Patients selection and demographics
Following approval by our local ethical committee (no. 276/2020/CE), we conducted a retrospective review of 192 consecutive patients who underwent lumbar arthrodesis with PPS for DLSS with two different techniques: fluoroscopy-assisted (FA group) and CT navigation (CTNav group). Surgical procedures were performed by the senior author (G.S.) and his assistant (G.L.R.). All patients signed a written informed consent before surgery. Inclusion criteria for the study were: (1) age at surgery between 18 and 75 years, (2) low back pain (LBP) with radicular irradiation in the lower limbs, (3) claudicatio neurogena, and (4) failed conservative treatment for at least 6 months. Diagnosis of DLSS was confirmed with standard standing AP and lateral X-ray of the lumbar spine, flexion–extension X-rays, and lumbar spine magnetic resonance imaging (MRI). Patients with a previous history of instrumented spine surgery as well as osteopenia defined as lumbar T-score < −1 SD on dual energy X-ray absorptiometry were excluded from the study. All patients operated on before March 2020 were assigned to the FA group; after March 2020, all patients were operated on with CT navigation technique. Demographics, intraoperative, clinical outcome, and radiological data were recorded.
Surgical technique
Surgery was performed under general anesthesia with patients positioned prone on a radiolucent table (TruSystem 7000, TRUMPF Medizin Systeme GmbH, Saalfeld, Germany). Intraoperative neuromonitoring was used for all surgical procedures (NVM5, NuVasive, Memphis, TN, USA). In the FA group, a standard C-arm fluoroscopy (OEC Brivo Plus, GE Healthcare, USA) was used for pedicle screw placement. We proceeded from rostral to caudal, putting both screws of the same level at the same time to reduce intraoperative radiation exposure. In the CTNav group, an AIRO mobile intraoperative CT scanner (v 2.1.0.2, Mobius Imaging LLC, Shirley, MA, USA) was used for pedicle screw placement. In brief, a small midline lumbar incision at the level of the intercristal line [6] was performed, and a spinous process clamp (Brainlab AG, Munich, Germany) was placed. The spinous process clamp was used as a reference guide for CT scanning. Following 3D reconstruction of the surgical area, a navigated drill guide (Brainlab AG, Munich, Germany) was used to drill the holes for PPS placement. After PPS placement, an intraoperative CT scan was performed again to check for the accuracy of screws placement. After screw placement, a midline incision for laminoartrectomy and dural sac/roots decompression was performed in both groups. Transforaminal interbody fusion with TLIF cage (Trabecular Metal, TM Ardis, Zimmer Biomet, IN, USA) was added where deemed appropriate. At the end of the surgical procedure, AP and lateral X-ray was performed to confirm the correct position of the implants in all patients.
Radiological and clinical outcomes
Intraoperative radiation exposure data were collected according to the imaging modality used. In the FA group, radiation exposure was recorded by the C-arm software in terms of: (1) cumulative radiation exposure (mGy), defined as the kinetic energy per unit mass of air provided to a defined point in space, (2) dose–area product (DAP) (Gy·cm2), defined as patient’s dose per area of skin irradiated within the radiation field, and (3) radiation time (seconds) total time of X-ray beam activation [7]. In the CTNav group, radiation exposure was measured using BrainLab Curve 1.2 navigation system software (Brainlab AG, Munich, Germany) provided with the AIRO Mobile intraoperative CT scanner (Mobius Imaging LLC, Shirley, MA, USA). Radiological data from AIRO included (1) dose–length product (DLP) (mGy·cm), a measure of CT tube radiation output, and (2) CT dose index volume (CTDIvol) (mGy), the radiation intensity used to perform a specific CT exam. To compare radiation exposure between the two groups, effective dose (ED) in mSv was calculated. In FA group, ED was calculated by multiplying DAP measurement by the conversion factor 0.27 mSv/Gy·cm2, as previously reported [8, 9]. Similarly, in the CTNav group, ED was calculated by multiplying DLP by the dose–length conversion factor 0.014, as previously reported [10,11,12].
All patients underwent X-ray and CT scan examination at 6 months after surgery. Screw positioning was assessed by a radiologist and a spine surgeon not involved in the surgical care of the patients. Screw placement accuracy was measured according to Gertzbein and Robbins classification and was graded from A to E, according to the extent by which every single screw breached the cortex of the pedicle [13, 14]. Screws were graded as follows: (A) fully intrapedicular position without breach of the pedicle cortex; (B) breach of the pedicle cortex < 2 mm; (C) breach of the pedicle cortex 2–4 mm; (D) breach of the pedicle cortex 4–6 mm; (E) breach of the pedicle cortex > 6 mm or screw outside of the pedicle. Grades A and B were considered satisfactory results, whereas grades C–E were considered unsatisfactory results.
All patients were asked to complete Oswestry disability index (ODI), visual analog scale leg pain (VAS-LP), and visual analog scale back pain (VAS-BP) questionnaires before surgery and at 6-month follow-up.
Statistical analysis
Statistical analysis was performed using GraphPad Prism 5.01 software (GraphPad Software Inc., San Diego, CA). Radiological and clinical data were expressed as median (range) and count (percentages), as appropriate. Means and percentages between the two groups were compared using the Mann–Whitney U and the χ2 tests, as appropriate. A p value < 0.05 was considered significant.