Description of Cancer
Cancer is a group of diseases of higher multicellular
organisms. It is characterized by alterations
in the expression of multiple genes, leading to
dysregulation of the normal cellular program for
cell division and cell differentiation. This results
in an imbalance of cell replication and cell death
that favors growth of a tumor cell population.
The characteristics that delineate a malignant
cancer from a benign tumor are the abilities to
invade locally, to spread to regional lymph nodes,
and to metastasize to distant organs in the body.
Clinically, cancer appears to be many different
diseases with different phenotypic characteristics.
As a cancerous growth progresses, genetic
drift in the cell population produces cell heterogeneity
in such characteristics as cell antigenicity,
invasiveness, metastatic potential, rate
of cell proliferation, differentiation state, and
response to chemotherapeutic agents. At the
molecular level, all cancers have several things
in common, which suggests that the ultimate
biochemical lesions leading to malignant transformation
and progression can be produced by a
common but not identical pattern of alterations
of gene readout. In general, malignant cancers
cause significant morbidity and will be lethal to
the host if not treated. Exceptions to this appear
to be latent, indolent cancers that may remain
clinically undetectable (or in situ), allowing the
host to have a standard life expectancy.
Some points in the description may not seem
intuitively obvious. For example, cancer doesn’t
just occur in humans, or just mammals for that
matter. Cancer (or at least tumorous growths—
these may or may not have been observed to
metastasize) has been observed in phyla as old as
Cnidaria, which appeared almost 600 million
years before the present, and in other ancient
phylasuchasEchinodermata(>500millionyears
old), Cephalopoda (500 million years old), Amphibia
(300 million years old), and Aves (150
million years old). Curiously, cancer has never
been seen (or at least reported) in a number of
phyla such as Nematoda, Tradigrada, and Rotifera.
It is intriguing to consider that these organisms
may have some protective mechanisms
that prevent them from getting tumors. If so, it
would be important to find out what these
mechanisms are.
One thing is clear, though, which is that
cancer is a disease of multicellular organisms.
This trait implies that there is something inherent
in the ability of cells to proliferate in
4 CANCER BIOLOGY
clumps or to differentiate into different cell
types and move around in the body to sites of
organogenesis that is key to the process of tumorigenesis.
Problems occur when these processes
become dysregulated.
One might also argue that evolution itself has
played some tricks on us because some of the
properties selected for may themselves be processes
that cancer cells use to become invasive
and metastatic. Or to phrase it differently: Is
cancer an inevitable result of a complex evolutionary
process that has advantages and disadvantages?
Some of these processes might be the
following:
1. The mechanism of cell invasiveness that
allows the implantation of the early embryo
into the uterine wall and the development
of a placenta.
2. Cell motility that allows neural cells, for
example, to migrate from the original neural
crest to form the nervous system.
3. The development of a large, complex genome
of up to 40,000 genes that must be
replicated perfectly every time a cell divides.
4. The large number of cells in a human or
higher mammal that must replicate and
differentiate nearly perfectly every time
(some can be destroyed if they become
abnormal).
5. The long life span of humans and higher
mammals, increasing the chance for a
genetic ‘‘hit’’ to occur and lead a cell down
a malignant path.
As we shall see in later chapters of this book,
cancer cells take advantage of a number of these
events and processes.
Other questions that arose at the gathering
above from scientists not in the field of cancer
were the following:
1. Is there a single trait or traits that all
cancer cells have?
2. How many genetic ‘‘hits’’ does it take to
make a cancer cell?
3. What kinds of genes are involved in these
hits?
These questions are all dealt with in later
chapters. Suffice it to say here that for a cell to
become cancerous or at least take the first steps
to becoming cancerous, at least two genetic hits
are required. One may be inherited and another
accrued after birth or both may be accrued after
birth (so-called somatic, or spontaneous, hits).
The kinds of genes involved are oncogenes,
which when activated lead to dysregulated cell
proliferation, and tumor suppressor genes, which
become inactivated or deleted, producing a loss
of the cell’s checks and balances controlling cell
proliferation and differentiation.
The single most common, if not universal,
trait that occurs in all cancers is genetic drift. or
the ability of cells to lose the stringent requirement
for preciseDNA replication and to acquire
the ability to undergo sequential progressive
changes in their genome, through mutations,
gene rearrangement, or gene deletion. This
has sometimes been called the acquisition of a
‘‘mutator phenotype.’’
WHAT SIGNIFICANT EVENTS HAVE
HAPPENED IN CANCER RESEARCH
IN THE LAST 25 YEARS?
As I was beginning to gather my thoughts for the
fourth edition of Cancer Biology, one of my
colleagues mentioned that he thought it would
be of interest to describe the significant things
that have happened in cancer biology in the
25 years since the first edition was published
(1981). Many things have happened since then,
of course, and everyone has their favorite list.
But looking back at the table of contents for the
first edition and at the outline for this edition,
several things struck me, as listed below.
1. Cancer susceptibility genes. In 1981 we
knew that familial clustering of some cancers
occurred, for example, with colon cancer,
but the genes involved in this hadn’t
been determined. The APC, BRCA-1,
BRCA-2, and p53 inherited mutations, for
example, were not known at that time. Research
in this area has identified a number
of genes involved in cancer susceptibility,
andwithmoderncloningtechniques,more
are identified every few months.
2. The techniques of modern molecular
biology were in their infancy at that time.
Polymerase chain reaction (PCR), DNA
CHARACTERISTICS OF HUMAN CANCER 5
microarrays, protein chips, and bioinformatics
were not terms in anybody’s
dictionary.
3. Genes involved in cancer initiation and
promotion were very poorly defined. Although
we knew that chemicals and irradiation
could damage DNA and initiate
cancer in animals and humans, the specific
genes altered were almost completely
unknown. We now know a lot about the
genes involved at various stages of a number
of cancers. For example, the work of
Bert Vogelstein and colleagues has defined
a pathway sometimes called the
‘‘Vogelgram’’ for the progression of colon
cancer (see Chapter 5). We knew that
DNArepairwas important and that heritable
conditions of defective DNA repair
(e.g., xeroderma pigmentosum) could lead
to cancer, but the ideas about the mechanisms
of DNA repair were primitive.
4. The identification of oncogenes didn’t
really start until the early 1980s. The src
gene was identified in 1976 by Stehelin
et al., and erb, myc, and myb oncogenes
were identified in the late 1970s, but this
was about the limit of our knowledge
(see Chapter 5).
5. The term tumor suppressor gene wasn’t
even coined until the early 1980s, although
their existence had been implied
from the cell fusion experiments of
Henry Harris, (Chapter 5) who showed
that if a normal cell was fused with a
malignant cell, the phenotype was usually
nonmalignant. The RB gene was the
first one cloned, in 1983 by Cavenee et al.
(Chapter 5) p53 was originally thought
of as an oncogene. It wasn’t realized until
1989 that wild-type p53 could actually
suppress malignant transformation. A
number of tumor suppressor genes have,
of course, been identified since then.
6. Starting in the 1970s, cell cycle checkpoints
were identified in yeast by Lee
Hartwell and colleagues, but the identification
of human homologs of these genes
didn’t occur until the late 1980s (seeChapter
4).
7. Tumor immunology was still poorly understood
in 1981—both the mechanism
of the immune response and the ability
to manipulate it with cytokines, activated
dendritic cells, and vaccines. Such manipulation
was not in the treatment armamentarium.
8. The first treatment of a patient with
gene therapy occurred in 1990. Several
gene therapy clinical trials for cancer are
under way and some gene therapy modalities
will likely be approved in the next
few years.
9. The viral etiology of cancer was still being
widely debated in 1981. The involvement
of Epstein-Barr virus in Burkitt’s
lymphoma and of hepatitis B virus in
liver cancer was becoming accepted, but
the role of viruses in these diseases and
in cervical cancer, Kaposis’ sarcoma, and
in certain T-cell lymphomas became
clearer much later.
10. Although some growth factors that affect
cancer cell replication, such as IGF-1
and IGF-2, FGF, NGF, PDGF, and
EGF, were known in 1981, knowledge
about their receptors and signal transduction
mechanisms was primitive indeed.
Tumor growth factor a was known
as sarcoma growth factor (SGF), and the
existence of its partner, TGF-b, was only
implied from what was thought to be
a contaminating HPLC peak from the
purification procedure. The explosion
of knowledge about signal transduction
mechanisms and how these pathways interact
has been a tremendous boon to our
understanding of how cells respond to
signals in their environment and communicate
with each other.
11. Knowledge about the regulation of gene
expression has greatly increased in the
past 25 years, on the basis of our current
information on the packaging of chromatin,
transcription factors, coinducers
and corepressors, and inhibitory RNA
(siRNA).
12. While not topics discussed in detail in
the earlier editions of Cancer Biology, advances
in diagnostic imaging such as magnetic
resonanceimaging (MRI), computed
tomography (CT), and positron emission
tomography (PET) have significantly im-
6 CANCER BIOLOGY
proved cancer diagnosis. Improved radiation
therapy, combined modality therapy,
bone marrow transplant, and supportive
care have also improved significantly.
BASIC FACTS ABOUT CANCER
Cancer is a complex family of diseases, and carcinogenesis,
the events that turn a normal cell in
the body into a cancer cell, is a complex multistep
process. From a clinical point of view, cancer
is a large group of diseases, perhaps up to a
hundred or more, that vary in their age of onset,
rate of growth, state of cellular differentiation,
diagnostic detectability, invasiveness, metastatic
potential, response to treatment, and prognosis.
From a molecular and cell biological point of
view, however, cancer may be a relatively small
number of diseases caused by similar molecular
defects in cell function resulting from common
types of alterations to a cell’s genes. Ultimately,
cancer is a disease of abnormal gene expression.
There are a number of mechanisms by which
this altered gene expression occurs. These mechanisms
may occur via a direct insult to DNA,
such as a gene mutation, translocation, amplification,
deletion, loss of heterozygosity, or via a
mechanism resulting from abnormal gene transcription
or translation. The overall result is an
imbalance of cell replication and cell death in
a tumor cell population that leads to an expansion
of tumor tissue. In normal tissues, cell proliferation
and cell loss are in a state of equilibrium.
Cancer is a leading cause of death in the
Western world. In the United States and a number
of European countries, cancer is the secondleading
killer after cardiovascular disease, although
in the United States since 1999 cancer
has surpassed heart disease as the number one
cause of death in people younger than 85.1 Over
1.3 million new cases of cancer occur in the
United States each year, not including basal cell
and squamous cell skin cancers, which add another
1 million cases annually. These skin cancers
are seldom fatal, do not usually metastasize, and
are curable with appropriate treatment, so they
are usually considered separately. Melanoma,
by contrast, is a type of skin cancer that is more
dangerous and can be fatal, so it is considered
with the others. The highest mortality rates are
seen with lung, colorectal, breast, and prostate
cancers (Fig. 1–1). Over 570,000 people die
each year in the United States from these and
other cancers. More people die of cancer in 1
year in the United States than the number of
people killed in all the wars in which the United
States was involved in the twentieth century
(Fig. 1–2).
In many cases the causes of cancer aren’t
clearly defined, but both external (e.g., environmental
chemicals and radiation) and internal
(e.g., immune system defects, genetic predisposition)
factors play a role (see Chapter 2). Clearly,
cigarette smoking is a major causative factor.
These causal factors may act together to initiate
(the initial genetic insult) and promote (stimulation
of growth of initiated cells) carcinogenesis.
Often 10 to 20 years may pass before an
initiated neoplastic cell grows into a clinically
detectable tumor.
Although cancer can occur at any age, it is
usually considered a disease of aging. The average
age at the time of diagnosis for cancer of
all sites is 67 years, and about 76% of all cancers
are diagnosed at age 55 or older. Although cancer
is relatively rare in children, it is the secondleading
cause of death in children ages 1–14. In
this age group leukemia is the most common
cause of death, but other cancers such as osteosarcoma,
neuroblastoma, Wilms’ tumor (a kidney
cancer), and lymphoma also occur.
Over eight million Americans alive today have
had some type of cancer. Of these, about half
are considered cured. It is estimated that about
one in three people now living will develop some
type of cancer.
There has been a steady rise in cancer death
rates in the United States during the past 75
years. However, the major reason why cancer
accounts for a higher proportion of deaths now
than it did in the past is that today more people
live long enough to get cancer, whereas earlier
in the twentieth century more people died of
infectious disease and other causes. For example,
in 1900 life expectancy was 46 years for men
and 48 years for women. By 2000, the expectancy
had risen to age 74 for men and age 80 for
women. Thus, even though the overall death
rates due to cancer have almost tripled since
1930 for men and gone up over 50% for women,
CHARACTERISTICS OF HUMAN CANCER 7
the age-adjusted cancer death rates in men have
only increased 54% in men and not at all for
women.2
The major increase has been in deaths due to
lung cancer. Thus, cigarette smoking is a highly
suspect culprit in the observed increases. In addition,
pollution, diet, and other lifestyle changes
may have contributed to this increase in cancer
mortality rates (Chapter 3). The mortality rates
for some cancers has decreased in the past 50
years (e.g., stomach, uterine cervix); however, the
mortality rates have been essentially flat for many
of the major cancers such as breast, colon, and
prostate, although 5-year survival rates have improved
for these cancers (see Chapter 3).
It is instructive to examine the trends in cancer
mortality over time to get some clues about
the causes of cancer. For males, lung cancer
remains the number one cancer killer (Fig. 1–3).
With a lag of about 20 years, its rise in mortality
parallels the increase in cigarette smoking
among men, which has an almost identical curve
starting in the early 1900s. Lung cancer mortality
rates for men have decreased somewhat
since 1990, and death rates for colorectal cancer
have dropped slightly in recent years, whereas
prostate cancer mortality has increased somewhat.
Stomach cancer mortality has dropped
significantly since the early 1900s, presumably
because of better methods of food preservation
Males
Prostate
Lung and Bronchus
Colon and Rectum
Urinary Bladder
Melanoma of the Skin
Non-Hodgkin Lymphoma
Kidney and Renal Pelvis
Leukemia
Oral Cavity and Pharynx
Pancreas
All Sites
Breast
Lung and Bronchus
Colon and Rectum
Uterine Corpus
Non-Hodgkin Lymphoma
Melanoma of the Skin
Ovary
Thyroid
Urinary Bladder
Pancreas
All Sites
Lung and Bronchus
Breast
Colon and Rectum
Ovary
Pancreas
Leukemia
Non-Hodgkin Lymphoma
Uterine Corpus
Multiple Myeloma
Brain and Other Nervous System
All Sites
Lung and Bronchus
Prostate
Colon and Rectum
Pancreas
Leukemia
Esophagus
Liver and Intrahepatic Bile Duct
Non-Hodgkin Lymphoma
Urinary Bladder
Kidney and Renal Pelvis
All Sites
232,090
93,010
71,820
47,010
33,580
29,070
22,490
19,640
19,100
16,100
710,040
33%
13%
10%
7%
5%
4%
3%
3%
3%
2%
100%
211,240
79,560
73,470
40,880
27,320
26,000
22,220
19,190
16,200
16,080
662,870
32%
12%
11%
6%
4%
4%
3%
3%
2%
2%
100%
73,020
40,410
25,750
16,210
15,980
10,030
9050
7310
5640
5480
275,000
27%
15%
10%
6%
6%
4%
3%
3%
2%
2%
100%
90,490
30,350
28,540
15,820
12,540
10,530
10,330
10,150
8970
8020
295,280
31%
10%
10%
5%
4%
4%
3%
3%
3%
3%
100%
Females
Males
Estimated Deaths
Estimated New Cases*
Females
Figure 1–1. Ten leading cancer types for estimated new cancer cases and
deaths, by sex, United States, 2005. *Excludes basal and squamous cell skin
cancers and in situ carcinoma except urinary bladder. Estimates are rounded
to the nearest 10. Percentage may not total 100% due to rounding. (From
American Cancer Society, Surveillance Research, 2005. CA Cancer J Clin
2005; 55:10–30, with permission.)
8 CANCER BIOLOGY
(e.g., better refrigeration, less addition of nitrate
and nitrate preservatives). Cancer of the gastroesophageal
junction, however, has risen significantly
in recent years, perhaps due to obesity
and increased incidence of gastric reflux into the
esophagus in the U.S. population.
Somewhat surprising, perhaps, is the fact that
lung cancer has overtaken breast cancer as the
number one cancer killer in women (Fig. 1–4).
This increase occurred in the late 1980s and, as
was the case for males, parallels the rise in the
percentage of women who smoke. Smoking
started to increase dramatically during World
War II. Rosie the Rivetter picked up some bad
male habits along with increased access to traditionally
male jobs.
Breast cancer mortality rates have remained
stubbornly stable, although a small decrease
(5%) has occurred since 1990. Uterine cancer
death rates have been going down, primarily
through earlier detection and treatment of cervical
cancer. Female colon cancer mortality has
been decreasing, but the reasons for this aren’t
clear. As in males, stomach cancer mortality in
women has been going down for many years.
The good news is that more and more people
are being cured of their cancers today. In the
1940s, for example, only one in four persons
diagnosed with cancer lived at least 5 years after
treatment; in the 1990s that figure rose to 40%.
When normal life expectancy is factored into
this calculation, the relative 5-year survival rate
is about 64% for all cancers taken together.1
Thus, the gain from 1 in 3 to 4 in 10 survivors
means that almost 100,000 people are alive now
who would have died from their disease in less
than 5 years if they had been living in the 1940s.
This progress is due to better diagnostic and
treatment techniques, many of which have come
about from our increasing knowledge of the
biology of the cancer cell.
HALLMARKS OF MALIGNANT
DISEASES
Malignant neoplasms or cancers have several
distinguishing features that enable the pathologist
or experimental cancer biologist to characterize
them as abnormal. The most common
600
500
400
300
200
100
0
0.01
Gulf Vietnam
Death in Thousands
Total Battle Deaths Each Year
WWI Korea WWII Cancer AIDS Murder
48 54
104
292
550
41 25
Figure 1–2. Total battle deaths from all wars with U.S. involvement in the
twentieth century, compared to number of deaths each year from cancer,
AIDS, and murder in the United States. (Personal communication from Don
Coffey, Johns Hopkins University, with permission.)
CHARACTERISTICS OF HUMAN CANCER 9
types of human neoplasms derive from epithelium,
that is, the cells covering internal or external
surfaces of the body. These cells have a supportive
stroma of blood vessels and connective
tissue. Malignant neoplasms may resemble normal
tissues, at least in the early phases of their
growth and development. Neoplastic cells can
develop in any tissue of the body that contains
cells capable of cell division. Though they may
grow fast or slowly, their growth rate frequently
exceeds that of the surrounding normal tissue.
This is not an invariant property, however, because
the rate of cell renewal in a number of
normal tissues (e.g., gastrointestinal tract epithelium,
bone marrow, and hair follicles) is as
rapid as that of a rapidly growing tumor.
The term neoplasm, meaning new growth, is
often used interchangeably with the term tumor
to signify a cancerous growth. It is important to
keep in mind, however, that tumors are of two
basic types: benign and malignant. The ability to
distinguish between benign and malignant tumors
is crucial in determining the appropriate
treatment and prognosis of a patient who has
a tumor. The following are features that differentiate
a malignant tumor from a benign tumor:
1. Malignant tumors invade and destroy adjacent
normal tissue; benign tumors grow
by expansion, are usually encapsulated,
and do not invade surrounding tissue.
Benign tumors may, however, push aside
30
40
50
60
Stomach
Rate per 100,000 Males
Leukemia
1930
1932
1934
1936
1938
1940
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Colon and Rectum
Liver
Year of Death
Pancreas
Prostate
Lung and Bronchus
70
80
90
100
20
10
0
Figure 1–3. Annual age-adjusted cancer death rates* among males for selected
cancer types, United States, 1930 to 2001. *Rates are age adjusted to
the 2000 U.S. standard population. Because of changes in ICD coding, numerator
information has changed over time rates for cancers of the lung and
bronchus, colon and rectum, and liver are affected by these changes. (From
U.S. Mortality Public Use Data Tapes, 1960 to 2001, U.S. Mortality Volumes,
1930 to 1959, National Center for Health Statistics, Centers for Disease
Control and Prevention, with permission.)
10 CANCER BIOLOGY
normal tissue and may become life threatening
if they press on nerves or blood
vessels or if they secrete biologically active
substances, such as hormones, that alter
normal homeostatic mechanisms.
2. Malignant tumorsmetastasize through lymphatic
channels or blood vessels to lymph
nodes and other tissues in the body. Benign
tumors remain localized and do not
metastasize.
3. Malignant tumor cells tend to be ‘‘anaplastic,’’
or less well differentiated than normal
cells of the tissue in which they arise. Benign
tumors usually resemble normal tissue
more closely than malignant tumors do.
Some malignant neoplastic cells at first
structurally and functionally resemble the
normal tissue in which they arise. Later, as
themalignant neoplasmprogresses, invades
surrounding tissues, and metastasizes, the
malignant cells may bear less resemblance
to the normal cell of origin. The development
of a less well-differentiated malignant
cell in a population of differentiated normal
cells is sometimes called dedifferentiation.
This term is probably a misnomer for the
process, because it implies that a differentiated
cell goes backwards in its developmental
process after carcinogenic insult. It
is more likely that the anaplastic malignant
30
25
15
5
40
35
50
45
60
55
65
Stomach
Rate per 100,000 Females
Ovary
1930
1932
1934
1936
1938
1940
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Colon and Rectum
Uterus
Year of Death
Pancreas
Breast
Lung and Bronchus
70
75
80
85
90
95
100
20
10
0
Figure 1–4. Annual age-adjusted cancer death rates* among females for
selected cancer types, United States, 1930 to 2001. *Rates are age adjusted to
the 2000 U.S. standard population. Because of ICD coding, numerator
information has changed over time, rates for cancers of the uterus, ovary,
lung and bronchus, and colon and rectum are affected by these changes.
Uterus cancers are for uterine cervix and uterine corpus combined. (From
U.S. Mortality Public Use Data Tapes, 1960 to 2001, U.S. Mortality Volumes,
1930 to 1959, National Center for Health Statistics, Centers for Disease
Control and Prevention, with permission.)
CHARACTERISTICS OF HUMAN CANCER 11
cell type arises from the progeny of a tissue
‘‘stem cell’’ (one that still has a capacity for
renewal and is not yet fully differentiated),
which has been blocked or diverted in its
pathway to form a fully differentiated cell.
Examples of neoplasms that maintain a
modicumof differentiation include islet cell
tumors of the pancreas that still make insulin,
colonic adenocarcinoma cells that form
glandlike epithelial structures and secrete
mucin, and breast carcinomas that make
abortive attempts to form structures resembling
mammary gland ducts. Hormoneproducing
tumors, however, do not respond
to feedback controls regulating normal tissue
growth or to negative physiologic feedback
regulating hormonal secretion. For
example, an islet cell tumor may continue
to secrete insulin in the face of extreme
hypoglycemia, and an ectopic adrenocortiocotropic
hormone (ACTH)-producing lung
carcinomamay continue to produce ACTH
even though circulating levels of adrenocortical
steroids are sufficient to cause
Cushing’s syndrome (see Chapter 6).Many
malignant neoplasms, particularly the more
rapidly growing and invasive ones, only
vaguely resemble their normal counterpart
tissue structurally and functionally. They
are thus said to be ‘‘undifferentiated’’ or
‘‘poorly differentiated.’’
4. Malignant tumors usually, though not invariably,
grow more rapidly than benign
tumors. Once they reach a clinically detectable
stage,malignant tumors generally show
evidence of significant growth,with involvement
of surrounding tissue, over weeks or
months, whereas benign tumors often grow
slowly over several years.
Malignant neoplasms continue to grow even
in the face of starvation of the host. They press
on and invade surrounding tissues, often interrupting
vital functions; they metastasize to vital
organs, for example, brain, spine, and bone marrow,
compromising their functions; and they invade
blood vessels, causing bleeding. The most
common effects on the patient are cachexia
(extreme body wasting), hemorrhage, and infection.
About 50% of terminal patients die from
infection (see Chapter 8).
Differential diagnosis of cancer from a benign
tumor or a nonneoplastic disease usually involves
obtaining a tissue specimen by biopsy, surgical
excision, or exfoliative cytology. The latter is an
examination of cells obtained from swabbings,
washings, or secretions of a tissue suspected to
harbor cancer: the ‘‘Pap test’’ involves such an
examination.
CLASSIFICATION OF HUMAN
CANCERS
Although the terminology applied to neoplasms
can be confusing for a number of reasons, certain
generalizations can be made. The suffix oma,
applied by itself to a tissue type, usually indicates
a benign tumor. Some malignant neoplasms,
however, may be designated by the oma suffix
alone; these include lymphoma, melanoma, and
thymoma. Rarely, the oma suffix is used to describe
a nonneoplastic condition such as granuloma,
which is often not a true tumor, but a mass
of granulation tissue resulting from chronic inflammation
or abscess. Malignant tumors are
indicted by the terms carcinoma (epithelial in
origin) or sarcoma (mesenchymal in origin) preceded
by the histologic type and followed by the
tissue of origin. Examples of these include adenocarcinoma
of the breast, squamous cell carcinoma
of the lung, basal cell carcinoma of skin,
and leiomyosarcoma of the uterus. Most human
malignancies arise from epithelial tissue. Those
arising from stratified squamous epithelium are
designated squamous cell carcinomas, whereas
those emanating from glandular epithelium are
termed adenocarcinomas. When a malignant
tumor no longer resembles the tissue of origin, it
may be called anaplastic or undifferentiated. If a
tumor is metastatic from another tissue, it is
designated, for example, an adenocarcinoma of
the colon metastatic to liver. Some tumors arise
from pluripotential primitive cell types and may
contain several tissue elements. These include
mixed mesenchymal tumors of the uterus, which
containcarcinomatousandsarcomatous elements,
and teratocarcinomas of the ovary, which may
contain bone, cartilage, muscle, and glandular
epithelium.
Neoplasms of the hematopoietic system usually
have no benign counterparts. Hence the
12 CANCER BIOLOGY
terms leukemia and lymphoma always refer to a
malignant disease and have cell-type designations
such as acute or chronic myelogenous
leukemia, Hodgkin’s or non-Hodgkin’s lymphoma,
and so on. Similarly, the term melanoma
always refers to a malignant neoplasm derived
from melanocytes.
MACROSCOPIC AND MICROSCOPIC
FEATURES OF NEOPLASMS
The pathologist can gain valuable insights about
the nature of a neoplasm by careful examination
of the overall appearance of a surgical specimen.
Often, by integrating the clinical findings with
macroscopic characteristics of a tumor, a tentative
differential diagnosis can be reached.
Also, notation of whether the tumor is encapsulated,
has extended through tissue borders, or
reached to the margins of the excision provides
important diagnostic information.
The location of the anatomic site of the neoplasm
is important for several reasons. The site of
the tumor dictates several things about the clinical
course of the tumor, including (1) the likelihood
and route of metastatic spread, (2) the
effects of the tumor on body functions, and (3)
the type of treatment that can be employed. It is
also important to determine whether the observed
tumor mass is the primary site (i.e., tissue
of origin) of the tumor or a metastasis. A primary
epidermoid carcinoma of the lung, for example,
would be treated differently and have a different
prognosis than an embryonal carcinoma of the
testis metastatic to the lung. It is not always easy
to determine the primary site of a neoplasm,
particularly if the tumor cells are undifferentiated.
The first signs of a metastatic tumor may
be a mass in the lung noted on CT scan or a
spontaneous fracture of a vertebra that had been
invaded by cancer cells. Because the lungs and
bones are frequent sites of metastases for a variety
of tumors, the origin of the primary tumor
may not be readily evident. This is a very difficult
clinical situation, because to cure the patient
or to produce long-term remission, the
oncologist must be able to find and remove or
destroy the primary tumor to prevent its continued
growth and metastasis. If histologic examination
does not reveal the source of the
primary tumor, or if other diagnostic techniques
fail to reveal other tumor masses, the clinician
has to treat blindly, and thus might not choose
the best mode of therapy.
Another consideration is the accessibility of a
tumor. If a tumor is surgically inaccessible or too
close to vital organs to allow complete resection,
surgical removal is impossible. For example, a
cancer of the common bile duct or head of the
pancreas is often inoperable by the time it is
diagnosed because these tumors invade and attach
themselves to vital structures early, thus
preventing curative resection. Similarly, if administered
anticancer drugs cannot easily reach
the tumor site, as is the case with tumors growing
in the pleural cavity or in the brain, these agents
might not be able to penetrate in sufficient
quantities to kill the tumor cells.
The site of the primary tumor also frequently
determines the mode of, and target organs for,
metastatic spread. In addition to local spread,
cancers metastasize via lymphatic channels or
blood vessels. For example, carcinomas of the
lung most frequently metastasize to regional
lymph nodes, pleura, diaphragm, liver, bone,
kidneys, adrenals, brain, thyroid, and spleen.
Carcinomas of the colon metastasize to regional
lymph nodes, and by local extension, they ulcerate
and obstruct the gastrointestinal tract.
The most common site of distant metastasis of
colon carcinomas is the liver, via the portal vein,
which receives much of the venous return from
the colon and flows to the liver. Breast carcinomas
most frequently spread to axillary lymph
nodes, the opposite breast through lymphatic
channels,lungs,pleura,liver,bone,adrenals,brain,
and spleen.
Some tissues are more common sites of metastasis
than others. Because of their abundant
blood and lymphatic supply, as well as their
function as ‘‘filters’’ in the circulatory system,
the lungs and the liver are the most common
sites of metastasis from tumors occurring in
visceral organs. Metastasis is usually the single
most important criterion determining the patient’s
prognosis. In breast carcinoma, for example,
the 5-year survival rate for patients with
localized disease and no evidence of axillary
lymph node involvement is about 85%; but when
more than four axillary nodes are involved, the
5-year survival is about 30%, on average.3
CHARACTERISTICS OF HUMAN CANCER 13
The anatomic site of a tumor will also determine
its effect on vital functions. A lymphoma
growing in the mediastinum may press on major
blood vessels to produce the superior vena
caval syndrome, manifested by edema of the
neck and face, distention of veins of the neck,
chest, and upper extremities, headache, dizziness,
and fainting spells. Even a small tumor
growing in the brain can produce such dramatic
central nervous system effects as localized weakness,
sensory loss, aphasia, or epileptic-like seizures.
A lung tumor growing close to a major
bronchus will produce airway obstruction earlier
than one growing in the periphery of the
lung. A colon carcinoma may invade surrounding
muscle layers of the colon and constrict the
lumen, causing intestinal obstruction. One of the
frequent symptoms of prostatic cancer is inability
to urinate normally.
The cytologic criteria that enable the pathologist
to confirm the diagnosis, or at least to
suspect that cancer is present (thus indicating
the need for further diagnostic tests), are as
follows:
1. The morphology of cancer cells is usually
different from and more variable than that
of their counterpart normal cells from the
same tissue. Cancer cells are more variable
in size and shape.
2. The nucleus of cancer cells is often larger
and the chromatin more apparent (‘‘hyperchromatic’’)
than the nucleus in normal
cells; the nuclear-to-cytoplasmic ratio
is often higher; and the cancer cell nuclei
contain prominent, large nucleoli.
3. The number of cells undergoing mitosis is
usually greater in a population of cancer
cells than in a normal tissue population.
Twenty or more mitotic figures per 1000
cells would not be an uncommon finding
in cancerous tissue, whereas less than 1
per 1000 is usual for benign tumors or
normal tissue.4 This number, of course,
would be higher in normal tissues that
have a high growth rate, such as bone
marrow and crypt cells of the gastrointestinal
mucosa.
4. Abnormal mitosis and ‘‘giant cells,’’ with
large, pleomorphic (variable size and
shape) or multiple nuclei, are much more
common in malignant tissue than in normal
tissue.
5. Obvious evidence of invasion of normal
tissue by a neoplasm may be seen, indicating
that the tumor has already become
invasive and may have metastasized.
GRADE AND STAGE OF NEOPLASMS
Histologic Grade of Malignancy
The histologic grading of malignancy is based on
the degree of differentiation of a cancer and on
an estimate of the growth rate as indicated by
the mitotic index. It was generally believed that
less differentiated tumors were more aggressive
and more metastatic than more differentiated
tumors. It is now appreciated that this is an
oversimplification and, in fact, not a very accurate
way to assess the degree of malignancy for
certain kinds of tumors. However, for certain
epithelial tumors, such as carcinomas of the
cervix, uterine endometrium, colon, and thyroid,
histologic grading is a fairly accurate index
of malignancy and prognosis. In the case of
epidermoid carcinomas, for example, in which
keratinization occurs, keratin production provides
a relatively facile way to determine the
degree of differentiation. On the basis of this
criterion, and others like it, tumors have been
classified as grade I (75% to 100% differentiation),
grade II (50% to 75%), grade III (25% to
50%), and grade IV (0% to 25%).4 More recent
methods of malignancy grading also take into
consideration mitotic activity, amount of infiltration
into surrounding tissue, and amount of
stromal tissue in or around the tumor. The chief
value of grading is that it provides, for certain
cancers, a general guide to prognosis and an
indicator of the effectiveness of various therapeutic
approaches.
Tumor Staging
Although the classification of tumors based on
the preceding descriptive criteria helps the oncologist
determine the malignant potential of a
tumor, judge its probable course, and determine
the patient’s prognosis, a method of discovering
the extent of disease on a clinical basis and a
14 CANCER BIOLOGY
universal language to provide standardized criteria
among physicians are needed. Attempts to
develop an international language for describing
the extent of disease have been carried out by
two major agencies—the Union Internationale
Contre le Cancer (UICC) and the American Joint
Committee for Cancer Staging and End Results
Reporting (AJCCS). Some of the objectives of the
classification system developed by these groups
are (1) to aid oncologists in planning treatment;
(2) to provide categories for estimating prognosis
and evaluating results of treatment; and (3) to
facilitate exchange of information.5 Both the
UICC and AJCCS schemes use the T, N, M
classification system, in which T categories define
the primary tumor; N, the involvement of regional
lymph nodes; and M, the presence or absence
of metastases. The definition of extent of
malignant disease by these categories is termed
staging. Staging defines the extent of tumor
growth and progression at one point in time; four
different methods are involved:
1. Clinical staging: estimation of disease progression
based on physical examination,
clinical laboratory tests, X-ray films, and
endoscopic examination.
2. Tumor imaging: evaluation of progression
based on sophisticated radiography—for
example, CT scans, arteriography, lymphangiography,
and radioisotope scanning;
MRI; and PET.
3. Surgical staging: direct exploration of the
extent of the disease by surgical procedure.
4. Pathologic staging: use of biopsy procedures
to determine the degree of spread,
depth of invasion, and involvement of
lymph nodes.
These methods of staging are not used interchangeably,
and their use depends on agreedupon
procedures for each type of cancer. For
example, operative findings are used to stage certain
types of cancer (e.g., ovarian carcinomas) and
lymphangiography is required to stage Hodgkin’s
disease. Although this means that different staging
methods are used to stage different tumors,
each method is generally agreed on by oncologists,
thus allowing a comparison of data from
different clinical centers. Once a tumor is clinically
staged, it is not usually changed for that
patient; however, as more information becomes
available following a more extensive workup,
such as a biopsy or surgical exploration, this information
is, of course, taken into consideration
in determining treatment and estimating prognosis.
Staging provides a useful way to estimate at
the outset what a patient’s clinical course and
initial treatment should be. The actual course of
the disease indicates its true extent. As more is
learned about the natural history of cancers, and
as more sophisticated diagnostic techniques become
available, the criteria for staging will likely
change and staging should become more accurate
(see Chapter 7).
It is important to remember that staging does
not mean that any given cancer has a predictable,
ineluctable progression. Although some
tumors may progress in a stepwise fashion from
a small primary tumor to a larger primary tumor,
and then spread to regional nodes and distant
sites (i.e., progressing from stage I to stage IV),
others may spread to regional nodes or have
distant metastases while the primary tumor is
microscopic and clinically undetectable. Thus,
staging is somewhat arbitrary, and its effectiveness
is really based on whether it can be used as
a standard to select treatment and to predict the
course of disease.
Although the exact criteria used vary with
each organ site, the staging categories listed below
represent a useful generalization.6
Stage I (T1 N0 M0): Primary tumor is limited to
the organ of origin. There is no evidence of
nodal or vascular spread. The tumor can
usually be removed by surgical resection.
Long-term survival is from 70% to 90%.
Stage II (T2 N1 M0): Primary tumor has spread
into surrounding tissue and lymph nodes
immediately draining the area of the tumor
(‘‘first-station’’ lymph nodes). The tumor is
operable, but because of local spread, it may
not be completely resectable. Survival is 45%
to 55%.
Stage III (T3 N2 M0): Primary tumor is large,
with fixation to deeper structures. First-station
lymph nodes are involved; they may be more
than 3 cm in diameter and fixed to underlying
tissues. The tumor is not usually resectable,
and part of the tumor mass is left behind.
Survival is 15% to 25%.
CHARACTERISTICS OF HUMAN CANCER 15
Stage IV (T4 N3 Mþ): Extensive primary tumor
(may be more than 10cm in diameter) is present.
It has invaded underlying or surrounding
tissues. Extensive lymph node involvement has
occurred, and there is evidence of distant metastases
beyond the tissue of origin of the primary
tumor. Survival is under 5%.
The criteria for establishing lymph node involvement
(N categories) are based on size, firmness,
amount of invasion, mobility, number of
nodes involved, and distribution of nodes involved
(i.e., ipsilateral, contralateral, distant involvement):
N0 indicates that there is no evidence
of lymph node involvement; N1 indicates that
there are palpable lymph nodes with tumor involvement,
but they are usually small (2 to 3 cm
in diameter) and mobile; N2 indicates that there
are firm, hard, partially movable nodes (3 to 5cm
in diameter), partially invasive, and they may feel
as if they were matted together; N3 indicates that
there are large lymph nodes (over 5 cm in diameter)
with complete fixation and invasion into
adjacent tissues; N4 indicates extensive nodal involvement
of contralateral and distant nodes.
The criteria applied to metastases (M categories)
are as follows: M0, no evidence of metastasis;
M1, isolated metastasis in one other organ;
M2, multiple metastases confined to one organ,
with minimal functional impairment; M3, multiple
organs involved with no to moderate
functional impairment; M4, multiple organ involvement
with moderate to severe functional
impairment. Occasionally a subscript is used to
indicate the site of metastasis, such as Mp, Mh,
Mo for pulmonary, hepatic, and osseous metastases,
respectively.
Diagnostic procedures are getting more sophisticated
all the time. Improved CT, MRI, and
PET scanners, as well as ultrasound techniques,
are being developed to better localize tumors
and determine their metabolic rate. One can
visualize the day when ‘‘noninvasive biopsies,’’
based on the ability to carry out molecular and
cellular imaging by means of external detection
of internal signals, may at least partially replace
the need for biopsy or surgical specimens to get
diagnostic information (see Chapter 7). There
will always be the need, however, for clinical
pathologists to examine tissue specimens to confirm
noninvasive procedures, at least for the
foreseeable future. The ultimate diagnosis, prognosis,
and selection of a treatment course will
depend on this.
Although the TNMsystem is useful for staging
malignant tumors, it is primarily based on a temporal
model that assumes a delineated progression
over time from a small solitary lesion to
one that is locally invasive, then involves lymph
nodes, and finally spreads through the body.
While this is true for some cancers, the linearity
of this progression model is an oversimplification.
For example, some patients have aggressive
tumors almost from the outset and may die before
lymph node involvement becomes evident,
whereas others may have indolent tumors that
grow slowly and remain localized for a long time,
even though they may become large.
In addition, the TNM staging system does not
take into account the molecular markers that we
now know can more clearly define the status of a
cancer, e.g., its gene array and proteomic profiles
(see Chapter 7). Nor does the TNM system,
as a prognostic indicator, take into account the
varied responsiveness of tumors to various
therapeutic modalities. Thus, treatment choices
and prognostic estimates should be based more
on the molecular biology of the tumor than the
tumor’s size, location, or nodal status at the time
of diagnosis.7
References
1. A. Jemal, T. Murray, E. Ward, A. Samuels, R. C.
Tivari, A. Ghafoor, E. J. Feuer, and M. J. Thun:
Cancer statistics, 2005. CA Cancer J Clin 55:10,
2005.
2. P. A. Wingo, C. J. Cardinez, S. H. Landis, R. T.
Greenlee, A. G. Ries, R. N. Anderson, and M. J.
Thun: Long-term trends in cancer mortality in the
United States, 1930–1998. Cancer 97:3133, 2003.
3. I. C. Henderson and G. P. Canellos: Cancer of the
breast—The past decade. N Engl J Med 302:17,
1980.
4. S. Warren: Neoplasms. In W. A. D. Anderson, ed.:
Pathology. St. Louis: C. V. Mosby, 1961, pp. 441–
480.
5. P. Rubin: A unified classification of cancers: An
oncotaxonomy with symbols. Cancer 31:963, 1973.
6. P. Rubin: Statement of the clinical oncologic problem.
In P. Rubin, ed.: Clinical Oncology. Rochester:
American Cancer Society, 1974, pp. 1–25.
7. H. B. Burke: Outcome prediction and the future
of the TNM staging system. J Natl Cancer Inst
Cancer is a group of diseases of higher multicellular
organisms. It is characterized by alterations
in the expression of multiple genes, leading to
dysregulation of the normal cellular program for
cell division and cell differentiation. This results
in an imbalance of cell replication and cell death
that favors growth of a tumor cell population.
The characteristics that delineate a malignant
cancer from a benign tumor are the abilities to
invade locally, to spread to regional lymph nodes,
and to metastasize to distant organs in the body.
Clinically, cancer appears to be many different
diseases with different phenotypic characteristics.
As a cancerous growth progresses, genetic
drift in the cell population produces cell heterogeneity
in such characteristics as cell antigenicity,
invasiveness, metastatic potential, rate
of cell proliferation, differentiation state, and
response to chemotherapeutic agents. At the
molecular level, all cancers have several things
in common, which suggests that the ultimate
biochemical lesions leading to malignant transformation
and progression can be produced by a
common but not identical pattern of alterations
of gene readout. In general, malignant cancers
cause significant morbidity and will be lethal to
the host if not treated. Exceptions to this appear
to be latent, indolent cancers that may remain
clinically undetectable (or in situ), allowing the
host to have a standard life expectancy.
Some points in the description may not seem
intuitively obvious. For example, cancer doesn’t
just occur in humans, or just mammals for that
matter. Cancer (or at least tumorous growths—
these may or may not have been observed to
metastasize) has been observed in phyla as old as
Cnidaria, which appeared almost 600 million
years before the present, and in other ancient
phylasuchasEchinodermata(>500millionyears
old), Cephalopoda (500 million years old), Amphibia
(300 million years old), and Aves (150
million years old). Curiously, cancer has never
been seen (or at least reported) in a number of
phyla such as Nematoda, Tradigrada, and Rotifera.
It is intriguing to consider that these organisms
may have some protective mechanisms
that prevent them from getting tumors. If so, it
would be important to find out what these
mechanisms are.
One thing is clear, though, which is that
cancer is a disease of multicellular organisms.
This trait implies that there is something inherent
in the ability of cells to proliferate in
4 CANCER BIOLOGY
clumps or to differentiate into different cell
types and move around in the body to sites of
organogenesis that is key to the process of tumorigenesis.
Problems occur when these processes
become dysregulated.
One might also argue that evolution itself has
played some tricks on us because some of the
properties selected for may themselves be processes
that cancer cells use to become invasive
and metastatic. Or to phrase it differently: Is
cancer an inevitable result of a complex evolutionary
process that has advantages and disadvantages?
Some of these processes might be the
following:
1. The mechanism of cell invasiveness that
allows the implantation of the early embryo
into the uterine wall and the development
of a placenta.
2. Cell motility that allows neural cells, for
example, to migrate from the original neural
crest to form the nervous system.
3. The development of a large, complex genome
of up to 40,000 genes that must be
replicated perfectly every time a cell divides.
4. The large number of cells in a human or
higher mammal that must replicate and
differentiate nearly perfectly every time
(some can be destroyed if they become
abnormal).
5. The long life span of humans and higher
mammals, increasing the chance for a
genetic ‘‘hit’’ to occur and lead a cell down
a malignant path.
As we shall see in later chapters of this book,
cancer cells take advantage of a number of these
events and processes.
Other questions that arose at the gathering
above from scientists not in the field of cancer
were the following:
1. Is there a single trait or traits that all
cancer cells have?
2. How many genetic ‘‘hits’’ does it take to
make a cancer cell?
3. What kinds of genes are involved in these
hits?
These questions are all dealt with in later
chapters. Suffice it to say here that for a cell to
become cancerous or at least take the first steps
to becoming cancerous, at least two genetic hits
are required. One may be inherited and another
accrued after birth or both may be accrued after
birth (so-called somatic, or spontaneous, hits).
The kinds of genes involved are oncogenes,
which when activated lead to dysregulated cell
proliferation, and tumor suppressor genes, which
become inactivated or deleted, producing a loss
of the cell’s checks and balances controlling cell
proliferation and differentiation.
The single most common, if not universal,
trait that occurs in all cancers is genetic drift. or
the ability of cells to lose the stringent requirement
for preciseDNA replication and to acquire
the ability to undergo sequential progressive
changes in their genome, through mutations,
gene rearrangement, or gene deletion. This
has sometimes been called the acquisition of a
‘‘mutator phenotype.’’
WHAT SIGNIFICANT EVENTS HAVE
HAPPENED IN CANCER RESEARCH
IN THE LAST 25 YEARS?
As I was beginning to gather my thoughts for the
fourth edition of Cancer Biology, one of my
colleagues mentioned that he thought it would
be of interest to describe the significant things
that have happened in cancer biology in the
25 years since the first edition was published
(1981). Many things have happened since then,
of course, and everyone has their favorite list.
But looking back at the table of contents for the
first edition and at the outline for this edition,
several things struck me, as listed below.
1. Cancer susceptibility genes. In 1981 we
knew that familial clustering of some cancers
occurred, for example, with colon cancer,
but the genes involved in this hadn’t
been determined. The APC, BRCA-1,
BRCA-2, and p53 inherited mutations, for
example, were not known at that time. Research
in this area has identified a number
of genes involved in cancer susceptibility,
andwithmoderncloningtechniques,more
are identified every few months.
2. The techniques of modern molecular
biology were in their infancy at that time.
Polymerase chain reaction (PCR), DNA
CHARACTERISTICS OF HUMAN CANCER 5
microarrays, protein chips, and bioinformatics
were not terms in anybody’s
dictionary.
3. Genes involved in cancer initiation and
promotion were very poorly defined. Although
we knew that chemicals and irradiation
could damage DNA and initiate
cancer in animals and humans, the specific
genes altered were almost completely
unknown. We now know a lot about the
genes involved at various stages of a number
of cancers. For example, the work of
Bert Vogelstein and colleagues has defined
a pathway sometimes called the
‘‘Vogelgram’’ for the progression of colon
cancer (see Chapter 5). We knew that
DNArepairwas important and that heritable
conditions of defective DNA repair
(e.g., xeroderma pigmentosum) could lead
to cancer, but the ideas about the mechanisms
of DNA repair were primitive.
4. The identification of oncogenes didn’t
really start until the early 1980s. The src
gene was identified in 1976 by Stehelin
et al., and erb, myc, and myb oncogenes
were identified in the late 1970s, but this
was about the limit of our knowledge
(see Chapter 5).
5. The term tumor suppressor gene wasn’t
even coined until the early 1980s, although
their existence had been implied
from the cell fusion experiments of
Henry Harris, (Chapter 5) who showed
that if a normal cell was fused with a
malignant cell, the phenotype was usually
nonmalignant. The RB gene was the
first one cloned, in 1983 by Cavenee et al.
(Chapter 5) p53 was originally thought
of as an oncogene. It wasn’t realized until
1989 that wild-type p53 could actually
suppress malignant transformation. A
number of tumor suppressor genes have,
of course, been identified since then.
6. Starting in the 1970s, cell cycle checkpoints
were identified in yeast by Lee
Hartwell and colleagues, but the identification
of human homologs of these genes
didn’t occur until the late 1980s (seeChapter
4).
7. Tumor immunology was still poorly understood
in 1981—both the mechanism
of the immune response and the ability
to manipulate it with cytokines, activated
dendritic cells, and vaccines. Such manipulation
was not in the treatment armamentarium.
8. The first treatment of a patient with
gene therapy occurred in 1990. Several
gene therapy clinical trials for cancer are
under way and some gene therapy modalities
will likely be approved in the next
few years.
9. The viral etiology of cancer was still being
widely debated in 1981. The involvement
of Epstein-Barr virus in Burkitt’s
lymphoma and of hepatitis B virus in
liver cancer was becoming accepted, but
the role of viruses in these diseases and
in cervical cancer, Kaposis’ sarcoma, and
in certain T-cell lymphomas became
clearer much later.
10. Although some growth factors that affect
cancer cell replication, such as IGF-1
and IGF-2, FGF, NGF, PDGF, and
EGF, were known in 1981, knowledge
about their receptors and signal transduction
mechanisms was primitive indeed.
Tumor growth factor a was known
as sarcoma growth factor (SGF), and the
existence of its partner, TGF-b, was only
implied from what was thought to be
a contaminating HPLC peak from the
purification procedure. The explosion
of knowledge about signal transduction
mechanisms and how these pathways interact
has been a tremendous boon to our
understanding of how cells respond to
signals in their environment and communicate
with each other.
11. Knowledge about the regulation of gene
expression has greatly increased in the
past 25 years, on the basis of our current
information on the packaging of chromatin,
transcription factors, coinducers
and corepressors, and inhibitory RNA
(siRNA).
12. While not topics discussed in detail in
the earlier editions of Cancer Biology, advances
in diagnostic imaging such as magnetic
resonanceimaging (MRI), computed
tomography (CT), and positron emission
tomography (PET) have significantly im-
6 CANCER BIOLOGY
proved cancer diagnosis. Improved radiation
therapy, combined modality therapy,
bone marrow transplant, and supportive
care have also improved significantly.
BASIC FACTS ABOUT CANCER
Cancer is a complex family of diseases, and carcinogenesis,
the events that turn a normal cell in
the body into a cancer cell, is a complex multistep
process. From a clinical point of view, cancer
is a large group of diseases, perhaps up to a
hundred or more, that vary in their age of onset,
rate of growth, state of cellular differentiation,
diagnostic detectability, invasiveness, metastatic
potential, response to treatment, and prognosis.
From a molecular and cell biological point of
view, however, cancer may be a relatively small
number of diseases caused by similar molecular
defects in cell function resulting from common
types of alterations to a cell’s genes. Ultimately,
cancer is a disease of abnormal gene expression.
There are a number of mechanisms by which
this altered gene expression occurs. These mechanisms
may occur via a direct insult to DNA,
such as a gene mutation, translocation, amplification,
deletion, loss of heterozygosity, or via a
mechanism resulting from abnormal gene transcription
or translation. The overall result is an
imbalance of cell replication and cell death in
a tumor cell population that leads to an expansion
of tumor tissue. In normal tissues, cell proliferation
and cell loss are in a state of equilibrium.
Cancer is a leading cause of death in the
Western world. In the United States and a number
of European countries, cancer is the secondleading
killer after cardiovascular disease, although
in the United States since 1999 cancer
has surpassed heart disease as the number one
cause of death in people younger than 85.1 Over
1.3 million new cases of cancer occur in the
United States each year, not including basal cell
and squamous cell skin cancers, which add another
1 million cases annually. These skin cancers
are seldom fatal, do not usually metastasize, and
are curable with appropriate treatment, so they
are usually considered separately. Melanoma,
by contrast, is a type of skin cancer that is more
dangerous and can be fatal, so it is considered
with the others. The highest mortality rates are
seen with lung, colorectal, breast, and prostate
cancers (Fig. 1–1). Over 570,000 people die
each year in the United States from these and
other cancers. More people die of cancer in 1
year in the United States than the number of
people killed in all the wars in which the United
States was involved in the twentieth century
(Fig. 1–2).
In many cases the causes of cancer aren’t
clearly defined, but both external (e.g., environmental
chemicals and radiation) and internal
(e.g., immune system defects, genetic predisposition)
factors play a role (see Chapter 2). Clearly,
cigarette smoking is a major causative factor.
These causal factors may act together to initiate
(the initial genetic insult) and promote (stimulation
of growth of initiated cells) carcinogenesis.
Often 10 to 20 years may pass before an
initiated neoplastic cell grows into a clinically
detectable tumor.
Although cancer can occur at any age, it is
usually considered a disease of aging. The average
age at the time of diagnosis for cancer of
all sites is 67 years, and about 76% of all cancers
are diagnosed at age 55 or older. Although cancer
is relatively rare in children, it is the secondleading
cause of death in children ages 1–14. In
this age group leukemia is the most common
cause of death, but other cancers such as osteosarcoma,
neuroblastoma, Wilms’ tumor (a kidney
cancer), and lymphoma also occur.
Over eight million Americans alive today have
had some type of cancer. Of these, about half
are considered cured. It is estimated that about
one in three people now living will develop some
type of cancer.
There has been a steady rise in cancer death
rates in the United States during the past 75
years. However, the major reason why cancer
accounts for a higher proportion of deaths now
than it did in the past is that today more people
live long enough to get cancer, whereas earlier
in the twentieth century more people died of
infectious disease and other causes. For example,
in 1900 life expectancy was 46 years for men
and 48 years for women. By 2000, the expectancy
had risen to age 74 for men and age 80 for
women. Thus, even though the overall death
rates due to cancer have almost tripled since
1930 for men and gone up over 50% for women,
CHARACTERISTICS OF HUMAN CANCER 7
the age-adjusted cancer death rates in men have
only increased 54% in men and not at all for
women.2
The major increase has been in deaths due to
lung cancer. Thus, cigarette smoking is a highly
suspect culprit in the observed increases. In addition,
pollution, diet, and other lifestyle changes
may have contributed to this increase in cancer
mortality rates (Chapter 3). The mortality rates
for some cancers has decreased in the past 50
years (e.g., stomach, uterine cervix); however, the
mortality rates have been essentially flat for many
of the major cancers such as breast, colon, and
prostate, although 5-year survival rates have improved
for these cancers (see Chapter 3).
It is instructive to examine the trends in cancer
mortality over time to get some clues about
the causes of cancer. For males, lung cancer
remains the number one cancer killer (Fig. 1–3).
With a lag of about 20 years, its rise in mortality
parallels the increase in cigarette smoking
among men, which has an almost identical curve
starting in the early 1900s. Lung cancer mortality
rates for men have decreased somewhat
since 1990, and death rates for colorectal cancer
have dropped slightly in recent years, whereas
prostate cancer mortality has increased somewhat.
Stomach cancer mortality has dropped
significantly since the early 1900s, presumably
because of better methods of food preservation
Males
Prostate
Lung and Bronchus
Colon and Rectum
Urinary Bladder
Melanoma of the Skin
Non-Hodgkin Lymphoma
Kidney and Renal Pelvis
Leukemia
Oral Cavity and Pharynx
Pancreas
All Sites
Breast
Lung and Bronchus
Colon and Rectum
Uterine Corpus
Non-Hodgkin Lymphoma
Melanoma of the Skin
Ovary
Thyroid
Urinary Bladder
Pancreas
All Sites
Lung and Bronchus
Breast
Colon and Rectum
Ovary
Pancreas
Leukemia
Non-Hodgkin Lymphoma
Uterine Corpus
Multiple Myeloma
Brain and Other Nervous System
All Sites
Lung and Bronchus
Prostate
Colon and Rectum
Pancreas
Leukemia
Esophagus
Liver and Intrahepatic Bile Duct
Non-Hodgkin Lymphoma
Urinary Bladder
Kidney and Renal Pelvis
All Sites
232,090
93,010
71,820
47,010
33,580
29,070
22,490
19,640
19,100
16,100
710,040
33%
13%
10%
7%
5%
4%
3%
3%
3%
2%
100%
211,240
79,560
73,470
40,880
27,320
26,000
22,220
19,190
16,200
16,080
662,870
32%
12%
11%
6%
4%
4%
3%
3%
2%
2%
100%
73,020
40,410
25,750
16,210
15,980
10,030
9050
7310
5640
5480
275,000
27%
15%
10%
6%
6%
4%
3%
3%
2%
2%
100%
90,490
30,350
28,540
15,820
12,540
10,530
10,330
10,150
8970
8020
295,280
31%
10%
10%
5%
4%
4%
3%
3%
3%
3%
100%
Females
Males
Estimated Deaths
Estimated New Cases*
Females
Figure 1–1. Ten leading cancer types for estimated new cancer cases and
deaths, by sex, United States, 2005. *Excludes basal and squamous cell skin
cancers and in situ carcinoma except urinary bladder. Estimates are rounded
to the nearest 10. Percentage may not total 100% due to rounding. (From
American Cancer Society, Surveillance Research, 2005. CA Cancer J Clin
2005; 55:10–30, with permission.)
8 CANCER BIOLOGY
(e.g., better refrigeration, less addition of nitrate
and nitrate preservatives). Cancer of the gastroesophageal
junction, however, has risen significantly
in recent years, perhaps due to obesity
and increased incidence of gastric reflux into the
esophagus in the U.S. population.
Somewhat surprising, perhaps, is the fact that
lung cancer has overtaken breast cancer as the
number one cancer killer in women (Fig. 1–4).
This increase occurred in the late 1980s and, as
was the case for males, parallels the rise in the
percentage of women who smoke. Smoking
started to increase dramatically during World
War II. Rosie the Rivetter picked up some bad
male habits along with increased access to traditionally
male jobs.
Breast cancer mortality rates have remained
stubbornly stable, although a small decrease
(5%) has occurred since 1990. Uterine cancer
death rates have been going down, primarily
through earlier detection and treatment of cervical
cancer. Female colon cancer mortality has
been decreasing, but the reasons for this aren’t
clear. As in males, stomach cancer mortality in
women has been going down for many years.
The good news is that more and more people
are being cured of their cancers today. In the
1940s, for example, only one in four persons
diagnosed with cancer lived at least 5 years after
treatment; in the 1990s that figure rose to 40%.
When normal life expectancy is factored into
this calculation, the relative 5-year survival rate
is about 64% for all cancers taken together.1
Thus, the gain from 1 in 3 to 4 in 10 survivors
means that almost 100,000 people are alive now
who would have died from their disease in less
than 5 years if they had been living in the 1940s.
This progress is due to better diagnostic and
treatment techniques, many of which have come
about from our increasing knowledge of the
biology of the cancer cell.
HALLMARKS OF MALIGNANT
DISEASES
Malignant neoplasms or cancers have several
distinguishing features that enable the pathologist
or experimental cancer biologist to characterize
them as abnormal. The most common
600
500
400
300
200
100
0
0.01
Gulf Vietnam
Death in Thousands
Total Battle Deaths Each Year
WWI Korea WWII Cancer AIDS Murder
48 54
104
292
550
41 25
Figure 1–2. Total battle deaths from all wars with U.S. involvement in the
twentieth century, compared to number of deaths each year from cancer,
AIDS, and murder in the United States. (Personal communication from Don
Coffey, Johns Hopkins University, with permission.)
CHARACTERISTICS OF HUMAN CANCER 9
types of human neoplasms derive from epithelium,
that is, the cells covering internal or external
surfaces of the body. These cells have a supportive
stroma of blood vessels and connective
tissue. Malignant neoplasms may resemble normal
tissues, at least in the early phases of their
growth and development. Neoplastic cells can
develop in any tissue of the body that contains
cells capable of cell division. Though they may
grow fast or slowly, their growth rate frequently
exceeds that of the surrounding normal tissue.
This is not an invariant property, however, because
the rate of cell renewal in a number of
normal tissues (e.g., gastrointestinal tract epithelium,
bone marrow, and hair follicles) is as
rapid as that of a rapidly growing tumor.
The term neoplasm, meaning new growth, is
often used interchangeably with the term tumor
to signify a cancerous growth. It is important to
keep in mind, however, that tumors are of two
basic types: benign and malignant. The ability to
distinguish between benign and malignant tumors
is crucial in determining the appropriate
treatment and prognosis of a patient who has
a tumor. The following are features that differentiate
a malignant tumor from a benign tumor:
1. Malignant tumors invade and destroy adjacent
normal tissue; benign tumors grow
by expansion, are usually encapsulated,
and do not invade surrounding tissue.
Benign tumors may, however, push aside
30
40
50
60
Stomach
Rate per 100,000 Males
Leukemia
1930
1932
1934
1936
1938
1940
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Colon and Rectum
Liver
Year of Death
Pancreas
Prostate
Lung and Bronchus
70
80
90
100
20
10
0
Figure 1–3. Annual age-adjusted cancer death rates* among males for selected
cancer types, United States, 1930 to 2001. *Rates are age adjusted to
the 2000 U.S. standard population. Because of changes in ICD coding, numerator
information has changed over time rates for cancers of the lung and
bronchus, colon and rectum, and liver are affected by these changes. (From
U.S. Mortality Public Use Data Tapes, 1960 to 2001, U.S. Mortality Volumes,
1930 to 1959, National Center for Health Statistics, Centers for Disease
Control and Prevention, with permission.)
10 CANCER BIOLOGY
normal tissue and may become life threatening
if they press on nerves or blood
vessels or if they secrete biologically active
substances, such as hormones, that alter
normal homeostatic mechanisms.
2. Malignant tumorsmetastasize through lymphatic
channels or blood vessels to lymph
nodes and other tissues in the body. Benign
tumors remain localized and do not
metastasize.
3. Malignant tumor cells tend to be ‘‘anaplastic,’’
or less well differentiated than normal
cells of the tissue in which they arise. Benign
tumors usually resemble normal tissue
more closely than malignant tumors do.
Some malignant neoplastic cells at first
structurally and functionally resemble the
normal tissue in which they arise. Later, as
themalignant neoplasmprogresses, invades
surrounding tissues, and metastasizes, the
malignant cells may bear less resemblance
to the normal cell of origin. The development
of a less well-differentiated malignant
cell in a population of differentiated normal
cells is sometimes called dedifferentiation.
This term is probably a misnomer for the
process, because it implies that a differentiated
cell goes backwards in its developmental
process after carcinogenic insult. It
is more likely that the anaplastic malignant
30
25
15
5
40
35
50
45
60
55
65
Stomach
Rate per 100,000 Females
Ovary
1930
1932
1934
1936
1938
1940
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Colon and Rectum
Uterus
Year of Death
Pancreas
Breast
Lung and Bronchus
70
75
80
85
90
95
100
20
10
0
Figure 1–4. Annual age-adjusted cancer death rates* among females for
selected cancer types, United States, 1930 to 2001. *Rates are age adjusted to
the 2000 U.S. standard population. Because of ICD coding, numerator
information has changed over time, rates for cancers of the uterus, ovary,
lung and bronchus, and colon and rectum are affected by these changes.
Uterus cancers are for uterine cervix and uterine corpus combined. (From
U.S. Mortality Public Use Data Tapes, 1960 to 2001, U.S. Mortality Volumes,
1930 to 1959, National Center for Health Statistics, Centers for Disease
Control and Prevention, with permission.)
CHARACTERISTICS OF HUMAN CANCER 11
cell type arises from the progeny of a tissue
‘‘stem cell’’ (one that still has a capacity for
renewal and is not yet fully differentiated),
which has been blocked or diverted in its
pathway to form a fully differentiated cell.
Examples of neoplasms that maintain a
modicumof differentiation include islet cell
tumors of the pancreas that still make insulin,
colonic adenocarcinoma cells that form
glandlike epithelial structures and secrete
mucin, and breast carcinomas that make
abortive attempts to form structures resembling
mammary gland ducts. Hormoneproducing
tumors, however, do not respond
to feedback controls regulating normal tissue
growth or to negative physiologic feedback
regulating hormonal secretion. For
example, an islet cell tumor may continue
to secrete insulin in the face of extreme
hypoglycemia, and an ectopic adrenocortiocotropic
hormone (ACTH)-producing lung
carcinomamay continue to produce ACTH
even though circulating levels of adrenocortical
steroids are sufficient to cause
Cushing’s syndrome (see Chapter 6).Many
malignant neoplasms, particularly the more
rapidly growing and invasive ones, only
vaguely resemble their normal counterpart
tissue structurally and functionally. They
are thus said to be ‘‘undifferentiated’’ or
‘‘poorly differentiated.’’
4. Malignant tumors usually, though not invariably,
grow more rapidly than benign
tumors. Once they reach a clinically detectable
stage,malignant tumors generally show
evidence of significant growth,with involvement
of surrounding tissue, over weeks or
months, whereas benign tumors often grow
slowly over several years.
Malignant neoplasms continue to grow even
in the face of starvation of the host. They press
on and invade surrounding tissues, often interrupting
vital functions; they metastasize to vital
organs, for example, brain, spine, and bone marrow,
compromising their functions; and they invade
blood vessels, causing bleeding. The most
common effects on the patient are cachexia
(extreme body wasting), hemorrhage, and infection.
About 50% of terminal patients die from
infection (see Chapter 8).
Differential diagnosis of cancer from a benign
tumor or a nonneoplastic disease usually involves
obtaining a tissue specimen by biopsy, surgical
excision, or exfoliative cytology. The latter is an
examination of cells obtained from swabbings,
washings, or secretions of a tissue suspected to
harbor cancer: the ‘‘Pap test’’ involves such an
examination.
CLASSIFICATION OF HUMAN
CANCERS
Although the terminology applied to neoplasms
can be confusing for a number of reasons, certain
generalizations can be made. The suffix oma,
applied by itself to a tissue type, usually indicates
a benign tumor. Some malignant neoplasms,
however, may be designated by the oma suffix
alone; these include lymphoma, melanoma, and
thymoma. Rarely, the oma suffix is used to describe
a nonneoplastic condition such as granuloma,
which is often not a true tumor, but a mass
of granulation tissue resulting from chronic inflammation
or abscess. Malignant tumors are
indicted by the terms carcinoma (epithelial in
origin) or sarcoma (mesenchymal in origin) preceded
by the histologic type and followed by the
tissue of origin. Examples of these include adenocarcinoma
of the breast, squamous cell carcinoma
of the lung, basal cell carcinoma of skin,
and leiomyosarcoma of the uterus. Most human
malignancies arise from epithelial tissue. Those
arising from stratified squamous epithelium are
designated squamous cell carcinomas, whereas
those emanating from glandular epithelium are
termed adenocarcinomas. When a malignant
tumor no longer resembles the tissue of origin, it
may be called anaplastic or undifferentiated. If a
tumor is metastatic from another tissue, it is
designated, for example, an adenocarcinoma of
the colon metastatic to liver. Some tumors arise
from pluripotential primitive cell types and may
contain several tissue elements. These include
mixed mesenchymal tumors of the uterus, which
containcarcinomatousandsarcomatous elements,
and teratocarcinomas of the ovary, which may
contain bone, cartilage, muscle, and glandular
epithelium.
Neoplasms of the hematopoietic system usually
have no benign counterparts. Hence the
12 CANCER BIOLOGY
terms leukemia and lymphoma always refer to a
malignant disease and have cell-type designations
such as acute or chronic myelogenous
leukemia, Hodgkin’s or non-Hodgkin’s lymphoma,
and so on. Similarly, the term melanoma
always refers to a malignant neoplasm derived
from melanocytes.
MACROSCOPIC AND MICROSCOPIC
FEATURES OF NEOPLASMS
The pathologist can gain valuable insights about
the nature of a neoplasm by careful examination
of the overall appearance of a surgical specimen.
Often, by integrating the clinical findings with
macroscopic characteristics of a tumor, a tentative
differential diagnosis can be reached.
Also, notation of whether the tumor is encapsulated,
has extended through tissue borders, or
reached to the margins of the excision provides
important diagnostic information.
The location of the anatomic site of the neoplasm
is important for several reasons. The site of
the tumor dictates several things about the clinical
course of the tumor, including (1) the likelihood
and route of metastatic spread, (2) the
effects of the tumor on body functions, and (3)
the type of treatment that can be employed. It is
also important to determine whether the observed
tumor mass is the primary site (i.e., tissue
of origin) of the tumor or a metastasis. A primary
epidermoid carcinoma of the lung, for example,
would be treated differently and have a different
prognosis than an embryonal carcinoma of the
testis metastatic to the lung. It is not always easy
to determine the primary site of a neoplasm,
particularly if the tumor cells are undifferentiated.
The first signs of a metastatic tumor may
be a mass in the lung noted on CT scan or a
spontaneous fracture of a vertebra that had been
invaded by cancer cells. Because the lungs and
bones are frequent sites of metastases for a variety
of tumors, the origin of the primary tumor
may not be readily evident. This is a very difficult
clinical situation, because to cure the patient
or to produce long-term remission, the
oncologist must be able to find and remove or
destroy the primary tumor to prevent its continued
growth and metastasis. If histologic examination
does not reveal the source of the
primary tumor, or if other diagnostic techniques
fail to reveal other tumor masses, the clinician
has to treat blindly, and thus might not choose
the best mode of therapy.
Another consideration is the accessibility of a
tumor. If a tumor is surgically inaccessible or too
close to vital organs to allow complete resection,
surgical removal is impossible. For example, a
cancer of the common bile duct or head of the
pancreas is often inoperable by the time it is
diagnosed because these tumors invade and attach
themselves to vital structures early, thus
preventing curative resection. Similarly, if administered
anticancer drugs cannot easily reach
the tumor site, as is the case with tumors growing
in the pleural cavity or in the brain, these agents
might not be able to penetrate in sufficient
quantities to kill the tumor cells.
The site of the primary tumor also frequently
determines the mode of, and target organs for,
metastatic spread. In addition to local spread,
cancers metastasize via lymphatic channels or
blood vessels. For example, carcinomas of the
lung most frequently metastasize to regional
lymph nodes, pleura, diaphragm, liver, bone,
kidneys, adrenals, brain, thyroid, and spleen.
Carcinomas of the colon metastasize to regional
lymph nodes, and by local extension, they ulcerate
and obstruct the gastrointestinal tract.
The most common site of distant metastasis of
colon carcinomas is the liver, via the portal vein,
which receives much of the venous return from
the colon and flows to the liver. Breast carcinomas
most frequently spread to axillary lymph
nodes, the opposite breast through lymphatic
channels,lungs,pleura,liver,bone,adrenals,brain,
and spleen.
Some tissues are more common sites of metastasis
than others. Because of their abundant
blood and lymphatic supply, as well as their
function as ‘‘filters’’ in the circulatory system,
the lungs and the liver are the most common
sites of metastasis from tumors occurring in
visceral organs. Metastasis is usually the single
most important criterion determining the patient’s
prognosis. In breast carcinoma, for example,
the 5-year survival rate for patients with
localized disease and no evidence of axillary
lymph node involvement is about 85%; but when
more than four axillary nodes are involved, the
5-year survival is about 30%, on average.3
CHARACTERISTICS OF HUMAN CANCER 13
The anatomic site of a tumor will also determine
its effect on vital functions. A lymphoma
growing in the mediastinum may press on major
blood vessels to produce the superior vena
caval syndrome, manifested by edema of the
neck and face, distention of veins of the neck,
chest, and upper extremities, headache, dizziness,
and fainting spells. Even a small tumor
growing in the brain can produce such dramatic
central nervous system effects as localized weakness,
sensory loss, aphasia, or epileptic-like seizures.
A lung tumor growing close to a major
bronchus will produce airway obstruction earlier
than one growing in the periphery of the
lung. A colon carcinoma may invade surrounding
muscle layers of the colon and constrict the
lumen, causing intestinal obstruction. One of the
frequent symptoms of prostatic cancer is inability
to urinate normally.
The cytologic criteria that enable the pathologist
to confirm the diagnosis, or at least to
suspect that cancer is present (thus indicating
the need for further diagnostic tests), are as
follows:
1. The morphology of cancer cells is usually
different from and more variable than that
of their counterpart normal cells from the
same tissue. Cancer cells are more variable
in size and shape.
2. The nucleus of cancer cells is often larger
and the chromatin more apparent (‘‘hyperchromatic’’)
than the nucleus in normal
cells; the nuclear-to-cytoplasmic ratio
is often higher; and the cancer cell nuclei
contain prominent, large nucleoli.
3. The number of cells undergoing mitosis is
usually greater in a population of cancer
cells than in a normal tissue population.
Twenty or more mitotic figures per 1000
cells would not be an uncommon finding
in cancerous tissue, whereas less than 1
per 1000 is usual for benign tumors or
normal tissue.4 This number, of course,
would be higher in normal tissues that
have a high growth rate, such as bone
marrow and crypt cells of the gastrointestinal
mucosa.
4. Abnormal mitosis and ‘‘giant cells,’’ with
large, pleomorphic (variable size and
shape) or multiple nuclei, are much more
common in malignant tissue than in normal
tissue.
5. Obvious evidence of invasion of normal
tissue by a neoplasm may be seen, indicating
that the tumor has already become
invasive and may have metastasized.
GRADE AND STAGE OF NEOPLASMS
Histologic Grade of Malignancy
The histologic grading of malignancy is based on
the degree of differentiation of a cancer and on
an estimate of the growth rate as indicated by
the mitotic index. It was generally believed that
less differentiated tumors were more aggressive
and more metastatic than more differentiated
tumors. It is now appreciated that this is an
oversimplification and, in fact, not a very accurate
way to assess the degree of malignancy for
certain kinds of tumors. However, for certain
epithelial tumors, such as carcinomas of the
cervix, uterine endometrium, colon, and thyroid,
histologic grading is a fairly accurate index
of malignancy and prognosis. In the case of
epidermoid carcinomas, for example, in which
keratinization occurs, keratin production provides
a relatively facile way to determine the
degree of differentiation. On the basis of this
criterion, and others like it, tumors have been
classified as grade I (75% to 100% differentiation),
grade II (50% to 75%), grade III (25% to
50%), and grade IV (0% to 25%).4 More recent
methods of malignancy grading also take into
consideration mitotic activity, amount of infiltration
into surrounding tissue, and amount of
stromal tissue in or around the tumor. The chief
value of grading is that it provides, for certain
cancers, a general guide to prognosis and an
indicator of the effectiveness of various therapeutic
approaches.
Tumor Staging
Although the classification of tumors based on
the preceding descriptive criteria helps the oncologist
determine the malignant potential of a
tumor, judge its probable course, and determine
the patient’s prognosis, a method of discovering
the extent of disease on a clinical basis and a
14 CANCER BIOLOGY
universal language to provide standardized criteria
among physicians are needed. Attempts to
develop an international language for describing
the extent of disease have been carried out by
two major agencies—the Union Internationale
Contre le Cancer (UICC) and the American Joint
Committee for Cancer Staging and End Results
Reporting (AJCCS). Some of the objectives of the
classification system developed by these groups
are (1) to aid oncologists in planning treatment;
(2) to provide categories for estimating prognosis
and evaluating results of treatment; and (3) to
facilitate exchange of information.5 Both the
UICC and AJCCS schemes use the T, N, M
classification system, in which T categories define
the primary tumor; N, the involvement of regional
lymph nodes; and M, the presence or absence
of metastases. The definition of extent of
malignant disease by these categories is termed
staging. Staging defines the extent of tumor
growth and progression at one point in time; four
different methods are involved:
1. Clinical staging: estimation of disease progression
based on physical examination,
clinical laboratory tests, X-ray films, and
endoscopic examination.
2. Tumor imaging: evaluation of progression
based on sophisticated radiography—for
example, CT scans, arteriography, lymphangiography,
and radioisotope scanning;
MRI; and PET.
3. Surgical staging: direct exploration of the
extent of the disease by surgical procedure.
4. Pathologic staging: use of biopsy procedures
to determine the degree of spread,
depth of invasion, and involvement of
lymph nodes.
These methods of staging are not used interchangeably,
and their use depends on agreedupon
procedures for each type of cancer. For
example, operative findings are used to stage certain
types of cancer (e.g., ovarian carcinomas) and
lymphangiography is required to stage Hodgkin’s
disease. Although this means that different staging
methods are used to stage different tumors,
each method is generally agreed on by oncologists,
thus allowing a comparison of data from
different clinical centers. Once a tumor is clinically
staged, it is not usually changed for that
patient; however, as more information becomes
available following a more extensive workup,
such as a biopsy or surgical exploration, this information
is, of course, taken into consideration
in determining treatment and estimating prognosis.
Staging provides a useful way to estimate at
the outset what a patient’s clinical course and
initial treatment should be. The actual course of
the disease indicates its true extent. As more is
learned about the natural history of cancers, and
as more sophisticated diagnostic techniques become
available, the criteria for staging will likely
change and staging should become more accurate
(see Chapter 7).
It is important to remember that staging does
not mean that any given cancer has a predictable,
ineluctable progression. Although some
tumors may progress in a stepwise fashion from
a small primary tumor to a larger primary tumor,
and then spread to regional nodes and distant
sites (i.e., progressing from stage I to stage IV),
others may spread to regional nodes or have
distant metastases while the primary tumor is
microscopic and clinically undetectable. Thus,
staging is somewhat arbitrary, and its effectiveness
is really based on whether it can be used as
a standard to select treatment and to predict the
course of disease.
Although the exact criteria used vary with
each organ site, the staging categories listed below
represent a useful generalization.6
Stage I (T1 N0 M0): Primary tumor is limited to
the organ of origin. There is no evidence of
nodal or vascular spread. The tumor can
usually be removed by surgical resection.
Long-term survival is from 70% to 90%.
Stage II (T2 N1 M0): Primary tumor has spread
into surrounding tissue and lymph nodes
immediately draining the area of the tumor
(‘‘first-station’’ lymph nodes). The tumor is
operable, but because of local spread, it may
not be completely resectable. Survival is 45%
to 55%.
Stage III (T3 N2 M0): Primary tumor is large,
with fixation to deeper structures. First-station
lymph nodes are involved; they may be more
than 3 cm in diameter and fixed to underlying
tissues. The tumor is not usually resectable,
and part of the tumor mass is left behind.
Survival is 15% to 25%.
CHARACTERISTICS OF HUMAN CANCER 15
Stage IV (T4 N3 Mþ): Extensive primary tumor
(may be more than 10cm in diameter) is present.
It has invaded underlying or surrounding
tissues. Extensive lymph node involvement has
occurred, and there is evidence of distant metastases
beyond the tissue of origin of the primary
tumor. Survival is under 5%.
The criteria for establishing lymph node involvement
(N categories) are based on size, firmness,
amount of invasion, mobility, number of
nodes involved, and distribution of nodes involved
(i.e., ipsilateral, contralateral, distant involvement):
N0 indicates that there is no evidence
of lymph node involvement; N1 indicates that
there are palpable lymph nodes with tumor involvement,
but they are usually small (2 to 3 cm
in diameter) and mobile; N2 indicates that there
are firm, hard, partially movable nodes (3 to 5cm
in diameter), partially invasive, and they may feel
as if they were matted together; N3 indicates that
there are large lymph nodes (over 5 cm in diameter)
with complete fixation and invasion into
adjacent tissues; N4 indicates extensive nodal involvement
of contralateral and distant nodes.
The criteria applied to metastases (M categories)
are as follows: M0, no evidence of metastasis;
M1, isolated metastasis in one other organ;
M2, multiple metastases confined to one organ,
with minimal functional impairment; M3, multiple
organs involved with no to moderate
functional impairment; M4, multiple organ involvement
with moderate to severe functional
impairment. Occasionally a subscript is used to
indicate the site of metastasis, such as Mp, Mh,
Mo for pulmonary, hepatic, and osseous metastases,
respectively.
Diagnostic procedures are getting more sophisticated
all the time. Improved CT, MRI, and
PET scanners, as well as ultrasound techniques,
are being developed to better localize tumors
and determine their metabolic rate. One can
visualize the day when ‘‘noninvasive biopsies,’’
based on the ability to carry out molecular and
cellular imaging by means of external detection
of internal signals, may at least partially replace
the need for biopsy or surgical specimens to get
diagnostic information (see Chapter 7). There
will always be the need, however, for clinical
pathologists to examine tissue specimens to confirm
noninvasive procedures, at least for the
foreseeable future. The ultimate diagnosis, prognosis,
and selection of a treatment course will
depend on this.
Although the TNMsystem is useful for staging
malignant tumors, it is primarily based on a temporal
model that assumes a delineated progression
over time from a small solitary lesion to
one that is locally invasive, then involves lymph
nodes, and finally spreads through the body.
While this is true for some cancers, the linearity
of this progression model is an oversimplification.
For example, some patients have aggressive
tumors almost from the outset and may die before
lymph node involvement becomes evident,
whereas others may have indolent tumors that
grow slowly and remain localized for a long time,
even though they may become large.
In addition, the TNM staging system does not
take into account the molecular markers that we
now know can more clearly define the status of a
cancer, e.g., its gene array and proteomic profiles
(see Chapter 7). Nor does the TNM system,
as a prognostic indicator, take into account the
varied responsiveness of tumors to various
therapeutic modalities. Thus, treatment choices
and prognostic estimates should be based more
on the molecular biology of the tumor than the
tumor’s size, location, or nodal status at the time
of diagnosis.7
References
1. A. Jemal, T. Murray, E. Ward, A. Samuels, R. C.
Tivari, A. Ghafoor, E. J. Feuer, and M. J. Thun:
Cancer statistics, 2005. CA Cancer J Clin 55:10,
2005.
2. P. A. Wingo, C. J. Cardinez, S. H. Landis, R. T.
Greenlee, A. G. Ries, R. N. Anderson, and M. J.
Thun: Long-term trends in cancer mortality in the
United States, 1930–1998. Cancer 97:3133, 2003.
3. I. C. Henderson and G. P. Canellos: Cancer of the
breast—The past decade. N Engl J Med 302:17,
1980.
4. S. Warren: Neoplasms. In W. A. D. Anderson, ed.:
Pathology. St. Louis: C. V. Mosby, 1961, pp. 441–
480.
5. P. Rubin: A unified classification of cancers: An
oncotaxonomy with symbols. Cancer 31:963, 1973.
6. P. Rubin: Statement of the clinical oncologic problem.
In P. Rubin, ed.: Clinical Oncology. Rochester:
American Cancer Society, 1974, pp. 1–25.
7. H. B. Burke: Outcome prediction and the future
of the TNM staging system. J Natl Cancer Inst
