June 2006
ANATOMIC PATHOLOGY
Immunohistochemical
Classification of Amyloid in Surgical Pathology Revisited
Kebbel, Anja ; Rocken, Christoph
American Journal of Surgical
Pathology. 30(6):673-683, June 2006.
Podoplanin: A
Novel Diagnostic Immunohistochemical Marker
[New
Antibody/Techniques]
Podoplanin is a transmembrane mucoprotein recognized by the recently commercially available D2-40 monoclonal antibody. Recent investigations have shown that podoplanin is selectively expressed in lymphatic endothelium as well as lymphangiomas, Kaposi sarcomas, and in a subset of angiosarcomas with probable lymphatic differentiation. Podoplanin has also been shown to be strongly expressed in seminomas, epithelioid mesotheliomas, and hemangioblastomas, and immunostaining for this marker can assist in the diagnosis of these tumors. This article reviews the current information on the applications of podoplanin immunostaining in surgical pathology.
It is a 38 kd mucin-type transmembrane glycoprotein with extensive O-glycosylation and a high content of sialic acid that was first reported in lymphatic endothelial cells, epithelial cells of the choroid plexus, alveolar type I cells, osteoblasts, and peritoneal mesothelial cells, by Wetterwald et al in 1996, under the designation E11 antigen. This protein was subsequently identified on the surface of rat glomerular epithelial cells (podocytes) and because it was found to be involved in the flattening of foot processes in puromycin-induced nephrosis, it was named podoplanin Recent investigations have demonstrated that podoplanin, the so-called oncofetal M2A antigen expressed in testicular germ cell tumors that is recognized by the recently commercially available D2-40 antibody, and the type I alveolar cell marker hT1[alpha]-2 (also known as human Aggrus) are identical proteins.
Advances
in Anatomic Pathology, Volume 13(2), 83-88,2006
Current Prostate Biopsy Interpretation
Criteria
for Cancer, Atypical Small Acinar Proliferation,
High-Grade
Prostatic Intraepithelial Neoplasia,
and Use
of Immunostains
Context. —The past decade has brought major changes in
prostate biopsy sampling, interpretation, and reporting.
Objective. —To summarize current information on diagnostic
decision making, Gleason grading, ‘‘atypical’’ diagnoses, and use of
immunostaining.
Data Sources. —Pertinent literature from 1985 to 2005 is
reviewed, emphasizing recent findings.
Conclusions. —Diagnosis begins by evaluating a focus of
atypical single-cell layer lined acini according to the 3 minimal diagnostic
criteria for cancer: an infiltrative pattern, nuclear enlargement and
hyperchromasia, and prominent nucleoli. The Gleason score and linear extent or
percent of each core containing cancer should be reported. Atypical small
acinar proliferation suspicious for malignancy designates foci that have either
qualitative or quantitative limitations in atypia precluding a definite cancer
diagnosis. It has about a 3% incidence as an isolated finding. Contemporary
studies indicate a 39% predictive value for cancer on repeat biopsy. Isolated
high-grade prostatic intraepithelial neoplasia has a 3% to 14% incidence and
predicts cancer on repeat biopsy in 23% of cases. Immunostaining for a marker
of benign prostate (cytoplasmic keratin 34bE12 or nuclear p63) and a marker of
cancer (a-methylacyl coA racemase, clone P504S) may or may not resolve atypical
small acinar proliferation diagnoses. Performance of 34bE12 and P504S
immunostains resolved 76% of atypical small acinar proliferation diagnoses per
consensus of 3 urologic pathologists studied; a technical limitation is
preservation of the focus in question on the levels used for immunostaining.
The past decade has ushered in major changes in
prostate needle biopsy sampling, interpretation, and reporting: urologists are
sometimes using serum prostatespecific antigen (PSA) cutoffs of 2.5 mg/dL
rather than 4.0 mg/dL for needle biopsy1; awareness of the need
for greater sampling sensitivity has prompted an increase in the number of
needle cores from 1 to 2 per specimen to 6, then to 12 to 142;
the criteria to diagnose a minimal focus of cancer have been defined3;
numerous studies have been carried out on the importance of atypical small
acinar proliferation (ASAP) suspicious for malignancy4–10 and high-grade prostatic intraepithelial neoplasia (HGPIN)10,11; a-methylacyl coA racemase (P504S) has emerged as a positive
immunostain to diagnose cancer12–14; and p63, a nuclear stain, has joined basal
cell cytokeratin (34bE12) as an immunostain whose loss of reactivity suggests
cancer.15 Have better sampling, diagnostic criteria, and
immunostains eased our task and reduced ASAP diagnoses? Judging from the diagnostic
dilemmas my colleagues show me every few days, not totally! Thus, the goal of
this review is to highlight current knowledge in prostate biopsy interpretation
so that diagnoses can be made with greater ease, efficiency, standardization,
and appreciation of their clinical implications.
CANCER DIAGNOSIS: THE MINIMAL CRITERIA
A definite cancer
diagnosis has far-reaching implications for the patient such as prostatectomy
or ablative radiotherapy. Hence the decision to diagnose minimal cancer
(defined as involving ,5% of a biopsy core3) versus ASAP is most
consequential. Minimal cancer can assume numerous patterns of acini and
invasion, illustrated in a review by Thorson and Humphrey.16 Acini with a single cell layer usually fulfill 3 criteria as the sine
qua non for cancer: an infiltrative pattern of acini, nuclear enlargement and
hyperchromasia, and prominent nucleoli (Table 1).3 These judgments may be
facilitated by comparing the acini in question to adjacent, definitely benign
acini (Figure 1). It should be cautioned that these 3 main findings are
sometimes not obvious. The infiltrative pattern may not be readily apparent.
Larger acini with papillary infoldings or cystic dilatation can mimic benign
hyperplasia at low power, a finding with a 2% incidence in needle biopsies.17 Visualization of prominent nucleoli may require high illumination owing
to nuclear hyperchromasia. Prominent nucleoli have been reported in 94% of
cancers in an unselected biopsy series,18 and cases lacking
prominent nucleoli tend to have Gleason score $7. Minute Gleason score 8 to 10
cancers may have no prominent nucleoli in 45% of cases and have ‘‘frequent’’
nucleoli in only 11%.19 Total lack of nucleoli plus minimal nuclear
enlargement can easily allow tumor cells to mimic xanthoma cells or
lymphocytes.19 Certain histologic findings can favor cancer
but can be found in benign acini. Luminal blue mucin was seen in 33% of cancers
but in 6% of those called ASAP3 and in 63% with atypical adenomatous
hyperplasia.20 Extravasation of this acidic mucin into the
stroma can form collagenous micronodules. These micronodules were described in
13% of prostate cancers at prostatectomy,21 and their incidence in
needle biopsy has been reported as 0.6%21 to 2%16 to 5%.3 Collagenous micronodules were seen in 2% of
ASAP cases.3 They are essentially never seen in definitely
benign acini,21 making them a rare but specific feature for
cancer. Crystalloids occurred in 19% of cancers, 16% of ASAP cases,3 and
5% of benign biopsy specimens.22 Perineural indentation by benign acini is a
well-known phenomenon, but perineural space invasion, too, has been reported in
benign biopsies in 26 rare cases. In almost half these cases, basal cells were
not evident by immunostaining in the perineurally situated acini.23 Finally, unlike in most malignancies, the finding of mitotic figures in
a prostatic needle biopsy is rarely diagnostically useful, as they were found
in only 10% of all cancers3 and in 8% of high-grade cancers.18 However, in benign acini with postatrophic hyperplasia, a mimic of
cancer, mitotic figures were not seen.24 Gleason grading has
been reviewed elsewhere,25 but 2 points deserve emphasis. First, we have
observed a tendency of general (nonspecialized) pathologists to undergrade
cancer submitted for consultation.26
This tendency is greatest when
the tumor focus is minimal. While working on a busy consultation service, I
reviewed the submitting diagnoses of 1248 biopsies, of which 393 had been
graded. Gleason scores of 2 to 4 or descriptors of ‘‘well differentiated’’ or
‘‘low grade’’ were assigned in 221 (56%) of cases; we considered only 2% of
these cases well differentiated; most (57%) were moderately differentiated and
10% were poorly differentiated.26
Gleason grade 2 should be
assigned very rarely to needle biopsy material and grade 1 generally cannot be
assigned because grade 1 requires circumscription of the focus of cancer acini
(Figure 2), a feature that is visualized only on transurethral resection or
prostatectomy material. Even in transurethral resections, sampling the
transition zone where low-grade cancer is more common, only 8 (11%) of 88 cases
were diagnosed by us as low grade (Gleason score 4 or less). Low-grade cancer
is sufficiently rare on needle biopsies that some recommend not diagnosing it27;
I consider the diagnosis reasonable if based on strict criteria. Second, at the
other extreme of grading is Gleason pattern 5, which may contain sheets of
separate cells without acinus formation. Notably, any focus of cancer that
contains comedonecrosis, with ghost tumor cells visible in the lumen of the
acinus,25 qualifies as Gleason pattern 5 (Figure 3)
regardless of the cribriform, papillary, or other pattern of the surrounding
cells. Once cancer is diagnosed, attention must be paid to the format of
reporting it. Contemporary protocols have increased the number of cores sampled
from 6 to 8 (sextant biopsy) 28 to 8 to 14 (extended biopsy).2 For
precise communication of the extent of the tumor, each site sampled should be
reported separately, with its own Gleason score.29 A cancer diagnosis
should contain either the linear extent of cancer or the percent of
the core involved by cancer. The linear extent of carcinoma, the greatest
percent of carcinoma in a single core, and the total percent of carcinoma in
all cores are all significant predictors of total tumor volume at radical
prostatectomy as calculated by image analysis.30 The percent of tumor is
generally more convenient to estimate and gives the clinician a fair idea of
the tumor volume. However, in our study of needle biopsy cores from 1847
patients, the length of tissue sampled per biopsy, taking the sum of fragments,
averaged 10.8 6 2.7 mm (range, 3.0–27.5 mm). We documented a 7.6-fold variation
in needle core length, with a 3.6-fold variation among the middle 95% of core
lengths.31 This variation (in addition to significantly
affecting diagnostic sensitivity31) reduces the correlation between estimated
percentage of minimal cancer in biopsies and the measured extent of cancer;
thus, some recommend that actual length of cancer be reported instead of
percent.32 Current College of American Pathologists
recommendations are that at least the percent of the core biopsy involved by
tumor should be reported for all specimens.29 In addition, biopsy
samples aggregating to 5 mm or less probably warrant a comment about adequacy.31
Arch Pathol Lab Med. 2006;130:835–843
Clinical and
Pathologic Features of Fibroadenoma of the Mastopathic Type
Purpose Fibroadenoma with mastopathic change (FAM) is a relatively uncommon subtype of fibroadenoma of the breast, with a high incidence of pathological misdiagnosis. This histological subtype remains poorly understood because of its rarity. Many questions remain unanswered about its clinicopathological importance, especially in the differential diagnosis of breast cancers.
Methods Among 218 breast fibroadenomas surgically resected as excisional biopsies at authors institute between 1990 and 2004, 19 were pathologically diagnosed as FAM. Authors reviewed these 19 patients.
Results The ages of the patients ranged from 20 to 51 years (mean 36.8 years). The tumor sizes ranged from 0.8 to 7 cm (mean 2.1 cm). Six of the 19 patients underwent core needle biopsy, resulting in a diagnosis of fibroadenoma in four patients and atypical ductal hyperplasia in two patients. Ultrasonography showed findings suggestive of solid tubular carcinoma in seven patients, fibroadenoma in ten patients, and unspecific malignant tumors in two. They were not specified clinically.
Conclusion Recognition of this
distinctive variant of fibroadenoma is important because it resembles
intraductal carcinoma and is increasing in incidence. It is crucial to
distinguish FAM from intraductal carcinoma in biopsy specimens. Thus, not only
pathologists but also clinicians must be able to recognize this type of
fibroadenoma, and cooperate closely to establish an accurate diagnosis.
Surgery Today, Volume 36, Number
7 , July 2006, pg: 590 - 595
CLINICAL
PATHOLOGY
New Urine Test Helps Diagnose Prostate Cancer: Presented at AUA
By Fran Lowry
ATLANTA, G.A. -- June 1, 2006 -- A new urine test may help urologists decide
which men need to go for a biopsy for diagnosis of prostate cancer more
effectively than the standard prostate specific antigen (PSA) test, researchers
reported in a poster presented at the annual meeting of the American Urological
Association Annual Meeting (AUA)
The urine test, called Aptima®, measures levels of PCA3, a prostate-specific mRNA molecule that is upregulated 60- to 100-fold in prostate cancer cells and increases when prostate cancer cells are present.
"There is a well recognized need for prostate cancer diagnostic tests with improved specificity relative to current methods, particularly for men with one or more negative biopsies with raised serum PSA levels," said Yves Fradet, MD, professor of surgery and urology and director, department of surgery, Laval University, Quebec, Canada, during a presentation on May 21st.
Men with elevated serum PSA levels and at least 1 previous negative biopsy pose a diagnostic "PSA dilemma", according to Dr. Fradet. "The serum PSA assay has no diagnostic value for this population," he said.
To assess the ability of the PCA3 assay to predict prostate biopsy outcome compared with the serum PSA assay, Dr. Fradet and colleagues collected whole urine specimens from 225 men who were scheduled for prostate biopsy. All the men had serum PSA levels 2.5 ng/mL or greater and all had at least 1 prior negative biopsy.
All men underwent attentive digital rectal exam, which involves 3 sweeps on each side of the prostate. The urine was kept on ice and processed within 4 hours.
"The DRE was performed by applying firm pressure from the base to apex and from the lateral to the median line for each lobe, exactly 3 strokes per lobe," Dr. Fradet explained.
The amounts of PSA3 and serum PSA were measured in all urine samples and the results were correlated with the results of the prostate biopsies. The researchers reported that the PCA3 test had significantly greater predictive value than serum PSA, with a P value equal to.002.
The serum PSA test did not predict biopsy outcomes for this population, Dr. Fradet noted. The PCA3 assay yielded sensitivity of 58%, specificity of 74% and an odds ratio of 3.9 compared with a 17% specificity and odds ratio of 1.2 in men with serum PSA levels of 4.0 ng/mL, he said.
The researchers also found that higher the ratio of PCA3 to PSA mRNA, the greater the risk of a positive biopsy.
"The results of this research study indicate that the PCA3 assay could provide a tool to aid in decisions to biopsy the prostate in men with elevated serum PSA and previous negative biopsy results," he concluded.
Survivin: Role in Diagnosis,
Prognosis, and Treatment of Bladder Cancer.[Review]
Akhtar, Mohammed;
Gallagher, Lisa; Rohan, Stephen
Survivin belongs to a family of proteins, which serve as inhibitors of apoptosis. Survivin inhibits apoptosis by blocking activation of effector caspases in both extrinsic and intrinsic pathways of apoptosis. Expression of survivin has been demonstrated in several malignant neoplasms and is generally associated with adverse prognosis. In the case of bladder cancer, survivin is expressed in the neoplastic epithelium but not in the uninvolved mucosa. Several studies on bladder cancer have indicated that there may be a relationship between survivin expression and ultimate behavior of the carcinoma, although the exact nature of this relationship is still not fully understood, because the results of some of these studies seem to be contradictory. As survivin is differentially expressed in bladder cancer and not in the normal urothelium, several studies have demonstrated efficacy of urine testing of survivin as a diagnostic tool for an early detection of bladder cancer. Survivin has also been suggested as a suitable target for developing specific therapy for local treatment of bladder cancer with encouraging initial results. Thus, survivin is a potentially significant protein with a crucial role in the diagnosis, prognosis, and treatment of bladder cancer.
Advances in Anatomic Pathology.
13(3):122-126, May 2006.
Anti-CCP abs in the management and
pathogenesis of RA
Walther J.
van Venrooij and Dr. Erik Vossenaar
Attention has recently been drawn to a novel serological test for the diagnosis of rheumatoid arthritis (RA). This test, marketed by Axis-Shield Diagnostics (Scotland) and Euro-Diagnostica (The Netherlands), measures antibodies directed to cyclic citrullinated peptides (CCP), and is generally referred to in the scientific literature as the anti-CCP test.
Specificity and sensitivity of anti-CCP for RA
From a clinical point of view, the ideal serological marker for RA should not only be highly specific for the disease but also be able to distinguish RA from other arthritides that mimic RA. The marker should also be present in the majority of patients (good sensitivity). Furthermore, the marker should be present very early in the disease and have the ability to predict disease outcome. From the viewpoint of laboratory management, the serological marker should be detectable with a reproducible and easily performed test. Recent data from several laboratories indicate that the CCP2 system meets all these requirements. The first version of the CCP test (CCP1), although very specific for RA, was not very sensitive (40-60%; reviewed in. Following further development a second generation test (CCP2) was introduced in 2002.
This newer test was not only very specific for RA (97-98%), but also demonstrated good sensitivity (75-80%). As detailed in several recent reviews e.g. and publications, there is general agreement about the excellent diagnostic properties of the CCP2 test. Universally the performance of the CCP2 test has been reported to be superior to that of rheumatoid factor (RF), the antibody system that has been used for more than 50 years, and is currently the only serological parameter in the ACR criteria for the classification of RA. The high specificity of the CCP2 test enables it to distinguish RA from other rheumatic diseases such as systemic lupus erythematosus (SLE) or Sjögren's syndrome or other forms of joint disease. For example, patients with chronic hepatitis C virus (HCV) infection often display extra-hepatic manifestations among which arthropathy is common, affecting up to 20% of HCV-infected individuals. Since many HCV infected patients are RF positive, this test has limitations in discriminating HCV-related arthritis from RA. In two very recent studies, reviewed in, it was shown that in a group of randomly selected HCV patients none were positive for anti-CCP2, while more than 31% were positive for RF. These results further support the utility of the CCP2 test as a specific and reliable diagnostic tool for RA. The sensitivity of the CCP2 test for RA is comparable to the sensitivity of the IgM-RF test, generally around 75-80% [3, 5]. It is however important to appreciate that the apparent sensitivity is dependent upon the composition of the patient population. Typically, the sensitivity is high when the cohort is comprised of established RA patients, and lower when more early RA patients are included (by definition these cohorts will contain a relatively high proportion of patients with other forms of arthritis).
Anti-CCP is present early in disease
In healthy individuals the occurrence of anti-CCP antibodies is less than 1%. When a random population attending a rheumatology clinic were tested for anti- CCP, about 2-5% of the patients tested positive, but did not appear to have RA (author’s unpublished observations). A similar phenomenon was observed when the population of an early arthritis clinic was tested. Two recent studies have provided clear evidence that such supposedly "false positive" individuals might actually be in the process of developing RA. Rantapää-Dahlqvist and collaborators analysed blood samples from 83 blood donors who subsequently developed RA. These investigators reported that anti-CCP2 antibody could be detected in some patients 10 years before appearance of the first clinical symptoms. The percentage of anti-CCP2 positive individuals (25% were positive more than 1.5 years before onset of the first symptoms) increased sharply to 52% in the last 1.5 years before manifestation of the first symptoms of the disease. RFs were also detectable in the pre-disease serum samples, although at lower frequencies than anti-CCP antibodies (IgM-RF: 15% >1.5 yr, 30% =1.5 yr; IgG-RF: 12% and 27%; IgA-RF: 29% and 39%). More than 70% of the patients were anti-CCP2 positive at their first visit to the rheumatology clinic. In a similar type of study, Nielen and co-workers measured anti-CCP1 and IgM-RF levels in serial blood samples of 72 blood donors that later developed RA. Anti-CCP positivity could be observed up to 14 years before the first clinical symptoms and 41% of the patients were CCP1 positive at presentation to the clinician. For IgM-RF the corresponding parameters were 10 years and 28% positivity. Thus, anti-CCP detected more positive subjects and detected them longer before the start of the complaints compared to IgM-RF. The conclusion from these studies was that anti-CCP antibody is superior to RF in its ability to predict the development of RA. Other recent studies point to the same conclusion. Jansen and co-workers tried to discriminate RA from undifferentiated polyarthritis at presentation. They reported that the combined presence of IgM-RF and anti-CCP1 is able to predict which patients with early arthritis ultimately develop RA with a sensitivity of 55% and a specificity of close to 97%. A similar study using the CCP2 test was performed by van Gaalen and co-workers. In this study 318 patients out of 936 patients attending an early arthritis clinic could not be diagnosed as having RA at first presentation and were thus classified as undifferentiated arthritis (UA). After 3 years of follow-up 40% of the UA patients were clinically classified as RA. Of the UA patients that were negative for anti-CCP at baseline, 25% developed RA in 3 years. By contrast, 93% of the UA patients with a positive anti-CCP2 test at baseline, developed RA within 3 years (79% after 1 year), giving an odds ratio (OR) of 38. A very recent study of Vittecoq and co-workers studying 314 early arthritis patients confirmed these observations. In this study within 12 months of follow-up 90% of the patients that were CCP positive at baseline were classified as established RA patients. The conclusion from these studies is that anti-CCP antibodies are present early in disease, and that their presence is able to accurately predict the development of RA.
Prognostic ability of anti-CCP
It has been known for some time that IgM-RF antibodies
are able to predict joint erosions in RA patients. Several studies performed
with the relatively insensitive first generation CCP1 test also showed that the
presence of anti-CCP may predict erosive disease. Similar results have now been
obtained using the CCP2 test. Forslind and colleagues assessed anti-CCP2 in 379
early-RA patients and measured radiological joint damage and disease
progression after two years follow-up. They found that the presence of
anti-CCP2 at baseline was associated with significantly higher Larsen scores
both at baseline and at endpoint compared to RF and other disease parameters.
In another study, Kastbom and co-workers followed 242 patients with recent-onset
RA for three years. Their results showed that anti-CCP2 antibody had similar
diagnostic sensitivity to RF in early RA, but was superior as a predictor of
the disease course over three years. All these studies indicate that the
presence of anti- CCP antibodies may predict erosive disease. However, the
weight of evidence that anti-CCP antibody is present preferentially in patients
with erosive disease can not exclude the fact that this may not be the case in
all individual patients. How can we be sure that (erosive) RA is developing in
a CCP-positive individual without obvious clinical complaints? The answer to
this question is the Holy Grail in RA and certainly needs additional study.
Given the fact that RA is likely to be a multi-factorial disease, future
investigations will focus on combination models to identify individuals who
have the highest probability of developing RA. Arguably such an "RA
passport" should contain serological data (IgM-RF, OR of about 2 and anti-
CCP2, OR of about 25-39), genetic data (e.g. HLADR4, OR of about 2) and some
clinical parameters as suggested by Visser and co-workers. At this time it is
clear that given its high odds ratio, the measurement of anti-CCP2 antibody
adds an important parameter in such an RA passport.
Clinical Laboratory
International, May ,2006
MICROBIOLOGY
Mycobacterium abscessus: an emerging rapid-growing
potential pathogen.
Petrini, Bjorn
Mycobacterium abscessus is the most pathogenic and
chemotherapy-resistant rapid-growing mycobacterium. It is commonly associated
with contaminated traumatic skin wounds and with post-surgical soft tissue
infections. It is also one of the mycobacteria that are most often isolated
from cystic fibrosis patients. It is essential to differentiate this species
from the formerly indistinct "M. chelonae-complex", as chemotherapy
is especially difficult in M. abscessus. Clarithromycin or azithromycin are the
only regular oral antimycobacterial agents with an effect on M. abscessus, and
should preferably be supplemented with other drugs since long-term monotherapy
may cause resistance. Amikacin is a major parenteral drug against M. abscessus
that should also be given in combination with another drug. The recently
introduced drug tigecycline may prove to be an important addition to
chemotherapy, but has yet to be fully clinically evaluated as an
antimycobacterial agent. Surgery can be curative, or at least helpful, in the
healing of M. abscessus infection, and if conducted, it should include the
removal of all foreign or necrotic material. There is increasing awareness of
M. abscessus as an emerging pathogen.
APMIS. 114(5):319-328, May
2006.
BOTTOM LINE
Dissecting the pathologist’s brain: mental processes
that lead
to pathological diagnoses
Gianni Bussolati
On the interest of morphology
The specificity
of the discipline of pathological anatomy is the study of seats and causes of
diseases through morphological interpretation at a macroscopical and
microscopical level. Such interpretation and categorization of visible features
are being effected through mental processes, which are still poorly understood,
despite their interest. First of all their study might allow to reach a more
sound, objective and motivated selection of residents and applicants to the
discipline. We know very little about the characteristics which will make a
good pathologist, but tests devised for assessing visual memory might help in
the selection. We do not even know, for instance, if gender makes any
difference and if women are better fit for this speciality than men or vice
versa. Indeed, recent evidence shows that women perform better in some forms of
memory, and this might have to be taken in consideration. In addition, again
for the recruitment, we might find ways to better motivate students and applicants.
The second reason why this study is important is to improve reproducibility in
pathological classifications and diagnoses because we might thus be able to
select and focus our attention on those specific steps of the mental process
that are more prone to affect reproducibility. Other medicine specialists
involved in morphological interpretation such as radiologist have already been
pursuing an analysis of the process leading to interpretation of images.
Following the percourse of perception and decision making during morphological
interpretation, they concluded that long decision times work against
performance. It is worth mentioning that this might apply as well to
cytological reading of pap smears. The third reason, which makes an analysis worthwhile,
is that we might find a way to improve the image of pathology. Morphology is
presently regarded as being too subjective, not being based on numbers, and we
have to acknowledge that its ranking among scientific disciplines is not high
despite the legacy of morphological sciences to medicine and their great impact
on the process of medical evolution.
Sources
The feasibility of such study is presently linked to indirect information because no studies were so far specifically addressed to clarify the process leading pathologists to formulate diagnoses. As a result, we have to derive suggestions and information from various alternative sources. One source is philosophy and psychology. Since Aristotle, from the very beginning of philosophy, human beings have been trying to understand how form is interpreted. In the nineteenth century, mainly German psychologists and philosophers worked on the “Gesthalt theorie,” the theory of the form, and they have been trying to establish laws whereby people can understand the forms. For instance, Wertheimer’s seventh law establishes the principle that the past experience addresses to the correct interpretation. A clarifying example comes from Kanizsa’s figure (Fig. 1), indicating that our mind does not only consider what is present but what is missing as well.
This
applies so well to the interpretation of histopathological slides: what you do
not see has to be taken into consideration, and what is missing is sometimes
more important than what you actually see in the microscopic figure. A
considerable source of information are art critics, which have been trying to
understand how people look at pictures and perceive art. What are the mental
mechanisms that make us appreciate a masterpiece? Martin Kemp has been
publishing about this subject, and in recent times, Semir Zeki, a
neurophysiologist, has been trying to explore the relationship between art and
the brain from a neurophysiological point of view. He came out with a very
interesting theory which might apply as well to the interpretation of
histological slides or of pathological lesions in general: the theory of
multistage integration of visual consciousness. Visual brain does not process
images as a whole, but through several multistage processing systems. Each
system is specialized in a given attribute such as colour or motion or form and
so on, working either in parallel or in sequence, finally leading to the
interpretation (diagnosis). Support to this theory comes from
neurophysiological investigations, which, in recent times, have been providing
very important information, especially through functional magnetic resonance
images (fMRIs) showing for instance that in right-handed people observing faces
rather than objects, it is the medial temporal lobe which is involved and
activated. Category-specific breakdowns provide some information: already in
the past, it was known that ischaemic infarcts in the medial temporal region of
the right hemisphere cause a failure in the rapid recognition of familiar
faces, and that patients with bilateral insult to the medial temporal lobe
cannot transform present experience into future conscious recollection.
Radiologists have been trying to understand how expert radiologist process
X-ray images, and it appears that they process such images in the way that we
all process faces: by quickly devoting processing resources to features that
distinguish one stimulus from another. Such process is likely to apply as well
to the morphological interpretation of histological slides.
Analysis of the mental process
When trying to analyse the mental processes leading to interpretation
of images in pathology, we have to select and analyse different and sequential
steps. Four steps can be identified in sequence: (1) to look, (2) to see, (3)
to recognize and (4) to understand. The first step, that of looking, is easily
understood because of course, the stimulus goes from the eye to the calcarine
sulcus in the occipital lobe. Thereafter, the stimulus is transferred to other
areas devoted to the interpretation of colour, form and faces, mainly located
in the fusiform area and in the medialtemporal lobe. The second step, that of
seeing, is deeply intermingled with the understanding process. The art critic
Martin Kemp observed that “content and recognition are inextricably involved in
all our visual acts, since in no occasion you have seeing without meaning,
without recollecting”. So let us consider the step of seeing together with that
of recognizing. A good example is the following. In 1609, a British astronomer,
Thomas Harriot, looked at the moon with his telescope. What he saw and what he
described was a sphere, dark and bright areas being separated by an irregular
line. He gave no other information or interpretation. A few weeks later Galileo
Galilei looked at the moon with his telescope, and what he saw and described in
his book Sidereus Nuncius (Star Announcement) were mountains and valleys and
seas and craters and so on. He saw an irregular line and he could interpret it.
Why? Most probably because Galileo had been growing in the atmosphere of
Italian Renaissance at the time when the laws of perspective has been
discovered and applied to paintings, so he could understand what the
irregularities meant and he could interpret them. The process of seeing and that
of recognition are strictly linked to the fusiform gyrus, as shown by fMRI, and
confirmative evidence comes from the observation that such nucleus is
selectively activated by faces. Specifically, if we analyse the diagnostic
process of pathological interpretation, we can subdivide different sequences:
first of all, a visual search of the image, then interpretation of the
perceived information and finally, a combination of the collected information.
Such process has been analysed and experimentally investigated by Tiersma and
coworkers. The experiment was carried out using a helmet endowed with a beam
that could follow the direction of the eye and the gazing in different points
so that such a scan eye tracking system was enabling the visualisation of the eye
points of gaze of a subject and thus provided the possibilities to create
computerised graphic representations of scanning patterns. This study indicates
that the well-known (almost obvious) fact that experienced pathologists reach
diagnoses (i.e. categorization of the lesions) in a shorter time is related to
their focusing on the most significant areas. Moreover, these processes
(searching, focusing and moving) constitute the unconscious phase, to be
followed and completed by the final, conscious stage of diagnosis. We have to
conclude that the diagnostic process is not just a sequence of three steps
(search, interpretation and association), but such steps have to be repeated
several times back and forth, again and again, till the process reaches the conscious
state. We have to conclude that while making a pathological diagnosis, we
experience first an unconscious stage and then a conscious state. The
unconscious stage is fast and laborious: in a few seconds, the eye goes here
and there, interpreting the different areas and images. Thereafter, the process
reaches consciousness, and the diagnosis is made. Such sequence had already
been investigated by radiologist who described the steps of visual scanning,
pattern recognition and decision making as the sequence of events taking place
during pulmonary nodule detection. They also concluded on the influence of
expertise on X-ray image processing. Following the analysis of the unconscious
steps of looking, seeing and recognizing, we have to take into consideration
the last step, the conscious one, that of understanding. This stage is most
probably linked to the pre-frontal area, as shown by the recent observation by
fMRI that when people look at an image, at a painting for instance, and then
reach the conclusion that they like it, the pre-frontal area is selectively
activated. We have to conclude that the process of looking is linked to the
calcarine sulcus, that of seeing and recognizing to the medial temporal lobe,
to hippocampus and amygdala, fusiform gyrus and parahippocampus gyrus, and that
finally, when pathologists reach the diagnosis, it is the pre-frontal area that
is selectively involved (Fig. 2).
On
the complexity of pathology.
We have to acknowledge that the analytical processes leading to the
interpretation of pathological images are not comprehensive of the whole
complexity of pathology. The processes relate to pattern recognition, which
represents a basic part of pathology, but not the whole of it. Let us now
consider the factors that seem essential for forming a good pathologist. Good
sight is important, but “we see what we know”, so culture is important.
Knowledge of previous discoveries and observations is fundamental in our
discipline, but knowledge is to remember. Memory and experience are extremely
important in pathology. Now, what exactly is memory? The study of memory and of
its different forms was conducted by several philosophers and psychologists,
but mainly, I’m referring to a philosopher, Henry Bergson, whose main
contribution was related to the unravelling of the different types of memory.
He lived and worked in Paris, at the beginning of the twentieth century, in the
same town and at the same time when Marcel Proust, who was a relative of
Bergson, was engaged in the Recherche du Temp Perdu. According to Bergson, we
can subdivide a memory by repetition, and that of recollection/ recording via
intensity, via affection. This represents a common experience for pathologists,
who can remember cases they observed years before, because they were so
impressed and interested so that every single particular is vividly remembered.
From a neurophysiological point of view, during the memorizing process, by
repetition, the area of hippocampus is directly stimulated, while in the
process, via intensity, the hippocampus is indirectly stimulated through the
area of the amygdala. However, not everything is known about memory and its
mechanism. A recent study shows that if you give a stimulus, its performance
deteriorates during the day, but if you have a short nap, then the performance
improves, and after a good sleep, it improves again and again till it reaches a
plateau. Our experience as pathologists teaches us something else. Sometimes,
not only memory is consoli- dated, but even recollection is enhanced by sleep.
We look at a peculiar case and we cannot give a proper interpretation, then we
go back home, and the day after, the solution immediately is obvious. Moreover,
a recent observation indicates that some neurones in the medial temporal lobe
are selectively activated by strikingly different pictures of given objects,
and it cannot be excluded that single neurones are dedicated to a given face or
image. Such observations indicate that we are far from having a complete and
well-assessed knowledge of the extremely complex and fascinating process of
memory, a process of fundamental importance in the process of diagnosis of
pathological lesions. Besides good sight, culture, memory and experience, of
course, technique is important for pathologists, and of course, a school which
gives a method, a system. But a list of the different components of the
characteristics which make up a good pathologist would not be complete without
taking into consideration a special and dedicated ability, a gift or a knack,
what John Azzopardi used to call “oculus patologicus,” the pathological eye,
the ability to select and recognize 
morphological subtleties (Fig. 3). This was
also the opinion of a German-Swiss pathologist, Paul Ernst, who subdivided his
students into three groups: the great majority unable to see what is shown,
then a relatively small, but valuable group, able to see and understand what is
shown, and then a minute minority, who can even see what is not shown.
On the essence of pathology
The priming factor, the most important component for making up a good
pathologist, is still failing to our search, but we should agree that among the
different factors, three characteristics are really essential for forming a
good pathologist: the “eye,” the experience and finally, the school. These are
clearly the pillars that make up a good pathologist. When I presented these
three alternatives to colleagues collected at the Meeting along the way to
Santiago (May 2004) and asked the colleagues’ opinion, the audience divided into
three groups. Some people said that the school is the most important factor,
somebody else said that the eye, sort of an inborn ability, is the most
valuable one and then, of course, experience is fundamental. The only
dissenting opinion was expressed by Francisco Gonzalez-Crussi, a prominent
pathologist, author of several scientific papers and 
a successful writer of rather hectic books
such as On the nature of things erotic. He came out with the observation that
“Love,” love for pathology is the priming factor. This opinion has merits and
reasons because in fact, memory is stimulated, motivated by interest, by
affection, by love. What is the eye, if not a total dedication to morphology,
faith in morphology. Finally, the school, in its ultimate essence, is nothing
but the transfer of an interest, the transfer of a passion, the transfer
through generations of a small flame, the tantalizing but vital flame of
curiosity. However, while searching for a proper term to indicate dedicated
interest and passion, the English term “love” seems inappropriate, being
perhaps too abused and uncommitted. A better term is the one proposed by Karl
Lennert: he used the term “freude an der morphologie,” enjoyment for
morphology. Such inspiring German expression is inextricably linked to
Schiller’s Hymn to Joy and of course, to the music from Beethoven’s 9th
Symphony. So let us conclude that by dissecting a pathologist’s brain,
ultimately, what you find is a dedicated heart (Fig. 4).
Virchows
Arch (2006) 448: 739–743