Interobserver variations in gross tumor volume delineation of brain tumors on computed tomography and impact of magnetic resonance imaging

Improved brain metastases segmentation using generative adversarial network and conditional random field optimization mask R-CNN

BACKGROUND

In radiotherapy, the delineation of the gross tumor volume (GTV) in brain metastases using computed tomography (CT) simulation localization is very important. However, despite the criticality of this process, a pronounced gap exists in the availability of tools tailored for the automatic segmentation of the GTV based on CT simulation localization images.

PURPOSE

This study aims to fill this gap by devising an effective tool specifically for the automatic segmentation of the GTV using CT simulation localization images.

METHODS

A dual-network generative adversarial network (GAN) architecture was developed, wherein the generator focused on refining CT images for more precise delineation, and the discriminator differentiated between real and augmented images. This architecture was coupled with the Mask R-CNN model to achieve meticulous GTV segmentation. An end-to-end training process facilitated the integration between the GAN and Mask R-CNN functionalities. Furthermore, a conditional random field (CRF) was incorporated to refine the initial masks generated by the Mask R-CNN model to ensure optimal segmentation accuracy. The performance was assessed using key metrics, namely, the Dice coefficient (DSC), intersection over union (IoU), accuracy, specificity, and sensitivity.

RESULTS

The GAN+Mask R-CNN+CRF integration method in this study performs well in GTV segmentation. In particular, the model has an overall average DSC of 0.819 ± 0.102 and an IoU of 0.712 ± 0.111 in the internal validation. The overall average DSC in the external validation data is 0.726 ± 0.128 and the IoU is 0.640 ± 0.136. It demonstrates favorable generalization ability.

CONCLUSION

The integration of the GAN, Mask R-CNN, and CRF optimization provides a pioneering tool for the sophisticated segmentation of the GTV in brain metastases using CT simulation localization images. The method proposed in this study can provide a robust automatic segmentation approach for brain metastases in the absence of MRI.

Empirical planning target volume modeling for high precision MRI guided intracranial radiotherapy

Purpose

Magnetic resonance image-guided radiotherapy for intracranial indications is a promising advance; however, uncertainties remain for both target localization after translation-only MR setup and intrafraction motion. This investigation quantified these uncertainties and developed a population-based planning target volume (PTV) model to explore target and organ-at-risk (OAR) volumetric coverage tradeoffs.

Methods

Sixty-six patients, 49 with a primary brain tumor and 17 with a post-surgical resection cavity, treated on a 1.5T-based MR-linac across 1329 fractions were included. At each fraction, patients were setup by translation-only fusion of the online T1 MRI to the planning image. Each fusion was independently repeated offline accounting for rotations. The six degree-of-freedom difference between fusions was applied to transform the planning CTV at each fraction (CTV_fx). A PTV model parameterized by volumetric CTV_fx coverage, proportion of fractions, and proportion of patients was developed. Intrafraction motion was quantified in a 412 fraction subset as the fusion difference between post- and pre-irradiation T1 MRIs.

Results

For the left-right/anterior-posterior/superior-inferior axes, mean ± SD of the rotational fusion differences were 0.1 ± 0.8/0.1 ± 0.8/-0.2 ± 0.9°. Covering 98 % of the CTV_fx in 95 % of fractions in 95 % of patients required a 3 mm PTV margin. Margin reduction decreased PTV-OAR overlap; for example, the proportion of optic chiasm overlapped by the PTV was reduced up to 23.5 % by margin reduction from 4 mm to 3 mm.

Conclusions

An evidence-based PTV model was developed for brain cancer patients treated on the MR-linac. Informed by this model, we have clinically adopted a 3 mm PTV margin for conventionally fractionated intracranial patients.

Reproducibility for Hepatocellular Carcinoma CT Radiomic Features: Influence of Delineation Variability Based on 3D-CT, 4D-CT and Multiple-Parameter MR Images

Purpose

Accurate lesion segmentation is a prerequisite for radiomic feature extraction. It helps to reduce the features variability so as to improve the reporting quality of radiomics study. In this research, we aimed to conduct a radiomic feature reproducibility test of inter-/intra-observer delineation variability in hepatocellular carcinoma using 3D-CT images, 4D-CT images and multiple-parameter MR images.

Materials and Methods

For this retrospective study, 19 HCC patients undergoing 3D-CT, 4D-CT and multiple-parameter MR scans were included in this study. The gross tumor volume (GTV) was independently delineated twice by two observers based on contrast-enhanced computed tomography (CECT), maximum intensity projection (MIP), LAVA-Flex, T2W FRFSE and DWI-EPI images. We also delineated the peritumoral region, which was defined as 0 to 5 mm radius surrounding the GTV. 107 radiomic features were automatically extracted from CECT images using 3D-Slicer software. Quartile coefficient of dispersion (QCD) and intraclass correlation coefficient (ICC) were applied to assess the variability of each radiomic feature. QCD<10% and ICC≥0.75 were considered small variations and excellent reliability. Finally, the principal component analysis (PCA) was used to test the feasibility of dimensionality reduction.

Results

For tumor tissues, the numbers of radiomic features with QCD<10% indicated no obvious inter-/intra-observer differences or discrepancies in 3D-CT, 4D-CT and multiple-parameter MR delineation. However, the number of radiomic features (mean 89) with ICC≥0.75 was the highest in the multiple-parameter MR group, followed by the 3DCT group (mean 77) and the MIP group (mean 73). The peritumor tissues also showed similar results. A total of 15 and 7 radiomic features presented excellent reproducibility and small variation in tumor and peritumoral tissues, respectively. Two robust features showed excellent reproducibility and small variation in tumor and peritumoral tissues. In addition, the values of the two features both represented statistically significant differences among tumor and peritumoral tissues (P<0.05). The PCA results indicated that the first seven principal components could preserve at least 90% of the variance of the original set of features.

Conclusion

Delineation on multiple-parameter MR images could help to improve the reproducibility of the HCC CT radiomic features and weaken the inter-/intra-observer influence.

Improving Automated Glioma Segmentation in Routine Clinical Use Through Artificial Intelligence-Based Replacement of Missing Sequences With Synthetic Magnetic Resonance Imaging Scans

OBJECTIVES

Although automated glioma segmentation holds promise for objective assessment of tumor biology and response, its routine clinical use is impaired by missing sequences, for example, due to motion artifacts. The aim of our study was to develop and validate a generative adversarial network for synthesizing missing sequences to allow for a robust automated segmentation.

MATERIALS AND METHODS

Our model was trained on data from The Cancer Imaging Archive (n = 238 WHO II-IV gliomas) to synthesize either missing FLAIR, T2-weighted, T1-weighted (T1w), or contrast-enhanced T1w images from available sequences, using a novel tumor-targeting loss to improve synthesis of tumor areas. We validated performance in a test set from both the REMBRANDT repository and our local institution (n = 68 WHO II-IV gliomas), using qualitative image appearance metrics, but also segmentation performance with state-of-the-art segmentation models. Segmentation of synthetic images was compared with 2 commonly used strategies for handling missing input data, entering a blank mask or copying an existing sequence.

RESULTS

Across tumor areas and missing sequences, synthetic images generally outperformed both conventional approaches, in particular when FLAIR was missing. Here, for edema and whole tumor segmentation, we improved the Dice score, a common metric for evaluation of segmentation performance, by 12% and 11%, respectively, over the best conventional method. No method was able to reliably replace missing contrast-enhanced T1w images.

DISCUSSION

Replacing missing nonenhanced magnetic resonance sequences via synthetic images significantly improves segmentation quality over most conventional approaches. This model is freely available and facilitates more widespread use of automated segmentation in routine clinical use, where missing sequences are common.

Impact of Interobserver Variability in Manual Segmentation of Non-Small Cell Lung Cancer (NSCLC) Applying Low-Rank Radiomic Representation on Computed Tomography

Simple Summary

Discovery of predictive and prognostic radiomic features in cancer is currently of great interest to the radiologic and oncologic community. Tumor phenotypic and prognostic information can be obtained by extracting features on tumor segmentations, and it is typically imaging analysts, physician trainees, and attending physicians who provide these labeled datasets for analysis. The potential impact of level and type of specialty training on interobserver variability in manual segmentation of NSCLC was examined. Although there was some variability in segmentation between readers, the subsequently extracted radiomic features were overall well correlated. High fidelity radiomic feature extraction relies on accurate feature extraction from imaging that produce robust prognostic and predictive radiomic NSCLC biomarkers. This study concludes that this goal can be obtained using segmenters of different levels of training and clinical experience.

Abstract

This study tackles interobserver variability with respect to specialty training in manual segmentation of non-small cell lung cancer (NSCLC). Four readers included for segmentation are: a data scientist (BY), a medical student (LS), a radiology trainee (MH), and a specialty-trained radiologist (SK) for a total of 293 patients from two publicly available databases. Sørensen–Dice (SD) coefficients and low rank Pearson correlation coefficients (CC) of 429 radiomics were calculated to assess interobserver variability. Cox proportional hazard (CPH) models and Kaplan-Meier (KM) curves of overall survival (OS) prediction for each dataset were also generated. SD and CC for segmentations demonstrated high similarities, yielding, SD: 0.79 and CC: 0.92 (BY-SK), SD: 0.81 and CC: 0.83 (LS-SK), and SD: 0.84 and CC: 0.91 (MH-SK) in average for both databases, respectively. OS through the maximal CPH model for the two datasets yielded c-statistics of 0.7 (95% CI) and 0.69 (95% CI), while adding radiomic and clinical variables (sex, stage/morphological status, and histology) together. KM curves also showed significant discrimination between high- and low-risk patients (p-value < 0.005). This supports that readers’ level of training and clinical experience may not significantly influence the ability to extract accurate radiomic features for NSCLC on CT. This potentially allows flexibility in the training required to produce robust prognostic imaging biomarkers for potential clinical translation.

A national study on the inter-observer variability in the delineation of organs at risk in the brain

BACKGROUND

The Danish Neuro Oncology Group (DNOG) has established national consensus guidelines for the delineation of organs at risk (OAR) structures based on published literature. This study was conducted to finalise these guidelines and evaluate the inter-observer variability of the delineated OAR structures by expert observers.

MATERIAL AND METHODS

The DNOG delineation guidelines were formed by participants from all Danish centres that treat brain tumours with radiotherapy. In a two-day workshop, guidelines were discussed and finalised based on a pilot study. Following this, the ten participants contoured the following OARs on T1-weighted gadolinium enhanced MRI from 13 patients with brain tumours: optic tracts, optic nerves, chiasm, spinal cord, brainstem, pituitary gland and hippocampus. The metrics used for comparison were the Dice similarity coefficient (Dice), mean surface distance (MSD) and others.

RESULTS

A total of 968 contours were delineated across the 13 patients. On average eight (range six to nine) individual contour sets were made per patient. Good agreement was found across all structures with a median MSD below 1 mm for most structures, with the chiasm performing the best with a median MSD of 0.45 mm. The Dice was as expected highly volume dependent, the brainstem (the largest structure) had the highest Dice value with a median of 0.89 whereas smaller volumes such as the chiasm had a Dice of 0.71.

CONCLUSION

Except for the caudal definition of the spinal cord, the variances observed in the contours of OARs in the brain were generally low and consistent. Surface mapping revealed sub-regions of higher variance for some organs. The data set is being prepared as a validation data set for auto-segmentation algorithms for use within the Danish Comprehensive Cancer Centre - Radiotherapy and potential collaborators.

A Fully Automatic Artificial Intelligence System Able to Detect and Characterize Prostate Cancer Using Multiparametric MRI: Multicenter and Multi-Scanner Validation

In the last years, the widespread use of the prostate-specific antigen (PSA) blood examination to triage patients who will enter the diagnostic/therapeutic path for prostate cancer (PCa) has almost halved PCa-specific mortality. As a counterpart, millions of men with clinically insignificant cancer not destined to cause death are treated, with no beneficial impact on overall survival. Therefore, there is a compelling need to develop tools that can help in stratifying patients according to their risk, to support physicians in the selection of the most appropriate treatment option for each individual patient. The aim of this study was to develop and validate on multivendor data a fully automated computer-aided diagnosis (CAD) system to detect and characterize PCas according to their aggressiveness. We propose a CAD system based on artificial intelligence algorithms that a) registers all images coming from different MRI sequences, b) provides candidates suspicious to be tumor, and c) provides an aggressiveness score of each candidate based on the results of a support vector machine classifier fed with radiomics features. The dataset was composed of 131 patients (149 tumors) from two different institutions that were divided in a training set, a narrow validation set, and an external validation set. The algorithm reached an area under the receiver operating characteristic (ROC) curve in distinguishing between low and high aggressive tumors of 0.96 and 0.81 on the training and validation sets, respectively. Moreover, when the output of the classifier was divided into three classes of risk, i.e., indolent, indeterminate, and aggressive, our method did not classify any aggressive tumor as indolent, meaning that, according to our score, all aggressive tumors would undergo treatment or further investigations. Our CAD performance is superior to that of previous studies and overcomes some of their limitations, such as the need to perform manual segmentation of the tumor or the fact that analysis is limited to single-center datasets. The results of this study are promising and could pave the way to a prediction tool for personalized decision making in patients harboring PCa.

A rapid review of influential factors and appraised solutions on organ delineation uncertainties reduction in radiotherapy

Background and purpose.Accurate volume delineation plays an essential role in radiotherapy. Contouring is a potential source of uncertainties in radiotherapy treatment planning that could affect treatment outcomes. Therefore, reducing the degree of contouring uncertainties is crucial. The role of utilized imaging modality in the organ delineation uncertainties has been investigated. This systematic review explores the influential factors on inter-and intra-observer uncertainties of target volume and organs at risk (OARs) delineation focusing on the used imaging modality for these uncertainties reduction and the reported subsequent histopathology and follow-up assessment.Methods and materials.An inclusive search strategy has been conducted to query the available online databases (Scopus, Google Scholar, PubMed, and Medline). 'Organ at risk', 'target', 'delineation', 'uncertainties', 'radiotherapy' and their relevant terms were utilized using every database searching syntax. Final article extraction was performed following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline. Included studies were limited to the ones published in English between 1995 and 2020 and that just deal with computed tomography (CT) and magnetic resonance imaging (MRI) modalities.Results.A total of 923 studies were screened and 78 were included of which 31 related to the prostate 20 to the breast, 18 to the head and neck, and 9 to the brain tumor site. 98% of the extracted studies performed volumetric analysis. Only 24% of the publications reported the dose deviations resulted from variation in volume delineation Also, heterogeneity in studied populations and reported geometric and volumetric parameters were identified such that quantitative synthesis was not appropriate.Conclusion.This review highlightes the inter- and intra-observer variations that could lead to contouring uncertainties and impede tumor control in radiotherapy. For improving volume delineation and reducing inter-observer variability, the implementation of well structured training programs, homogeneity in following consensus and guidelines, reliable ground truth selection, and proper imaging modality utilization could be clinically beneficial.

Generative Adversarial Networks to Synthesize Missing T1 and FLAIR MRI Sequences for Use in a Multisequence Brain Tumor Segmentation Model

Background Missing MRI sequences represent an obstacle in the development and use of deep learning (DL) models that require multiple inputs. Purpose To determine if synthesizing brain MRI scans using generative adversarial networks (GANs) allows for the use of a DL model for brain lesion segmentation that requires T1-weighted images, postcontrast T1-weighted images, fluid-attenuated inversion recovery (FLAIR) images, and T2-weighted images. Materials and Methods In this retrospective study, brain MRI scans obtained between 2011 and 2019 were collected, and scenarios were simulated in which the T1-weighted images and FLAIR images were missing. Two GANs were trained, validated, and tested using 210 glioblastomas (GBMs) (Multimodal Brain Tumor Image Segmentation Benchmark [BRATS] 2017) to generate T1-weighted images from postcontrast T1-weighted images and FLAIR images from T2-weighted images. The quality of the generated images was evaluated with mean squared error (MSE) and the structural similarity index (SSI). The segmentations obtained with the generated scans were compared with those obtained with the original MRI scans using the dice similarity coefficient (DSC). The GANs were validated on sets of GBMs and central nervous system lymphomas from the authors' institution to assess their generalizability. Statistical analysis was performed using the Mann-Whitney, Friedman, and Dunn tests. Results Two hundred ten GBMs from the BRATS data set and 46 GBMs (mean patient age, 58 years ± 11 [standard deviation]; 27 men [59%] and 19 women [41%]) and 21 central nervous system lymphomas (mean patient age, 67 years ± 13; 12 men [57%] and nine women [43%]) from the authors' institution were evaluated. The median MSE for the generated T1-weighted images ranged from 0.005 to 0.013, and the median MSE for the generated FLAIR images ranged from 0.004 to 0.103. The median SSI ranged from 0.82 to 0.92 for the generated T1-weighted images and from 0.76 to 0.92 for the generated FLAIR images. The median DSCs for the segmentation of the whole lesion, the FLAIR hyperintensities, and the contrast-enhanced areas using the generated scans were 0.82, 0.71, and 0.92, respectively, when replacing both T1-weighted and FLAIR images; 0.84, 0.74, and 0.97 when replacing only the FLAIR images; and 0.97, 0.95, and 0.92 when replacing only the T1-weighted images. Conclusion Brain MRI scans generated using generative adversarial networks can be used as deep learning model inputs in case MRI sequences are missing. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Zhong in this issue. An earlier incorrect version of this article appeared online. This article was corrected on April 12, 2021.

A multi-institutional analysis of a general pelvis continuous Hounsfield unit synthetic CT software for radiotherapy

Purpose

To validate a synthetic computed tomography (sCT) software with continuous HUs and large field‐of‐view (FOV) coverage for magnetic resonance imaging (MRI)‐only workflow of general pelvis anatomy in radiotherapy (RT).

Methods

An sCT software for general pelvis anatomy (prostate, rectum, and female pelvis) has been developed by Philips Healthcare and includes continuous HUs assignment along with large FOV coverage. General pelvis sCTs were generated using a two‐stack T1‐weighted mDixon fast‐field echo (FFE) sequence with a superior‐inferior coverage of 36 cm. Seventy‐seven prostate, 43 rectum, and 27 gynecological cases were scanned by three different institutions. mDixon image quality and sCTs were evaluated for soft tissue contrast by using a confidence level scale from 1 to 5 for bladder, prostate/rectum interface, mesorectum, and fiducial maker visibility. Dosimetric comparison was performed by recalculating the RT plans on the sCT after rigid registration. For 12 randomly selected cases, the mean absolute error (MAE) between sCT and CT was calculated to evaluate HU similarity, and the Pearson correlation coefficients (PCC) between the CT‐ and sCT‐generated digitally reconstructed radiographs (DRRs) were obtained for quantitative comparison. To examine geometric accuracy of sCT as a reference for cone beam CT (CBCT), the difference between bone‐based alignment of CBCT to CT and CBCT to sCT was obtained for 19 online‐acquired CBCTs from three patients.

Results

Two‐stack mDixon scans with large FOV did not show any image inhomogeneity or fat‐water swap artifact. Fiducials, Foley catheter, and even rectal spacer were visible as dark signal on the sCT. Average visibility confidence level (average ± standard deviation) on the sCT was 5.0 ± 0.0, 4.6 ± 0.5, 3.8 ± 0.4, and 4.0 ± 1.1 for bladder, prostate/rectum interface, mesorectum and fiducial markers. Dosimetric accuracy showed on average < 1% difference with the CT‐based plans for target and normal structures. The MAE of bone and soft tissue between the sCT and CT are 120.9 ± 15.4 HU, 33.4 ± 4.1 HU, respectively. Average PCC of all evaluated DRR pairs was 0.975. The average offset between CT and sCT as reference was (LR, AP, SI) = (0.19 ± 0.35, 0.14 ± 0.60, 0.44 ± 0.54) mm.

Conclusions

The continuous HU sCT software‐generated realistic sCTs and DRRs to enable MRI‐only planning for general pelvis anatomy.

Advancements in Neuroimaging to Unravel Biological and Molecular Features of Brain Tumors

Simple Summary

Advanced neuroimaging is gaining increasing relevance for the characterization and the molecular profiling of brain tumor tissue. On one hand, for some tumor types, the most widespread advanced techniques, investigating diffusion and perfusion features, have been proven clinically feasible and rather robust for diagnosis and prognosis stratification. In addition, 2-hydroxyglutarate spectroscopy, for the first time, offers the possibility to directly measure a crucial molecular marker. On the other hand, numerous innovative approaches have been explored for a refined evaluation of tumor microenvironments, particularly assessing microstructural and microvascular properties, and the potential applications of these techniques are vast and still to be fully explored.

Abstract

In recent years, the clinical assessment of primary brain tumors has been increasingly dependent on advanced magnetic resonance imaging (MRI) techniques in order to infer tumor pathophysiological characteristics, such as hemodynamics, metabolism, and microstructure. Quantitative radiomic data extracted from advanced MRI have risen as potential in vivo noninvasive biomarkers for predicting tumor grades and molecular subtypes, opening the era of “molecular imaging” and radiogenomics. This review presents the most relevant advancements in quantitative neuroimaging of advanced MRI techniques, by means of radiomics analysis, applied to primary brain tumors, including lower-grade glioma and glioblastoma, with a special focus on peculiar oncologic entities of current interest. Novel findings from diffusion MRI (dMRI), perfusion-weighted imaging (PWI), and MR spectroscopy (MRS) are hereby sifted in order to evaluate the role of quantitative imaging in neuro-oncology as a tool for predicting molecular profiles, stratifying prognosis, and characterizing tumor tissue microenvironments. Furthermore, innovative technological approaches are briefly addressed, including artificial intelligence contributions and ultra-high-field imaging new techniques. Lastly, after providing an overview of the advancements, we illustrate current clinical applications and future perspectives.

Volumetric and dosimetric impact of MRI in delineation of gross tumor volume of non-spinal vertebral metastases treated with stereotactic ablative radiation therapy

PURPOSE

To investigate the Gross Tumor Volumes (GTV) and its dosimetric impact of magnetic resonance imaging (MRI) assisted contouring for non-spinal metastasis treated with stereotactic ablative body radiotherapy (SABR).

MATERIAL AND METHODS

Five observer contours on CT (GTV_CT) and CT+MR (GTV_CT+MR) were evaluated against expert team contours (GTV_EC) for 14 selected cases. Dice Similarity Index (DSC) and Geographical Miss Index (GMI) quantify observer variation. We also analyze the maximum dose (D_max) and dose received by 0.35cc (D_0.35cc) of the spinal cord (SC) for GTV_CT and GTV_CT+MR, where optimization parameters and priorities were unchanged. Percent rank function is also evaluated for SC doses.

RESULTS

The mean DSC and GMI scores for the CT-only dataset are 0.6974 and 0.2851 and for CT+MR dataset is 0.7764 and 0.1907 respectively. Statistically, significant results were found for mean GTV volumes between GTV_EC versus GTV_CT and GTV_CT versus GTV_CT+MR (P<0.001). Dosimetric analysis of D_max and D_0.35cc exceeded 84.2% and 88.5% of times its respective threshold doses for CT-only dataset, whereas for the CT+MR dataset, it exceeded only by 18% and 15.7% times. 'Percent rank' function analysis for SC doses also indicates the same.

CONCLUSION

This study supports MRI fusion for GTV and OAR delineation for non-spinal metastasis. Our study showed that the dosimetric analysis is vital for observer variation studies and the addition of the MR data set is significant to improve the confidence of Stereotactic treatments.

Importance and Evidence of Extent of Resection in Glioblastoma

Maximal safe resection is an essential part of the multidisciplinary care of patients with glioblastoma. A growing body of data shows that gross total resection is an independent prognostic factor associated with improved clinical outcome. The relationship between extent of glioblastoma (GB) resection and clinical benefit depends critically on the balance between cytoreduction and avoiding neurologic morbidity. The definition of the extent of tumor resection, how this is best measured pre- and postoperatively, and its relation to volume of residual tumor is still discussed. We review the literature supporting extent of resection in GB, highlighting the importance of a standardized definition and measurement of extent of resection to allow greater collaboration in research projects and trials. Recent developments in neurosurgical techniques and technologies focused on maximizing extent of resection and safety are discussed.

Use of MRI increases interobserver agreement on gross tumor volume for imaging-diagnosed canine intracranial meningioma

There is a lack of information regarding interobserver agreement on canine meningioma gross tumor volume (GTV) delineation, and on the impact of MRI on this agreement. The objectives of this retrospective, secondary analysis, observer agreement study were to describe agreement between veterinary radiation oncologists on GTV for canine intracranial meningioma, and to compare interobserver agreement between delineation based on CT alone and delineation based on fused CT-MRI. Eighteen radiation oncologists delineated GTV for 13 dogs with an imaging diagnosis of meningioma on pre- and postcontrast CT, pre- and postcontrast T1-weighted magnetic resonance, and T2-weighted magnetic resonance images. Dice similarity coefficient (DSC), concordance index (CI), and center of volume (COV) were used to quantify interobserver agreement. Multilevel mixed models were used to examine the difference in volume, DSC, CI and COV 3D distance between CT and CT-MR imaging. The mean volume for GTV contours delineated using fused CT-MRI was larger than when CT alone was used for delineation (mean difference CT-MR - CT = 0.89 cm3, 95% CI 0.66 to 1.12, P < .001). Interobserver agreement on GTV was improved when MRI was used; the mean DSC and CI were higher, and the mean COV 3D distance was lower, when fused CT-MRI was used than when CT alone was used (P < .001 for all differences). Based on our results, fused CT-MRI is recommended for radiation therapy planning of canine intracranial meningioma.

Improving Delineation of True Tumor Volume With Multimodal MRI in a Rat Model of Brain Metastasis

PURPOSE

Brain metastases are almost universally lethal with short median survival times. Despite this, they are often potentially curable, with therapy failing only because of local relapse. One key reason relapse occurs is because treatment planning did not delineate metastasis margins sufficiently or accurately, allowing residual tumor to regrow. The aim of this study was to determine the extent to which multimodal magnetic resonance imaging (MRI), with a simple and automated analysis pipeline, could improve upon current clinical practice of single-modality, independent-observer tumor delineation.

METHODS AND MATERIALS

We used a single rat model of brain metastasis (ENU1564 breast carcinoma cells in BD-IX rats), with and without radiation therapy. Multimodal MRI data were acquired using sequences either in current clinical use or in clinical trial and included postgadolinium T1-weighted images and maps of blood flow, blood volume, T1 and T2 relaxation times, and apparent diffusion coefficient.

RESULTS

In all cases, independent observers underestimated the true size of metastases from single-modality gadolinium-enhanced MRI (85 ± 36 μL vs 131 ± 40 μL histologic measurement), although multimodal MRI more accurately delineated tumor volume (132 ± 41 μL). Multimodal MRI offered increased sensitivity compared with independent observer for detecting metastasis (0.82 vs 0.61, respectively), with only a slight decrease in specificity (0.86 vs 0.98). Blood flow maps conferred the greatest improvements in margin detection for late-stage metastases after radiation therapy. Gadolinium-enhanced T1-weighted images conferred the greatest increase in accuracy of detection for smaller metastases.

CONCLUSIONS

These findings suggest that multimodal MRI of brain metastases could significantly improve the visualization of brain metastasis margins, beyond current clinical practice, with the potential to decrease relapse rates and increase patient survival. This finding now needs validation in additional tumor models or clinical cohorts.

Three-Plane-assembled Deep Learning Segmentation of Gliomas

PURPOSE

To design a computational method for automatic brain glioma segmentation of multimodal MRI scans with high efficiency and accuracy.

MATERIALS AND METHODS

The 2018 Multimodal Brain Tumor Segmentation Challenge (BraTS) dataset was used in this study, consisting of routine clinically acquired preoperative multimodal MRI scans. Three subregions of glioma-the necrotic and nonenhancing tumor core, the peritumoral edema, and the contrast-enhancing tumor-were manually labeled by experienced radiologists. Two-dimensional U-Net models were built using a three-plane-assembled approach to segment three subregions individually (three-region model) or to segment only the whole tumor (WT) region (WT-only model). The term three-plane-assembled means that coronal and sagittal images were generated by reformatting the original axial images. The model performance for each case was evaluated in three classes: enhancing tumor (ET), tumor core (TC), and WT.

RESULTS

On the internal unseen testing dataset split from the 2018 BraTS training dataset, the proposed models achieved mean Sørensen-Dice scores of 0.80, 0.84, and 0.91, respectively, for ET, TC, and WT. On the BraTS validation dataset, the proposed models achieved mean 95% Hausdorff distances of 3.1 mm, 7.0 mm, and 5.0 mm, respectively, for ET, TC, and WT and mean Sørensen-Dice scores of 0.80, 0.83, and 0.91, respectively, for ET, TC, and WT. On the BraTS testing dataset, the proposed models ranked fourth out of 61 teams. The source code is available at https://github.com/GuanLab/Brain_Glioma.

CONCLUSION

This deep learning method consistently segmented subregions of brain glioma with high accuracy, efficiency, reliability, and generalization ability on screening images from a large population, and it can be efficiently implemented in clinical practice to assist neuro-oncologists or radiologists. Supplemental material is available for this article. © RSNA, 2020.

Multimodality image registration in the head-and-neck using a deep learning-derived synthetic CT as a bridge

PURPOSE

To develop and demonstrate the efficacy of a novel head-and-neck multimodality image registration technique using deep-learning-based cross-modality synthesis.

METHODS AND MATERIALS

Twenty-five head-and-neck patients received magnetic resonance (MR) and computed tomography (CT) (CTaligned ) scans on the same day with the same immobilization. Fivefold cross validation was used with all of the MR-CT pairs to train a neural network to generate synthetic CTs from MR images. Twenty-four of 25 patients also had a separate CT without immobilization (CTnon-aligned ) and were used for testing. CTnon-aligned 's were deformed to the synthetic CT, and compared to CTnon-aligned registered to MR. The same registrations were performed from MR to CTnon-aligned and from synthetic CT to CTnon-aligned . All registrations used B-splines for modeling the deformation, and mutual information for the objective. Results were evaluated using the 95% Hausdorff distance among spinal cord contours, landmark error, inverse consistency, and Jacobian determinant of the estimated deformation fields.

RESULTS

When large initial rigid misalignment is present, registering CT to MRI-derived synthetic CT aligns the cord better than a direct registration. The average landmark error decreased from 9.8 ± 3.1 mm in MR→CTnon-aligned to 6.0 ± 2.1 mm in CTsynth →CTnon-aligned deformable registrations. In the CT to MR direction, the landmark error decreased from 10.0 ± 4.3 mm in CTnon-aligned →MR deformable registrations to 6.6 ± 2.0 mm in CTnon-aligned →CTsynth deformable registrations. The Jacobian determinant had an average value of 0.98. The proposed method also demonstrated improved inverse consistency over the direct method.

CONCLUSIONS

We showed that using a deep learning-derived synthetic CT in lieu of an MR for MR→CT and CT→MR deformable registration offers superior results to direct multimodal registration.

Determination of optimal virtual monochromatic energy level for target delineation of brain metastases in radiosurgery using dual-energy CT

OBJECTIVE

Determination of the optimal energy level of virtual monochromatic image (VMI) for brain metastases in contrast-enhanced dual-energy CT (DECT) for radiosurgery and assessment of the subjective and objective image quality of VMI at the optimal energy level.

METHODS

20 patients (total of 42 metastases) underwent contrast-enhanced DECT. Spectral image analysis of VMIs at energy levels ranging from 40 to 140 keV in 1 keV increments was performed to determine the optimal VMI (VMI_opt) as the one corresponding to the highest contrast-to-noise ratio (CNR) between brain parenchyma and the metastases. The objective and subjective values of VMI_opt were compared to those of the VMI with 120 kVp equivalent, defined as reference VMI (VMI_ref, 77 keV). The objective measurement parameters included mean HU value and SD of tumor and brain parenchyma, absolute lesion contrast (LC), and CNR. The subjective measurements included five-point scale assessment of "overall image quality" and "tumor delineation" by three radiation oncologists.

RESULTS

The VMI at 63 keV was defined as VMI_opt. The LC and CNR of VMI_opt were significantly (p < 0.01) higher than those of VMI_ref (LC: 37.4 HU vs 24.7 HU; CNR: 1.1 vs 0.8, respectively). Subjective analysis rated VMI_opt significantly (p < 0.01) superior to VMI_ref with respect to the overall image quality (3.2 vs 2.9, respectively) and tumor delineation (3.5 vs 2.9, respectively).

CONCLUSION

The VMI at 63 keV derived from contrast-enhanced DECT yielded the highest CNR and improved the objective and subjective image quality for radiosurgery, compared to VMI_ref.

ADVANCES IN KNOWLEDGE

This paper investigated for the first time the optimal energy level of VMI in DECT for brain metastases. The findings will lead to improvement in tumor visibility with optimal VMI and consequently supplement accuracy delineation of brain metastases.

Feasibility of synthetic computed tomography generated with an adversarial network for multi-sequence magnetic resonance-based brain radiotherapy

The aim of this work is to generate synthetic computed tomography (sCT) images from multi-sequence magnetic resonance (MR) images using an adversarial network and to assess the feasibility of sCT-based treatment planning for brain radiotherapy. Datasets for 15 patients with glioblastoma were selected and 580 pairs of CT and MR images were used. T1-weighted, T2-weighted and fluid-attenuated inversion recovery MR sequences were combined to create a three-channel image as input data. A conditional generative adversarial network (cGAN) was trained using image patches. The image quality was evaluated using voxel-wise mean absolute errors (MAEs) of the CT number. For the dosimetric evaluation, 3D conformal radiotherapy (3D-CRT) and volumetric modulated arc therapy (VMAT) plans were generated using the original CT set and recalculated using the sCT images. The isocenter dose and dose-volume parameters were compared for 3D-CRT and VMAT plans, respectively. The equivalent path length was also compared. The mean MAEs for the whole body, soft tissue and bone region were 108.1 ± 24.0, 38.9 ± 10.7 and 366.2 ± 62.0 hounsfield unit, respectively. The dosimetric evaluation revealed no significant difference in the isocenter dose for 3D-CRT plans. The differences in the dose received by 2% of the volume (D2%), D50% and D98% relative to the prescribed dose were <1.0%. The overall equivalent path length was shorter than that for real CT by 0.6 ± 1.9 mm. A treatment planning study using generated sCT detected only small, clinically negligible differences. These findings demonstrated the feasibility of generating sCT images for MR-only radiotherapy from multi-sequence MR images using cGAN.

Imaging for Target Delineation and Treatment Planning in Radiation Oncology: Current and Emerging Techniques

Imaging in radiation oncology has a wide range of applications. It is necessary not only for tumor staging and treatment response assessment after therapy but also for the treatment planning process, including definition of target and organs at risk, as well as treatment plan calculation. This article provides a comprehensive overview of the main imaging modalities currently used for target delineation and treatment planning and gives insight into new and promising techniques.

The rationale for MR-only treatment planning for external radiotherapy

•MR-only treatment planning could improve the spatial accuracy of radiotherapy.•The benefit compared to a mixed MR-CT workflow will vary between patient groups.•Further development of QA tools is needed before the procedure will save resources.

Stability of radiomic features of apparent diffusion coefficient (ADC) maps for locally advanced rectal cancer in response to image pre-processing

Quantitative imaging features (radiomics) extracted from apparent diffusion coefficient (ADC) maps of rectal cancer patients can provide additional information to support treatment decision. Most available radiomic computational packages allow extraction of hundreds to thousands of features. However, two major factors can influence the reproducibility of radiomic features: interobserver variability, and imaging filtering applied prior to features extraction. In this exploratory study we seek to determine to what extent various commonly-used features are reproducible with regards to the mentioned factors using ADC maps from two different clinics (56 patients). Features derived from intensity distribution histograms are less sensitive to manual tumour delineation differences, noise in ADC images, pixel size resampling and intensity discretization. Shape features appear to be strongly affected by delineation quality. On the whole, textural features appear to be poorly or moderately reproducible with respect to the image pre-processing perturbations we reproduced.

Inter-rater agreement in glioma segmentations on longitudinal MRI

Background

Tumor segmentation of glioma on MRI is a technique to monitor, quantify and report disease progression. Manual MRI segmentation is the gold standard but very labor intensive. At present the quality of this gold standard is not known for different stages of the disease, and prior work has mainly focused on treatment-naive glioblastoma. In this paper we studied the inter-rater agreement of manual MRI segmentation of glioblastoma and WHO grade II-III glioma for novices and experts at three stages of disease. We also studied the impact of inter-observer variation on extent of resection and growth rate.

Methods

In 20 patients with WHO grade IV glioblastoma and 20 patients with WHO grade II-III glioma (defined as non-glioblastoma) both the enhancing and non-enhancing tumor elements were segmented on MRI, using specialized software, by four novices and four experts before surgery, after surgery and at time of tumor progression. We used the generalized conformity index (GCI) and the intra-class correlation coefficient (ICC) of tumor volume as main outcome measures for inter-rater agreement.

Results

For glioblastoma, segmentations by experts and novices were comparable. The inter-rater agreement of enhancing tumor elements was excellent before surgery (GCI 0.79, ICC 0.99) poor after surgery (GCI 0.32, ICC 0.92), and good at progression (GCI 0.65, ICC 0.91). For non-glioblastoma, the inter-rater agreement was generally higher between experts than between novices. The inter-rater agreement was excellent between experts before surgery (GCI 0.77, ICC 0.92), was reasonable after surgery (GCI 0.48, ICC 0.84), and good at progression (GCI 0.60, ICC 0.80). The inter-rater agreement was good between novices before surgery (GCI 0.66, ICC 0.73), was poor after surgery (GCI 0.33, ICC 0.55), and poor at progression (GCI 0.36, ICC 0.73). Further analysis showed that the lower inter-rater agreement of segmentation on postoperative MRI could only partly be explained by the smaller volumes and fragmentation of residual tumor. The median interquartile range of extent of resection between raters was 8.3% and of growth rate was 0.22 mm/year.

Conclusion

Manual tumor segmentations on MRI have reasonable agreement for use in spatial and volumetric analysis. Agreement in spatial overlap is of concern with segmentation after surgery for glioblastoma and with segmentation of non-glioblastoma by non-experts.

•

Inter-rater agreement for longitudinal glioma segmentation was determined.

•

Agreement between 4 experts was higher than between 4 novices.

•

Three time-points of glioblastoma (WHO IV) and diffuse glioma (WHO II-III) are studied.

•

Impact on extent of resection and growth rate measurements was determined.

Automatic brain lesion segmentation on standard MRIs of the human head: a scoping review protocol

INTRODUCTION

Automatic brain lesion segmentation from medical images has great potential to support clinical decision making. Although numerous methods have been proposed, significant challenges must be addressed before they will become established in clinical and research practice. We aim to elucidate the state of the art, to provide a synopsis of competing approaches and identify contrasts between them.

METHODS AND ANALYSIS

We present the background and study design of a scoping review for automatic brain lesion segmentation methods for conventional MRI according to the framework proposed by Arksey and O'Malley. We aim to identify common image processing steps as well as mathematical and computational theories implemented in these methods. We will aggregate the evidence on the efficacy and identify limitations of the approaches. Methods to be investigated work with standard MRI sequences from human patients examined for brain lesions, and are validated with quantitative measures against a trusted reference. PubMed, IEEE Xplore and Scopus will be searched using search phrases that will ensure an inclusive and unbiased overview. For matching records, titles and abstracts will be screened to ensure eligibility. Studies will be excluded if a full paper is not available or is not written in English, if non-standard MR sequences are used, if there is no quantitative validation, or if the method is not automatic. In the data charting phase, we will extract information about authors, publication details and study cohort. We expect to find information about preprocessing, segmentation and validation procedures. We will develop an analytical framework to collate, summarise and synthesise the data.

ETHICS AND DISSEMINATION

Ethical approval for this study is not required since the information will be extracted from published studies. We will submit the review report to a peer-reviewed scientific journal and explore other venues for presenting the work.

Integration of rotational angiography enables better dose planning in Gamma Knife radiosurgery for brain arteriovenous malformations

OBJECTIVE

In Gamma Knife radiosurgery (GKS) for arteriovenous malformations (AVMs), CT angiography (CTA), MRI, and digital subtraction angiography (DSA) are generally used to define the nidus. Although the AVM angioarchitecture can be visualized with superior resolution using rotational angiography (RA), the efficacy of integrating RA into the GKS treatment planning process has not been elucidated.

METHODS

Using data collected from 25 consecutive patients with AVMs who were treated with GKS at the authors’ institution, two neurosurgeons independently created treatment plans for each patient before and after RA integration. For all patients, MR angiography, contrasted T1 imaging, CTA, DSA, and RA were performed before treatment. The prescription isodose volume before (PIVB) and after (PIVA) RA integration was measured. For reference purposes, a reference target volume (RTV) for each nidus was determined by two other physicians independent of the planning surgeons, and the RTV covered by the PIV (RTVPIV) was established. The undertreated volume ratio (UVR), overtreated volume ratio (OVR), and Paddick’s conformal index (CI), which were calculated as RTVPIV/RTV, RTVPIV/PIV, and (RTVPIV)2/(RTV × PIV), respectively, were measured by each neurosurgeon before and after RA integration, and the surgeons’ values at each point were averaged. Wilcoxon signed-rank tests were used to compare the values obtained before and after RA integration. The percentage change from before to after RA integration was calculated for the average UVR (%ΔUVRave), OVR (%ΔOVRave), and CI (%ΔCIave) in each patient, as ([value after RA integration]/[value before RA integration] − 1) × 100. The relationships between prior histories and these percentage change values were examined using Wilcoxon signed-rank tests.

RESULTS

The average values obtained by the two surgeons for the median UVR, OVR, and CI were 0.854, 0.445, and 0.367 before RA integration and 0.882, 0.478, and 0.463 after RA integration, respectively. All variables significantly improved after compared with before RA integration (UVR, p = 0.009; OVR, p < 0.001; CI, p < 0.001). Prior hemorrhage was significantly associated with larger %ΔOVRave (median 20.8% vs 7.2%; p = 0.023) and %ΔCIave (median 33.9% vs 13.8%; p = 0.014), but not %ΔUVRave (median 4.7% vs 4.0%; p = 0.449).

CONCLUSIONS

Integrating RA into GKS treatment planning may permit better dose planning owing to clearer visualization of the nidus and, as such, may reduce undertreatment and waste irradiation. Further studies examining whether the observed RA-related improvement in dose planning also improves the radiosurgical outcome are needed.

Role and future of MRI in radiation oncology

Technical innovations and developments in areas such as disease localization, dose calculation algorithms, motion management and dose delivery technologies have revolutionized radiation therapy resulting in improved patient care with superior outcomes. A consequence of the ability to design and accurately deliver complex radiation fields is the need for improved target visualization through imaging. While CT imaging has been the standard of care for more than three decades, the superior soft tissue contrast afforded by MR has resulted in the adoption of this technology in radiation therapy. With the development of real time MR imaging techniques, the problem of real time motion management is enticing. Currently, the integration of an MR imaging and megavoltage radiation therapy treatment delivery system (MR-linac or MRL) is a reality that has the potential to provide improved target localization and real time motion management during treatment. Higher magnetic field strengths provide improved image quality potentially providing the backbone for future work related to image texture analysis-a field known as Radiomics-thereby providing meaningful information on the selection of future patients for radiation dose escalation, motion-managed treatment techniques and ultimately better patient care. On-going advances in MRL technologies promise improved real time soft tissue visualization, treatment margin reductions, beam optimization, inhomogeneity corrected dose calculation, fast multileaf collimators and volumetric arc radiation therapy. This review article provides rationale, advantages and disadvantages as well as ideas for future research in MRI related to radiation therapy mainly in adoption of MRL.

Multi-institutional study of the variability in target delineation for six targets commonly treated with radiosurgery

BACKGROUND

Although accurate delineation of the target is a key factor of success in radiosurgery there are no consensus guidelines for target contouring.

AIM

The aim of the present study was therefore to quantify the variability in target delineation and discuss the potential clinical implications, for six targets regarded as common in stereotactic radiosurgery.

MATERIAL AND METHODS

Twelve Gamma Knife centers participated in the study by contouring the targets and organs at risks and performing the treatment plans. Analysis of target delineation variability was based on metrics defined based on agreement volumes derived from overlapping structures following a previously developed method. The 50% agreement volume (AV₅₀), the common and the encompassing volumes as well as the Agreement Volume Index (AVI) were determined.

RESULTS

Results showed that the lowest AVI (0.16) was found for one of the analyzed metastases (range of delineated volumes 1.27-3.33 cm³). AVI for the other two metastases was 0.62 and 0.37, respectively. Corresponding AVIs for the cavernous sinus meningioma, pituitary adenoma and vestibular schwannoma were 0.22, 0.37 and 0.50.

CONCLUSIONS

This study showed that the variability in the contouring was much higher than expected and therefore further work in standardizing the contouring practice in radiosurgery is warranted.

Improved Brain Tumor Segmentation via Registration-Based Brain Extraction

Automated brain tumor segmenters typically run a “skull-stripping” pre-process to extract the brain from the 3D image, before segmenting the area of interest within the extracted volume. We demonstrate that an effective existing segmenter can be improved by replacing its skull-stripper component with one that instead uses a registration-based approach. In particular, we compare our automated brain segmentation system with the original system as well as three other approaches that differ only by using a different skull-stripper—BET, HWA, and ROBEX: (1) Over scans of 120 patients with brain tumors, our system’s segmentation accuracy (Dice score with respect to expert segmentation) is 8.6% (resp. 2.7%) better than the original system on gross tumor volumes (resp. edema); (2) Over 103 scans of controls, the new system found 92.9% (resp. 57.8%) fewer false positives on T1C (resp. FLAIR) volumes. (The other three methods were significantly worse on both tasks). Finally, the new registration-based approach is over 15% faster than the original, requiring on average only 178 CPU seconds per volume.

Magnetic resonance imaging in radiotherapy treatment target volumes definition for brain tumours: a systematic review and meta-analysis

Purpose

The aim of this study is to establish clinical evidence regarding the use of magnetic resonance imaging (MRI) in target volume definition for radiotherapy treatment planning of brain tumours.

Methods

Primary studies were systematically retrieved from six electronic databases and other sources. Studies included were only those that quantitatively compared computed tomography (CT) and MRI in target volume definition for radiotherapy of brain tumours. Study characteristics and quality were assessed and the data were extracted from eligible studies. Effect estimates for each study was computed as mean percentage difference based on individual patient data where available. The included studies were then combined in meta-analysis using Review Manager (RevMan) software version 5.0.

Result

Five studies with a total number of 72 patients were included in this review. The quality of the studies was rated strong. The percentages mean differences of the studies were 7·47, 11·36, 30·70, 41·69 and −24·6% using CT as the baseline. The result of statistical analysis showed small-to-moderate heterogeneity; τ2=36·8; χ2=6·23; df=4 (p=0·18); I2=36%. The overall effect estimate was −1·85 [95% confidence interval (CI); −7·24, 10·94], Z=0·40 (p=0·069>0·5).

Conclusion

Brain tumour volumes measured using MRI-based method for radiotherapy treatment planning were larger compared with CT defined volumes but the difference lacks statistical significance.

Fully Automatic Lesion Localization and Characterization: Application to Brain Tumors Using Multiparametric Quantitative MRI Data

When analyzing brain tumors, two tasks are intrinsically linked, spatial localization, and physiological characterization of the lesioned tissues. Automated data-driven solutions exist, based on image segmentation techniques or physiological parameters analysis, but for each task separately, the other being performedmanually or with user tuning operations. In this paper, the availability of quantitative magnetic resonance (MR) parameters is combined with advancedmultivariate statistical tools to design a fully automated method that jointly performs both localization and characterization. Non trivial interactions between relevant physiologicalparameters are capturedthanks to recent generalized Student distributions that provide a larger variety of distributional shapes compared to the more standard Gaussian distributions. Probabilisticmixtures of the former distributions are then consideredto account for the different tissue types and potential heterogeneity of lesions. Discriminative multivariate features are extracted from this mixture modeling and turned into individual lesion signatures. The signatures are subsequently pooled together to build a statistical fingerprintmodel of the different lesion types that captures lesion characteristics while accounting for inter-subject variability. The potential of this generic procedure is demonstrated on a data set of 53 rats, with 36 rats bearing 4 different brain tumors, for which 5 quantitative MR parameters were acquired.

New approaches to the analysis of eye movement behaviour across expertise while viewing brain MRIs

Brain tumour detection and diagnosis requires clinicians to inspect and analyse brain magnetic resonance images. Eye-tracking is commonly used to examine observers' gaze behaviour during such medical image interpretation tasks, but analysis of eye movement sequences is limited. We therefore used ScanMatch, a novel technique that compares saccadic eye movement sequences, to examine the effect of expertise and diagnosis on the similarity of scanning patterns. Diagnostic accuracy was also recorded. Thirty-five participants were classified as Novices, Medics and Experts based on their level of expertise. Participants completed two brain tumour detection tasks. The first was a whole-brain task, which consisted of 60 consecutively presented slices from one patient; the second was an independent-slice detection task, which consisted of 32 independent slices from five different patients. Experts displayed the highest accuracy and sensitivity followed by Medics and then Novices in the independent-slice task. Experts showed the highest level of scanning pattern similarity, with medics engaging in the least similar scanning patterns, for both the whole-brain and independent-slice task. In the independent-slice task, scanning patterns were the least similar for false negatives across all expertise levels and most similar for experts when they responded correctly. These results demonstrate the value of using ScanMatch in the medical image perception literature. Future research adopting this tool could, for example, identify cases that yield low scanning similarity and so provide insight into why diagnostic errors occur and ultimately help in training radiologists.

[Delineation of the surgical bed of operated brain metastases treated with adjuvant stereotactic irradiation: A review]

Stereotactic radiotherapy of the surgical bed of brain metastases is a technique that comes supplant indications of adjuvant whole brain radiotherapy after surgery. After a growing number of retrospective studies, a phase III trial has been presented and validated this indication. However, several criteria such as the dose, the fractionation, the use of a margin and definition of volumes remain to be defined. Our study consisted in making a literature review in order to provide a guideline of delineation of surgical beds of brain metastases, as well as the different modalities of their implementation process.

Expertise Affects Inter-Observer Agreement at Peripheral Locations within a Brain Tumor

Magnetic resonance imaging (MRI) is a crucial tool for clinical brain tumor detection and delineation. Since the process of gross tumor volume delineation resides with clinicians, a better understanding of how they perform this task is required if improvements in life expectancy are to be made. Novice-expert comparison studies have been used to examine the effect of expertise on abnormality detection, but little research has investigated expertise-related differences in brain tumor delineation. In this study, undergraduate students (novices) and radiologists (experts) inspected a combination of T1 and T2 single and whole brain MRI scans, each containing a tumor. Using a tablet and stylus to provide an interactive environment, participants had an unlimited amount of time to scroll freely through the MRI slices and were instructed to delineate (i.e., draw a boundary) around any tumorous tissue. There was no reliable evidence for a difference in the gross tumor volume or total number of slices delineated between experts and novices. Agreement was low across both expertise groups and significantly lower at peripheral locations within a tumor than central locations. There was an interaction between expertise level and location within a tumor with experts displaying higher agreement at the peripheral slices than novices. An effect of brain image set on the order in which participants inspected the slices was also observed. The implications of these results for the training undertaken by early career radiologists and current practices in hospitals are discussed.

Inter-observer agreement in GTV delineation of bone metastases on CT and impact of MR imaging: A multicenter study

BACKGROUND AND PURPOSE

The use of Stereotactic Body Radiotherapy (SBRT) for bone metastases is increasing rapidly. Therefore, knowledge of the inter-observer differences in tumor volume delineation is essential to guarantee precise dose delivery. The aim of this study is to compare inter-observer agreement in bone metastases delineated on different imaging modalities.

MATERIAL AND METHODS

Twenty consecutive patients with bone metastases treated with SBRT were selected. All patients received CT and MR imaging in treatment position prior to SBRT. Five observers from three institutions independently delineated gross tumor volume (GTV) on CT alone, CT with co-registered MRI and MRI alone. Four contours per imaging modality per patient were available, as one set of contours was shared by 2 observers. Inter-observer agreement, expressed in generalized conformity index [CIgen], volumes of contours and contours center of mass (COM) were calculated per patient and imaging modality.

RESULTS

Mean GTV delineated on MR (45.9±52.0cm³) was significantly larger compared to CT-MR (40.2±49.4cm³) and CT (34.8±41.8cm³). A considerable variation in CIgen was found on CT (mean 0.46, range 0.15-0.75) and CT-MRI (mean 0.54, range 0.17-0.71). The highest agreement was found on MRI (mean 0.56, range 0.20-0.77). The largest variations of COM were found in anterior-posterior direction for all imaging modalities.

CONCLUSIONS

Large inter-observer variation in GTV delineation exists for CT, CT-MRI and MRI. MRI-based GTV delineation resulted in larger volumes and highest consistency between observers.

Statistical evaluation of manual segmentation of a diffuse low-grade glioma MRI dataset

Software-based manual segmentation is critical to the supervision of diffuse low-grade glioma patients and to the optimal treatment's choice. However, manual segmentation being time-consuming, it is difficult to include it in the clinical routine. An alternative to circumvent the time cost of manual segmentation could be to share the task among different practitioners, providing it can be reproduced. The goal of our work is to assess diffuse low-grade gliomas' manual segmentation's reproducibility on MRI scans, with regard to practitioners, their experience and field of expertise. A panel of 13 experts manually segmented 12 diffuse low-grade glioma clinical MRI datasets using the OSIRIX software. A statistical analysis gave promising results, as the practitioner factor, the medical specialty and the years of experience seem to have no significant impact on the average values of the tumor volume variable.

Altered Network Topology in Patients with Primary Brain Tumors After Fractionated Radiotherapy

Radiation therapy (RT) is a critical treatment modality for patients with brain tumors, although it can cause adverse effects. Recent data suggest that brain RT is associated with dose-dependent cortical atrophy, which could disrupt neocortical networks. This study examines whether brain RT affects structural network properties in brain tumor patients. We applied graph theory to MRI-derived cortical thickness estimates of 54 brain tumor patients before and after RT. Cortical surfaces were parcellated into 68 regions and correlation matrices were created for patients pre- and post-RT. Significant changes in graph network properties were tested using nonparametric permutation tests. Linear regressions were conducted to measure the association between dose and changes in nodal network connectivity. Increases in transitivity, modularity, and global efficiency (n = 54, p < 0.0001) were all observed in patients post-RT. Decreases in local efficiency (n = 54, p = 0.007) and clustering coefficient (n = 54, p = 0.005) were seen in regions receiving higher RT doses, including the inferior parietal lobule and rostral anterior cingulate. These findings demonstrate alterations in global and local network topology following RT, characterized by increased segregation of brain regions critical to cognition. These pathological network changes may contribute to the late delayed cognitive impairments observed in many patients following brain RT.

MRI Reduces Variation of Contouring for Boost Clinical Target Volume in Breast Cancer Patients Without Surgical Clips in the Tumour Bed

BACKGROUND

Omitting the placement of clips inside tumour bed during breast cancer surgery poses a challenge for delineation of lumpectomy cavity clinical target volume (CTVLC). We aimed to quantify inter-observer variation and accuracy for CT- and MRI-based segmentation of CTVLC in patients without clips.

PATIENTS AND METHODS

CT- and MRI-simulator images of 12 breast cancer patients, treated by breast conserving surgery and radiotherapy, were included in this study. Five radiation oncologists recorded the cavity visualization score (CVS) and delineated CTVLC on both modalities. Expert-consensus (EC) contours were delineated by a senior radiation oncologist, respecting opinions of all observers. Inter-observer volumetric variation and generalized conformity index (CIgen) were calculated. Deviations from EC contour were quantified by the accuracy index (AI) and inter-delineation distances (IDD).

RESULTS

Mean CVS was 3.88 +/- 0.99 and 3.05 +/- 1.07 for MRI and CT, respectively (p = 0.001). Mean volumes of CTVLC were similar: 154 +/- 26 cm3 on CT and 152 +/- 19 cm3 on MRI. Mean CIgen and AI were superior for MRI when compared with CT (CIgen: 0.74 +/- 0.07 vs. 0.67 +/- 0.12, p = 0.007; AI: 0.81 +/- 0.04 vs. 0.76 +/- 0.07; p = 0.004). CIgen and AI increased with increasing CVS. Mean IDD was 3 mm +/- 1.5 mm and 3.6 mm +/- 2.3 mm for MRI and CT, respectively (p = 0.017).

CONCLUSIONS

When compared with CT, MRI improved visualization of post-lumpectomy changes, reduced interobserver variation and improved the accuracy of CTVLC contouring in patients without clips in the tumour bed. Further studies with bigger sample sizes are needed to confirm our findings.

CT- and MRI-based gross target volume comparison in vestibular schwannomas

AIM

This study represents an enumeration and comparison of gross target volumes (GTV) as delineated independently on contrast-enhanced computed tomography (CT) and T1 and T2 weighted magnetic resonance imaging (MRI) in vestibular schwannomas (VS).

BACKGROUND

Multiple imaging in radiotherapy improves target localization.

METHODS AND MATERIALS

42 patients of VS were considered for this prospective study with one patient showing bilateral tumor. The GTV was delineated separately on CT and MRI. Difference in volumes were estimated individually for all the 43 lesions and similarity was studied between CT and T1 and T2 weighted MRI.

RESULTS

The male to female ratio for VS was found to be 1:1.3. The tumor was right sided in 34.9% and left sided in 65.1%. Tumor volumes (TV) on CT image sets were ranging from 0.251 cc to 27.27 cc. The TV for CT, MRI T1 and T2 weighted were 5.15 ± 5.2 cc, 5.8 ± 6.23 cc, and 5.9 ± 6.13 cc, respectively. Compared to MRI, CT underestimated the volumes. The mean dice coefficient between CT versus T1 and CT versus T2 was estimated to be 68.85 ± 18.3 and 66.68 ± 20.3, respectively. The percentage of volume difference between CT and MRI (%VD: mean ± SD for T1; 28.84 ± 15.0, T2; 35.74 ± 16.3) and volume error (%VE: T1; 18.77 ± 10.1, T2; 23.17 ± 13.93) were found to be significant, taking the CT volumes as the baseline.

CONCLUSIONS

MRI with multiple sequences should be incorporated for tumor volume delineation and they provide a clear boundary between the tumor and normal tissue with critical structures nearby.

Reducing radiation dose and enhancing imaging quality of 4DCT for radiation therapy using iterative reconstruction algorithms

Purpose

Four-dimensional computed tomography (CT) images are typically used to quantify the necessary internal target volumes for thoracic and abdominal tumors. However, 4-dimensional CT is typically associated with excessive imaging dose to patients and the situation is exacerbated when using repeat 4-dimensional CT imaging on a weekly or daily basis throughout fractionated therapy. The aim of this work is to evaluate an iterative reconstruction (IR) algorithm that helps reduce the imaging dose to the patient while maintaining imaging quality as quantified by point spread function and contrast-to-noise ratios (CNRs).

Methods and materials

An IR algorithm, SAFIRE, was applied to CT data of a phantom and patients with varying CT doses and reconstruction kernels. Phantom data enable measurements of spatial resolution, contrast, and noise. The impact of SAFIRE on 4-dimensional CT was assessed with patient data acquired at 2 different dose levels during image guided radiation therapy with an in-room CT.

Results

Phantom data demonstrate that IR reduces noise approximately in proportion to the number of iterations indicated by the strength (SAFIRE 1 to SAFIRE 5). Spatial resolution and contrast are conserved independent of dose and reconstruction parameters. The CNR increases with an increase of imaging dose or an increase in the number of iterations. The use of IR on CT sets confirms the results that were derived from phantom scans. The IR significantly enhances single breathing phase CTs in 4-dimensional CT sets as assessed by CT number discrimination. Furthermore, the IR of the low dose 4-dimensional CT features a 45% increase in the CNR in comparison with the standard dose 4-dimensional CT.

Conclusions

The use of IR algorithms reduces noise while preserving spatial resolution and contrast, as evaluated from both phantom and patient CT data sets. For 4-dimensional CT, the IR can significantly improve image quality and reduce imaging dose without compromising image quality.

Simultaneous 68Ga DOTATATE Positron Emission Tomography/Magnetic Resonance Imaging in Meningioma Target Contouring: Feasibility and Impact Upon Interobserver Variability Versus Positron Emission Tomography/Computed Tomography and Computed Tomography/Magnetic Resonance Imaging

AIMS

The increasing use of highly conformal radiation techniques to treat meningioma confers a greater need for accurate targeting. Several groups have shown that positron emission tomography/computed tomography (PET/CT) information alters meningioma targets contoured by single observers, but whether this translates into improved accuracy has not been defined. As magnetic resonance imaging (MRI) is the cornerstone of meningioma target contouring, simultaneous PET/MRI may be superior to PET/CT. We assessed whether ⁶⁸Ga DOTATATE PET imaging (from PET/CT and PET/MRI) reduced interobserver variability (IOV) in meningioma target volume contouring.

MATERIALS AND METHODS

Ten patients with meningioma underwent simultaneous ⁶⁸Ga DOTATATE PET/MRI followed by PET/CT. They were selected as it was anticipated that target volume definition in their cases would be particularly challenging. Three radiation oncologists contoured target volumes according to an agreed protocol: gross tumour volume (GTV) and clinical target volume (CTV) on CT/MRI alone, CT/MRI+PET(CT) and CT/MRI+PET(MRI). GTV/CTV Kouwenhoven conformity levels (KCL), regions of contour variation and qualitative differences between PET(CT) and PET(MRI) were evaluated.

RESULTS

There was substantial IOV in contouring. GTV mean KCL: CT/MRI 0.34, CT/MRI+PET(CT) 0.38, CT/MRI+PET(MRI) 0.39 (P = 0.06). CTV mean KCL: CT/MRI 0.31, CT/MRI+PET(CT) 0.35, CT/MRI+PET(MRI) 0.35 (P = 0.04 for all groups; P > 0.05 for individual pairs). One observer consistently contoured largest and one smallest. Observers rarely decreased volumes in relation to PET. Most IOV occurred in bone followed by dural tail, postoperative bed and venous sinuses. Tumour edges were qualitatively clearer on PET(MRI) versus PET(CT), but this did not affect contouring.

CONCLUSION

IOV in contouring challenging meningioma cases was large and only slightly improved with the addition of ⁶⁸Ga DOTATATE PET. Simultaneous PET/MRI for meningioma contouring is feasible, but did not improve IOV versus PET/CT. Whether volumes can be safely reduced according to PET requires evaluation.

Cerebral Cortex Regions Selectively Vulnerable to Radiation Dose-Dependent Atrophy

PURPOSE AND OBJECTIVES

Neurologic deficits after brain radiation therapy (RT) typically involve decline in higher-order cognitive functions such as attention and memory rather than sensory defects or paralysis. We sought to determine whether areas of the cortex critical to cognition are selectively vulnerable to radiation dose-dependent atrophy.

METHODS AND MATERIALS

We measured change in cortical thickness in 54 primary brain tumor patients who underwent fractionated, partial brain RT. The study patients underwent high-resolution, volumetric magnetic resonance imaging (T1-weighted; T2 fluid-attenuated inversion recovery, FLAIR) before RT and 1 year afterward. Semiautomated software was used to segment anatomic regions of the cerebral cortex for each patient. Cortical thickness was measured for each region before RT and 1 year afterward. Two higher-order cortical regions of interest (ROIs) were tested for association between radiation dose and cortical thinning: entorhinal (memory) and inferior parietal (attention/memory). For comparison, 2 primary cortex ROIs were also tested: pericalcarine (vision) and paracentral lobule (somatosensory/motor). Linear mixed-effects analyses were used to test all other cortical regions for significant radiation dose-dependent thickness change. Statistical significance was set at α = 0.05 using 2-tailed tests.

RESULTS

Cortical atrophy was significantly associated with radiation dose in the entorhinal (P=.01) and inferior parietal ROIs (P=.02). By contrast, no significant radiation dose-dependent effect was found in the primary cortex ROIs (pericalcarine and paracentral lobule). In the whole-cortex analysis, 9 regions showed significant radiation dose-dependent atrophy, including areas responsible for memory, attention, and executive function (P≤.002).

CONCLUSIONS

Areas of cerebral cortex important for higher-order cognition may be most vulnerable to radiation-related atrophy. This is consistent with clinical observations that brain radiation patients experience deficits in domains of memory, executive function, and attention. Correlations of regional cortical atrophy with domain-specific cognitive functioning in prospective trials are warranted.

Progressive disease in glioblastoma: Benefits and limitations of semi-automated volumetry

Purpose

Unambiguous evaluation of glioblastoma (GB) progression is crucial, both for clinical trials as well as day by day routine management of GB patients. 3D-volumetry in the follow-up of GB provides quantitative data on tumor extent and growth, and therefore has the potential to facilitate objective disease assessment. The present study investigated the utility of absolute changes in volume (delta) or regional, segmentation-based subtractions for detecting disease progression in longitudinal MRI follow-ups.

Methods

165 high resolution 3-Tesla MRIs of 30 GB patients (23m, mean age 60.2y) were retrospectively included in this single center study. Contrast enhancement (CV) and tumor-related signal alterations in FLAIR images (FV) were semi-automatically segmented. Delta volume (dCV, dFV) and regional subtractions (sCV, sFV) were calculated. Disease progression was classified for every follow-up according to histopathologic results, decisions of the local multidisciplinary CNS tumor board and a consensus rating of the neuro-radiologic report.

Results

A generalized logistic mixed model for disease progression (yes / no) with dCV, dFV, sCV and sFV as input variables revealed that only dCV was significantly associated with prediction of disease progression (P = .005). Delta volume had a better accuracy than regional, segmentation-based subtractions (79% versus 72%) and a higher area under the curve by trend in ROC curves (.83 versus .75).

Conclusion

Absolute volume changes of the contrast enhancing tumor part were the most accurate volumetric determinant to detect progressive disease in assessment of GB and outweighed FLAIR changes as well as regional, segmentation-based image subtractions. This parameter might be useful in upcoming objective response criteria for glioblastoma.

Radiation Dose-Dependent Hippocampal Atrophy Detected With Longitudinal Volumetric Magnetic Resonance Imaging

PURPOSE

After radiation therapy (RT) to the brain, patients often experience memory impairment, which may be partially mediated by damage to the hippocampus. Hippocampal sparing in RT planning is the subject of recent and ongoing clinical trials. Calculating appropriate hippocampal dose constraints would be improved by efficient in vivo measurements of hippocampal damage. In this study we sought to determine whether brain RT was associated with dose-dependent hippocampal atrophy.

METHODS AND MATERIALS

Hippocampal volume was measured with magnetic resonance imaging (MRI) in 52 patients who underwent fractionated, partial brain RT for primary brain tumors. Study patients had high-resolution, 3-dimensional volumetric MRI before and 1 year after RT. Images were processed using software with clearance from the US Food and Drug Administration and Conformité Européene marking for automated measurement of hippocampal volume. Automated results were inspected visually for accuracy. Tumor and surgical changes were censored. Mean hippocampal dose was tested for correlation with hippocampal atrophy 1 year after RT. Average hippocampal volume change was also calculated for hippocampi receiving high (>40 Gy) or low (<10 Gy) mean RT dose. A multivariate analysis was conducted with linear mixed-effects modeling to evaluate other potential predictors of hippocampal volume change, including patient (random effect), age, hemisphere, sex, seizure history, and baseline volume. Statistical significance was evaluated at α = 0.05.

RESULTS

Mean hippocampal dose was significantly correlated with hippocampal volume loss (r=-0.24, P=.03). Mean hippocampal volume was significantly reduced 1 year after high-dose RT (mean -6%, P=.009) but not after low-dose RT. In multivariate analysis, both RT dose and patient age were significant predictors of hippocampal atrophy (P<.01).

CONCLUSIONS

The hippocampus demonstrates radiation dose-dependent atrophy after treatment for brain tumors. Quantitative MRI is a noninvasive imaging technique capable of measuring radiation effects on intracranial structures. This technique could be investigated as a potential biomarker for development of reliable dose constraints for improved cognitive outcomes.