Practice Evolution
Decentralized Computer-Assisted Immunohistochemical Image Analysis
Richard C. Friedberg, MD, PhD; Liron Pantanowitz, MD
I n 2007, the Futurescape of Anatomic Pathology meeting
provided plenty of discussion about the technologic fu-
ture of the field. Toward the end of that meeting, a ‘‘Re-
KEY TREND NO. 1—THE EVOLUTION OF ANATOMIC
PATHOLOGY ALONG CLINICAL PATHOLOGY LINES
The first, relevant, long-term key trend for anatomic pa-
ality Check’’ discussion provided a critique of the meeting thology is its evolution along clinical pathology lines. Spe-
with respect to the day-to-day activities of practicing pa- cifically, anatomic pathology is moving away from its tra-
thologists.1 At this year’s conference, we are providing an ditional teaching, which indicates that a particular image
update on our institution’s attempts to apply some of the is diagnostic of a particular disease primarily because the
tools and techniques presented here last year. Some tools viewing pathologist was once instructed by a more emi-
and techniques have worked, and some have not. The in- nent pathologist that similar images were diagnostic of
troduction of new technology often involves a steep learn- that condition. This mode of knowledge transfer is akin
ing curve. Because these tools and technologies are not to a guild model in which an entry-level artisan is ap-
simply ‘‘plug-and-play,’’ pathologists, technologists, and prenticed to and taught by a more senior artisan who is
administrators must recognize that we cannot simply deemed an expert in the art of the craft. The fundamental
adopt and implement them. We must validate, verify, and techniques based on hematoxylin, eosin, paraffin, and
make operational these technologies, which are indeed formaldehyde have undergone minimal changes during
works-in-progress. the past century. In recent decades, however, progress has
The internal environment at Baystate Health (Spring- been made, and pathologists are using newer stains and
field, Massachusetts) is well suited to implementing new tools, such as immunohistochemistry (IHC) and fluores-
technologies. We are a subspecialty-focused and academ- cent in situ hybridization, to analyze formalin-fixed, par-
ic-private practice and hybrid model, including extensive affin-embedded tissue. Recent trends are clearly headed
toward a more quantitative, reproducible, validated, spe-
anatomic, clinical, and molecular pathology sections inte-
cific, and reliable approach.
grated into 1 department. Annually, the department pro-
Image analysis in anatomic pathology has already start-
cesses 40 000 to 50 000 surgical specimens and 6 to 7 mil-
ed shifting from qualitative (eg, positive or negative), to
lion laboratory tests. We also have an average of 1 new semiquantitative (eg, 0 , 1 , 2 , 3 ), to even more quan-
breast cancer case per day. Thus, we have ample raw ma- titative in some areas (eg, copies per cell). Clinical pa-
terial to test the emerging tools and technologies. thology underwent this same transition many years ago.
Our external environment is much the same as it is for all Curiously, a brief review of the history of IHC and en-
of pathology: increased technologic innovation coupled with zyme-linked immunosorbent assays shows that both start-
increased biologic information in a background of increased ed around the same time as methods of using an antibody
demands from patients, payors, and providers. All of us are to identify an antigen. With IHC, the antigen expression
dealing with more and more information to ingest, digest, was identified by using an antibody-linked enzyme, such
and synthesize. From a high-level perspective, these in- as alkaline phosphatase, to generate color at the site of the
creased clinical demands are leading to a convergence of 2 bound antigen. Under the microscope, the presence or ab-
independent, long-term trends in anatomic pathology; these sence of the generated color was assessed by an anatomic
trends are evident in the application of computer-aided im- pathologist to determine whether the stain was positive
age analysis (CAIA) in cases of breast cancer. or negative at the location in question. Determining the
adequacy of the ‘‘darkness’’ (ie, immunoreactivity) of the
stain and the location within the architecture of the cell or
tissue became fundamental skills of the pathologist. For
Accepted for publication October 1, 2008. years, this qualitative approach was deemed diagnostical-
From the Departments of Pathology (Dr Friedberg) and Pathology
Informatics (Dr Pantanowitz), Baystate Health, Springfield, Massachu- ly sufficient. In time, the strength of the staining became
setts; and Tufts University School of Medicine, Boston, Massachusetts clinically significant, and therefore, anatomic pathologists
(Drs Friedberg and Pantanowitz). typically began to grade staining intensity on a scale from
The authors have no relevant financial interest in the products or 0 to 3 . As in the classic example of HER2/neu, such
companies described in this article. semiquantitative assessments may be inadequate and sub-
Presented at the College of American Pathologists Futurescape of