan Meier estimates.Results: Three hundred and twenty-six patients were included: 177 (54.3%) with and 149 (45.7%) without MGMT methylation. Multivariable Cox proportional hazards models stratified for MGMT methylation identified age ≤ 65y (HR 0.63; 95% CI, 0.49–0.81; p < 0.0001), chemoradiation (HR 0.13; 95% CI, 0.09–0.19; p < 0.0001), maximal CE tumor resection (HR 0.58; 95% CI, 0.39–0.87; p = 0.009), ≥ 30% NCE tumor resection (HR 0.71; 95% CI, 0.53–0.93; p = 0.014), and minimal residual CE tumor volume (HR 0.64; 95% CI, 0.46–0.88 p = 0.007) as being associated with longer overall survival. Kaplan Meier estimates showed that extensive surgery was more beneficial for patients with MGMT methylated glioblastoma.Conclusions: This study shows an association between maximal CE tumor resection, ≥30% NCE tumor resection, minimal residual CE tumor volume, and longer overall survival in patients with newly diagnosed IDH wildtype glioblastoma. Intraoperative imaging and stimulation mapping may be used to pursue safe and maximal resection. In future research, the safety aspect of maximizing tumor resection needs to be addressed.

Radiogenomics correlation between MR imaging features and mRNA-based subtypes in lower-grade glioma

BackgroundTo investigate associations between lower-grade glioma (LGG) mRNA-based subtypes (R1-R4) and MR features.MethodsmRNA-based subtyping was obtained from the LGG dataset in The Cancer Genome Atlas (TCGA). We identified matching patients (n = 145) in The Cancer Imaging Archive (TCIA) who underwent MR imaging. The associations between mRNA-based subtypes and MR features were assessed.ResultsIn the TCGA-LGG dataset, patients with the R2 subtype had the shortest median OS months (P < 0.05). The time-dependent ROC for the R2 subtype was 0.78 for survival at 12 months, 0.76 for survival at 24 months, and 0.76 for survival at 36 months. In the TCIA-LGG dataset, 41 (23.7%) R1 subtype, 40 (23.1%) R2 subtype, 19 (11.0%) R3 subtype and 45 (26.0%) R4 subtype cases were identified. Multivariate analysis revealed that enhancing margin (ill-defined, OR: 9.985; P = 0.003) and T1 + C/T2 mismatch (yes, OR: 0.091; P = 0.023) were associated with the R1 subtype (AUC: 0.708). The average accuracy of the ten-fold cross validation was 71%. Proportion of contrast-enhanced (CE) tumour (> 5%, OR: 14.733; P < 0.001) and necrosis/cystic changes (yes, OR: 0.252; P = 0.009) were associated with the R2 subtype (AUC: 0.832). The average accuracy of the ten-fold cross validation was 82%. Haemorrhage (yes, OR: 8.55; P < 0.001) was positively associated with the R3 subtype (AUC: 0.689). The average accuracy of the ten-fold cross validation was 87%. Proportion of CE tumour (> 5%, OR: 0.14; P < 0.001) was negatively associated with the R4 subtype (AUC: 0.672). The average accuracy of the ten-fold cross validation was 71%. For the prediction of the R2 subtype, the nomogram showed good discrimination and calibration. Decision curve analysis demonstrated that prediction with the R2 model was clinically useful.ConclusionsPatients with the R2 subtype had the worst prognosis. We demonstrated that MRI features can identify distinct LGG mRNA-based molecular subtypes.