Sinonasal inverted papilloma (IP) has a propensity for malignant transformation. Although the IP-associated squamous cell carcinoma (SCC) is rare, it has a poor prognosis. To the best of our knowledge, this is the first study to assess IMP3 immunohistochemical (IHC) expression in sinonasal tumors and to compare it to the Ki-67 IHC expression and to other established clinicopathological parameters. A retrospective study was conducted on three groups which consisted of 72 cases of sinonasal IP, 20 age-matched samples of normal respiratory epithelium, and 15 cases of sinonasal SCC associated with IP, which were obtained from the archives of the Pathology Lab of Ain Shams University Specialized and Ain Shams University Hospitals during the period from January 2012 to December 2019. An IHC study was performed to evaluate IMP3 and Ki-67 expression in the three groups, with correlation of IMP3 expression to established clinicopathological parameters of sinonasal SCC on top of IP. Both IMP3 and Ki-67 showed a sharp rise in expression in the sinonasal SCC group. In addition, there were statistically significant differences in expression values between the 3 groups (
Sinonasal inverted papilloma (IP) is one of the most common nonmalignant lesions in the nasal cavity and paranasal sinuses. Although benign, it has a high recurrence rate (up to 78% of cases), destroys surrounding structures (including bone erosion and orbital wall involvement), and may undergo malignant transformation (in 5–15% of cases). In fact, sinonasal IPs are concomitantly diagnosed in 1.7–56% of patients with sinonasal squamous cell carcinoma (SCC). Moreover, malignant transformation in such lesions has worse long-term clinical outcomes that include metastasis, postoperative disfigurement, and/or death [
The oncofetal RNA-binding protein insulin-like growth factor 2 mRNA binding protein 3 (IMP3) is expressed during the early stages of embryogenesis and plays a role in cell growth and migration. These functions are mediated by posttranscriptional regulation of cell proliferation via regulation of IGF-2 transcription. After birth, IMP3 becomes epigenetically silenced, such that normal adult tissues show no detectable expression of this protein. However, its reexpression was noted in a series of human malignancies but not in the adjacent normal epithelium [
Only a few studies have assessed the immunohistochemical (IHC) expression of IMP3 in head and neck SCC [
The current study was a retrospective study conducted on three groups: 72 cases of sinonasal IP, 20 age-matched samples of normal respiratory epithelium obtained from cases with inflammatory/allergic nasal polypi and cases submitted to exclude fungal infection that proved to be free of fungal infection by periodic acid Schiff, and 15 cases of sinonasal SCC associated with IP. They were obtained from the archives of the Pathology Lab of Ain Shams University Specialized and Ain Shams University Hospitals. These were received and diagnosed during the period from January 2012 to December 2019. All cases underwent surgical intervention such as transnasal endoscopic resection and open surgical resection. The archival files were reviewed to determine the age and sex of patients. Hematoxylin and Eosin (H&E) stained slides were examined to reevaluate and verify the histopathologic diagnosis and histologic grade according to the World Health Organization (WHO) tumor differentiation [
ENT reports of the patients of the third group of IP-associated SCC were reviewed to determine (a) survival time, which was calculated based on the date of major surgery and the date of last follow-up or death, and (b) progression-free survival time, which was calculated based on the date of major surgery or last session of adjuvant postoperative radiotherapy and the date of relapse (local recurrence/distant metastasis) at the last follow-up.
Signed written and informed consent was obtained from all participants prior to surgery. The study was approved by the Research Ethical Committee at the Faculty of Medicine, Ain Shams University.
Four-micrometer sections of formalin-fixed and paraffin-embedded samples of normal respiratory epithelium, sinonasal IP, and sinonasal SCC associated with IP were prepared. IHC staining was performed on samples of the three groups using rabbit monoclonal anti-IMP3 (Clone: EP286; Cell Marque, Sigma-Aldrich Co., California, USA; 1 : 100 dilution) and rabbit monoclonal anti-Ki-67 (Clone: SP6; Cell Marque, Sigma-Aldrich Co., California, USA; 1 : 200 dilution). Avidin-Biotin immunoperoxidase complex technique was used by applying the sensitive detection kit (BioGenex, Fermont, California, USA) [
IHC analysis of IMP3 and Ki-67 were blindly performed by three pathologists (the authors). Cytoplasmic staining of IMP3 and nuclear staining of Ki-67 in cells in any of the lesions of the three groups were regarded as positive staining. Slides were scanned by three pathologists at 20x magnification to identify representative hot spots for counting positive cells. The counting of cells with cytoplasmic positivity for IMP3 and cells with nuclear positivity for Ki-67 was performed at 400x magnification. Next, the proportion of positive cells over the total number of counted cells was independently estimated by each of the three pathologists, and the average was calculated. A multiheaded microscope was used to resolve any discrepancies.
The assessment of the intensity of cytoplasmic staining of IMP3 and the nuclear staining of Ki-67 was carried out by subjective consideration of staining intensity, which was scored as follows: 0: negative, 1: weak positivity, 2: moderate positivity, and 3: strong positivity. Any discrepancies were resolved by consensus using a multiheaded microscope [
All readings were performed independently and without any prior knowledge of the clinical or histopathological characteristics of the cases.
Data was tested for normality with a Shapiro-Wilk test and expressed as
Data describing age and gender of the three studied groups are demonstrated in Table
Description of age and gender among the studied groups.
Group | ||||||
---|---|---|---|---|---|---|
Normal epithelium | Inverted papilloma | SCC | ||||
Mean | ±SD | Mean | ±SD | Mean | ±SD | |
Age | 47.95 | 6.03 | 48.43 | 5.42 | 57.80 | 9.30 |
Gender | ||||||
Male | 9 | 45.0% | 49 | 68.1% | 13 | 86.7% |
Female | 11 | 55.0% | 23 | 31.9% | 2 | 13.3% |
SD: standard deviation; SCC: squamous cell carcinoma.
Among the 15 cases in the sinonasal SCC group, 2 cases (13.3%) were T1, 4 cases were T2 (26.7%), 4 were T3 (26.7%), and 5 were T4 (33.3%). Eight cases were N0/N1 (53.3%), while the nodal stage of the rest of the cases was N2/N3. Three cases (20%) were grade 1, 10 cases (66.7%) were grade 2, and 2 cases (13.3%) were grade 3.
The overall IMP3 IHC expression revealed a pattern of cytoplasmic expression of low intensity in a few cells of the normal epithelium and sinonasal IP groups. On the other hand, a higher percentage of tumor cells showed IMP3 expression of high intensity among the third group of IP-associated SCC. Ki-67 IHC expression in the normal epithelium and sinonasal IP was characterized by nuclear staining of low to moderate intensity in a few cells. In the sinonasal SCC group, a high percentage of Ki-67 nuclear positivity was detected in the tumor cells (Figures
Normal respiratory epithelium: (a) by H&E (H&E ×200); (b) by IMP3 showing low intensity of cytoplasmic staining in scant epithelial cells (IHC ×200); (c) by Ki-67 showing low to moderate nuclear expression in nuclei of epithelial cells (IHC ×200).
Sinonasal inverted papilloma: (a) by H&E (H&E ×100); (b) by IMP3 showing low intensity of cytoplasmic staining in outermost epithelial cell layer of the IP masses (IHC ×100); (c) by Ki-67 showing scattered nuclear expression of moderate intensity in outermost layer of IP masses (IHC ×100).
Different cases of IP-associated sinonasal SCC: (a, c, e) IMP3 IHC cytoplasmic expression in tumor cells (IHC ×100); (b, d, f) Ki-67 IHC nuclear expression in tumor cells ((b, f) IHC ×100, (d) IHC ×200).
IMP3 expression was characterized by heterogeneity among SCC cases, such that negative areas alternate with positive areas of low intensity cytoplasmic staining of tumor cells among well-differentiated SCC. This pattern of heterogeneity gradually increases among moderately differentiated cases, until a diffuse cytoplasmic staining of moderate to strong intensity is seen among poorly differentiated SCC (Figure
Heterogenous expression of IMP3 in IP-associated sinonasal SCC: (a) in well-differentiated SCC, negative areas of tumor were seen alternating with areas of moderate intensity in well-differentiated SCC (IHC ×100); (b) in moderately differentiated sinonasal SCC, focal strong cytoplasmic staining alternate with foci of cytoplasmic staining of low intensity (IHC ×100); (c) poorly differentiated SCC showed diffuse moderate to strong cytoplasmic staining (IHC ×100).
The mean and median values of IMP3 and Ki-67 IHC expression are presented in Table
Comparison between the studied groups with regard to IHC expression of IMP3 and Ki-67 values.
Group | Sig | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Normal epithelium | Inverted papilloma | SCC | |||||||||
Mean | ±SD | Median (IQR) | Mean | ±SD | Median (IQR) | Mean | ±SD | Median (IQR) | |||
IMP3 | 6.86 | 2.63 | 6.9 (5.1–8) | 6.41 | 3.1 | 5.2 (4.6–8) | 31.93 | 22.11 | 17 (12–48) | 0.001 | HSA |
Ki-67 | 21.76 | 7.86 | 21 (16.5–28.9) | 20.05 | 6.43 | 19.9 (18–25.5) | 31.43 | 8.55 | 26 (25–38) | 0.001 | HSB |
ROC analysis revealed that both IMP3 and Ki-67 could be used to discriminate sinonasal SCC cases from normal/IP cases, with IMP3 at a cut-off value of ≥9.5 with 100% and 81.5% sensitivity and specificity, respectively, and Ki-67 at a cut-off value of ≥22.5 with 100% and 62.5% sensitivity and specificity, respectively (
ROC curve using IMP3 and Ki-67 to differentiate SCC cases from noncancerous cases.
IHC marker | AUC (CI) | Sensitivity | Specificity | +PV | -PV | |
---|---|---|---|---|---|---|
0.971 (0.918 to 0.994) | 100.00 | 81.52 | 46.9 | 100.0 | <0.0001 (HS) | |
0.828 (0.743 to 0.894) | 100.00 | 65.22 | 31.9 | 100.0 | <0.0001 (HS) |
AUC: area under the curve; CI: confidence interval; +PV: positive predictive value; -PV: negative predictive value; sig: significance; HS: highly significant.
Logistic regression to study independent factors associated with SCC.
Odds ratio (OR) | Sig | 95% CI for OR | |||
---|---|---|---|---|---|
Lower | Upper | ||||
Age | 1.044 | .670 | NS | 0.857 | 1.270 |
Gender (female) | 0.051 | .585 | NS | 0.0041 | 1.10 |
IMP3 expression | 2.443 | .005 | HS | 1.309 | 4.560 |
Ki-67 expression | 0.759 | .113 | NS | .539 | 1.068 |
Both markers showed statistically significant correlation with each other within each of the three groups (
Correlation between IMP3 and Ki-67 expression among the three studied groups.
Normal epithelium group | Ki-67 expression | |
IMP3 expression | 0.737 | |
0.0001 | ||
Sig | HS | |
Inverted papilloma group | Ki-67 expression | |
IMP3 expression | 0.818 | |
0.0001 | ||
Sig | HS | |
Sinonasal SCC | Ki-67 expression | |
IMP3 expression | 0.935 | |
0.0001 | ||
Sig | HS |
Spearman’s rho analysis revealed that both IMP3 and Ki-67 expressions were significantly related to lymph node and tumor stages, but not significantly related to tumor grade (Table
Correlation between IMP3 and Ki-67 expression and clinicopathological parameters of SCC cases.
IMP3 expression | Ki-67 expression | ||
---|---|---|---|
Lymph nodal stage | Rho | 0.842 | 0.788 |
0.0001 | 0.0001 | ||
Sig | HS | HS | |
Tumor histologic grade | Rho | 0.488 | 0.364 |
0.065 | 0.182 | ||
Sig | NS | NS | |
Tumor stage | Rho | 0.596 | 0.572 |
0.019 | 0.026 | ||
Sig | S | S |
ROC analysis was used to select clinically important cut-off scores of progression and survival for IMP3 in the 15 cases of sinonasal SCC on top of IP. At each percentage score, the sensitivity and specificity for each survival outcome under study were plotted, thus generating a ROC curve. The score having the closest distance to the point with both maximum sensitivity and specificity was selected as the cut-off score leading to the greatest number of tumors which were correctly classified as having or not having the clinical outcome. Table
ROC analysis using IMP3 to predict mortality, local recurrence, and distant metastasis among SCC cases.
IMP3 | AUC (CI) | Sensitivity | Specificity | +PV | -PV | |
---|---|---|---|---|---|---|
1.0 (1.0 to 1.0) | 100.00 | 100.00 | 100.00 | 100.00 | <0.027 (S) | |
1.0 (1.0 to 1.0) | 100.00 | 100.00 | 100.00 | 100.00 | <0.001 (HS) | |
1.0 (1.0 to 1.0) | 100.00 | 100.00 | 100.00 | 100.00 | <0.009 (HS) |
Kaplan-Meier analysis showed the correlation between IMP3 and overall survival (OS), where OS was 100% at 70 months follow-up at a
(a) Kaplan-Meier analysis showing correlation between IMP3 IHC expression and OS at a
Although a small number of sinonasal IP are associated with a synchronous or metachronous squamous cell carcinoma, malignant transformation of these lesions leads to long-term morbidity and/or mortality. This necessitates better and novel diagnostic, prognostic, and therapeutic approaches [
The current study showed that IMP3 and Ki-67 significantly correlated among the three studied groups. Moreover, both proteins had a specific pattern of expression that was nearly similar for the normal epithelium and IP groups and sharply increasing in the IP-associated SCC group. This was in agreement with the pattern described by Maržić et al. [
With its diverse morphology, IPs may be associated with varying degrees of atypia, dysplasia, carcinoma in situ, and SCC. By computed tomography scan, the sensitivity and specificity of the diagnosis of SCC in IP are low, and orbital wall invasion can be found in recurrent benign IP tumors [
Thus, the results of the current study are novel in this area, where ROC analysis revealed that both IMP3 and Ki-67 could be used to discriminate sinonasal SCC cases from IP cases. Furthermore, logistic regression demonstrated that IMP3 is the only statistically significant factor associated with sinonasal SCC after adjustment of other factors. This was in concordance with a previous study [
In the current study, IMP3 sensitivity and specificity in differentiating sinonasal SCC from noncancerous cases were 100% and 81.5%, respectively. This was unlike Maržić et al. and Chen et al. [
Our current work revealed that IMP3 expression was significantly related to lymph node and tumor stages, but not to tumor grade. This was in agreement with Clauditz et al. [
Guided by Zlobec et al. [
A previous study has demonstrated some promising results that support the use of IMP3 as a potential vaccination therapy in patients with head and neck SCC, which showed an immune response that improved the OS of patients [
The correlation between IMP3 expression to OS and higher incidence of local recurrence suggests targeting IMP3 might decrease the incidence of local recurrence, and thereby improve the prognosis of such cases [
The current study is the first to evaluate IMP3 IHC expression in three separate groups of samples from the sinonasal tract-normal sinonasal epithelium, IP, and SCC on top of IP. Among the points of strength of the current study was comparing the IMP3 IHC expression to an established biomarker of carcinogenesis, namely, Ki-67. It showed a potential diagnostic role for IMP3 in discriminating SCC on top of IP from noncancerous cases with sensitivity and specificity of 100% and 81.5%, respectively. This might suggest that in problematic cases, IMP3 IHC expression could serve as a low-cost ancillary aid. Nevertheless, more studies including dysplastic changes should be performed.
Moreover, the present study has demonstrated that IMP3 expression correlated with poor prognostic indicators and tumor progression in cases of SCC on top of IP. A potential therapeutic value in such cases could be achieved by targeting IMP3 in order to improve prognosis. However, only multi-institutional prognostic studies, to encompass a larger sample size, would be able to confirm the prognostic and therapeutic impact of IMP3 in this context.
Therefore, a limitation to this study is the small number of patients with IP-associated sinonasal SCC (15 cases). This might be attributed to the rarity of this diagnosis. Hence, we found it intriguing to explore all the data we had and to fully present it as a cornerstone for further studies to build upon.
This study is one of a few IHC studies investigating IP-associated sinonasal SCC. In addition, it proposed novel insights about the attributes of IMP3 in cancer initiation and progression through evaluation of its IHC expression in normal, benign neoplastic, and frankly malignant sinonasal SCC. Such results not only help in offering a better understanding of sinonasal carcinogenesis but also provide a potential low-cost ancillary tool to resolve enigmatic differential diagnoses with high yield diagnostic accuracy in IP-associated sinonasal SCC detection. Furthermore, in the malignant spectrum, IMP3 might also provide a promising prognostic marker and a potential therapeutic target.
All data generated or analyzed during this study are included in this published article.
The study was approved by the Research Ethical Committee (REC) at the Faculty of Medicine, Ain Shams University.
All patients who participated in this study signed a written informed consent before surgery.
The authors declare that they have no conflicts of interest/competing interests.
Sarah A. Hakim designed and coordinated the study, performed data collection and statistical analysis, reviewed the histological diagnosis, evaluated immunohistochemistry, and drafted the manuscript. Nermine M. Abd Raboh participated in the study design, reviewed the histological diagnosis, evaluated immunohistochemistry, carried out photographing, and coordinated and critically reviewed the manuscript. Lobna S. Shash participated in the study design, reviewed the histological diagnosis, evaluated immunohistochemistry, and coordinated and critically reviewed the manuscript. The authors read and approved the final manuscript.