Category: Journal Club

In the world of hybrid and integrated imaging systems, you can’t get much more incompatible than MR and X-ray, at least until now. I’d been hearing about work on such an x-ray/MR hybrid system for a few years now, and now there’s a paper by Fahrig et al in this month’s Medical Physics (R. Fahrig, Z. Wen, A. Ganguly, G. DeCrescenzo, J. A. Rowlands, G. M. Stevens, R. F. Saunders, and N. J. Pelc, “Performance of a static-anode/flat-panel x-ray fluoroscopy system in a diagnostic strength magnetic field: A truly hybrid x-ray/MR imaging system“, Med Phys 32, 1775-1784 (2005)) describing the performance characteristics of a flat panel fluoroscopy and open MRI blended together.

Talk about bleeding edge imaging. Not something that would have been possible just a few short years ago before flat panel image receptors were available. With MR, you get great soft tissue contrast, but you can’t see solid objects like bone, implants or wires. With fluoroscopy, you have decent resolution, good visibility of bone, implants, and wires/catheters, but soft tissue contrast isn’t so great.

Not entirely sure what kind of clinical demand there is for something like this, but from the quick glance I gave at the paper, there have been a number of clinical procedures performed on it already. I look forward to digging into this paper and finding out more about the system.

Abstract:

Minimally invasive procedures are increasing in variety and frequency, facilitated by advances in imaging technology. Our hybrid imaging system (GE Apollo™ flat panel, custom Brand x-ray static anode x-ray tube, GE Lunar high-frequency power supply and 0.5 T Signa SP™) provides both x-ray and MR imaging capability to guide complex procedures without requiring motion of the patient between two distant gantries. The performance of the x-ray tube in this closely integrated system was evaluated by modeling and measuring both the response of the filament to an externally applied field and the behavior of the electron beam for field strengths and geometries of interest. The performance of the detector was assessed by measuring the slanted-edge modulation transfer function (MTF) and when placed at zero field and at 0.5 T. Measured resonant frequencies of filaments can be approximated using a modified vibrating beam model, and were at frequencies well below the 25 kHz frequency of our generator for our filament geometry. The amplitude of vibration was not sufficient to cause shorting of the filament during operation within the magnetic field. A simple model of electrons in uniform electric and magnetic fields can be used to estimate the deflection of the electron beam on the anode for the fields of interest between 0.2 and 0.5 T. The MTF measured at the detector and the DQE showed no significant difference inside and outside of the magnetic field. With the proper modifications, an x-ray system can be fully integrated with a MR system, with minimal loss of image quality. Any x-ray tube can be assessed for compatibility when placed at a particular location within the field using the models. We have also concluded that a-Si electronics are robust against magnetic fields. Detailed knowledge of the x-ray system installation is required to provide estimates of system operation. ©2005 American Association of Physicists in Medicine

The second article again comes from the Journal of Nuclear Medicine and discusses the optimization of PET/CT protocols for large patients, something that is quickly becoming a problem in all imaging modalities. Large patients are the bane of all imaging modalities. Noisy, low contrast x-ray images and low count-density nuclear medicine images from increased attenuation and scatter make for images of marginal diagnostic quality.

Most PET imaging protocols call for a fixed amount of activity to be administered to the patient (typically 5-10 mCi) regardless of the patient’s weight. For thin to normal sized patients, this makes for decent images. For larger patients though, often the result is a noisy image. This is typically compensated for by increasing the imaging time per bed position, which of course increases the total imaging time, the possibility of patient motion artifacts and may not be tolerated well by all patients.

Going to a weight based method (0.21 mCi/kg in the paper) for determining the amount of activity to administer the patient along with a small increase in imaging time per bed position is one method proposed by the authors. The idea behind this is that larger patients get a larger dose (up to 20 mCi), which compensates for photons lost due to increased attenuation. This also means that the imaging time per bed position can be kept the same, or increased by a smaller amount. At the 5 min/bed position time recommended by the authors, this comes out to roughly 30-40 minutes/study.

Benjamin S. Halpern, MD, Magnus Dahlbom, PhD, Martin A. Auerbach, MD, Christiaan Schiepers, MD, PhD, Barbara J. Fueger, MD, Wolfgang A. Weber, MD, Daniel H.S. Silverman, MD, PhD, Osman Ratib, MD, PhD and Johannes Czernin, MD, “Optimizing Imaging Protocols for Overweight and Obese Patients: A Lutetium Orthosilicate PET/CT Study“, JNM 46: 603-607

Abstract:

High photon attenuation and scatter in obese patients affect image quality. The purpose of the current study was to optimize lutetium orthosilicate (LSO) PET image acquisition protocols in patients weighing ≥91 kg (200 lb).
Methods: Twenty-five consecutive patients (16 male and 9 female) weighing ≥91 kg (200 lb; range, 91-168 kg [200-370 lb]) were studied with LSO PET/CT. After intravenous injection of 7.77 MBq (0.21 mCi) of ¹⁸F-FDG per kilogram of body weight, PET emission scans were acquired for 7 min/bed position. Single-minute frames were extracted from the 7 min/bed position scans to reconstruct 1-7 min/bed position scans for each patient. Three reviewers independently analyzed all 7 reconstructed whole-body images of each patient. A consensus reading followed in cases of disagreement. Thus, 175 whole-body scans (7 per patient) were analyzed for number of hypermetabolic lesions. A region-of-interest approach was used to obtain a quantitative estimate of image quality. Results: Fifty-nine hypermetabolic lesions identified on 7 min/bed position scans served as the reference standard. Interobserver concordance increased from 64% for 1 min/bed position scans to 70% for 3 min/bed position scans and 78% for 4 min/bed position scans. Concordance rates did not change for longer imaging durations. Region-of-interest analysis revealed that image noise decreased from 21% for 1 min/bed position scans to 14%, 13%, and 11% for, respectively, 4, 5, and 7 min/bed position scans. When compared with the reference standard, 14 lesions (24%) were missed on 1 min/bed position scans but only 2 (3%) on 4 min/bed position scans. Five minute/bed position scans were sufficient to detect all lesions identified on the 7 min/bed position scans. Conclusion: Lesion detectability and reader concordance peaked for 5 min/bed position scans, with no further diagnostic gain achieved by lengthening the duration of PET emission scanning. Thus, 5 min/bed position scans are sufficient for optimal lesion detection with LSO PET/CT in obese patients.

It’s been a while since the last journal club article. Partly because I hadn’t come across too many articles I thought were interesting enough, mostly because I haven’t had much time to do much journal reading lately. And now from this month’s Journal of Nuclear Medicine come two articles that I found very interesting and informative.

The first one looks at radiation exposures to patients from combined PET/CT scans, an increasingly popular (and quickly becoming ‘standard of care’) method of diagnosing cancer and monitoring therapy efficacy. CT and PET radiation doses were examined at 4 hospitals each employing a variety of techniques: low-dose CT for attenuation correction, diagnostic CT for attenuation correction and localization, contrast and non-contrast studies. Total radiation dose came out to around 25 mSv (about 7 mSv from PET, 18 mSv from CT) and was surprisingly mostly independent of the protocol used.

A handy table of dose coefficients for various organs is also provided, which will make it easy to estimate the radiation dose to various organs from an exam given the injected activity and CTDI from the CT scan.

What was not clear was if any of the scanners had any of the CT dose reduction methods being used in the newest scanners (the ones that dynamically adjust tube current throughout the scan). These have been shown to effectively reduce patient dose while maintaining a desired image quality. I’m sure these methods incorporated into the newest PET/CT units can bring down the radiation dose a little more.

This ought to be a useful paper for any medical physicist or radiologist finding the need to estimate radiation dose from a PET/CT scan, or wanting to optimize their protocols to minimize dose.

Gunnar Brix, PhD, Ursula Lechel, MS, Gerhard Glatting, PhD, Sibylle I. Ziegler, PhD, Wolfgang Münzing, PhD, Stefan P. Müller, MD and Thomas Beyer, PhD, “Radiation Exposure of Patients Undergoing Whole-Body Dual-Modality ¹⁸F-FDG PET/CT Examinations“, JNM 46 608-613 (2005)

Abstract:

We investigated radiation exposure of patients undergoing whole-body ¹⁸F-FDG PET/CT examinations at 4 hospitals equipped with different tomographs. Methods: Patient doses were estimated by using established dose coefficients for ¹⁸F-FDG and from thermoluminescent measurements performed on an anthropomorphic whole-body phantom.
Results: The most relevant difference between the protocols examined was the incorporation of CT as part of the combined PET/CT examination: Separate low-dose CT scans were acquired at 2 hospitals for attenuation correction of emission data in addition to a contrast-enhanced CT scan for diagnostic evaluation, whereas, at the other sites, contrast-enhanced CT scans were used for both purposes. Nevertheless, the effective dose per PET/CT examination was similar, about 25 mSv.
Conclusion: The dosimetric concepts presented in this study provide a valuable tool for the optimization of whole-body ¹⁸F-FDG PET/CT protocols. Further reduction of patient exposure can be achieved by modifications to the existing hardware and software of PET/CT systems.

A news article titled “Patients, physicians unaware of CT radiation exposure” in Radiology 2004;231:393-398 and posted at Aunt Minnie talks about just how much patients and physicians don’t know about radiation exposures from CT.

The article raises a some interesting points about the need to educate health professionals about radiation. About the only radiology education most physicians receive is a 3 or 4 month rotation through radiology where they learn what it is they’re looking at when looking at an x-ray. Even some radiologists don’t know as much about radiation as you might think. The trouble is, squeezing more education into the curriculum is difficult at best.

Yes, radiation exposures from CT are significantly higher than regular radiographic exposures. Yes, there might be an increase in the risk of getting cancer. It’s a small risk but one that is dwarfed by risks from other activities such as smoking or environmental exposure to chemical carcinogens. Radiation can be bad for you, but saves lives when used properly.

Ultimately, when it comes to any diagnostic imaging involving radiation, it comes down to a risk versus benefit thing. This is one of the things I try to emphasize to residents when I’m teaching them. Does the benefit of diagnosing an immediate problem in the patient outweigh the small increased long term cancer risk? Will the exam provide significantly more information useful for treatment than another exam that uses less or now radiation?

For the past 7 years I’ve been involved in teaching radiology residents all about the physics of medical imaging. I teach them (try to anyway) about radiation, how it interacts with matter, the different imaging modalities and how they work and generate images, radiation safety and biology. It’s part of the board exams they need to pass. But I don’t teach them because they have to pass the board exam. I teach them because down the road, they will be the “radiation experts” in whatever community they end up working in. They’re the ones that people expect to know about the risks and benefits of getting radiation. They’re the ones that people will go to for answers.

And my education efforts go beyond the classroom as well. Dose/exposure charts for all the equipment I survey are generated and distributed for posting so that people can see (if they care to look) what kinds of exposures the machine produces. I probably need to advertise this a little more though.

It’s one of the activities I enjoy about my job.

Abstract:

PURPOSE: To determine the awareness level concerning radiation dose and possible risks associated with computed tomographic (CT) scans among patients, emergency department (ED) physicians, and radiologists.
MATERIALS AND METHODS: Adult patients seen in the ED of a U.S. academic medical center during a 2-week period with mild to moderate abdominopelvic or flank pain and who underwent CT were surveyed after acquisition of the CT scan. Patients were asked whether or not they were informed about the risks, benefits, and radiation dose of the CT scan and if they believed that the scan increased their lifetime cancer risk. Patients were also asked to estimate the radiation dose for the CT scan compared with that for one chest radiograph. ED physicians who requested CT scans and radiologists who reviewed the CT scans were surveyed with similar questions and an additional question regarding the number of years in practice. The 2 test of independence was used to compare the three respondent groups regarding perceived increased cancer risk from one abdominopelvic CT scan.
RESULTS: Seven percent (five of 76) of patients reported that they were told about risks and benefits of their CT scan, while 22% (10 of 45) of ED physicians reported that they had provided such information. Forty-seven percent (18 of 38) of radiologists believed that there was increased cancer risk, whereas only 9% (four of 45) of ED physicians and 3% (two of 76) of patients believed that there was increased risk (X²₂ = 41.45, P < .001). All patients and most ED physicians and radiologists were unable to accurately estimate the dose for one CT scan compared with that for one chest radiograph.
CONCLUSION: Patients are not given information about the risks, benefits, and radiation dose for a CT scan. Patients, ED physicians, and radiologists alike are unable to provide accurate estimates of CT doses regardless of their experience level.
© RSNA, 2004

An article published in the April 28 issue of JAMA tries to make a link between dental radiation and infant low birth weight.

It’s been making the news lately and stirring up a lot of fuss.

At first I didn’t think much of it and wasn’t going to pay much attention to it. Just another study trying to link tiny insignificant exposures of radiation to a much larger problem (correlated effect vs causal relationship). But after seeing it make the news I thought I’d take a closer look and see what all the commotion was all about.

The main hypothesis behind the paper is that dental x-rays result in non-trivial radiation dose the the thyroid, which affects thyroid function, which in turn has an effect on infant birth weight. There is no radiation exposure to the fetus from dental radiation.

One thing I have to take issue with is their radiation dosimetry. There doesn’t appear to have been any effort to do any dose reconstruction for the study. Rather, the authors looked at what dental radiation procedures people in the study got, looked up radiation doses for those procedures from a 1993 NEXT survey of dental exposures and said “This is the radiation exposure these patients got”. And if a patient had multiple x-ray procedures, they just got added up.

Now, NEXT is a fine tool to see how x-ray exposures compare for different types of procedures or equipment. The FDA (or state delegates) sends out their inspectors to randomly selected hospitals and clinics every now and then, they make some radiation measurements with various phantoms, and all of it gets compiled into the survey report.
Have a look at the 1999 NEXT Dental survey. Intraoral skin entrance exposures using D speed film (page 9) averages 194.6 mR, but with a seriously wide distribution (standard deviation 103.5). If dental units in Washington State (where the study population comes from) actually produce lower exposures than what’s published in the NEXT survey, then radiation exposures are seriously overestimated in the study.

Would more accurate dosimetry affect the results of the study? Not sure. I’m also not too clear on how the authors accounted for other risk factors associated with low birth weight (smoking, alcohol consumption, hypertension, etc).

I think the most interesting table of the paper is Table 2 which lists the dental procedures, and the numbers of low birth weight (LBW), term low birth weight (TLBW) and normal birth weight (NBW) infants. The last two columns give P values for comparisons between LBW/NBW and TLBW/NBW, although I’m not sure if this is a Student’s T-test correlation or some other type of test. It’s not stated in the paper as far as I can see. A strong correlation between LBW/TLBW vs NBW is shown for thyroid exopsures > 0.4 mGy, although n is very low for this group. But there is also a strong correlation for TLBW vs NBW for no radiation exposure (P = 0.02) where n is much larger for both groups.

As for the main hypothesis, the study does give some support to the idea that dental radiation might have an effect on thyroid function, affecting birth weight. I don’t think it’s anywhere near definite though. I think to really see something, you’d have to do thyroid function tests and check hormone levels on a whole lot of people receiving dental x-rays and be able to accurately measure the radiation dose to the thyroid in order to get any sort of valid correlation. I think it’s more likely that dental radiation to the thyroid is a contributing factor rather than a causal factor in low birth weight, acting in conjunction with other low birth weight risk factors.

Abstract:

Context Both high- and low-dose radiation exposures in women have been associated with low-birth-weight offspring. It is unclear if radiation affects the hypothalamus-pituitary-thyroid axis and thereby indirectly birth weight, or if the radiation directly affects the reproductive organs.

Objective To investigate whether antepartum dental radiography is associated with low-birth-weight offspring.

Design A population-based case-control study.

Participants and Setting Enrollees of a dental insurance plan with live singleton births in Washington State between January 1993 and December 2000. Cases were 1117 women with low-birth-weight infants (<2500 g), of whom 336 were term low-birth-weight infants (1501-2499 g and gestation 37 weeks). Four control pregnancies resulting in normal-birth-weight infants (2500 g) were randomly selected for each case (n = 4468).
Main Outcome Measures Odds of low birth weight and term low birth weight by dental radiographic dose during gestation.

Results An exposure higher than 0.4 milligray (mGy) during gestation occurred in 21 (1.9%) mothers of low-birth-weight infants and, when compared with women who had no known dental radiography, was associated with an adjusted odds ratio (OR) for a low-birth-weight infant of 2.27 (95% confidence interval [CI], 1.11-4.66, P = .03). Exposure higher than 0.4 mGy occurred in 10 (3%) term low-birth-weight pregnancies and was associated with an adjusted OR for a term low-birth-weight infant of 3.61 (95% CI, 1.46-8.92, P = .005).

Conclusion Dental radiography during pregnancy is associated with low birth weight, specifically with term low birth weight.

Antepartum Dental Radiography and Infant Low Birth Weight JAMA. 2004;291:1987-1993