Image-guided LP in the radiology department is performed for one of four reasons.

The principal indication for an image-guided LP is a failed bedside attempt or the belief that a bedside attempt will be unsuccessful. Ordering providers may not be adequately trained in LP technique or may have not sought credentialing, opting to send the patient for an image-guided procedure. Occasionally patients request image guidance. Typical factors contributing to a failed bedside procedure are obesity, severe degenerative disk disease, or scoliosis.

In the current health care climate, it is important to note that LP is a procedure with low reimbursement and that scheduling a procedure room with fluoroscopic guidance and a dedicated radiology technician comes at considerable cost over the bedside procedure. Fluoroscopic image guidance can add several levels of complexity to the procedure of LP, requiring proper placement of the order with the radiology department, communication of the order to the staff member who will perform the procedure, scheduling of the procedure room, and transport of the patient. Because LP is poorly reimbursed, hospitals rarely have dedicated facilities for this procedure, and LP cases may compete with complex neurointerventional cases for access to expensive angiography suites. Alternatively, fluorography room time must be negotiated with other divisions of the radiology department. Finally, fluoroscopy entails a radiation dose, which may be unnecessary. For these reasons, the bedside procedure remains the first line of approach, and the radiology department typically requires that a bedside attempt be made before the image-guided procedure is undertaken.

Fluoroscopically guided LP entails a review of the patient's medical history and coordination of a significant number of hospital personnel. Completion of a preprocedure checklist (Appendix 1) can expedite preparation and can ensure a safe procedure.

Medically unstable patients such as patients receiving mechanical ventilation should be accompanied by emergency radiology or ICU personnel who can monitor vital signs. Outpatients should have someone available to drive them home after they have been released from the postprocedure recovery area. In-house staff are generally aware that patients should take nothing by mouth 2 hours before LP, and outpatients should be contacted by radiology support staff with procedure guidelines, which includes asking standard questions regarding major medical conditions, medications, and allergies.

Diagnostic myelography remains a procedure performed only by a diagnostic neuroradiologist. An intrathecal contrast injection is followed by CT of the spine in the area of interest (Fig. 2). The diagnostic study is generally interpreted by the radiologist who performed myelography.

In the absence of metallic interference, MRI is superior to myelography for the visualization of the contents of the spinal canal because MRI allows direct visualization of the spinal cord and nerve roots. Myelography is an invasive technique and offers little image detail or soft-tissue contrast. Therefore, MRI is generally the imaging modality of choice, and myelography is used only to address very specific questions.

Older fluoroscopy equipment often has one fixed overhead camera. Imaging in radiology is best performed with two orthogonal views: one to assess the target and the other to assess depth. The fixed overhead camera can be complemented by adding a second camera that allows a lateral view or by equipping the facility to obtain a cross-table lateral view. Newer facilities offer a biplane camera, which can greatly facilitate the ability to assess a lateral view. State-of-the-art angiography suites can have multiple biplane cameras and large imaging screens. Generally, however, diagnostic radiologists are not familiar with advanced angiography facilities, and the basic biplane camera is adequate for fluoroscopic guidance of LP.

Several different approaches are common for fluoroscopically guided LP and are often a matter of preference [35]. With any approach, proper patient positioning is necessary to optimize the likelihood of procedure success. The invasive portion of the procedure should not be attempted until the patient is stably and comfortably positioned and the target for access to the spinal canal is clearly identified on fluoroscopy. Time and effort invested in optimal patient positioning can greater simplify the invasive portion of the procedure.

A target for puncture access is chosen to be below the conus tip, but the canal narrows toward the sacrum and the distance from the skin to the thecal sac increases. The conus tip is typically at the L1 level but can be verified by assessing previous imaging studies if any are available. Given these factors, the optimal levels are L2–L3 and L3–L4. Ultimately the level is chosen by fluoroscopic inspection as the level that appears most un-obstructed by osteophytes, degenerative loss of disk height, or changes from previous surgery. The ideal level has a vertebral body as the “backstop” rather than a disk, and the bone ensures that the needle tip is not positioned too deeply.

LP is difficult for the patient but requires patient cooperation. Patients who are fearful or hesitant may require sedation. Patients who move during the procedure can pose risk to both themselves and the physician performing the procedure. I find that a continued dialogue with patients can be reassuring to them. The feeling of loss of control can be disturbing to patients, and it may be helpful to reassure anxious patients that they can choose to discontinue the procedure if they wish to do so.

Needle options are generally 20- and 22-gauge. A standard needle is 3.0 inches (7.6 cm) long; LP of obese patients may require a longer needle (5.5 inches, 14.0 cm). Smaller-caliber needles have been shown to result in a lower incidence of post-LP headache; needles of higher gauge than 22-gauge are easily bent during LP and are not typically used [39]. The tip of the needle has a bevel that can be used to help steer the needle. The stylette should always be fully in place when advancing or removing the needle to prevent cutting a nerve root or creating a suction to the nerve root when the needle is removed.

Efforts should be made to minimize the fluoroscopy radiation dose to the patient and health care personnel. Studies in the literature report a wide range of radiation dose values for LP, showing radiation doses from LP to be comparable to those from lumbar CT, but these radiation values are no doubt on the high end because they are reported for obese patients [40]. In addition, a wide range of fluoroscopy times have been reported [40, 41].

Certainly a number of techniques can serve to dramatically reduce fluoroscopy time and radiation dose. First, use imaging only to guide the procedure. Avoid continuous fluoroscopy; instead, use spot images. If possible, use the Screen Capture mode so that the image is retained on the screen after fluoroscopy has ended.

Second, cone the image to the necessary FOV. Magnification can be helpful but increases the radiation dose to the smaller area that is irradiated.

Third, remember that recording an image for the PACS requires increased radiation, so use discretion when obtaining stored images.

Fourth, some older machines have an incandescent light (i.e., white light) directly overhead within the gantry. If prone positioning is used, this light can be helpful because the needle will cast a shadow. When the hub of the needle is in line with the tip, there is no shadow. Once the target is confirmed with fluoroscopy, the fluoroscope may be used only sparingly and observing the needle shadow may be sufficient to guide the needle.

Fifth, currently most machines report a total procedure fluoroscopy time and dose delivered (in mGy and mGy · cm²). In a straightforward procedure, the fluoroscopy time can be approximately 0–0.3 minutes, and the dose delivered can be 20–25 mGy and 4000–5000 mGy · cm². These values are comparable to the dose from a single chest radiograph [42].

If a puncture is proving difficult, alternative approaches may be helpful. Changing the level of approach or trying an alternative approach, such as prone oblique, can be productive. One should avoid making multiple passes through the thecal sac if possible because this increases the risk of hematoma, dural tear, and CSF leak. Proper intrathecal positioning of the needle tip is confirmed with the flow of the CSF. Occasionally the needle is in the correct position, but the CSF does not flow because of low CSF pressure or a local obstruction to flow. If a lateral view confirms that the needle is in the proper position, very subtly advancing or withdrawing the needle or rotating the needle may solve the problem. Injecting a small amount of myelographic contrast material and visualizing the contrast material with the fluoroscope can be informative. It is possible for the needle to puncture the dura and tent the thecal sac, resulting in a subdural injection into the potential space (Fig. 5), and contrast material will accumulate at the site of injection and will not highlight the lumbar nerve roots. A venous plexus is found anterior to the vertebral bodies, and venous blood may be seen if the needle has passed through the posterior dura.

A review of previous spinal imaging studies should precede choosing a plan of action. Additional imaging may be helpful, and the requesting provider might be contacted to order CT of the lumbar spine, which can better show the bony landmarks and possible points of access than fluoroscopy. Cases that are unsuccessful using fluoroscopy with a single overhead camera may be scheduled in a biplane suite to be performed by a diagnostic neuroradiologist or interventional neuroradiologist. When one radiologist is unsuccessful in performing LP, another may have a slightly different technique that results in success. In refractory cases in which the tap is critical, CT guidance is an option [35, 43]. Although an uncommon technique, cervical puncture is occasionally used for myelography or CSF collection [44].

Complicated cases, such as patients with scoliosis or a history of spinal surgery, may require a more detailed review of previous imaging studies or a request for CT of the lumbar spine before the procedure. Scoliotic patients can have extremely strong lumbar musculature and fascia that may require a larger-gauge needle to prevent bending of the needle, and the prone oblique approach may be required. Previous bone removal surgery such as laminectomy should facilitate the puncture; however, if the surgery was remote, strong scar tissue can cover the laminectomy site and can be difficult to puncture. Again, a larger-gauge needle may be needed in these cases.

An opening pressure may be requested for the evaluation of idiopathic intracranial hypertension or normal-pressure hydrocephalus, among other conditions. It is generally accepted that an accurate pressure requires the patient to be in the lateral decubitus position [35]; however, this position can pose a problem when using older fluoroscopy equipment with a single overhead camera because the patient needs to be repositioned after needle placement. Repositioning the patient with the needle in place can be difficult, and there is a possibility of losing CSF access or widening the dural puncture location, thus potentially increasing the risk of bleeding or CSF leak. Patient positioning for obtaining an accurate opening pressure has been discussed given that ordering physicians have historically recorded opening pressures from a lateral decubitus position, but the constraints of many fluoroscopy units necessitate recording the pressure while the patient is in a prone position. One recent study reported no significant difference in opening pressure values as a function of patient position [45], and another study reported a small increase in recorded pressure values when patients were in the prone position [46]. It is important to remember that the “zero” point of the manometer is roughly at the level of the heart. If the opening pressure is measured with the patient in prone position, the distance from the zero point of the manometer and the heart must be estimated and added. Free flow of CSF can be verified by elevating and lowering the manometer to verify that the CSF meniscus within the manometer rises and falls appropriately. A normal pressure ranges from 60 to 200 mm H₂O in patients older than 8 years and up to 250 mm H₂O in obese patients [47]. Opening pressures of greater than 250 mm H₂O are diagnostic of intracranial hypertension [48].

For diagnostic LP, laboratory orders should be in place for the studies requested by the ordering physician. With a standard diagnostic LP, CSF is collected in four tubes, for a total volume of 8–15 mL [15]. The common practice is to put 3 mL of CSF in each of the first three tubes and to put 5 mL of CSF in the last tube. CSF from the plastic tubing connected to the needle can be retained in the last tube.

A high-volume tap (removal of > 30 mL of CSF) generally implies the diagnostic use of CSF removal to assess improvement of symptoms, as might be seen in idiopathic intracranial hypertension or normal-pressure hydrocephalus [49]; however, a high-volume tap has been shown to be of limited therapeutic value in patients with idiopathic intracranial hypertension [50]. A closing pressure may be requested and may be helpful to the ordering physician in determining the volume of the CSF reservoir [51]. A closing pressure can also be helpful in determining the volume of CSF removed. A conversation with the patient's neurologist before the procedure can help to optimize the diagnostic value of a high-volume tap.

The patient is typically advised to remain horizontal for the remainder of the day after LP; this position is thought to give the puncture a chance to heal, although no advantages of bed rest have been found in a number of randomized controlled trials [39]. The patient should not engage in strenuous activity for at least 24 hours after LP. If the patient is to return home, it is advisable that someone who can drive him or her home after a short stay in postprocedure recovery be available. Dehydration can exacerbate a post-LP headache, so hydration is advised through the day after the procedure. Alcohol causes dehydration, so refraining from alcohol consumption is also advised through the day after the procedure. Further details for the management of post-LP headache are discussed in a recent online review [39].

If hemorrhage is suspected, the patient's history may suggest the most immediate logical course of action. A patient with borderline coagulation panel results or with corrected coagulopathy may best be treated with immediate supplemental FFP and close clinical follow-up including attention to lower extremity neurologic status. A suspected hematoma in a patient with abnormal clinical findings might be confirmed using MRI. If the patient has a contraindication to MRI, myelography may be necessary. The value of unenhanced CT in this setting should not be underestimated. Although unenhanced CT offers little soft-tissue contrast, a spinal hematoma may prove conspicuous even in the presence of hardware. Correction of coagulopathy and surgical decompression are the standard of care for patients with symptomatic spinal hematoma because it is generally believed that the extent of irreversible damage to the spinal cord or nerve roots is a function of the duration of the compression [4]. Better neurologic outcomes are reported in patients who undergo surgery within 12 hours of symptom onset than in those who undergo surgery later [11]; however, an increasing number of reports describe the resolution of spinal epidural hematoma with nonoperative management [52].

LP is a common and necessary procedure that sometimes requires fluoroscopic guidance. Familiarity with the indications for LP, methods of fluoroscopic guidance, and potential complications of LP can optimize the safety and success of this procedure.

I thank V. Miloushev for assistance with the images and C. G. Filippi for helpful commentary.