Bo Nordin

PREDICTING BREAST CANCER RECURRENCE

The prognostic value of 435 cytochemical, cytometrical, morphological, epidemiological, and clinical variables was analyzed in a prospective study of 179 breast cancer patients followed for five years after mastectomy. A variable reduction was obtained by first selecting variables correlated with recurrence rate in direct (Students t test) or correlation analysis with consideration of the type of variable analyzed (nominal, interval, ordinal). The 20 variables most strongly correlated with recurrence were analyzed by logistic stepwise regression analysis in order to find out what combination of variables had most discriminatory power in predicting recurrence. It was found that axillary metastization as such was correlated with a combination of variables describing mitotic frequency, size of primary tumor and differentiation of the primary tumor (average cluster size in fine‐needle biopsies). It was also found that there was a strong time dependency in the predictive power of the variables, so that different variable combinations predicted the recurrence rate during the first 2.5 year period (size of axillary metastases and primary tumor, number of lymphocytes around the tumor, mitotic frequency, and degree of differentiation) compared with the second 2.5 year period (variance of DNA content among tumor cell nuclei, number of lymphocytes around the tumor, occurrence of multiple tumors in the operated breast and occurrence of breast cancer among relatives). While other factors previously shown to be correlated with risk of recurrence were also found to be positively correlated here, they were neither as highly predictive as, nor did they increase the predictive value of the above mentioned combined variables. The current study strongly emphasizes that, at the present time, studies of recurrence prediction in human breast cancer should be based on an optimal combination of a number of variables which, independently, influence the prognosis. Further, the current study indicates that prerequisite methods for predicting breast cancer recurrence exist today.

A new method for segmentation of colour images applied to immunohistochemically stained cell nuclei

A new method for segmenting images of immunohistochemically stained cell nuclei is presented. The aim is to distinguish between cell nuclei with a positive staining reaction and other cell nuclei, and to make it possible to quantify the reaction. First, a new supervised algorithm for creating a pixel classifier is applied to an image that is typical for the sample. The training phase of the classifier is very user friendly since only a few typical pixels for each class need to be selected. The classifier is robust in that it is non‐parametric and has a built‐in metric that adapts to the colour space. After the training the classifier can be applied to all images from the same staining session. Then, all pixels classified as belonging to nuclei of cells are grouped into individual nuclei through a watershed segmentation and connected component labelling algorithm. This algorithm also separates touching nuclei. Finally, the nuclei are classified according to their fraction of positive pixels.

A new three-dimensional connected components labeling algorithm with simultaneous object feature extraction capability

Abstract A new algorithm for performing connected component labeling of volume data is presented in this paper. The algorithm uses a table that combines efficient handling of label equivalences with the flexibility to add the calculation of features for each labeled component as well as to set various feature thresholds. The volume of each component is calculated in our implementation and it is possible to set a volume threshold for discarding small regions. The reuse of storage in the table is implemented in a simple but natural way.

COMPUTER ANALYSIS OF CERVICAL CELLS AUTOMATIC FEATURE EXTRACTION AND CLASSIFICATION

A prescreening instrument for cervical smears using computerized image processing and pattern recognition techniques requires that single cells in the specimen can be automatically isolated and analyzed. This paper describes a dual wavelength method for automatic isolation of the cytoplasm and nuclei of cells. Density-oriented, shape-oriented and texture-oriented parameters were calculated and evaluated for more than 600 cells. It is shown that the computer can be taught to distinguish between normal and atypical cells with an accuracy of ca. 97%, while human classification reproducibility is ca. 95%. In addition, an attempt to assign a measure of atypia to individual cells is described.

Multifocal breast carcinoma

The present study comprises 173 breast cancers in women living in Sweden. Mastectomy was performed and the surgical specimens were thoroughly scrutinized histopathologically with special attention given to the mammary tissue outside the dominant mass. Fifty-two patients (30 percent) had multifocal growth in the same breast as the dominant breast cancer. The patients have been followed up for 3 years after operation and compared with age-matched controls. The multifocal growth was not correlated with age, size of the tumor or death of the patient within 3 years. A previous diagnosis of benign breast disease was significantly more common among the cancer patients than among the controls, although it was unrelated to multifocal growth. This study stresses the importance of considering the high incidence of multifocal growth of breast cancer when discussing treatment by operation less radical than mastectomy.

Watershed segmentation of binary images using distance transformations

A common problem with binary images generated by a segmentation algorithm is to split the domains either into different objects or an object into different parts. While it is easy to do this interactively by drawing lines in the image it is a much more difficult task to formulate rules for this operation in a computer language and thus automate the procedure. This paper presents a fast algorithm that yields results very similar to an interactive splitting procedure for the domains of a binary image. The algorithm is based on watershed segmentation using distance transformations. We let a pixel belong to a watershed line if at least two neighbors belong to differently labeled segments. We have criteria for relabeling segments which do not become large enough to form segments of their own. After having labeled all pixels we replace every watershed line with the lines with shortest distances. The algorithm preserves the shape and the number of segments with good accuracy and is also independent of how the domains are rotated in the image.

High resolution segmentation of cervical cells.

A major problem in the automation of cervical cytology screening is the segmentation of cell images. This paper presents the present status of the work on that problem at the University of Uppsala. A dual resolution system is used. Suspect malignant cells are located at 4 mu resolution. Each such cell is rescanned at 0.5 mu resolution at two different wavelengths, 530 and 570 nm. The nucleus and the cytoplasm are isolated each by two independent methods. For the nucleus adaptive thresholding in the histogram of the 570 nm image and a contouring in a radially transformed version of that image is used. For the cytoplasm a two dimensional thresholding in the 2D histogram and a contouring in a radially transformed version of the 530 nm image is used. If the two nuclear masks agree the surrounding area is checked for disturbing objects. If also the cytoplasm masks agree and are without disturbing objects the whole cell is accepted. The result of the cytoplasm masks agree and are without disturbing objects the whole cell is accepted. The result of the segmentation is thus three categories; free cells, free nuclei and rejected objects. The shape of the objects belonging to the former two categories is checked and irregularly shaped ones are rejected as probably consisting of several overlapping nuclei. Cells passing also this test are classified as normal or malignant. The experience from using this algorithm is discussed and areas for further research are pointed out.

Segmentation of cervical cells: Detection of overlapping cell nuclei☆

Abstract A method for detecting overlapping cell nuclei in Pap-stained cervical smears is described. The algorithm uses information both from the nuclear contour and from the density profile of the nucleus. For the analysis of the nuclear contour the smoothed difference chain code is used. From this code any significant concavities along the contour are found and a number of features describing their size and relative location are computed. If these clearly indicate an overlap situation the object is classified as an overlap. Otherwise a density profile is generated along a line orthogonal to the line joining the two major concavities. This profile is checked for peaks and valleys indicative of an overlap situation and a new set of features are generated and used to classify the object as single or overlapping. The algorithm performed reasonably well when tested on an independent test set of about 240 cell images.

MUSE — a new tool for interactive image analysis and segmentation based on multivariate statistics

MUSE--a new software tool for the interactive exploration of multivariate images and the development of image segmentation methods has been designed, implemented, and tested in a number of real application projects. The multivariate statistical classification and projection methods in MUSE can be used not only to analyze multispectral images but also, in special cases, multitemporal images and volume images. Additionally MUSE can be applied to normal greyscale images provided they are made multivariate through an initial processing step. This step may consist in the application of filters designed to enhance any existing texture differences between different regions in the images. MUSE has been successfully applied to medical images (color photographs, MR, PET, SPECT) as well as to satellite images (Landsat TM).

CELLO - An Interactive System for Image Analysis

CELLO is an interactive, command oriented image analysis system. It has been developed for applications in automated cytology i.e. for the development of microscopic cell image analysis algorithms but is very flexible and can be used for a wide range of image analysis applications.