Serial step sections at narrow intervals with immunohistochemistry are required for accurate histological assessment of sentinel lymph node biopsy in oral squamous cell carcinoma

Sentinel lymph node (SLN) biopsy is an accurate staging modality in early oral squamous cell carcinoma (OSCC), but its accuracy relies on labor‐intensive histopathology protocols. We sought to determine whether serial step sections with immunohistochemistry (SSSIHC) at narrow intervals of the entire SLN are required to accurately exclude metastasis.


| INTRODUCTION
Accurate staging of the neck is essential in the management of oral cavity squamous cell carcinoma (OSCC) because lymph node metastasis is described as the single most important prognostic factor. [1][2][3] Clinical and radiological methods alone are inadequate staging modalities in early OSCC as approximately 20%-40% of T1-2 tumors harbor occult nodal metastasis. [4][5][6] The high rates of occult metastasis have driven the move toward elective neck dissection (END) for all T1-2 N0 OSCCs. However, routine END is likely to be unnecessary surgery in 70%-80% in this group of patients. Against this background, several recent meta-analyses and clinical trials heave demonstrated sentinel lymph node biopsy (SLNB) to be a viable neck staging procedure that can reduce unnecessary END. [7][8][9][10][11] The reduced morbidity of SLNB compared to END has led several authorities to recommend the former as standard of care in early stage OSCC. [12][13][14] The sensitivity and negative predictive value of SLNB is known to be dependent on laboratory technique. 8,15 While serial step sections (SSS) with immunohistochemistry (IHC) improves diagnostic accuracy, it creates additional burden on laboratory resources and pathologist's time. Furthermore, there is considerable laboratory methodological variation in the literature (Table 1) with no consensus regarding the number of SSS required, nor the optimal interval thickness between step levels. Reducing the number, and/or increasing the interval thickness between step sections may be expedient in relation to laboratory workload but runs the risk of missing metastatic deposits thereby reducing the sensitivity of the technique. Our standard practice has been to undertake SSS with IHC at 150 μm intervals of the entire sentinel node (SN) according to the method used in a large multicenter European trial. 6 The aim of this study was to determine whether this labor-intensive protocol is required to accurately exclude the presence of metastatic carcinoma cells within the SN. To achieve this aim, we reviewed consecutive cases of OSCC SLNBs to identify (1) the first histological section level containing carcinoma cells, and (2) the total number of section levels containing carcinoma cells.

| PATIENTS AND METHODS
All SNs in this single-institution study were processed according to a standard operating procedure. SNs of <3 mm thickness were submitted whole, nodes of 3- Consecutive patients undergoing SLNB for T1-2 N0 OSCC between May 2007 and October 2020 were retrospectively identified using a pathology database. All slides for cases previously reported as positive for carcinoma were reviewed and the first level to contain carcinoma was identified. For cases progressing to step sections with IHC, the number of levels containing carcinoma cells were also recorded. This study was registered as part of a service evaluation audit (Guy's & St Thomas' NHS Foundation reference 10965) and exempt from formal research ethics committee approval.

| RESULTS
Two hundred and seventy-two patients underwent SNLB during the study period giving a total of 901 nodes (mean 3.3, range 1-10 SNs per patient). Per SN, the mean, median, and range of slides were 41. 6

| Number of positive levels
Of the positive SNs requiring SSS, carcinoma cells were present in only a single level in 29 nodes (47.5% of positive SNs, Figures 3 and 4), of which 8 (13.1%) and 21 (34.4%) of all SN progressing to SSS were Mi and ITC, respectively. Where tumor cells were present in multiple levels, there was discontinuous tumor (multiple deposits) in 7 (5.9%) nodes, of which 2 (28.6%) were Ma and 5 (71.4%) ITC ( Figure 5).

| DISCUSSION
Our study identified occult lymph node metastasis in 33.1% of patients with cT1-2 N0 OSCC, a rate commensurate with that reported in the literature. [4][5][6]15 The high rate of occult metastasis mandates active management of the neck in this group of patients. However, subjecting all patients with clinically undetectable neck disease to END is likely to result in unnecessary surgery in an estimated 60%-80% of individuals. Against this background, SLNB is now considered by many to be the standard of care since it accurately stages the neck thereby avoiding the need for neck dissection in patients who do not require it. 13,14 The benefits of SLNB over END are only apparent if the sensitivity of the former exceeds 83.7%. 13 However, while SSS and IHC increases the sensitivity rates of SLNB, this results in considerable burden on histopathology services and may explain the lack of widespread availability and reluctance to implement this technique. 28 In this context, some groups have suggested that simplifying the laboratory procedure with fewer section levels may be clinically expedient. 29,30 For example, Bell et al. reported that a single H&E and IHC section of the SN was sufficient to provide an acceptable negative predictive value for the nodal basin status in a cohort of 35 patients. 29 Similarly, Jefferson et al. re-evaluated SNs from 10 patients previously assessed as tumor free on a single H&E with IHC and showed no micrometastases when these nodes were subsequently subjected to SSS with IHC. 30 Both these studies are limited by low patient numbers, low numbers or lack of positive SN in the test cohort and absence of negative controls. In contrast, our data show that while the majority (83%, Figure 2) of metastatic carcinomas are identified within the first four section levels, failure to sample beyond this may result in 16.7% of tumor deposits being missed, the majority of which were ITCs. We acknowledge that comparison of laboratory methods between studies are limited by their retrospective nature. Furthermore, the clinical significance of differing laboratory protocols remains largely unknown since there is a lack of comparative patient outcome data. Nevertheless, until such data becomes available, and in keeping with previous consensus guidelines, we recommend that exhaustive SSS with IHC should be undertaken on the entire SN. 31,32 The presence of ITCs in SLNB for OSCC remains controversial primarily since the clinical significance of this category is poorly understood. Two European prospective studies demonstrated that patients with ITCs had adverse outcomes compared to SN negative individuals despite going on to receive completion neck dissections. 6,33 Similarly, Trivedi et al. reported a neck recurrence rate of 30% in patients with ITCs. 34 By contrast, a recent prospective clinical trial showed that patients with ITCs who do not proceed to neck dissection had similar outcomes to those with negative SNs. 10 These apparent contradictory findings may be explained, at least in part, by the low ITC cohort sizes (n = 10-12) in these studies. The clinical significance of ITCs is further complicated by the somewhat arbitrary maximum size definition of <200 μm which is largely a result of extrapolation from other cancer sites. Moreover, some authorities also exclude contact with vessel or lymph sinus wall, extravasation, extravascular stromal reaction, and extravascular tumor cell proliferation as part of the definition of ITCs. 35 However, these additional criteria apply to other cancer sites and have not currently been validated in OSCC.
Since there is currently no low-end size cut-off definition, this category includes a broad histomorphological range from a single cell to clusters of up to 200. The fate of the single metastatic oral squamous carcinoma cell remains unknown and its histological detection is likely to be poorly reproducible, which may further explain the apparently contradictory clinical significance of ITCs. In our study, the majority (50.8%) of SNs diagnosed as positive by SSS with IHC were classified as ITC. Our findings indicate that SSS with IHC of the entire SN remains necessary because the current standard of care stipulates completion neck dissection following identification of ITCs. 15,36 Further work is required to elucidate the clinical significance of ITCs against patient outcomes, in particular whether the presence of a single cell mandates completion neck dissection. Until this is known, prospective acquisition of detailed morphological features of ITCs, including size, estimated number of cells, stromal appearance, and possible tumor and/or microenvironment molecular profiling may help to define a low-end cut-off value, above which completion neck dissection is indicated.
This study also questioned whether a reduction in workload achieved by increasing the interval thickness between section levels would compromise diagnostic accuracy. The rationale for intervals of 150 μm stems from the definition of ITCs (i.e., deposits of carcinoma cells up to 200 μm) and possibly assumes that disregarding any cells that would be otherwise present in the discarded material between section levels would not impact on patient outcomes. 32 Our data show that 47.5% of positive SNs requiring SSS contained carcinoma cells in a single section level. Interestingly, although the majority of carcinoma cells present in a single were Mi, the latter unequivocally requiring completion neck dissection.
The retrospective nature of our study raises several limitations: (1) any conclusions relating to interval thickness are based on multiples of 150 μm, and (2) ITCs may be present in the tissue discarded between section levels. However, the presence of Mi in single section levels, together with the undetermined size cut-off for clinically significant ITCs, indicate that the interval thickness should remain no greater than 150 μm. The labor-intensive nature of this protocol and its resultant burden on histopathology workload highlight the necessity for careful resource planning prior to implementing this service.
Multifocal deposits of carcinoma within a single SN have been previously described. 37 However, our data indicate that their occurrence in nonadjacent section levels is rare (5.9% of positive nodes). Classification of multifocal metastasis as Ma, Mi or ITC remains problematic as it is uncertain whether pathologists should categorize the SN according to the largest deposit or if a sum estimation of all foci should be made. Categorization based on total number of deposits are likely to upstage a subset of positive SNs (e.g., ITCs to Mi). This unresolved issue currently does not impact on subsequent patient management since any category of positivity necessitates completion neck dissection. However, as the low-end cut-off for ITCs becomes more clearly defined, the cumulative total size of multifocal deposits may be required to inform patient management. As such, accurate categorization according to total size estimates in multifocal deposits within a single SN is not possible without SSS and IHC.

| CONCLUSION
SSS with IHC of the entire SN 150 μm intervals are required to identify occult metastasis in OSCC. Further work is needed to determine the criteria for clinically significant ITCs.

ACKNOWLEDGMENT
The authors acknowledge the contribution of all biomedical science staff at Viapath Head & Neck Pathology, Guy's Hospital who undertook all laboratory technical work for this study.