Effective Dose of CT- and Fluoroscopy-Guided Perineural/Epidural Injections of the Lumbar Spine: A Comparative Study
The objective of this study was to compare the effective radiation dose of perineural and epidural injections of the lumbar spine under computed tomography (CT) or fluoroscopic guidance with respect to dose-reduced protocols. We assessed the radiation dose with an Alderson Rando phantom at the lumbar segment L4/5 using 29 thermoluminescence dosimeters. Based on our clinical experience, 4–10 CT scans and 1-min fluoroscopy are appropriate. Effective doses were calculated for CT for a routine lumbar spine protocol and for maximum dose reduction; as well as for fluoroscopy in a continuous and a pulsed mode (3–15 pulses/s). Effective doses under CT guidance were 1.51 mSv for 4 scans and 3.53 mSv for 10 scans using a standard protocol and 0.22 mSv and 0.43 mSv for the low-dose protocol. In continuous mode, the effective doses ranged from 0.43 to 1.25 mSv for 1–3 min of fluoroscopy. Using 1 min of pulsed fluoroscopy, the effective dose was less than 0.1 mSv for 3 pulses/s. A consequent low-dose CT protocol reduces the effective dose compared to a standard lumbar spine protocol by more than 85%. The latter dose might be expected when applying about 1 min of continuous fluoroscopy for guidance. A pulsed mode further reduces the effective dose of fluoroscopy by 80–90%.
KeywordsComputed tomography (CT), guidance Computed tomography (CT), radiation exposure Fluoroscopy Phantoms Spine, lumbar Intervention Thermoluminescence dosimeter
Epidural and perineural injections with local anaesthetics and corticosteroids are widely used in the conservative treatment of low back pain and radicular pain due to degenerative disease of the spine and disc herniation [1, 2]. Although experienced clinicians might perform this procedure without image guidance, the use of fluoroscopy or computed tomography (CT) might improve the quality of the procedure by directing the injection properly . The procedure is performed using a direct approach to the epidural space or the neuroforamen at the lumbar spine or through the sacral hiatus. Renfrew et al.  published a prospective evaluation of 316 caudal-approach epidural steroid injections via the hiatus sacralis. Correct nonfluoroscopically guided placement of the needle varied from 48% to 62% depending on the experience of the physician. Even with correct needle placement in the sacral canal, there was venous injection in 9.2% of the procedures. Computed tomography (CT) is a perfect tool to guide the needle to the epidural and perineural space, but the radiation dose is a serious concern. It was shown from a dose survey of the British National Radiological Protection Board in 1989 that CT contributes approximately 20% to the collective dose of the population from diagnostic imaging even though only 2% of the radiological examinations are CT scans . In single departments, the collective dose from CT might be as high as 67%. For injection therapy under CT guidance, it is possible to scan in a low-dose mode with significant mAs reduction, because bone and soft tissue differentiation still is sufficient for needle guidance. Low-dose protocols can significantly reduce the effective dose to the patient. New techniques in fluoroscopy such as pulsed fluoroscopy have also led to further reduction of radiation to the patient and the physician .
The goal of this study was to compare effective doses in lumbar spine injection therapy under CT guidance with standard and low-dose protocols to fluoroscopy-guided procedures with continuous and pulsed fluoroscopy.
Materials and Methods
Alderson Rando Phantom
We used an Alderson Rando Phantom (The Phantom Laboratory, Salem, NY) according to the ICRU-44 standard, which represents a man of 1.75 m height with a weight of 73.5 kg. In the axial plane, the phantom is cut in 34 slices with a thickness of 2.5 cm. In these slices, small holes with a diameter of 2 mm and a depth of 6 mm were drilled, where the thermoluminescence dosimeters were placed. The distance between these holes was 1.5 cm.
We used lithium fluoride thermoluminescent dosimeters (TLDs), series 100H (Harshaw, Solon, OH, USA) with diameters of 1 × 1 × 6 mm. They are able to detect photons with more than 5 keV, neutrons >100 keV, and β-radiation >70 keV. The dose that can be measured linearly ranges from 10 μGy to 1 Gy. The linear reading process can be adjusted between 5 and 100 s, with a temperature between 0°C and 400°C.
Before starting the experiments, we performed a CT scan of the whole Alderson Rando phantom to determine the relevant organs for dose calculation. The distribution of the TLDs to organ regions was made on the basis of the scheme proposed by Huda und Sandison . We labeled the drilling holes to ensure that each individual TLD was placed at the identical site in every measurement.
We used 29 TLDs for all measurements. Each radiation sensitive organ according to IRCP Report Nr. 60  was represented by at least one TLD. The organs and tissues belonging to the “remainder” located next to the radiated field were armed with TLDs as well. Red bone marrow is regarded as being highly sensitive to irradiation, with a weighted organ factor of 0.12. Therefore, we specifically evaluated the organ dose of active bone marrow in different parts of the skeleton. TLDs were placed in one thoracic and in one lumbar vertebra, pelvis, rib, scapula, and clavicle. The total organ dose (H) of active bone marrow was calculated using the following formula:
This formula takes into account differences of the distribution of red bone marrow in these bones . Considering the CT examination, we expected a steep decrease of the absorbed dose at both sides of the collimated X-ray beam, so some organs will be only partially involved. In order to avoid overestimation of the total absorbed dose, these organs were armed with multiple TLDs in different locations, and we calculated the organ dose as an average of the single measurements within the specific organ. For organs that were only partially irradiated, like the colon and the skin, special care was taken to estimate the irradiated mass fraction of these organs. The effective dose in females was estimated by positioning TLDs to the position of the ovaries and the breast. The center slice for the CT measurements was located at the L4/5 level of the Rando phantom (slice 28), where we simulated the therapeutic injection.
Each TLD chip has its own sensitivity factor, and we calculated the absorbed dose on the basis of the individual calibration factor of each TLD, which was determined using a 1-MeV cobalt-60 source. The reading of the TLDs was performed 1 day after the experiment. The reading included a 10-s preheating where the TLD was exposed to a constant temperature of 100°C. This was followed by the reading process in which the light intensity is measured while the TLDs were heated up to 240°C in steps of 10°C per second. The measured charge of the TLDs in nanocoulombs was multiplied by the individual calibration factor for each TLD to obtain the energy dose (in mGy). Each reading was followed by a “post-read-anneal process,” heating the TLDs up to 400°C to remove any remaining charge. The effective dose was calculated on the basis of the measured organ dose and the individual weighting factors supplied by ICRP.
Epidural/Perineural Injection Simulation
The segment L4/5 is the most mobile segment of the lumbar spine; thus, it is the most frequently affected segment with regard to a disc herniation. Therefore, we have chosen this segment for simulation of the injection therapy.
All examinations were performed on a fourth-generation Somaton Plus 4 spiral CT scanner (Siemens, Erlangen, Germany). Tube voltage can be adjusted to 80 kV, 120 kV, or 140 kV, the tube current being adjustable from 50 to 420 mA. Slice thickness can be chosen between 1 and 10 mm. The field of view is 500 mm.
Standard Lumbar Spine Protocol (CT)
The Rando phantom was brought into prone position. A short lateral scout scan of the lower lumbar spine was obtained with the following parameters: length, 256 mm; scan time, 2.9 s; tube voltage, 120 kV; tube current, 111 mA; slice thickness, 3 mm; gantry tilt, 0°. To ensure identical repositioning, the scout scan always started with the upper border of Rando slice No. 23. An axial slice was planned with a minimal gantry tilt of 1.5°. Using this tilt, the radiation beam was parallel to Rando slice No. 28 in prone position. This slice represented the segment L4/5 at the level of the neuroforamen. The scan was centered in the upper third of slice No. 28, where the drilling holes are placed. By doing so, we also made sure that the only 6-mm-long TLDs were completely within the radiation beam. Table movement was set to zero so that all scans were made at identical positions. Scan parameters (scan time, 1.5 s; tube voltage, 140 kV; 146 mA; 219 mAs per scan) were taken from our nonspiral routine lumbar spine protocol, except that we used a slice thickness of 8 mm rather than the standard 3 mm. A thicker slice is useful for the visualization of the needle. In our own experience with a large number of epidural/perineural injections at the lumbar spine, we found that approximately four axial scans are necessary to do a complete epidural/perineural injection: The first scan allows for adequate anatomic orientation, the second visualizes the needle halfway down, the third documents correct needle positioning in the epidural or perineural space, and the fourth scan demonstrates epidural distribution of contrast media. In patients with more complicated access to the neuroforamen, the number of scans will increase but rarely will exceed 10 scans. We repeated these measurements five times using the same setting to obtain results with adequate reproducibility. The resulting values in nanocoulombs are given as an average of the five measurements.
Low-Dose Protocol (CT)
We performed fluoroscopy on a Polystar T.O.P. (Siemens, Erlangen, Germany), a C-arm multifunctional fluoroscopy unit with the ability to switch the position of the X-ray tube between over and under the table. The photomultiplier unit is a Sirecon 40-4 HDR 4 with an adjustable field of view (FOV) (40 cm, 30 cm, 22 cm, and 17 cm); the X-ray generator unit (SiemensPolydoros SX 80) allows for continuous or generator-pulsed fluoroscopy without grid control.
Comparision of the effective dose using a standard lumbar spine protocol and a maximum low-dose protocol for simulated CT-guided injection therapy of the lumbar spine
No. of scans
Effective dose (mSv) standard protocol
Effective dose (mSv) low-dose protocol
Effective doses under continuous fluoroscopy
Effective doses (mSv) male
Effective doses (mSv) female
In the male phantom, the average effective dose was 0.43 mSv for 1 min of continuous fluoroscopy and 0.81 and 1.25 mSv for 2 and 3 min, respectively (Table 2). The effective dose in the female phantom was slightly higher.
Effective doses for pulsed fluoroscopy (1 min)
Effective dose (mSv) male
Effective dose (mSv) female
Sciatica and low back pain due to degenerative spine disease are common disorders representing an important medical and socioeconomic problem. Ever since the first report by Evans in 1930, epidural injections for the treatment of back pain and/or radicular pain are widely used . The method has been further refined, and, currently, corticosteroid injection in the epidural space can be regarded as one main treatment option in the conservative management of sciatica and low back pain [1, 11, 12, 13]. Being an efficient short-term treatment option, long-term effects are discussed controversially . Imaging techniques to guide the approach to the epidural space help avoid incorrect needle placement and subsequent injection of the steroids and the local anaesthetics into undesired locations, such as the subarachnoid space [4, 11]. Correct needle placement and accurate delivery of the drugs to the compressed nerve roots are crucial steps with respect to the safety and efficacy of this procedure, and many physicians perform the procedure under fluoroscopic [1, 3] or CT guidance [12, 15]. CT guidance offers a fast and accurate approach to the epidural space via the interlaminar route. With dedicated techniques, the corticosteroids can even be placed at the site of nerve root compression in the ventral epidural space . CT guidance also offers advantages in patients with previous spine surgery now suffering from epidural fibrosis. Nevertheless, there is some concern about the higher radiation dose associated with CT scanning as compared to fluoroscopy. This is of special interest, because patients treated for low back pain and sciatica are often quite young. Moreover, therapeutic injection will often be repeated several times to achieve maximum treatment benefit. The effective dose of abdominal CT is around 6 mSv, varying between 2 and 20 mSv [16, 17, 18]. New CT scanners allow for maximum dose reduction while providing an image quality that is still sufficient for visualization of the needle and the distribution of the injected contrast media. Thus, the applied dose might be significantly reduced.
Fluoroscopy does not contribute as much as CT to the general radiation exposure of man ; still, single examinations might reach or exceed the effective doses measured with standard CT protocols. For example, biphasic colon examination is known to be one of the fluoroscopic examinations with the highest radiation exposure in diagnostic radiology. As Geleijns and co-workers have shown , the mean effective dose is 4.7 mSv (range: 2.7–8.4 mSv) and it roughly contributes 15% to the collective dose resulting from medical diagnostic radiology. In recent years, some manufacturers of fluoroscopy units have introduced pulsed fluoroscopy for dose reduction. The two basic techniques are generator-controlled pulsed fluoroscopy and grid-controlled pulsed fluoroscopy. The main difference is that with the grid-controlled technique, the pulses originate in the X-ray tube rather than the X-ray generator. The resulting pulses in the grid-controlled technique are in the form of discrete and uniform current without ramping and trailing of the pulse and, therefore, without wasteful low-energy radiation . Pulsed fluoroscopy has proved to reduce fluoroscopy time and consecutive radiation dose significantly. Scanavacca and co-workers  have shown that pulsed fluoroscopy can reduce radiation exposure time to 80% during cardiac catheter ablation without increasing procedure duration and without decreasing success rate. Similar results were found by Nikolic et al.  comparing pulsed and continuous fluoroscopy together with other radiation-reducing techniques in uterine artery embolization. They could reduce the radiation dose to less than 50%. Hernandez and Goodsitt  reported a dose reduction of 75% in a pediatric population comparing pulsed to continuous fluoroscopy. This markedly reduced radiation exposure does not necessarily result in diminished image quality. In a series of different abdominal fluoroscopy procedures, Boland et al.  found no significant difference between pulse rates with 3.75, 7.5, and 15 frames/s and continuous fluoroscopy. Vetter and co-workers  could show that although radiation dose was reduced between 10% and 46% in grid-controlled fluoroscopy, the image quality of the last image hold was improved compared to the continuous mode. Apart from dose reduction through pulsed fluoroscopy, there are many other factors influencing radiation dose. Among the most important, image magnification and oblique views should be prevented, whenever possible exposures should be replaced by last image holds, and the operator should be familiar with the technical parameters of his fluoroscopy unit [25, 26].
In a survey of 288 noncoronary diagnostic and interventional procedures under fluoroscopic control, McParland  has estimated the effective dose to be between 3.9 and 12 mSv for interventions in the abdomen (nephrostomy, biliary stent, percutaneous transhepatic cholangiography). Teeuwiise and co-workers reported effective doses of 13.5 mSv with conventional CT scanners and 9.3 mSv with spiral CT fluoroscopy in a set of 16 drainages, 8 mSv and 6.1 mSv in 49 biopsies, and 2.1 mSv and 0.8 mSv in 31 coagulations of osteoid osteomas, respectively . Few studies are available comparing the effective dose of examinations under CT and under fluoroscopy control for a single intervention. Slomczykowski et al.  compared radiation dose in pedicle screw insertion with CT navigation and a fluoroscopic method. They found a higher effective dose for the CT procedure even when using optimized protocols. A similar study was performed by Schaeren and Dick , who found a 15 times higher effective dose with CT-based navigation (7.27 mSv) versus fluoroscopy (0.48 mSv) for pedicle screw instrumentation.
We have chosen 1 min as an appropriate time to simulate fluoroscopy from our experience of the duration of the last 20 diagnostic and therapeutic radiculographies of the lumbar spine performed by our orthopedic surgeons. This relative long duration of fluoroscopy is mainly due to our status as a teaching hospital and can probably be optimized, as other studies have described median screening times of 15–50 s [31, 32] in similar procedures.
Our own results show that with standard protocols that are not optimized for a low radiation dose, the radiation exposure of CT-guided epidural/perineural injections is much higher than with 1 min of fluoroscopy and is similar to what has been described for pedicle screw instrumentation by Schaeren and Dick . With low-dose protocols as could be performed on our CT scanner, the effective dose applied to the patient decreased to values between 0.22 and 0.43 mSv for 4 scans and 10 scans, respectively. Compared to the standard CT protocol, dose reduction is approximately 86–89%. The effective dose of 1 min of continuous fluoroscopy had a similar effective dose of 0.43 mSv; thus we conclude that low-dose CT and continuous fluoroscopy are comparable with respect to radiation exposure in lumbar spine injection therapy. In institutions where pulsed fluoroscopy is available, this technique can further reduce the effective dose down to less than 0.1 mSv, representing a further reduction of the radiation dose by more than 80%. However, in difficult cases or in other locations (e.g., cervical or thoracic spine), CT guidance might be preferred because of its better resolution and precision.
A limitation of the study is that under fluoroscopy, the influence of oblique views, which might be used for perineural injection, a potentially increasing radiation dose was not examined. The radiation dose added by taking radiographs to document correct needle placement and contrast-agent distribution in fluoroscopically guided procedures was not considered in this study. Although we have shown that image quality of our low-dose protocols is sufficient to perform the intervention and we do use these protocols in our daily routine, this study does not systematically evaluate image quality. We did not examine epidural injection therapy under CT fluoroscopy guidance in this article because we did not have this option on our scanner. CT fluoroscopy is known to have a higher radiation dose than conventional CT in interventional procedures ; therefore, it is likely that CT fluoroscopy also has higher effective doses in epidural injection therapy.
In conclusion we found that a consequent low-dose CT protocol reduces the effective dose compared to a standard lumbar spine protocol by more than 85%. The radiation dose of a low-dose CT protocol is comparable to that of 1 min of continuous fluoroscopy. Fluoroscopy in the pulsed mode might further reduce the effective dose compared to the continuous mode by 80–90%.
- 6.Boland GW, Murphy B, Arellano R, et al. (2000) Dose reduction in gastrointestinal and genitourinary fluoroscopy: use of grid-controlled pulsed fluoroscopy. Am J Roentgenol 175(5):1453–1457Google Scholar
- 8.International Commission on Radiological Protection (1990) 1990 Recommendations of the International Commission on Radiological Protection. Pergamon. OxfordGoogle Scholar
- 9.Anon. (1976) Environmental radioactivity and radiation exposure. Annual Report. The Secretary of State for Home Affairs of GermanyGoogle Scholar
- 10.Evans W (1930) Intrasacral epidural injection therapy in the treatment of sciatica. Lancet 2:1225–1229Google Scholar
- 13.CT guided perineural injection technique. In: Kraemer J, Koester K (eds) MR Imaging of the Lumbar Spine. A Teaching Atlas. Thieme: New York, pp 18–26 Google Scholar
- 19.Galanski M, Nagel HD, Stamm G (2001) CT-Expositionspraxis in derBundesrepublik Deutschland: Ergebnisse einer bundesweiten Umfrage im Jahr 1999. Fortschr Röntgenstr 173:1–66Google Scholar
- 22.Nicolic B, Spies JB, Campbell L, et al. (2001) Uterine artery embolization: reduced radiation with refined technique. J Vasc Intervent Radiol 12(1):39–44Google Scholar
- 23.Hernandez RJ, Goodsitt MM (1996) Reduction of radiation dose in pediatric patients using pulsed fluoroscopy. Am J Roentgenol 167:1247–1253Google Scholar
- 24.Vetter S, Heckmann H, Strecker EP, et al. (1998) Klinische Aspekte zu Bildqualität und Dosis bei gittergesteuerter gepulster Durchleuchtung. Akt Radiol 8:191–195 Google Scholar
- 26.Nicolic B, Abbara S, Levy E, et al. (2000) Influence of radiographic technique and equipment on absorbed ovarian dose associated with uterine artery embolization. J Vasc Intervent Radiol 11(9):1173–1178Google Scholar