Warning: Declaration of My_Walker::start_el(&$output, $item, $depth, $args) should be compatible with Walker_Nav_Menu::start_el(&$output, $data_object, $depth = 0, $args = NULL, $current_object_id = 0) in /home2/ajnrblog/public_html/ajnrdigest/wp-content/themes/ajnr/functions.php on line 258
Multiparametric Analysis of Permeability and ADC Histogram Metrics for Classification of Pediatric Brain Tumors by Tumor Grade - AJNR News Digest
May-June 2018
Pediatrics

Multiparametric Analysis of Permeability and ADC Histogram Metrics for Classification of Pediatric Brain Tumors by Tumor Grade

Vajapeyam pic

Sridhar Vajapeyam

Pediatric brain tumors are a unique group of heterogeneous tumors. MR imaging remains the modality of choice for diagnosis including tumor extent, tumor grade, treatment planning, and image-guided therapies. Physiologic characterization of the tumor related to MR permeability and diffusion characteristics (eg, cellularity) may be used for further evaluation.

In our previous work, we reported the successful use of tissue permeability metrics derived from dynamic contrast-enhanced MR imaging in discriminating between low-grade (grades 1–2) and high-grade (grades 3–4) pediatric brain tumors.1 In particular, we found that Ktrans (transfer constant from the blood plasma into the extracellular extravascular space [EES]) and kep (rate constant from EES back into blood plasma) were significantly higher in high-grade tumors (P <.001), whereas ve (extravascular extracellular volume fraction) was significantly lower (P <.001). At the Neuroimaging Center for the Pediatric Brain Tumor Consortium at Boston Children’s Hospital, we recently reported ADC histogram metrics in diffuse intrinsic pontine glioma, which significantly correlated with survival; lower diffusion values (ie, increased cellularity), increased skewness, and enhancement were associated with shorter survival.2

This current work uses a multiparametric analysis of both permeability and ADC histogram metrics to achieve a more accurate discrimination between low- and high-grade tumors.

Receiver operating characteristic (ROC) analyses of the 5 most significant variables, namely Ktrans, kep, ve, FL_ADC_mean (mean ADC of 3D FLAIR tumor volume), and PG_ADC_mean (mean ADC of 3D enhancing volume) showed that all displayed high sensitivity and specificity in discriminating between low- and high-grade tumors, with the areas of ROC (AROC) values above 0.88 (P <.001). AROC values for ve and FL_ADC_mean were 0.865 and 0.925, respectively, while the AROC estimate for the 2-variable combination of these metrics was 0.965, proving the improved accuracy obtained by multiparametric analyses. These results also underscore the importance of using 3D tumor analyses to effectively characterize heterogeneous tumors. This work was initially presented at the American Society of Neuroradiology (ASNR) 55th Annual Meeting in Long Beach, California.

Molecular subtyping of pediatric brain tumors has become increasingly important, as well as more readily available in recent years with the 2016 WHO Classification of Brain Tumors. Our future work will focus on the use of multiparametric analyses to further characterize molecular subtypes using imaging data such as MR perfusion and diffusion, MR spectroscopy, and PET imaging. We expect to present this work at future ASNR meetings.

References

  1. Vajapeyam S, Stamoulis C, Ricci K, et al. Automated processing of dynamic contrast-enhanced MRI: correlation of advanced pharmacokinetic metrics with tumor grade in pediatric brain tumors. AJNR Am J Neuroradiol 2017;38:170–75, 10.3174/ajnr.A4949.
  2. Poussaint TY, Vajapeyam S, Ricci KI, et al. Apparent diffusion coefficient histogram metrics correlate with survival in diffuse intrinsic pontine glioma: a report from the Pediatric Brain Tumor Consortium. Neuro Oncol 2016;18:725–34, 10.1093/neuonc/nov256.

Read this article at AJNR.org …