Robert T. Pu

BACKGROUND. Tissue microarray allows rapid and efficient evaluation of gene expression at the protein level and of immunochemical markers. To our knowledge, there has been no report of constructing cytology microarray using effusion cell blocks and testing its utility in immunochemical marker validation. METHODS. A total of 23 malignant effusions (primary tumor of breast [5], GI tract [5], lung [5] and ovary [8]) were used to construct a cytology microarray so that 3 cores of 0.6 mm in diameter were taken from the original cell blocks. Antibodies including AE1/AE3, EMA, and Ki-67 were applied to all cases, and CK7, CK20, TTF-1, WT-1, ER, and PR antibodies were used for selected cases. The cellularity, composition of cells, the staining pattern, and the intensity of each antibody were compared between corresponding cell block sections and CMA cores. RESULTS. The composition of tumor cells in the original block and the cores (including Sections 1 and 45) on cytology microarray were similar, ranging from 5% to 90%. Immunostains of AE1/AE3 and EMA were all positive and 100% concordant between the originals and cytology microarray. Similarly, CK7, CK20, ER, PR, TTF-1, and WT-1 stained both original blocks and cytology microarray with a high level of agreement with respect to percentage of positive cells, staining pattern (cytoplasm or nuclear), and intensity. Ki-67 stain showed slightly lower concordance (84%) with a few cases not in agreement because of low tumor burden in the original block coupled with low percentage of staining by antibody. CONCLUSIONS. Three 0.6 mm cores of cytology microarray are representative of the original cell block with cellularity and antibody staining pattern, intensity, and percentage. Therefore, CMA has a great potential in clinical research and practice as it allows rapid validation of immunocytochemical markers. Cancer (Cancer Cytopathol) 2008. © 2008 American Cancer Society.

The study was aimed to investigate the impact of using standard criteria for assessing specimen adequacy and diagnostic terminology (CAST) on fine‐needle aspiration (FNA) diagnosis and clinical management of thyroid nodules. The study included similar numbers of FNAs performed in 2 year before (group A) and 1.5 year after (group B) implementing the standard CAST. In comparison to group A, group B showed a significantly lower rate of nondiagnostic specimens (RND) (16.1% vs. 21.6%, P ≤ 0.01) and rate of descriptive diagnoses (RDD) (3.8% vs. 14.5%, P ≤ 0.001) and greater non‐neoplastic (70.0% vs. 64.1%, P < 0.05) and follicular cell lesions (7.4% vs. 4.3%, P < 0.05) but a similar percentage of neoplastic diagnoses. The rate of surgical follow‐up (RSF) was significantly higher in group B than in group A, overall (21.6% vs. 17.0%, P < 0.05), or in subgroups of non‐neoplastic (12.6% vs. 5.4%, P < 0.01) and neoplastic categories (81.0% vs. 61.0%, P < 0.05). The rate of cytohistologic concordance was higher in group B although the difference was not statistically significant. We concluded that use of the standard CAST on FNA diagnosis of thyroid nodules significantly reduced RND and RDD, providing more consistent diagnoses among the pathologists as well as better and more uniform communication between the pathologists and the clinicians. Furthermore, the cytohistological concordance was slightly better after CAST implementation, indicating that the improvement of diagnostic consistency among pathologists did not sacrifice the diagnostic accuracy. Diagn. Cytopathol. 2008;36:161–166.

Cherubism is a rare benign, non‐neoplastic condition characterized by facial alteration due to symmetrically expansion of bilateral mandibles and maxillas. A fine‐needle aspiration of a right cheek mass was performed in a 17‐year‐old teenager who was previously diagnosed Cherubism. The smears showed moderate cellularity with spindle cells mixed with multinucleated giant cells of osteoclast type. The cells appeared bland with normal nuclear/cytoplasm ratio. The nuclei were round or oval with smooth nuclear membranes and fine, evenly distributed chromatin. These cytologic features were considered consistent with but not diagnostic of Cherubism. Clinical correlations are needed in distinguishing Cherubism from other lesions containing giant cells. Diagn. Cytopathol. 2008;36:188–189.

BACKGROUND. Missed papillary thyroid carcinoma (PTC) diagnoses on fine-needle aspiration (FNA) can result from many causes. To the authors' knowledge, the issue of whether the detection of PTC is correlated with nodule heterogeneity has not been studied to date. METHODS. The authors identified all thyroidectomy specimens with a diagnosis of PTC that had undergone at least 1 prior FNA in the study institution between 1998 and 2003. The tumor size at the time of the resection, the ultrasound (US)-determined nodule size, and other parameters were compared between the 2 groups in which PTC was or was not diagnosed on FNA. RESULTS. Of a total of 89 specimens, 47 were diagnosed on FNA with an average tumor size of 1.7 cm and an US-determined nodule size of 2.1 cm (a difference of 0.4 cm). Forty-two specimens with a smaller average tumor size of 0.9 cm (P <.0001) and a US-determined nodule size of 2.4 cm (a difference of 1.5 cm) were missed. The differences with regard to the US-determined nodule size and tumor size between the 2 groups were significant (0.4 cm vs 1.5 cm; P < .0001). In the missed group, 29 specimens were found to have PTC foci that measured ≤1.0 cm and 26 had a reasonable size difference (RSD; defined as a PTC size outside the range of ±50% of the US-determined nodule size) as the indicator of the mixed nature of nodules targeted for FNA, whereas in the diagnostic group, 9 foci measured ≤1.0 cm and 6 had RSD. There was no cytologic evidence with which to render a diagnosis of PTC on further review in the missed group. CONCLUSIONS. The major reason for a missed diagnosis of PTC on FNA is because of inadequate tumor sampling due to the heterogeneity of the nodule targeted for FNA. This is illustrated by the RSD noted between the targeted nodule and the actual PTC tumor focus in the resection specimen. Cancer (Cancer Cytopathol) 2008. © 2007 American Cancer Society.

To distinguish carcinoma, either adenocarcinoma (ADC) or squamous cell carcinoma (SCC), and malignant mesothelioma (MM) in effusion can be a diagnostic challenge based on morphology alone. This study evaluates the utility of WT‐1, p63, MOC31, mesothelin, and cytokeratin (K903 and CK5/6) immunostains in effusions when ADC and SCC of the lung are in the differential diagnosis with MM. A cohort of 43 effusions consisting of lung ADC (N = 10), SCC (N = 15), and MM (N = 18, mostly (16) pleural based), was subjected to immunostains using the above mentioned antibodies. WT‐1 was positive in 100% MM, 0% ADC, and 0% SCC cases while p63 was positive in 0% MM, 30% ADC, and 80% SCC cases. Stain for MOC31 was positive in 100% ADC, 67% SCC, and 35% MM cases. Similarly, mesothelin antibody stained 100% ADC, 60% SCC, and 47% MM cases. Antibodies for K903 and CK5/6 stained 100% SCC cases but fewer ADC cases (40 and 10%, respectively). In conclusion, in this cohort of mostly pleural malignant effusion, MM can be identified with positive staining for WT‐1 and negative staining for p63. Conversely, negative staining with WT‐1 and positive staining for p63 exclude MM. Used as part of an immunostain panel, cytokeratin markers (CK5/6 and K903) are useful in differentiating SCC from ADC when MM is already excluded, and MOC31 might have limited value in differentiating ADC from MM. A negative stain with MOC31 can exclude lung ADC. Mesothelin, on the other hand, is not useful in the differential diagnosis of ADC, SCC, and MM. Diagn. Cytopathol. 2008;36:20–25.

We tested whether methylation profiles generated by real‐time methylation‐specific PCR (MSP) can be useful in differentiating benign, reactive mesothelial cell proliferation (RM) from malignant mesothelioma (MM). Forty‐two of the 63 cases (67%) yielded informative results for RARβ2, GPC3, CDKN2A (p16), TERT, and CCND2 (cyclinD2) gene methylation. DNA methylation of any gene was observed in much higher frequency in MM cases than RM cases (63% vs. 33%, P < 0.05). Individual gene methylation was higher in the MM than the RM cases for most of the genes; however, this was not statistically significant (RARβ2: 58% vs. 33%, P > 0.05; GPC3: 36% vs. 27%, P > 0.05; CDKN2A: 4% vs. 0%; TERT: 4% vs. 0%), while CCND2 methylation was not detected in any case. Although preliminary, we demonstrate that real‐time MSP can be applied to archival specimens and gene methylation profiling may have potential to be a useful ancillary tool to help distinguish MM from RM. Diagn. Cytopathol. 2007;35:498–502.

Thyroid fine‐needle aspiration (FNA) is a standard procedure for the clinical triage of thyroid nodules. The diagnosis of an adequately sampled thyroid FNA is generally grouped into three categories: benign, malignant, and indeterminate. The latter group usually includes follicular neoplasm, follicular lesion, and sometimes a more specific diagnosis such as Hurthle cell neoplasm or follicular lesion/neoplasm with Hurthle cell change. Whether a FNA diagnosis of Hurthle cell lesion/neoplasm (HLN) denotes a worse clinical outcome than follicular lesion/neoplasm (FLN) remains controversial. A cohort of 303 thyroid FNA cases with follow‐up thyroidectomy in our institutes was identified, with the follow‐up excision diagnosis compared to the FNA diagnosis in order to address this issue. Of this cohort, 87 cases had an FNA diagnosis of HLN while 216 cases had a diagnosis of FLN. Upon excision, the FNA diagnosis of HLN group had 14 cases of goiter/nodular hyperplasia (16%), 46 cases of adenoma (12 follicular adenoma (14%) and 34 cases of Hurthle cell adenoma (39%)), and 27 cases of carcinoma (31%, 12 papillary carcinoma and 15 Hurthle cell carcinoma). The FLN group had 74 cases of goiter/nodular hyperplasia (34.3%), 8 cases of Hashimoto thyroiditis (3.7%), 73 cases of follicular adenoma (33.8%), one case of granular cell tumor, and 60 cases of carcinoma (27.8%, 46 papillary carcinoma, 12 follicular carcinoma, and 1 Hurthle cell carcinoma and 1 parathyroid carcinoma) upon excision. There is no significant difference in predicting cancer between the two cytology diagnosis groups (HLN versus FLN, 31% versus 27.8%, P = 0.5771). When sorting all the cases by the surgical diagnosis, while comparable for age at diagnosis, the cancer group having the higher proportion of male patients than the non‐cancer group (28.7% versus 16.7%, P = 0.0259). Hurthle cell carcinoma patients are typically older than patients with other cancer diagnoses (59 versus 44, P = 0.0077). Our results suggest that an FNA diagnosis of HLN does not predict more malignancy than FLN. Males and older patients with a HLN FNA diagnosis carry a higher risk of Hurthle cell carcinoma upon thyroidectomy. Diagn. Cytopathol. 2006;34:330–334.