A review on ultrasound-based thyroid cancer tissue characterization and automated classification.

In this paper, we review the different studies that developed Computer Aided Diagnostic (CAD) for automated classification of thyroid cancer into benign and malignant types. Specifically, we discuss the different types of features that are used to study and analyze the differences between benign and malignant thyroid nodules. These features can be broadly categorized into (a) the sonographic features from the ultrasound images, and (b) the non-clinical features extracted from the ultrasound images using statistical and data mining techniques. We also present a brief description of the commonly used classifiers in ultrasound based CAD systems. We then review the studies that used features based on the ultrasound images for thyroid nodule classification and highlight the limitations of such studies. We also discuss and review the techniques used in studies that used the non-clinical features for thyroid nodule classification and report the classification accuracies obtained in these studies.

Diagnosis of Hashimoto's thyroiditis in ultrasound using tissue characterization and pixel classification.

Hashimoto's thyroiditis is the most common type of inflammation of the thyroid gland, and accurate diagnosis of Hashimoto's thyroiditis would be helpful to better manage the disease process and predict thyroid failure. Most of the published computer-based techniques that use ultrasound thyroid images for Hashimoto's thyroiditis diagnosis are limited by lack of procedure standardization because individual investigators use various initial ultrasound settings. This article presents a computer-aided diagnostic technique that uses grayscale features and classifiers to provide a more objective and reproducible classification of normal and Hashimoto's thyroiditis-affected cases. In this paradigm, we extracted grayscale features based on entropy, Gabor wavelet, moments, image texture, and higher order spectra from the 100 normal and 100 Hashimoto's thyroiditis-affected ultrasound thyroid images. Significant features were selected using t-test. The resulting feature vectors were used to build the following three classifiers using tenfold stratified cross validation technique: support vector machine, k-nearest neighbor, and radial basis probabilistic neural network. Our results show that a combination of 12 features coupled with support vector machine classifier with the polynomial kernel of order 1 and linear kernel gives the highest accuracy of 80%, sensitivity of 76%, specificity of 84%, and positive predictive value of 83.3% for the detection of Hashimoto's thyroiditis. The proposed computer-aided diagnostic system uses novel features that have not yet been explored for Hashimoto's thyroiditis diagnosis. Even though the accuracy is only 80%, the presented preliminary results are encouraging to warrant analysis of more such powerful features on larger databases.

Mutations in gastrointestinal stromal tumors diagnosed by endoscopic ultrasound-guided fine needle aspiration.

Gastrointestinal stromal tumors (GISTs) are characterized by overexpression and mutations of c-Kit. Approximately 80% of c-Kit mutations occur in exon 11, being a response factor to imatinib (Gleevec) therapy. We aimed to assess whether c-Kit and PDGFRA mutation analysis of GISTs obtained by endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) could be routinely performed. Mutation analysis of c-Kit hotspot exons (9, 11, 13 and 17) and PDGFRA hotspot exons (12 and 18) was performed in aspirates 51 mesenchymal tumors. We identified c-Kit mutations in 61% of GIST cases, in accordance with previously published ranges (30-90%). Nearly 95% (19/20) of c-kit-mutant tumors carried exon 11 mutations. Mutation analysis is possible in FNA cell blocks and can assist in the diagnosis and therapeutic decisions in GIST cases.

Utility of urinary cytology for diagnosing human polyoma virus infection in transplant recipients: a study of 37 cases with electron microscopic analysis.

Human polyoma virus (HPOV) infection is associated with hemorrhagic cystitis, tubulointerstitial nephritis, and renal transplant dysfunction/allograft loss. We evaluated the utility of cytologic examination to detect HPOV infection in 37 urinary cytology (UC) samples (3 bladder washings, and 34 voided samples) from 29 transplant patients, compared to electron microscopic studies (EMS). Evidence of viral infection was found in 11 specimens (30%). Five cases were diagnosed as HPOV by both UC and EMS. One was positive for HPOV by EMS only. Two cases diagnosed as HPOV by UC were demonstrated to be adenovirus (AV) with EMS. Two cases diagnosed as cytomegalovirus (CMV) by EMS had negative UC. One was called HPOV by UC; EMS in this case was negative. Compared to EMS, the sensitivity and specificity of UC for detecting HPOV were 83% and 90%, respectively, with a positive predictive value of 63% and a negative predictive value of 96%. We conclude that UC is a relatively sensitive and specific method for detecting active HPOV infection in transplant patients, and is important in light of the clinical significance of HPOV infection in transplant recipients. The sensitivity and accuracy of UC for diagnosing HPOV can be increased by adding EMS.

Soft tissue masses in patients with multiple myeloma: a fine-needle aspiration study of 30 cases with flow cytometry and clinical correlation.

BACKGROUND: To the authors' knowledge, the prognostic significance of plasma cell cytology in soft tissue (ST) masses from patients with multiple myeloma (MM) is unknown. Myeloma patients usually are monitored by bone marrow (BM) aspirates and biopsies to assess plasma cell differentiation, tumor burden, and response to treatment. Monitoring of ST lesions by fine-needle aspiration (FNA) is not performed routinely. The objective of the current study was to examine ST masses in MM patients using FNA and to classify and determine the prognostic significance of MM in these lesions based on cytologic features. METHODS: FNAs of 30 ST masses from 27 patients with a history of MM were examined for disease involvement. In the patients with MM, the cytologic features were evaluated and the lesions were graded as low grade, intermediate grade, or high grade based on the classification of Bartl et al. for MM in BM specimens. Concurrent BM samples as well as cytogenetic and flow cytometric results also were reviewed. RESULTS: Twenty-seven of the FNA specimens (90%) were positive for MM, and three specimens (10%) were negative (one case each of lipoma, keratinous cyst, and aspergillosis). Among the MM cases, 5 (18.5%) were low grade, 15 (55.6%) were intermediate grade, and 7 (25.9%) were high grade (blastic MM). Simultaneous BM involvement was present in 23.5% of low-grade MM (4 of 17 cases), 35.3% of intermediate-grade MM (6 of 17 cases), and 71% of high-grade MM (5 of 7 cases). Clinically, 10 of 24 patients (42%) died within 9 months (median, 2 months). Patients with high-grade myeloma (blastic MM) in ST masses appeared to have worse survival; 43% (3 of 7 patients) died by a median time of 2 months, compared with 12% of patients with low-grade and intermediate-grade MM (2 of 17 patients). CONCLUSIONS: FNA of ST masses appears to improve the management of MM patients by providing diagnostic material, samples for ancillary studies, and prognostic information. ST MM can be classified reliably into grades of prognostic significance utilizing the classification of Bartl et al. Intermediate-grade MM was the most frequent subtype present in ST masses.

Tissue effects of salivary gland fine-needle aspiration. Does this procedure preclude accurate histologic diagnosis?

Recent reports have alluded to various tissue effects secondary to fine-needle aspiration (FNA), particularly infarction observed in resected salivary gland masses, precluding accurate histologic diagnosis. Our experience with the use of 25-gauge needles indicates otherwise. We retrospectively reviewed 94 resected salivary gland masses previously sampled by FNA, looking for infarction, hemorrhage, needle track tumor seeding, and fibrosis. We assessed the significance of these complications and their impact on the histologic diagnosis. The median interval from FNA to excision was 25 days. Variable degrees of infarction and hemorrhage were present in 7 cases (7%) and 9 cases (10%), respectively. Infarction ranged from 5% to 80% (average, 20%), while hemorrhage averaged less than 20% of the material on the tissue sections. Significant infarction was present in acinic cell carcinomas (3/7), but histologic diagnosis was not compromised, and tissue alterations were absent. We conclude that FNA of salivary gland lesions using 25-gauge needles is safe and does not significantly alter the histologic diagnosis. The tissue effects observed did not preclude accurate diagnostic interpretation in any case.

Fine-needle aspiration cytology of mucous retention cyst of the tongue: distinction from other cystic lesions of the tongue.

Mucous retention cyst (MRC) is a common submucosal lesion of the oral cavity that, when deeply seated, simulates a neoplasm. This report describes the fine-needle aspiration cytology findings of a lingual MRC of complex architecture and with metaplastic epithelium. In addition, we emphasize its cytologic differential diagnosis from other benign and malignant cystic lesions of the tongue. To the best of our knowledge, this is the first report of aspiration cytology of a complex MRC of the tongue.

Fine-needle aspiration cytology findings in 214 cases of nonparotid lesions of the head.

The use and limitations of fine-needle aspiration (FNA) of lesions of the parotid gland are known, but those of nonparotid lesions of the head have been described only sporadically. We conducted this study to evaluate the utility of FNA and to analyze the causes of diagnostic discrepancies for these lesions. A total of 6,898 FNAs of different sites was performed at our institutions between January 1991-August 1998, and 214 (3.1%) of the cases were FNAs of nonparotid lesions of the head. The most common diagnosis of nonparotid lesions was squamous-cell carcinoma, in 22% (n = 48), and the most common site aspirated was the scalp, in 34% (n = 73). Lipomas and keratinous cysts comprised 5% (n = 9) of the total. A statistical analysis was conducted on 98 paired cytology and histology (n = 83) and cytology and flow cytometry (n = 15) specimens (70 malignant and 28 benign). FNA recognized the malignant and benign nature of the lesion in 60 and 26 cases, respectively with 86% sensitivity 93% specificity and 88% accuracy. Causes of false-negative FNA diagnoses (n = 10) included sampling error (n = 6), bloody smears with scant cellularity (n = 3), and bland cytomorphology (n = 1). Florid granulation tissue and a mucocele of the tongue accounted for the two false-positive cases. We conclude that FNA is an effective tool for triage of surgery candidates with nonparotid lesions of the head. Adequate samples with sufficient cellularity are required for avoiding false-negative diagnoses. Occasionally, tissue biopsy is needed for diagnosis of equivocal cases.

Cytology and flow cytometry of malignant effusions of multiple myeloma.

Identifying malignant plasma cells in body fluids from multiple myeloma patients is important for therapeutic and prognostic considerations. This can be difficult when plasma cells are mature in appearance or low in number. We examined the cytological and flow cytometric findings of myelomatous pleural and pericardial effusions from 8 patients with advanced multiple myeloma. Cytoplasmic immunoglobulin light chain excess vs. DNA ploidy in the plasma-cell population was evaluated by flow cytometry (FCM). The cytology smears of one pericardial and 14 pleural effusions from the 8 patients were reviewed. Screening Papanicolaou-stained smears facilitated the detection of malignant nuclear features; however, morphology of plasma cells was best seen in Diff-Quik-stained smears. Low cellularity and inadequate air-drying of smears accounted for the false-negative cytology seen in two fluids from a single patient. A malignant plasma cell population was identified in 9 of 10 fluids submitted for FCM, including the two fluids with negative cytology. The false-negative FCM was from a suboptimal specimen with high background staining. Six fluids had an aneuploid DNA content, and four were diploid. A combination of Papanicolaou- and Diff-Quik-stained smears is recommended for the evaluation of plasma cells in effusions from patients with multiple myeloma. Cytology and flow cytometry confirmed malignancy in 87% and 90% of fluids evaluated, respectively; all cases were diagnosed by either one or both methods. Our results suggest that FCM and cytology of serous effusions in multiple myeloma patients are complementary and should be used in difficult cases. Diagn. Cytopathol. 2000;22:147-151.

Peripheral primitive neuroectodermal tumor of the parotid gland region: report of a case with fine-needle aspiration findings.

A case of peripheral primitive neuroectodermal tumor of the parotid gland region in a 38-yr-old woman is reported. She had a 1-yr history of a large, firm, and slightly tender left parotid-region mass. CT scan showed an invasive tumor involving the parotid gland, mandible, infratemporal fossa, and parapharyngeal space. Fine-needle aspiration cytology of the mass showed a highly cellular, poorly cohesive smear pattern exhibiting small cuboidal cells, with fibrillary cytoplasm forming occasional rosette-like structures. Numerous intact single cells with fragile cytoplasm, finely granular chromatin, and inconspicuous nucleoli were present together with free-lying nuclei in the background. Histologic, immunohistochemical, and ultrastructural findings confirmed the diagnosis. Diagn. Cytopathol. 2000;22:161-166. Published 2000 Wiley-Liss, Inc.

Intravascular lymphomatosis with bone marrow involvement.

We describe the case of a 56-year-old man who presented with numbness and tingling of the extremities, weakness, and fatigue. Laboratory findings included anemia and thrombocytopenia. A diagnosis of intravascular lymphomatosis was established when liver, omentum, and bone marrow samples were examined. A review of the literature reveals that most cases of intravascular lymphomatosis have cytopenias, mainly anemia and thrombocytopenia, but bone marrow involvement is rare. In our case, a subtle neoplastic infiltrate in the marrow sinusoids was highlighted with a B-cell marker. While immunohistochemical analysis was not performed in most reported cases in the literature, our studies suggest that a systematic search in bone marrow of cases of intravascular lymphomatosis may reveal unsuspected neoplastic cells. We conclude that bone marrow involvement in intravascular lymphomatosis appears to be rare, has subtle features, and is difficult to diagnose if unsuspected and not searched for.

Q-switched neodymium:yttrium-aluminum-garnet laser treatment of lentigo maligna.

Lentigo maligna is a premalignant lesion of atypical melanocytes that typically arises on the head and neck of elderly patients. It is considered a melanoma in situ with a significant risk for transformation to invasive lentigo maligna melanoma. Surgery is the preferred method of treatment; however, because of the advanced age of the typical patient with lentigo maligna, the frequency of complicating medical problems, and the cosmetic or functional aspects of treatment, surgical excision is not always feasible. The purpose of this pilot study was to evaluate the efficacy and safety of Q-switched neodymium:yttrium-aluminum-garnet laser treatment of lentigo maligna. Eight patients were treated with 532 and/or 1064 nm wavelengths from the laser. All patients showed a response to laser therapy, and 2 patients treated with 1 treatment from each wavelength had complete eradication of the LM, with no evidence of recurrence in 42 months. Further study is warranted, but Q-switched neodymium:yttrium-aluminum-garnet laser is a promising alternative treatment for lentigo maligna.

Fine needle aspiration cytology of papillary neoplasms.

This article covers the fine needle aspiration biopsy cytomorphology of papillary carcinomas of different organs, differential diagnoses, and clinical correlation. Diagnostic problems and helpful cytologic features are emphasized. The purpose is to have a concise source of information that helps the pathologist to evaluate these neoplasms.

Diagnostic value of brush cytology in the diagnosis of duodenal, biliary, and ampullary neoplasms.

Endoscopy is a valuable tool in the diagnosis and management of duodenal lesions and biliary strictures. We assessed the value of cytology in the evaluation of these lesions and analyzed the causes of discrepancy among clinical, histologic, and cytologic parameters. The study included 118 patients with duodenal ulcers, ampullary neoplasms, or biliary strictures who were examined between 1975 and 1995; 120 cytologic examinations were performed. The specimens included brushings of the duodenum (DB, n = 50), ampulla (AB, n = 32), and biliary ducts (BB, n = 38). Endoscopic biopsies performed concurrently included the duodenum (n = 37), the ampulla (n = 22), and the biliary ducts (n = 23). Comparison of cytologic and histologic results showed the following sensitivity and specificity: DB, 40% and 97%, respectively; AB, 100% each; BB, 75% and 93%, respectively. The DB, AB, and BB revealed malignant neoplasms in 2 of 5, 7 of 7, and 6 of 8 cases, respectively. Twenty-three duodenal neoplasms were diagnosed by either modality and included 11 adenocarcinomas, 9 villous tumors, 2 metastatic renal cell carcinomas, and 1 large cell non-Hodgkin's lymphoma. Endoscopic brush cytology is an effective means of diagnosing ampullary neoplasms, and it complements tissue biopsy in cases of bile duct stricture. Location, predominance of tumor-induced stroma, an extramucosal growth pattern, sampling error, and interpretative experience influence the diagnostic evaluation. Cytologic diagnosis of an adenoma does not exclude an underlying malignant neoplasm in ampullary tumors. In some instances, it may be difficult to distinguish between villous tumors with severe dysplasia and adenocarcinomas by cytology alone.

Urine cytology of primary and secondary urinary bladder adenocarcinoma.

BACKGROUND: Primary and secondary adenocarcinomas of the urinary bladder are uncommon, and the urine cytology of these tumors has rarely been described. Familiarity with the cytomorphology of these neoplasms may facilitate their detection in urine cytology specimens. METHODS: The authors reviewed 46 urine samples (19 voided, 19 instrumented, and 8 bladder washings) from 41 patients with biopsy-proven primary urinary bladder adenocarcinoma (n = 11) or metastatic adenocarcinoma (n = 35) from the prostate (n = 17), colon (n = 10), breast (n = 3), kidney (n = 3), or uterus (n = 1), or from unknown origin (n = 1). Cytomorphology, the role of cytology, and causes for negative diagnoses were evaluated. RESULTS: Cytologic diagnoses of malignancy, adenocarcinoma not otherwise specified, and adenocarcinoma of a specific type were given in 87%, 28%, and 39% of cases, respectively. Columnar cells, coarse chromatin, and necrosis were found in adenocarcinoma of the colon. Syncytial and acinar arrangements, round or oval nuclei, vesicular chromatin, and prominent nucleoli were commonly found in adenocarcinoma of the prostate. These features permitted us to make a specific diagnosis in 90% of cases of adenocarcinoma of the colon and 41% of cases of adenocarcinoma of the prostate. Cytologic examination failed to lead to a diagnosis of malignancy in 18% of primary adenocarcinoma cases. CONCLUSIONS: A large number of adenocarcinomas of the colon and prostate have sufficient cytologic features to suggest the correct diagnosis in urine samples. The cytomorphology of primary bladder adenocarcinoma is not as easily characterized. The submucosal nature of some metastatic deposits and tumor differentiation influence the diagnostic accuracy.

Cytology of germ cell tumors: extragonadal, extracranial masses and intraoperative problems.

BACKGROUND: Germ cell tumors (GCTs) and their metastases may be found in numerous sites that are accessible to cytologic sampling, and many are responsive to chemotherapy. METHODS: The authors reviewed 20 examples of GCT cytology from 16 males and 3 females ranging in age from 1.5 to 61 years (median, 34 years). With two exceptions, one benign cystic ovarian teratoma in which intraoperative cytology was used to diagnose an associated adult-type carcinoma and one undescended testis in which seminoma presented as an abdominal mass, the material reviewed included no examples of primary gonadal GCT. RESULTS: The authors studied 7 primary and 13 metastatic GCTs; these studies were based on 13 in vivo aspirations, 4 intraoperative preparations, and 3 samples of body cavity fluids. All samples were correctly interpreted as malignant, and only one was incorrectly classified as a non-GCT malignancy. CONCLUSIONS: Clinical and cytologic findings are useful in the diagnosis of GCTs and their metastases. Incorrect interpretation of these neoplasms as poorly differentiated malignancies of other types may deprive the patient of effective chemotherapy. Air-dried, Romanowsky-stained smear material and cell block sections may contribute to the resolution of diagnostic dilemmas.

Basal cell carcinoma metastatic to the salivary glands: differential diagnosis in fine-needle aspiration cytology.

A disparate group of salivary gland neoplasms is characterized by small, uniform, hyperchromatic, basaloid cells. This "small blue cell" pattern is most common in non-Warthin's types of monomorphic adenoma, or in adenoid cystic carcinoma. Small cell anaplastic carcinoma (primary or metastatic), metastatic basaloid squamous cell carcinoma, basal cell adenocarcinoma, and metastatic nasopharyngeal carcinoma are rarely encountered but may present a cytologically similar appearance. We report one female and two male patients (median age = 84 yr) with cutaneous-type basal cell carcinoma (BCC) aspirated from metastatic deposits in the parotid (2 cases) or the submandibular (1 case) gland. One was correctly classified at the time of aspiration, based on a previous history of multiple facial BCC. One was interpreted as carcinoma, the previous history being unavailable at the time of FNA. Smears in these two cases show necrosis and rare keratotic cells. The third cases was mistaken for pleomorphic adenoma (PA); the smears showed metachromatic fragments of collagenous tumor stroma that were misinterpreted as the matrix material typical PA. Similar material was identified in the other two cases. When the "small blue cell" pattern is encountered in salivary bland cytology, one should consider BCC, especially if necrosis is identified. The desmoplastic tumor stroma of BCC may mimic the chondroid matrix of PA. Careful consideration of previous history is very important.

In situ DNA fragmentation assay for detection of apoptosis in paraffin-embedded tissue sections. Technical considerations.

Detection, by light microscopy, of cells in situ undergoing apoptosis has been improved by use of an in situ apoptosis (DNA fragmentation) assay on formalin-fixed and paraffin-embedded tissue sections. We studied conditions of tissue preparation and fixation that may affect the test results. In this study, we intended to determine whether archival tissues prepared under unknown conditions can be used for the in situ apoptosis assay. All tissue sections were pretreated with Proteinase K, followed by incubation with biotinylated 11-deoxyuridine triphosphate in terminal deoxynucleotidyl transferase and then avidin-biotin-peroxidase complex. The following formalin-fixed and paraffin-embedded histologic sections were tested: (1) normal tissues from surgically resected specimens fixed immediately or stored at 4 degrees C and then fixed after 1, 2, 4, 6, or 24 hours; (2) archival autopsy material from histologically normal tissues; and (3) freshly prepared normal tissues from C57 mice. We observed that fixation- and prefixation-elapsed times do not adversely affect the results of the assay. Similar, if not identical results were seen in archival human tissues stored for up to 25 years, the normal tissues freshly prepared from surgical specimens, and the tissues from C57 mice. We conclude that the in situ assay of DNA fragmentation is rapid, sensitive, and reproducible. The use of formalin-fixed and paraffin-embedded archival material as old as 25 years opens the way for a variety of studies of apoptosis in diverse pathologic states.

Diagnostic pitfalls of aspiration cytology of salivary duct carcinoma.

BACKGROUND: Salivary duct carcinoma (SDC) is a highly aggressive primary salivary gland neoplasm that resembles intraductal and infiltrating breast carcinoma histologically. The purpose of this study was to review the cytologic features of histologically proven salivary duct carcinomas and to identify the potential pitfalls in cytologic diagnosis. METHODS: Fine-needle aspiration cytology of nine histologically proven salivary duct carcinomas was reviewed. RESULTS: The patients' age ranged from 62 to 89 years (median, 69 years). There were eight males and one female. The cytologic diagnoses of these cases were as follows: pleomorphic adenoma (PA) (three cases), high grade carcinoma, not otherwise specified (three cases), mucoepidermoid carcinoma (MEC) (two cases), and atypical cytology with differential diagnosis including MEC, oncocytoma, and acinic cell carcinoma (one case). The spectrum of cytologic findings included broad flat sheets and three-dimensional clusters. There was mild to severe cellular pleomorphism and nuclear atypia. Papillary clusters and cribriforming occasionally were identified. Bland cytologic features and prominent hyaline stroma, shown to represent the dense fibrosis on histologic sections, were identified in three cases cytologically interpreted as PA. CONCLUSIONS: Fine-needle aspiration of SDC may be difficult to interpret accurately, and bland cytomorphologic features in some cases may lead to a false-negative interpretation; several clinically important pitfalls are demonstrated in our series.

Cytologic findings in thyroglossal duct carcinoma.

Approximately 150 cases of thyroglossal duct carcinoma, predominantly of the papillary type, have been reported, but the preoperative fine-needle aspiration (FNA) diagnosis of such neoplasms has rarely been cited. The authors describe FNA findings in four samples obtained from three patients who were 29, 50, and 83 years of age, histologically diagnosed as papillary (n = 2) and squamous (n = 1) thyroglossal duct carcinomas. Atypia and squamous cell carcinoma were the FNA diagnoses in the patients with papillary carcinomas. The remaining case was correctly diagnosed as keratinizing squamous cell carcinoma. Cellularity was scant in two cases and moderate in one, and all displayed a cystic background. The authors also reviewed FNA features in 11 papillary and 2 Hurthle cell carcinomas from the English language literature; diagnostic findings were present in less than one third of the cases. In conclusion, familiarity with the FNA findings of thyroglossal duct carcinoma is limited by its rarity. The presence of large, atypical squamous cells, or psammoma bodies, in the FNA material of a midline anterior cystic neck mass should suggest papillary thyroglossal duct carcinoma.