Fine needle aspiration biopsy diagnosis of mucoepidermoid carcinoma. Statistical analysis

Fine needle aspiration (FNA) biopsy is an increasingly popular method for the evaluation of salivary gland tumors. Of the common salivary gland tumors, mucoepidermoid carcinoma is probably the most difficult to diagnose accurately by this means. A series of 96 FNA biopsy specimens of salivary gland masses, including 34 mucoepidermoid carcinomas, 51 other benign and malignant neoplasms, 7 nonneoplastic lesions and 4 normal salivary glands, were analyzed in order to identify the most useful criteria for diagnosing mucoepidermoid carcinoma. Thirteen cytologic criteria were evaluated in the FNA specimens, and a stepwise logistic regression analysis was performed. The three cytologic features selected as most predictive of mucoepidermoid carcinoma were intermediate cells, squamous cells and overlapping epithelial groups. Using these three features together, the sensitivity and specificity of accurately diagnosing mucoepidermoid carcinoma were 97% and 100%, respectively.     

Evaluation of cystic salivary gland lesions by fine needle aspiration: an analysis of 21 cases

OBJECTIVE: To analyze the potential sources of diagnostic errors and overall accuracy rate of the fine needle aspiration biopsy (FNAB) diagnosis of cystic salivary gland neoplasms. STUDY DESIGN: A 10-year (1993-2002) retrospective review of the cytopathology slides from the Division of Cytopathology, Department of Pathology, Long Island Jewish Medical Center, New Hyde Park, New York, identified a total of 97 consecutive salivary gland FNAB cases that microscopically were interpreted as representing cystic lesions. Of these, 21 cases had histologic follow-up at our institution. RESULTS: A correct diagnosis was rendered by FNAB in 15 of 21 (72%) cases. This included 9 Warthin's tumors, 2 mucoepidermoid carcinomas, 2 simple cysts, 1 cystadenoma and 1 abscess. Clinically insignificant discrepancies were identified in 3 of 21 (14%) FNABs. Clinically significant misdiagnoses were identified in a further 3 of 21 (14%) cases. CONCLUSION: A systematic approach to the diagnosis of cystic salivary gland lesions by FNAB can result in a correct diagnosis in > 70% of cases. Careful attention should be directed at identifying the extracellular fluid components present (mucoid vs. watery proteinaceous) as well as the predominant cellular component (e.g., lymphocytes, histiocytes, epithelial cells and oncocytes). It is important to recognize, however, that occasionally epithelial cells may not be detected on FNAB of cystic salivary gland lesions, as a result of either cellular dilution by cystfluid or inadequate sampling. Regardless, with all FNABs tentatively diagnosed as a mucinous cystic lesion, the referring clinician should be informed that a low grade mucoepidermoid carcinoma cannot be ruled out.     

Intranuclear inclusions in fine needle aspirates of bronchial low grade mucoepidermoid carcinoma with clear cell change: a report of two cases

BACKGROUND: Mucoepidermoid carcinoma of the bronchus is a rare neoplasm that can be recognized on histology as well as cytology by the presence of three characteristic cell types: mucus secreting, epidermoid and intermediate. We encountered two cases displaying unusual cytologic features, including clear intranuclear inclusions. CASES: Two females, aged 33 and 39, presented with an intrabronchial tumor and pulmonary parenchymatous mass, respectively. Fine needle aspiration of both tumors showed similar cytologic features, with a dominant population of cells with bland nuclei and wide cytoplasm, and frequent intranuclear inclusions. A minor component of mucus-secreting cells was also recognized. Histologically, both tumors corresponded to the clear cell variant of mucoepidermoid carcinoma. CONCLUSION: The cytologic picture in our cases has not been described previously in fine needle aspirates of mucoepidermoid carcinoma, in neither the bronchus nor salivary gland. The differential diagnosis of a monotonous population of epithelial cells with intranuclear inclusions involves bronchioloalveolar carcinoma, but the absence of the characteristic sheet pattern, as well as the clinical and image findings, excludes this possibility. The lack of atypia and intrabronchial location limits the scope to carcinoid and salivary gland-type tumors of the bronchus. Since we were aware of the possibility of unusual cytologic presentations of mucoepidermoid carcinomas, search for different cellular populations suggested the precise diagnosis.     

Cytomorphologic spectrum of carcinoma ex pleomorphic adenoma

OBJECTIVE: To delineate the cytomorphologic features of carcinoma ex pleomorphic adenoma (CPA) and identify the diagnostic pitfalls. STUDY DESIGN: Smears of 14 cases suspected as CPA on fine needle aspiration over a period of 15 years were reviewed. Cytohistologic correlation was done in 10 cases. RESULTS: All cases had a salivary gland mass of 1-16 years' duration, with a rapid increase in size in 10 cases. Epithelial cells predominated over stroma in 11 of 14 cases. Group I showed unequivocal malignant cells admixed with benign epithelial and stromal components of pleomorphic adenoma (PA), which were considered diagnostic of CPA on review. The cytologic differential diagnosis in these cases included mucoepidermoid carcinoma, carcinosarcoma and metastatic adenocarcinoma. Group II comprised 7 cases suspected to be cellular PA with atypia or CPA. These showed mild to moderate degrees of pleomorphism, absence of unequivocal malignant cells, and a variable proportion of benign epithelial and stromal components. Four of them were histologically confirmed as CPA. CONCLUSION: Sampling error is an important cause of diagnostic pitfalls. Correlation with clinical data is essential in diagnosis of CPA on cytology. In a proper clinical setting, extensive fine needle aspiration sampling should be done initially. Any degree of nuclear atypia in PA should be documented, alerting the clinician and histopathologist to the possibility of CPA.     

Comparison of ThinPrep and conventional preparations on fine needle aspiration cytology material

OBJECTIVE: To compare the various cytologic features on ThinPrep 2000 (TP) (Cytyc Corporation, Marlborough, Massachusetts, U.S.A.) and conventional preparation (CP) specimens from fine needle aspiration cytology (FNAC) material by a semiquantitative scoring system. STUDY DESIGN: In this prospective study a total of 71 consecutive cases were included. In each case, two passes were performed. The first pass was used for conventional preparations, with direct smears made and fixed immediately in 95% alcohol for Papanicolaou stain. For TP preparation a second pass produced material for processing in the ThinPrep 2000. The TP and CP slides were studied independently by two observers and representative slides of CP and TP compared for cellularity, background blood and necrotic cell debris, cell architecture, informative background, presence of monolayer cells, and nuclear and cytoplasmic details by a semiquantitative scoring system. Statistical analysis was performed by Wilcoxon's signed rank test on an SPSS program (Chicago, Illinois, U.S.A.). RESULTS: TP preparations contained adequate diagnostic cells in all cases and were tangibly superior to CP preparations concerning monolayer cells, absence of blood and necrosis, and preservation of nuclear and cytoplasmic detail (statistically significant, Wilcoxon's signed rank test, P < .000). CONCLUSION: TP preparations are superior to conventional preparations with regard to clear background, monolayer cell preparation and cell preservation. It is easier and less time consuming to screen and interpret TP preparations because the cells are limited to smaller areas on clear backgrounds, with excellent cellular preservation. However, TP preparations are more expensive than CP and require some experience for interpretation.     

Endoscopic ultrasound-guided fine needle aspiration of the pancreas. Diagnostic utility and accuracy

OBJECTIVE: Endoscopic ultrasound-guided fine needle aspiration biopsy (EUS-FNAB) is regarded as a safe and reliable procedure for diagnosing and staging of pancreatic neoplasms. This study retrospectively evaluated both the diagnostic utility and accuracy of pancreatic EUS-FNABs and potential cytologic pitfalls when using Diff-Quik stain for on-site evaluation. STUDY DESIGN: Pancreatic EUS-FNABs performed between 1995 and 1998 were identified from the files of the Department of Pathology. All patients were studied via a linear-array ultrasound endoscope with an FNAB device. Immediate evaluation of the specimen by a pathologist using air-dried slides and Diff-Quik stain was done on all cases. An average of five passes (range, three to nine) were performed. Five cytologic categories were identified: nondiagnostic, benign, atypical, suspicious and malignant. EUS disease staging, histologic correlation and clinical follow-up were reviewed. RESULTS: Sixty-nine consecutive pancreative FNABs were evaluated in the study period. The patients comprised 38 females and 31 males with a mean age of 65 years (range, 36-83). Histologic correlation was available on 40 patients, and follow-up was available on the remaining 29. The cytologic diagnoses included: 31 malignant, 8 suspicious, 6 atypical, 20 benign and 4 nondiagnostic. Forty-three cases were true positive, 9 were true negative, 2 were false positive, and 11 were false negative. The overall sensitivity was 80% and specificity was 82%. CONCLUSION: The study showed that cytologic evaluation of pancreatic EUS-FNABs has 80% sensitivity and 82% specificity. False negative diagnosis was usually due to sampling error. A nondiagnostic cytologic diagnosis should be rendered in the absence of adequate sampling of a lesion. On-site cytologic evaluation of EUS-FNABs aids in guaranteeing specimen adequacy, and the pathologist should be trained to evaluate Diff-Quik-stained samples.     

Diagnostic efficacy of immunocytochemistry on fine needle aspiration biopsies processed by thin-layer cytology

OBJECTIVE: To evaluate the efficacy of immunocytochemistry (ICC) performed on smears processed by thin-layer cytology (TLC). STUDY DESIGN: During the period January 2001-September 2003, 3,573 consecutive fine needle aspiration biopsies were processed with both conventional smears (CSs) and TLC diagnosed by a single pathologist; 113 required immunocytochemical study. CSs were fixed in ethanol whereas TLC slides were processed with the ThinPrep 2000 method (Cytyc Co., Marlborough, Massachusetts, U.S.A); both were stained with Papanicolaou stain. ICC staining was carried out on only TLC slides. RESULTS: The 113 cytologic cases were grouped as follows: 32 thyroid nodules with 16 histologic controls (HCs), 24 lymph nodes (regardless of location) with 15 HCs, 18 liver and pancreatic lesions (3 HCs), 11 lung nodules (6 HCs), 5 kidney and adrenal gland lesions (1 HC), 6 abdominal (2 HCs) and 4 mediastinal masses (1 HC), 6 salivary gland tumors (3 HCs), 4 bone masses (2 HCs) and 3 subcutaneous lesions (1 HC). ICC contributed to the diagnosis in 104 cases (92%), whereas it was inconclusive in 9. The cytologic diagnoses were histologically confirmed in 46 of 50 cases (92%). CONCLUSION: ICC can be successfully applied on TLC slides with better results than on CSs, and its yield can be useful in making the correct diagnosis on fine needle aspiration biopsy.     

Transvaginal fine needle aspiration biopsy.

OBJECTIVE: To assess the role of transvaginal fine needle aspiration biopsy (FNAB) in the evaluation of palpable gynecologic masses. STUDY DESIGN: Transvaginal FNABs from 1994 to 1999 were identified from the files of Barnes-Jewish Hospital. Histologic correlation was obtained using the Pathology Department's computer database. Two pathologists reviewed the pathologic samples. Pertinent clinical information was obtained by reviewing the medical records. RESULTS: Twenty-two transvaginal FNABs from 22 patients were studied. The patients' mean age was 59 years (range, 29-84). Most patients (77%) had a previous history of a gynecologic malignancy, and 73% had a previous total abdominal hysterectomy and bilateral salpingo-oophorectomy. The size of the lesion sampled was provided in 15 cases and ranged from <1 to 5.4 cm in diameter. The location of the mass was reported as follows: vaginal (10 cases), vaginal cuff (5), rectovaginal septum (2), cul-de-sac (1), fornix (1), vaginal apex (1), right side of pelvis (1), and not specified (1). The cytologic diagnoses were: negative for malignancy (10 cases), positive for malignancy (9) and unsatisfactory (3). Most cases (77%) had histologic correlation or clinical follow-up. There was one false negative and no false positive cytologic diagnosis. CONCLUSION: Cytologic interpretation of transvaginal FNAB is an effective toolfor the evaluation of palpable pelvic and vaginal masses. Its specificity and sensitivity are 100% and 88%, respectively.     

Gastrin--a normal and pathologic regulator of gastric function.

Fine needle aspiration biopsy (FNAB) is an underused diagnostic procedure in children, particularly in the evaluation of superficial masses. A total of 54 FNABs of superficial masses were performed in children aged 1 month to 15 years. Adequate material for diagnosis was obtained in 50 attempts. The cytologic diagnosis increased clinical understanding and provided a guide for treatment in 46 of the 50 cases. The cytologic diagnosis was confirmed in 15 of 19 patients who underwent an operation. Surgical intervention was obviated in 31 patients. There was one false-positive diagnosis of cancer. We describe the role of FNAB in children and its technique, accuracy, and diagnostic problems.     

Liquid-based vs. conventional smears in fine needle aspiration of bone and soft tissue tumors

OBJECTIVE: To compare the accuracy of fine needle aspiration cytology of bone and soft tissue tumors utilizing ThinPrep (TP) (Cytyc Corporation, Boxborough, Massachusetts, U.S.A.) vs. conventional smears (CS). STUDY DESIGN: Fine needle aspiration cytology from bone and soft tissue tumors was processed and assessed for cellularity, nuclear and cytoplasmic preservation, cellular architecture and stromal background with both the TP liquid-based smear technique and conventional methods. RESULTS: An accurate diagnosis was made in 13% of TP cases as compared to 64% in CS cases. CONCLUSION: CS of fine needle aspiration sample is far superior to TP in diagnosing tumors of bone and soft tissues. Preservation of cytoplasmic features and cellular architecture was superior in conventionally prepared smears.     

Cytodiagnosis of 112 salivary gland lesions. Correlation with histologic and frozen section diagnosis.

Aspirates of 112 cases of salivary gland lesions with histologic correlation were reviewed. Fifty-five cases (49%) had frozen sections made. The 112 cases included 76 cases of benign lesions (31 cases of pleomorphic adenoma, 19 of Warthin's tumor and 26 of nonneoplastic lesions), 22 of primary salivary gland malignancy and 14 of metastatic malignant lesions. The overall accuracy in diagnosing benign and malignant lesions was 95%. The accuracy in diagnosing the exact category of neoplastic lesions was 70%. The diagnostic sensitivity for malignant lesions was 86% and the specificity, 99%. There was one false positive, in which a pleomorphic adenoma was diagnosed as small cell carcinoma. Five false-negative cases were encountered that were due to underdiagnosis of mucoepidermoid carcinoma and adenoid cystic carcinoma. The smears were reviewed, and the diagnostic pitfalls are discussed. A comparison of the cytodiagnosis and frozen section diagnosis was made. In frozen sections there were two false negatives, and two diagnoses were deferred. The overall diagnostic accuracy was 91%. The accuracy in diagnosing the exact category of neoplastic disease was 77%. The diagnostic sensitivity for malignant disease was 70% and specificity, 100%. Frozen section, however, did supplement the fine needle aspiration diagnosis in 13 cases.     

Fine needle aspiration biopsy of Hashimoto's thyroiditis. Sources of diagnostic error.

OBJECTIVE: To determine the accuracy of cytologic interpretation in the diagnosis of Hashimoto's thyroiditis (HT). STUDY DESIGN: At Ottawa Hospital from 1987 to 1994, 1,638 fine needle aspiration biopsies (FNABs) from thyroid were performed. HT was suggested in 184 FNAB samples taken from 157 patients. Of the 184 aspirates diagnosed with HT, 39 had corresponding surgical specimens taken from 31 patients. A retrospective review of these FNABs and surgical pathology slides formed the basis of this study. RESULTS: In 27 (69%) aspirates, HT was diagnosed on both the FNAB and surgical specimens. In 10 of 27 FNABs an associated lesion was not sampled by FNAB. In four of these 10 aspirates some of the cellular features of HT were misinterpreted, and the possibility of an associated neoplasm could not be ruled out. This resulted in four false positive diagnoses. In 12 (31%) FNABs from nine patients, the cytologic diagnosis of HT was not confirmed histologically. These cases included five Hürthle cell adenomas and one case each of follicular adenoma, nodular goiter, macrofollicular adenoma and malignant lymphoma. This resulted in five false negative diagnoses. CONCLUSION: These results support the value of FNAB in the diagnosis of HT. The presence of hyperplastic follicular cells on FNAB samples from HT may mimic a follicular neoplasm and result in a false positive interpretation. Adequate sampling of the thyroid is important, particularly when there is an associated lesion. The diagnosis of lymphocytic thyroiditis should not be made when only a few lymphocytes are present. Finally, pleomorphic Hürthle cells may be present in aspirates from Hürthle cell neoplasms and underdiagnosed as HT, especially when they are associated with a few lymphocytes.     

Pulmonary fine needle aspiration biopsy. Assessing the negative diagnosis.

OBJECTIVE: To assess the significance of the "negative for malignancy" category when applied to pulmonary transthoracic fine needle aspiration biopsy (FNAB). STUDY DESIGN: Transthoracic lung FNABs diagnosed as "negative for malignancy" were identified from the files of Barnes-Jewish Hospital's South and North Campus over a period of five and nine years, respectively. Histologic correlation and clinical follow-up were obtained. RESULTS: Of the 1,181 lung FNABs performed during the study period, 108 cases (9%) had a negative cytologic diagnosis. Histologic correlation was available in 46 cases (43%), of which 23 cases had benign histologic findings, and 19 cases were malignant. Thirty-five of the 62 cases without histologic correlation had clinical follow-up consistent with a benign process. CONCLUSION: Based on the histologic correlation and clinical data available, the negative predictive value was 77%. Inadequate sampling was responsible for all false negative cytologic diagnoses in this series.     

Value of the filter imprint technique in aspiration biopsy cytology

In 468 fine needle aspiration (FNA) biopsies, after smearing of the aspirate on glass slides, additional material was obtained by flushing the needle with a fixative. The cells were collected on Millipore filters, from which imprints were prepared. The filter imprints were found to be sufficiently cellular in 60% of the cases, which reduced by 33 (7.3%) the number of cases with unsatisfactory aspirates. Diagnostic specificity and sensitivity were not influenced by the technique. The filter imprint technique was of particular value in breast aspirates.     

Focal adenomatoid hyperplasia of salivary glands. A differential diagnostic problem in fine needle aspiration biopsy

Adenomatoid hyperplasia of minor salivary glands is an idiopathic lesion usually involving the palate and mimicking a primary salivary gland neoplasm. In particular, the lesions may resemble low-grade mucoepidermoid tumors clinically and by fine needle aspiration (FNA) cytology. An illustrative case is presented, and the nature of this unusual entity is reviewed. Its importance in the differential diagnosis of FNA biopsies of minor salivary gland lesions is also discussed.     

Computed tomography-guided fine needle aspiration biopsy of deep-seated lesions. A four-year experience

With the increased sophistication of radiologic imaging techniques, the sensitivity of detecting nonpalpable, deep-seated lesions has greatly improved. Coupling these techniques with fine needle aspiration biopsy (FNAB) provides a cost-effective, minimum-risk, highly sensitive and specific method of diagnostically evaluating the lesions. Over a four-year period (1985-1989) a total of 2,229 FNABs were performed at Loyola University Medical Center, Chicago; 539 of these biopsies were computed tomographically guided. The geographic computed tomographic biopsy sites were: thorax, 267 (49.54%); abdomen, 175 (32.47%); and retroperitoneum, 97 (18%). Four hundred eighty FNAB cases were diagnostic, with subsequent histologic follow-up in 284 (58%) cases. Fifty-nine (10.9%) FNABs were unsatisfactory; of them, 31 had subsequent diagnostic histology, and 9 remained unsatisfactory. A good correlation between FNAB and histology was observed, with an overall sensitivity of 93.2%, specificity of 98.8%, false-negative rate of 6.8% and false-positive rate of 1.2%. Diagnostic pitfalls and biopsy adequacy in computed tomographically guided FNABs are discussed.     

Expression of cytokeratin 19 and protein p63 in fine needle aspiration biopsy of papillary thyroid carcinoma

OBJECTIVE: To analyze the immunocytochemical distribution of CK19 and p63 on archival cytologic smears of 27 papillary thyroid carcinomas (PTCs), 22 benign thyroid lesions and 5 malignant non-PTC lesions. STUDY DESIGN: Archival cytologic smears of 27 papillary carcinomas, 22 benign thyroid lesions and 5 malignant nonpapillary carcinomas were processed for immunocytochemical detection of CK19 and p63, and results were compared. RESULTS: CK19 showed strong cytoplasmic staining in 22/27 fine needle aspiration biopsies (FNABs) of PTCs, in 5 benign lesions and in 4 malignant lesions of the control group. p63 Positivity was present in FNABs of 20/27 PTC and in 1 FNAB of nodular hyperplasia. Eighteen FNABs of PTC (66.6%) showed both strong CK19 staining and p63-positive cells, whereas none of the control cases showed coexpression of CK19 and p63. CONCLUSION: Coexistence of strong CK19 positivity and p63-positive cells can enhance the cytologic diagnosis of PTC.     

Assessment of thin-layer breast aspirates for immunocytochemical evaluation of HER2 status

OBJECTIVE: To determine whether immunocytochemistry (ICC) for HER2 on ThinPrep (TP)-processed breast fine needle aspiration biopsies (Cytyc Corp., Boxborough, Massachusetts, U.S.A.) is comparable to the findings of immunohistochemistry on corresponding surgically removed tissue. STUDY DESIGN: Immunostaining was performed on 63 malignant breast fine needle aspirates and compared to immunostaining on paraffin sections (PSs) from the subsequent biopsies. The HercepTest (Dako, Carpinteria, California, U.S.A.) and TAB250 antibodies were utilized. Cases in which the TP and paraffin HER2 results did not correlate were further assessed for gene amplification by differential polymerase chain reaction (dPCR). RESULTS: HER2 overexpression was found in 9 of the 63 cases (14%). TAB250 had higher specificity on PS versus TP (P = .008), and TAB250 had higher specificity on PS versus the HercepTest on PS and TP (P = .004 and .0001, respectively). CONCLUSION: HER2 immunostaining with both the HercepTest and TAB250 on TP is unreliable due to low specificity (72% and 83% for HercepTest and TAB250, respectively). However, both antibodies have high sensitivity (89% and 100%, respectively); suggesting that this method may have some utility as a preliminary screening test for HER2 status. Negative HER2 staining by ICC is highly predictive of the absence of HER2 overexpression, whereas positive HER2 staining on TP would require further validation by either dPCR of fluorescence in situ hybridization.     

Interobserver variability in the fine needle aspiration biopsy diagnosis of follicular lesions of the thyroid gland

OBJECTIVE: To study the degree of interobserver variability in the interpretation of fine needle aspiration (FNA) biopsies of the thyroid, specifically in the categorization of follicular lesions (FLs), and to examine the accuracy of FNA diagnosis of FLs with surgical follow-up. STUDY DESIGN: Fifty cases were chosen with surgical follow-up and a cytologic diagnosis of either FL (21) or follicular neoplasms (29). Representative slides were selected for each case and circulated to 4 pathologists for review. Interobserver variability was assessed using pairwise K statistics. Accuracy of the cytologic diagnoses in predicting a nonneoplastic or neoplastic outcome was determined by measuring sensitivity and specificity. Likelihood ratios and receiver operator characteristic curves were calculated for each reviewer. RESULTS: Interobserver agreement between the 4 pathologists was fair to substantial (K scores, 0.199-0.617). The accuracy of the 4 pathologists' cytologic diagnoses in predicting the surgical outcome was 77-90% for follicular neoplasms and 53-74% for nonneoplastic diagnoses. CONCLUSION: FLs present diagnostic difficulties as to cytologic categorization. A wide range of interobserver agreement was found in this study of 4 pathologists from the same institution. Some pathologists make greater use of intermediate categories, such as FL, favor nonneoplastic, or FL, favor neoplastic, whereas others show more definitive categorization into benign and neoplastic groups.     

Accuracy of cytology vs. microbiopsy for the diagnosis of well-differentiated hepatocellular carcinoma and macroregenerative nodule. Definition of standardized criteria from a study of 100 cases

OBJECTIVE: To determine the accuracy of ultrasound (US)-guided fine needle aspiration (FNA) for the diagnosis of well-differentiated hepatocellular carcinoma (wd HCC) and macroregenerative nodule (MRN) and to identify the most useful cytologic and histologic criteria to distinguish between those two diagnoses. STUDY DESIGN: Cytologic and histologic specimens of 50 wd HCC and 50 MRN were reviewed blindly and the diagnosis compared to the final clinical diagnosis. Twenty-eight cytologic and 25 histologic criteria were examined and subjected to statistical analysis. RESULTS: Among 100 cases studied, the final diagnosis was available for 43. In those 43 cases, combining analysis of cytologic and histologic specimens, the sensitivity of US-guided FNA was of 75% and the specificity 100%. Cytologic analysis was better than isolated histologic analysis, with a sensitivity of 75% vs. 68%, respectively. Sensitivity of cytologic diagnosis was lower for smaller nodules and for those located in poorly accessible hepatic segments. With the use of stepwise logistic regression analysis, four cytologic features (increased nuclear/cytoplasmic ratio, cellular monomorphism, nuclear crowding, loss of bile duct cells) and four histologic features (increased nuclear/cytoplasmic ratio, decreased Kupffer cells, cellular monomorphism, increased trabeculae thickness) were identified as predictive of HCC.