A single-center study was conducted in the musculoskeletal imaging department of the University Hospital of Lille, France. This is a referral center for the management of difficult implants, as described in the literature , with physicians (radiologists, gynecologists, surgeons) who have expertise in localizing and removing nonpalpable implants. Between November 26, 2019, and January 15, 2021, 45 patients were referred to the department specifically for the ultrasound-guided removal of a deep implant, defined as clinically non-palpable and/or with a failed attempt of removal in consultation or surgery. All consecutive 45 patients who underwent this intervention were included in this retrospective analysis. There was no loss to follow-up. Written informed consent was obtained for all patients for the analysis of the data from their procedure. Institutional Review Board approval was obtained under the reference CRM-2002–116.
Sterile equipment comprised a pair of gloves, an ultrasound probe cover, ultrasonography gel, compresses, antiseptic (povidone iodine), two syringes (10 cc and 20 cc), lidocaine chlorhydrate 1% (10 mg/mL), sodium chloride (NaCl 0.9%, 9 mg/mL), a scalpel, a 5-cm 21-G needle, wound closure strips, and 1 mm Hartmann grasping microforceps (MCO13A, Integra MicroFrance). The procedure was performed by a senior musculoskeletal radiologist experienced in ultrasound-guided interventions (T.J.). Procedures were performed on an outpatient basis in a dedicated interventional ultrasound room.
Patients were positioned supine, arm in 90–100° abduction, and external rotation, exposing the medial side. Ultrasound localization and removal procedures were performed using a high frequency hockey stick probe (i22LH8, Aplio i800, Canon Medical Systems). Under strict aseptic conditions and continuous ultrasound guidance, the radiologist implemented local anesthesia using lidocaine 1% (Fig. 1a) using a 5 cm 21-G needle, in all soft tissue up to the chosen grasping site of the implant. The needle was kept in the same position then used for hydrodissection by NaCl 0.9%, releasing adhesions around the grasping site of the implant and distancing local critical structures .
A small skin incision (tip of the scalpel) was made at the same insertion point to introduce the microforceps under ultrasound control (Fig. 1b). In contact with the implant, the forceps was opened to grasp it and remove it in a single piece, without fragmentation (Fig. 1c) (supplementary material 1).
The ultrasound probe was continuously kept parallel to the long axis of the forceps (“in-plane” forceps) and in the short axis of the implant (“out-of-plane” implant), which confers several advantages. First, it gives the operator a wider range of potential grasping sites along the 4 cm of the implant, whereas an “in-plane” approach of the implant would only enable a maximum of 2 grasping points (the implant edges). Moreover, due to its insertion technique, the implant runs parallel to the surrounding neurovascular structures in the medial side of the arm, a transverse approach thus enables a permanent control of both the implant and the surrounding structures. Finally, due to its flexibility and the soft nature of the surrounding tissues, the implant can move when in contact with the forceps, resulting in its disappearance out of the field of view if monitored in its long axis, which is not the case in its short axis since the implant is always on screen, even when pushed away or grasped by the forceps.
The size of the incision at the end of the procedure was measured (Fig. 1d) and skin closure was performed with wound closure strips.
Telephone follow-up at 1 week and 1 month screened for complications: local or general signs of infection, healing status, local or neuropathic pain, and local ecchymosis.
Pre-procedural data comprised patient age, body-mass index (BMI: kg/m2), implant palpability, reasons for removal, and intended subsequent contraception in gynecological follow-up.
Data collected during ultrasound location comprised implant position (supra- or sub-fascial), implant depth with respect to the skin, and at-risk neurovascular structures within 3 mm of the implant (corresponding to the forceps opening distance).
Intra-procedural data comprised timing of the various steps (location, anesthesia, removal) and quantities of lidocaine 1% and NaCl 0.9% used. Pain was monitored throughout the procedure, on a 0–10 numeric scale.
Analyses were performed with Prism 9 software (GraphPad). Quantitative data were reported as mean ± standard deviation (SD). Normal distribution was tested using a D’Agostino-Pearson test. Normally distributed data were compared using the Welch t-test, and non-normal data using the Kolmogorov–Smirnov test. Qualitative data were reported as raw number and percentage (%) and compared using the two-tailed Fisher exact test. The significance threshold was set at p < 0.05.