|Year : 2019 | Volume
| Issue : 1 | Page : 131-140
Insulinoma-associated protein 1 reveals an eminent performance for the diagnosis of pulmonary neuroendocrine tumors in cytological materials
Nesreen H Hafez MD 1, Manal Z Shalaby2, Mohamed M Ahmed3
1 Department of Pathology, National Cancer Institute, Cairo University, Giza, Egypt
2 Department of Medicine, National Research Center, Giza, Egypt
3 Department of Reproductive Health and Family Planning Research, National Research Center, Giza, Egypt
|Date of Submission||02-Apr-2019|
|Date of Acceptance||29-Apr-2019|
|Date of Web Publication||29-Nov-2019|
Nesreen H Hafez
Pathology Department, National Cancer Institute, Cairo University, Giza
Source of Support: None, Conflict of Interest: None
Background Although cytological interpretation of classic pulmonary neuroendocrine tumors (NETs) is straightforward, their identification remains a diagnostic challenge in daily practice, especially if cytomorphological features are unclear. The currently used immunostaining panel has some limitations. insulinoma-associated protein 1 (INSM1) is a transcription factor that is reported in different NETs. The aim of the current study was to identify INSM1 expression in cytological materials of lung NETs, to compare its performance with that of the currently used markers and to compare the expression in primary and corresponding metastatic NETs.
Materials and methods This retrospective study included 40 cytological samples of primary pulmonary NETs and nine metastatic extrapulmonary materials of the same patients as well as 20 samples of non-NETs (control group). Image-guided fine-needle aspiration was performed. Cell block sections were stained for INSM1, synaptophysin, CD56, and chromogranin A. For all markers, at least 1% positive cells were considered positive. Diffuse/strong positive was reported if at least 50% of the cells were detected at low power magnification. Focal/weak positivity was defined if less than 50% of tumor cells were positive but not easily identified at low power.
Results INSM1 was positive in 92.5% (37/40). The sensitivity for small-cell lung carcinoma, large-cell neuroendocrine carcinoma, and carcinoid was 92.6, 91.7, and 100%, respectively. No significant statistical difference was identified regarding expression in various NET subtypes (P>0.05). INSM1 was negative in all non-NETs. The INSM1 sensitivity was similar to CD56 sensitivity but more than that of synaptophysin and chromogranin. The INSM1 specificity was higher than those of CD56 and chromogranin A and equal to synaptophysin. The INSN1 sensitivity was similar to the sensitivity of the combined conventional markers as a group with a superior specificity. INSM1 staining in the nine metastatic NETs was matched up to their primary tumors.
Conclusion INSM1 combined high sensitivity, specificity, and accuracy when compared with the currently used markers. Nuclear INSM1 expression is easier and faster to interpret than the cytoplasmic or membranous expression of the other markers. It should be used in neuroendocrine marker panels. INSM1 staining was retained in all metastatic sites.
Keywords: cytology, immunocytochemistry, insulinoma-associated protein 1, pulmonary neuroendocrine tumors
|How to cite this article:|
Hafez NH, Shalaby MZ, Ahmed MM. Insulinoma-associated protein 1 reveals an eminent performance for the diagnosis of pulmonary neuroendocrine tumors in cytological materials. Egypt J Pathol 2019;39:131-40
|How to cite this URL:|
Hafez NH, Shalaby MZ, Ahmed MM. Insulinoma-associated protein 1 reveals an eminent performance for the diagnosis of pulmonary neuroendocrine tumors in cytological materials. Egypt J Pathol [serial online] 2019 [cited 2022 Jan 26];39:131-40. Available from: http://www.xep.eg.net/text.asp?2019/39/1/131/271994
| Introduction|| |
Neuroendocrine tumors (NETs) are epithelial tumors with a major neuroendocrine (NE) differentiation. NETs of the lung originate from enterochromaffin cells that are located in the bronchial mucosa as an element of disperse NE system all over the body. NETs account for 20–25% of all primary lung tumors (Mukhopadhyay et al., 2019). At the Egyptian National Cancer Institute, NETs of the respiratory system constituted 13.6% of all primary and metastatic tumors to the respiratory system, with the lung being the main primary site in 68% of cases, and small-cell lung carcinoma (SCLC) was the most prevalent histological subtype, representing 11.2% (Ibrahim, 2016).
Fine-needle aspiration cytology (FNAC) is a common established technique for the diagnosis of many tumors, including accessible pulmonary nodules. Cytological and small biopsy materials allow a precise diagnosis of most NETs when the cytomorphologic features and clinical findings are typical. However, large-cell neuroendocrine carcinoma (LCNEC) might be identified in cytological materials as non-SCLC, not otherwise specified, or adenocarcinoma (French, 2009). The differentiation of typical from atypical carcinoid is only achieved on resected specimens (Rodriguez et al., 2018). In addition, crushing artifacts, necrosis, and scarce representative materials present a difficulty in interpretation owing to the intrinsic characteristics of the submitted materials (Fujino et al., 2017).
Immunocytochemistry (ICC) can improve the diagnostic reproducibility of NETs when applied in association with morphologic criteria (Rodriguez et al., 2018). It could save time if ICC markers are performed on cytological material at the time of diagnosis. Moreover, ICC marker detection could be applied in patients with metastatic or locally advanced tumors for whom excision is not indicated. Some patients may be unfit for anesthesia and operative procedures; others may refuse any surgical intervention. In such cases, cytological materials may be the only way to reach a diagnosis (Doxtader and Mukhopadhyay, 2018).
The established NE markers, synaptophysin, chromogranin A, and CD56, have some limitations. Chromogranin A is highly specific for NETs, but it lacks sensitivity. Synaptophysin and CD56 are sensitive markers but not entirely specific. In addition, cytoplasmic expression of these markers might cause difficulty in the differentiation from nonspecific background in cases with weak and focal expression and in the occurrence of marked crushing artifacts or massive necrosis (Rosenbaum et al., 2015). Therefore, using a panel of ICC markers could be helpful. However, the panel of these three conventional markers has failed to detect NE differentiation in 10–25% of high-grade NE carcinomas (Rooper et al., 2017).
Therefore, there is a need for a new NE marker that is more accurate, simple to interpret, and can reduce the requirement for large expensive panels. Molecular profiling of NETs has recognized some proteins that have a vital role in NE differentiation and could be used as possible indicative markers. One of these is insulinoma-associated protein 1 (INSM1). INSM1 is a zinc-finger transcription factor involved in the normal development of NE cell all over the body. Recently, INSM1 has been reported to be identified in NETs of different body systems with relatively high sensitivity and specificity like in insulinoma, medullary thyroid carcinoma, medulloblastoma, and pituitary adenoma (Fujino et al., 2015; Rodriguez et al., 2018). Recent research studies on the excised specimens have proposed the use of INSM1 as the only NE marker for lung NETs (Taniwaki et al., 2006; Rodriguez et al., 2018); nevertheless, others have recommended that INSM1 can present a helpful adjunct to established markers (Rooper et al., 2017; Doxtader and Mukhopadhyay, 2018). In addition, a previous study reported a positive relation of INSM1 expression to chemosensitivity in SCLC cell lines (McColl et al., 2017).
To our knowledge, the expression of INSM1 in pulmonary cytological materials has been very limited in literature (Doxtader and Mukhopadhyay, 2018; Rodriguez et al., 2019).
The aim of this study was to identify the usefulness of INSM1 as a diagnostic ICC marker for NE differentiation in primary lung tumors using cytological materials, to compare its sensitivity and specificity with those of routinely used NE markers, and finally to recognize INSM1 expression in metastatic NETs in comparison with its expression in the corresponding primary lung tumors.
| Materials and methods|| |
This retrospective study on 40 cases with pulmonary NETs was conducted in Cytology Unit, Pathology Department, National Cancer Institute, Cairo University. These cases presented with primary pulmonary nodules during the period from January 2013 to August 2018. Of these cases, nine had metastatic extrapulmonary tumors from the pulmonary NETs, for which cytological materials from the metastatic sites were available as well. They included six pleural effusion and three lymph node metastases. A total of 20 samples of non-NETs were used as a control group: 11 adenocarcinomas and nine squamous cell carcinomas.
Patients’ cytological files were revised, and all cases of pulmonary NETs diagnosed by FNAC during the previous period were retrieved. Results of ICC staining of synaptophysin, CD56, and chromogranin A were detected from cytology reports, if performed. Patients’ medical records were reviewed for the corresponding final histopathological diagnosis and data related to age and sex of the patients, tumor site and size, smoking status, and metastases and their sites, if present at the time of diagnosis.
The cases were initially referred to the Radiology Department where image-guided FNAC was performed using a 23-G needle. According to accessibility and feasibility, patients underwent ultrasound-guided FNAC or underwent endobronchial US-guided FNAC. Two to three passes were done for every case to prepare slide smears and cell blocks. A cytopathologist was available in the radiology department for rapid on-site evaluation to guide the number of FNA passes required to obtain an adequate material that helps in reaching a definitive diagnosis.
The aspirated materials were immediately spread on the glass slides labeled in patient’s name. One slide of each case was rapidly stained for rapid on-site evaluation and the remaining slides were fixed in 95% ethyl alcohol without staining. The needle washing was collected for cell block preparation.
The fixed unstained slides and cell block materials were referred to the cytology unit of the pathology department. The slides were left in alcohol for at least 30 min at room temperature and subsequently stained using modified Papanicolaou stains. Cell block was prepared by fixing the material in 10% neutral-buffered formalin and then it was embedded in paraffin. Thin section was cut and stained by hematoxylin-eosin stain. Initial ICC studies were performed on cell block sections.
The slides of all cases were examined microscopically and diagnosed as pulmonary NETs according to the standard cytomorphologic features (French, 2009) alone or with the aid of ICC NE markers (chromogranin A, synaptophysin, and CD56). The cases were positive for at least one marker to be reported as NETs. All cases were later on resected or biopsied and had final histopathological reports confirming the NE differentiation.
Inclusion criteria for enrolment were patients with cytological reports as well as confirming final corresponding histopathological reports of primary pulmonary NETs, availability of adequate cell block materials for ICC, and no prior local or systemic therapy before FNAC. Patients were excluded from the study when cases had no cell blocks or inadequate tissue in cell blocks for ICC assessment. Cases showing morphologic features of non-NETs or showed negative reactions to all previously mentioned NE markers were used as control.
All archival slides of selected cases, including smears, cell block sections, and ICC slides were retrieved from the archives of the cytology unit of the pathology department and reviewed to confirm the diagnosis and to score the immunostaining expression of markers. Cell block sections were evaluated to determine the adequacy for applying ICC. From each block, 4-µm-thick sections were prepared on positively charged glass slides.
Immunostaining procedure was applied using the labeled streptavidin-biotin-peroxidase complex method according to the manufacturer’ specifications. The immunostaining technique used the Benchmark XT auto-immunostaining system (Ventana Medical System, Tucson, Arizona, USA). All sections were deparaffinized, rehydrated, and treated with 0.3% H2O2 for 5 min at room temperature to block endogenous peroxidase activity. Heat-based antigen retrieval was performed to obtain optimal results following standard protocol. A mouse monoclonal antibody against INSM1 (clone A-8: sc-271408, in dilution 1 : 250; Santa Cruz Biotechnology Inc.), chromogranin A (clone H-300, in dilution 1 : 100; Santa Cruz, California, USA), synaptophysin (clone 27G12, 1 : 400 dilution; Novacastra/Leica Biosystems, Buffalo Grove, Illinois, USA), and CD56 (clone MS-204-R7; Lab Vision Corporation, Fremont, California, USA) was applied. Appropriate positive and negative controls were included in each run. Normal pancreatic tissue was used as positive controls for all markers. Negative controls were obtained by replacing the primary antibody with PBS.
Interpretation of immunohistochemical findings
INSM1-positive result appeared as nuclear staining, whereas synaptophysin and chromogranin gave granular cytoplasmic staining. CD56 staining was cytoplasmic and membranous. For all markers, IHC was classified as positive if there was at least 1% positive tumor cells and negative if no staining was detected or less than 1% of the tumor cells were positive. The staining was considered diffuse/strong positive if at least 50% of the cells were positive and staining was simply detected at the low power magnification (×10). Focal/weak positivity was defined if less than 50% of tumor cells were positive but not easily identified at the low power (Rodriguez et al., 2019).
Statistical analysis was performed using statistical package for the social sciences (IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp). Statistical significance was determined by Fisher’s exact and χ2 analysis. The level of significance was set at 0.05 or less. In addition, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the marker were calculated.
| Results|| |
Clinicopathological data of the selected cases
A total of 40 cytological samples of primary pulmonary NETs and nine materials with metastatic extrapulmonary NETs of the same patients as well as 20 samples of non-NETs (control cases) were included in this study. The mean age of the included NET cases was 51.2±10.2 years (range: 34–72). There was a male predominance (29 males), with male : female ratio of 2.6 : 1. A total of 31 (77.5%) cases were located in the right lung and nine (22.5%) were in the left lung. The mean tumor size was 7.1±2.4 cm (range: 3–12 cm). History of smoking was detected in 27 (67.5%) cases. The studied cases included 27 SCLC (67.5%), 12 (30%) LCNEC, and one (2.5%) carcinoid. Metastasis was detected in nine (22.5%) cases; six had pleural effusion metastasis and three cases had lymph node metastasis. These cases had cytological materials with cell blocks from the metastatic sites to compare INSM1 expression in primary and corresponding metastatic tumors.
Overall, 20 samples of non-NETs were used as a control group: 11 adenocarcinomas and nine squamous cell carcinomas. The mean age was 55±8.1 years. Male : female ratio was 1.2 : 1. Seven (70%) cases were detected in the right lung. The mean tumor size was 8.3±3.9. Smoking was detected in six (60%) cases.
Insulinoma-associated protein 1 immunocytochemical results
The INSM1 immunostaining results on NET cytological materials are summarized in [Table 1]. INSM1 was positive in 37 of 40 primary NETs of the lung (92.5%). Overall, 25 of 27 studied SCLC cases ([Figure 1] and [Figure 2]), 11 of 12 studied LCNEC cases, and the only studied carcinoid case were positive for INSM1. The sensitivity of INSM1 for SCLC, LCNEC, and carcinoid was 92.6, 91.7, and 100%, respectively. No significant statistical difference was identified between INSM1 expression in various NET subtypes (P>0.05).
|Table 1 Insulinoma-associated protein 1 immunostaining interpretation among positive neuroendocrine tumor cases|
Click here to view
|Figure 1 Cytological smear of small-cell lung carcinoma case demonstrating cluster of small-cell with scant cytoplasm. The chromatin is finely granular (×400, Papanicolaou stain).|
Click here to view
|Figure 2 Insulinoma-associated protein 1 immunocytochemical staining on cell block section of the same small-cell lung carcinoma case showing strong/diffuse nuclear positivity (×400, immunocytochemistry).|
Click here to view
INSM1 had diffuse/strong expression in most positive cases (23/37, 62.2%) and focal/weak in 14 (37.8%) cases. No INSM1 staining in the inflammatory cells or in the background was detected. The three negative cases for INSM1 were SCLC (two cases) and LCNEC (one case).
All 20 non-NETs of the lung were INSM1 negative. Four control samples expressed background staining and faint focal nonspecific cytoplasmic expression. This reactivity was distinctive from the strong nuclear expression identified in NETs. A statistically significant difference was demonstrated between INSM1 staining in NETs and non-NETs cases (P<0.001).
The INSM1 sensitivity, specificity, positive predictive value, and negative predictive value were 92.5% [95% confidence interval (CI): 79.4–98.1%], 100% (95% CI: 81–100%), 100% (95% CI: 88.8–100%), and 87% (95% CI: 67–96.3%), respectively. The accuracy was 93.6% (95% CI: 85.8–98.8%).
Immunocytochemical staining of synaptophysin, CD56, and chromogranin A
In 24 (60%) cases, ICC was performed at the time of the initial cytological interpretation to confirm the NE differentiation: 14 SCLC cases, nine LCNEC cases, and one carcinoid case. The archival immunostained slides were retrieved and evaluated to determine the marker percentages and intensities. In the remaining 16 cases where the initial cytological diagnosis was based on cytomorphological features alone, ICC staining for synaptophysin, CD56, and chromogranin was carried out. Different conventional marker immunostaining results are summarized in [Table 2] and [Table 3].
|Table 2 The results of the neuroendocrine markers among the studied pulmonary neuroendocrine tumors on cell block materials|
Click here to view
|Table 3 Conventional immunocytochemistry marker interpretation among positive cases|
Click here to view
Overall, synaptophysin was positive in 33 (82.5%) cases. They included positivity in 25 SCLC ([Figure 3]) and eight LCNEC cases ([Table 2]). The staining was diffuse/strong in 20 (60.6%) cases and focal/weak in 13 (39.4%) cases ([Table 3]). CD56 expression was detected in 37 (92.5%): 26 cases were SCLC ([Figure 4]), 10 cases were LCNEC, and one case was carcinoid ([Table 2]). Diffuse strong staining was reported in 26 (70.3%) cases and focal/weak in 11 (29.7%) cases ([Table 3]). Chromogranin A was positive in 29 (72.5%) cases, where diffuse/strong staining was reported in 21 (72.4%) cases and focal/weak staining in eight (27.6%) cases.
|Figure 3 Synaptophsin immunocytochemical staining of the same previous case of small-cell lung carcinoma demonstrating strong/diffuse cytoplasmic staining (×400, immunocytochemistry).|
Click here to view
|Figure 4 CD56 immunocytochemical staining on cell block section of the same previous small-cell lung carcinoma case demonstrating strong/diffuse cytoplasmic and membranous expression (×400, immunocytochemistry).|
Click here to view
These established markers were implemented on 20 non-NETs of the lung. CD56 was positive in three adenocarcinoma cases. Chromogranin A was positive in two cases (one adenocarcinoma and one squamous cell carcinoma). All of these cases revealed focal weak staining. The only chromogranin A-positive adenocarcinoma case was also positive for CD56. Subsequent histopathological examination of this case was found to be adenocarcinoma with no NE differentiation. Synaptophysin is 100% specific but less sensitive (82.5%). CD56 is more sensitive (92.5%) but less specific (85%). Chromogranin was more specific (90%) and less sensitive (72.5%) ([Table 4]).
|Table 4 Statistical values of insulinoma-associated protein 1 compared with other conventional neuroendocrine markers|
Click here to view
Insulinoma-associated protein 1 immunostaining compared with synaptophysin, CD56, and chromogranin A
INSM1 immunostaining was then compared with the staining of the three established NE markers, each one individually or as a group.
The sensitivity of INSM1 for lung NETs was similar to the sensitivity of CD56 but more than the sensitivity of synaptophysin and chromogranin A. The specificity of INSM1 was higher than those of CD56 and chromogranin A and equal to synaptophysin. CD56 had more diffuse/strong staining (70.3%) than INSM1 (62.2%). INSM1 revealed the highest sensitivity (92.5%), specificity (100%), and accuracy (95%) when compared with the other markers. The lowest sensitivity and accuracy were reported for chromogranin A, whereas the lowest specificity was detected for CD56 ([Table 4]).
For SCLCs, the sensitivity of INSM1 was 92.5% which was less than the sensitivity of CD56 but more than that of chromogranin A and equivalent to synaptophysin. For LCNEC, the sensitivity of INSM1 was higher than that of other markers ([Table 3]).
To detect the performance of INSM1 as an NE marker, we compared INSM1 positivity to the combined expression of the previously established markers as a group. For this assessment, any case was counted in the positive group if synaptophysin, CD56, or chromogranin A was expressed.
Although the sensitivity of INSN1 (37/40, 92.5%) was similar to the sensitivity of the combined conventional markers as a group (37/40, 92.5%), the specificity of INSM1 (100%) was superior to that of the combined markers (5/15, 75%) ([Table 2]).
[Table 5] reveals different staining alternatives of the studied ICC markers in the pulmonary NETs. Expression of the four markers together was detected in 23 (57.5%) cases, including 18 SCLC and five LCNEC. Ten (25%) cases expressed three markers. Expression of INSM1, synaptophysin, and CD56 was reported in four cases. Immunostaining for INSM1, CD56, and chromogranin was also reported in six cases. The remaining positive seven cases were positive for two markers ([Figure 5] and [Figure 6]). The three INSM1-negative cases were positive for CD56 and synaptophysin but were also negative for chromogranin A. No studied case revealed expression of only one marker, and no case revealed negative expression for the four studied markers.
|Table 5 Different staining alternatives of the studied immunocytochemistry markers in the different neuroendocrine tumor subtypes|
Click here to view
|Figure 5 Hypercellular smear of carcinoid case showing loosely cohesive aggregate of round cells with moderate to scant amount of cytoplasm and finely granular chromatin pattern. Minimal pleomorphism is detected. Attempts to form rosettes-like arrangement are noticed (×400, Papanicolaou stain).|
Click here to view
|Figure 6 Insulinoma-associated protein 1 immunocytochemical staining on cell block section of the same carcinoid case showing strong/diffuse nuclear positivity (×400, immunocytochemistry).|
Click here to view
Insulinoma-associated protein 1 in primary tumors and corresponding metastases
INSM1 staining in nine metastatic NETs was matched up to their primary tumors. In all cases, INSM1 was positive in both primary and metastasis tumors. Diffuse/strong staining was detected in six of nine (66.7%) primary tumors, whereas focal/weak staining was detected in only four (44.4%) metastatic cases. However, the staining intensity differences between primary and metastatic tumors were not statistically significant (P=0.656).
| Discussion|| |
NETs are a diverse group of tumors with particular morphological criteria, clinical findings and biological behavior. NETs can originate in several body organs. Nevertheless, they are more frequent in the gastrointestinal tract and the lung. Owing to the presence of several therapeutic strategies for pulmonary tumors, there is a pressing need for the accurate tumor interpretation and classification (Chai et al., 2018).
Numerous classification systems have been applied for reporting NETs (Oronsky et al., 2017; Rindi et al., 2018) with a target of dividing them into prognostic groups that help in their management. As a general rule, NETs are categorized into well-differentiated neoplasms, which include low-grade typical carcinoid (2% of lung tumors) or intermediate-grade atypical carcinoid (<1%) and poorly differentiated tumors that contain high-grade SCLC (20%) and LCNEC (3%) (Rodriguez et al., 2018).
On cytological materials, well-differentiated tumors are described by the existence of a rather uniform round or oval-sized, small-sized to medium-sized cells arranged singly and in loosely cohesive clusters. The tumor cells have smooth nuclear contour, salt and pepper chromatin and eccentrically located nuclei with no necrosis, and less than one mitosis per 2 mm2. In such cases, a diagnosis of ‘well-differentiated NETs/carcinoid’ should be used. A diagnosis of ‘atypical carcinoid’ is probable just if focal necrosis or mitosis from 2 to 9 is recorded (Righi et al., 2017). On the contrary, poorly differentiated carcinomas are identified by cells with notable atypia and the presence of necrosis, mitosis, and nuclear molding. The classic NE chromatin is detected as well. In such cases, a diagnosis of ‘high-grade NE carcinoma’ should be used with the detection of large or small cells whenever possible (Saeed et al., 2018). Ki-67 is detected in more than 50%, and it is not mandatory for the diagnosis, but it could be useful for the treatment decision (Righi et al., 2017).
Although the cytological interpretation of classic pulmonary NETs may be straightforward using standard cytological stains, their identification remains a diagnostic challenge in daily practice especially in poorly differentiated tumors (Rooper et al., 2017). The cytopathologist needs to enhance the confidence of interpretation using ICC markers (Mukhopadhyay et al., 2019).
The specificity of the traditionally used NE markers is limited. A marker with superior sensitivity and specificity is required. No novel NE markers have been established in the diagnostic field since 1990s, when CD56 was recognized (Mukhopadhyay et al., 2019). Recently in the pulmonary pathology scope, INSM1 has been described as a sensitive and specific NE marker that outperforms the currently established markers with the potential to replace the standard panel (Fujino et al., 2017; Rooper et al., 2017). These studies encourage us to study the diagnostic performance of INSM1 in cytological materials of lung NETs.
The main objective of this study was to clarify the significance of INSM1 immunostaining in pulmonary NET diagnosis compared with non-NETs on cell block materials.
In this study, the overall sensitivity of INSM1 was 92.5%, which was similar to the sensitivity that was demonstrated in the previous comparable study on pulmonary cytological materials (92%) (Doxtader and Mukhopadhyay, 2018). In another previous work on cytology, only one case was INSM1 negative (Rodriguez et al., 2019). In this case, all other studied NE markers were also negative. They explained these negativities to the massive necrosis in the slides that interfered with the expression of all markers. However, our result was less than that reported in the former studies on excised histopathological specimens. Rooper et al. (2017) reported 96.4% overall sensitivity for INSM1 in lung NETs. They recommended that INSM1 might be utilized as a single marker for lung NETs. A sensitivity of 100% was reported by Fujino et al. (2017) in surgical pathology materials. Kriegsmann et al. (2018), demonstrated a lower sensitivity of 76%.
Regarding the NET subtypes, the sensitivity of INSM1 for SCLC was 92.6% in this study, which is nearly similar to the sensitivity of 93% reported by Doxtader and Mukhopadhyay (2018). In literature, the sensitivity ranged from 93 to 100% (Fujino et al., 2017; Rooper et al., 2017; Mukhopadhyay et al., 2019; Rodriguez et al., 2019). Fujino et al. (2015) found that INSM1 was identified in all SCLC excised specimens as well as in 57.1% of SCLC cell lines. They noted no expression in non-SCLC in excised tissues or cell lines. A lower sensitivity of INSM1 in SCLC (81%) was detected by others (Švajdler et al., 2018).
In this study, INSM1 sensitivity for LCNEC was 91.7%, which was similar to that reported by Rooper et al. (2017). However, a sensitivity of 75% was reported in another study (Mukhopadhyay et al., 2019).
These discrepancies in sensitivity among different studies can be explained by the differences in numbers of study population, patients’ heterogeneity, types of materials used (cytology, whole-tissue sections, or microarray specimens), staining techniques (manual or automated), using different antibody clones, differences in technical methods (dilution or antigen retrieval), and the use of different scoring systems (Fujino et al., 2017; Mukhopadhyay et al., 2019).
In this study, the studied carcinoid case was diffusely positive to INSM1. Rooper et al. (2017) detected diffuse positivity in all well-differentiated NETs, including carcinoid. Mukhopadhyay et al. (2019) reported a weak focal positivity in 98% of carcinoid cases. The sensitivity reported by Doxtader and Mukhopadhyay (2018), was 90% in carcinoid. In this study, no significant statistical difference was identified regarding INSM1 expression in various NET subtypes (P>0.05).
This study demonstrated 100% specificity for INSM1 as all studied pulmonary non-NETs were INSM1 negative. A statistical significant difference was demonstrated between INSM1 staining in NETs and non-NETs cases (P<0.001). These results agreed with what was reported in a previous research (Doxtader and Mukhopadhyay, 2018). Others noted INSM1 expression in a small number of non-NETs (Rooper et al., 2017; Kriegsmann et al., 2018; Rodriguez et al., 2019). Some reported that this expression did not cause any diagnostic problem as the staining is focal with absence of NE features, and they explained this positivity by nonspecific antibody cross-reactivity or by focal NE differentiation in such tumors (Rooper et al., 2017).
In this study, the expression of INSM1 was diffuse/strong in most studied NETs (62.2%). This result was in line with most previously published data (Fujino et al., 2017; Rooper et al., 2017; Rodriguez et al., 2019). Others reported focal and weak staining of INSM1 in all NETs for unexplained reasons (Mukhopadhyay et al., 2019).
The second objective of this study was to compare INSM1 expression in pulmonary NETs with that of the currently used NE markers, each one individually and in a group.
In our series, INSM1 and CD56 demonstrated the highest sensitivity (92.5%). INSM1 also demonstrated the highest specificity and accuracy for the detection of NETs, whereas CD56 was the least specific. Synaptophysin had the same specificity of INSM1 but had a lower sensitivity. Chromogranin A was the least sensitive and the least accurate marker. Our results almost agreed with what reported by others (Rodriguez et al., 2018). Doxtader and Mukhopadhyay (2018), in their study detected that the overall INSM1 sensitivity was less than the sensitivities of CD56 and synaptophysin but more than the sensitivity of chromogranin A. Fujino et al. (2015) detected that INSM1 sensitivity was higher than chromogranin A and synaptophysin.
According to NET subtypes, it was reported that the poorly differentiated NETs (SCLC and LSNEC) were categorized as difficult cases to be diagnosed (Chai et al., 2018). We demonstrated that CD56 had the highest sensitivity for SCLC followed by INSM1, whereas INSM1 revealed the highest sensitivity for LCNEC followed by CD56. Doxtader and Mukhopadhyay (2018) reported nearly similar results. Our results were in accordance with others; Rooper et al. (2017) and Fujino et al. (2017), detected INSM1 expression in all well-differentiated neoplasms as well as in 90–100% of poorly differentiated tumors that might be negative for other conventional markers.
Among the studied 40 primary pulmonary NETs, for which cytological materials were available, three cases were not properly classified by INSM1. However, these cases were classified correctly by CD56 and synaptophysin. Combination of INSM1 with CD56 correctly achieved the highest overall sensitivity and tumor subtype sensitivity, specificity, and staining intensity. Nearly similar results were found by others (Švajdler et al., 2018).
Preliminary data reported that the nuclear expression of INSM1 staining was easier to be interpreted than the established markers and is less prone to nonspecific reactivity (Rosenbaum et al., 2015; Rooper et al., 2017). Our data supported this observation.
When INSM1 was compared with the currently used conventional markers as a group, it was noted that the sensitivity of INSN1 (92.5%) was similar to the sensitivity of the combined conventional markers as a group (92.5%). However, the specificity of INSM1 (100%) was significantly superior to that of the combined markers (75%). These results could support that the substitution of currently used markers by INSM1 is feasible or at least INSM1 should be used as a useful adjunct to the existing markers. Doxtader and Mukhopadhyay (2018), detected a lower sensitivity of INSM1 compared with the group of conventional markers (92 vs. 100%) with a higher specificity (100 vs. 90%). Rooper et al. (2017) recommended the use of INSM1 as a standalone first-line marker for identification of all grades of NETs and reported that this marker would save tissues for critical genetic and immunologic methods and could save money. In the same line, other authors recommended the use of INSM1 as a single marker in cytology material when the morphologic criteria are indicative of NETs, and if INSM1 is negative but the differential diagnosis includes NETS, a panel should be used (Rodriguez et al., 2019). In contrast, Mukhopadhyay et al. (2019), did not suggest total replacement of conventional markers with INSM1, as the expression of conventional markers was stronger and more diffuse, whereas INSM1 expression, in their work, was patchy and weak. Kriegsmann et al. (2018) found that all already established NE markers as well as their grouping had an elevated sensitivity and an inferior specificity compared with INSM1. Hence, they documented that replacement of currently used markers by INSM1 was not justified. Švajdler et al. (2018) reported that p16 alone and a combined panel including CD56, TTF1, and p16 showed significantly higher sensitivity than INSM1. In their study, INSM1 had a higher sensitivity than synaptophysin and chromogranin A. On the contrary, they found a higher specificity for INSM1 than p16 or the combined panel.
Mukhopadhyay et al. (2019), noted the expression of INSM1 in benign NE cells and carcinoid tumorlets; therefore, it should not be used as a marker of neoplasia or malignancy. Moreover, it is not a specific marker of pulmonary NETs. In contrast, Rosenbaum et al. (2015) and Fujino et al. (2015) found significant elevated INSM1 expression in neoplastic versus non-neoplastic specimens.
Our last objective was to recognize if INSM1 expression in metastatic sites is comparable to its expression in the corresponding primary tissues or is lost during tumor progression and metastasis. Our results showed the highest concordance of INSM1 positivity between paired cytological primary NETs and corresponding cytological metastatic materials. INSM1 staining was retained in all metastatic sites. Our finding was in accordance with a previous work (Mukhopadhyay et al., 2019), where the authors concluded that INSM1 expression in metastatic sites could be a consistent substitute for the INSM1 status of the primary tumor. Švajdler et al. (2018), reported that 87.5% had INSM1 expression in both primary tumors and metastatic deposits.
| Conclusion|| |
INSM1 revealed an eminent performance in the diagnosis of pulmonary NETs in cytological materials. INSM1 combined high sensitivity, specificity, and accuracy when compared with the currently used conventional markers. Nuclear expression of INSM1 is easier and faster to interpret than the cytoplasmic or membranous expression of the currently used markers. Thus, INSM1 is considered a promising NE marker in cytological samples and should be used as a main complementary marker to the established ICC panel for NE differentiation. Furthermore, we confirmed that INSM1 staining was retained in all metastatic tumors when compared with the corresponding primary tumors.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Chai SM, Brown IS, Kumarasinghe MP (2018). Gastroenteropancreatic neuroendocrine neoplasm: selected pathology review and molecular updates. Histopathology 72:153–167.
Doxtader EE, Mukhopadhyay S (2018). Insulinoma-associated protein 1 is a sensitive and specific marker of neuroendocrine lung neoplasm in cytology specimens. Cancer Cytopathol 126:243–252.
French CA (2009). Respiratory tract, pulmonary neuroendocrine neoplasms. In: Cibas ES, Ducatman BS, editors. Cytology: diagnostic principles and clinical correlates. 3rd edition. Philadelphia, PA: Elsevier. pp. 65–104.
Fujino K, Motooka Y, Hassan WA, Ali Abdalla MO, Sato Y, Kudoh S et al.
(2015). Insulinoma-associated protein 1 is a crucial regulator of neuroendocrine differentiation in lung cancer. Am J Pathol 185:3164–3177.
Fujino K, Yasufuku K, Kudoh S, Motooka Y, Sato Y, Wakimoto J et al.
(2017). INSM1 is the best marker for the diagnosis of neuroendocrine tumors: comparison with CGA, SYP and CD56. Int J Clin Exp Pathol 10:5393–5405.
Ibrahim M (2016). Respiratory system and pleural tumors. In: Mokhtar N, Salama A, Badawy O, Khorshed E, Mohamed G, Ibrahim M, Abdelazim H, editors. Cancer pathology registry, A12-year registry 2000–2011. Cairo, Egypt: Cairo Press. pp. 115–135.
Kriegsmann K, Zgorzelski C, Kazdal D, Cremer M, Muley T, Winter H et al.
(2018). Insulinoma-associated protein 1 (INSM1) in thoracic tumors is less sensitive but more specific compared with synaptophysin, chromogranin A and CD56. Appl Immunohistochem Mol Morphol 2018:24.
McColl K, Wildey G, Sakre N, Lipka MB, Behtaj M, Kresak A et al.
(2017). Reciprocal expression of INSM1 and YAP1 defines subgroups in small cell lung cancer. Oncotarget 8:73745–73756.
Mukhopadhyay S, Dermawan JK, Lanigan CP, Farver CF (2019). Insulinoma-associated protein 1 (INSM1) is a sensitive and highly specific marker of neuroendocrine differentiation in primary lung neoplasms: an immunohistochemical study of 345cases, including 292 whole-tissue sections. Mod Pathol 32:100–109.
Oronsky B, Ma PC, Morgensztern D, Carter CA (2017). Nothing but NET: A review of neuroendocrine tumors and carcinomas. Neoplasia 19:991–1002.
Righi L, Gatti G, Volante M, Papotti M (2017) Lung neuroendocrine tumors: pathological characteristics. J Thorac Dis 9:S1442–S1447.
Rindi G, Klimstra DS, Abedi-Ardekani B, Asa SL, Bosman FT, Brambilla E et al.
(2018). A common classification framework for neuroendocrine neoplasms: an International Agency for Research on Cancer (IARC) and World Health Organization (WHO) expert consensus proposal. Mod Pathol 31:1770–1786.
Rodriguez EF, Chowsilpa S, Maleki Z (2018). Insulinoma associated protein 1 immunostain: a diagnostic tool for pulmonary small cell carcinoma in cytology. Acta Cytol 62:333–338.
Rodriguez EF, Judd Fite J, Chowsilpa S, Maleki Z (2019). Insulinoma-associated protein 1 immunostaining on cytology specimens: an institutional experience. Hum Pathol 85:128–135.
Rooper LM, Sharma R, Li QK, Illei PB, Westra WH (2017). INSM1 demonstrate superior performance to the individual and combined use of synaptophysin, chromogranin and CD56 for diagnosing neuroendocrine tumors of the thoracic cavity. Am J Surg Pathol 41:1561–1569.
Rosenbaum JN, Guo Z, Baus RM, Werner H, Rehrauer WM, Lloyd RV (2015). INSM1: a novel immunohistochemical and molecular marker for neuroendocrine and neuroepithelial neoplasms. Am J Clin Pathol 144:579–591.
Saeed OAM, Cramer H, Wang X, Wu HH (2018). Fine needle aspiration cytology of hepatic metastases of neuroendocrine tumors: a 20-year retrospective, single institutional study. Diagn Cytopathol 46:35–39.
Švajdler M, Mezencev R, Šašková B, Ondič O, Mukenšnábl P, Michal M (2018). Triple marker composed of P16, CD56 and TTF1 shows higher sensitivity than INSM1 for diagnosis of pulmonary small cell carcinoma: proposal for a rational immunohistochemical algorism for diagnosis of small cell carcinoma in small biopsy and cytology specimens. Hum Pathol 85:58–64.
Taniwaki M, Daigo Y, Ishikawa N, Takano A, Tsunoda T, Yasui W et al.
(2006). Gene expression profiles of small-cell lung cancers: molecular signatures of lung cancer. Int J Oncol 29:567–575.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]