Thoracic paravertebral block assessed with 3D Cone Beam
computed tomography. Part one: patterns of diffusion.
Background: Local anesthetic (LA) spread and somatosensory extension after thoracic paravertebral block (TPVB) are not predictable. Four patterns of LA spread had been classically described: longitudinal (L) or paravertebral (PV), intercostal (IC), cloud-like (CL) or mixed (M). Diffusion studies have been conducted in cadaver models or after contrast injection in radiography, using anatomical landmarks, LOR or neuroestimulation to perform TPVB. 
Goal of Study: to describe ‘in vivo’ LA spread after US-guided TPVB through 3D Cone Beam computed tomography (CTCB) analysis. Relationship with clinical efectiveness is discussed in part 2.
Material and methods: Observational prospective study. Institutional Ethics Committee and patient informed consent was obtained. 76 bilateral TPVB were performed in 44 patients (>18 yo, ASA II-III) (Table 1) scheduled for transarterial chemoembolization of liver tumor lesions. 
Procedure: standard monitoring and councious sedation. In the prone position, bilateral US single-shot TPVB at 6th/7th level was performed. 5 ml of Iomeprol 350, 5 ml of PS, 5 ml of LA per side was injected with in plane tranversal or paramedian approach Afterwards CTCB was performed.  A single, blinded radiologist evaluated CT images. US image  was evaluated as good, adequate, suboptimal. TPVB was only permormed  if image quality was “good” or “adequate”.Results: US landmarks were identified and TPVB performed in all cases (Table 2). 69 bilateral TPVB (138 single-shot TPVB) were analysed; 7 bilateral TPVB were excluded due to CT image loss. LA spread pattern combined the IC level of puncture with variable medial and lateral extensión into the PVS and above IC spaces. CT analysis suggests that spread defined as CL is situated ahead of the intercostal space, without direct contact with the intercostal nerve. The median (IQR) distribution of LA at the PVS was 3 (2,3) vertebral levels. Video 1 and figures 1-4 ilustrate LA spread CT images. 12 (17.4%) cases also presented epidural diffusion and 54 (71.4%) prevertebral diffusion (Figure 3). Interpleural spread was present in 3 TPVB (2.1%) (Figure 6-7). Video 1 and figure 5 ilustrate in vivo thoracoscopic spread.
Conclusions: After TPVB LA presents a unique spread pattern which extends into the IC space at puncture site with variable PV and lateral CL extension. In vivo thoracoscopic spread and latex spread in cadaver models reproduce this pattern.
Studies conducted with advanced imaging techniques (C, MNR) could optimize the puncture technique to achieve the predominant LA spread within the target PVS.
Volumetric image studies are needed to determine LA proportion that diffuses into the target PVS and the clinical relevance of CL spread. 
Relationship with clinical efectiveness is discussed in part 2 of the study

Thoracic paravertebral block assessed with computed tomography Cone Beam 3D. Part one: different patterns of diffusion