Evaluating the Predictive Value of Glioma Growth Models for Low-Grade Glioma After Tumor Resection

Tumor growth models have the potential to model and predict the spatiotemporal evolution of glioma in individual patients. Infiltration of glioma cells is known to be faster along the white matter tracts, and therefore structural magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI) can be used to inform the model. However, applying and evaluating growth models in real patient data is challenging. In this work, we propose to formulate the problem of tumor growth as a ranking problem, as opposed to a segmentation problem, and use the average precision (AP) as a performance metric. This enables an evaluation of the spatial pattern that does not require a volume cut-off value. Using the AP metric, we evaluate diffusion-proliferation models informed by structural MRI and DTI, after tumor resection. We applied the models to a unique longitudinal dataset of 14 patients with low-grade glioma (LGG), who received no treatment after surgical resection, to predict the recurrent tumor shape after tumor resection. The diffusion models informed by structural MRI and DTI showed a small but significant increase in predictive performance with respect to homogeneous isotropic diffusion, and the DTI-informed model reached the best predictive performance. We conclude there is a significant improvement in the prediction of the recurrent tumor shape when using a DTI-informed anisotropic diffusion model with respect to istropic diffusion, and that the AP is a suitable metric to evaluate these models. All code and data used in this publication are made publicly available.

Federated learning enables big data for rare cancer boundary detection

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

Combined molecular subtyping, grading, and segmentation of glioma using multi-task deep learning

BACKGROUND

Accurate characterization of glioma is crucial for clinical decision making. A delineation of the tumor is also desirable in the initial decision stages but is time-consuming. Previously, deep learning methods have been developed that can either non-invasively predict the genetic or histological features of glioma, or that can automatically delineate the tumor, but not both tasks at the same time. Here, we present our method that can predict the molecular subtype and grade, while simultaneously providing a delineation of the tumor.

METHODS

We developed a single multi-task convolutional neural network that uses the full 3D, structural, preoperative MRI scans to predict the IDH mutation status, the 1p/19q co-deletion status, and the grade of a tumor, while simultaneously segmenting the tumor. We trained our method using a patient cohort containing 1508 glioma patients from 16 institutes. We tested our method on an independent dataset of 240 patients from 13 different institutes.

RESULTS

In the independent test set, we achieved an IDH-AUC of 0.90, an 1p/19q co-deletion AUC of 0.85, and a grade AUC of 0.81 (grade II/III/IV). For the tumor delineation, we achieved a mean whole tumor Dice score of 0.84.

CONCLUSIONS

We developed a method that non-invasively predicts multiple, clinically relevant features of glioma. Evaluation in an independent dataset shows that the method achieves a high performance and that it generalizes well to the broader clinical population. This first-of-its-kind method opens the door to more generalizable, instead of hyper-specialized, AI methods.

The impact of different volumetric thresholds to determine progressive disease in patients with recurrent glioblastoma treated with bevacizumab

Background

The optimal volumetric threshold for determining progressive disease (PD) in recurrent glioblastoma is yet to be determined. We investigated a range of thresholds in association with overall survival (OS).

Methods

First recurrent glioblastoma patients treated with bevacizumab and/or lomustine were included from the phase II BELOB and phase III EORTC26101 trials. Enhancing and nonenhancing tumor volumes were measured at baseline, first (6 weeks), and second (12 weeks) follow-up. Hazard ratios (HRs) for the appearance of new lesions and several thresholds for tumor volume increase were calculated using cox regression analysis. Results were corrected in a multivariate analysis for well-established prognostic factors.

Results

At first and second follow-up, 138 and 94 patients respectively, were deemed eligible for analysis of enhancing volumes, while 89 patients were included in the analysis of nonenhancing volumes at first follow-up. New lesions were associated with a significantly worse OS (3.2 versus 11.2 months, HR = 7.03, P < .001). At first follow-up a threshold of enhancing volume increase of ≥20% provided the highest HR (5.55, p = .001. At second follow-up, any increase in enhancing volume (≥0%) provided the highest HR (9.00, p < .001). When measuring nonenhancing volume at first follow-up, only 6 additional patients were scored as PD with the highest HR of ≥25% increase in volume (HR=3.25, p = .008).

Conclusion

Early appearing new lesions were associated with poor OS. Lowering the volumetric threshold for PD at both first and second follow-up improved survival prediction. However, the additional number of patients categorized as PD by lowering the threshold was very low. The per-RANO added change in nonenhancing volumes to the analyses was of limited value.

The Erasmus Glioma Database (EGD): Structural MRI scans, WHO 2016 subtypes, and segmentations of 774 patients with glioma

The Erasmus Glioma Database (EGD) contains structural magnetic resonance imaging (MRI) scans, genetic and histological features (specifying the WHO 2016 subtype), and whole tumor segmentations of patients with glioma. Pre-operative MRI data of 774 patients with glioma (281 female, 492 male, 1 unknown, age range 19-86 years) treated at the Erasmus MC between 2008 and 2018 is available. For all patients a pre-contrast T1-weighted, post-contrast T1-weighted, T2-weighted, and T2-weighted FLAIR scan are available, made on a variety of scanners from four different vendors. All scans are registered to a common atlas and defaced. Genetic and histological data consists of the IDH mutation status (available for 467 patients), 1p/19q co-deletion status (available for 259 patients), and grade (available for 716 patients). The full WHO 2016 subtype is available for 415 patients. Manual segmentations are available for 374 patients and automatically generated segmentations are available for 400 patients. The dataset can be used to relate the visual appearance of the tumor on the scan with the genetic and histological features, and to develop automatic segmentation methods.

Multiple-correlation similarity for block-matching based fast CT to ultrasound registration in liver interventions

In this work we present a fast approach to perform registration of computed tomography to ultrasound volumes for image guided intervention applications. The method is based on a combination of block-matching and outlier rejection. The block-matching uses a correlation based multimodal similarity metric, where the intensity and the gradient of the computed tomography images along with the ultrasound volumes are the input images to find correspondences between blocks in the computed tomography and the ultrasound volumes. A variance and octree based feature point-set selection method is used for selecting distinct and evenly spread point locations for block-matching. Geometric consistency and smoothness criteria are imposed in an outlier rejection step to refine the block-matching results. The block-matching results after outlier rejection are used to determine the affine transformation between the computed tomography and the ultrasound volumes. Various experiments are carried out to assess the optimal performance and the influence of parameters on accuracy and computational time of the registration. A leave-one-patient-out cross-validation registration error of 3.6 mm is achieved over 29 datasets, acquired from 17 patients.

The impact of surgery in molecularly defined low-grade glioma: an integrated clinical, radiological, and molecular analysis

Background

Extensive resections in low-grade glioma (LGG) are associated with improved overall survival (OS). However, World Health Organization (WHO) classification of gliomas has been completely revised and is now predominantly based on molecular criteria. This requires reevaluation of the impact of surgery in molecularly defined LGG subtypes.

Methods

We included 228 adults who underwent surgery since 2003 for a supratentorial LGG. Pre- and postoperative tumor volumes were assessed with semiautomatic software on T2-weighted images. Targeted next-generation sequencing was used to classify samples according to current WHO classification. Impact of postoperative volume on OS, corrected for molecular profile, was assessed using a Cox proportional hazards model.

Results

Median follow-up was 5.79 years. In 39 (17.1%) histopathologically classified gliomas, the subtype was revised after molecular analysis. Complete resection was achieved in 35 patients (15.4%), and in 54 patients (23.7%) only small residue (0.1-5.0 cm3) remained. In multivariable analysis, postoperative volume was associated with OS, with a hazard ratio of 1.01 (95% CI: 1.002-1.02; P = 0.016) per cm3 increase in volume. The impact of postoperative volume was particularly strong in isocitrate dehydrogenase (IDH) mutated astrocytoma patients, where even very small postoperative volumes (0.1-5.0 cm) already negatively affected OS.

Conclusion

Our data provide the necessary reevaluation of the impact of surgery in molecularly defined LGG and support maximal resection as first-line treatment for molecularly defined LGG. Importantly, in IDH mutated astrocytoma, even small postoperative volumes have negative impact on OS, which argues for a second-look operation in this subtype to remove minor residues if safely possible.

Comparison of 2D (RANO) and volumetric methods for assessment of recurrent glioblastoma treated with bevacizumab-a report from the BELOB trial

Background

The current method for assessing progressive disease (PD) in glioblastoma is according to the Response Assessment in Neuro-Oncology (RANO) criteria. Bevacizumab-treated patients may show pseudo-response on postcontrast T1-weighted (T1w) MRI, and a more infiltrative non-enhancing growth pattern on T2w/fluid attenuated inversion recovery (FLAIR) images. We investigated whether the RANO criteria remain the method of choice for assessing bevacizumab-treated recurrent glioblastoma when compared with various volumetric methods.

Methods

Patients with assessable MRI data from the BELOB trial (n = 148) were included. Patients were treated with bevacizumab, lomustine, or both. At first and second radiological follow-up (6 and 12 wk), PD was determined using the 2D RANO criteria and various volumetric methods based on enhancing tumor only and enhancing plus non-enhancing tumor. Differences in overall survival (OS) between PD and non-PD patients were assessed with the log-rank test and a Cox model. Hazard ratios (HRs) and their 95% CIs were determined.

Results

For all patients together, all methods (except subtraction of non-enhancing from enhancing volume at first follow-up) showed significant differences in OS between PD and non-PD patients (P < .001). The largest risk increase for death in case of PD at both first and second follow-up was found with the RANO criteria: HR = 2.81 (95% CI, 1.92-4.10) and HR = 2.80 (95% CI, 1.75-4.49), respectively. In the bevacizumab-treated patients, all methods assessed showed significant differences in OS between PD and non-PD patients. There were no significant differences between methods.

Conclusions

In the first 12 weeks, volumetric methods did not provide significant improvement over the RANO criteria as a posttreatment prognostic marker.

Diffusion MRI Phenotypes Predict Overall Survival Benefit from Anti-VEGF Monotherapy in Recurrent Glioblastoma: Converging Evidence from Phase II Trials

Purpose: Anti-VEGF therapies remain controversial in the treatment of recurrent glioblastoma (GBM). In the current study, we demonstrate that recurrent GBM patients with a specific diffusion MR imaging signature have an overall survival (OS) advantage when treated with cediranib, bevacizumab, cabozantinib, or aflibercept monotherapy at first or second recurrence. These findings were validated using a separate trial comparing bevacizumab with lomustine.Experimental Design: Patients with recurrent GBM and diffusion MRI from the monotherapy arms of 5 separate phase II clinical trials were included: (i) cediranib (NCT00035656); (ii) bevacizumab (BRAIN Trial, AVF3708g; NCT00345163); (iii) cabozantinib (XL184-201; NCT00704288); (iv) aflibercept (VEGF Trap; NCT00369590); and (v) bevacizumab or lomustine (BELOB; NTR1929). Apparent diffusion coefficient (ADC) histogram analysis was performed prior to therapy to estimate "ADC_L," the mean of the lower ADC distribution. Pretreatment ADC_L, enhancing volume, and clinical variables were tested as independent prognostic factors for OS.Results: The coefficient of variance (COV) in double baseline ADC_L measurements was 2.5% and did not significantly differ (P = 0.4537). An ADC_L threshold of 1.24 μm²/ms produced the largest OS differences between patients (HR ∼ 0.5), and patients with an ADC_L > 1.24 μm²/ms had close to double the OS in all anti-VEGF therapeutic scenarios tested. Training and validation data confirmed that baseline ADC_L was an independent predictive biomarker for OS in anti-VEGF therapies, but not in lomustine, after accounting for age and baseline enhancing tumor volume.Conclusions: Pretreatment diffusion MRI is a predictive imaging biomarker for OS in patients with recurrent GBM treated with anti-VEGF monotherapy at first or second relapse. Clin Cancer Res; 23(19); 5745-56. ©2017 AACR.