Expert-centered Evaluation of Deep Learning Algorithms for Brain Tumor Segmentation.

Purpose To present results from a literature survey on practices in deep learning segmentation algorithm evaluation and perform a study on expert quality perception of brain tumor segmentation. Materials and Methods A total of 180 articles reporting on brain tumor segmentation algorithms were surveyed for the reported quality evaluation. Additionally, ratings of segmentation quality on a four-point scale were collected from medical professionals for 60 brain tumor segmentation cases. Results Of the surveyed articles, Dice score, sensitivity, and Hausdorff distance were the most popular metrics to report segmentation performance. Notably, only 2.8% of the articles included clinical experts' evaluation of segmentation quality. The experimental results revealed a low interrater agreement (Krippendorff α, 0.34) in experts' segmentation quality perception. Furthermore, the correlations between the ratings and commonly used quantitative quality metrics were low (Kendall tau between Dice score and mean rating, 0.23; Kendall tau between Hausdorff distance and mean rating, 0.51), with large variability among the experts. Conclusion The results demonstrate that quality ratings are prone to variability due to the ambiguity of tumor boundaries and individual perceptual differences, and existing metrics do not capture the clinical perception of segmentation quality. Keywords: Brain Tumor Segmentation, Deep Learning Algorithms, Glioblastoma, Cancer, Machine Learning Clinical trial registration nos. NCT00756106 and NCT00662506 Supplemental material is available for this article. © RSNA, 2023.

Not without Context-A Multiple Methods Study on Evaluation and Correction of Automated Brain Tumor Segmentations by Experts.

RATIONALE AND OBJECTIVES: Brain tumor segmentations are integral to the clinical management of patients with glioblastoma, the deadliest primary brain tumor in adults. The manual delineation of tumors is time-consuming and highly provider-dependent. These two problems must be addressed by introducing automated, deep-learning-based segmentation tools. This study aimed to identify criteria experts use to evaluate the quality of automatically generated segmentations and their thought processes as they correct them. MATERIALS AND METHODS: Multiple methods were used to develop a detailed understanding of the complex factors that shape experts' perception of segmentation quality and their thought processes in correcting proposed segmentations. Data from a questionnaire and semistructured interview with neuro-oncologists and neuroradiologists were collected between August and December 2021 and analyzed using a combined deductive and inductive approach. RESULTS: Brain tumors are highly complex and ambiguous segmentation targets. Therefore, physicians rely heavily on the given context related to the patient and clinical context in evaluating the quality and need to correct brain tumor segmentation. Most importantly, the intended clinical application determines the segmentation quality criteria and editing decisions. Physicians' personal beliefs and preferences about the capabilities of AI algorithms and whether questionable areas should not be included are additional criteria influencing the perception of segmentation quality and appearance of an edited segmentation. CONCLUSION: Our findings on experts' perceptions of segmentation quality will allow the design of improved frameworks for expert-centered evaluation of brain tumor segmentation models. In particular, the knowledge presented here can inspire the development of brain tumor-specific metrics for segmentation model training and evaluation.

Abnormal vascular structure and function within brain metastases is linked to pembrolizumab resistance.

BACKGROUND: We recently conducted a phase 2 trial (NCT028865685) evaluating intracranial efficacy of pembrolizumab for brain metastases (BM) of diverse histologies. Our study met its primary efficacy endpoint and illustrates that pembrolizumab exerts promising activity in a select group of patients with BM. Given the importance of aberrant vasculature in mediating immunosuppression, we explored the relationship between checkpoint inhibitor (ICI) efficacy and vascular architecture in the hopes of identifying potential mechanisms of intracranial ICI response or resistance for BM. METHODS: Using Vessel Architectural Imaging (VAI), a histologically validated quantitative metric for in vivo tumor vascular physiology, we analyzed dual echo DSC/DCE MRI for 44 patients on trial. Tumor and peri-tumor cerebral blood volume/flow, vessel size, arterial- and venous-dominance, and vascular permeability were measured before and after treatment with pembrolizumab. RESULTS: BM that progressed on ICI were characterized by a highly aberrant vasculature dominated by large-caliber vessels. In contrast, ICI-responsive BM possessed a more structurally balanced vasculature consisting of both small and large vessels, and there was a trend towards a decrease in under-perfused tissue, suggesting a reversal of the negative effects of hypoxia. In the peri-tumor region, development of smaller blood vessels, consistent with neo-angiogenesis, was associated with tumor growth before radiographic evidence of contrast enhancement on anatomical MRI. CONCLUSIONS: This study, one of the largest functional imaging studies for BM, suggests that vascular architecture is linked with ICI efficacy. Studies identifying modulators of vascular architecture, and effects on immune activity, are warranted and may inform future combination treatments.

Structural and functional vascular dysfunction within brain metastases is linked to pembrolizumab inefficacy.

Structurally and functionally aberrant vasculature is a hallmark of tumor angiogenesis and treatment resistance. Given the synergistic link between aberrant tumor vasculature and immunosuppression, we analyzed perfusion MRI for 44 patients with brain metastases (BM) undergoing treatment with pembrolizumab. To date, vascular-immune communication, or the relationship between immune checkpoint inhibitor (ICI) efficacy and vascular architecture, has not been well-characterized in human imaging studies. We found that ICI-responsive BM possessed a structurally balanced vascular makeup, which was linked to improved vascular efficiency and an immune-stimulatory microenvironment. In contrast, ICI-resistant BM were characterized by a lack of immune cell infiltration and a highly aberrant vasculature dominated by large-caliber vessels. Peri-tumor region analysis revealed early functional changes predictive of ICI resistance before radiographic evidence on conventional MRI. This study was one of the largest functional imaging studies for BM and establishes a foundation for functional studies that illuminate the mechanisms linking patterns of vascular architecture with immunosuppression, as targeting these aspects of cancer biology may serve as the basis for future combination treatments.

Positron Emission Tomography/Magnetic Resonance Imaging (PET/MRI) Versus the Standard of Care Imaging in the Diagnosis of Peritoneal Carcinomatosis.

OBJECTIVE: To compare positron emission tomography (PET)/magnetic resonance imaging (MRI) to the standard of care imaging (SCI) for the diagnosis of peritoneal carcinomatosis (PC) in primary abdominopelvic malignancies. SUMMARY BACKGROUND DATA: Identifying PC impacts prognosis and management of multiple cancer types. METHODS: Adult subjects were prospectively and consecutively enrolled from April 2019 to January 2021. Inclusion criteria were: 1) acquisition of whole-body contrast-enhanced (CE) 18F-fluorodeoxyglucose PET/MRI, 2) pathologically confirmed primary abdominopelvic malignancies. Exclusion criteria were: 1) greater than 4 weeks interval between SCI and PET/MRI, 2) unavailable follow-up. SCI consisted of whole-body CE PET/computed tomography (CT) with diagnostic quality CT, and/or CE-CT of the abdomen and pelvis, and/or CE-MRI of the abdomen±pelvis. If available, pathology or surgical findings served as the reference standard, otherwise, imaging followup was used. When SCI and PET/MRI results disagreed, medical records were checked for management changes. Follow-up data were collected until August 2021. RESULTS: One hundred sixty-four subjects were included, 85 (52%) were female, and the median age was 60 years (interquartile range 50-69). At a subject level, PET/MRI had higher sensitivity (0.97, 95% CI 0.86-1.00) than SCI (0.54, 95% CI 0.37-0.71), P < 0.001, without a difference in specificity, of 0.95 (95% CI 0.90-0.98) for PET/MRI and 0.98 (95% CI 0.93-1.00) for SCI, P » 0.250. PET/MRI and SCI results disagreed in 19 cases. In 5/19 (26%) of the discordant cases, PET/MRI findings consistent with PC missed on SCI led to management changes. CONCLUSION: PET/MRI improves detection of PC compared with SCI which frequently changes management.

Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of Glioblastoma.

PURPOSE: To determine the influence of preprocessing on the repeatability and redundancy of radiomics features extracted using a popular open-source radiomics software package in a scan-rescan glioblastoma MRI study. MATERIALS AND METHODS: In this study, a secondary analysis of T2-weighted fluid-attenuated inversion recovery (FLAIR) and T1-weighted postcontrast images from 48 patients (mean age, 56 years [range, 22-77 years]) diagnosed with glioblastoma were included from two prospective studies (ClinicalTrials.gov NCT00662506 [2009-2011] and NCT00756106 [2008-2011]). All patients underwent two baseline scans 2-6 days apart using identical imaging protocols on 3-T MRI systems. No treatment occurred between scan and rescan, and tumors were essentially unchanged visually. Radiomic features were extracted by using PyRadiomics (https://pyradiomics.readthedocs.io/) under varying conditions, including normalization strategies and intensity quantization. Subsequently, intraclass correlation coefficients were determined between feature values of the scan and rescan. RESULTS: Shape features showed a higher repeatability than intensity (adjusted P < .001) and texture features (adjusted P < .001) for both T2-weighted FLAIR and T1-weighted postcontrast images. Normalization improved the overlap between the region of interest intensity histograms of scan and rescan (adjusted P < .001 for both T2-weighted FLAIR and T1-weighted postcontrast images), except in scans where brain extraction fails. As such, normalization significantly improves the repeatability of intensity features from T2-weighted FLAIR scans (adjusted P = .003 [z score normalization] and adjusted P = .002 [histogram matching]). The use of a relative intensity binning strategy as opposed to default absolute intensity binning reduces correlation between gray-level co-occurrence matrix features after normalization. CONCLUSION: Both normalization and intensity quantization have an effect on the level of repeatability and redundancy of features, emphasizing the importance of both accurate reporting of methodology in radiomics articles and understanding the limitations of choices made in pipeline design. Supplemental material is available for this article. © RSNA, 2020See also the commentary by Tiwari and Verma in this issue.

PET/MRI assessment of lung nodules in primary abdominal malignancies: sensitivity and outcome analysis.

PURPOSE: To evaluate PET/MR lung nodule detection compared to PET/CT or CT, to determine growth of nodules missed by PET/MR, and to investigate the impact of missed nodules on clinical management in primary abdominal malignancies. METHODS: This retrospective IRB-approved study included [18F]-FDG PET/MR in 126 patients. All had standard of care chest imaging (SCI) with diagnostic chest CT or PET/CT within 6 weeks of PET/MR that served as standard of reference. Two radiologists assessed lung nodules (size, location, consistency, position, and [18F]-FDG avidity) on SCI and PET/MR. A side-by-side analysis of nodules on SCI and PET/MR was performed. The nodules missed on PET/MR were assessed on follow-up SCI to ascertain their growth (≥ 2 mm); their impact on management was also investigated. RESULTS: A total of 505 nodules (mean 4 mm, range 1-23 mm) were detected by SCI in 89/126 patients (66M:60F, mean age 60 years). PET/MR detected 61 nodules for a sensitivity of 28.1% for patient and 12.1% for nodule, with higher sensitivity for > 7 mm nodules (< 30% and > 70% respectively, p < 0.05). 75/337 (22.3%) of the nodules missed on PET/MR (follow-up mean 736 days) demonstrated growth. In patients positive for nodules at SCI and negative at PET/MR, missed nodules did not influence patients' management. CONCLUSIONS: Sensitivity of lung nodule detection on PET/MR is affected by nodule size and is lower than SCI. 22.3% of missed nodules increased on follow-up likely representing metastases. Although this did not impact clinical management in study group with primary abdominal malignancy, largely composed of extra-thoracic advanced stage cancers, with possible different implications in patients without extra-thoracic spread.

Consensus recommendations for a dynamic susceptibility contrast MRI protocol for use in high-grade gliomas.

Despite the widespread clinical use of dynamic susceptibility contrast (DSC) MRI, DSC-MRI methodology has not been standardized, hindering its utilization for response assessment in multicenter trials. Recently, the DSC-MRI Standardization Subcommittee of the Jumpstarting Brain Tumor Drug Development Coalition issued an updated consensus DSC-MRI protocol compatible with the standardized brain tumor imaging protocol (BTIP) for high-grade gliomas that is increasingly used in the clinical setting and is the default MRI protocol for the National Clinical Trials Network. After reviewing the basis for controversy over DSC-MRI protocols, this paper provides evidence-based best practices for clinical DSC-MRI as determined by the Committee, including pulse sequence (gradient echo vs spin echo), BTIP-compliant contrast agent dosing (preload and bolus), flip angle (FA), echo time (TE), and post-processing leakage correction. In summary, full-dose preload, full-dose bolus dosing using intermediate (60°) FA and field strength-dependent TE (40-50 ms at 1.5 T, 20-35 ms at 3 T) provides overall best accuracy and precision for cerebral blood volume estimates. When single-dose contrast agent usage is desired, no-preload, full-dose bolus dosing using low FA (30°) and field strength-dependent TE provides excellent performance, with reduced contrast agent usage and elimination of potential systematic errors introduced by variations in preload dose and incubation time.

Vascular dysfunction promotes regional hypoxia after bevacizumab therapy in recurrent glioblastoma patients.

BACKGROUND: Hypoxia is a driver of treatment resistance in glioblastoma. Antiangiogenic agents may transiently normalize blood vessels and decrease hypoxia before excessive pruning of vessels increases hypoxia. The time window of normalization is dose and time dependent. We sought to determine how VEGF blockade with bevacizumab modulates tumor vasculature and the impact that those vascular changes have on hypoxia in recurrent glioblastoma patients. METHODS: We measured tumor volume, vascular permeability (Ktrans), perfusion parameters (cerebral blood flow/volume, vessel caliber, and mean transit time), and regions of hypoxia in patients with recurrent glioblastoma before and after treatment with bevacizumab alone or with lomustine using [18F]FMISO PET-MRI. We also examined serial changes in plasma biomarkers of angiogenesis and inflammation. RESULTS: Eleven patients were studied. The magnitude of global tumor hypoxia was variable across these 11 patients prior to treatment and it did not significantly change after bevacizumab. The hypoxic regions had an inefficient vasculature characterized by elevated cerebral blood flow/volume and increased vessel caliber. In a subset of patients, there were tumor subregions with decreased mean transit times and a decrease in hypoxia, suggesting heterogeneous improvement in vascular efficiency. Bevacizumab significantly changed known pharmacodynamic biomarkers such as plasma VEGF and PlGF. CONCLUSIONS: The vascular signature in hypoxic tumor regions indicates a disorganized vasculature which, in most tumors, does not significantly change after bevacizumab treatment. While some tumor regions showed improved vascular efficiency following treatment, bevacizumab did not globally alter hypoxia or normalize tumor vasculature in glioblastoma.

Management implications of fluorodeoxyglucose positron emission tomography/magnetic resonance in untreated intrahepatic cholangiocarcinoma.

PURPOSE: Intrahepatic cholangiocarcinoma (ICC) is associated with a poor prognosis with surgical resection offering the best chance for long-term survival and potential cure. However, in up to 36% of patients who undergo surgery, more extensive disease is found at time of operation requiring cancellation of surgery. PET/MR is a novel hybrid technology that might improve local and whole-body staging in ICC patients, potentially influencing clinical management. This study was aimed to investigate the possible management implications of PET/MR, relative to conventional imaging, in patients affected by untreated intrahepatic cholangiocarcinoma. METHODS: Retrospective review of the clinicopathologic features of 37 patients with iCCC, who underwent PET/MR between September 2015 and August 2018, was performed to investigate the management implications that PET/MR had exerted on the affected patients, relative to conventional imaging. RESULTS: Of the 37 patients enrolled, median age 63.5 years, 20 (54%) were female. The same day PET/CT was performed in 26 patients. All patients were iCCC-treatment-naïve. Conventional imaging obtained as part of routine clinical care demonstrated early-stage resectable disease for 15 patients and advanced stage disease beyond the scope of surgical resection for 22. PET/MR modified the clinical management of 11/37 (29.7%) patients: for 5 patients (13.5%), the operation was cancelled due to identification of additional disease, while 4 "inoperable" patients (10.8%) underwent an operation. An additional 2 patients (5.4%) had a significant change in their operative plan based on PET/MR. CONCLUSIONS: When compared with standard imaging, PET/MR significantly influenced the treatment plan in 29.7% of patients with iCCC. TRIAL REGISTRATION: 2018P001334.

Clinical impact of PET/MR in treated colorectal cancer patients.

PURPOSE: The primary aim of the present study was to evaluate if PET/MR induced management changes versus standard of care imaging (SCI) in treated colorectal cancer patients. The secondary aim was to assess the staging performance of PET/MR and of SCI versus the final oncologic stage. METHODS: Treated CRC patients who underwent PET/MR with 18F-FDG and SCI between January 2016 and October 2018 were enrolled in this retrospective study. Their medical records were evaluated to ascertain if PET/MR had impacted on their clinical management versus SCI. The final oncologic stage, as reported in the electronic medical record, was considered the true stage of disease. RESULTS: A total of 39 patients who underwent 42 PET/MR studies were included, mean age 56.7 years (range 39-75 years), 26 males, and 13 females. PET/MR changed clinical management 15/42 times (35.7%, standard error ± 7.4%); these 15 changes in management were due to upstaging in 9/42 (21.5%) and downstaging in 6/42 (14.2%). The differences in management prompted by SCI versus PET/MR were statistically significant, and PET/MR outperformed SCI (P value < 0.001; odds ratio = 2.8). In relation to the secondary outcome, PET/MR outperformed the SCI in accuracy of oncologic staging (P value = 0.016; odds ratio = 4.6). CONCLUSIONS: PET/MR is a promising imaging tool in the evaluation of treated CRC and might change the management in these patients. However, multicenter prospective studies with larger patient samples are required in order to confirm these preliminary results.

Automatic assessment of glioma burden: a deep learning algorithm for fully automated volumetric and bidimensional measurement.

BACKGROUND: Longitudinal measurement of glioma burden with MRI is the basis for treatment response assessment. In this study, we developed a deep learning algorithm that automatically segments abnormal fluid attenuated inversion recovery (FLAIR) hyperintensity and contrast-enhancing tumor, quantitating tumor volumes as well as the product of maximum bidimensional diameters according to the Response Assessment in Neuro-Oncology (RANO) criteria (AutoRANO). METHODS: Two cohorts of patients were used for this study. One consisted of 843 preoperative MRIs from 843 patients with low- or high-grade gliomas from 4 institutions and the second consisted of 713 longitudinal postoperative MRI visits from 54 patients with newly diagnosed glioblastomas (each with 2 pretreatment "baseline" MRIs) from 1 institution. RESULTS: The automatically generated FLAIR hyperintensity volume, contrast-enhancing tumor volume, and AutoRANO were highly repeatable for the double-baseline visits, with an intraclass correlation coefficient (ICC) of 0.986, 0.991, and 0.977, respectively, on the cohort of postoperative GBM patients. Furthermore, there was high agreement between manually and automatically measured tumor volumes, with ICC values of 0.915, 0.924, and 0.965 for preoperative FLAIR hyperintensity, postoperative FLAIR hyperintensity, and postoperative contrast-enhancing tumor volumes, respectively. Lastly, the ICCs for comparing manually and automatically derived longitudinal changes in tumor burden were 0.917, 0.966, and 0.850 for FLAIR hyperintensity volume, contrast-enhancing tumor volume, and RANO measures, respectively. CONCLUSIONS: Our automated algorithm demonstrates potential utility for evaluating tumor burden in complex posttreatment settings, although further validation in multicenter clinical trials will be needed prior to widespread implementation.

Publisher Correction: Probing tumor microenvironment in patients with newly diagnosed glioblastoma during chemoradiation and adjuvant temozolomide with functional MRI.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has been fixed in the paper.

Probing tumor microenvironment in patients with newly diagnosed glioblastoma during chemoradiation and adjuvant temozolomide with functional MRI.

Functional MRI may identify critical windows of opportunity for drug delivery and distinguish between early treatment responders and non-responders. Using diffusion-weighted, dynamic contrast-enhanced, and dynamic susceptibility contrast MRI, as well as pro-angiogenic and pro-inflammatory blood markers, we prospectively studied the physiologic tumor-related changes in fourteen newly diagnosed glioblastoma patients during standard therapy. 153 MRI scans and blood collection were performed before chemoradiation (baseline), weekly during chemoradiation (week 1-6), monthly before each cycle of adjuvant temozolomide (pre-C1-C6), and after cycle 6. The apparent diffusion coefficient, volume transfer coefficient (Ktrans), and relative cerebral blood volume (rCBV) and flow (rCBF) were calculated within the tumor and edema regions and compared to baseline. Cox regression analysis was used to assess the effect of clinical variables, imaging, and blood markers on progression-free (PFS) and overall survival (OS). After controlling for additional covariates, high baseline rCBV and rCBF within the edema region were associated with worse PFS (microvessel rCBF: HR = 7.849, p = 0.044; panvessel rCBV: HR = 3.763, p = 0.032; panvessel rCBF: HR = 3.984; p = 0.049). The same applied to high week 5 and pre-C1 Ktrans within the tumor region (week 5 Ktrans: HR = 1.038, p = 0.003; pre-C1 Ktrans: HR = 1.029, p = 0.004). Elevated week 6 VEGF levels were associated with worse OS (HR = 1.034; p = 0.004). Our findings suggest a role for rCBV and rCBF at baseline and Ktrans and VEGF levels during treatment as markers of response. Functional imaging changes can differ substantially between tumor and edema regions, highlighting the variable biologic and vascular state of tumor microenvironment during therapy.

Comparison of the clinical performance of upper abdominal PET/DCE-MRI with and without concurrent respiratory motion correction (MoCo).

PURPOSE: To compare the clinical performance of upper abdominal PET/DCE-MRI with and without concurrent respiratory motion correction (MoCo). METHODS: MoCo PET/DCE-MRI of the upper abdomen was acquired in 44 consecutive oncologic patients and compared with non-MoCo PET/MRI. SUVmax and MTV of FDG-avid upper abdominal malignant lesions were assessed on MoCo and non-MoCo PET images. Image quality was compared between MoCo DCE-MRI and non-MoCo CE-MRI, and between fused MoCo PET/MRI and fused non-MoCo PET/MRI images. RESULTS: MoCo PET resulted in higher SUVmax (10.8 ± 5.45) than non-MoCo PET (9.62 ± 5.42) and lower MTV (35.55 ± 141.95 cm3) than non-MoCo PET (38.11 ± 198.14 cm3; p < 0.005 for both). The quality of MoCo DCE-MRI images (4.73 ± 0.5) was higher than that of non-MoCo CE-MRI images (4.53±0.71; p = 0.037). The quality of fused MoCo-PET/MRI images (4.96 ± 0.16) was higher than that of fused non-MoCo PET/MRI images (4.39 ± 0.66; p < 0.005). CONCLUSION: MoCo PET/MRI provided qualitatively better images than non-MoCo PET/MRI, and upper abdominal malignant lesions demonstrated higher SUVmax and lower MTV on MoCo PET/MRI.

Pharmacodynamics of mutant-IDH1 inhibitors in glioma patients probed by in vivo 3D MRS imaging of 2-hydroxyglutarate.

Inhibitors of the mutant isocitrate dehydrogenase 1 (IDH1) entered recently in clinical trials for glioma treatment. Mutant IDH1 produces high levels of 2-hydroxyglurate (2HG), thought to initiate oncogenesis through epigenetic modifications of gene expression. In this study, we show the initial evidence of the pharmacodynamics of a new mutant IDH1 inhibitor in glioma patients, using non-invasive 3D MR spectroscopic imaging of 2HG. Our results from a Phase 1 clinical trial indicate a rapid decrease of 2HG levels by 70% (CI 13%, P = 0.019) after 1 week of treatment. Importantly, inhibition of mutant IDH1 may lead to the reprogramming of tumor metabolism, suggested by simultaneous changes in glutathione, glutamine, glutamate, and lactate. An inverse correlation between metabolic changes and diffusion MRI indicates an effect on the tumor-cell density. We demonstrate a feasible radiopharmacodynamics approach to support the rapid clinical translation of rationally designed drugs targeting IDH1/2 mutations for personalized and precision medicine of glioma patients.

Diagnostic performance of PET/MR in the evaluation of active inflammation in Crohn disease.

This study investigates the performance of PET/MR versus each sub-modality alone in the assessment of active inflammation in patients with Crohn disease, when compared to surgery as standard of reference. Sensitivity for detecting active inflammation was 91.5% for PET, 80% for MR, and 88% for PET/MR. Specificity for active inflammation was 74% for PET, 87% for MR, and 93% for PET/MR. Diagnostic accuracy was 84% for PET, 83% for MR, and 91% for PET/MR. In conclusion, PET/MR is significantly more accurate than either sub-modality alone and more specific than PET alone in the detection of active inflammation in patients with Crohn disease.

Residual Convolutional Neural Network for the Determination of IDH Status in Low- and High-Grade Gliomas from MR Imaging.

Purpose: Isocitrate dehydrogenase (IDH) mutations in glioma patients confer longer survival and may guide treatment decision making. We aimed to predict the IDH status of gliomas from MR imaging by applying a residual convolutional neural network to preoperative radiographic data.Experimental Design: Preoperative imaging was acquired for 201 patients from the Hospital of University of Pennsylvania (HUP), 157 patients from Brigham and Women's Hospital (BWH), and 138 patients from The Cancer Imaging Archive (TCIA) and divided into training, validation, and testing sets. We trained a residual convolutional neural network for each MR sequence (FLAIR, T2, T1 precontrast, and T1 postcontrast) and built a predictive model from the outputs. To increase the size of the training set and prevent overfitting, we augmented the training set images by introducing random rotations, translations, flips, shearing, and zooming.Results: With our neural network model, we achieved IDH prediction accuracies of 82.8% (AUC = 0.90), 83.0% (AUC = 0.93), and 85.7% (AUC = 0.94) within training, validation, and testing sets, respectively. When age at diagnosis was incorporated into the model, the training, validation, and testing accuracies increased to 87.3% (AUC = 0.93), 87.6% (AUC = 0.95), and 89.1% (AUC = 0.95), respectively.Conclusions: We developed a deep learning technique to noninvasively predict IDH genotype in grade II-IV glioma using conventional MR imaging using a multi-institutional data set. Clin Cancer Res; 24(5); 1073-81. ©2017 AACR.

Computation of ultimate SAR amplification factors for radiofrequency hyperthermia in non-uniform body models: impact of frequency and tumour location.

PURPOSE: We introduce a method for calculation of the ultimate specific absorption rate (SAR) amplification factors (uSAF) in non-uniform body models. The uSAF is the greatest possible SAF achievable by any hyperthermia (HT) phased array for a given frequency, body model and target heating volume. METHODS: First, we generate a basis-set of solutions to Maxwell's equations inside the body model. We place a large number of electric and magnetic dipoles around the body model and excite them with random amplitudes and phases. We then compute the electric fields created in the body model by these excitations using an ultra-fast volume integral solver called MARIE. We express the field pattern that maximises the SAF in the target tumour as a linear combination of these basis fields and optimise the combination weights so as to maximise SAF (concave problem). We compute the uSAFs in the Duke body models at 10 frequencies in the 20-900 MHz range and for twelve 3 cm-diameter tumours located at various depths in the head and neck. RESULTS: For both shallow and deep tumours, the frequency yielding the greatest uSAF was ∼900 MHz. Since this is the greatest frequency that we simulated, we hypothesise that the globally optimal frequency is actually greater. CONCLUSIONS: The uSAFs computed in this work are very large (40-100 for shallow tumours and 4-17 for deep tumours), indicating that there is a large room for improvement of the current state-of-the-art head and neck HT devices.

Early changes in glioblastoma metabolism measured by MR spectroscopic imaging during combination of anti-angiogenic cediranib and chemoradiation therapy are associated with survival.

Precise assessment of treatment response in glioblastoma during combined anti-angiogenic and chemoradiation remains a challenge. In particular, early detection of treatment response by standard anatomical imaging is confounded by pseudo-response or pseudo-progression. Metabolic changes may be more specific for tumor physiology and less confounded by changes in blood-brain barrier permeability. We hypothesize that metabolic changes probed by magnetic resonance spectroscopic imaging can stratify patient response early during combination therapy. We performed a prospective longitudinal imaging study in newly diagnosed glioblastoma patients enrolled in a phase II clinical trial of the pan-vascular endothelial growth factor receptor inhibitor cediranib in combination with standard fractionated radiation and temozolomide (chemoradiation). Forty patients were imaged weekly during therapy with an imaging protocol that included magnetic resonance spectroscopic imaging, perfusion magnetic resonance imaging, and anatomical magnetic resonance imaging. Data were analyzed using receiver operator characteristics, Cox proportional hazards model, and Kaplan-Meier survival plots. We observed that the ratio of total choline to healthy creatine after 1 month of treatment was significantly associated with overall survival, and provided as single parameter: (1) the largest area under curve (0.859) in receiver operator characteristics, (2) the highest hazard ratio (HR = 85.85, P = 0.006) in Cox proportional hazards model, (3) the largest separation (P = 0.004) in Kaplan-Meier survival plots. An inverse correlation was observed between total choline/healthy creatine and cerebral blood flow, but no significant relation to tumor volumetrics was identified. Our results suggest that in vivo metabolic biomarkers obtained by magnetic resonance spectroscopic imaging may be an early indicator of response to anti-angiogenic therapy combined with standard chemoradiation in newly diagnosed glioblastoma.