A framework for evaluating image segmentation algorithms.

The purpose of this paper is to describe a framework for evaluating image segmentation algorithms. Image segmentation consists of object recognition and delineation. For evaluating segmentation methods, three factors-precision (reliability), accuracy (validity), and efficiency (viability)-need to be considered for both recognition and delineation. To assess precision, we need to choose a figure of merit, repeat segmentation considering all sources of variation, and determine variations in figure of merit via statistical analysis. It is impossible usually to establish true segmentation. Hence, to assess accuracy, we need to choose a surrogate of true segmentation and proceed as for precision. In determining accuracy, it may be important to consider different 'landmark' areas of the structure to be segmented depending on the application. To assess efficiency, both the computational and the user time required for algorithm training and for algorithm execution should be measured and analyzed. Precision, accuracy, and efficiency factors have an influence on one another. It is difficult to improve one factor without affecting others. Segmentation methods must be compared based on all three factors, as illustrated in an example wherein two methods are compared in a particular application domain. The weight given to each factor depends on application.     

Holliday junctions in FLP recombination: resolution by step-arrest mutants of FLP protein.

The FLP "recombinase" of the 2-micron circle yeast plasmid can resolve synthetic FLP site-Holliday junctions. Mutants of the FLP protein that are blocked in recombination but are normal in substrate cleavage can also mediate resolution. The products of resolution by these mutants are almost exclusively nicked molecules with a protein-bound 3' end. There is no significant asymmetry in strand cleavage (top versus bottom) by the mutants in linear or in circular FLP substrates; nor is there a bias in resolution (toward parentals or toward recombinants) of Holliday junctions (corresponding to top- or to bottom-strand exchange) by wild-type FLP. During normal FLP recombination, a small amount of the expected Holliday intermediate can be detected.     

The value of fine-needle aspiration biopsy in the cytodiagnosis of salivary gland lesions

Fine-needle aspiration cytology (FNAC) was performed on 195 cases of salivary gland lesions. The smears were technically adequate in 178 cases. Tissue examination was available for subsequent histocytologic correlation in 57 cases. The cytodiagnosis included inflammatory lesions (59) and benign (68) and malignant (51) tumors (total, 119). The accuracy of cytodiagnosis was 87.7% with a sensitivity of 80.9% and a specificity of 94.3%. Exact histologic typing was possible in 61.9% of the malignant tumors. Mucoepidermoid tumors and cellular-atypical pleomorphic adenoma posed difficulties in cytodiagnosis.     

The diagnostic usefulness of tumour markers CEA and CA-125 in pleural effusion

Pleural effusion is a common diagnostic problem. The analysis of serum and pleural fluid for tumour markers is widely used as a diagnostic aid in clinical practice. The aim of the present study was to determine the usefulness of simultaneous quantification of carcinoembryonic antigen (CEA) and carbohydrate antigen (CA-125) in distinction of malignant from benign effusion. Data from a total of 78 patients including 53 patients with benign and 25 patients with malignant effusion was evaluated. The cut-off values for differentiating benign from malignant effusions were determined using results obtained from patients with known benign effusions (mean + 2 SD, 95% confidence interval). The cut-off for CEA and CA-125 were 5.1 ng/ml and 1707 IU/ml respectively. CEA assay in pleural fluid had an acceptable sensitivity and good specificity of 64% and 98% respectively. CA-125 had a sensitivity of 36% and specificity of 94%. The combination of the two tumour markers gave a sensitivity of 72% and specificity of 92.4%. We suggest a good clinical strategy may be to begin with CEA measurement (assay specificity 98%); if CEA is below the cut-off value (negative), CA-125 could then be measured to improve the sensitivity of detection of malignant effusions. However, measurement of these tumour markers is not cost effective from the point of view that it does not give information on the type of malignancy present. The latter has to be determined either by histological or cytological study.     

Cytology of mucinous carcinoma of breast: a report of 28 cases with histological correlation

Pure mucinous carcinoma (MC) of the breast is a relatively uncommon variant of breast carcinoma with distinctive histological and cytological features. In this study we have analysed fine needle aspiration (FNA) cytological material from 28 cases of MC of breast and correlated the cytomorphological features with histopathology. The 28 patients consisted of 27 females and one male patient. 14 patients were Chinese, 10 were Indian and four were Malay. Their ages ranged from 38 to 90 with a mean at 52 years. The left breast was involved in 17 and the right in 11 cases. The duration of the lump varied from two weeks to 10 years. The cytological picture was characterized by abundant extracellular mucin giving a "sea of mucin" or "whirlpools of mucin" effect, in which were seen floating clusters of tumour cells with relatively bland cytological features. Myxo-vascular fragments were seen in 12 cases. Dissociated tumour cells showed a plasmacytoid appearance with eccentric nuclei. In four cases, the mucin was scanty in amount and the cellularity was high while in two cases, the cellularity was very low. Psammoma bodies were seen in cytological smears in one case. Histological study of excision or mastectomy specimens confirmed all 28 cases to be pure MC. Knowledge of the distinctive cytomorphological appearance of MC would enable correct identification of these lesions as malignant and prompt treatment that could further enhance the survival of these prognostically good breast cancers.     

Total hemolytic complement (CH50) and its fractions (C3 and C4) in the sera of patients with carcinoma of the oral cavity,uterine cervix,and breast

The total hemolytic complement activity of CH50 and its fractions C3 and C4 was determined in the sera of 196 patients with carcinoma of the oral cavity, 172 patients with carcinoma of the uterine cervix, and 166 patients with breast cancer. The values were compared with those of 18 patients with mammary dysplasia, 32 patients with mild to moderate dysplasia of the cervix, and 100 healthy, normal age- and sex-matched controls. No alterations in CH50, C3, and C4 were observed in the sera of patients with benign lesions, whereas a significant rise in the three factors was observed in all the cancer patients studied. The complement activity increased significantly with the progression of the disease up to stage III and remained persistently elevated thereafter. Patients who had a clinical cure had normal levels of CH50, C3, and C4, whereas the values remained elevated in patients who were still undergoing treatment for residual lesions.     

Fine-needle aspiration cytology of phyllodes tumors

Breast lesions with a significant spindle cell or mesenchymal component are not commonly encountered in fine-needle aspiration (FNA) cytologic material and include a heterologous variety of benign and malignant conditions, with phyllodes tumors (PTs) being the foremost differential diagnostic consideration. This study comprises 28 tumors diagnosed histologically as PT in which FNAC material was available for review. Histological sections and cytological smears from these cases were retrieved and subjected to detailed morphological review. Cytological parameters assessed included ratio of stroma to epithelium, pattern characteristics and cytological characteristics of the stromal, and epithelial components and the background cells. Large and hypercellular stroma fragments, dissociated spindle and plump stromal cells, often accompanied by large, folded sheets of epithelium were cytological features that characterized PT. Smears from malignant PT showed predominantly or solely mesenchymal components. FNAC was a highly reliable procedure for the diagnosis of PT, giving an accuracy rate of 92.8%.     

Papillary cystic type of acinic cell carcinoma of parotid: fine needle aspiration cytological features of a high grade variant with oncocytic metaplasia

A 60-year-old female developed a right parotid swelling six months after surgery for intra-oral squamous cell carcinoma. Fine needle aspiration (FNA) cytological smears showed dissociated large and small pleomorphic tumour cells with abundant mitoses and oncocytic features. A cytological diagnosis of parotid acinic cell carcinoma (ACC) was made. Histological study of the subtotal parotidectomy specimen showed a papillary cystic variety of acinic cell carcinoma (ACC-PCV). FNA cytological features in this case of ACC-PCV differs from the two previously reported cases in that it showed prominent oncocytic and high grade features and absence of papillary pattern in the cytological smears. ACC-PCV is an uncommon tumour and knowledge of its varied FNA cytological features is important for the diagnosis of this neoplasm.