Multiplex Intraoperative Rapid Immunohistochemistry with Noncontact Antibody Mixing for Distinguishing the Histologic Phenotype of Lung Cancer

Early-stage non-small cell lung cancer (NSCLC) is generally treated with surgical resection, and radical lobectomy has been the standard procedure [1]. However, the JCOG0802/WJOG4607L phase III trial showed that 5-year overall survival for segmentectomy was superior to standard lobectomy for clinical stage IA NSCLC [2]. On the other hand, segmentectomy is associated with a twofold higher incidence of local recurrence than lobectomy, which highlights the importance of selecting suitable patients for segmentectomy. Whether lobectomy, segmentectomy, or other surgical procedures are used often depends on the intraoperative pathological diagnosis. The preoperative diagnosis of small pulmonary nodules can be difficult through transbronchial or transthoracic biopsy, depending on the size and location of the nodules. Intraoperative consultation with frozen sections (FS) is frequently employed in clinical practice to establish the initial surgical strategy for primary lung cancer or lung metastases in these patients [3]. In addition, a large National Cancer Database (NCDB) study of 143,040 patients with lymph node-negative early NSCLC, which is well-feasible radical segmentectomy indication, yielded that the desirable threshold for tumor size requiring lobectomy varies depending on histological findings [4]. Thus, accurate intraoperative histologic diagnosis is essential for selecting the optimal oncologic treatment and surgical approach.

For prompt intraoperative diagnosis, hematoxylin and eosin (HE) staining of FS represents the gold standard when evaluating lung lesions for tumor tissue sampling. However, with poorly differentiated tumors and small nodules less than 1 cm in size, it is sometimes difficult to accurately classify the histologic type based on HE staining alone [4‒6]. By contrast, immunohistochemistry (IHC) is a reliable screening and molecular analysis method. However, using IHC for intraoperative diagnosis has not been possible because IHC involves time-consuming processing. To overcome that limitation, we have been developing a rapid IHC method that makes use of an alternating current (AC) electric field to facilitate the antigen-antibody reaction by stirring the antibody-containing solution on the sections without a stirrer through repeated transformation of the microdroplet’s shape [7‒13]. The resultant AC mixing enables more prompt and accurate diagnosis/molecular analysis by increasing the opportunity for contact between the antigen and antibody, irrespective of the antibody type [10, 12, 13]. With this rapid IHC technique, we could stably detect target cells within FS and intraoperatively diagnose within 20 min, as opposed to the 3–6 h required for conventional IHC. In our earlier prospective multicenter study, rapid IHC with AC mixing on FS improved accuracy when diagnosing the histological type and/or the organ origin of undiagnosed pulmonary tumors, yielding an overall definitive diagnosis rate of 88.76% [13]. In addition, we have reported that this technique is applicable to multiple tumor types and organs, such as breast cancer and central nervous system tumors [8, 14]. The AC mixing technique has the potential to enable more accurate pathological diagnosis and guide toward desired therapeutic strategies.

Various molecular biomarkers associated with NSCLC have been identified in recent years, and innovative therapeutic agents for these biomarkers have been developed. As a result, molecular biomarker testing has become crucial for determining treatment approaches for NSCLC, and accordingly, more tissue slides have been required [15]. Pathologists and surgeons have continuously dealt with an unresolved dilemma of exploring a growing number of molecular biomarkers with just a single small tumor sample to improve tailor-made therapies. In that context, multiplex IHC has emerged as a potent tool for simultaneous detection of multiple biomarkers on the same tissue section, thereby expanding molecular/immune profiling while preserving tumor material [16]. This has improved pathological diagnosis by enabling the evaluation of tumor biological characteristics that are hard to reliably access through HE staining alone, including comprehensive information on biomarker expression levels, co-localization, and compartmentalization.

As a successful pulmonary sublobar resection, including segmentectomy, must be oncologically adequate, the accuracy of a FS diagnosis is more important than ever. Here, we describe the development and validation of a rapid multiplex IHC staining protocol labeling thyroid transcription factor-1 (TTF-1) + cytokeratin (CK) 5, desmoglein 3 (DSG3) + Napsin A, and p63 + tripartite motif containing (TRIM) 29. We aimed to evaluate the clinical reliability of an added rapid multiplex IHC technique with AC mixing for intraoperative FS diagnosis in lung cancers.

We first sought to establish rapid multiplex IHC protocols by comparing single and multiplex rapid IHC staining using formalin-fixed, paraffin-embedded (FFPE) tissue blocks from 10 patients who received radical lung cancer surgery (5 adenocarcinomas [ADs], 5 squamous cell carcinomas [SCCs]) at our institute in 2020 (Table 1). We then sought to verify the diagnostic performance of multiplex rapid IHC staining using 43 FS samples from patients who underwent radical lung cancer surgery between April 2021 and June 2022. These patients’ clinical characteristics are listed in Table 2. A diagram of the process by which cases were selected for study is shown in Figure 1.

All patients received standard preoperative and intraoperative care. Lung cancer was treated with segmentectomy or lobectomy plus systematic lymph node dissection under one-lung ventilation. Using resected lung tissue or a core needle biopsy specimen, tumor histologic type was assessed with FS examination and rapid IHC in all eligible patients. Our pathologists made intraoperative diagnoses of undiagnosed pulmonary tumors based solely on HE staining but retained the optional diagnosis made with rapid IHC. The rapid IHC was only a guide in the present study.

We used the Histo-Tek R-IHC^™ device to apply an AC electric field (Sakura Finetek Japan Co., Ltd, Tokyo, Japan) as described in earlier reports [7‒9]. The theory behind AC electric field mixing and the method for its application were previously described in detail (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000539640) [7‒13]. Briefly, we use the device to apply a high-voltage (4 [−4.5] kV, offset 2.4 kV), low-frequency (15 [5–90] Hz) AC electric field to the sections. This causes the antibodies to be mixed within microdroplets as the voltage is switched on and off at regular intervals, changing the droplet’s shape. The available antibodies and protocols are referenced at the R-IHC Study Group website (http://www.rihc.jp).

Online supplementary Table 1 summarizes the staining protocol for conventional IHC and the new rapid IHC using AC mixing. A panel comprising 6 primary monoclonal antibodies (TTF-1, Napsin A, CK5, DSG3, p63, TRIM29) was used for multiplex IHC on FS or FFPE tissue blocks. This commercial-use panel was designed by Tacha et al. [17], and six antibodies with higher sensitivity and specificity were selected based on the results of their pilot study for 15 antibody staining. It is composed of 3 double-stained marker panels (TTF-1 + CK5, DSG3 + Napsin A, and p63 + TRIM29; Biocare Medical, inc.). TTF-1 (nuclear), DSG 3 (membrane), and p63 (nuclear) were visualized 3,3′diaminobenzidine (DAB), while CK5 (cytoplasm), Napsin A (cytoplasm), and TRIM29 (cytoplasm) were visualized fast red. This panel was previously shown to accurately distinguish between AD and SCC in 93% of cases [17].

FS slides for HE staining and rapid IHC were simultaneously prepared intraoperatively from the same pulmonary tumor samples. Pulmonary tissues were cut at optimal intervals, then immediately embedded in optimum cutting temperature compound, frozen for 30 seconds in liquid acetone at −80° and transferred to a cryostat for sectioning. Serial FS were cut at 5 μm and placed on slides, fixed in acetone for 30 s, air-dried for 30 s at room temperature, and 3 slides were used per case. After rinsing the slide with water, the endogenous peroxidase and alkaline phosphatase are blocked. For staining, each slide was incubated for 5 min with a primary antibody cocktail (TTF-1 + CK5, DSG3 + Napsin A, and p63 + TRIM29; Biocare Medical, Inc.) under AC electric field (total 3 slides). The sections were then washed with Tris-buffered saline and incubated for 5 min with Biocare’s MACH2 Double Stain 2 (Biocare Medical, Inc.) under the same AC electric field. After again washing with Tris-buffered saline, the sections were developed with DAB substrate and Biocare’s Warp Red (Biocare Medical, Inc.), counterstained with hematoxylin, dehydrated, and mounted on coverslips. Online supplementary Table 1 summarizes the procedures for conventional IHC and the new rapid IHC using AC mixing.

Two board-certified pathologists at our hospital evaluated the specimens for this study. All dissected tumors and surgical margins were sectioned and examined using HE staining and conventional IHC with FFPE tissue blocks. The intraoperative multiplex IHCs result requires approximately 30–40 min between the removal of the surgical specimen and the communication of the final result. For each stain, positive IHC staining was defined based on the WHO Classification of Tumors Series (https://tumourclassification.iarc.who.int/welcome/).

Statistical analysis was performed using JMP IN 17.0.0 software (SAS Institute, Cary, NC, USA). The overall definitive diagnosing rate providing both histological type and organ origin was calculated by excluding cases that had suspicious diagnoses. Cohen’s kappa coefficient (κ), along with 95% confidence intervals, were used to assess agreement of 4 × 2 contingency tables between protocols. Cohen’s κ can be interpreted as: 0.81–1, almost perfect agreement; 0.61–0.80, substantial agreement; 0.41–0.60, moderate agreement; 0.20–0.40, fair agreement; and 0–0.20, slight agreement [18].

To develop the multiplex IHC standard protocols for lung cancer, we first compared the results obtained with conventional IHC and rapid multiplex IHC with AC mixing using FFPE samples from 10 NSCLC patients (5 ADs, 5 SCCs). We found that the results obtained with rapid multiplex IHC were consistent with those obtained using conventional IHC with each antibody (Fig. 2).

Next, the accuracy of rapid multiplex IHC for diagnosing lung cancer histology with FS was evaluated using 43 undiagnosed lung specimens collected at our institute. No false diagnosis of malignancy was rendered in any case when intraoperative HE staining was used alone. In addition, there was no difference in the diagnoses of AD and SCC between conventional IHC and rapid IHC with AC mixing (Fig. 3).

Among the 28 cases of AD diagnosed using FFPE tissue blocks, 22 (78.6%) were positive for TTF-1, and 26 (92.9%) were positive for Napsin A (Fig. 4a). Among the 12 cases of SCC diagnosed using FFPE samples, all were positive for CK5, DSG3, and TRIM29, while 8 (67%) were positive for p63 (Fig. 4b).

The addition of rapid multiplex IHC led to the accurate diagnosis of three cases that were diagnostically challenging with HE staining alone. Figure 5 presents two cases where there were noteworthy discrepancies in the histological type diagnosed with HE staining alone and the two IHC techniques (conventional IHC with FFPE and rapid multiplex IHC with FS/FFPE). One case was a poorly differentiated grade 3 tumor diagnosed as AD based on HE staining. The diagnosis from both IHCs was SCC, which was confirmed by a final pathological examination. The second case was a moderately differentiated grade 2 tumor initially diagnosed as SCC based on HE staining of FS alone, but the diagnosis from the IHCs was AD, which again was confirmed by the final pathological examination.

We found 93.0% (40/43) agreement between intraoperative FS diagnoses obtained using rapid multiplex IHC and the final pathological diagnosis with FFPE-IHC (Cohen’s 𝜅 coefficient = 0.860 and 95% CI: 0.727–0.993). In three cases, the tumor could not be diagnosed using rapid multiplex IHC. These included one case of pleomorphic carcinoma and two cases of small cell carcinoma (SCLC). This multiplex IHC did not include antibodies for pleomorphic carcinoma and did not show staining with any of the six antibodies. Two cases had the potential to be diagnosed as AD or SCLC due to the only TTF-1 positivity on rapid IHC alone, with all other markers being negative. However, these cases were diagnosed as one SCLC and one AD using both HE staining and rapid multiplex IHC. Limiting to the diagnosis of AD or SCC, the diagnosis of lung cancer histological type was achieved with an accuracy of 100%.

In the present study, we demonstrated that rapid multiplex IHC with AC mixing (using TTF-1 + CK5, DSG3 + Napsin A, and p63 + TRIM29) on FS improves the accuracy of intraoperative diagnoses as compared to HE staining alone. When rapid multiplex IHC was added, the definitive diagnosis rate distinguishing between AD and SCC was 100%. Rapid multiplex IHC thus appears to be an accurate method to guide diagnosis of histological type in NSCLC.

Most highly or moderately differentiated lung carcinomas can be classified into histologic types based on morphologic diagnosis with HE staining alone. However, accurately classifying poorly differentiated lung carcinomas requires IHC staining using multiple antibodies. Several reports suggest that a panel of several antibodies can be used to classify tumors as AD or SCC with high accuracy [17, 19, 20]. Using a previously described six-antibody panel in the present study, we were able to classify poorly differentiated lung carcinomas with 87% accuracy [17]. Using the rapid multiplex IHC method with AC mixing and the same antibody panel, we were also able to classify all ADs and SCCs accurately.

TTF-1 is a member of the NKX2 gene family [21] and the most useful marker for distinguishing lung AD from other subtypes of lung cancer or cancers originating from other sites [22, 23]. TTF-1 positivity is detected in 60–80% of lung ADs [24, 25], but it has also been reported in 80–97% of SCLCs and 50–75% of large cell neuroendocrine carcinomas [26‒28]. Consequently, careful interpretation is necessary when diagnosing pulmonary/extrapulmonary neuroendocrine cancers [23, 29]. In the present study, two tumors with TTF-1 positivity alone had the potential to be either AD or SCLC, and one was diagnosed as AD based on HE staining. Thus, combining HE staining with rapid multiplex IHC leads to more reliable diagnoses.

Napsin A is a well-established immunohistochemical marker that has garnered attention as a highly specific marker for distinguishing lung AD from SCC [30‒32]. In our study, 92.9% (26/28) of patients diagnosed with lung AD in the final pathology were positive for Napsin A. Consistent with our results, several other studies have reported similar sensitivities ranging from 73.8% to 90.7%. However, with rapid multiplex IHC on FS, 1 patient (1/28) with lung AD had a false negative result. In our earlier multicenter study, we also experienced difficulty staining for Napsin A [13], emphasizing the need for careful interpretation to avoid false negatives. Nonetheless, considering its potential usefulness in cases where TTF-1 staining is unclear, we recommend including Napsin A when using rapid IHC.

CK5/6 and p63 are markers specific for most SCCs [20]. A study of mRNA expression found that moderate to high expression of CK6 was detected in 28% ADs, while CK5 expression was more specific for SCC [33]. Therefore, CK5 antibody rather than CK5/6 was selected for this panel. Human p63 is a homolog of the p53 tumor suppressor and is a nuclear antigen found in basal epithelial cells [34]. Among lung cancers, p63 immunoreactivity is most common in SCCs and high-grade neuroendocrine carcinomas [35]. However, p63 staining was weak when using multiplex IHC on FS. While it is common to detect p63 expression in ADs, expression of p40, a biomarker corresponding to the non-transactivating isoform of p63 (delta Np63), is specific for SCC [13, 36]. Therefore, with multiplex IHC, p40 is a potentially more reliable marker for lung SCC.

DSG3 and TRIM29 are newer antibodies yet to be studied to the extent of those addressed above [33, 37‒40]. DSG 3 is a member of the cadherin cell adhesion molecule superfamily and a calcium-binding transmembrane glycoprotein component [33]. DSG3 had 88% sensitivity and 98% specificity in classifying AD or SCC, making it a potential biomarker for lung cancer. In addition, it has been suggested that DSG3-positive lung cancers are associated with a favorable prognosis [37]. TRIM29 is a member of the TRIM protein family and has been suggested to play a role in cancer progression, differentiation, and proliferation in gastric, bladder, and gynecological cancers [38, 39]. Its role in lung cancer remains unclear, however, with only one report associating high expression of TRIM29 with shortened survival in lung cancer [40]. When previously used in the same six-antibody panel of markers used in the present study, TRIM29 was reportedly stained in 7.0% of ADs but in 93.7% of SCCs. In the present study, all cases of SCC showed TRIM positivity, while one case (3.6%) of AD was positive. These findings suggest that TRIM29 may be suitable for distinguishing between AD and SCC, but its utility requires further validation.

We have previously reported on the utility of AC mixing technique for determining the origin of metastatic lung tumors, detecting metastases from breast cancer, and distinguishing central nervous system tumors [8, 14]. Additionally, the AC mixing technique is beneficial for expediting and stabilizing in situ hybridization detection of target genes such as echinoderm microtubule-associated protein-like 4 and anaplastic lymphoma kinase (EML4-ALK) and human epidermal growth factor receptor 2 (HER2) [41, 42]. Moreover, stable PD-L1 IHC with AC mixing has contributed to improving tumor proportion score (TPS) scoring and patient selection for immune checkpoint inhibitors through compatible assays [12]. Rapid Ki-67 IHC with AC mixing may be useful for intraoperative assessment of tumor malignancy [43]. In our earlier prospective multicenter study, we reported that rapid IHC on FS improves the accuracy when diagnosing histological type/origin of undiagnosed pulmonary tumors as compared to HE staining alone. When rapid intraoperative IHC was added, the overall definitive diagnosis rate was improved from 76.92% to 88.76% [13]. Therefore, the AC mixing technique can serve as a versatile clinical tool.

The AC mixing IHC technique has several potential limitations. First, rapid IHC, including the multiplex setting, entails universal use of antibodies. Consequently, the diagnosticians must determine whether all scheduled antibodies are adaptable to each molecular target before their clinical use. However, the advantages of this procedure are its simplicity, high degree of accuracy, and preservation of surgical tissue for subsequent pathology, including molecular assessments. A second limitation is that the surgical strategy would need to change if there are discrepancies between HE and IHC staining, such as the negativity of TTF-1 or neuroendocrine markers observed in the present study. IHC is crucial for distinguishing among malignant neoplasms with similar morphologies, but it can also be misleading depending on the antibodies selected and staining quality. This is the central challenge when applying IHC to surgical pathology. To accurately assess their suitability for testing when using rapid IHC in daily practice, it will be essential for clinicians to fully understand the variation in target protein expression among different sample sites. An important third limitation is sampling bias due to tissue heterogeneity, which is the main pitfall of histological tissue comparison studies. IHC status may show intra- and intertumoral heterogeneity, and it is important to understand the variation in target protein expression among different sample sites so as to assess their suitability for testing. The fourth limitation is the additional cost of multiplex IHC techniques. The Histo-Tek R-IHC^TM device costs 3,980,000 Japanese yen. Additionally, rapid AC mixing IHC entails universal use of antibodies, and the multiplex IHC cocktails (6 mL) on FS cost a total of 216,000 Japanese yen/40 tests AC mixing IHC achieved stable and reagent-saving IHC, even when reagent concentrations were smaller concentration than standard [10]. As a fifth limitation, we employed commercial antibody cocktails in the present study and did not compare their diagnostic accuracy with other antibodies. Therefore, it remains unclear whether this combination of antibodies is optimal for distinguishing AD from SCC. For instance, the addition of highly specific SCC marker p40 or neuroendocrine markers may further enhance accuracy and broaden the diagnostic spectrum. To complete this new diagnostic technique and system, future research will be needed to provide the available antibodies/protocols.

In summary, we have shown that rapid multiplex AC mixing IHC (using TTF-1 + CK5, DSG 3 + Napsin A, and p63 + TRIM29) can effectively distinguish AD from SCC. Clinical pathology currently relies on singleplex IHC in lung cancer. Multiplex IHC will be important as personalized medicine increasingly entails prompt, simultaneous assessment of multiple molecular/immune profiles.

We thank Professor Shinya Tanaka of the Department of Cancer Pathology, Hokkaido University, who provided advice on the principles of AC electric field mixing.

Opt-out informed consent protocol was used for use of participant data for research purposes. The study protocol and this consent procedure and were reviewed and approved by the Institutional Review Board (IRB) at Akita University Hospital, Approval No. [896, 929, 2679].

The authors have no conflict of interest.

No specific funding was disclosed.

Shoji Kuriyama: data curation, formal analysis, investigation, methodology, and writing – original draft. Kazuhiro Imai: conceptualization, investigation, methodology, project administration, supervision, validation, and writing – review and editing. Hiroshi Nanjo: conceptualization, data curation, investigation, methodology, project administration, validation, and visualization. Yuki Wakamatsu: data curation and investigation. Shinogu Takashima, Tsubasa Matsuo, Hidenobu Iwai, Ryo Demura, Haruka Suzuki, Yuzu Harata, and Sumire Shibano : investigation, supervision, and validation. Akiyuki Wakita and Yusuke Sato: supervision and validation. Kyoko Nomura: formal analysis, supervision, and validation. Yoshihiro Minamiya: conceptualization, formal analysis, methodology, project administration, supervision, and validation.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants, but they are available from the corresponding author, K.I. (Department of Thoracic Surgery, Akita University Graduate School of Medicine; e-mail: [email protected]) upon reasonable request.