A novel integrated platform for the identification of surgical margins in oral squamous cell carcinoma: results from a prospective single-institution series

This pilot prospective non-randomized study was approved by the local ethics committee (approval #19_2018bis). A total of 16 patients (10 males, 6 females) with OSCC eligible for surgical treatment were enrolled. Informed consent was obtained from all patients. Only patients diagnosed and managed in IRCCS Ca′ Granda Foundation – Policlinico Maggiore Hospital, Milan, Italy, > 18 years old, chemotherapy- and radiotherapy-naïve, with no history of cancer were included. All patients underwent surgical treatment with a program established according to the guidelines of the American Joint Committee on Cancer (7th edition) [2].

The MM was assessed by a craniofacial surgeon at a distance of 1,5–2 cm from the tumor, as described (Fig. 1a) [7, 14]. Subsequently, two ears, nose and throat (ENT) surgeons performed intraoperative NBI endoscopic evaluation using a scope of 4 mm outside diameter (Olympus Visera Pro system, Center Valley, PA America, with OTV-S7Pro camera and CLV-S40Pro light source). This intraoperative analysis allowed for the identification of the interface between the likely neoplastic/dysplastic and likely healthy areas, i.e. NBI margin. Next, the maxillofacial surgeon performed multiple biopsies of both the MM and the NBI margins. For each patient, from 2 to 6 bioptic samples were obtained. The NBI margins were then classified as overlapping, external, or internal, compared to the MM. In case of overlap between the MM and NBI margins, only one biopsy was performed (Fig. 1).

Hematoxylin and eosin (H&E) stained frozen sections of the MM margins were intraoperatively analyzed by a pathologist as part of the standard protocol to drive the surgical intervention. All surgical samples, including the tumor, the MM and NBI margins, were then analyzed after tissue processing by two pathologists (NF and MM), as shown in Fig. 1a. Specifically, all cases were classified and graded following the latest World Health Organization criteria [15]. Pathologic staging was assessed according to the current TNM staging system [16]. The presence of tumor infiltrating lymphocytes (TILs) was assessed as described [17]. The MM and NBI surgical margins were defined as positive in the presence of OSCC and/or dysplasia; otherwise, they were marked as negative.

MD was investigated by immunohistochemistry (IHC) in the MM and NBI surgical margins. Representative 4-μm-thick sections were cut from the MM, NBI, and tumor blocks and subjected to IHC using pre-diluted antibodies against CD34 as previously described [18]. Positive and negative controls were included in each slide run. Briefly, the protocols use an automated staining system (Dako Omnis) and anti-human prediluted antibodies [19]. Protein expression was analyzed in all different samples by two independent pathologists (NF and SF). Discordant results were resolved during dedicated consensus sessions. Sections were first observed at low magnification (40x) to identify the areas with the higher concentration of vessels. Then, the vessels count was performed at 200x by means of a customized digital image analysis algorithm using the Aperio CS2 instrument (Leica Microsystems Srl) [20]. The MD value was expressed as a percentage. Each CD34-positive structure (round, oval, and irregular) separated from other profiles or tissue elements was counted as a single vessel, regardless of the presence of a clear lumen.

Data were analyzed using Prism 4.0 (GraphPad Inc., La Jolla, CA, USA). Differences among sample groups were analyzed using the unpaired Student’s t-test as previously described [21]. The association between positive margins was evaluated by Fisher’s exact test according to the classification proposed by Piazza and collaborators [22]. Statistical significance was assumed for a probability value (p) less than 0.05.

Taken together, 104 specimens were analyzed, including 16 (15.4%) tumors, 34 (32.7%) MM, 34 (32.7%) NBI margins, and 20 (19.2%) MM-NBI overlapping margins (Fig. 1b). The NBI margins were closer to the lesion in 17 (50%) cases (Fig. 1b) compared to the MM assessment. However, this method showed no propensity to allow for a more conservative resection, given that in the same number of margins (n = 17, 50%) was the MM the more conservative approach. Furthermore, this heterogeneity was irrespective of the tumor site and was not present at a single-patient level. At the histological examination, the margins collected with the MM intraoperative assessment revealed dysplasia in 3 (8.8%) cases and OSCC in 1 (2.9%) case, while 30 (88.2%) samples were negative as represented in Table 2 and Fig. 2a. The analysis of the NBI margins showed dysplasia and OSCC in 2 (5.9%) and 1 (2.9%) cases; respectively, while 31 (91.2%) margins were negative, as confirmed by histological examination (Table 2 and Fig. 2b). Among the 20 overlapping MM-NBI margins, 2 (10%) cases were positive. In particular, positive margins showed a significant association with thick and thin non-keratinized epithelial cells [23] (p = 0.027). These data suggest that the intraoperative integration of MM and NBI analysis might allow for a more conservative excision of OSCC compared to each of the two methods alone.

MD has been investigated in 83 margins and matched OSCCs. This analysis showed significantly high CD34 levels in pathological margins compared to the normal ones (p < 0.0001, Fig. 2c). This observation was unrelated to the intraoperative method of surgical margins assessment (i.e. MM and NBI). Internal and external margins didn’t show a statistically significant different MD, akin to the tumor site.