John A. Malko

Hepatic iron overload: diagnosis and quantification by noninvasive imaging.

The diagnostic efficacy of magnetic resonance (MR) and computed tomography (CT) for detection and quantification of hepatic iron was assessed in a series of patients under investigation for clinical or biochemical evidence of hepatic iron overload. Thirty patients underwent MR imaging (SE 30,60/1000 or SE 30,60/2000) at 0.5 Tesla with calculation of hepatic T2 and liver to paraspinous muscle signal intensity ratios. Twenty-nine patients also had measurement of hepatic attenuation on noncontrast CT images. Results of these imaging studies were correlated in all patients with quantitative iron determination from liver biopsy specimens. The best predictor of liver iron among parameters studied was the ratio of the signal intensities of liver and paraspinous muscle (L/M) on a SE 60/1000 sequence. Both MR using L/M ratios and CT were sensitive methods for detection of severe degrees of hepatic iron overload with 100% of patients with hepatic iron on biopsy > 600 Μg/ 100 mg liver dry weight detected on the basis of L/M <0.6 or CT attenuation >70 Hounsfield units (HU). The MR parameter, however, was more specific than CT (100 vs 50%) and showed a higher degree of correlation with quantitated hepatic iron from biopsy. T2 measurements showed poor correlation with hepatic iron, due to difficulty in obtaining precise T2 measurements in vivo when the signal intensity is low. None of the parameters utilized was sensitive for detecting mild or moderate degrees of hepatic iron overload.We conclude that MR and CT are sensitive techniques for noninvasive detection of severe hepatic iron overload, with MR providing greater specificity than CT. Lesser degrees of iron deposition, however, may go undetected by our current imaging techniques.

Manganese Dipyridoxyl Diphosphate: Effect of Dose, Time, and Pulse Sequence on Hepatic Enhancement in Rats

We used an animal model to investigate the hepatic enhancement characteristics of manganese dipyridoxyl diphosphate (MnDPDP) related to time, dose, and pulse sequence. The contrast doses selected were in the human tolerance range. Using an SE 300/15 pulse sequence, maximum mean hepatic enhancement of 45% (8 mumols/kg) and 58% (12 mumols/kg) over baseline was seen during a plateau maintained between 5 and 50 minutes postinjection in the 8 mumols/kg group, and between 10 and 90 minutes in the 12 mumols/kg group. This plateau was followed by a very gradual decline in hepatic enhancement. Using either 4 or 8 mumols/kg, there was a significant increase in postcontrast hepatic intensity on all relatively T1-weighted pulse sequences (spin echo [SE] 300/15, inversion recovery [IR] 1400/20/400, gradient echo [GE] 47/13/80 degrees, and GE 60/20/30 degrees) except GE 47/13/80 degrees at 4 mumols/kg. At 8 mumols/kg there was superior enhancement, with IR 1400/20/400 and SE 300/15, but at 4 mumols/kg there was no consistently superior sequence. None of the relatively T2-weighted pulse sequences (SE 2000/50, SE 2000/100, or GE 100/30/20 degrees) demonstrated a significant change in hepatic intensity using either dose of contrast. The data suggest that the best combination of dose, pulse sequence, and time for hepatic imaging with MnDPDP is 8 mumols/kg using heavily T1-weighted sequences 5 to 60 minutes following contrast administration.

Are hepatic and muscle T2 values different at 0.5 and 1.5 Tesla

In an effort to determine whether T2 values of liver and muscle change with increasing field strength, 144 abdominal MR examinations were retrospectively evaluated. These patients were evaluated with a dual echo T2-weighted spin-echo sequence. Eighty-two of the examinations were performed at 0.5 Tesla and 72 at 1.5 Tesla (T). Eleven of the patients were evaluated with both MR systems with the same sequences. T2 values were also obtained from a Fe NH4(SO4)2 12H2O phantom. The T2 values of liver decreased from 57.8 +/- 11.3 at 0.5 T to 43.7 +/- 8.3 at 1.5 T. The T2 values of muscle decreased from 44.2 +/- 9 at 0.5 T to 35.4 +/- 7.2 at 1.5 T. Patients who were examined on both systems also demonstrated a decrease in both liver and muscle T2 values. For concentrations in the range of hepatic T2s, the phantom demonstrated a decrease in T2 values from 0.5 to 1.5 T ranging from 20.3 to 23.4%. All the T2 changes were statistically significant (p less than .05). The findings suggest that T2 values may depend on field strength, or may vary due to other hardware-related differences.

Controlled Eddy Currents: Applications to MR Imaging

We describe how the artifact caused by eddy currents generated in free standing copper coils may be controlled and used to advantage in magnetic resonance (MR) imaging. The eddy currents distort the excitation field of the MR imager in the vicinity of the coil; the coils may then be used to reduce the signal intensity of tissues adjacent to them. Two examples of how this signal reduction may be used to advantage are given. In one example a coil was used to eliminate an aliasing artifact by removing the signal from an unwanted object. In another example a coil was placed on the anterior abdominal wall of a subject, thereby reducing the high-intensity signal from subcutaneous fat and the resulting ghosting from respiratory motion.

Renal corticomedullary junction. Performance of T1-weighted MR pulse sequences.

Inability to demonstrate the renal corticomedullary junction (CMJ) on magnetic resonance (MR) images has been reported in connection with several medical renal diseases. T1-weighted spin echo pulse sequences have been advocated to demonstrate a signal intensity difference between cortex and medulla. This study was undertaken to determine which of several T1-weighted spin echo (SE) and gradient echo (GE) sequences are better for delineation of the CMJ. The MR studies were performed at 0.5 Tesla on 27 normal volunteers. Multi-slice axial images of both kidneys were obtained in all subjects at each of the following five pulse sequences: SE 250/20, SE 500/30, SE 900/30, and GE 300/15 with 80 degrees and 64 degrees flip angles. Contrast/noise ratios were calculated for the signal intensity differences between cortex and medulla; the average standardized contrast/noise ratios ranked as follows: GE 300/15/80 degrees = 3.01 +/- 0.74, GE 300/15/64 degrees = 2.72 +/- 0.74, SE 250/20 = 2.02 +/- 0.33, SE 500/30 = 1.96 +/- 0.51, and SE 900/30 = 1.71 +/- 0.39. In addition, the five sequences for each patient were randomized and the images were independently ranked for delineation of CMJ by three MR radiologists. The cumulative subjective ranking for all observers from best to worst is as follows: SE 500/30, GE 300/15/80 degrees, GE 300/15/64 degrees, SE 900/30, SE 250/20. Although better contrast/noise ratios are achieved with the GE sequences and the more T1-weighted SE sequences, as a practical matter this does not seem to be the only significant factor when compared with the visual image evaluation by independent observers.(ABSTRACT TRUNCATED AT 250 WORDS)

A nonplanimetric technique for measuring fluid volumes using MR imaging--phantom results.

A nonplanimetric algorithm for calculating the volume of homogeneous fluids, using data from a single slice MRI scan, is discussed. The algorithm uses the fact that the total MR signal from a homogeneous fluid placed inside an MR scanner is directly proportional to the volume of the fluid. A simple ratio of fluid volumes and signal strengths thus allows the determination of an unknown fluid volume from a known fluid volume and the measured signals from the known and unknown volumes. Signal strengths are obtained from a single image of the known and unknown volumes, by a simple summation of pixel intensities in the image. The volume algorithm was tested on a 0.5 T clinical imager using fluid-filled volume phantoms (85–500 ml) that ranged in complexity from simple bulk volumes (flasks) to a mock-brain phantom composed of two fluid-filled bottles (to represent ventricles) embedded in a tangle of small-diameter fluid-filled tubes (to represent sulci). The known volume was a 5 ml syringe filled with the same fluid as contained in the phantoms. Volumes were calculated using various single-slice spin-echo pulse sequences and were in all cases found to be within 3% of the known volumes. The phantom results imply that intracranial CSF volumes might be determined by comparing signals derived from intracranial CSF with those from known volumes of CSF. Such a procedure would require the use of a pulse sequence that returns a signal from the CSF only, such as a very long TE spin-echo sequence. Among the questions still to be addressed are the possible differences between intracranial and in vitro CSF and the effects of CSF flow. (AJNR 8: 267–270, 1987)