INTRODUCTION
Salivary gland tumors present considerable diagnostic challenges due to their heterogeneous nature and complex histopathology. The fifth edition of the World Health Organization Classification of Head and Neck Tumors lists 15 benign and 21 malignant epithelial tumors of the salivary glands [1]. This diversity stems from the specific composition of these glands [2]. The malignancy rate for parotid gland tumors ranges from 15% to 25% [3,4]. Accurate preoperative assessment is crucial for differentiating between malignant and benign tumors, as well as for determining the appropriate surgical method and extent.
Fine needle aspiration (FNA) cytology is widely utilized to evaluate salivary gland lesions. However, its diagnostic utility remains a subject of debate [5]. In clinical practice, FNA often fails to deliver the level of precision required for optimal surgical planning, leading to unexpected malignancy due to preoperative false negatives [6,7]. In contrast, core needle biopsy (CNB) can obtain sufficient tissue, preserve histological architecture, and provide more accurate histological assessments than FNA [8]. A recent meta-analysis demonstrated that CNB has higher sensitivity and specificity than FNA in distinguishing between malignant and benign tumors, with fewer non-diagnostic results [9].
We hypothesized that if CNB could provide a preoperative diagnosis with high accuracy, the prognosis would differ between cancers with preoperative “malignant” and “less-than-malignant” CNB diagnoses. In this study, we aimed to explore the clinical utility of preoperative CNB diagnosis in patients undergoing primary surgery of the parotid gland and to assess its prognostic implications for parotid gland cancers.
MATERIALS AND METHODS
This study was conducted in accordance with the principles stated in the Declaration of Helsinki. Approval for this research was obtained from Institutional Review Board of Seoul National University Bundang Hospital (No. B-2402-885-103). Informed consent for this retrospective analysis was waived.
Study population and data collection
We conducted a retrospective review of patients who either ultrasound-guided CNB or ultrasound-guided FNA, and primary parotidectomy at Seoul National University Bundang Hospital, a tertiary referral hospital in Gyeonggi-do, Korea, between August 2003 and August 2023. Patients who underwent either ultrasound-guided CNB or ultrasound-guided FNA, and primary parotidectomy were included in the cohort. A majority of the FNA and CNB procedures were performed in our hospital and by dedicated specialist radiologists. While a small number of procedures were performed at external affiliated hospitals due to delays in the schedule, all specimens were analyzed, interpreted, and made diagnoses at our institution by dedicated pathologists. Patients with any of the following categories were excluded: those lacking preoperative CNB or FNA results, individuals who underwent revision parotidectomy, and patients ultimately diagnosed with lymphoma or metastatic cancer involving the parotid gland. Demographic data of the patients, results of preoperative CNB or FNA, definitive postoperative histopathologic diagnosis, grade, and features, treatment modalities, cancer staging, and information on recurrence and survival were obtained from our institutional clinical data warehouse.
CNB and FNA report and diagnostic classification
Preoperative diagnostic results from CNB and FNA were categorized into 6 groups: inadequate, benign, nonneoplastic, indeterminate, suspicious for malignancy (SFM), and malignant. When analyzing prognostic factors for disease-free survival (DFS), preoperative CNB diagnoses were classified into “malignant on CNB” and “less-than-malignant on CNB,” and the latter included SFM, indeterminate, nonneoplastic, benign, and inadequate results on CNB. When calculating for accuracy, inadequate biopsy results were excluded. The diagnosis of SFM was considered as positive for malignancy in calculating diagnostic performance measures. The final histopathologic reports after parotidectomy were used as the reference standard. Histologic grade of parotid cancer was categorized into either low/intermediate or high grade according to the postoperative histopathological report. Malignant tumors were staged according to the AJCC Cancer Staging Manual (7th edition, AJCC-7).
Statistical analysis
Categorical data were reported as numbers and percentages and continuous data as mean and standard deviations (SD). All statistical tests were two-tailed, and the statistical significance was defined as a P-value less than 0.05 (P<0.05). Clinicopathological characteristics were compared according to preoperative results, between “malignant on CNB” and “less-than-malignant on CNB,” using t-test and Mann-Whitney U-test for continuous variables and chi-square test and Fisher’s exact test for categorical variables.
To identify prognostic factors associated with DFS in the CNB group, univariable Cox proportional hazards regression analyses were performed based on log-rank test. To identify independent predictors of DFS among preoperatively available factors, including age, sex, and preoperative diagnosis, variables found to have a P-value less than 0.1 were further included in multivariable Cox proportional hazards regression models. Cox regression results were presented as hazard ratios (HRs) with 95% CI and P-value calculated from log-rank tests.
To assess the accuracy assessments of CNB and FNA in diagnosing malignancies, the following parameters were calculated: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy ([true positive+true negative]/[true positive+true negative+false positive+false negative]). SFM and malignant results were considered as test positive for malignancy. Inadequate results were excluded from the calculation. Exact binomial confidence limits were calculated for test sensitivity, specificity, and positive and negative predictive value. All analyses were performed using R software (version 4.2.2) and R studio (version 2023.6.0.421).
RESULTS
Patient characteristics
Among 880 patients assessed using CNB, 615 underwent primary parotidectomy (termed the CNB group). This group included 331 male (53.8%) and 284 female (46.2%) participants, with a mean age of 53.4 years (SD, 15.8 years). Preoperatively, 142 cases (23.1%) were diagnosed as SFM or malignant. Of 953 patients evaluated with FNA, 547 underwent primary parotidectomy. The median time interval between CNB or FNA and surgery was 61.0 days for both techniques (interquartile range [IQR], 32.5–109 days for CNB and 31.0–98.5 days for FNA). The FNA group consisted of 547 patients, 300 of whom were male (54.8%), with a mean age of 52.3 years (SD, 15.9 years); 64 cases (11.7%) had preoperative diagnoses of SFM or malignancy. Post-parotidectomy histologic diagnoses revealed malignancies in 102 (16.6%) CNB cases and 81 (14.8%) FNA cases. In the CNB group, the risk of malignancy was 43.3% for SFM and 96.2% for malignant diagnoses. By contrast, in the FNA group, the risk of malignancy was 31.8% for SFM and 95.0% for malignant diagnoses (Table 1).
Among the 102 patients (52 male [51%]; mean age, 56.7 years [SD, 15.7 years]) ultimately diagnosed with parotid gland malignancies in the CNB group, 70 patients (69%) were 65 years old or younger, while 32 (31%) were older than 65 years. Preoperative CNB categorized 50 cases (49%) as malignant, with the remaining 52 (51%) receiving other diagnoses, such as inadequate, nonneoplastic, benign, indeterminate, or SFM; collectively, these were termed “less than malignant.” Tumor grading revealed that 84 (82%) cases were low- or intermediate-grade, while 18 (18%) were high-grade. The distribution of pT stages was as follows: 29 cases at T1, 40 at T2, 24 at T3, and nine at T4. In terms of pN stage, 85 cases were classified as x/0, five as N1, nine as N2, and three as N3. Extraparenchymal extension was observed in 28 cases, while perineural invasion was present in 18 cases. Negative surgical margins were achieved in 82 cases, and 43 patients received postoperative radiotherapy.
Correlation between preoperative CNB diagnosis and histologic factors
A finding of malignancy on CNB was associated with higher rates of advanced pT stage (T3 or T4), positive lymph nodes (pN+), high tumor grade, and positive extraparenchymal extension relative to a CNB finding of “less than malignant.” No significant correlation was found between preoperative CNB diagnosis and margin status (Table 2).
Survival analysis of parotid gland malignancy in the CNB cohort
In the CNB group, 10 of 102 patients experienced locoregional recurrence or distant metastasis during the follow-up period, which had a median duration of 37.5 months (IQR, 17.9–60.9 months). Of these, six patients encountered locoregional recurrence or distant metastasis within the first year. Three patients developed locoregional recurrence, while the other seven exhibited distant metastasis. Overall, among patients with parotid gland malignancy in the CNB group, the 5-year predicted DFS was 86.9% (95% CI, 79.2%–95.3%). Patients with a diagnosis of malignancy on CNB had a significantly poorer 5-year predicted DFS (73.0%; 95% CI, 57.2%–93.2%) compared to those with diagnoses in the less-than-malignant category (98.0%; 95% CI, 94.1%–100%) (P<0.01) (Fig. 1).
Univariable Cox regression analysis for DFS indicated that male sex (HR, 11.47; P<0.01) and a CNB finding of malignancy (HR, 11.03; P<0.01) were associated with comparatively poor DFS. Among postoperative factors, high-grade malignancy, advanced pT stage, lymph node involvement, extraparenchymal extension, and perineural invasion were associated with poor DFS (Supplementary Table 1). Multivariable Cox regression showed that male sex (HR, 8.48; 95% CI, 1.05–68.76) and a finding of malignancy on CNB (HR, 8.20; 95% CI, 1.01–66.15) remained significant predictors of poorer DFS (Table 3). When including all potential postoperative prognostic factors in the multivariable Cox regression analysis, male sex, histologic grade, and perineural invasion were significant predictors of unfavorable DFS.
Diagnostic performance of CNB compared to FNA
After excluding inadequate results, CNB demonstrated a higher sensitivity for malignancy (89.0%) than FNA (45.2%). This indicates that CNB has a superior capacity to identify potential malignancies. Additionally, CNB exhibited a higher NPV, at 97.6%, compared to FNA (90.8%) (Table 4).
DISCUSSION
Based on these findings, CNB is valuable not only for accurate preoperative diagnosis but also as a prognostic predictor of DFS in patients with parotid gland cancer. A preoperative diagnosis of malignancy on CNB was identified as a prognostic factor linked to comparatively poor DFS in cases of parotid gland cancer. Additionally, CNB was significantly more sensitive than FNA in detecting parotid gland malignancies. To our knowledge, this study is the first to explore the prognostic value of preoperative CNB in parotid gland malignancy. Among the available preoperative clinical data, male sex and a preoperative diagnosis of malignancy on CNB were found to be independent prognostic factors.
A preoperative diagnosis of malignancy obtained via CNB was associated with a poorer DFS than diagnoses termed “less than malignant,” which included nonneoplastic, benign, indeterminate, or SFM findings. Thus, the results of preoperative CNB offer valuable prognostic information. This information may be critical in determining the therapeutic approach and extent of surgery, as well as in counseling patients. Our findings indicate that a diagnosis of malignancy on CNB was associated with high tumor grade, advanced pT stage (T3 or T4), pN+ status, and extraparenchymal extension. Relative to FNA, CNB employs larger-bore needles to obtain intact tissue cores, thus preserving the histological architecture and enabling a more thorough assessment [8]. The superior diagnostic accuracy of CNB, particularly in distinguishing malignant from benign tumors with fewer nondiagnostic results, was confirmed in a recent meta-analysis [9]. Consequently, the preoperative CNB result may be associated with tumor aggressiveness, thus impacting the prognosis of parotid cancer. Meaningful changes in and distortions of the normal glandular architecture, which are characteristic of highly aggressive salivary gland cancers, may be detected in CNB specimens but not in FNA samples. These features lead to a more definitive diagnosis of malignancy, rather than SFM or other less definitive diagnoses. Therefore, preoperative CNB diagnosis can be considered reflective of tumor grade. Conversely, early-stage tumors may exhibit less aggressive morphology in CNB samples, potentially resulting in false negative or SFM diagnoses. Thus, CNB could serve as a potential surrogate for identifying postoperative histological markers of aggression.
Despite a less-than-malignant diagnosis on CNB failing to confirm cancer prior to surgery, it did not adversely affect prognostic outcomes. A previous multicenter case series study, which included 85 cases of unexpected malignant tumors of the parotid gland, found that these preoperatively undiagnosed parotid cancers were subject to excellent local control and overall survival. The study reported 5-year overall and relapse-free survival rates of 100.0% and 95.2%, respectively [10]. Our study largely concurs with these findings, indicating that malignancies not confirmed preoperatively still demonstrated relatively favorable prognostic outcomes.
In our study, sex, along with preoperative CNB diagnosis, was identified as a preoperative predictive factor for DFS. Previous research has indicated that male sex is a poor prognostic factor in salivary gland tumors [11-13]. These studies have suggested that male sex is associated with reduced survival, although only one identified sex as an independent prognostic factor [12]. Several biological and behavioral factors may contribute to the link between male sex and lower DFS in parotid gland cancers [14]. From a biological perspective, differences in genetics, epigenetics, and the influence of sex hormones could impact cancer outcomes [15]. Women typically exhibit stronger immune responses than men, which may support their defense against cancer cells and lead to more favorable prognoses [16]. Behaviorally, lifestyle factors such as smoking and alcohol use, which may be more prevalent among men, as well as sex differences in health care-seeking behavior have been suggested to contribute to the observed survival disparities [17].
Among potential postoperative prognostic factors, multivariable Cox regression analysis identified male sex, histologic grade, and perineural invasion as significant predictors of reduced DFS. A multicenter study involving 195 patients with parotid gland carcinoma found that high-grade histology was an independent prognostic factor in multivariate analysis, along with the presence of positive neck lymph nodes [18]. Additionally, one meta-analysis revealed that perineural invasion was associated with poorer overall survival, DFS, disease-specific survival, and distant-metastasis-free survival [19].
In our study, we compared the diagnostic performance of CNB and FNA in 616 and 547 patients, respectively, with parotid gland tumors. The results indicated that CNB was superior to FNA, particularly in sensitivity (89.0% vs. 45.2%) and NPV (97.6% vs. 90.8%). Although the risk of malignancy in parotid gland tumors is lower than in other salivary glands, the former are more prevalent [3,4]. A meta-analysis of 6,169 patients reported the sensitivity and specificity of FNA for differentiating malignancy in parotid gland lesions as 80% (95% CI, 76%–83%) and 97% (95% CI, 96%–98%), respectively. The prevalence of malignant disease was 25%, with a PPV of 0.90 and an NPV of 0.94 [5]. In contrast, a meta-analysis of 277 cases indicated that CNB had a sensitivity of 0.92 (95% CI, 0.77–0.98) and a specificity of 1.00 (95% CI, 0.76–1.00), suggesting that CNB’s sensitivity is slightly higher than the sensitivity of 0.89 observed in our study [6]. Another recent meta-analysis demonstrated that CNB had a higher sensitivity (0.92; 95% CI, 0.88–0.96) and specificity (1.00; 95% CI, 0.99–1.00) for differentiating malignant salivary gland neoplasms from benign lesions compared to FNA, which had a sensitivity of 65% (95% CI, 55%–73%) and a specificity of 97% (95% CI, 94%–98%) [8]. Our data align with the literature in that CNB displayed a higher sensitivity than FNA. This finding suggests that CNB is more reliable for detecting malignancies in parotid gland tumors, which could lead to more appropriate treatment decisions. The slightly lower measures of diagnostic accuracy in our study compared to previous research may stem from the fact that many of the FNA and CNB procedures were performed at other hospitals, with only the cytologic or histologic analyses conducted at our institution. This could have introduced heterogeneity in the procedures, potentially yielding reduced sensitivity and specificity.
The primary limitations of this study stem from its single-institution, non-randomized, retrospective design, which may limit the generalizability of the findings. Furthermore, our inclusion criteria were limited to patients who underwent primary parotidectomy with confirmed parotid cancer diagnoses, thus excluding those who did not receive primary surgery. This exclusion could have introduced selection bias. In the group with a “less-than-malignant” finding on CNB, limited information was available regarding preoperative clinical T staging or N staging, since these patients were not confirmed to have malignancy before surgery. While not a limitation per se, our study did employ the seventh edition of the AJCC staging system. However, the only substantial change from the seventh to the eighth edition regarding salivary gland cancers was the incorporation of extranodal extension in N staging. Since our analysis was based on N positivity, our findings are fully applicable to the eighth edition. Finally, long-term follow-up is essential for salivary gland cancers due to their potential for indolent tumor biology and the possibility of late recurrence in low-grade tumors. Consequently, future multicenter prospective studies with extended follow-up periods are warranted.
CNB is valuable not only for accurate preoperative diagnosis but also as a prognostic indicator in patients with parotid gland cancer. Preoperative CNB diagnosis was shown to be a prognostic factor for DFS. Furthermore, the sensitivity and NPV of CNB were higher than those of FNA in the diagnosis of parotid gland malignancy.
HIGHLIGHTS
▪ A diagnosis of malignancy on core needle biopsy (CNB) was associated with poor prognosis in cancers of the parotid gland.
▪ CNB demonstrated higher sensitivity and negative predictive value compared to fine needle aspiration.
▪ Preoperative CNB can be a valuable tool for diagnosing malignancy and predicting prognosis in parotid gland cancers.
Download PDF File
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement
Print 
