Performance of Standardized Relative CBV for Quantifying Regional Histologic Tumor Burden in Recurrent High-Grade Glioma: Comparison against Normalized Relative CBV Using Image-Localized Stereotactic Biopsies.

BACKGROUND AND PURPOSE: Perfusion MR imaging measures of relative CBV can distinguish recurrent tumor from posttreatment radiation effects in high-grade gliomas. Currently, relative CBV measurement requires normalization based on user-defined reference tissues. A recently proposed method of relative CBV standardization eliminates the need for user input. This study compares the predictive performance of relative CBV standardization against relative CBV normalization for quantifying recurrent tumor burden in high-grade gliomas relative to posttreatment radiation effects. MATERIALS AND METHODS: We recruited 38 previously treated patients with high-grade gliomas (World Health Organization grades III or IV) undergoing surgical re-resection for new contrast-enhancing lesions concerning for recurrent tumor versus posttreatment radiation effects. We recovered 112 image-localized biopsies and quantified the percentage of histologic tumor content versus posttreatment radiation effects for each sample. We measured spatially matched normalized and standardized relative CBV metrics (mean, median) and fractional tumor burden for each biopsy. We compared relative CBV performance to predict tumor content, including the Pearson correlation (r), against histologic tumor content (0%-100%) and the receiver operating characteristic area under the curve for predicting high-versus-low tumor content using binary histologic cutoffs (≥50%; ≥80% tumor). RESULTS: Across relative CBV metrics, fractional tumor burden showed the highest correlations with tumor content (0%-100%) for normalized (r = 0.63, P < .001) and standardized (r = 0.66, P < .001) values. With binary cutoffs (ie, ≥50%; ≥80% tumor), predictive accuracies were similar for both standardized and normalized metrics and across relative CBV metrics. Median relative CBV achieved the highest area under the curve (normalized = 0.87, standardized = 0.86) for predicting ≥50% tumor, while fractional tumor burden achieved the highest area under the curve (normalized = 0.77, standardized = 0.80) for predicting ≥80% tumor. CONCLUSIONS: Standardization of relative CBV achieves similar performance compared with normalized relative CBV and offers an important step toward workflow optimization and consensus methodology.

Accurate Patient-Specific Machine Learning Models of Glioblastoma Invasion Using Transfer Learning.

BACKGROUND AND PURPOSE: MR imaging-based modeling of tumor cell density can substantially improve targeted treatment of glioblastoma. Unfortunately, interpatient variability limits the predictive ability of many modeling approaches. We present a transfer learning method that generates individualized patient models, grounded in the wealth of population data, while also detecting and adjusting for interpatient variabilities based on each patient's own histologic data. MATERIALS AND METHODS: We recruited patients with primary glioblastoma undergoing image-guided biopsies and preoperative imaging, including contrast-enhanced MR imaging, dynamic susceptibility contrast MR imaging, and diffusion tensor imaging. We calculated relative cerebral blood volume from DSC-MR imaging and mean diffusivity and fractional anisotropy from DTI. Following image coregistration, we assessed tumor cell density for each biopsy and identified corresponding localized MR imaging measurements. We then explored a range of univariate and multivariate predictive models of tumor cell density based on MR imaging measurements in a generalized one-model-fits-all approach. We then implemented both univariate and multivariate individualized transfer learning predictive models, which harness the available population-level data but allow individual variability in their predictions. Finally, we compared Pearson correlation coefficients and mean absolute error between the individualized transfer learning and generalized one-model-fits-all models. RESULTS: Tumor cell density significantly correlated with relative CBV (r = 0.33, P < .001), and T1-weighted postcontrast (r = 0.36, P < .001) on univariate analysis after correcting for multiple comparisons. With single-variable modeling (using relative CBV), transfer learning increased predictive performance (r = 0.53, mean absolute error = 15.19%) compared with one-model-fits-all (r = 0.27, mean absolute error = 17.79%). With multivariate modeling, transfer learning further improved performance (r = 0.88, mean absolute error = 5.66%) compared with one-model-fits-all (r = 0.39, mean absolute error = 16.55%). CONCLUSIONS: Transfer learning significantly improves predictive modeling performance for quantifying tumor cell density in glioblastoma.

Parapharyngeal Space Venous Malformation: An Imaging Mimic of Pleomorphic Adenoma.

Venous malformations in the parapharyngeal space are rare and may be challenging to diagnose with imaging secondary to multiple overlapping features with pleomorphic adenoma, which is much more commonly found in this region. While both lesions are T1 isointense and T2 hyperintense relative to skeletal muscle and demonstrate contrast enhancement, more uniform T2 hyperintensity and progressive contrast pooling on delayed postcontrast T1WI may allow the radiologist to include venous malformation in the differential diagnosis. This is important because it has the potential to alter management from surgical resection to observation. The primary aim of this study was to review the imaging appearance of parapharyngeal venous malformations through a retrospective case series.

Retrospective Validation of a Computer-Assisted Quantification Model of Intracerebral Hemorrhage Volume on Accuracy, Precision, and Acquisition Time, Compared with Standard ABC/2 Manual Volume Calculation.

BACKGROUND AND PURPOSE: Intracerebral hemorrhage accounts for 6.5%-19.6% of all acute strokes. Initial intracerebral hemorrhage volume and expansion are both independent predictors of clinical outcomes and mortality. Therefore, a rapid, unbiased, and precise measurement of intracerebral hemorrhage volume is a key component of clinical management. The most commonly used method, ABC/2, results in overestimation. We developed an interactive segmentation program, SegTool, using a novel graphic processing unit, level set algorithm. Until now, the speed, bias, and precision of SegTool had not been validated. MATERIALS AND METHODS: In a single stroke academic center, 2 vascular neurologists and 2 neuroradiologists independently performed a test-retest experiment that involved repeat measurements of static, unchanging intracerebral hemorrhage volumes on CT from 76 intracerebral hemorrhage cases. Measurements were made with SegTool and ABC/2. True intracerebral hemorrhage volumes were estimated from a consensus of repeat manual tracings by 2 operators. These data allowed us to estimate measurement bias, precision, and speed. RESULTS: The measurements with SegTool were not significantly different from the true intracerebral hemorrhage volumes, while ABC/2 overestimated volume by 45%. The interrater measurement variability with SegTool was 50% less than that with ABC/2. The average measurement times for ABC/2 and SegTool were 35.7 and 44.6 seconds, respectively. CONCLUSIONS: SegTool appears to have attributes superior to ABC/2 in terms of accuracy and interrater reliability with a 9-second delay in measurement time (on average); hence, it could be useful in clinical trials and practice.

MRI-Based Texture Analysis to Differentiate Sinonasal Squamous Cell Carcinoma from Inverted Papilloma.

BACKGROUND AND PURPOSE: Because sinonasal inverted papilloma can harbor squamous cell carcinoma, differentiating these tumors is relevant. The objectives of this study were to determine whether MR imaging-based texture analysis can accurately classify cases of noncoexistent squamous cell carcinoma and inverted papilloma and to compare this classification performance with neuroradiologists' review. MATERIALS AND METHODS: Adult patients who had inverted papilloma or squamous cell carcinoma resected were eligible (coexistent inverted papilloma and squamous cell carcinoma were excluded). Inclusion required tumor size of >1.5 cm and preoperative MR imaging with axial T1, axial T2, and axial T1 postcontrast sequences. Five well-established texture analysis algorithms were applied to an ROI from the largest tumor cross-section. For a training dataset, machine-learning algorithms were used to identify the most accurate model, and performance was also evaluated in a validation dataset. On the basis of 3 separate blinded reviews of the ROI, isolated tumor, and entire images, 2 neuroradiologists predicted tumor type in consensus. RESULTS: The inverted papilloma (n = 24) and squamous cell carcinoma (n = 22) cohorts were matched for age and sex, while squamous cell carcinoma tumor volume was larger (P = .001). The best classification model achieved similar accuracies for training (17 squamous cell carcinomas, 16 inverted papillomas) and validation (7 squamous cell carcinomas, 6 inverted papillomas) datasets of 90.9% and 84.6%, respectively (P = .537). For the combined training and validation cohorts, the machine-learning accuracy (89.1%) was better than that of the neuroradiologists' ROI review (56.5%, P = .0004) but not significantly different from the neuroradiologists' review of the tumors (73.9%, P = .060) or entire images (87.0%, P = .748). CONCLUSIONS: MR imaging-based texture analysis has the potential to differentiate squamous cell carcinoma from inverted papilloma and may, in the future, provide incremental information to the neuroradiologist.

Is Hypoglossal Nerve Palsy Caused by Craniocervical Junction Degenerative Disease an Underrecognized Entity?

Isolated hypoglossal nerve palsy is uncommon, and underlying craniocervical junction degenerative disease has rarely been reported as an underlying cause. To improve understanding of this entity, we present a retrospective series of 18 patients with hypoglossal palsy in whom twelfth cranial nerve compression within the premedullary cistern or hypoglossal canal, or both, was found secondary to craniocervical junction juxta-articular cysts, retro-odontoid fibrous pseudotumors, and osteophytes. The imaging techniques and characteristic craniocervical junction degenerative disease lesion imaging findings presented here might help clinicians interpreting hypoglossal palsy imaging studies avoid perceptual and interpretive errors commonly found in the present series.

The Impact of Middle Turbinate Concha Bullosa on the Severity of Inferior Turbinate Hypertrophy in Patients with a Deviated Nasal Septum.

BACKGROUND AND PURPOSE: Inferior turbinate hypertrophy and concha bullosa often occur opposite the direction of nasal septal deviation. The objective of this retrospective study was to determine whether a concha bullosa impacts inferior turbinate hypertrophy in patients who have nasal septal deviation. MATERIALS AND METHODS: The electronic medical record was used to identify sinus CT scans exhibiting nasal septal deviation for 100 adult subjects without and 100 subjects with unilateral middle turbinate concha bullosa. Exclusion criteria included previous sinonasal surgery, tumor, sinusitis, septal perforation, and craniofacial trauma. Nasal septal deviation was characterized in the coronal plane by distance from the midline (severity) and height from the nasal floor. Measurement differences between sides for inferior turbinate width (overall and bone), medial mucosa, and distance to the lateral nasal wall were calculated as inferior turbinate hypertrophy indicators. RESULTS: The cohorts with and without concha bullosa were similarly matched for age, sex, and nasal septal deviation severity, though nasal septal deviation height was greater in the cohort with concha bullosa than in the cohort without concha bullosa (19.1 ± 4.3 mm versus 13.5 ± 4.1 mm, P < .001). Compensatory inferior turbinate hypertrophy was significantly greater in the cohort without concha bullosa than in the cohort with it as measured by side-to-side differences in turbinate overall width, bone width, and distance to the lateral nasal wall (P < .01), but not the medial mucosa. Multiple linear regression analyses found nasal septal deviation severity and height to be significant predictors of inferior turbinate hypertrophy with positive and negative relationships, respectively (P < .001). CONCLUSIONS: Inferior turbinate hypertrophy is directly proportional to nasal septal deviation severity and inversely proportional to nasal septal deviation height. The effect of a concha bullosa on inferior turbinate hypertrophy is primarily mediated through influence on septal morphology, because the nasal septal deviation apex tends to be positioned more superior from the nasal floor in these patients.

Radiation dose reduction in paranasal sinus CT using model-based iterative reconstruction.

BACKGROUND AND PURPOSE: CT performed with Veo model-based iterative reconstruction has shown the potential for radiation-dose reduction. This study sought to determine whether Veo could reduce noise and improve the image quality of low-dose sinus CT. MATERIALS AND METHODS: Twenty patients consented to participate and underwent low- and standard-dose sinus CT on the same day. Standard-dose CT was created with filtered back-projection (120 kV[peak], 210 mA, 0.4-second rotation, and 0.531 pitch). For low-dose CT, mA was decreased to 20 (the remaining parameters were unchanged), and images were generated with filtered back-projection and Veo. Standard- and low-dose datasets were reconstructed by using bone and soft-tissue algorithms, while the low-dose Veo reconstruction only had a standard kernel. Two blinded neuroradiologists independently evaluated the image quality of multiple osseous and soft-tissue craniofacial structures. Image noise was measured by using multiple regions of interest. RESULTS: Eight women and 12 men (mean age, 63.3 years) participated. Volume CT dose indices were 2.9 mGy (low dose) and 31.6 mGy (standard dose), and mean dose-length products were 37.4 mGy-cm (low dose) and 406.1 mGy-cm (standard dose). Of all the imaging series, low-dose Veo demonstrated the least noise (P < .001). Compared with filtered back-projection low-dose CT using soft-tissue and bone algorithms, Veo had the best soft-tissue image quality but the poorest bone image quality (P < .001). CONCLUSIONS: Veo significantly reduces noise in low-dose sinus CT. Although this reduction improves soft-tissue evaluation, thin bone becomes less distinct.

Localization of a rapid CSF leak with digital subtraction myelography.

A 53-year-old woman with superficial siderosis underwent spinal MR imaging, which demonstrated a large cervicothoracic epidural fluid collection compatible with a CSF leak. Conventional and dynamic CT myelography failed to localize the dural tear because of rapid equilibration of myelographic contrast between the thecal sac and the extradural collection. The superior temporal resolution of digital subtraction myelography precisely localized the CSF leak preoperatively and led to the successful surgical correction of the dural tear.

Focal opacification of the olfactory recess on sinus CT: just an incidental finding?

BACKGROUND AND PURPOSE: The CT appearance of the anterior skull base has been investigated but with limited attention directed to the olfactory recess. As defined by opacity abutting the undersurface of the cribriform plate, the prevalence of olfactory recess opacity (ORO) on sinus CT was examined to clarify whether this should raise suspicion for an unsuspected pathologic process. MATERIALS AND METHODS: Outpatient sinus CTs were evaluated for ORO in 500 consecutive patients (mean age, 46.9 years; 52.6% women). On a per-side basis (n = 1000), the presence of surgical changes, inflammatory sinus disease, and concha bullosa was determined by 2 neuroradiologists. Logistic regression was used to examine the association of ORO with these variables. RESULTS: ORO was identified in 59 (11.8%) patients, bilateral in 27 (5.4%), and unilateral in 32 (6.4%). There were 343 of 1000 ethmoid sides that were diseased, and 66 (27.2%) showed ipsilateral ORO. In contrast, only 20 (3.0%) of 657 clear ethmoid sides showed ORO (P < .0001). ORO was significantly (P = .013) more common with previous surgery (18/75; 24.0%) than without (68/925; 7.4%). Ipsilateral concha bullosa was not associated with ORO. Of 32 patients with unilateral ORO, 5 (15.6%) had no ethmoid opacification or previous surgery, and 1 of these patients had an encephalocele causing the ORO. Finally, unilateral ORO was present in only 1 of 122 patients with completely clear sinuses (the encephalocele that was just mentioned). CONCLUSION: ORO is distinctly uncommon without sinonasal inflammation or previous surgery. Isolated unilateral ORO raises suspicion for an underlying neoplasm or cephalocele and warrants further evaluation.

Active synovial matrix metalloproteinase-2 is associated with radiographic erosions in patients with early synovitis.

INTRODUCTION: In cancer the gelatinases [matrix metalloproteinase (MMP)-2 and MMP-9] have been shown to be associated with tissue invasion and metastatic disease. In patients with inflammatory arthritis the gelatinases are expressed in the synovial membrane, and have been implicated in synovial tissue invasion into adjacent cartilage and bone. It is hypothesized that an imbalance between the activators and inhibitors of the gelatinases results in higher levels of activity, enhanced local proteolysis, and bone erosion. OBJECTIVES: To determine whether the expression and activity levels of MMP-2 and MMP-9, and their regulators MMP-14 and tissue inhibitor of metalloproteinase (TIMP), are associated with early erosion formation in patients with synovitis of recent onset. PATIENTS AND METHOD: A subset of 66 patients was selected from a larger early synovitis cohort on the basis of tissue availability for the study of synovial tissue and serum gelatinase expression. Patients with peripheral joint synovitis of less than 1 years' duration were evaluated clinically and serologically on four visits over a period of 12 months. At the initial visit, patients underwent a synovial tissue biopsy of one swollen joint, and patients had radiographic evaluation of hands and feet initially and at 1year. Serum MMP-1, MMP-2, MMP-9, MMP-14, and TIMP-1 and TIMP-2 levels were determined, and synovial tissue was examined by immunohistology for the expression of MMP-2 and MMP-9, and their molecular regulators. Gelatinolytic activity for MMP-2 and MMP-9 was quantified using a sensitive, tissue-based gel zymography technique. Four healthy individuals underwent closed synovial biopsy and their synovial tissues were similarly analyzed. RESULTS: Of the 66 patients studied, 45 fulfilled American College of Rheumatology criteria for rheumatoid arthritis (RA), with 32 (71%) being rheumatoid factor positive. Of the 21 non-RA patients, seven had a spondylarthropathy and 14 had undifferentiated arthritis. Radiographically, 12 of the RA patients had erosions at multiple sites by 1 year, whereas none of the non-RA patients had developed erosive disease of this extent. In the tissue, latent MMP-2 was widely expressed in the synovial lining layer and in areas of stromal proliferation in the sublining layer and stroma, whereas MMP-9 was expressed more sparsely and focally. MMP-14, TIMP-2, and MMP-2 were all detected in similar areas of the lining layer on consecutive histologic sections. Tissue expression of MMP-14, the activator for pro-MMP-2, was significantly higher in RA than in non-RA patients (8.4 +/- 5 versus 3.7 +/- 4 cells/high-power field; P = 0.009). In contrast, the expression of TIMP-2, an inhibitor of MMP-2, was lower in the RA than in the non-RA samples (25 +/- 12 versus 39 +/- 9 cells/high-power field; P = 0.01). Synovial tissue expressions of MMP-2, MMP-14, and TIMP-2 were virtually undetectable in normal synovial tissue samples. The synovial tissue samples of patients with erosive disease had significantly higher levels of active MMP-2 than did those of patients without erosions (Fig. 1). Tissue expression of MMP-2 and MMP-9, however, did not correlate with the serum levels of these enzymes. With the exception of serum MMP-2, which was not elevated over normal, serum levels of all of the other MMPs and TIMPs were elevated to varying degrees, and were not predictive of erosive disease. Interestingly, MMP-1 and C-reactive protein, both of which were associated with the presence of erosions, were positively correlated with each other (r = 0.42; P < 0.001). DISCUSSION: MMP-2 and MMP-9 are thought to play an important role in the evolution of joint erosions in patients with an inflammatory arthritis. Most studies have concentrated on the contribution of MMP-9 to the synovitis, because synovial fluid and serum MMP-9 levels are markedly increased in inflammatory arthropathies. Previously reported serum levels of MMP-9 have varied widely. In the present sample of patients with synovitis of recent onset, serum MMP-9 levels were elevated in only 21%. Moreover, these elevations were not specific for RA, the tissue expression of MMP-9 was focal, and the levels of MMP-9 activity were not well correlated with early erosions. Although serum MMP-2 levels were not of prognostic value, high synovial tissue levels of MMP-2 activity were significantly correlated with the presence of early erosions. This may reflect augmented activation of MMP-2 by the relatively high levels of MMP-14 and low levels of TIMP-2 seen in these tissues. We were able to localize the components of this trimolecular complex to the synovial lining layer in consecutive tissue sections, a finding that is consistent with their colocalization. In conclusion, we have provided evidence that active MMP-2 complexes are detectable in the inflamed RA synovium and may be involved in the development of early bony erosions. These results suggest that strategies to inhibit the activation of MMP-2 may have the potential for retarding or preventing early erosions in patients with inflammatory arthritis.

Cerebral metabolic profile, selective neuron loss, and survival of acute and chronic hyperglycemic rats following cardiac arrest and resuscitation.

Cortical metabolites and regional cerebral intracellular pH (pHi) were measured in normoglycemic (NM), acute hyperglycemic (AH), and chronic hyperglycemic (CH, 2 week duration, streptozotocin-induced) Wistar rat brains during cardiac arrest and resuscitation. During total ischemia in AH and CH rats (plasma glucose approximately 30 mM), cortical ATP, PCr, glucose, and glycogen all fell significantly as expected. Lactate levels increased dramatically in association with a concomitant intracellular acidosis. Although lactate reached higher concentrations in AH and CH than NM, pHi was significantly lower only in the AH group. With 5 min of reperfusion, all groups recovered to near baseline in all variables, though lactate remained elevated. In a separate aspect of the study, animals from each experimental group were allowed to recover for 4 days following resuscitation, with outcome being gauged by mortality rate and hippocampal CA1 neuron counts. NM survival rate was significantly better than AH and CH. In particular, no CH rats survived for 4 days despite rapid initial recovery. After 4 days, the AH group had suffered significantly greater CA1 neuron loss than the NM rats. In summary, our research identified differences in intra-ischemic acid-base status in the two hyperglycemic groups, suggesting that chronic hyperglycemia may alter the brain's buffering capacity. These observations may account for differences between acutely and chronically hyperglycemic subjects regarding outcome, and they suggest that factors other than hydrogen ion production during ischemia are responsible for modulating outcome.