One-step nucleic acid amplification for intraoperative analysis of sentinel lymph node in papillary thyroid carcinoma

Introduction

Prophylactic lymphadenectomy of the central neck compartment is a subject of controversy in the surgical approach to papillary thyroid carcinoma (PTC) (1). The high incidence of lymph node metastases (M1n) in this compartment; the difficulty of diagnosing such metastases before surgery and the potential avoidance of second surgery are cited as advantages of the procedure. In contrast, the doubtful prognostic benefits; the increased rate of surgical complications and the longer duration of surgery are arguments against prophylactic lymphadenectomy. The current recommendation is to perform lymphadenectomy whenever lymphatic metastatic disease (stage cN1 of the TNM classification) (2) has been demonstrated before or during surgery and as prophylactic treatment in high-risk patients. The American Thyroid Association (ATA) 2015 guide refers to these cases as corresponding to stages T3 and T4 of the TNM classification (3). Thus, the identification of lymphatic M1n disease of the neck is crucial for adequate patient treatment. However, 30–55% of the cases classified as cN0 and undergoing lymphadenectomy become pN1 disease (4, 5). As with other tumors, selective sentinel lymph node biopsy (SLNB) has been proposed with a view to improving diagnostic performance in M1n disease (6). This technique is used to identify the sentinel lymph node (SN), which represents the first tumor lymphatic drainage station. Different techniques based on the tumor or peritumor injection of a vital dye or radioisotope is used to locate the SN (7). The SNs are removed during surgery and after thyroidectomy and are analyzed intraoperatively. If the SN proves positive for tumor invasion, lymphadenectomy of the affected compartment is performed. In contrast, a negative SN implies the absence of M1n and avoids the need for lymphadenectomy. In a previous study we were able to confirm the viability and efficacy of the technique (8). However, a possible limitation of the procedure was an 8% incidence of false-negative (FN) results in the intraoperative analysis by conventional pathological evaluation on lymph node cytological imprints and/or on frozen section. Intraoperative diagnostic errors have received a number of explanations, including an inevitable random factor in lymph node sampling, and interpretation errors due to the low quality of the morphological images afforded by the frozen section. A recent review on the value of SLNB in PTC reports an even higher incidence of FN results (14.5%) (9). Thus, the definitive implantation of SLNB requires optimization of the intraoperative study of the SN, with the aim of avoiding FN results. This limitation is common to other neoplasms in which SLNB is used, with similar FN rates (10, 11, 12). In recent years, one-step nucleic acid amplification (OSNA) has been introduced in the intraoperative analysis with the purpose of reducing these rates, particularly in application to SN in breast cancer (13, 14). This quantitative molecular technique allows the real-time detection of mRNA encoding for cytokeratin 19 (CK19), specifically amplifying mRNA in a single step through reverse transcriptase loop-mediated isothermal amplification (RT-LAMP) (15). One-step nucleic acid amplification is highly reliable in analyzing the entire lymph node in the same intraoperative procedure, affording rapid, definitive, precise and automated information about lymph node involvement. The fact that OSNA is a quantitative technique offers added value, since it can not only identify the presence or absence of metastasis, but can also quantify lymph node tumor burden. Intense CK19 expression is known to be present at membrane level in approximately 70% of all PTCs (16), and such positivity is therefore also seen in the metastatic cells within the lymph node. No other elements of the node show CK19. This has led to consideration of the application of OSNA in this context. A previous study in collaboration with other Spanish centers demonstrated good correlation between OSNA and the conventional lymph node study in patients undergoing surgery due to PTC and to standardized neck lymphadenectomy, with a very similar to cut-off value of positivity for the copies of CK19 mRNA/µL accepted in breast cancer (17, 18). The present study was carried out to analyze the results obtained with OSNA in the intraoperative evaluation of SN in a group of patients subjected to SLNB.

Subjects and methods

Between July 2013 and April 2016 we prospectively and consecutively included all patients over 18 years of age that met the following criteria: (1) Newly diagnosed PTC with cytological confirmation through fine needle aspiration biopsy (FNAB); (2) No evidence of neck M1n before surgery (cN0); (3) No evidence of locally advanced disease or distant metastatic spread; (4) Immunohistochemically confirmed tumor cell CK19 expression in the diagnostic FNAB sample and (5) Signing of the informed consent. This document was approved (with date 01/07/13) by the Clinical Research Ethics Committee of the center (number of reference PR (SC) 192/2013). A total of 35 patients were finally enrolled in the study.

Sentinel lymph node biopsy technique: On the day before surgery, 0.1–0.2 mL of 99mTc-nanocolloid (Nanocoll) of particle size 50–80 nm and with an activity of 148 MBq was injected into the tumor under ultrasound guidance. This was followed 2–4 h later by SPECT/CT (GE, Infinia Hawkeye 4). The projections of the SNs detected by the imaging technique were subsequently marked on the skin. On the next day, total thyroidectomy was performed, and the SNs were located with a gamma ray detector probe (Europrobe; Eurorad, Eckbolsheim, France). The SNs were resected and sent to the pathology laboratory. In the case of confirmed M1n in the SNs, lymphadenectomy of the corresponding compartment was performed. When the affected SN was located in the lateral compartment (LC), lymphadenectomy was completed with lymph node resection of the central compartment (CC). Lymphadenectomy was left at the discretion of the surgeon when suspect lymph nodes were detected independently of the results of the SN study.

Pathological processing and OSNA technique: The SNs were sent fresh to the pathology laboratory. The lymph nodes were immediately dissected, with separation of the possible accompanying non-lymph node adipose tissue for later conventional histological study. The lymph nodes were measured and weighed. Central sectioning of each lymph node was carried out, and two cytological imprints were obtained from the exposed surface. One was stained with Diff-Quik (Química Clínica Aplicada, S.A., Spain) and evaluated immediately to morphologically confirm the lymph node nature of the sample. The complementary imprint was later stained with the Papanicolaou method (Merck KGaA, Germany). Following initial cytological verification, the SN was subjected to OSNA processing, and in the event of a positive result the surgeon was immediately informed in order to complete the corresponding lymphadenectomy. The approximate waiting time for the surgeon with OSNA is around 45 min, discretely superior to that necessary for the study of the nodes by conventional way. In those isolated cases where the cytological study of the SN unequivocally confirmed the presence of metastasis, the corresponding lymphadenectomy was indicated from the start, and since surgical decision did not depend on the metastatic burden of the lymph node, its molecular study could be deferred. The histological material of the lymphadenectomies was likewise sent fresh to the pathology laboratory, with immediate dissection and identification of the lymph nodes for deferred OSNA study. In the case of deferred studies, the lymph nodes were cryopreserved in liquid nitrogen or cold isopentane and kept at −80°C until the time of analysis. In all cases the lymph node tissue was processed by mixing it in a separate tube with 4 mL of Lynorhag homogenization buffer (pH 3.5) that stabilizes the mRNA and precipitates the genomic DNA with no previous nucleic acid extraction or purification procedure. The lymph nodes were homogenized (Polytron System PT1300D, Kinematica AG, Switzerland) during 60 s. From the resulting sample we pipetted 1 mL into an Eppendorf tube, followed by centrifugation at 10 000 g at room temperature for 1 min. After centrifugation, the supernatant was diluted 1:10 and 1:100 with Lynorhag buffer and analyzed with the RD-100i−OSNA system (Sysmex Corp., Japan), which performs fully automated and standardized amplification and detection of CK19 mRNA within 16 min, using the Lynoamp reagent. The system measures the time taken by the test sample to reach specific turbidity (rise time), and a standard rise time vs concentration (copies of CK19 mRNA/μL) curve previously obtained from three calibration controls is used to calculate the concentration of CK19 mRNA molecules in the study sample. The OSNA technique defines positivity as ≥250 copies/μL. Additional cut-off points are defined to distinguish between micrometastasis – (mM, ≥250 copies/μL and <5000 copies/μL) and macrometastasis – (MM, ≥5000 copies/μL). The RD-100i−OSNA system is able to process up to four samples simultaneously. From the homogenized sample in the conical base tube, we pipetted another 1 mL into an Eppendorf tube (backup sample). This tube was frozen at −80°C for eventual reanalysis of the lymph node if needed (Fig. 1).

Pathology workflow and surgical decision algorithm (OSNA: one-step nucleic acid amplification). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Pathology workflow and surgical decision algorithm (OSNA: one-step nucleic acid amplification). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

Pathology workflow and surgical decision algorithm (OSNA: one-step nucleic acid amplification). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

Results

The clinical and histological characteristics of the patients are shown in Table 1. Immunohistochemical CK19 positivity of the primary tumor was confirmed in all the diagnostic FNAB samples of the patients. None of the candidate subjects were therefore discarded for this reason. A total of 110 SNs were obtained, with a median (IQR) of 3 (1–4) per patient (minimum 1 and maximum 7). Five of the samples submitted as SN were discarded for intraoperative analysis, since the initial confirmatory cytological evaluation failed to evidence the presence of lymphatic tissue. The subsequent histological study was able to identify sub-millimetric lymph nodes in three of these five samples; another consisted only of adipose tissue and the fifth corresponded to parathyroid gland tissue. The parathyroid tissue was subjected to OSNA on a deferred basis, yielding an expected false-positive result of the molecular technique in the form of mM.

Conventional cytology found 25 of the 110 SNs (22.7%) to be positive for metastasis with one or both stains (Diff-Quik and/or Papanicolaou), while 42 (38.2%) were found to be positive with OSNA (P < 0.001). The OSNA-positive cases corresponded to 21 mM and 21 MM. All the discrepant SNs were OSNA positive and cytologically negative, and corresponded to 15 mM and 2 MM. These results indicate that SLNB proved positive in 14 patients (40%) in the cytological study and in 23 patients (65.7%) with the OSNA technique (P < 0.001).

Lymphadenectomy was performed in 29 patients. These included 23 cases with positive SLNB findings and 6 in which lymphadenectomy was performed despite the fact that the SNs were reported to be negative. In these six cases the surgeon suspected pathological lymph nodes. Five of the six cases did not reveal M1n following lymphadenectomy, while the sixth case proved positive for MM. This latter patient represented a special case, since, despite the intraoperative submission of histological material identified as SN from the central compartment (CC), the gross examination failed to reveal the presence of lymphatic tissue, and no molecular analysis was therefore decided. The deferred study of this material, fixed in formalin solution and conventionally processed with embedding in paraffin, confirmed the presence of a positive sub-millimetric lymph node from the same compartment of the MM already suspected by the surgeon. Of the 23 patients undergoing lymphadenectomy due to positive SLNB, the SNs were found to be the only positive nodes in 6 cases.

In the 29 patients subjected to lymphadenectomy we obtained a total of 360 lymph nodes (92 SNs and 268 non-SNs) – 61% corresponding to CC and the rest to LC. The median (IQR) number of total lymph nodes per patient was 11 (6.5–17.5) (minimum 3 and maximum 31). The median lymph node weight was 0.09 g (0.05–0.17) and the diameter was 0.5 cm (0.3–0.7). There were no differences in these parameters between SNs and non-SNs. Of the total 360 lymph nodes, 52 were positive for M1n in the cytological study (14.4%) and 107 in the OSNA analysis (29.7%). There were 59 discordant cases: OSNA positivity and cytological negativity in 57 cases (47 mM and 10 MM), and OSNA negativity and cytological positivity in the remaining two cases (both corresponding to non-SNs). Sixty-four of the 107 OSNA-positive cases corresponded to mM and 43 to MM. The median number of CK19 mRNA copies/μL in these lymph nodes was 2700 (895–20 938) (minimum 250 and maximum 1 500 000). There were no differences in weight (P = 0.62) or diameter (P = 0.37) between the OSNA-positive and -negative lymph nodes. The OSNA-positive lymph nodes likewise showed no significant correlation between the number of mRNA copies and weight (r = 0.004; P = 0.9) or size (r = −0.05; P = 0.58).

These results indicate that lymphadenectomy proved positive in 16 patients according to the cytological study (55%) and in 24 patients according to the OSNA technique (83%) (P = 0002). Therefore, the FN ratio for cytology in lymphadenectomies was 28%. There were no significant differences in the total number of resected lymph nodes between the positive and negative lymphadenectomies. In the positive lymphadenectomies, the median number of M1n cases detected by OSNA was 4 (1–7). Of the 23 cases with positive SLNB according to OSNA, MM was detected in 12 of the SNs, while the remaining 11 only revealed mM in the SNs. Among the former, nine patients also showed MM in the lymphadenectomy, while among the latter only one lymphadenectomy revealed MM (P < 0.001). In accordance, 10 out 24 patients with M1n carried only mMs.

Figure 2 summarizes the findings of the study, comparing the results of the cytological study vs those of the OSNA technique.

Comparison between cytology and OSNA results (LN, lymph node; OSNA, one-step nucleic acid amplification; SN, sentinel node).

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Comparison between cytology and OSNA results (LN, lymph node; OSNA, one-step nucleic acid amplification; SN, sentinel node).

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

Comparison between cytology and OSNA results (LN, lymph node; OSNA, one-step nucleic acid amplification; SN, sentinel node).

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

Discussion

The value of SN using the OSNA technique (SLNB-OSNA) is widely accepted in breast cancer (www.nice.org.uk/guidance/dg8; www.rcpath.org/resourceLibrary/g148-breastdataset-hires-jun16-pdf.html). More recently, OSNA has started to be applied in other tumors (19, 20, 21, 22, 23), though its application to PTC has not been evaluated to date. This is therefore the first study to examine the usefulness of SLNB-OSNA in this type of tumor. In a previous collaborative study, we found a good correlation between the conventional cytological and histological data and the OSNA technique in the analysis of lymph nodes from patients with PTC (17). Similar results were reported by Kaczka et al. (24). The present study analyzes the results obtained with OSNA in application to SLNB in a group of 35 patients with PTC.

In contrast to the situation in other types of tumors, the use of SLNB for the study of M1n in PTC is not widely accepted (25). One of the arguments against its use is the high FN rates in the intraoperative study, i.e., SN analysis proves negative, while the definitive paraffin-embedded study proves positive. In a recent systematic review, Portinari and Carcoforo (9) identified three publications including a total of 194 patients subjected to intraoperative evaluation of the SN. The authors recorded an FN rate of 14.5% in the pathology study. In a study by our group, the frozen section study yielded a FN rate of 8.3% (8). Assessment and the drawing of firm conclusions in this regard are difficult, however, since consensus is lacking in the histological study of the lymph nodes. This is so for both intraoperative analysis (frozen section study alone, frozen section plus cytology) and for definitive deferred analysis (single central section, serial sectioning of the node, immunohistochemical study). In breast cancer, where the SLNB technique is more widely used and greater experience has been gained, a review conducted before introduction of the molecular techniques identified up to 123 different protocols for studying the SN (26). In this regard, improving the results obtained and standardizing the intraoperative diagnosis could contribute to expand the use of this technique.

One-step nucleic acid amplification has been successfully used for over 10 years for the intraoperative analysis of SN involvement in breast cancer (27, 28, 29). The first OSNA validation studies in this context defined precise equivalence between the amount of CK19 mRNA in the lymph node and the magnitude of M1n and made it possible to distinguish between mM (<2 mm) and MM (>2 mm) (13). The concepts of mM and MM according to OSNA were defined from the histological definition originally established in breast cancer, with subsequent extrapolation to other types of tumors. Although strictly speaking the criteria of mM and MM have not been categorically defined in PTC, the concept of small or large lymph node metastasis does appear to have clinical significance (3). We therefore consider it adequate to extrapolate these criteria to our clinical context.

As in other tumors, an essential requirement before applying OSNA to PTC is the preoperative verification of immunohistochemical CK19 expression in the tumor cells – such expression being maintained as well in the metastatic epithelial cells of a hypothetically affected lymph node. For this reason immunohistochemical analysis was performed in the initial diagnostic FNAB sample in all the patients – no subjects having been discarded due to CK19 negativity. This 100% positivity rate is higher than the values published in the literature (around 74%) (16). In breast cancer, different clinical guides specifically recommend preoperative immunohistochemical verification of the expression of the protein in the initial diagnostic sample (www.nice.org.uk/guidance/dg8, 30).

We considered cytological study of the lymph nodes before OSNA to be essential, since molecular processing of the entire lymph node impaired processing the tissue for histological examination. Cytological evaluation of the sample was the first step in the intraoperative study for confirming the tissue as corresponding to lymph node and discarding other neck structures such as parathyroid gland tissue, thymus or salivary gland tissue, characterized by CK19 expression and anatomically located close to the resected lymphatic compartments. In our series, one purported SN that was discarded intraoperatively was finally found to correspond to parathyroid gland. This particular case therefore would have represented a false-positive OSNA result if previous intraoperative cytological study had not been performed. On the other hand, in all cases where the intraoperative cytological study of the SNs proved unequivocally positive for metastasis, and since at this point M1n volume did not influence the subsequent surgical decision, we indicated lymphadenectomy without waiting for the molecular study and without observing subsequent discrepancies. This reaffirms the strong positive predictive value of the cytological study.

Our series includes two cases of non-SN nodes in which cytology proved positive while the OSNA findings were negative. This apparent discrepancy can be explained by the fact that we are comparing a quantitative technique (OSNA), which requires a minimum expression threshold in order to classify a sample as metastatic, versus a qualitative technique (cytology) where any evidence of malignancy is classified as a positive result. In both cases the presence of M1n was verified by the immunohistochemical study of the cytological imprint, with the observation of very few atypical epithelial cells (Fig. 3). The two mentioned lymph nodes were classified by OSNA as ‘isolated tumor cells’. While there is agreement that this diagnostic category should be classified as negative, it in fact indicates the presence of very few metastatic cells in the lymph node (≤0.02 mm according to the histological evaluation – the OSNA equivalent being 100–250 CK19 mRNA copies). Consequently, there would be no true discrepancy but rather a disadjustment in the established molecular positivity thresholds.

Immunohistochemical Cytoqueratin 19 positivity in imprint cytology (Cell Marke A53-B/A2.26, USA, ×600). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Immunohistochemical Cytoqueratin 19 positivity in imprint cytology (Cell Marke A53-B/A2.26, USA, ×600). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

Immunohistochemical Cytoqueratin 19 positivity in imprint cytology (Cell Marke A53-B/A2.26, USA, ×600). A full colour version of this figure is available at https://doi.org/10.1530/EJE-18-0624.

Citation: European Journal of Endocrinology 180, 1; 10.1530/EJE-18-0624

• Download Figure
• Download figure as PowerPoint slide

In our study we found OSNA to be more sensitive than conventional cytology in detecting M1n. In effect, among the more than 350 analyzed lymph nodes, the OSNA positivity rate doubled that of the cytological study, with an increase from 40% to 65% in the number of metastatic SLNB and from 55% to 83% in the number of positive lymphadenectomies. The cytology FN ratio of 28% observed in lymphadenectomies is higher than previously published data with respect to conventional morphology (usually cytology plus frozen section) (8, 9). This is explained because we compare the morphological technique with lower diagnostic yield (cytology only) with the most advanced technique (OSNA), with greater ability to detect metastasis even than histology. Most of the discordant cases were due to OSNA positivity and cytological negativity. This high sensitivity results in an increase in the number of lymph node resection procedures; as a result, the technique could lead to unnecessary lymphadenectomies. The prognostic relevance of M1n in PTC has not been fully established. Some studies indicate that only patients with high tumor load metastases are at an increased risk of disease recurrence. This tumor load has been characterized in terms of the number and size of the affected lymph nodes (3). No precise quantitative assessment of lymph node tumor mass has therefore been established to date. One-step nucleic acid amplification could contribute the required parameter. Although our sample is small, it should be emphasized that in 75% of the cases in which MM was detected in the SNs, we also detected MM in the subsequent lymphadenectomy. In comparison, when only mM was present in the SNs, we only found MM in 9% of the corresponding lymphadenectomies. These findings suggest that the distinction between mM and MM in SLNB could guide the decision to perform lymphadenectomy, and in this regard we could consider the possibility of avoiding node resection if MM is not confirmed in the SN. Therefore, what might initially seem to be an over-detection of metastasis ends up being, with the formulated approach, a more precise decision algorithm and aimed at carrying out lymphadenectomy only in those patients with evidence of high volume metastases. Based on this reasoning, ‘verifying’ cytology before OSNA will continue to play a role in that it can confirm that the sample consists of lymph node tissue. However, in view of its qualitative rather than quantitative nature, cytological analysis would no longer prove sufficient to determine whether subsequent lymphadenectomy is required or not. In this regard, the quantitative nature of OSNA would be more reliable than cytology, and diagnosis moreover may be standardized and automated.

One-step nucleic acid amplification also has some limitations. One such limitation would be the decision whether or not to perform the molecular analysis in those intraoperative SN samples in which thorough gross examination fails to identify lymph node tissue. In our series there were four samples meeting this criterion in which the deferred histological study confirmed the presence of non-palpable sub-millimetric SNs. In one of these cases the undetected SN revealed mM in the histological study and was moreover accompanied by MM in the same lymphatic compartment resected in response to surgeon suspicion, despite the absence of intraoperative evidence of an affected SN in the corresponding compartment. Thus, when the surgeon firmly suspects lymph node metastasis, his or her criterion must prevail over any decision algorithm, and the suspect lymph node should be removed. Another limitation of the OSNA technique is the impossibility of conducting an adequate histopathological study of the lymph node. This for example precludes the evaluation of extra-nodal disease spread in M1n cases or the identification of concomitant node disease. This latter aspect explains why, for example, some guides recommending SN analysis with OSNA in breast cancer explicitly advocate a parallel morphological study (www.https://link.springer.com/chapter/10.1007/978-3-642-13666-5_873).

In conclusion, the present study shows SLNB with intraoperative analysis of the SNs using the OSNA technique to be feasible in the clinical context of PTC. It do not imply any modification in the standard SLNB surgery for thyroid cancer. Furthermore, the results obtained indicate that, with very few exceptions, the presence of mM in the SN is associated to negative lymph node resection findings or lymphadenectomy exhibiting only mM. This fact, and demonstration of the usefulness of OSNA as an intraoperative strategy for the quantification of M1n, with the existing evidence of the scant clinical impact of small-volume M1n in PTC, pave the way toward a more personalized surgical approach – limiting lymphadenectomy to cases with intraoperative evidence of MM in the SNs. Definitive confirmation of our study hypothesis will require the analysis of larger patient samples, as well as corroboration of the findings by other investigators.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study.

Funding

The study was sponsored by Sysmex España S.L.

Author contribution statement

C I F, C Z L, I R B, J M M, J C C, J M F L-B, S R y C A and O G L designed the study; A G B, X S C, I R B and J C C injected the sentinel nodes; O G L, E C B and J M F L-B performed the surgery; C I F and J T-S carried out pathological and molecular processing; C I F and C Z L collected and analyzed the data, made the literature search and wrote the manuscript; C I F, C Z L, O G L and R P-B interpreted the results; C I F, C Z L and J T-S generated the figures; R P-B carried out a critical revision of the manuscript.