J Amjoto 2019 102279
J Amjoto 2019 102279
J Amjoto 2019 102279
Am J Otolaryngol
journal homepage: www.elsevier.com/locate/amjoto
Department of Otolaryngology – Head and Neck Surgery, Medical University of South Carolina, Charleston, SC, USA
Keywords: Purpose: Identify variables that are independent predictors of survival in carcinoma ex pleomorphic adenoma
Carcinoma ex pleomorphic adenoma (CXPA) of the major salivary glands using a population-based database and evaluate the incidence and man-
Salivary malignancy agement strategies for this rare malignancy.
Parotid gland Materials and Methods: The Surveillance, Epidemiology, and End Results (SEER) database was queried for all
Survival
cases of major salivary gland CXPA from 1973 to 2015.
Results: Of the 619 patients identified, the parotid gland was the most common site of involvement (76.9%, 476/
619). The reported incidence of CXPA has risen in the past decade (2005–2015, 0.24 to 0.63 per 1,000,000). The
2-year and 5-year disease-specific survival (DSS) rates were 90.3% and 80.4%, respectively. On univariate
analysis, facial nerve sacrifice was not a statistically significant predictor of survival (HR = 1.213, 95% CI
[0.588–2.058], P = 0.602). Patients with a tumor size > 4 cm, multiple positive lymph nodes, and distant
metastatic disease had a 2 to 4-fold statistically significant increase in mortality using a multivariate analysis.
Statistical significance was not demonstrated in the DSS of patients who underwent partial versus total par-
otidectomy procedures.
Conclusions: CXPA is a rare salivary malignancy that has a reported increased incidence in the last decade.
Tumor size > 4 cm, multiple positive lymph nodes, and distant metastatic disease are predictors of disease-
specific mortality. Further research should be conducted to improve early detection and survival strategies for
this salivary cancer.
Level of Evidence: 4.
1. Introduction should raise suspicion for CXPA. The gold standard for diagnosing
CXPA is with histopathological evaluation of the suspected lesion [6].
Carcinoma ex pleomorphic adenoma (CXPA) is a rare malignant Fine needle aspiration cytology is typically used for preoperative eva-
tumor of the salivary glands [1]. CXPA is part of the malignant mixed luation. Diagnosis can be challenging if the entire PA component is
tumors category, which contains both benign and malignant compo- replaced by the malignant tumor. On the other hand, the malignant
nents on histologic analysis. This malignant tumor arises from the component may be sparse or scattered and missed when analyzing the
epithelial component of either primary or recurrent pleomorphic ade- pathological specimen [1]. Both scenarios can lead to misdiagnosis. As
noma (PA) [1]. Based on prior studies, the average prevalence of CPXA a result, this high-grade tumor can potentially lead to distant metas-
ranges from approximately 3% to 15% among all malignant salivary tasis. In terms of prognosis, survival rates reported in previous litera-
tumors [2–5]. ture range from 30% to over 70% [1–4,7].
CXPA most commonly presents as a firm mass in the parotid gland, Currently, there is limited body of evidence regarding the man-
although the submandibular and minor salivary glands may also be agement of these malignant salivary tumors. The purpose of this study
involved [6]. This presentation can be mistaken for a PA; however, a is to provide a current and comprehensive update on the management
rapid increase in size observed in a previously slow growing mass of these rare tumors by querying the Surveillance, Epidemiology, and
☆
No disclosures, financial or otherwise.
☆☆
No conflicts of interest.
☆☆☆
No financial support.
⁎
Corresponding author at: 135 Rutledge Avenue, MSC 550, Charleston, SC 29425, USA.
E-mail address: [email protected] (A. Gupta).
https://doi.org/10.1016/j.amjoto.2019.102279
Received 17 July 2019
0196-0709/ © 2019 Elsevier Inc. All rights reserved.
Please cite this article as: Avigeet Gupta, et al., Am J Otolaryngol, https://doi.org/10.1016/j.amjoto.2019.102279
A. Gupta, et al. Am J Otolaryngol xxx (xxxx) xxxx
2.1. Data
2
A. Gupta, et al. Am J Otolaryngol xxx (xxxx) xxxx
Table 2
Carcinoma ex Pleomorphic Adenoma treatment modalities.
Treatment Number of Patients
N = 619a %
b
Surgery
No surgery 17 2.8
Surgery 600 97.2
Type of surgery
Partial parotidectomy 273 52.5
Total parotidectomy 213 42.3
Facial nerveb
Spared 181 68.8
Sacrificed 82 31.2
Radiation
None 240 38.7
External beam 379 61.2
Chemotherapy
None/Unknown 549 88.7
Chemotherapy 70 11.3
a
Total number of cases for certain variables may be less than N because of
missing values.
b
Surgery type and facial nerve were identified for patients from 1998 to Fig. 2. Partial versus Total Parotidectomy Kaplan-Meier Disease Specific
2015. Survival Comparison.
The overall 2-year and 5-year survival rates were noted to be 84.6%
and 68.5%, respectively. Overall rates of DSS at 2-year and 5-year in-
Fig. 3. Kaplan-Meier Disease Specific Survival Stratified by Tumor Size.
tervals were 90.3% and 80.4%, respectively. Survival analysis of those
who underwent partial or total parotidectomy demonstrated no sig-
nificant differences in survival using a log-rank test (P = 0.278) (HR = 5.313, 95% CI [2.424–11.647]), late stage tumors (HR = 5.678,
(Fig. 2). Stratification of DSS by tumor size demonstrated a decreased 95% CI [2.562–12.583]), and multiple lymph node involvement
DSS with an increased tumor size. There was a significant difference in (HR = 5.411, 95% CI [3.374–8.677]) were associated with a > 5-fold
survival among patients with tumors of 4 cm or greater and those with a increase in mortality. On the other hand, partial parotidectomy was
tumor size of < 4 cm using a log-rank test (P < 0.001) (Fig. 3). associated with decreased mortality (HR = 0.636, 95% CI
[0.413–0.979]). Of note, facial nerve sacrifice was not a demonstrated
to be a statistically significant predictor of survival (HR = 1.213, 95%
3.5. Predictors of survival
CI [0.588–2.058], P = 0.602).
After adjusting for relevant covariates using multivariable analysis,
The following tumor characteristics were demonstrated to be sta-
tumor size of 4 cm or greater (HR = 2.394, 95% CI [1.121–5.113]),
tistically significant (P < 0.05) predictors of mortality by univariate
multiple positive lymph nodes (HR = 3.886, 95% CI [1.806–8.364]),
analyses: high grade, late stage, distant metastasis, tumor size, extra-
and distant metastatic disease (HR = 3.661, 95% CI [1.352–9.915])
parenchymal extension, multiple lymph node involvement (lymph node
were found to be significant (P < 0.05) independent predictors of
levels I–V), and partial parotidectomy (Table 3). Notably, distant me-
mortality (Table 4).
tastasis was associated with a 13-fold increase in mortality (hazard ratio
(HR) = 12.984, 95% CI [6.926–24.338]). High grade tumors
3
A. Gupta, et al. Am J Otolaryngol xxx (xxxx) xxxx
Table 3 noted to be 0.68 per 1,000,000 by Rebhun et al. [8] using population-
Univariate analyses of disease-specific survival. based data through 2011. Based upon our study, the reported incidence
Risk factor Hazard ratio 95% CI P value of CXPA has been rising, especially in the last decade (0.24 to 0.63 per
1,000,000 from 2005 to 2015). A possible interpretation of this result
Female 1.213 0.826–1.782 0.325 may arise from the increased recognition and awareness of CXPA as a
Age 1.012 0.998–1.026 0.086
specific diagnostic entity by pathologists that specialize in salivary
African-American 0.724 0.318–1.652 0.444
High grade 5.313 2.424–11.647 < 0.0001⁎
gland malignancies. The pathological diagnosis of CXPA can prove to be
Late stage 5.678 2.562–12.583 < 0.0001⁎ difficult in unusual cases and lead to potential misclassifications. Dif-
Distant metastasis 12.984 6.926–24.338 < 0.0001⁎ ferential diagnoses for CXPA include, but are not limited to, benign
Tumor size pleomorphic adenomas, metastatic mixed tumors, high grade salivary
≤2 cm 0.511 0.288–0.908 0.022⁎ adenocarcinomas, and other salivary gland malignancies.
2-4 cm 0.589 0.366–0.946 0.029⁎ The pathogenesis of this disease is poorly understood, but geo-
≥4 cm 2.642 1.714–4.073 < 0.0001⁎
graphical differences in the prevalence of this tumor have been pre-
Extraparenchymal Extension 3.677 2.423–5.579 < 0.0001⁎
viously described. In the United Kingdom, Malata et al. [9] demonstrated
Number of positive nodes
that CXPA represents 25% of all primary parotid malignancies. In con-
One 1.347 0.645–2.812 0.428
Multiple 5.411 3.374–8.677 < 0.0001⁎
trast, CXPA had a prevalence of 14% of primary parotid malignancies in
Switzerland [4], and only 10% in Japan [13]. In the United States, a
Lymph node level
Level 1 4.041 2.204–7.410 < 0.0001⁎
study by Byrne and Spector [14] noted CXPA to represent 12% of pri-
Level 2 4.707 4.707–7.648 < 0.0001⁎ mary parotid cancers. Further work is necessary to understand the
Level 3 6.785 3.984–11.555 < 0.0001⁎ burden of disease, risk factors of pathogenesis, geographical differences,
Level 4 5.463 2.365–12.618 < 0.0001⁎ and this phenomenon of a reported rising incidence in the United States.
Level 5 7.688 3.675–16.080 < 0.0001⁎
4
A. Gupta, et al. Am J Otolaryngol xxx (xxxx) xxxx