Concordance of histopathologic and cytologic grading in musculoskeletal sarcomas: can grades obtained from analysis of the fine-needle aspirates serve as the basis for therapeutic decisions?

Abstract

Utilization of fine-needle aspiration (FNA) for the diagnosis of musculoskeletal lesions has been de-emphasized by many clinicians because of concerns regarding the technique's ability to supply all information necessary for prognostication and appropriate selection of therapy. Paramount among the concerns is the belief that FNA is incapable of supplying precise subtyping and grading in many cases. Secondary concerns regarding the use of FNA involve its perceived inability to supply adequate tissue for ancillary studies including immunohistochemistry and molecular diagnostic analysis. The authors investigated the ability of FNA to accurately subtype and grade a series of 107 primary sarcomas of the musculoskeletal system. The files of the Departments of Pathology at Duke University Medical Center and the University of California at Los Angeles and the private consultation files of one of the authors were searched for all fine-needle aspirates of sarcomas arising within the musculoskeletal system. A total of 107 cases were obtained and reviewed by three board-certified cytopathologists. Each cytopathologist independently assigned subtype when possible and generated a grade for each of the sarcomas. Corresponding surgical material was available for 77 cases. The surgical material was re-reviewed for accuracy of diagnosis and assignment of grade independently of the cytologic examination. Correlation of cytologic grade with histopathologic grade was made and analyzed by the kappa test. In addition, agreement on grade between cytopathologists was analyzed, and accuracy of histologic subtype prediction by cytologic analysis was studied. Most of the sarcomas were cytologically graded as Grade 1 or 2 but were assigned a histologic grade of 2 or 3. Cytopathologist A graded 93% of the sarcomas as either Grade 1 or 2, cytopathologist B graded 89% of cases as Grade 1 or 2, and cytopathologist C graded 94% of cases as Grade 1 or 2. Histologic evaluation demonstrated 82% of cases to be Grade 2 or 3. Correlation of cytologic and surgical grade varied among the observers. Cytopathologist A showed the highest degree of correlation between cytologic and surgical grade with an r value of 0.5. The corresponding r values for cytopathologists B and C were 0.46 and 0.41, respectively. Correlation between cytologic and surgical grade was significant for all three observers (P < 0.001). Only nuclear grade showed a consistent correlation in predicting final surgical pathology grade. Cellularity, mitotic rate, and the presence of necrosis were not statistically significant for predicting histopathologic grade. Accurate exact subtyping by cytologic examination was achieved in approximately 55% of cases. Prediction of histopathologic subtype by cytologic analysis was most successful when a distinctive stroma was present or high-grade features indicative of malignant fibrous histiocytoma were observed. There was little agreement as to histopathologic type as predicated by cytology for low-grade spindle cell sarcomas. Although there was a statistically significant correlation between cytologically assigned grade and final histopathologic grade, statistical analysis revealed only a moderate correlation between the two with an overall r value of approximately 0.57. Cytologic analysis tended to undergrade in comparison to final histopathologic grading. Only analysis of nuclear atypia showed good correlation with final surgical grade. Fine-needle aspiration was only moderately successful at predicting histopathologic subtype for musculoskeletal sarcomas in this series.