Probing for Thoracic Pedicle Screw Tract Violation(s): Is it Valid?

Introduction: Preparation of the thoracic pedicle screw tract is a critical step prior to the placement of screws. The ability to detect pedicle wall violation(s) prior to insertion of thoracic pedicles screws, however, has not been studied.

Purpose: To evaluate the inter- and intra-observer validity, accuracy, and agreement of detecting thoracic pedicle screw tract violation(s) among surgeons at various levels of training.

Methods: Using a straight-forward trajectory, under direct visualization, 108 thoracic pedicle screw tracts (54 cadaveric thoracic vertebrae) were prepared in a standard fashion by undertapping with a 4.5 mm cannulated tap. A deliberate pedicle violation was randomly created by an independent investigator in either the anterior, medial, lateral, inferior, or superior wall in 65 pedicles. Following this, four blinded, independent surgeons at various levels of training probed the specimens on 3 separate occasions in order to determine if a breach was present (1296 discrete data points). Surgeon findings were then recorded as breach present or absent, and if present, by breach location. Inter- and intra-observer agreement was computed using Cohen’s Kappa correlation coefficient (K) with a 95% Confidence Interval (CI).

Results: The validity of correctly determining if a breach was present or absent (% correct), as well as the accuracy in detecting the location of a breach (breach location) varied by level of training/experience (Observer 1-most experienced; Observer 4-least experienced).

* p<0 .05; PPV="positive" predictive value, NPV="negative" FNPV="false" negative FPPV="false" positive Sens.="sensitivity," Spec.="specificity"

Intra-rater agreement also varied by level of experience: Observer 1 (good/excellent; K=0.74, 95%CI=0.64-0.85), Observer 2 and 3 (good; K=0.63, CI=0.50-0.76 and K=0.67, CI=0.55-0.79 respectively), Observer 4 (fair; K=0.44, CI=0.36- 0.62).(p<0 .05). There was also low inter-observer agreement with Kappa ranging from 0.027-0.054. (p<0.05).

Conclusions: All surgeons (except for Observer 4 – PGY1) had good or excellent validity, accuracy, and intra-observer agreement in detecting the presence or absence of a breach. (p<0 .05). This was dependent on the surgeon’s level of training, with most experienced surgeon (Observer 1) having highest validity/accuracy. Agreement between surgeons (inter-observer) low, and likely is attributable to varying levels experience. Probing prior placement pedicle screws in thoracic spine highly depends