17. Pathology Division
The Pathology Division of the Research Institute East was established in
April 1994. Its research activities are centered on morphological study of
cancer tissue. Major projects include quality control of pathological
diagnosis, a study of prognostic factors of cancers and pathology of
precancerous lesions and early cancers. Techniques available are
morphological study, immunohistochemistry, tissue culture, analysis of
genes and their expression by the polymerase chain reaction, and image
analysis.
Quality Control of Pathological Diagnosis
The accuracy of pathological diagnosis depends on the knowledge and
experience of the pathologists. In daily practice, the dependability of
pathological diagnosis is seldom questioned. However, in
multi-institutional protocol studies, differences in diagnostic criteria
between institutions and individual pathologists may be large enough to
cause inclusion of cases ineligible for the study. This may distort the
treatment results. During the last few years, the extent of discrepancy
between pathologists has been examined for breast cancers, lymphomas, soft
tissue tumors and other neoplasms. It has become apparent that the
discrepancy among pathologists is large enough to warrant central review of
pathological material of the cases registered in protocol studies. In order
to minimize the discrepancy in diagnostic criteria, it seems necessary to
apply new techniques such as gene analysis, cytogenetics and image
analysis.
Prognostic Factors for Cancer
Certain pathological features are associated with good or poor prognosis of
cancer. These include histological parameters, data of image analysis, and
results of examination of phenotypes and genotypes of cancer. This
Division is studying pathological prognostic factors for various cancers.
Examples include prognostic factors for sebaceous gland carcinoma and sweat
gland carcinoma of the skin. Our data demonstrated that the p53 positivity
index and labeling index of the proliferating cell nuclear antigen are good
prognostic factors.(9) Tumors with higher than average indices of these two
antigens had a poorer prognosis than those with less than average indices.
Prognostic factors for breast cancer are also being studied. Patients with
invasive ductal carcinomas of the breast with a fibrotic focus have been
shown to have a shorter disease-free period than those without a fibrotic
focus. The phenotypic and morphological differences between these two
sub-categories of breast cancer are now being studied.
Pathology of Precancerous Lesions and Early Cancers
It is important to study precan-cerous lesions and early cancers, as they
may provide a clue for early recognition of the carcinogenic process. It is
also important from the clinical point of view because studies of the
clinicopathological correlation of these lesions may contribute to the
development of techniques necessary for early detection of cancers.
Atypical adeno-matous hyperplasia (AAH) of the lung has been thought to be
a precancerous lesion of adenocarcinoma of the lung. The similarities and
differences between AAH and well-differentiated adenocarcinoma of the lung
are being studied from the histological point of view and from data on DNA
histograms. Three-dimensional reconstruction of AAH and well-differentiated
adenocarcinoma will be initiated soon. Intraductal papillary lesion of the
breast is another subject under study. Differential diagnosis of benign and
malignant papillary lesions is often difficult. By using data of image
analysis, the differentiation ability of an artificial intelligence (AI)
program is being tested. It has been demonstrated that the AI program can
recognize typical cancer. Further study with difficult cases will
follow.
Effective Use of Image Data
Routine pathological examination generates a large number of images that
include macroscopic and microscopic views, schematic representation of
cancer and electron micrographs. These image data are a valuable resource
for not only pathologists but also clinicians. In order to utilize the
image data effectively, an electronic image storage system is now under
development. This system can store image data with several indices such as
patients' demography and diagnoses. It is expected that this system will
allow pathologists to search for and retrieve necessary images easily and
use them for diagnostic, research and educational purposes.
The images stored in the database are also utilized for computer-assisted
image diagnosis. In this project, characteristics of image data are
converted to numerical data and the latter are analyzed by AI. It is now
being determined whether this system is capable of separating malignant
cells from benign cells. This system is expected to contribute to the
standardization of pathological diagnosis.
List of papers
from this division
Table of Contents