RSNA 2003 

Abstract Archives of the RSNA, 2003


502-p

Detection of Pulmonary Nodules in MRI: 3D Gradient Echo Sequence and HASTE Sequence vs. Spiral CT

Scientific Posters

Presented on December 3, 2003
Presented as part of L03: Chest Lung Nodules

Participants

Philipp Kamm MD, PRESENTER: Nothing to Disclose

Abstract: HTML Purpose: To test the accuracy of two breathhold MR imaging sequences, a 3D gradient echo sequence and a HASTE sequence, as a tool for lung cancer screening in detection of pulmonary nodules in comparison to Spiral CT. Methods and Materials: 31 patients underwent an MRI exam at an 1.5 T MR scanner (Magnetom Sonata, Siemens). All patients had a Spiral CT (in 36% a Multirow CT) not older than two weeks, which was carried out because of the clinical suspicion for lung nodules. MRI was performed with the following parameters for the 3D gradient echo sequence (3D GRE): TR / TE / flip = 2.9 ms / 1.1 ms / 5º using two overlapping transversal stacks of 160 mm width with 40 partitions. Voxel size was 1.7 x 1.4 x 4 mm. The HASTE sequence was performed in three planes with the following parameters: TR / TE / flip = 800 ms / 25 ms / 150º, voxel size was 2.0 x 1.4 x 5 mm. Analysis was performed separately for both sequences by two independent radiologists without knowledge of the CT exam results and retrospectively in knowledge of the CT. Results: 4 patients showed no lesion, 15 patients had a solitary nodule and in 12 cases several nodules were found in CT, ranging in size from 2 to 35 mm. 14 nodules were smaller than 5 mm. For this lesion size the detection rates in the HASTE sequence were 0, 3, 3 (reader 1, reader 2, retrospective reading), in the 3D GRE the detection rates were 2, 1, 3. Of 12 nodules ranging in size from 5 to 10 mm the detection rates in the HASTE Sequence were 5, 6 ,6 and in the 3D GRE 6, 7, 9 (reader 1, reader 2, retrospective reading). For 18 nodules larger than 10 mm size the detection rates in the HASTE sequence were 17, 18, 18 and in the 3D GRE the detection rates were 17, 18, 18 (reader 1, reader 2, retrospective reading). The overall sensitivity by retrospective reading was 61% (27/44) for the HASTE sequence and 68% (30/44) for the 3D GRE. For lesions larger than 4 mm the sensitivities were 80% (24/30) for the HASTE and 90% (27/30) for the 3D GRE. The rate of false positive findings in the 3D GRE was 40% and 47% (reader 1 and reader 2) lower than in the HASTE Sequence. All four healthy patients were rated correctly in both sequences by reader 1, whereas reader 2 detected 1 of 4 in the HASTE, and 3 of 4 in the 3D GRE. Conclusion: The 3D GRE sequence is superior to the HASTE sequence in the detection of pulmonary nodules larger than 4 mm. The rate of false positive findings is considerably lower in the 3D GRE. Therefore the 3D GRE seems to be a promising approach for radiation free lung cancer screening.       Questions about this event email: juergen.schaefer@med.uni-tuebingen.de

Cite This Abstract

Kamm MD, P, Detection of Pulmonary Nodules in MRI: 3D Gradient Echo Sequence and HASTE Sequence vs. Spiral CT.  Radiological Society of North America 2003 Scientific Assembly and Annual Meeting, November 30 - December 5, 2003 ,Chicago IL. http://archive.rsna.org/2003/3107193.html