MRI in Breast Imaging Atiyah Yahya, Ph.D., FCCPM

1 MRI in Breast Imaging Atiyah Yahya, Ph.D., FCCPMDepartm...
Author: Lee George
0 downloads 0 Views

1 MRI in Breast Imaging Atiyah Yahya, Ph.D., FCCPMDepartment of Medical Physics, Cross Cancer Institute Department of Oncology, University of Alberta Edmonton, Canada MRI in Breast Imaging

2 Introduction to NMR Preparation of the nuclear system1H nuclei z, Bo Mo y Excitation of the nuclear system Mo Bo y x z o = Bo x Absence of magnetic field Presence of a static field Bo RF coil: A resonating LC circuit tuned to o creates an excitation pulse. 90 pulse

3 A 90° B1 RF pulse rotates Mo to the transverse plane.Apply a magnetic field B1 (RF pulse) which rotates at the Larmor frequency. z Bo Mo B1 y x o Mo Bo x z y A 90° B1 RF pulse rotates Mo to the transverse plane. o

4 Bo C Surface coil: A resonating LC circuit tuned to o I(t)=Iocos(ot) C C B1(t) = B1cos(ot) C

5 Acquisition of the nuclear signalRF coil: A resonating LC circuit tuned to o C C To Receiver M C o spectrum FT time signal

6 MRI Spatial encoding is achieved by the use of magnetic field gradients. A gradient enables slice selection in one direction. Phase and frequency encoding gradients in the orthogonal directions enables spatial encoding across the slice. A 2D inverse Fourier transform is performed to convert the spatial frequency map to an image.

7 Three magnetic fields in MRI:Main magnetic field, Bo (T) Excitation RF field, B1 (T) x, y, and z magnetic field gradients (mT/m) Figure from: Shim coils produce small currents that compensate for Bo inhomogeneities.

8 Superconducting MagnetB1 Bo Superconducting Magnet Transmit/receive head RF coil I Bo Superconducting Magnet

9 Example: Spin Echo MRI Pulse Sequence180y TR is the repetition time (time from one 90° excitation pulse to the next). RF time TE/ TE/2 Gz time Gx time Gy time = slice selective pulses = phase encode gradients = slice selection gradients = frequency encode gradients = spoiler gradients

10 K-space (spatial frequencies)x image-space kx y 2DFT N x N ky N x N kx Frequency Encoding

11 y image-space N x N N x N is commonly 256 x 256kx y N x N N x N is commonly 256 x 256 If field of view = 250 mm, spatial resolution = 0.98 mm

12 T2 Weighting T2 relaxation: dephasing of spins in the transverse plane. T2 weighted sequence: long TR and relatively long TE. Tissue with longer T2 appears brighter Mo Bo y x z M Bo y x z t time (t) M Moexp(-t/T2) Mo

13 T1 Weighting T1 relaxation: Return of longitudinal magnetization back to its thermal equilibrium state. T1 weighted sequence: short TE and relatively short TR Tissue with shorter T1 appears brighter time (t) Mz Mo Mz = Mo(1-exp(-t/T1))

14 Dynamic Contrast Enhanced MRIGadolinium (Gd) based contrast agent is injected into the patient. The contrast agent largely reduces T1 of the surrounding tissue. Highly vascularized tumour tissue has a higher initial uptake of the agent. Tumour appears brighter or enhanced on a T1 weighted image. Subtract post contrast images from pre-contrast image. Initial rise typically occurs in less than two minutes. Figure 1 from Kuhl et al. Radiology, 211: , 1999

15 Fat Suppression Employed to minimize signal from breast fatty tissue.Inversion recovery is a method of fat suppression. SPAIR (SPectral Attenuated Inversion Recovery) 180 frequency selective inversion (targets fat) 90x 180y RF time TI TE/ TE/2 Gz time Gx time Gy time

16 Mo time (t) Mz fat -Mo RF time Gz time Gx time Gy time TI TE/2 TE/2180 frequency selective inversion (targets fat) 90x 180y RF time TI TE/ TE/2 Gz time Gx time Gy time time (t) Mo -Mo Mz fat

17 Invasive cancer T2 weighted non-fat suppressedT1 weighted, pre-contrast fat suppressed T1 weighted, post-contrast fat suppressed Figure 1 from:

18 T1 weighted, post-contrastFigure 2 from: T1 weighted, post-contrast

19 Maximum Intensity Projection (MIP)Use post-contrast subtracted images. MIP projects brightest pixels through a volume on a two dimensional projection. MIPs are generated from contrast enhanced breast MRI images. Provides a projection of vessels and enhanced tissue. Yields a distribution of enhancement. Figure 3 from:

20 MRI in Breast ScreeningMRI is not recommended for general screening because it: Results in more false positives (lower specificity) which leads to unnecessary biopsies. Is expensive. Is not able to detect cancers indicated by microcalcifications. Scans during luteal phase of menstrual cycle are avoided due to background enhancement from higher levels of estrogen and progesterone1. Uses a contrast agent. Longer acquisition times. C. Klifa, J. Magn. Reson. Imaging, v. 33, p. 1229, 2011

21 A number of guidelines1 recommend MRI in addition to mammography annually for screening high risk women including those that: Have an at least 20 – 25 % lifetime risk of developing breast cancer (e.g have a BRCA 1 and 2 gene mutation or family history). Received chest radiation treatment between the ages of (for example to treat Hodgkin Lymphoma). Figure from: A. Scaranelo, Canadian Medical Association Journal, v. 184, 2012

22 MRI in Diagnosis Extent of disease Lesion characterizationMRI has shown to be useful for identifying multicentric and multifocal breast cancer in some patients with invasive carcinoma and ductal carcinoma in situ (DCIS). bilateral disease the extent of invasive lobular carcinoma Lesion characterization MRI may be used to characterize an indeterminate lesion (after assessing with mammography, ultrasound and physical inspection).

23 Metastatic disease with unknown primaryMRI may be able to identify a primary breast tumour when metastatic disease of unknown origin is present. MRI guided biopsy MRI guided biopsy may be used when a lesion is only identifiable by MRI and can be used for fine needle aspiration, core needle, vacuum-assisted biopsy and pre-operative wire localization procedures.

24 Figures from:

25 MRI in Planning TreatmentMRI evaluation may be performed prior to surgical treatment (mastectomy or breast conservation) to evaluate the tumour position and extent and the presence of contralateral cancer. used to assess residual disease post surgery in patients with positive or close margins. MRI useful for detecting residual disease even in cases when no residual calcifications are detected on the mammogram1 performed before, during and after chemotherapy treatment to assess response. used to assess for recurrence when mammography and ultrasound yield inconclusive results (differentiation of scar tissue from recurrence). L. Lalonde et al., CARJ, v. 56, p. 301, 2005

26 MRI for Assessing Breast Implant RupturesMRI provides high contrast between implants and soft tissue. Can use T2 weighted images or inversion recovery sequences where either the breast tissue or implant signal is saturated. time (t) Mo -Mo Mz water

27 Left breast implant ruptureFigure 2 from M.S. di Santolo et al., Quantitative Imaging in Medicine and Surgery, v. 4, p. 462, 2014

28 Abbreviated MRI for Breast Screening1Abbreviated protocol to reduce MRI acquisition time to improve the feasibility of MRI for breast screening. Protocol includes one pre-contrast scan, one-post contrast scan and a MIP of the subtracted post-contrast images. Acquisition time is approximately 3 minutes, the time for a radiologist to read the MIP is 3 seconds and the time to read the post contrast images is under 30 seconds. Was found that the abbreviated protocol yielded results comparable to a full MRI screening protocol (about 40 minutes). C. Kuhl et al., Journal of Clinical Oncology, v. 32, p. 2304, 2014

29 Contraindications to Breast MRIImplantable devices that are not MRI compatible (patients are screened) Inability to lie prone Severe obesity Claustrophobia Pregnancy (risk/benefit analysis) Contraindications to gadolinium contrast agents

30 Contraindications to Gd Contrast AgentsAllergic reactions to Gd based contrast agents are rare ( 0.15%)1 Patients with a previous allergic reaction can be administered with a prophylaxis before another Gd contrast agent exposure. Patients with advanced renal disease are at risk of developing nephrogenic systemic fibrosis (NSF) if administered with a Gd based contrast agent. NSF is a fibrosis disease which causes skin thickening which can reduce mobility and joint motion. May affect internal organs. Believed that slow clearance of Gd results in dissociation of Gd (toxic) from its chelate. Some contrast agents (Omniscan, Magnevist and OptiMARK) have been associated with more cases of NSF.1 MultiHance, Gadavist, Dotarem and ProHance have been associated with fewer, if any, cases.1 ACR Manual on Contrast Media, 2016

31 Screening for Renal Function ImpairmentIf a patient has a history of kidney disease, renal function should be assessed and a measure of the estimated glomerular filtration rate (eGFR) should be obtained. It is recommended to avoid administering Gd based contrast agents to patients with eGFR < 30 mL/min/1.73 m2.1 ACR Manual on Contrast Media, 2016

32 Retention of Gd in the BrainRetention of Gd in the brain has been found in patients with multiple (four or more) Gd contrast scans.1,2 Not related to kidney function.2 Unknown if harmful.2 Found to occur with the use of linear chelate based Gd contrast agents.3 1. Kanda et al. Radiology , v. 270, p. 834, 2014 2.http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm htm 3. Kanda et al. Radiology, v. 275, p. 803, 2015

33 MRI Quality Control ACR (American College of Radiology) has a breast MRI accreditation program with certain quality control requirements. Standard ACR MRI tests. Breast coil tests.

34 ACR MRI Quality Control“Large” ACR MRI phantom Inner dimensions: diameter 19 cm Length about 15 cm Phantom figure from:

35 Weekly technologists testsTable positioning Centre the phantom in the RF head coil and using the alignment laser. Centre the phantom at the magnet iso-centre. Acquire a T1 weighted sagittal spin echo image (FOV = 25 cm). FOV = 25 cm F H 12.5 cm ± 0.5 cm Phantom centre

36 Record the centre frequency and the transmitter gainWeekly fluctuations should be within limits set by the medical physicist. F H Image quality assessment from 11 axial slices (FOV = 25 cm) Check that images are free of artifacts

37 Geometric accuracy 190 mm ± 2 mm

38 High-contrast spatial resolutionHorizontal resolution Vertical resolution 1.1 mm 0.9 mm 1.0 mm

39 Low-contrast detectability

40 Visual checklist Verify that the following are electrically and/or mechanically stable: Patient bed Alignment laser Indicator lights RF room Signage monitors

41 Annual medical physicist testsMagnetic field homogeneity o spectrum Spectral peak method Diameter to match manufacturer specifications Measure FWHM and convert to ppm by dividing by (42.6*Bo) Compare to manufacturer specifications

42 Slice position accuracyF H If slice is centred exactly at the vertices then the bar lengths will be equal Bar length difference should be within 5 mm

43 Slice thickness accuracyTwo crossed ramps at 5.71 ° with respect to the slice Length of ramp appears as 10 x slice thickness Slice thickness = 0.2 x (top x bottom)(top + bottom)

44 Measure signal over 75 % of image Measure noise standard deviation RF coil checks Volume coils For volume coil test SNR, image uniformity and percent signal ghosting. SNR Axial slice from ACR phantom Measure signal over 75 % of image Measure noise standard deviation SNR = signal/(noise st. dev.)

45 Uniformity Adjust window and level to find brightest region Measure the mean signal in that region with a small ROI Repeat to find mean signal in darkest region Percent uniformity = 100 x (1- 𝑀𝑎𝑥 𝑅𝑂𝐼 −𝑀𝑖𝑛 𝑅𝑂𝐼 𝑀𝑎𝑥 𝑅𝑂𝐼+𝑀𝑖𝑛 𝑅𝑂𝐼 ) Should be greater than 87.5 % for 1.5 T and below (head coil).

46 Ghosting Measure mean signal Measure mean signal in each of the four noise ROIs Percent ghosting = 100 x | 𝐿𝑒𝑓𝑡+𝑅𝑖𝑔ℎ𝑡 −(𝑇𝑜𝑝+𝐵𝑜𝑡𝑡𝑜𝑚) 2 𝑥 𝑚𝑒𝑎𝑛 𝑠𝑖𝑔𝑛𝑎𝑙 | Should be less than 2.5 % (head coil and T1 weighted spin echo)

47 Breast coil Consist of multiple elements each with its own RF channel. A 2-L bottle filled with 10 mM CuSO4 can be inserted in each coil side. “Small” ACR phantom can also be used.

48 - Measure SNR from all elements collectively and if Breast coil testing should verify that all coil elements are functioning, that reasonable uniform images are obtained and that signal from the left and right breasts are comparable. - Measure SNR from all elements collectively and if possible from each individual element.

49 Questions? Concluding Remarks Overview of MRITechniques used for breast MRI Role of MRI in breast screening, diagnosis and treatment Contraindications to MRI Safety concerns associated with gadolinium based contrast agents MRI quality control Questions?