Performance Characteristics of Incisional and Core Needle Biopsies for Diagnosis in Parotid Gland: Single-Institutional Experience and Assessment of the Value of a Milan System for Reporting Salivary Gland Cytopathology-Like Risk Stratification Model.

Abstract

Context.—: Unlike parotid fine-needle aspiration biopsy, standardized reporting for core needle biopsy (CNB) and incisional biopsy (IB) is not established.

Objective.—: To examine the value of risk stratification by a Milan System for Reporting Salivary Gland Cytopathology (MSRSGC)-like classifier for parotid CNB/IB.

Design.—: Five hundred ninety-two parotid biopsy records (CNB = 356, IB = 236) were retrieved (1994-2022) along with clinicopathologic data. Diagnoses were transformed to an MSRSGC-like classifier and compared with end points including risk of malignancy.

Results.—: Over time, CNB was progressively more used compared with IB. Overall malignancy call rate was 223 of 592 (37.7%). Common specific diagnoses included Warthin tumor, lymphoma subtypes, and metastatic squamous cell carcinoma for CNB and IB, in addition to pleomorphic adenoma for CNB. Descriptive diagnoses were still frequent. Nondiagnostic rates were higher in CNB (26 of 356; 7.30%) than IB (5 of 236; 2.12%; P <.001). Tissue volumes significantly influenced CNB adequacy, with minimum and optimal volumes of 4.76 mm³ (J index, receiver operating characteristic curve) and 12.92 mm³ (95th percentile of distribution), respectively. One hundred forty-four patients (112 CNBs) had follow-up resections; diagnoses were concordant for 66 of 73 adequate CNBs (90.41%). Our restructured risk grouping of MSRSGC categories performed robustly in terms of risk of malignancy (sensitivity = 85.5%, specificity = 100%, accuracy = 92.3%, area under the curve = 0.9677).

Conclusions.—: Although CNB and IB are amenable to a risk stratification system, there are some differences as compared with fine-needle aspiration biopsy, particularly given the high baseline prevalence of malignancy. Specific diagnoses are often feasible and concordant with resection. CNB tissue volume can inform optimal and minimal sampling recommendations for adequacy.