Diagnostic accuracy of combined optical- and radio-guided SNB for neck staging of oral squamous cell carcinoma lesions in the anterior oral cavity

A prospective single-arm clinical study was conducted at Copenhagen University Hospital-Rigshospitalet, a tertiary cancer center in Denmark, between November 2017 and December 2020. Consecutive patients with primary early-stage cT1-T2N0M0 OSCC, scheduled for SNB staging were enrolled. Only patients with biopsy-verified squamous cell carcinoma located in the anterior oral cavity with the following sub-sites were included: FOM, inferior surface of tongue, buccal mucosa, and lower gum. These sub-sites were selected due to a possible higher reported risk of false negative results of SNB caused by the shine-through phenomenon. The nodal neck status prior to surgery was evaluated with clinical examination, ultrasonography, MRI, and/or CT. Exclusion criteria were previous head and neck cancer, previous, radiotherapy or surgery to the neck, and pregnancy or allergy to ICG or iodine. From all patients, a written informed consent to participate was obtained. The study was approved by the Danish Regional Scientific Ethical Committee (H-1-7001881) and conducted in accordance with the Helsinki Declaration (2002).

At the day of surgery, or the day before, the patients had a peri-tumoral submucosal injection of the bimodal tracer ICG:Tc99m:Nanocoll. The preparation of the tracer and the imaging protocol has previously been reported in detail [19]. Briefly, immediately after tracer injection by the surgeon, dynamic and static lymphoscintigraphy was performed, followed by a single-photon emission computed tomography/computed tomography (SPECT/CT). The preoperative imaging was reviewed by a nuclear medicine physician, and number and anatomical location of SNs were reported to the surgical team. A lymph node clearly identified on lymphoscintigraphy and SPECT/CT was reported as a SN. If no SN could be detected on initial imaging, a second scanning protocol was undertaken 90–120 min post-injection (see Figs. 1 and 2).

The tumor ablation and the reconstruction were undertaken prior to SNB neck procedure. No free flaps were used in this series. For optical near-infrared fluorescence (NIRF) imaging, one of two available clinically approved optical camera systems, Fluobeam 800 (Fluoptics, France) or EleVision (Medtronics, USA) was used. The Fluobeam 800 generated NIRF imaging merged on video recording of the surgical field in gray scale. The camera head was handheld during the procedure, and imaging was presented on a screen on a chart for the surgical team. The EleVision recorded video in color-merged with the NIRF imaging presented in color or signal-intensity graded color. The camera head was mounted on a maneuverable surgical arm attached to chart with a screen. During imaging, surgical lights and headlights were turned away to diminish interference from near-infrared light from external light sources. Prior to incision, the SNs identified on preoperative imaging, were searched with a handheld gamma probe (Neo2000, Neoprope Corporation, USA), and the locations were marked on the skin. All patients had an upper submandibular neck incision to allow a super-selective ND of level 1 to be performed. If the tumor involved the midline, both neck sides were entered to access level 1. Depending on the location of the detected SNs on the preoperative imaging, second small incision in the lower part of the neck was needed in some cases to harvest SNs located in levels II–IV. To maintain oncological safety in the assessment of the possible diagnostic benefit of optical navigation for SN detection in level 1, it was decided, that all patients also had a super-selective ND of level 1 performed, instead of an optical exploration only. With this approach, possible subclinical nodal metastasis in level 1, not detected by optical guidance would be removed, and the histological examination of the neck dissection specimen served as reference for the accuracy of optical-guided SN detection. Navigation with a handheld gamma probe directed by the radioactive signal and NIRF imaging for the optical signal was performed in parallel to detect SNs. Specifically, for level 1, the surgical field was initially searched for possible SNs, not detected on preoperative imaging, with NIRF imaging, before the super-selective ND was performed. The anatomical borders of the super-selective ND in level 1b were the preglandular triangle defined by the lower edge of the mandible, the anterior border of the submandibular gland, the anterior belly of the digastric muscle and the mylohyoid muscle medially. The anatomical boundaries for level 1a were the midline space below the mylohyoid muscle between the anterior bellies of the digastric muscles and the hyoid bone inferiorly. Before closure, the surgical field was searched systematically with the gamma probe and the optical camera for any remnant signals. The dissected specimens from level 1 were also searched for radioactive or optical signals ex vivo. A lymph node was denoted a sentinel node if the radioactive count was at least 10% of the count of the hottest lymph node detected [9]. The individual gamma counts were registered, and optical images of all harvested SNs were stored. The exact anatomical location of all harvested SNs was recorded on a pictogram in a clinical report form. Contralateral drainage was defined as drainage to the opposite neck site from a tumor without midline involvement (see Fig. 3).

Harvested SNs were processed in a standard step-serial sectioning protocol based on H&E and cytokeratin antibody staining, as previously described [19]. Frozen sectioning technique was not applied. The level 1 ND specimens were examined for lymph nodes by routine histological analysis. Detected metastatic deposits were classified as macro-metastasis, micro-metastasis, or isolated tumor cells (ITC), respectively [20]. In case of a positive SNB, a completion ND was subsequently performed. All patients entered a 5-year surveillance program and were seen for regular clinical follow-up, every 3 month the first year, every 6 month the second and third year, and every 12 month the fourth and fifth year. Based on medical chart review, last follow-up was conducted on March 15, 2022.

The software SAS Enterprise Version 9.4 (SAS Institute, Inc) was used for data analysis. Survival data was analyzed by use of the Kaplan–Meier method. Overall survival (OS) was defined as time from diagnosis until death of any course or censored at the last known follow-up. Recurrence-free survival (RFS) was defined as time from diagnosis until biopsy-validated recurrence or censored at the last known follow-up. For assessment of a possible association between depth of invasion (DOI) as a continuous variable and presence of a positive SN, a Student’s t test was applied. To access the diagnostic performance of SNB, the primary endpoint was false negative events defined as N-site recurrence in the initially SNB-negative group of patients. False negative rate (FNR, (FN/FN + TP), false omission rate (FOR, FN/FN + TN), and negative predictive value (NPV = 1-FOR) were calculated.