water phantom is a watertight container with a submerged target model. Water makes the background hypoechoic, allowing easy visualization of the target and needle. Water phantoms are readily available, inexpensive, and reusable. However, they provide no tactile feedback, since water cannot hold the needle in place. Therefore, water phantoms are useless for teaching and practicing ultrasound-guided needle performance, except for the identification of sonoanatomy and the development of new needle techniques (Fig. 1) [345].

A gelatin-based phantom is produced by placing the target model in a gelatin solution. These designs have been used for ultrasound biopsy and ultrasound-guided vascular access training phantoms [67]. Gelatin-based and gelatin/agar-based phantoms of the lumbosacral spine were introduced recently [18]. Brascher et al. [9] used a phantom made of guar gum for training in ultrasound-guided intercostal nerve and stellate ganglion blocks. Materials such as flour, cellulose powder, or psyllium husk can be added to the gelatin solution to simulate the appearance of soft tissue in the phantom [3610]. Psyllium husk is a dietary fiber supplement that we usually add to the gelatin solution for spine phantoms to create an opaque, echogenic background. Increasing the psyllium husk content increases the firmness of the phantom and the level of difficulty for novices. Such phantoms are relatively inexpensive and simple to produce. Gelatin phantoms produce sufficient tactile feedback for practicing needle handling (Fig. 2). These phantoms can be stored in a refrigerator for weeks. However, needle-track marks are left in the phantom after practicing injections. The parenchyma of the material can be broken when using a bent or blunt needle. A hyperechogenic needle cannot be used in this phantom, because air pockets instead of fluid remain between the phantom tissue and needle. The gelatin background will deteriorate and decompose over time. A hydrocolloid skin dressing can be used to cover the surface of the phantom to minimize biological breakdown [11] but this can decrease the image resolution.

The commercial Blue Phantom (CAE Health Care, Seattle, WA) is made from an elastomeric rubber. Its firm texture gives more tactile feedback than does an agar/gelatin/guar gum model. When a needle is inserted into the elastomeric material, it pushes the material aside and then rebounds to its original location after the needle is withdrawn [3]. Consequently, this phantom has less chance of generating needle-track artifacts. As in a gelatin model, a bent, blunt, or hyperechogenic needle is not used. Although many centers use this phantom, it is expensive compared with homemade phantoms. And its low background echogenicity can lead to false confidence in performance. In addition, it is difficult to incorporate a target model within the structure.

A meat phantom is moderately cheap compared with the Blue phantom. Since it has anatomic structures, it gives some tactile needle feedback, and the echogenicity of the background mimics human tissue. Practice of needle guidance is possible with less needle-track artifacts, because the tracks fill with tissue fluid after needle removal. Simulation of local anesthetic injection and dissection after injection are available. The pig phantom is popular (Fig. 3). Beef has a thick fat layer, and turkey is too slippery [3].

Cadaver phantoms are expensive and not readily available without the cooperation of the Anatomy Department. When available, they provide excellent images and tactile feedback that mimic living human tissues. They also permit local anesthetic injection and dissection of the target tissue after injection [312]. However, they have no normal vascular anatomy, because the vessels are collapsed. Hocking and McIntyre [12] demonstrated restoration of the normal vascular anatomy following gelatin perfusion.

In three-dimensional (3D) printing, successive layers of material are laid down under computer control. This has the potential to produce objects from multiple materials that either transmit or block ultrasound. West et al. [13] described a method for producing ultrasound phantoms of the spine using 3D printing. Their model had a component representing the ligamentum flavum that transmitted ultrasound. The osseous and ligament components of the model had sonographic appearances similar to those of human patients. The spinal model was secured in a microwave-safe rectangular container, which was filled with gelled agar.