Johan Coolen MD, Presenter: Nothing to Disclose

Frederik De Keyzer, Abstract Co-Author: Nothing to Disclose

Dirk Vanraemdonck MD, PhD, Abstract Co-Author: Nothing to Disclose

Eric Verbeken MD, PhD, Abstract Co-Author: Nothing to Disclose

Walter De Wever MD, Abstract Co-Author: Nothing to Disclose

Johan Vansteenkiste MD, PhD, Abstract Co-Author: Nothing to Disclose

Ilse Roebben BMEDSC, Abstract Co-Author: Nothing to Disclose

Christophe Dooms MD, PhD, Abstract Co-Author: Nothing to Disclose

Willy COOSEMANS MD, PHD, Abstract Co-Author: Nothing to Disclose

Paul De Leyn MD, PhD, Abstract Co-Author: Nothing to Disclose

Kristiaan Nackaerts MD, PhD, Abstract Co-Author: Speaker, Pfizer Inc Speaker, Novartis AG Advisory Board, Pfizer Inc Travel support, AstraZeneca PLC Travel support, Eli Lilly and Company

Johny A. Verschakelen MD, PhD, Abstract Co-Author: Nothing to Disclose

Steven Dymarkowski MD, Abstract Co-Author: Nothing to Disclose

Invasive surgical procedures (mediastinoscopy/-tomy) are cornerstones of mediastinal oncologic staging, but can lead to patient discomfort and morbidity. In this study, we attempted to use a noninvasive MR imaging approach to differentiate benign from malignant mediastinal lesions and at the same time attempt to provide an imaging-based preoperative evaluation of lesion resectability.

76 consecutive patients (47 male, 29 female), with a suspicious mediastinal lesion on clinical or imaging-based (CT, PET, EBUS or EUS) examinations, were included. Additional to the anatomical MRI sequences on a 3T MR scanner, both a dynamic contrast-enhanced (DCE)- and a diffusion-weighted (DW) acquisition were performed one day prior to surgery. All MR images were evaluated by visual inspection and by calculating the mean ADC values of the whole lesion and of the most suspect zone (based on b1000 and ADC map). Finally, DCE curves were evaluated. Histological examinations of operative specimens served as reference.

A total of 32 benign and 44 malignant lesions were included in this study. Diagnosis based on the mean ADC calculated from the whole lesion, using a threshold of 2.1 x10-3 mm²/s to differentiate benign from malignant, was disappointing (sens 71%, spec 56% and acc 64%). Contrarily, the ADC-based evaluation of the target zone evaluation alone performed better (sens 75%, spec 88% and acc 80%) with an optimal cut-off value of 1.5 x10-3 mm²/s. However, the best results were found using a combination of DW (b1000 and ADC-map interpretation) with DCE-MRI leading to 65 correct and only 11 incorrect (6FN and 5 FP) diagnoses (sens 86%, spec 84% and acc 86%). Over all patients, a very good correlation was found between MR signs of local invasion and surgical resectability (kappa 0.84, p<0.0001).

In the hard-to-image mediastinal region, noninvasive imaging techniques (DWI and DCE) might provide additional information for preoperative lesion characterization and assessment of lesion resectability. However, volume and especially heterogeneity of the mediastinal lesions require the use of target zone calculations of the ADC maps, and improvement seems possible by combining this with a visual DCE interpretation.

MR imaging using DW- and DCE-MRI analysis might help to noninvasively predict lesion resectability prior to surgery, and could help to improve lesion characterization.

Coolen, J, De Keyzer, F, Vanraemdonck, D, Verbeken, E, De Wever, W, Vansteenkiste, J, Roebben, I, Dooms, C, COOSEMANS, W, De Leyn, P, Nackaerts, K, Verschakelen, J, Dymarkowski, S, How functional MR Imaging Can Take the Hurdle of Mediastinal Lesion Characterization and Improve Prediction of Mediastinal Lesion Resectability: From Whole Lesion Assessment to Target Zone Delineation.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045748.html