Neural Network Ensemble Based CAD System for Focal Liver Lesions from B-Mode Ultrasound

A neural network ensemble (NNE) based computer-aided diagnostic (CAD) system to assist radiologists in differential diagnosis between focal liver lesions (FLLs), including (1) typical and atypical cases of Cyst, hemangioma (HEM) and metastatic carcinoma (MET) lesions, (2) small and large hepatocellular carcinoma (HCC) lesions, along with (3) normal (NOR) liver tissue is proposed in the present work. Expert radiologists, visualize the textural characteristics of regions inside and outside the lesions to differentiate between different FLLs, accordingly texture features computed from inside lesion regions of interest (IROIs) and texture ratio features computed from IROIs and surrounding lesion regions of interests (SROIs) are taken as input. Principal component analysis (PCA) is used for reducing the dimensionality of the feature space before classifier design. The first step of classification module consists of a five class PCA-NN based primary classifier which yields probability outputs for five liver image classes. The second step of classification module consists of ten binary PCA-NN based secondary classifiers for NOR/Cyst, NOR/HEM, NOR/HCC, NOR/MET, Cyst/HEM, Cyst/HCC, Cyst/MET, HEM/HCC, HEM/MET and HCC/MET classes. The probability outputs of five class PCA-NN based primary classifier is used to determine the first two most probable classes for a test instance, based on which it is directed to the corresponding binary PCA-NN based secondary classifier for crisp classification between two classes. By including the second step of the classification module, classification accuracy increases from 88.7 % to 95 %. The promising results obtained by the proposed system indicate its usefulness to assist radiologists in differential diagnosis of FLLs.     

Characterization of Primary and Secondary Malignant Liver Lesions from B-Mode Ultrasound

Characterization of hepatocellular carcinomas (HCCs) and metastatic carcinomas (METs) from B-mode ultrasound presents a daunting challenge for radiologists due to their highly overlapping appearances. The differential diagnosis between HCCs and METs is often carried out by observing the texture of regions inside the lesion and the texture of background liver on which the lesion has evolved. The present study investigates the contribution made by texture patterns of regions inside and outside of the lesions for binary classification between HCC and MET lesions. The study is performed on 51 real ultrasound liver images with 54 malignant lesions, i.e., 27 images with 27 solitary HCCs (13 small HCCs and 14 large HCCs) and 24 images with 27 MET lesions (12 typical cases and 15 atypical cases). A total of 120 within-lesion regions of interest and 54 surrounding lesion regions of interest are cropped from 54 lesions. Subsequently, 112 texture features (56 texture features and 56 texture ratio features) are computed by statistical, spectral, and spatial filtering based texture features extraction methods. A two-step methodology is used for feature set optimization, i.e., feature pruning by removal of nondiscriminatory features followed by feature selection by genetic algorithm–support vector machine (SVM) approach. The SVM classifier is designed based on optimum features. The proposed computer-aided diagnostic system achieved the overall classification accuracy of 91.6 % with sensitivity of 90 % and 93.3 % for HCCs and METs, respectively. The promising results obtained by the proposed system indicate its usefulness to assist radiologists in diagnosing liver malignancies.     

SVM-Based CAC System for B-Mode Kidney Ultrasound Images

The present study proposes a computer-aided classification (CAC) system for three kidney classes, viz. normal, medical renal disease (MRD) and cyst using B-mode ultrasound images. Thirty-five B-mode kidney ultrasound images consisting of 11 normal images, 8 MRD images and 16 cyst images have been used. Regions of interest (ROIs) have been marked by the radiologist from the parenchyma region of the kidney in case of normal and MRD cases and from regions inside lesions for cyst cases. To evaluate the contribution of texture features extracted from de-speckled images for the classification task, original images have been pre-processed by eight de-speckling methods. Six categories of texture features are extracted. One-against-one multi-class support vector machine (SVM) classifier has been used for the present work. Based on overall classification accuracy (OCA), features from ROIs of original images are concatenated with the features from ROIs of pre-processed images. On the basis of OCA, few feature sets are considered for feature selection. Differential evolution feature selection (DEFS) has been used to select optimal features for the classification task. DEFS process is repeated 30 times to obtain 30 subsets. Run-length matrix features from ROIs of images pre-processed by Lee’s sigma concatenated with that of enhanced Lee method have resulted in an average accuracy (in %) and standard deviation of 86.3 ± 1.6. The results obtained in the study indicate that the performance of the proposed CAC system is promising, and it can be used by the radiologists in routine clinical practice for the classification of renal diseases.     

Accumulative copy number increase of MET drives tumor development and histological progression in a subset of ovarian clear-cell adenocarcinomas

Our previous study demonstrated that, among ovarian carcinomas, amplification of the MET gene and overexpression of MET specifically and commonly occur in clear-cell adenocarcinoma histology. This study was conducted to address how these alterations contribute to development and progression of this highly chemoresistant form of ovarian cancer. We histologically reviewed 21 previously described MET amplification-positive clear-cell adenocarcinoma cases, and selected 11 tumors with synchronous endometriosis and 2 tumors with adjacent clear-cell adenofibroma (CCAF) components. Using double in situ hybridization and immunohistochemistry, copy number alterations of the MET gene and levels of MET protein expression were analyzed in these putative precursor lesions and the corresponding invasive carcinoma components in this selected cohort. All of the non-atypical precursor lesions analyzed (ie, non-atypical endometrioses and the benign CCAFs) were negative for MET gain. However, low-level (≥3 MET copies in ≥10% and ≥4 MET copies in 10-40% of tumor cells) gain of MET was detected in 4 (40%) of the 10 atypical endometrioses and 1 of the 2 borderline CCAFs. Moreover, high-level (≥4 MET copies in ≥40% of tumor cells) gain of MET were detected in five (50%) of the atypical endometrioses. In 4 (31%) of the 13 cases enrolled, intratumoral heterogeneity for MET gain was documented in invasive carcinoma components, wherein all the relatively differentiated carcinoma components showed low-level gain of MET and all the corresponding poorly differentiated carcinomas showed high-level gain. The overall incidence of MET overexpression gradually increased from the precursors of non-atypical form (0%), through those of atypical form (67%) and the relatively differentiated carcinoma components (92%), to the poorly differentiated carcinoma components (100%). These results suggest that accumulative MET gene copy number alterations causing MET overexpression are associated with higher tumor grade and might drive the development and progression of the MET amplification-positive ovarian clear-cell adenocarcinoma.     

不典型部位软骨母细胞瘤

目的 探讨长骨骨骺和骨突以外不典型部位软骨母细胞瘤的临床、影像学和病理形态学特征,分析误诊原冈。方法 收集上海市5家大型综合性医院近12年内经病理证实的18例不典型部位软骨母细胞瘤的临床、X线和病理资料,作回顾性分析。肿瘤细胞的软骨分化倾向作S-100标记证实。结果18例中10例位于足部跗骨和小管状骨(55．6％),单骨发病最常见的是跟骨和距骨,平均发病年龄27．8岁,年龄＞25岁的成年人占55．6％。X线大多表现为膨胀性,多房性边界清楚的溶骨性骨质破坏。5例(28％)有局部骨皮质破坏,1例骨外软组织浸润,10例(55．6％)合并动脉瘤性骨囊肿或单纯性骨囊肿。术前临床无1例确诊,病理2例误诊,大多误诊为动脉瘤性骨囊肿或巨细胞瘤。结论 应综合考虑临床表现、X线特征以及病理形态改变进行不典型部位软骨母细胞瘤的诊断,避免误诊。     

Improvement in specificity of ultrasonography for diagnosis of breast tumors by means of artificial intelligence.

A set of ultrasonograms of lesions from 200 patients between the ages of 14 and 93 years who underwent mammography followed by ultrasonographic examination and excisional biopsy has been studied with computer vision techniques to improve the ultrasonographic specificity of the diagnosis. Selected features representing the texture of the lesion were calculated and then classified by an artificial neural network. This network was biased toward correctly classifying all the malignant cases at the expense of some misclassification of the benign cases. The network diagnosed the malignant cases with 100% sensitivity and 40% specificity (compared with 0% specificity for the radiologists diagnosing the same set of cases in the breast imaging setting), and tests performed with a leave-one-out technique indicate that the network will generalize well to new cases. This suggests that methods based on neural network classification of texture features show promise for potentially decreasing the number of unnecessary biopsies by a significant amount in patients with sonographically identifiable lesions.     

Cytodiagnosis of hepatocellular carcinoma in fine-needle aspirates of the liver: its differentiation from reactive hepatocytes and metastatic adenocarcinoma

Fine-needle aspiration (FNA) cytology is a useful tool for diagnosis of primary malignancies and metastatic lesions of the liver. However, well-differentiated hepatocellular carcinoma (HCC) may resemble benign/reactive hepatocytes, and less differentiated HCC may simulate poorly differentiated adenocarcinoma, leading to difficulties in interpretation of aspirates from liver. To determine the subtle cytomorphological features which can differentiate these lesions, ultrasound-guided FNA smears from 86 cases of liver malignancy were subjected to detailed cytologic assessment. These included 20 cases of HCC, 38 cases of metastatic adenocarcinoma, and 28 cases of benign/reactive hepatocytes. The important features for separating HCC or well-differentiated HCC from benign/reactive hepatocytes were excessive cellularity, trabecular pattern vs. thin cords of hepatocytes, nuclear pleomorphism, atypical stripped nuclei, and macronucleoli (P < 0.001 to < 0.0001). The most significant features for differentiating HCC from metastatic adenocarcinoma were trabecular growth pattern, hepatocytic cells vs. columnar/cuboidal cells, eosinophilic granular cytoplasm, lipid vacuoles, bile pigments, and atypical stripped nuclei (P < 0.001 to < 0.0001). The cytomorphological features which may distinguish poorly differentiated HCC from poorly differentiated adenocarcinoma were polygonal (hepatocytic) cells, eosinophilic granular cytoplasm and lipid vacuoles in HCC, and columnar/cuboidal cells and acinar/glandular formation in adenocarcinoma (P < 0.05 to < 0.001). Diagn. Cytopathol. 1999;21:370-377.     

Early stages of multistep hepatocarcinogenesis: adenomatous hyperplasia and early hepatocellular carcinoma.

From a series of 320 heptocellular carcinoma (HCC) cases treated surgically, we selected small nodular lesions that had not destroyed the preexisting liver structure grossly. After excluding metastases and large regenerative nodules, 58 lesions from 41 cases were chosen. All the lesions were hypercellular. Among them, 33 lesions showing histologic features of very well-differentiated HCC (Edmondson grade I), that is, small hepatocytes with little cellular atypia but with structural atypia, such as a thin trabecular structure of acinar formation in some areas, were classified as early HCC (eHCC). In seven eHCCs, areas of overt carcinoma, classified as Edmondson grade II, were found in the background of Edmondson grade I carcinoma. The remaining 25 lesions lacked structural atypia and were classified as adenomatous hyperplasia (AH). Among the AHs, 10 nodules with a very focal abnormal structure were subclassified as atypical adenomatous hyperplasia (AAH). There was a tendency for the size and cellularity of the atypical lesions to increase in order from AH to AAH to eHCC. All nodules larger than 1.5 cm were eHCC. A degree of cellularity more than twice that of a regenerative nodular was suggested to be an indicator of HCC. All small nodular lesions were associated with chronic liver disease. These histologic observations appear to explain the stepwise development of overt HCC from very well-differentiated eHCC, and of eHCC from AH probably through AAH, at least in cases of HCC associated with chronic liver disease.     

Correspondence in texture features between two mammographic views

It is well established that radiologists are better able to interpret mammograms when two mammographic views are available. Consequently, two mammographic projections are standard: mediolateral oblique (MLO) and craniocaudal (CC). Computer-aided diagnosis algorithms have been investigated for assisting in the detection and diagnosis of breast lesions in digitized/digital mammograms. A few previous studies suggest that computer-aided systems may also benefit from combining evidence from the two views. Intuitively, we expect that there would only be value in merging data from two views if they provide complementary information. A measure of the similarity of information is the correlation coefficient between corresponding features from the MLO and CC views. The purpose of this study was to investigate the correspondence in Haralick's texture features between the MLO and CC mammographic views of breast lesions. Features were ranked on the basis of correlation values and the two-view correlation of features for subgroups of data including masses versus calcification and benign versus malignant lesions were compared. All experiments were performed on a subset of mammography cases from the Digital Database for Screening Mammography (DDSM). It was observed that the texture features from the MLO and CC views were less strongly correlated for calcification lesions than for mass lesions. Similarly, texture features from the two views were less strongly correlated for benign lesions than for malignant lesions. These differences were statistically significant. The results suggest that the inclusion of texture features from multiple mammographic views in a CADx algorithm may impact the accuracy of diagnosis of calcification lesions and benign lesions.     

Cytologic criteria to distinguish hepatocellular carcinoma from nonneoplastic liver.

The authors reviewed a series of fine-needle aspiration biopsy (FNAB) specimens of the liver to identify useful cytologic criteria to distinguish hepatocellular carcinoma (HCC) from nonneoplastic liver. Ten cytologic features were examined in this study: high cellularity, acinar pattern, trabecular pattern, hyperchromasia, pleomorphism, irregularly granular chromatin, uniformly prominent nucleoli, multiple nucleoli, increased nuclear/cytoplasmic ratio, and atypical naked hepatocytic nuclei. These features were examined in a series of 82 FNAB specimens from 52 patients with HCC and 30 patients with nonneoplastic lesions. With the use of a step-wise logistic regression analysis, three features were identified as predictive of HCC: increased nuclear/cytoplasmic ratio (P = 0.001), trabecular pattern (P = 0.002), and atypical naked hepatocytic nuclei (P = 0.03). When these three criteria were used, the sensitivity of diagnosing HCC by FNAB was 100%, and the specificity was 87%.     

Clinicocytopathology of breast cancers with a ring-like appearance on ultrasonography and/or magnetic resonance imaging

On breast cancer imaging some cancers have an anechoic or high-echoic zone in the tumor on ultrasonography and ring-shaped enhancement on contrast-enhanced magnetic resonance imaging (MRI) with high intensity in the central area of the tumor on T2-weighted imaging, necessitating their differentiation from benign disease. Thus, nine breast cancers with a ring-like appearance on imaging were analyzed on cytopathology. Histologically the cancer cells of these lesions showing a ring-like appearance were located in the periphery of the tumor, with a central hypocellular zone. Five such lesions with a thick, doughnut-like appearance were identified as cancers with acellular zones (CAC), and four lesions with a thinner, rim-like appearance as matrix-producing carcinomas (MPC). The percentage ratio of the cancer-zone width to the tumor diameter was 26.4 ± 7.8 and 8.0 ± 3.2 (mean ± SD), respectively ( P = 0.003). Cytologically, highly atypical, naked-nucleus cells were observed in eight of the nine cancers. In two MPC and three CAC, cartilage matrix and amorphous material, respectively, were observed in the background. In summary, the present series of breast cancers having a ring appearance on imaging did not have uniform cytopathological features. They were classified as MPC or CAC, and cytology was useful in their diagnosis and differentiation in some cases.     

Automated diagnosis of coronary artery diseased patients by heart rate variability analysis using linear and non-linear methods

Abstract

Coronary artery disease (CAD) is a highly considered dangerous disease which may lead to myocardial infarction and even sudden cardiac death. The objective of this work is to evaluate the diagnostic performance features derived from linear and non-linear methods of Heart Rate Variability (HRV) analysis for classification software modules with Normal (NOR) subjects and CAD patients. The proposed methodology follows the recording of electrocardiogram from 60 NOR subjects and 64 CAD patients, RR interval tachogram generation, computing the features from time domain, frequency domain, non-linear methods and its analysis, feature dimension reduction by Principal Component Analysis (PCA) and classification by probabilistic neural network, K nearest neighbour and Support Vector Machine (SVM) classifiers. The results of the study indicate a clear difference in NOR subjects and CAD affected patients by using PCA-SVM classifier with an accuracy of 91.67%, sensitivity of 86.67% and 96.67% for NOR and CAD classes, respectively.     

Quantitative Ultrasound Analysis for Classification of BI-RADS Category 3 Breast Masses

The accuracy of an ultrasound (US) computer-aided diagnosis (CAD) system was evaluated for the classification of BI-RADS category 3, probably benign masses. The US database used in this study contained 69 breast masses (21 malignant and 48 benign masses) that at blinded retrospective interpretation were assigned to BI-RADS category 3 by at least one of five radiologists. For computer-aided analysis, multiple morphology (shape, orientation, margin, lesions boundary, and posterior acoustic features) and texture (echo patterns) features based on BI-RADS lexicon were implemented, and the binary logistic regression model was used for classification. The receiver operating characteristic curve analysis was used for statistical analysis. The area under the curve (Az) of morphology, texture, and combined features were 0.90, 0.75, and 0.95, respectively. The combined features achieved the best performance and were significantly better than using texture features only (0.95 vs. 0.75, p value?=?0.0163). The cut-off point at the sensitivity of 86 % (18/21), 95 % (20/21), and 100 % (21/21) achieved the specificity of 90 % (43/48), 73 % (35/48), and 33 % (16/48), respectively. In conclusion, the proposed CAD system has the potential to be used in upgrading malignant masses misclassified as BI-RADS category 3 on US by the radiologists.     

Liver lesions in children and adolescents: cytopathologic analysis and clinical correlates in 44 cases

The role of fine needle aspiration (FNA) as a first-line diagnostic modality is well-established for neoplastic and nonneoplastic liver masses in adults. However, cytopathologic analysis of such lesions in children and adolescents has not been well studied. An 18-year retrospective review of the cytopathology archives at The Johns Hopkins Hospital identified 44 cases of hepatic FNA in children and adolescents (i.e., 21 years and younger). The cytopathologic material was reviewed and correlated with subsequent surgical pathology diagnoses and/or clinical follow-up. Among the 44 FNAs studied, 40 (90.9%) were deemed diagnostic for evaluation and 4 (9.1%) were nondiagnostic. Twenty four (60%) of the diagnostic FNAs were neoplasms, 10 (25%) were nonneoplastic lesions, and 6 (15%) were "atypical" or "suspicious." There were 21 (87.5%) malignant and 3 (12.5%) benign tumors. Among the malignant cases, metastases/secondary tumors accounted for 12 (54.5%) followed by hepatocellular carcinoma (HCC)--7 (38.1%). The types of metastatic/secondary neoplasms were Wilms tumor (2), germ cell tumor (2), acute leukemias (2), and miscellaneous others (6). Among the nonneoplastic lesions focal nodular hyperplasia was most common--4 (40%). Among the "atypical" cases, 4 (66.6%) turned out to be benign on follow-up, 2 were HCC, and 1 was Hodgkin lymphoma. Immunoperoxidase (IPOX) studies were done in 20 (45.5%) cases to confirm the cytopathologic diagnoses. Overall, FNA showed a sensitivity of 95.2% and a specificity of 100% for a malignant diagnosis.     

不典型肝细胞癌的CT表现与分析

目的探讨原发性肝细胞癌(HCC)在螺旋CT扫描中的不典型表现及其产生原因。方法选择17例HCC患者,其中男性14例,女性3例;年龄29～71岁,平均年龄52岁。分析经手术病理组织证实的螺旋CT扫描中呈不典型表现的CT征像和病理学改变。结果17例共发现18个病灶为不典型表现,其中10个病灶在动脉期强化明显,门静脉期和延迟期为等密度或高密度;6个病灶在动脉期、门静脉期和延迟期均呈低密度;2个病灶动脉期病灶边缘有环形强化,门静脉期和延迟期仍有持续强化。结论肝癌在螺旋CT 3期增强扫描中有多种不典型表现,病灶的血液供应情况、病理学基础等是其常见原因。     

On the utility of fine-needle aspiration in the diagnosis of primary scalp lesions

A study of the clinico-cytologic findings of 62 primary palpable lesions located in the scalp is reported. Trichilemmal cyst (16 cases) followed by lipoma (8 cases) and benign melanocytic proliferation (4 cases) were the most frequent benign conditions (n = 45); basal-cell carcinoma (6 cases) and squamous-cell carcinoma (5 cases) were the most frequently aspirated malignant tumors (n = 17). In the cytologic category of benign lesions (n = 45), cytohistologic correlation was carried out in 19 cases and the cytohistologic agreement was 100%, except for the three actinic keratosis cases that were diagnosed as atypical epidermal lesions. Cytohistologic agreement was 100% in the 16 malignant lesions in which excisional biopsy was performed. Based on the findings reported here, fine-needle aspiration is indicated as the first-choice technique for the clinical evaluation of primary scalp lesions. Experience with cutaneous cytopathology is essential in order to be able to carry out this task.     

Argyrophilic nucleolar organizer regions and alpha-fetoprotein in adenomatous hyperplasia in human cirrhotic livers.

Recently, adenomatous hyperplasia (AH) of the liver has been suspected as a precancerous lesion in human hepatocarcinogenesis. The authors examined 75 cases of AH from 42 cirrhotic livers, using staining of argyrophilic nucleolar organizer regions (AgNORs). These reflect proliferative cell activity. Findings in AH were compared with those seen in hepatocellular carcinoma (HCC) and other chronic liver diseases. Expression of alpha-fetoprotein (AFP) was also examined immunohistochemically. The authors classified AH into three types: ordinary (OAH), atypical (AAH), and AH with focal malignancy (FM). OAH implies a lack of atypia; AAH represents AH with structural and cellular atypia but without the features of overt carcinoma; and FM denotes AH with foci of overt HCC. Forty of the 75 cases of AH were categorized as OAH, 19 as AAH, and 16 as FM. The noncancerous areas of FM had features of AAH. The mean number of AgNORs in AH was intermediate between that seen in cirrhosis (2.93) and HCC (6.18) and showed a step-wise increase in the following order: OAH (2.95), AAH (3.89), noncancerous areas in FM (4.58), and malignant foci in FM (5.71). There was no significant difference in AgNOR counts between OAH and cirrhosis. AgNOR counts in AAH and FM were significantly higher than those of OAH, and lower than those of HCC. AFP was positive in 12 of 25 HCCs and in malignant foci of 3 FM lesions, but it was absent in OAH and AAH. These data suggest that OAH has a limited capacity for proliferation but that AAH and FM are much more replicative. The latter two conditions are probably preneoplastic lesions or early forms of HCC.     

p53 immunolocalization in cell block preparations of squamous lesions of the neck: an adjunct to fine-needle aspiration diagnosis of malignancy

OBJECTIVE: Mutations of the p53 tumor suppressor gene, with consequent nuclear p53 protein accumulation, are among the most common genetic abnormalities in human cancers. The purpose of this study was to determine the utility of p53 immunostaining as an adjunct to the diagnosis of malignancy in fine-needle aspirations of squamous lesions of the neck. MATERIALS AND METHODS: Using a monoclonal antibody to the p53 protein and a standard avidin-biotin complex technique, immunostaining was performed on paraffin-embedded cell blocks of 20 cases with the following cytologic diagnoses: (1) metastatic squamous cell carcinoma (SCC) (7 cases); (2) atypical squamous cells, SCC cannot be excluded (7 cases); and (3) cytologic findings consistent with branchial cleft cyst (6 cases). Tissue or clinical follow-up was available in all cases. RESULTS: Five (71%) of 7 cases with an unequivocal cytologic diagnosis of metastatic SCC were positive for p53 protein. Tissue follow-up confirmed metastatic SCC in all of these 7 cases. Of the 7 cases with cytologic diagnosis of atypical squamous cells, 2 were negative and 5 (71%) were positive for p53 protein. Subsequent excisional biopsies in these cases revealed metastatic SCC (6 cases) and branchial cleft cyst (1 case). The squamous cells in all 5 cases with cytologic findings consistent with branchial cleft cyst were negative for p53 protein; tissue follow-up confirmed the diagnoses of branchial cleft cyst in 4 cases. In the remaining 2 cases excision was not performed, as the cystic lesion was completely decompressed and, clinically, no recurrences were identified at 14 and 8 months of follow-up. CONCLUSIONS: Our findings suggest that p53 immunostaining is helpful in differentiating benign and malignant squamous lesions. While negative staining for p53 does not exclude malignancy, positive immunostaining may aid in accurate fine-needle aspiration diagnosis of malignancy in cytomorphologically equivocal squamous lesions of the neck.     

Fine-needle aspiration of cystic lesions of the kidney. Morphologic spectrum and diagnostic problems in 41 cases

Although imaging studies show the nature of most cystic lesions of the kidney (RCs), many RCs require fine-needle aspiration (FNA) for accurate diagnosis. Interpretation of the FNAs remains challenging. The FNA specimens of 41 RCs were reviewed and correlated with imaging studies. Final diagnoses for 30 cytologically benign lesions were simple cyst (28), acquired cystic kidney (1), and cystic renal carcinoma (1). The fluid from the benign cysts displayed macrophages, epithelial cells from the cyst lining, tubular cells, neutrophils, and Liesegang rings. Fluid from the acquired cystic kidney and the cystic renal cell carcinoma showed features similar to those of the benign cysts. The 9 cases with "suspicious" cytology included 5 complex cystic lesions displaying rare but atypical epithelial cell clusters, 3 low-grade renal cell carcinomas with many mildly atypical papillary clusters of epithelial cells, and 1 simple benign cyst with many tubular cells. The 2 cytologically malignant lesions were cystic renal cell carcinomas with abundant tumor cells with partially clear cytoplasm and atypical nuclei admixed with abundant macrophages and lymphocytes; 1 case developed in a kidney with acquired cystic disease. Simple cysts remain the most frequently aspirated RCs, but complex cystic lesions are increasingly recognized. Since many RCs are composed of independent loculi, a nonrepresentative sample is a potential problem, and cytologic-radiologic correlation becomes mandatory. The "suspicious" patterns identified in this study should serve as diagnostic guidelines and set the foundation for future validation.