Inter-observer variability in the measurement of diffuse intrinsic pontine gliomas

The relationship between imaging features, therapeutic response, and overall survival in pediatric diffuse intrinsic pontine glioma

We aimed to evaluate the relationship between imaging features, therapeutic responses (comparative cross-product and volumetric measurements), and overall survival (OS) in pediatric diffuse intrinsic pontine glioma (DIPG). A total of 134 patients (≤ 18 years) diagnosed with DIPG were included. Univariate and multivariate analyses were performed to evaluate correlations of clinical and imaging features and therapeutic responses with OS. The correlation between cross-product (CP) and volume thresholds in partial response (PR) was evaluated by linear regression. The log-rank test was used to compare OS patients with discordant therapeutic response classifications and those with concordant classifications. In univariate analysis, characteristics related to worse OS included lower Karnofsky, larger extrapontine extension, ring-enhancement, necrosis, non-PR, and increased ring enhancement post-radiotherapy. In the multivariate analysis, Karnofsky, necrosis, extrapontine extension, and therapeutic response can predict OS. A 25% CP reduction (PR) correlated with a 32% volume reduction (R2 = 0.888). Eight patients had discordant therapeutic response classifications according to CP (25%) and volume (32%). This eight patients' median survival time was 13.0 months, significantly higher than that in the non-PR group (8.9 months), in which responses were consistently classified as non-PR based on CP (25%) and volume (32%). We identified correlations between imaging features, therapeutic responses, and OS; this information is crucial for future clinical trials. Tumor volume may represent the DIPG growth pattern more accurately than CP measurement and can be used to evaluate therapeutic response.

The use of artificial intelligence tools in cancer detection compared to the traditional diagnostic imaging methods: An overview of the systematic reviews

BACKGROUND AND PURPOSE

In comparison to conventional medical imaging diagnostic modalities, the aim of this overview article is to analyze the accuracy of the application of Artificial Intelligence (AI) techniques in the identification and diagnosis of malignant tumors in adult patients.

DATA SOURCES

The acronym PIRDs was used and a comprehensive literature search was conducted on PubMed, Cochrane, Scopus, Web of Science, LILACS, Embase, Scielo, EBSCOhost, and grey literature through Proquest, Google Scholar, and JSTOR for systematic reviews of AI as a diagnostic model and/or detection tool for any cancer type in adult patients, compared to the traditional diagnostic radiographic imaging model. There were no limits on publishing status, publication time, or language. For study selection and risk of bias evaluation, pairs of reviewers worked separately.

RESULTS

In total, 382 records were retrieved in the databases, 364 after removing duplicates, 32 satisfied the full-text reading criterion, and 09 papers were considered for qualitative synthesis. Although there was heterogeneity in terms of methodological aspects, patient differences, and techniques used, the studies found that several AI approaches are promising in terms of specificity, sensitivity, and diagnostic accuracy in the detection and diagnosis of malignant tumors. When compared to other machine learning algorithms, the Super Vector Machine method performed better in cancer detection and diagnosis. Computer-assisted detection (CAD) has shown promising in terms of aiding cancer detection, when compared to the traditional method of diagnosis.

CONCLUSIONS

The detection and diagnosis of malignant tumors with the help of AI seems to be feasible and accurate with the use of different technologies, such as CAD systems, deep and machine learning algorithms and radiomic analysis when compared with the traditional model, although these technologies are not capable of to replace the professional radiologist in the analysis of medical images. Although there are limitations regarding the generalization for all types of cancer, these AI tools might aid professionals, serving as an auxiliary and teaching tool, especially for less trained professionals. Therefore, further longitudinal studies with a longer follow-up duration are required for a better understanding of the clinical application of these artificial intelligence systems.

TRIAL REGISTRATION

Systematic review registration. Prospero registration number: CRD42022307403.

A Review of Recent Advances in Brain Tumor Diagnosis Based on AI-Based Classification

Uncontrolled and fast cell proliferation is the cause of brain tumors. Early cancer detection is vitally important to save many lives. Brain tumors can be divided into several categories depending on the kind, place of origin, pace of development, and stage of progression; as a result, tumor classification is crucial for targeted therapy. Brain tumor segmentation aims to delineate accurately the areas of brain tumors. A specialist with a thorough understanding of brain illnesses is needed to manually identify the proper type of brain tumor. Additionally, processing many images takes time and is tiresome. Therefore, automatic segmentation and classification techniques are required to speed up and enhance the diagnosis of brain tumors. Tumors can be quickly and safely detected by brain scans using imaging modalities, including computed tomography (CT), magnetic resonance imaging (MRI), and others. Machine learning (ML) and artificial intelligence (AI) have shown promise in developing algorithms that aid in automatic classification and segmentation utilizing various imaging modalities. The right segmentation method must be used to precisely classify patients with brain tumors to enhance diagnosis and treatment. This review describes multiple types of brain tumors, publicly accessible datasets, enhancement methods, segmentation, feature extraction, classification, machine learning techniques, deep learning, and learning through a transfer to study brain tumors. In this study, we attempted to synthesize brain cancer imaging modalities with automatically computer-assisted methodologies for brain cancer characterization in ML and DL frameworks. Finding the current problems with the engineering methodologies currently in use and predicting a future paradigm are other goals of this article.

Current Role and Future Potential of CSF ctDNA for the Diagnosis and Clinical Management of Pediatric Central Nervous System Tumors

Most pediatric central nervous system (CNS) tumors are located in eloquent anatomic areas, making surgical resection and, in some cases, even biopsy risky or impossible. This diagnostic predicament coupled with the move toward molecular classification for diagnosis has exposed an urgent need to develop a minimally invasive means to obtain diagnostic information. In non-CNS solid tumors, the detection of circulating tumor DNA (ctDNA) in plasma and other bodily fluids has been incorporated into routine practice and clinical trial design for selection of molecular targeted therapy and longitudinal monitoring. For primary CNS tumors, however, detection of ctDNA in plasma has been challenging. This is likely related at least in part to anatomic factors such as the blood-brain barrier. Due to the proximity of primary CNS tumors to the cerebrospinal fluid (CSF) space, our group and others have turned to CSF as a rich alternative source of ctDNA. Although multiple studies at this time have demonstrated the feasibility of CSF ctDNA detection across multiple types of pediatric CNS tumors, the optimal role and utility of CSF ctDNA in the clinical setting has not been established. This review discusses the work-to-date on CSF ctDNA liquid biopsy in pediatric CNS tumors and the associated technical challenges, and reviews the promising opportunities that lie ahead for integration of CSF ctDNA liquid biopsy into clinical care and clinical trial design.

Sequential Diffusion Tensor Imaging and Magnetic Resonance Spectroscopy in Patients Undergoing Reirradiation for Progressive Diffuse Intrinsic Pontine Glioma

Purpose

Diffusion tensor imaging for evaluation of white matter tracts is used with magnetic resonance spectroscopy (MRS) to improve management of diffuse intrinsic pontine glioma (DIPG). Changes in the apparent diffusion coefficient (ADC), fractional anisotropy (FA), and tumor metabolite ratios have been reported after initial radiation for DIPG, but these markers have not been studied sequentially in patients undergoing reirradiation for progressive DIPG. Here, we report a case series of 4 patients who received reirradiation for progressive DIPG on a prospective clinical trial in which we evaluated quantitative changes in FA, ADC, and tumor metabolites and qualitative changes in white matter tracts.

Methods and Materials

The median reirradiation dose was 25.2 Gy (24-30.8 Gy). Fiber tracking was performed using standard tractography analysis. The FA and ADC values for the corticospinal and medial lemniscus tracts were calculated before and after reirradiation. Multivoxel MRS was performed. Findings were correlated with clinical features and conventional MRI of tumors.

Results

All patients had an initial response to reirradiation as shown by a decrease in tumor size. In general, FA increased with disease response and decreased with progression, whereas ADC decreased with disease response and increased with progression. At second progression, the FA fold change relative to values during disease response decreased in both patients with available imaging at second progression. Visualization of tracts demonstrated robust reconstitution of previously disrupted paths during tumor response; conversely, there was increased fiber tract disruption and infiltration during tumor progression. The MRS analysis revealed a decrease in choline:creatinine and choline:N-acetylaspartate ratios during tumor response and increase during progression.

Conclusions

Distinct changes in white matter tracts and tumor metabolism were observed in patients with DIPG undergoing reirradiation on a prospective clinical trial. Changes related to tumor response and progression were observed after 24 to 30.8 Gy reirradiation.

Volumetric endpoints in diffuse intrinsic pontine glioma: comparison to cross-sectional measures and outcome correlations in the International DIPG/DMG Registry

BACKGROUND

Cross-sectional tumor measures are traditional clinical trial endpoints; however volumetric measures may better assess tumor growth. We determined the correlation and compared the prognostic impact of cross-sectional and volumetric measures of progressive disease (PD) among patients with DIPG.

METHODS

Imaging and clinical data were abstracted from the International DIPG Registry. Tumor volume and cross-sectional product (CP) were measured with mint Lesion™ software using manual contouring. Correlation between CP and volume (segmented and mathematical [ellipsoid] model) thresholds of PD were assessed by linear regression. Landmark analyses determined differences in survival (via log-rank) between patients classified as PD versus non-PD by CP and volumetric measurements at 1, 3, 5, 7, and 9 months postradiotherapy (RT). Hazard ratios (HR) for survival after these time points were calculated by Cox regression.

RESULTS

A total of 312 MRIs (46 patients) were analyzed. Comparing change from the previous smallest measure, CP increase of 25% (PD) correlated with a segmented volume increase of 30% (R2 = 0.710), rather than 40% (spherical model extrapolation). CP-determined PD predicted survival at 1 month post-RT (HR = 2.77), but not other time points. Segmented volumetric-determined PD (40% threshold) predicted survival at all imaging timepoints (HRs = 2.57, 2.62, 3.35, 2.71, 16.29), and 30% volumetric PD threshold predicted survival at 1, 3, 5, and 9 month timepoints (HRs = 2.57, 2.62, 4.65, 5.54). Compared to ellipsoid volume, segmented volume demonstrated superior survival associations.

CONCLUSIONS

Segmented volumetric assessments of PD correlated better with survival than CP or ellipsoid volume at most time points. Semiautomated tumor volume likely represents a more accurate, prognostically-relevant measure of disease burden in DIPG.

Deep learning-based automatic tumor burden assessment of pediatric high-grade gliomas, medulloblastomas, and other leptomeningeal seeding tumors

BACKGROUND

Longitudinal measurement of tumor burden with magnetic resonance imaging (MRI) is an essential component of response assessment in pediatric brain tumors. We developed a fully automated pipeline for the segmentation of tumors in pediatric high-grade gliomas, medulloblastomas, and leptomeningeal seeding tumors. We further developed an algorithm for automatic 2D and volumetric size measurement of tumors.

METHODS

The preoperative and postoperative cohorts were randomly split into training and testing sets in a 4:1 ratio. A 3D U-Net neural network was trained to automatically segment the tumor on T1 contrast-enhanced and T2/FLAIR images. The product of the maximum bidimensional diameters according to the RAPNO (Response Assessment in Pediatric Neuro-Oncology) criteria (AutoRAPNO) was determined. Performance was compared to that of 2 expert human raters who performed assessments independently. Volumetric measurements of predicted and expert segmentations were computationally derived and compared.

RESULTS

A total of 794 preoperative MRIs from 794 patients and 1003 postoperative MRIs from 122 patients were included. There was excellent agreement of volumes between preoperative and postoperative predicted and manual segmentations, with intraclass correlation coefficients (ICCs) of 0.912 and 0.960 for the 2 preoperative and 0.947 and 0.896 for the 2 postoperative models. There was high agreement between AutoRAPNO scores on predicted segmentations and manually calculated scores based on manual segmentations (Rater 2 ICC = 0.909; Rater 3 ICC = 0.851). Lastly, the performance of AutoRAPNO was superior in repeatability to that of human raters for MRIs with multiple lesions.

CONCLUSIONS

Our automated deep learning pipeline demonstrates potential utility for response assessment in pediatric brain tumors. The tool should be further validated in prospective studies.

Radiologists and Clinical Trials: Part 2: Practical Statistical Methods for Understanding and Monitoring Independent Reader Performance

Though many clinical trials rely on medical image evaluations for primary or key secondary endpoints, the methods to monitor reader performance are all too often mired in the legacy use of adjudication rates. If misused, this simple metric can be misleading and sometimes entirely contradictory. Furthermore, attempts to overcome the limitations of adjudication rates using de novo or ad hoc methods often ignore well-established research conducted over the last half-century and can lead to inaccurate conclusions or variable interpretations. Underperforming readers can be missed, expert readers retrained, or worse, replaced. This paper aims to standardize reader performance evaluations using proven statistical methods. Additionally, these methods will describe how to discriminate between scenarios of concern and normal medical interpretation variability. Statistical methods are provided for inter-reader and intra-reader variability and bias, including the adjudicator's bias. Finally, we have compiled guidelines for calculating correct sample sizes, considerations for intra-reader memory recall, and applying alternative designs for independent readers.

Characterizing Brain Tumor Regions Using Texture Analysis in Magnetic Resonance Imaging

Purpose

To extract texture features from magnetic resonance imaging (MRI) scans of patients with brain tumors and use them to train a classification model for supporting an early diagnosis.

Methods

Two groups of regions (control and tumor) were selected from MRI scans of 40 patients with meningioma or glioma. These regions were analyzed to obtain texture features. Statistical analysis was conducted using SPSS (version 20.0), including the Shapiro-Wilk test and Wilcoxon signed-rank test, which were used to test significant differences in each feature between the tumor and healthy regions. T-distributed stochastic neighbor embedding (t-SNE) was used to visualize the data distribution so as to avoid tumor selection bias. The Gini impurity index in random forests (RFs) was used to select the top five out of all features. Based on the five features, three classification models were built respectively with three machine learning classifiers: RF, support vector machine (SVM), and back propagation (BP) neural network.

Results

Sixteen of the 25 features were significantly different between the tumor and healthy areas. Through the Gini impurity index in RFs, standard deviation, first-order moment, variance, third-order absolute moment, and third-order central moment were selected to build the classification model. The classification model trained using the SVM classifier achieved the best performance, with sensitivity, specificity, and area under the curve of 94.04%, 92.3%, and 0.932, respectively.

Conclusion

Texture analysis with an SVM classifier can help differentiate between brain tumor and healthy areas with high speed and accuracy, which would facilitate its clinical application.

5-aminolevulinic acid-guided surgery for focal pediatric brainstem gliomas: A preliminary study

Background:

There is a growing body of literature supporting the use of 5-aminolevulinic acid (5-ALA) in the pediatric population, however, its use is still considered “off label” in this setting. In this retrospective study, we report our experience using 5-ALA in pediatric patients with focal brainstem gliomas (BSGs).

Methods:

Patients younger than 16 years presenting with a newly diagnosed BSG that was focal in nature were considered suitable for treatment with 5-ALA-assisted surgery. Exclusion criteria included MRI features suggestive of a diffuse intrinsic pontine glioma. A single dose of 5-ALA was administered preoperatively. Intraoperative fluorescence was recorded as “solid,” “vague,” or “none.” The effectiveness of the fluorescence was graded as “helpful” or “unhelpful.”

Results:

Eight patients underwent 5-ALA-assisted surgery. There were four tumors located in the pons, two midbrain tumors, and two cervicomedullary tumors. Histological analysis demonstrated three diffuse astrocytomas, three pilocytic astrocytomas, and two anaplastic astrocytomas. Solid fluorescence was found in three of the eight cases, vague fluorescence was found in two cases, and no fluorescence was found in three cases. Fluorescence was useful in 3 (37%) cases. No patients experienced any complications attributable to the administration of the 5-ALA.

Conclusion:

With a total fluorescence rate of 62.5% but a subjectively assessed “usefulness” rate of only 37.5%, the role of 5-ALA in BSG surgery is limited. Given the toxicological safety, however, of the agent, caution is perhaps needed before dismissing the use of 5-ALA entirely.

Response assessment in diffuse intrinsic pontine glioma: recommendations from the Response Assessment in Pediatric Neuro-Oncology (RAPNO) working group

Optimising the conduct of clinical trials for diffuse intrinsic pontine glioma involves use of consistent, objective disease assessments and standardised response criteria. The Response Assessment in Pediatric Neuro-Oncology working group, consisting of an international panel of paediatric and adult neuro-oncologists, clinicians, radiologists, radiation oncologists, and neurosurgeons, was established to address issues and unique challenges in assessing response in children with CNS tumours. A working group was formed specifically to address response assessment in children and young adults with diffuse intrinsic pontine glioma and to develop a consensus on recommendations for response assessment. Response should be assessed using MRI of brain and spine, neurological examination, and anti-inflammatory or antiangiogenic drugs. Clinical imaging standards are defined. As with previous consensus recommendations, these recommendations will need to be validated in prospective clinical trials.

Tumor Response Assessment in Diffuse Intrinsic Pontine Glioma: Comparison of Semiautomated Volumetric, Semiautomated Linear, and Manual Linear Tumor Measurement Strategies

BACKGROUND AND PURPOSE

2D measurements of diffuse intrinsic pontine gliomas are limited by variability, and volumetric response criteria are poorly defined. Semiautomated 2D measurements may improve consistency; however, the impact on tumor response assessments is unknown. The purpose of this study was to compare manual 2D, semiautomated 2D, and volumetric measurement strategies for diffuse intrinsic pontine gliomas.

MATERIALS AND METHODS

This study evaluated patients with diffuse intrinsic pontine gliomas through a Phase I/II trial (NCT02607124). Clinical 2D cross-product values were derived from manual linear measurements (cross-product = long axis × short axis). By means of dedicated software (mint Lesion), tumor margins were traced and maximum cross-product and tumor volume were automatically derived. Correlation and bias between methods were assessed, and response assessment per measurement strategy was reported.

RESULTS

Ten patients (median age, 7.6 years) underwent 58 MR imaging examinations. Correlation and mean bias (95% limits) of percentage change in tumor size from prior examinations were the following: clinical and semiautomated cross-product, r = 0.36, -1.5% (-59.9%, 56.8%); clinical cross-product and volume, r = 0.61, -2.1% (-52.0%, 47.8%); and semiautomated cross-product and volume, r = 0.79, 0.6% (-39.3%, 38.1%). Stable disease, progressive disease, and partial response rates per measurement strategy were the following: clinical cross-product, 82%, 18%, 0%; semiautomated cross-product, 54%, 42%, 4%; and volume, 50%, 46%, 4%, respectively.

CONCLUSIONS

Manual 2D cross-product measurements may underestimate tumor size and disease progression compared with semiautomated 2D and volumetric measurements.

Brain pathology identification using computer aided diagnostic tool: A systematic review

Computer aided diagnostic (CAD) has become a significant tool in expanding patient quality-of-life by reducing human errors in diagnosis. CAD can expedite decision-making on complex clinical data automatically. Since brain diseases can be fatal, rapid identification of brain pathology to prolong patient life is an important research topic. Many algorithms have been proposed for efficient brain pathology identification (BPI) over the past decade. Constant refinement of the various image processing algorithms must take place to expand performance of the automatic BPI task. In this paper, a systematic survey of contemporary BPI algorithms using brain magnetic resonance imaging (MRI) is presented. A summarization of recent literature provides investigators with a helpful synopsis of the domain. Furthermore, to enhance the performance of BPI, future research directions are indicated.

Direct Involvement of Cranial Nerve V at Diagnosis in Patients With Diffuse Intrinsic Pontine Glioma: A Potential Magnetic Resonance Predictor of Short-Term Survival

Background: Diffuse intrinsic pontine glioma (DIPG) has a dismal prognosis. Magnetic resonance imaging (MRI) remains the gold standard for non-invasive DIPG diagnosis. MRI features have been tested as surrogate biomarkers. We investigated the direct involvement of cranial nerve V (CN V) in DIPG at diagnosis and its utility as predictor of poor overall survival. Materials and Methods: We examined MRI scans of 35 consecutive patients with radiological diagnosis of DIPG. Direct involvement of CN V was assessed on the diagnostic scans. Differences in overall survival (OS) and time to progression (TTP) were analyzed for involvement of CN V, sex, age, tumor size, ring enhancement, and treatment regimen. Correlations between involvement of CN V and disease dissemination, magnet strength and slice thickness were analyzed. Statistical analyses included Kaplan-Meier curves, log-rank test and Spearman's Rho. Results: After excluding six long-term survivors, 29 patients were examined (15 M, 14 F). Four patients presented direct involvement of CN V. Histological data were available in 12 patients. Median OS was 11 months (range 3-23 months). Significant differences in OS were found for direct involvement of CN V (median OS: 7 months, 95% CI 1.1-12.9 months for involvement of CN V vs. 13 months, 95% CI 10.2-15.7 for lack of involvement of CN V, respectively, p < 0.049). Significant differences in TTP were found for the two treatment regimens (median TTP: 4 months, 95% CI 2.6-5.3 vs. 7 months, 95% CI 5.9-8.1, respectively, p < 0.027). No significant correlation was found between involvement of CN V and magnet strength or slice thickness (r = -0.201; p = NS). A trend toward positive correlation was found between direct involvement of CN V at diagnosis and dissemination of disease at follow-up (r = 0.347; p < 0.065). Conclusions: In our cohort, direct involvement of CN V correlated with poor prognosis. Based on our data, we suggest that in DIPG direct involvement of CN V should be routinely evaluated on diagnostic scans.

A Review on a Deep Learning Perspective in Brain Cancer Classification

A World Health Organization (WHO) Feb 2018 report has recently shown that mortality rate due to brain or central nervous system (CNS) cancer is the highest in the Asian continent. It is of critical importance that cancer be detected earlier so that many of these lives can be saved. Cancer grading is an important aspect for targeted therapy. As cancer diagnosis is highly invasive, time consuming and expensive, there is an immediate requirement to develop a non-invasive, cost-effective and efficient tools for brain cancer characterization and grade estimation. Brain scans using magnetic resonance imaging (MRI), computed tomography (CT), as well as other imaging modalities, are fast and safer methods for tumor detection. In this paper, we tried to summarize the pathophysiology of brain cancer, imaging modalities of brain cancer and automatic computer assisted methods for brain cancer characterization in a machine and deep learning paradigm. Another objective of this paper is to find the current issues in existing engineering methods and also project a future paradigm. Further, we have highlighted the relationship between brain cancer and other brain disorders like stroke, Alzheimer’s, Parkinson’s, and Wilson’s disease, leukoriaosis, and other neurological disorders in the context of machine learning and the deep learning paradigm.

Nanotechnology Applications for Diffuse Intrinsic Pontine Glioma

Diffuse intrinsic pontine gliomas (DIPGs) are invariably fatal tumors found in the pons of elementary school aged children. These tumors are grade II-IV gliomas, with a median survival of less than 1 year from diagnosis when treated with standard of care (SOC) therapy. Nanotechnology may offer therapeutic options for the treatment of DIPGs. Multiple nanoparticle formulations are currently being investigated for the treatment of DIPGs. Nanoparticles based upon stable elements, polymer nanoparticles, and organic nanoparticles are under development for the treatment of brain tumors, including DIPGs. Targeting of nanoparticles is now possible as delivery techniques that address the difficulty in crossing the blood brain barrier (BBB) are developed. Theranostic nanoparticles, a combination of therapeutics and diagnostic nanoparticles, improve imaging of the cancerous tissue while delivering therapy to the local region. However, additional time and attention should be directed to developing a nanoparticle delivery system for treatment of the uniformly fatal pediatric disease of DIPG.

A novel magnetic resonance imaging segmentation technique for determining diffuse intrinsic pontine glioma tumor volume

OBJECTIVE Accurately determining diffuse intrinsic pontine glioma (DIPG) tumor volume is clinically important. The aims of the current study were to 1) measure DIPG volumes using methods that require different degrees of subjective judgment; and 2) evaluate interobserver agreement of measurements made using these methods. METHODS Eight patients from a Phase I clinical trial testing convection-enhanced delivery (CED) of a therapeutic antibody were included in the study. Pre-CED, post-radiation therapy axial T2-weighted images were analyzed using 2 methods requiring high degrees of subjective judgment (picture archiving and communication system [PACS] polygon and Volume Viewer auto-contour methods) and 1 method requiring a low degree of subjective judgment (k-means clustering segmentation) to determine tumor volumes. Lin's concordance correlation coefficients (CCCs) were calculated to assess interobserver agreement. RESULTS The CCCs of measurements made by 2 observers with the PACS polygon and the Volume Viewer auto-contour methods were 0.9465 (lower 1-sided 95% confidence limit 0.8472) and 0.7514 (lower 1-sided 95% confidence limit 0.3143), respectively. Both were considered poor agreement. The CCC of measurements made using k-means clustering segmentation was 0.9938 (lower 1-sided 95% confidence limit 0.9772), which was considered substantial strength of agreement. CONCLUSIONS The poor interobserver agreement of PACS polygon and Volume Viewer auto-contour methods highlighted the difficulty in consistently measuring DIPG tumor volumes using methods requiring high degrees of subjective judgment. k-means clustering segmentation, which requires a low degree of subjective judgment, showed better interobserver agreement and produced tumor volumes with delineated borders.

Prognostic value and kinetics of circulating endothelial cells in patients with recurrent glioblastoma randomised to bevacizumab plus lomustine, bevacizumab single agent or lomustine single agent. A report from the Dutch Neuro-Oncology Group BELOB trial

BACKGROUND

Angiogenesis is crucial for glioblastoma growth, and anti-vascular endothelial growth factor agents are widely used in recurrent glioblastoma patients. The number of circulating endothelial cells (CECs) is a surrogate marker for endothelial damage. We assessed their kinetics and explored their prognostic value in patients with recurrent glioblastoma.

METHODS

In this side study of the BELOB trial, 141 patients with recurrent glioblastoma were randomised to receive single-agent bevacizumab or lomustine, or bevacizumab plus lomustine. Before treatment, after 4 weeks and after 6 weeks of treatment, CECs were enumerated.

RESULTS

The number of CECs increased during treatment with bevacizumab plus lomustine, but not during treatment in the single-agent arms. In patients treated with lomustine single agent, higher absolute CEC numbers after 4 weeks (log₁₀CEC hazard ratio (HR) 0.41, 95% CI 0.18-0.91) and 6 weeks (log₁₀CEC HR 0.16, 95% CI 0.05-0.56) of treatment were associated with improved overall survival (OS). Absolute CEC numbers in patients receiving bevacizumab plus lomustine or bevacizumab single agent were not associated with OS.

CONCLUSION

CEC numbers increased during treatment with bevacizumab plus lomustine but not during treatment with either agent alone, suggesting that this combination induced the greatest vascular damage. Although the absolute number of CECs was not associated with OS in patients treated with bevacizumab either alone or in combination, they could serve as a marker in glioblastoma patients receiving lomustine single agent.

Diffuse intrinsic pontine glioma: is MRI surveillance improved by region of interest volumetry?

BACKGROUND

Paediatric diffuse intrinsic pontine glioma (DIPG) is noteworthy for its fibrillary infiltration through neuroparenchyma and its resultant irregular shape. Conventional volumetry methods aim to approximate such irregular tumours to a regular ellipse, which could be less accurate when assessing treatment response on surveillance MRI. Region-of-interest (ROI) volumetry methods, using manually traced tumour profiles on contiguous imaging slices and subsequent computer-aided calculations, may prove more reliable.

OBJECTIVE

To evaluate whether the reliability of MRI surveillance of DIPGs can be improved by the use of ROI-based volumetry.

MATERIALS AND METHODS

We investigated the use of ROI- and ellipsoid-based methods of volumetry for paediatric DIPGs in a retrospective review of 22 MRI examinations. We assessed the inter- and intraobserver variability of the two methods when performed by four observers.

RESULTS

ROI- and ellipsoid-based methods strongly correlated for all four observers. The ROI-based volumes showed slightly better agreement both between and within observers than the ellipsoid-based volumes (inter-[intra-]observer agreement 89.8% [92.3%] and 83.1% [88.2%], respectively). Bland-Altman plots show tighter limits of agreement for the ROI-based method.

CONCLUSION

Both methods are reproducible and transferrable among observers. ROI-based volumetry appears to perform better with greater intra- and interobserver agreement for complex-shaped DIPG.

Diffusion tensor histogram analysis of pediatric diffuse intrinsic pontine glioma

Purpose. To evaluate tumor structure in children with diffuse intrinsic pontine glioma (DIPG) using histogram analyses of mean diffusivity (MD), determine potential treatment and corticosteroid-related effects on MD, and monitor changes in MD distributions over time. Materials and Methods. DTI was performed on a 1.5T GE scanner. Regions of interest included the entire FLAIR-defined tumor. MD data were used to calculate histograms. Patterns in MD distributions were evaluated and fitted using a two-normal mixture model. Treatment-related effects were evaluated using the R² statistic for linear mixed models and Cox proportional hazards models. Results. 12 patients with DIPG underwent one or more DTI exams. MD histogram distributions varied among patients. Over time, histogram peaks became shorter and broader (P = 0.0443). Two-normal mixture fitting revealed large lower curve proportions that were not associated with treatment response or outcome. Corticosteroid use affected MD histograms and was strongly associated with larger, sharper peaks (R² = 0.51, P = 0.0028). Conclusions. MD histograms of pediatric DIPG show significant interpatient and intratumoral differences and quantifiable changes in tumor structure over time. Corticosteroids greatly affected MD and must be considered a confounding factor when interpreting MD results in the context of treatment response.

Measurements of the pons as a biomarker of progression for pediatric DIPG

Treatment of pediatric diffuse intrinsic pontine glioma (DIPG) remains challenging, and reliable biomarkers of response are lacking. Radiographic response is a primary endpoint in many investigational studies of brain tumors, but there is no standard method of tumor measurement for DIPG, significant inter-observer variability exists given the invasive nature of these tumors, and tumor measurements are not predictive of outcome. Because DIPGs involve a significant portion of the pons, we evaluated the reliability and prognostic value of one-dimensional (1D) and two-dimensional (2D) pons measurements using anatomical landmarks rather than tumor boundaries. Patients with DIPG (n = 75) were evaluated longitudinally at our institution using MRI. Four readers independently performed 1D and 2D measurements of the pons using FLAIR images. Agreement and inter-reader variability were evaluated using differences among the six reader pairs and the coefficient of variation (CV). Prognostic value of pons measurements was calculated using Cox proportional hazards models, where relative hazard (RH) represents risk of death. Readers evaluated 384 exams. Agreement of readers' 1D and 2D measurements was strong (median difference between reader pairs 3.1 and 5.4%, respectively), with low inter-reader variability (median CV = 3.1% and median CV = 4.8%, respectively). Increases in 1D and 2D pons measurements over time indicated poorer prognosis (RH = 2.29, p = 0.0025 and RH = 1.13, p = 0.0016, respectively), with shorter overall survival. Pons measurements had low inter-reader variability compared to previously reported tumor measurement techniques and correlated with outcome in children with DIPG. Measurements of the pons (as opposed to direct measurements of tumor) are a viable in vivo biomarker for DIPG.

Challenges with defining response to antitumor agents in pediatric neuro-oncology: a report from the response assessment in pediatric neuro-oncology (RAPNO) working group

Criteria for new drug approval include demonstration of efficacy. In neuro-oncology, this is determined radiographically utilizing tumor measurements on MRI scans. Limitations of this method have been identified where drug activity is not reflected in decreased tumor size. The RANO (Response Assessment in Neuro-Oncology) working group was established to address limitations in defining endpoints for clinical trials in adult neuro-oncology and to develop standardized response criteria. RAPNO was subsequently established to address unique issues in pediatric neuro-oncology. The aim of this paper is to delineate response criteria issues in pediatric clinical trials as a basis for subsequent recommendations.

Single- and multivoxel proton spectroscopy in pediatric patients with diffuse intrinsic pontine glioma

PURPOSE

To determine the feasibility of two magnetic resonance spectroscopy (MRS) techniques for treating pediatric patients with diffuse intrinsic pontine gliomas (DIPGs) and to evaluate the relationship of metabolic profiles determined by each technique. Utility of each technique for improving patient management is also discussed.

METHODS AND MATERIALS

Children with DIPG (n = 36) were evaluated using single-voxel spectroscopy (SVS) and magnetic resonance spectroscopic imaging (MRSI) during the same imaging session. Patients were followed longitudinally (n = 150 total studies). Technical feasibility was defined by sufficient water and lipid suppression for detection of metabolites. Correlation of metabolic data obtained by SVS and MRSI was determined using the Spearman rank method. Metabolite ratios, including choline:N-acetyl-aspartate (Cho:NAA) and Cho:creatine (Cho:Cr), were obtained from SVS and MRSI.

RESULTS

SVS and MRSI acquisitions were feasible in >90% of studies. Maximum Cho:NAA and Cho:Cr from MRSI analysis were strongly associated with Cho:NAA and Cho:Cr obtained by SVS (r = 0.67 and 0.76, respectively). MRSI Cho:NAA values were more heterogeneous than Cho:Cr values within the same lesion, and a strong linear relationship between the range and maximum Cho:NAA values was observed.

CONCLUSIONS

SVS and MRSI acquisitions were feasible, with a strong correlation in metabolic data. Both techniques may improve diagnostic evaluation and management of DIPG. SVS is recommended for global assessment of tumor metabolism before and after therapy. MRSI showed heterogeneous patterns of metabolic activity within these tumors and is recommended for planning and monitoring targeted therapies and evaluating nearby tissue for tumor invasion.

Assessment of intra-observer variability in measurement of high-grade brain tumors

A 25% increase in bidimensional products (BPs) of tumor diameter has been used as a criterion for brain tumor progression. We studied intra-observer variability in measurements of BPs. Ten patients with contrast-enhancing glioblastoma multiforme underwent baseline and follow-up MR imaging. Seven observers measured BPs in various planes. Differences in BPs between scans were expressed as a percentage of baseline. This calculation was performed for both readings of the baseline and follow-up scans. Differences between change from baseline to follow-up on each reading (termed D values) were calculated for each reader (total of 196 D values). Median D value in each plane was calculated for each reader. Range of D values was 12.36-33.64% in axial plane (average 10.63%), 12.18-38.62% in coronal plane (average 26.84%) and 15.12-35.48% in sagittal plane (average 26.11%). Across all planes, 88 (45%) D values were >25%. When all imaging planes for any single observation were combined, in 76% of cases, at least one D value of >25% was seen. Based on the high degree of intra-observer variability, tumor measurements producing an increase in BP of >25% can routinely be obtained solely by chance.

Role of early postradiation magnetic resonance imaging scans in children with diffuse intrinsic pontine glioma

PURPOSE

To determine optimal timing of assessing postradiation radiographic response on magnetic resonance imaging (MRI) scans in pediatric patients with diffuse intrinsic pontine glioma (DIPG).

METHODS AND MATERIALS

Patients were treated on a prospective study at the National Cancer Institute (Protocol #06-C-0219) evaluating the effects of radiotherapy (RT). Standard RT was administered in standard fractionation over 6 weeks. Postradiation MRI scans were performed at 2 and 6-8 weeks.

RESULTS

Eleven patients with DIPG were evaluated. Median age was 6 years (range, 4-13 years). Patients were treated with external-beam RT to 55.8 Gy (n = 10) or 54 Gy (n = 1), with a gross tumor volume to planning target volume expansion of 1.8-2.0 cm. All patients received prescribed dose and underwent posttreatment MRI scans at 2 and 6-8 weeks. Pretreatment imaging revealed compression of fourth ventricle (n = 11); basilar artery encasement (n = 9); tumor extension outside the pons (n = 11); and tumor hemorrhage (n = 2). At the 2-week scan, basilar artery encasement improved in 7 of 9 patients, and extent of tumor was reduced in 5 of 11 patients. Fourth ventricle compression improved in 6 of 11 patients but worsened in 3 of 11 patients. Presumed necrosis was observed in 5 of 11 patients at 2 weeks and in 1 additional patient at 6-8 weeks. There was no significant difference in mean anteroposterior and transverse diameters of tumor between the 2- and 6-8-week time points. Six of 11 patients had increasing ventricular size, with no evidence of obstruction.

CONCLUSIONS

There is no significant difference in tumor size of DIPG patients who have received standard RT when measured at 2 weeks vs. 6-8 weeks after RT. The majority of patients had the largest change in tumor size at the 2-week post-RT scan, with evolving changes documented on the 6-8-week scan. Six of 11 patients had progressive ventriculomegaly without obstruction, suggestive of communicating hydrocephalus. To the best of our knowledge, this is the first documentation of this phenomenon in this cohort of patients.

Proton magnetic resonance spectroscopy predicts survival in children with diffuse intrinsic pontine glioma

Patients with diffuse intrinsic pontine glioma (DIPG) face a grim prognosis with limited treatment options. Many patients will enroll on investigational trials though the role of chemotherapy or immunotherapy is unclear. Radiographic changes on conventional MRI are used to evaluate tumor response and progression, but are not predictive of outcome in these patients. More sensitive measures of tumor biology are needed to improve patient management. We evaluated changes in magnetic resonance spectroscopy (MRS) biomarkers in patients with DIPG. Thirty-eight patients were enrolled prospectively on an IRB-approved protocol, which included standard MRI, single voxel spectroscopy (SVS) and multi-slice multi-voxel spectroscopy (MRSI). Scans were performed at multiple time points during each patient's clinical course, with a total of 142 scans. The prognostic values of Choline:N-acetylaspartate (Cho:NAA), Cho:Creatine (Cho:Cr) and the presence of lactate and lipids (+Lac/Lip) were evaluated. Cho:NAA and variance in Cho:NAA values among different voxels within a tumor were each predictive of shorter survival. This prospective study shows that MRS can be used to identify high-risk patients and monitor changes in tumor metabolism, which may reflect changes in tumor behavior.

Predicting outcome of children with diffuse intrinsic pontine gliomas using multiparametric imaging

Noninvasive evaluation using MRI is the primary means to routinely assess children with diffuse intrinsic pontine gliomas (DIPGs). However, no standard MR sequence has correlated with outcome in these patients. In this study, patients with DIPGs were assessed to determine the combined prognostic value via dynamic susceptibility contrast (DSC) MRI, single-voxel spectroscopy (SVS), multivoxel MR spectroscopy (MRS), and T1-weighted post-gadolinium imaging. Eligible patients had clinical and radiographic findings consistent with a DIPG. Imaging studies were acquired on a 1.5T MRI at various time points during each patient's course. Data were evaluated using a Cox proportional hazard model, a time-dependent covariant Cox model, a Wald test, and a Kaplan-Meier analysis. Ninety-eight studies were performed on 34 patients of median age 5.5 years. Median survival from diagnosis was 468 days. At baseline imaging only, increased ratio of choline to n-acetylaspartate (Cho:NAA) on SVS and increased perfusion on DSC-MRI each predicted shorter survival (relative risk [RR] = 1.48, P = .015 and RR = 4.91, P = .0012, respectively). When analyzing all subsequent time points, increased maximum Cho:NAA on MRS (RR = 1.45, P = .042), increased Cho:NAA on SVS (RR = 1.69, P = .003), increased perfusion (RR = 4.68, P = .0016), and the presence of enhancement (RR = 5.69, P = .022) each predicted shorter survival. Kaplan-Meier analysis showed shorter survival associated with increased perfusion at baseline (P = .0004). Increased perfusion at any time point predicts a significantly shorter survival in children with DIPG. In addition, enhancement, increased Cho:NAA on SVS, and increased maximum Cho:NAA on chemical shift imaging are predictive of shorter survival over time. Routine baseline and subsequent imaging for children with DIPG should, at minimum, incorporate DSC-MRI and SVS.

Interpretation of magnetic resonance images in diffuse intrinsic pontine glioma: a survey of pediatric neurosurgeons

OBJECT

The current management paradigm for children with newly diagnosed diffuse intrinsic pontine glioma (DIPG) is to establish a diagnosis and begin therapy based on MR imaging findings correlated with an appropriate clinical presentation, and without a tissue diagnosis. This strategy assumes that pediatric neurosurgeons and neurooncologists uniformly interpret MR imaging findings in this population. This study sought to examine the consistency of North American pediatric neurosurgeons in assessing MR images in this patient population, and in their surgical plans based on the interpretation of those images.

METHODS

The authors created an online survey and invited all members of the American Society of Pediatric Neurosurgeons and the American Association of Neurological Surgeons/Congress of Neurological Surgeons Joint Section for Pediatric Neurosurgery to participate. The survey consisted of 58 questions, and 48 pertained to representative MR images from 16 children who presented to The Children's Hospital of Colorado with diffuse pontine tumors. Based on the imaging presented and a standard clinical scenario, the respondent was asked if she or he believed a lesion to be "typical" or "atypical," whether she or he would biopsy the lesion, and what surgical approach would be chosen. The remaining 10 questions pertained to respondent demographics and his or her practice regarding tissue preservation and interest in participating in a multicenter trial that included tumor biopsy in selected cases. Rates at which each lesion was considered to be typical or atypical and rates of recommended biopsy were calculated.

RESULTS

Surveys were received by 269 individuals. Eighty-six responses were received (32.0%). No tumor was judged to be either typical or atypical by all respondents. Individual surgeons varied widely regarding how many of the tumors were judged as typical or warranted a biopsy. The percentage of respondents who disagreed with the majority opinion regarding whether a tumor qualified as typical ranged from 2.3% to 48.8%, with a median of 28.6%. More than 75% agreement regarding whether a tumor was typical or atypical was found in 7 (43.8%) of 16 cases. The κ statistic regarding typicality was 0.297 ± 0.0004 (mean ± SEM), implying only fair agreement. For every tumor, at least 1 respondent who believed the lesion to be typical in appearance would nevertheless biopsy the lesion. Of those respondents who considered a lesion to be typical, a median of 5.1% (range 1.2%-66.7%) would choose to biopsy this lesion. Of those respondents who considered a lesion atypical, a median of 18.3% (range 3.7%-100%) would choose not to biopsy the lesion. Of 85 responses to the question, "Would you be willing to biopsy a typical diffuse pontine glioma as a part of a multicenter trial?," 59 (69.4%) of 85 respondents answered that they would.

CONCLUSIONS

Although making a diagnosis based on radiographic evidence alone represents a well-established management paradigm for children with suspected DIPG, this study demonstrates considerable inconsistency on the part of pediatric neurosurgeons in the application of this strategy to individual patients. As such, the practice of diagnosing DIPG based on imaging characteristics and clinical history alone does not reach the appropriate threshold to be considered a standard of care.

Phase II trial of tipifarnib and radiation in children with newly diagnosed diffuse intrinsic pontine gliomas

We performed a phase II study to assess the efficacy and toxicity of tipifarnib, a farnesyltransferase inhibitor, administered with radiation therapy (RT) in children with newly diagnosed diffuse intrinsic pontine gliomas. Children 3-21 years old with pontine gliomas (BSGs) were treated with concurrent tipifarnib and RT, followed by adjuvant tipifarnib. Tipifarnib was taken orally twice daily (125 mg/m(2)/dose) during RT; after RT, it was taken at 200 mg/m(2) twice daily for 21 days, in 28-day cycles. Initial and follow-up neuroimaging was centrally reviewed. Forty eligible patients (median age, 5.5 years; range, 3.3-16.5 years) had a median progression-free survival of 6.8 months (range, 0.2-18.6 months) and median overall survival of 8.3 months (range, 0.2-18.6 months). Kaplan-Meier estimates (± standard error) of 1-year progression-free and overall survival were 12.9% ±4.9% and 34.3% ±7.4%, respectively. A single patient remained on tipifarnib without progression at the completion of the study, two years after initiation of treatment. Seven patients were without disease progression for at least six months, three of whom remained controlled for more than a year. The most frequent toxicity was grade 3 lymphopenia. We documented a single instance of "pseudoprogression" by neuroimaging review. We found no discordance among 3 approaches to defining disease progression: as interpreted by treating institutions (based on clinical status and/or imaging) and by central review (using bi-dimensional tumor "area" versus volumetric measurements). For children with diffuse BSGs, tipifarnib administered with irradiation offered no clinical advantage over historical controls. Biopsies and molecular analyses of pediatric BSGs are vital for identification of new agents and for rational use of targeted agents.

Prospective neuraxis MRI surveillance reveals a high risk of leptomeningeal dissemination in diffuse intrinsic pontine glioma

Prognosis of diffuse intrinsic pontine gliomas (DIPGs) remains poor. Failure has been predominantly local, with leptomeningeal dissemination (LD) occurring in 4-33% of patients in pre-MRI era series. Routine craniospinal imaging after initial treatment may reveal other relapse patterns relapse. Sixteen consecutive pediatric patients with DIPG treated between 2006 and 2009 were retrospectively reviewed. Treatment regimens, recurrence patterns, survival, and pathologic diagnosis were recorded. Fourteen patients received involved-field radiotherapy to 54 Gy, and two patients received craniospinal irradiation for LD at presentation. Neuraxis MRI was performed at diagnosis and at 4 month intervals following radiotherapy. Fifteen patients have had progression of disease (median progression-free survival 5.0 ± 1.2 months), and 13 patients have died (median survival 9.0 ± 1.4 months). Local failure occurred in 12 patients (75%). LD occurred in nine patients (56%). LD was present at diagnosis in three patients, after initial staging and treatment in six patients, and during autopsy in two patients. Median overall survival was 12.0 ± 3.3 months without LD and 8.0 ± 2.1 months with LD (P = 0.059, log rank test). Median progression-free survival was 9.5 ± 3.9 months without LD and 3.0 ± 2.1 months with LD (P = 0.012, log rank test). The high incidence of LD probably reflects liberal use of spine MRI surveillance. All patients should undergo routine craniospinal imaging at diagnosis and follow-up. Central nervous system prophylaxis should be considered in future clinical trials.

Update and developments in the treatment of glioblastoma multiforme - focus on bevacizumab

Glioblastoma is the most common primary brain tumor with a relatively poor prognosis. This article reviews the current standard therapy and discusses new developments in treatment of this disease. Surgical resection followed by radiation and chemotherapy has proven to be the most effective initial therapy. Recent advancement in molecular targeted therapies has led to the Food and Drug Administration (FDA) approval of bevacizumab in the setting of recurrent glioblastoma. The molecular pathways of glioblastoma growth are highlighted in this review. While numerous molecular targets are currently being intensely investigated, vascular endothelial growth factor (VEGF) receptor targeted therapy has been the only one to have shown clinical effect. The role of bevacizumab in this context provides a dynamic breakthrough in cancer therapy. Clinical trials have demonstrated significantly increased overall survival and six month progression free survival (PFS) in recurrent glioblastoma treated with bevacizumab alone or in combination with irinotecan. The use of this agent has also dramatically changed the imaging characteristics of glioblastoma. The anti-angiogenesis effects of bevacizumab have complicated the criterion for determining tumor growth. This may lead to redefinition of progressive disease based on non-invasive monitoring.

Bidimensional measurements in brain tumors: assessment of interobserver variability

OBJECTIVE

Bidimensional tumor measurements indicating a greater than 25% increase in tumor size are generally accepted as indicating tumor progression. We hypothesized that use of digital images and a homogeneous reader population would have lower interobserver variability than in previous studies.

SUBJECTS AND METHODS

Eight board-certified radiologists measured tumor diameters in three planes in two consecutive MRI examinations of 22 patients with contrast-enhancing high-grade brain tumors. Products of tumor measurements were calculated, and determinations were made about tumor progression (> 25% increase in area). A variance components model was run on diameter products and the ratios of consecutive maximal diameter products. The variance components included patient examination effect, reader effect, and residual effect.

RESULTS

Complete agreement was found among readers in 10 cases (45%), all indicating stable disease. In the other 12 cases, at least one reader considered progressive disease present. The variance components model showed variance due to readers was small, indicating only modest bias among readers. The residual variance component was large (0.038), indicating that repeated measurements on the same image likely are variable even for the same reader. This variability in measurement implies that repeated measurements by the typical reader have an inherent 14% false-positive rate in the diagnosis of progression of tumors that are stable.

CONCLUSION

Our hypothesis was disproved. We found substantial interreader disagreement and indications that the very nature of the measurement method produces a high rate of false-positive readings of stable tumors. These findings should be considered in interpretation of images with this widely accepted criterion for brain tumor progression.

New approaches to pharmacotherapy of tumors of the nervous system during childhood and adolescence

Tumors of the nervous system are among the most common and most chemoresistant neoplasms of childhood and adolescence. Malignant tumors of the brain collectively account for 21% of all cancers and 24% of all cancer-related deaths in this age group. Neuroblastoma, a peripheral nervous system tumor, is the most common extracranial solid tumor of childhood, and 65% of children with this tumor have only a 10 or 15% chance of living 5 years beyond the time of initial diagnosis. Novel pharmacological approaches to nervous system tumors are urgently needed. This review presents the role of and current challenges to pharmacotherapy of malignant tumors of the nervous system during childhood and adolescence and discusses novel approaches aimed at overcoming these challenges.