[Assessment of pedicle screws following dorsal instrumentation of the spine: a comparison of conventional radiographs with multidetector CT]. / Lagebeurteilung transpedikulA currencyrer Schrauben nach dorsaler Spondylodese mittels konventioneller RA paragraph signntgenaufnahmen: ein Vergleich zur Mehrzeilen-Computertomografie.

ABSTRACT

PURPOSE: To evaluate the performance of conventional radiographs (CR) for the assessment of pedicle screws in comparison with CT including MPR. MATERIALS AND METHODS: Comparison of CR and CT for the evaluation of screw length, extracorporal perforation (grade A 0 - 2 mm, B 3 - 4 mm, C 5 - 6 mm, D > 6 mm, relevant grade C and D), screw contact/loosening, and subjective image quality (1 excellent to 4 non-diagnostic). The sensitivity (SE), specificity (SP), positive (PPV) and negative predictive value (NPV), and overall accuracy (OA) were calculated with 4x4 contingency tables. The subjective image quality was evaluated using the Wilcoxon test (p < 0.05). RESULTS: 411 screws were evaluated. The relevant medial perforation was diagnosed with CR/CT in 53(13 %)/ 11(3 %) screws, while the relevant lateral perforation was found in 3(1 %)/ 43(11 %) screws. Pedicle screw contact was diagnosed by CR/CT in 10(2 %)/ 18(4 %) cases and incorrect length was confirmed in 80(20 %)/ 90(22 %) screws. A diagnosis of loosening was established in 10(2 %) screws by CR and in 7(2 %) screws by MDCT. In summary, CR demonstrated SE, SP, PPV, NPV, and OA for the following variables relevant medial perforation = 64 / 89 / 13 / 99 / 88 %, relevant lateral perforation = 4 / 99 / 67 / 90 / 90 %, screw contact = 11 / 98 / 20 / 96 / 94 %, incorrect screw length = 86 / 58 / 89 / 51 / 81 %, and screw loosening = 71 %/ 99 %/ 50 %/ 100 %/ 98 %. The subjective image quality of CR was significantly lower compared to CT (2.19 A + or - 0.84 and 1.11 A + or - 0.31; p < 0.001). CONCLUSION: CR evaluation of pedicle screws following dorsal instrumentation of the spine demonstrated an OA < or = 90 % for relevant medial and lateral perforation and for incorrect screw length compared to CT. Thus, reliable evaluation of pedicle screws based solely on CR does not seem feasible.