The initial search returned 2278 results, with 562 articles eligible. After reviewing those articles, we found that only 30 were relevant for inclusion (Fig. 1). The results of the analysis are divided into three sections: (i) the accuracy of the spine interventions with fluoroscopy, (ii) the accuracy of US- and CT-guided spine interventions and (iii) the anatomic basis for the failure of the landmark-guided technique.

Fig. 1

PRISMA 2020 flow diagram for new systematic reviews which included searches of databases and registries only

Accuracy of Landmark- Versus Fluoroscopy-Guided Spinal Pain Procedures

The first published paper exploring the accuracy of landmark-based epidural injections was published in 1980 [17]. Since then, many authors have published their studies describing the accuracy and effectiveness of their techniques, comparing different imaging methods either with one another or with landmark-based approaches. With a systematic approach, we have identified those studies and present them in this review (Tables 1, 2, 3).

Epidural Steroid InjectionsCaudal Epidural Steroid Injections

We identified eight prospective studies [7, 17,18,19,20,21,22,23] assessing the accuracy of the blindly performed caudal epidural injections. The methodology of all involved a fluoroscopic image with contrast applied following the blind insertion of the needle to confirm the correct placement of the needle. As presented in Table 1, the rates of the incorrect needle placement varied significantly among the studies from 0% to more than 50% (median 25.5%). As expected, it also varied amongst clinicians with different level of experience (as explored in two studies [7, 20]—see Table 1). The intravascular needle placement also varies among the studies from 1.5% to 9.2% (median 4.5%).

Table 1 Landmark-based epidural steroid injections (caudal and lumbar)Lumbar Interlaminar Epidural Steroid Injection

Similarly, Table 1 includes seven studies [17, 23,24,25,26,27,28] that assessed epidural injections at the lumbar level. The rates of incorrect needle placement and unintended intravascular placement are seen in Table 1. It is evident that the former rate also ranges between studies (7–30.4%, median 12.3%) and can be as high as one in three injections [17]. The intravascular needle placement rate was assessed in only two studies [17, 23] and no cases were identified in either (Table 1).

Cervical Interlaminar Epidural Steroid Injections

No studies directly comparing the landmark with the fluoroscopic technique were identified. Stojanovic et al. assessed the number of loss of resistance (LOR) attempts required to achieve adequate epidural contrast spread, which is the gold standard to confirm the needle placement in the epidural space [29]. In this study a very high rate of false LOR was identified (53%) and, in some cases, even a third or a fourth attempt was required for the needle to reach the epidural space [29]. This is understandable given the discontinuation of ligamentum flavum at cervical levels [30, 31]. Other confounding factors, while using the LOR technique, can be cysts in interspinous ligaments, the paravertebral muscles, the thoracic paravertebral spaces, and intermuscular planes (such as the dorsal thoracolumbar ligament) [32]. Additionally, the spread of the contrast was unilateral in 51% of the cases, while in 28% of the cases the spread was ventral, emphasizing the importance of the use of fluoroscopy for patients with unilateral pain [29].

Another study showed that the use of contrast media in the cervical area is of utmost importance as its spread can show different patterns, while the extent of dispersion can be variable and unpredictable. Gill et al. suggested that the contralateral oblique view provides clear confirmation of epidural spread, while the antero-posterior (AP) view is an excellent view to measure both the craniocaudal and foraminal spread [33]. At low volumes of contrast, the AP view seems to provide the best estimate of the likelihood that the injectate will reach the pain-generating pathology and is considered the optimal for this purpose [33].

Lumbar and Cervical Transforaminal Steroid Injection

Transforaminal injections can only be performed with image guidance for two reasons: there is no established safe landmark-guided method to direct the needle to the foramen; and the high prevalence of vessels, potentially providing vascular support to the spinal cord, in the proximity of the nerve roots in the foramen. Table 2 shows prospective studies that evaluated the risk of intravascular [34,35,36,37,38] and intradiscal [36] needle placement while performing fluoroscopically guided transforaminal injections. The risk of intravascular injection varies amongst the spinal levels and amongst the included studies. At the cervical level it was documented as approximately 20% [34, 35], in the thoracic region around 8% [34], at the lumbar levels it ranged from almost 6% [34] to 15% [36] and at the sacral level from 16.5% [34] to 21% [37]. One study reported a rate of intradiscal needle insertion of 2.3% [36].

Table 2 Fluoroscopically guided transforaminal epidural injections and comparison among DSA, real-time fluoroscopy, anteroposterior and oblique approach

To improve the safety of the injection, several techniques may be employed. Real-time fluoroscopy, digital subtraction angiography, and oblique views may provide various safety advantages, including improved needle guiding, better imaging of anatomical structures and a lower risk of procedural complications, when performing epidural injections [39,40,41,42,43,44,45,46,47,48].

Real-time fluoroscopy provides continuous viewing of needle insertion and contrast distribution, allowing for quick modifications to improve accuracy and reduce complications. Real-time fluoroscopy also lowers the risk of unintentional dural puncture, nerve damage and vascular trauma by ensuring accurate needle insertion in the epidural space [39].

Digital subtraction angiography (DSA) improves vascular structure imaging by removing background structures, resulting in improved clarity and definition of the epidural space. This technique includes acquiring images before and after contrast injection, then subtracting the pre-contrast image from the post-contrast image to highlight vascular structures. This approach helps to identify vascular abnormalities and reduces the danger of unintended vascular puncture during epidural treatments. DSA helps to avoid unintentional vascular injections, lowering the risk of haematoma development, arterial embolization and other vascular problems [39, 41]. It is important to point out that DSA cannot reliably confirm intra-arterial placement [49].

Oblique fluoroscopic images, by tilting the fluoroscopic C-arm, give additional viewpoints on needle trajectory and contrast spread, allowing for better visibility of the epidural area from various angles. These views can assist in confirming adequate needle insertion in the epidural area and highlight potential issues such as nerve impingement or intravascular injection. In general, this view improves procedural accuracy and lowers the risk of complications; it is particularly useful for the S1 level as seen in Table 2 [50].

Many studies compared the aforementioned techniques as seen in Table 2.

Live fluoroscopy has been deemed better than the static images, as the interpretation of the latter may miss up to 57% of the vascular injections [51]. DSA is not a panacea for preventing adverse outcomes during the performance of neuraxial procedures. Although not all studies identify the advantages of DSA [41], most studies find a clinically and statistically significant difference (see Table 2). Some disadvantages of this method are that DSA is limited by motion artifacts and the images are subject to human interpretation. Any motion between the initial scout film and subsequent images is detected as a change, impeding the subtraction process, and causing degradation of image quality. Thus, utilisation of this technology does not negate the potential for human error nor the potential for patient injury. DSA may provide greater sensitivity and specificity, but the exact limits of detection are unclear, and the safety profile is neither fully characterized nor validated [52]. One significant drawback of DSA is that it may substantially increase exposure to ionizing radiation, comparable to computed tomography angiography (CTA). The routine use of DSA is not warranted on the basis of current medical evidence [52].

Sacroiliac Joint Steroid Injections (SIJ)

Table 3 presents three studies [53,54,55] supporting that imaging techniques for SIJ are superior to non-imaging techniques. Without image guidance, the misplacement of needle is unacceptable (78–88%), while fluoroscopically guided injection results in not only superior accuracy but also better intermediate-term outcome [55].

Table 3 Sacroiliac joint injectionsMedial Branch Block and Facet Joint Injections

Although facet joint injections and medial branch blocks are commonly performed under imaging techniques (fluoroscopy, CT or US), blind techniques for lumbar levels have also been described [56]. To date, we did not identify any published paper to assess the accuracy of the landmark-based technique for facet injections. Furthermore, considering that the medial branches are deep in location, of small calibre and in close proximity to the nerve root, a blind technique is not recommended according to all current consensus guidelines. Currently, fluoroscopy is the gold standard for facet joint and medial branch blocks and is either recommended or required by multiple insurance companies [57]. The Spine Interventional Society guidelines also state that fluoroscopy is mandatory for the conduct of lumbar medial branch blocks’ as it provides an overview of the bony anatomy as well as the ability to confirm contrast spread [58]. CT guidance may be an alternative; however, it is not currently recommended as the imaging method of choice [57].

Accuracy of Identification of Level of Spine

Different to the context of regional anaesthesia, correct identification of the spine level corresponding to the spinal pathology is important. Without imaging, identification of the level of spine relies on the anatomical landmarks such as Tuffier’s line, the tenth rib line as well as posterior superior iliac spine. One study assessing the accuracy in 60 patients showed that the true level was noted in 48.33% when using Tuffier’s line, 53.33% when using the 10th rib line and 65% of cases when using the posterior superior iliac spine [59]. Another study examining the anatomical landmark for lumbar puncture (L4–L5 intervertebral space or L4 vertebra) reported that there was concordance of intervertebral space identification in 64% of the cases (78/122) as confirmed by US [60]. For the cervicothoracic intervertebral spaces one study showed that the level identification may deviate in up to 58% of cases [61]. The reason behind the poor accuracy is obvious as the recognition of the bony landmark is merely contingent on the tactile sensation, which can be severely hampered with increasing depth of the subcutaneous layer.

Use of CT and Ultrasound Guidance: Comparison with FluoroscopyUltrasoundFacet Joint or Medial Branch Block

A comprehensive review comparing US with CT and fluoroscopy reported that US guidance for injection of the cervical facet joints and their innervating nerves had reasonable accuracy (78–100%) with shorter procedural time compared to fluoroscopy or CT guidance, while offering comparable pain relief [62]. The same review found the accuracy of US-guided lumbar facet joint intra-articular injections to be 86–100%, more reliable than medial branch blocks (72–97%), with analgesia similar to that from fluoroscopy- and CT-guided blocks [62]. However, those conclusions need to be interpreted with care. Firstly, the US-guided procedures are skill dependent and a number of those studies are from specialized centres where a high number of US-guided spine injections are performed [63]. A recent meta-analysis showed that US-guided lumbar medial branch blocks and facet joint injections are associated with a significant risk of incorrect needle placement (7–14% among the studies), as confirmed by fluoroscopy or CT [64]. A recent cadaveric study on the accuracy of cervical medial branch block under US guidance even revealed that the accuracy was less than 79% [65]. Secondly, the success rates also depend on the patients’ body build. Early studies included patients with low body mass index (BMI) [66, 67]. However, when US-guided spine procedures were applied in patients with obesity (BMI > 30), the success rate dropped to 62%. Thirdly, these procedures are more challenging for deeper targets (e.g. lower cervical levels, L5 dorsal ramus) [68]. US technical limitations and individual patient factors also contribute to the risk of incorrect needle placement. The use of US may be indicated in a certain clinical setting or selected clinical scenarios where avoiding radiation exposure is a key outcome [64] or not accessible.

Epidural Steroid Injection

The current literature also supports that US-guided transforaminal epidural injection is an effective procedure with the equivalent efficacy to fluoroscopy-guided injections [62]. One recent meta-analysis supported that US-guided epidural injections are comparable to fluoroscopic guided injections regarding pain control and functional improvement, offering a decreased risk of inadvertent vascular puncture [69]. Another meta-analysis, though, supported that fluoroscopy-guided injections led to better functional status compared to the US-guided blocks [70]. In terms of effectiveness in treating back pain and complications, no difference was identified [70].

Again, it is important to recognize that the part of neurovascular structures within the bony component of vertebrae cannot be visualized by the US. That means that US cannot monitor the intravascular injection in the foraminal area as well as the transforaminal and epidural spread of the injectate, all of which is made possible by fluoroscopy with contrast.

CT Guidance

The use of CT for spinal injections is practiced in some pain centers but is not widespread. One prospective cohort study compared fluoroscopy-guided to CT-guided transforaminal epidural steroid injections and showed similar results at 3 months in both groups (P = 0.511) [71]. CT allows direct visualization of the nerve root and the needle tip, as well as of the distribution pattern of the injected contrast [12].

In addition to the accessibility of the CT machine across the intervention community, radiation dose is a limiting factor. Kamp et al. reported ten times higher radiation dose for patients who received CT-guided spinal procedures when compared to that from fluoroscopy-guided procedures [72]. In the same retrospective study, higher discharge rate and a lower procedural cost were reported [72]. Investigators have explored a low-dose protocol, limited in one or two spinal segments [12]. To support this, another study showed that ultralow-dose CT-fluoroscopy for image-guided lumbar spine epidural steroid injections can offer lower radiation dose compared to the fluoroscopy-guided technique. Ultralow-dose CT-fluoroscopy omits a planning CT scan, utilises CT-fluoroscopy and minimises radiation dose parameters [73]. The radiation dose is an important parameter for clinicians. Dietrich et al. showed that fluoroscopy-guided lumbar spinal injections resulted in lower radiation exposure for participants but higher radiation exposure for physicians when compared with CT-guided injections [74]. This is not surprising as the position of the provider needs to be in close proximity to the patients during the fluoroscopy procedure.

Anatomic Basis for Importance of Image Guidance

While it is clear that the transforaminal approach requires image guidance, practice survey suggests that interlaminar epidural steroid injection is frequently performed without image guidance even in developed countries [75]. Image guidance allows accurate delivery of injectate to the epidural space. In general, there are two reasons why the interlaminar epidural medication is delivered to the wrong location with the traditional LOR technique: failure to detect the wrong location outside the epidural space, and failure to detect the intravascular spread.

Traditionally, the epidural procedure is performed with loss of air/saline resistance technique [76]. The premise of this haptic technique is contingent on the dramatic drop in resistance in the epidural space relative to that inside the ligamentum flavum. The sensitivity of LOR in the lumbar area is 99% but with a specificity of 27% [28]. Risks of false LOR have been reported between 8% and 30% [17, 26, 28]. This can be related to the experience (e.g. failure to recognize the needle in paraspinal muscle) or equipment (smaller gauge epidural needle). Even in experienced hands, there are a few anatomic reasons accounting for the injectate found in the wrong place.

False Localization with LOR MethodInterspinous Ligament and Ligamentum Flavum

Once the needle is in the interspinous ligament, it gives the tactile feeling of high resistance to the procedure provider because of the solid architecture of fibres in the ligament. However, multiple anatomic studies revealed that two-thirds of interspaces had at least one gap in the lumbar interspinous ligament (LISL) filled with adipose tissue [77,78,