close

Вход

Забыли?

вход по аккаунту

?

346

код для вставкиСкачать
796
Diagnostic Discrepancies and Their Clinical Impact in
a Neuropathology Referral Practice
Janet M. Bruner, M.D.
Lila Inouye, M.D.
Gregory N. Fuller, M.D., Ph.D.
Lauren A. Langford, M.D., Dr.med
BACKGROUND. During the course of their neuropathology practice, the authors
received cases to review in consultation. In some cases, the patients came to
the authors’ hospital for therapy; in others, the primary pathologists requested a
consultation. Because changes in diagnosis might significantly alter patient management, protocol entry, care costs, or the potential for physician liability, the
Department of Pathology, The University of
Texas M. D. Anderson Cancer Center, Houston,
Texas.
authors determined the frequency and degrees of their disagreements with the
original diagnoses submitted to them.
METHODS. The authors reviewed the first 500 brain or spinal cord biopsy cases
that were submitted to their neuropathology consultation service for a second
opinion in 1995. Disagreements were coded into 10 categories, but were grouped
for this analysis as follows: serious (having immediate significance for therapy or
intervention), less serious but potentially substantial (calling for a change in type
or grade of glioma), minor (adding or deleting information), and those in which
the authors made the first diagnosis themselves.
RESULTS. There was some degree of disagreement between the original and review
diagnoses in 214 (42.8%) of the 500 cases. Disagreements were counted as serious
in 44 cases (8.8%), less serious but substantial in 96 cases (19.2%), and minor in
50 cases (10.0%); the authors made the first diagnosis in 24 cases (4.8%).
CONCLUSIONS. Clinically important diagnostic errors that can affect immediate
patient care decisions occur in a substantial number of brain and spinal cord
biopsy cases. Thus, seeking expert neuropathology consultation is prudent and
cost-effective for pathologists who are less experienced with these types of cases.
Cost savings in case management might result from confirmation of diagnosis
before definitive therapy is administered to the patients. The rates of discrepancy
between original diagnoses and second opinions in other subspecialties of pathology should be examined. [See editorial on pages 665–7, this issue.] Cancer 1997;
79:796–803. q 1997 American Cancer Society.
KEYWORDS: second opinion, brain biopsy, pathology, glioma, brain tumor.
Presented at the United States–Canadian Academy of Pathology Annual Meeting, Washington,
DC, March 23–26, 1996.
Supported by a grant from the Physicians Referral Service of the University of Texas M. D. Anderson Cancer Center.
Address for reprints: Janet M. Bruner, M.D.,
Neuropathology Section, Room B4.4808, Department of Pathology, Box 85, M. D. Anderson
Cancer Center, 1515 Holcombe Blvd., Houston,
TX 77030.
Received July 19, 1996; revision received September 18, 1996; accepted September 23, 1996.
I
n an era of increasing limitation of health care dollars, expansion
of managed health care, and decreasing reimbursement for medical
services, the primary goal of medical care in the United States must
continue to be safeguarding patients’ well-being and quality of life
while maximizing the efficiency of medical treatment and the conservation of resources. As a result, much debate has focused on the
issue of the necessity of referring patients or pathologic material for
consultation by specialists. For patients in whom the invasive procedure of surgical biopsy is performed in an attempt to define a diagnosis and plan definitive therapy, it is important to ensure that a correct
diagnosis is made.
We undertook this study to evaluate whether submission of
pathologic material for review by our neuropathology consultation
q 1997 American Cancer Society
/ 7b4c$$0894
01-24-97 15:20:58
cana
W: Cancer
Diagnostic Discrepancies in Pathology/Bruner et al.
service has had a significant clinical impact. We retrospectively compared the diagnoses submitted to us
with our own review diagnoses in a consecutive series
of 500 brain and spinal cord biopsies sent to us from
other hospitals. Our objective was to determine the
numbers and types of changes we made in diagnosis
and whether these changes could result in differences
in prediction of patient prognosis, planned therapy,
or collection and analysis of protocol data. This information might also be of value to third-party insurance
providers or managed care plans in justifying a second
opinion or independent review of pathology or biopsy
material before treatment plans for covered patients
are finalized. Recognizing errors in diagnosis early
would save money by preventing incorrect or inadequate treatment programs. Depending on the dollar
amount per patient and the frequency of the errors,
this potential for recognition might justify the routine
use of independent review by expert pathology subspecialists before initiation of definitive therapy. An additional stimulus for our review was the question of
whether, in an attempt to conserve resources in protocol studies, review of pathology materials could be
discontinued for patients who are to be treated externally using our institution’s clinical protocols.
MATERIALS AND METHODS
The University of Texas M. D. Anderson Cancer Center
is a 535-bed categoric cancer hospital associated with
an integrated research institute. The Department of
Pathology reviews approximately 30,000 surgical pathology cases per year. The Section of Neuropathology
reviews about 1500 cases per year, generally limited
to brain and spinal cord biopsies or resections of tumor or suspected tumor. Approximately 65% of our
cases are outside material sent for review. For the year
1995, our case distribution included approximately
60% gliomas, 10% metastatic tumors from various
sites, 5% meningiomas, and 3% other primary central
nervous system tumors (e.g., medulloblastomas, pituitary adenomas, lymphomas, and germ cell tumors).
Cases that included no neoplastic material accounted
for 10%. The remaining cases (12%) were accrued from
miscellaneous neurosurgical procedures for spine or
craniofacial reconstruction that sometimes included
locally invasive neoplasms not primarily associated
with the central nervous system. Demographically, our
cases included 7% of patients younger than 18 years.
Each of the three board-certified neuropathologists on
our faculty (J.M.B., G.N.F., and L.A.L.) reviews assigned
cases; then, for teaching purposes, the cases are reviewed again in conference during a daily ‘‘rounds’’
session at a multihead microscope, and a consensus
diagnosis is reached.
/ 7b4c$$0894
01-24-97 15:20:58
cana
797
For this study, we retrospectively reviewed the first
500 consecutive cases that had been submitted to our
neuropathology service from outside institutions for
review or consultation in 1995. These cases came from
two sources: referred patient cases, in which the patient had been referred to our center for definitive
initial therapy, adjuvant therapy, a second opinion,
or other therapy after tumor recurrence or treatment
failure; and consultation-only cases, in which a second
opinion in pathology was being sought but the patient
was not being referred. We chose to evaluate the differences between these two sources because the referred
patients had their material reviewed because of referral to our center, whereas the consultation-only cases
were submitted because of some doubt about the original diagnosis on the part of a physician at an outside
institution. In all cases, a brain or spinal cord biopsy
had been performed at another hospital because neoplastic disease was suspected. Slides and/or paraffin
tissue blocks accompanied by institutional pathology
reports were submitted to us. All the reviews were
completed before this study began.
Our diagnoses were confirmed by our three faculty
neuropathologists after conference review at a
multihead microscope. In some cases, additional information that the referring or original pathologist did
not have when making the initial diagnosis, such as
radiologic appearance or clinical symptoms, was made
available to us at our request. We also performed additional special histopathologic studies before making a
diagnosis when this was required.
Our review diagnoses were compared with the
submitted diagnoses, and discrepancies were grouped
into 10 categories (Table 1). For the purpose of analysis, we further consolidated these categories into five
general groups. The first group, serious discrepancies,
comprised those that would have caused an immediate alteration in therapy, excessive increased cost in
terms of patient quality of life or ineffective utilization
of resources, or potential for medical malpractice liability for the treating physician. This group included
cases in which the submitted diagnosis was changed
from neoplastic to nonneoplastic, from nonneoplastic
to neoplastic, from malignant to benign, or from benign to malignant. This group also included cases in
which the diagnosis was changed from one nonneoplastic condition to another or from one type of neoplasm to another. Since we consider all diffusely infiltrating gliomas potentially malignant, we did not use
this category when we changed only the grade of a
glioma, even if that change was from low grade to
anaplastic. Such cases were instead placed in the second group.
The second group comprised discrepancies that
W: Cancer
798
CANCER February 15, 1997 / Volume 79 / Number 4
TABLE 1
Categories of Discrepancies in the Diagnosis of 500 Cases
No. of cases
Category
Referred patient group
Consultation only group
Total
2
2
2
3
3
7
4
4
4
13
9
6
6
7
16
16
29
18
33
34
62
1
10
1
21
18
23
22
28
24
Total discrepancies
No discrepancy
69
147
145
139
214
286
Totals
216
284
500
Serious discrepancy
Tumor diagnosis changed to nontumor
Nontumor diagnosis changed to tumor
Malignant tumor changed to benign
Benign tumor changed to malignant
Diagnosis changed within benign or malignant
Less serious discrepancy
Glioma type changed
Glioma grade changed
Minor discrepancy
Tentative or doubtful diagnosis confirmed
Information added or deleted
First diagnosis made by our service
were less serious but still had substantial potential to
affect patient care decisions or prognosis. This group
included cases in which we changed the type or grade
of a primary glioma. We are currently using a threetiered system of astrocytoma grading,1 and we consider later in this article the implications of our changing grades in this category if other pathologists were
using the four-tiered system of St. Anne – Mayo.2
The third group, minor discrepancies, comprised
cases in which we added or deleted some information or
merely confirmed a tentative or doubtful diagnosis. The
fourth group comprised cases in which we made the initial diagnosis. These cases were submitted by pathologists
who could not arrive at a diagnosis and requested our
opinion before rendering theirs. The fifth group comprised cases in which there was no discrepancy between
the submitted diagnosis and ours.
RESULTS
A summary of our results is presented in Table 1. Referred patient cases comprised 43.2% of the 500 cases
(n Å 216), and consultation-only cases comprised
56.8% (n Å 284). Of the 284 consultation-only cases,
85 were sent as part of protocol studies for second
opinions with regard to therapy. In the remainder of
cases, we made or confirmed diagnoses at the request
of pathologists or other attending physicians, such as
surgeons, internists, or radiotherapists.
In 286 (57.2%) of the 500 cases, we completely
agreed with the referring pathologist; 147 of 216
(68.0%) of these were referred patient cases, and 139
/ 7b4c$$0894
01-24-97 15:20:58
cana
FIGURE 1. Numbers of cases, with and without discrepancies, from the
referred patient group and the consultation-only group are shown. The
500 cases were grouped into referred patient cases or consultation-only
cases, and the total discrepancy rate in each group was determined. The
overall rate of disagreement between the submitted and review diagnoses
was 42.8%.
of 284 (48.9%) were consultation-only cases. For all of
the remaining 214 of 500 cases (42.8%), there was at
least a minor disagreement between the submitting
pathologist and us, or the diagnosis had been in doubt
at the submitting institution (Fig. 1). Discrepancies
were thus noted in 69 of 216 referred patient cases
(31.9%) and in 145 of 284 consultation-only cases
(51.0%) (Fig. 1).
W: Cancer
Diagnostic Discrepancies in Pathology/Bruner et al.
Forty-four cases (8.8%) were classified as serious
errors (Table 2), 96 (19.2%) as less serious errors, and
50 (10.0%) as minor errors. We made the first diagnosis
in 24 cases (4.8%) (Fig. 2).
After subtracting those cases for which we made
the first diagnosis (which might not be considered
cases of ‘‘discrepancy,’’ although a second opinion
was necessary for the initiation of definitive therapy),
the rates of disagreement were slightly changed. Those
44 cases classified as serious errors accounted for 9.2%,
the 96 less serious errors for 20.2%, and the 50 minor
errors for 10.5%.
DISCUSSION
This study of diagnostic discrepancies in a series of
neuropathology referral cases demonstrates that clinically significant errors that could affect patient management and quality of life or incur physician/institutional medicolegal liability are not infrequent. Many
of the errors we found were less significant and might
not change immediate patient therapy or survival, but
could affect experimental protocol entry and/or statistical evaluation of experimental therapies. It should
be noted that errors in all categories occurred in both
the referred patient group and the consultation-only
group. The rates for the consultation-only group might
not be surprising, since these cases were voluntarily
submitted due to some recognized insecurity about
the external diagnosis, usually by the pathologist.
However, cases from the referred patient group were
only reviewed when the patient came to our institution
for treatment. All of these diagnoses had previously
been considered correct by the submitting pathologist.
Clinicians might not recognize a misdiagnosis as discordant with clinical findings. For example, an appearance of a ‘‘mass lesion’’ diagnosed histologically as a
tumor is acceptable, even though the correct diagnosis
might be an infection, an infarct, or a demyelinating
disease. The appropriate therapies for each of these
diagnoses are vastly different.
In the literature addressing discrepancies in histopathologic diagnoses, we found few studies that compared an expert consultant’s diagnosis with the referring pathologist’s diagnosis in surgical pathology
cases.3 – 7 These series reported discrepancies on review
in cases of ovarian carcinoma (8.2%),3 malignant
lymphoma (2.4 – 8.4%),4 and sarcomas (5%).5 In a mandatory second opinion program, review of anatomic
pathology cases revealed a 7.8% rate of discordant diagnoses.6 A study of second opinion reviews of prostate needle biopsies for confirmation of malignancy
prior to definitive radical prostatectomy emphasized
the cost savings of the review, even though the overall
error rate was only 1.3%.7 Several British authors have
/ 7b4c$$0894
01-24-97 15:20:58
cana
799
published retrospective reviews of pathologic diagnoses and commented on internal and external quality
assurance mechanisms.8 – 11 There have been studies
about interobserver variability in histopathology in
specific areas of systemic pathology12,13 and comparisons of frozen-section diagnosis with subsequent paraffin-section confirmation.14,15 A few studies have prospectively reviewed surgical pathology cases within an
institution as a mechanism for quality assurance, reporting rates of clinically significant errors ranging
from 0.26 to 1.8%.16 – 18
With respect to review of brain biopsies, we found
only one prior publication. This was a study that examined histopathologic misclassification among 680 patients entered in a Radiation Therapy Oncology Group
(RTOG) clinical trial for malignant gliomas.19 This
study examined reclassification of anaplastic astrocytomas and glioblastomas by the study review neuropathologist and the effect that it might have had on analysis of the trial data. Fifty-four (34%) of 159 tumors
submitted as anaplastic astrocytomas and 20 (4%) of
521 tumors submitted as glioblastomas were incorrectly classified by the submitting institutions. These
misclassifications resulted in skewed survival statistics
in the submitted group. It is noteworthy that the group
of patients in which the diagnosis was changed from
anaplastic astrocytoma to glioblastoma by the study
neuropathologist had a median survival time that was
nearly the same as that of patients correctly submitted
as having glioblastoma, and the median survival time
of the few glioblastoma patients reclassified as anaplastic astrocytoma was significantly better than that
of the glioblastoma group. The results of the misclassification caused substantial reductions in study power,
as estimated by the computer simulations. The authors confirmed the recognition among neuropathologists of the value of radiologic studies and clinical information in the accurate interpretation of brain biopsies.
In our study, we found serious errors in diagnosis
in 44 cases (8.8%) (Table 2). For example, a case of a
pineal region cystic mass was submitted with a diagnosis of germinoma or ganglioglioma; we changed it
to benign pineal cyst. We changed a diagnosis submitted as cerebral vasculitis to central nervous system
lymphoma. A case that was submitted after the patient
had undergone a course of brain radiation therapy
elsewhere for glioblastoma was rediagnosed as a benign central neurocytoma, a surgically curable disease.
A diagnosis submitted as benign meningioma was
changed to oligodendroglioma, a malignant primary
brain tumor. A diagnosis submitted as a cerebral infarct was changed to toxoplasmosis. The therapeutic
implications of these reclassified diagnoses are critical
W: Cancer
800
CANCER February 15, 1997 / Volume 79 / Number 4
TABLE 2
Major Diagnostic Discrepancies
Case
no.
Age/gender
Source of case
(patient or consult)
Submitting diagnosis
MDACC review diagnosis
Therapy given
before review a
P
C
C
C
C
C
C
C
P
Schwannoma
Low grade astrocytoma
Oligodendroglioma
Meningioma
Ganglioneuroma
Ganglioglioma or germinoma
Low grade astrocytoma
Low grade astrocytoma
Glioblastoma multiforme
Benign cyst
Histiocytes, gliosis
Normal brain
Sarcoidosis, meningeal
Normal ganglion
Normal pineal gland
Normal brain
Demyelination and gliosis
Reactive changes / necrosis
N
Y
N
N
N
N
U
N
N
P
P
C
C
C
C
Gliosis
Gliosis
Normal brain
Gliosis
Reactive changes
Vasculitis
Infiltrating glioma
Infiltrating glioma
Low grade astrocytoma
Langerhans cell histiocytosis
Ganglion cell tumor
Large cell lymphoma
N
U
N
N
N
N
P
P
C
C
C
C
Anaplastic astrocytoma
Gliosarcoma
Anaplastic astrocytoma
Glioblastoma multiforme
Anaplastic astrocytoma
Chordoma
Ganglioglioma
Ganglioglioma
Schwannoma
Central neurocytoma
Ganglioglioma
Paraganglioma
Y
Y
N
Y
Y
N
P
P
P
C
C
C
C
SEGA2
Meningioma
Atypical meningioma
Ganglioglioma
Meningioma
Central neurocytoma
SEGA2
Anaplastic astrocytoma
Malignant meningioma
Malignant meningioma
Oligodendroglioma
Anaplstic oligodendroglioma
Unclassified glioma
Anaplastic astrocytoma
N
N
N
U
N
N
Y
Neuroendocrine carcinoma
PML3
Adenocarcinoma, salivary gland origin
Metastatic small cell carcinoma
Tanycytic ependymoma
Demyelination, suggestive of multiple
sclerosis
Amyloid angiopathy
Rosai-Dorfman disease
Myxoid chondrosarcoma
Schwannoma
Toxoplasmosis
Germinoma
Atypical teratoid/rhabdoid tumor
Demyelination, not PML3
Pituitary adenoma
Pleomorphic xanthoastrocytoma
Y
N
U
N
N
N
Changed tumor to nontumor
1
2
3
4
5
6
7
8
9
57F
68M
65M
45M
34M
40F
26M
45F
41M
Changed nontumor to tumor
10
11
12
13
14
15
8M
31M
35F
36M
2M
63M
Changed malignant to benign
16
17
18
19
20
21
25F
30M
63M
37M
31M
74M
Changed benign to malignant
22
23
24
25
26
27
28
29M
43F
43M
31F
35F
18M
35F
Changed diagnosis within benign or malignant
29
30
31
32
33
34
47M
42F
47F
75F
40F
39F
P
P
P
C
C
C
Olfactory neuroblastoma
Gliosis
Olfactory neuroblastoma
Astrocytoma/ependymoma
Schwannoma
PML3
35
36
37
38
39
40
41
42
43
44
85F
54F
40M
71F
53F
31M
3F
38M
35M
28F
C
C
C
C
C
C
C
C
C
C
Vasculitis
Inflammatory/reactive
Chordoma
Myxopapillary ependymoma
Infarct
Glioma
Anaplastic astrocytoma
PML3
Craniopharyngioma
Meningioma
SEGA: subependymal giant cell astrocytoma of tuberous sclerosis; PML: progressive multifocal leukoencephalopathy; Y: yes; N: no; U: unknown.
a
Therapy other than biopsy or surgery given before the case was reviewed at MDACC.
/ 7b4c$$0894
01-24-97 15:20:58
cana
W: Cancer
N
N
N
N
N
N
N
N
N
N
Diagnostic Discrepancies in Pathology/Bruner et al.
801
FIGURE 2. Numbers of cases in the referred patient and consultation-only groups are shown by type of discrepancy and source of referral. The graph
compares the errors between the referred patient group and the consultation-only group for each category of discrepancy, including serious (immediate
significance for therapy or intervention), less serious but potentially substantial (change in type or grade of glioma), minor (added or deleted information),
or those in which we made the first diagnosis. The bottom bar compares the rates of complete agreement between the two groups.
in all of these cases, as inappropriate or inadequate
therapy would have immediate or delayed consequences for preserving life or maximizing quality of
life. Cases 34 and 42 involve the difference between
progressive multifocal leukoencephalopathy (PML)
and a nonspecific demyelinating process. While therapy might not change, with the distinction between
these two diagnoses the implications for patient survival and quality of life are substantial. PML is a uniformly fatal demyelinating disease with survival typically measured in months, whereas other demyelinating processes, such as multiple sclerosis or postinfectious demyelination, would have an extended survival possibly approaching a normal lifespan. Case 36
also might not involve a therapeutic change, but a
diagnosis of Rosai – Dorfman disease defines an entity
and a prognosis for the patient, whereas a less specific
diagnosis of ‘‘inflammatory/reactive changes’’ might
cause a further, possibly extensive diagnostic search
for infectious or other causes, thus adding to patient
discomfort and health care costs. The cost of obtaining
a second opinion from a pathologist is generally low
when compared with the cost of inappropriate therapy. In this study, we have not specifically addressed
the cost implications of inappropriate radiation therapy for benign diseases, withholding necessary antimicrobial therapy for specific infections, or delaying definitive therapy for malignant diseases improperly di-
/ 7b4c$$0894
01-24-97 15:20:58
cana
agnosed as benign. There are also substantial potential
costs of litigation involved, if sufficient damages to the
patients accrue to increase the liability of pathologists
for making diagnoses below the community standard
of care. Managed care reimbursement may act as a
disincentive for the additional costs of second opinion
reviews by pathologists, especially in capitated plans.
However, this seems to be a foolish economy for
health care providers when compared with the potential cost of inappropriate therapy.
We found that 44% of the discrepancies (96 cases,
or 19.2% of the total group) were of a less serious, but
still substantial, nature. Examples in this group include
cases submitted as anaplastic astrocytoma and
changed to anaplastic oligodendroglioma or cases
submitted as lower grades of astrocytoma and
changed to glioblastoma, based on histologic features.
We realize that multiple grading systems for gliomas
are currently in use. We compared the result using
our three-tiered grading system1 with that achieved
by using the St. Anne – Mayo four-tiered method2 and
found that this type of disagreement (change from
Grade 4 astrocytoma to anaplastic astrocytoma) was
present in only 3 of 62 cases in which we changed the
grade of the glioma. In 14 of 62 cases, we found mitotic
activity in astrocytomas that had been submitted as
low grade. We reclassified the glioma as higher grade
in these cases. In 18 of 62 cases, we changed the grade
W: Cancer
802
CANCER February 15, 1997 / Volume 79 / Number 4
from ‘‘anaplastic,’’ as submitted, to glioblastoma, due
to the presence of tumor necrosis. Although these less
serious errors were deemed less likely to have an impact on long term survival, they do have a high potential for immediate impact on quality of life. They could
affect the assignment of patients to therapeutic protocols in academic institutions and thus alter immediate
patient therapy, or, as found by Scott et al.,19 adversely
affect data collection and analysis.
This group also included 16 of 62 cases in which
limited brain biopsies for suspected tumor in older
patients were submitted with a diagnosis of low grade
astrocytoma. Our examination revealed that these
small biopsies were probably taken from the periphery
of higher grade tumors, so we changed the diagnosis
to infiltrating astrocytoma.20 The implication for therapy is that these patients should not be treated as or
given prognostic estimates based on the diagnosis of
low grade tumor, because this would constitute inadequate therapy or misleading the patient. Usually, survival for patients with low grade astrocytoma can extend beyond 7 years, whereas the median survival for
patients with anaplastic astrocytoma is about 36
months and about 8 months for patients with glioblastoma.1,21 Thus, it should be noted that, although we
considered these discrepancies in glioma grading or
classification to be minor, there were at least 32 cases
in which we increased the tumor grade, thus altering
survival prognosis. In an additional 16 cases, we were
able to caution the treating physician that the biopsy
was probably not representative of the actual tumor
grade and that definitive therapy should thus be based
on a combination of other clinical and/or radiologic
features, rather than on biopsy histology alone.
Although it may seem to be only of minor consequence, our confirmation of diagnoses that were in
some doubt at the referring institution added significant information supporting either the initiation or
the withholding of specific therapy. We confirmed the
presence of low grade astrocytoma that was in doubt
(vs. reactive gliosis) in 12 cases, confirmed a severely
gliotic or demyelinative process (vs. glioma) in 6 cases,
and confirmed a diagnosis of chordoma (vs. pituitary
adenoma) in 1 case.
We considered 23.4% of the discrepancies (50
cases, or 10.0% of the total group) minor errors that
were unlikely to change patient therapy, even though
we did not completely agree with the submitted diagnosis. Examples of this type of discrepancy included
cases in which the submitted diagnosis was brain
lymphoma and our additional immunohistochemistry
studies confirmed the presence of a large cell
lymphoma of B-cell type. Another example was one
in which we were able to classify a metastatic carci-
/ 7b4c$$0894
01-24-97 15:20:58
cana
noma as being of thyroid gland origin by using immunohistochemistry to demonstrate a positive cytoplasmic reaction for thyroglobulin. This patient also
had a primary carcinoma of nonthyroid origin, and
the distinction was deemed clinically important for
treatment.
We classified the cases for which we made the first
diagnosis as a separate group. All but 1 of these 24
cases were consultation-only cases, sent by pathologists who realized that they could not make a final
diagnosis without assistance. Although these cases
might not strictly be considered discrepancies, review
was necessary for initiation of appropriate patient
therapy, and thus they should be considered cases
for which a second opinion regarding diagnosis was
required. If these cases are subtracted from the total
group before statistical calculations are made, the
overall error rate is slightly higher.
We suggest that the greatest source of the errors
in diagnosis is the lack of experience in brain biopsy
interpretation in the practice of general surgical pathology. From personal experience and inquiries made
of pathologists in private practice, we estimate that
the case load of brain and spinal cord biopsies for
tumor or other mass lesions in a large private hospital
(ú500 beds) with an active neurosurgical service might
amount to 150 – 200 per year. With a general pathology
practice group of 5 or 6 pathologists who share the
service load equally, this means that each pathologist
will see about 30 – 40 brain or spinal cord diagnostic
biopsies per year, or less than 1 per week. Even in
university medical centers, the numbers of brain and
spinal cord biopsies seen in a pathology practice may
be small. In one study of prospective peer review of
all diagnostic biopsies received in a university medical
school hospital during a 6-month period, only 22 peripheral and central nervous system cases were
noted.18 In contrast, the 3 neuropathologists in our
neuropathology practice review over 1500 biopsies per
year. Since we are a teaching institution and review
most cases at a daily microscopic consensus conference, each of us actually views about 1200 cases per
year. We have specialized diagnostic techniques available to us, including immunohistochemistry, in situ
hybridization for viral pathogens, and electron microscopy. Our close relationship with our clinical neurosurgery, neurooncology, and neuroradiology colleagues and our participation in the Brain Tumor Center weekly interdisciplinary Tumor Board conferences
leave us, in our highly specialized practice, with a more
complete comprehension of the consequences of our
diagnoses and their implication for treatment.
The results of this study suggest that neuropathology review of diagnostic cases provides a valuable ser-
W: Cancer
Diagnostic Discrepancies in Pathology/Bruner et al.
vice for hospitals with pathologists who are less experienced in the interpretation of brain and spinal cord
biopsies or who do not have specialized techniques,
such as immunohistochemistry, available to them.
Since there were similar error rates in several types of
cases, including those submitted as benign or malignant and neoplastic or nonneoplastic, it is difficult to
recommend a strategy or to define criteria for identifying those brain and spinal cord biopsy cases in a
general pathology practice that should reasonably be
sent to an expert neuropathologist for a consultation.
It is our practice to give specific feedback to the referring pathologist in all cases that are submitted for consultation only. A letter of explanation is written, regardless of whether we disagree or agree with the submitted diagnosis. For cases in which the patient is
treated at our institution, we write a letter of explanation to the referring pathologist only for those cases
in which there is a disagreement between their diagnosis and ours.
As with the RTOG study,19 we recognized few discrepancies among those cases submitted as glioblastoma, although there were some significant errors in
that group. The decision to consult is left to the pathologist handling the case, but factors to consider include
his or her own experience with these types of biopsies
and the ‘‘rarity’’ of the proposed diagnosis. It should
be remembered that external confirmation of any surgical pathology diagnosis serves as documentation of
an active program of quality assurance. Another advantage of consultation is that of arriving at a correct
diagnosis that supports the initiation of appropriate
therapy with a minimum of delay. Our data may reassure many pathologists that submitting a high percentage of brain and spinal cord biopsies for expert
neuropathology consultation is reasonable and prudent. A similar system of case review might be useful
in other pathology subspecialties in efforts to assure
appropriate patient therapy, minimize liability for
physicians, and maximize the efficient utilization of
increasingly limited health care resources in the
United States.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
REFERENCES
1.
2.
3.
Nelson JS, Tsukada Y, Schoenfeld D, Fulling K, Lamarche J,
Peress N. Necrosis as a prognostic criterion in malignant
supratentorial astrocytic gliomas. Cancer 1983;52:550–4.
Daumas-Duport C, Scheithauer BW, O’Fallon J, Kelly P.
Grading of astrocytomas: a simple and reproducible
method. Cancer 1988;62:2152–65.
McGowan L, Norris HJ. The mistaken diagnosis of carcinoma of the ovary. Surg Gynecol Obstet 1991;173:211–5.
/ 7b4c$$0894
01-24-97 15:20:58
cana
20.
21.
803
Kim H, Zelman RJ, Fox MA, Bennett JM, Berard CW, Butler
JJ, et al. Pathology panel for lymphoma clinical studies: a
comprehensive analysis of cases accumulated since its inception. J Natl Cancer Inst 1982;68:43–67.
Harris M, Hartley AL, Blair V, Birch JM, Banerjee SS, Freemont AJ, et al. Sarcomas in North West England: I. Histopathological peer review. Br J Cancer 1991;64:315–20.
Abt AB, Abt LG, Olt GJ. The effect of interinstitution anatomic pathology consultation on patient care. Arch Pathol
Lab Med 1995;119:514–7.
Epstein JI, Walsh PC, Sanfilippo F. Clinical and cost impact
of second-opinion pathology: review of prostate biopsies
prior to radical prostatectomy. Am J Surg Pathol
1996;20:851–7.
Sherwood AJ, Hunt AC. An external quality assessment
scheme in histopathology. J Clin Pathol 1984;37:409–14.
Zuk JA, Kenyon WE, Myskow MW. Audit in histopathology:
description of an internal quality assessment scheme with
analysis of preliminary results. J Clin Pathol 1991;44:10–6.
Ramsay AD. Locally organized medical audit in histopathology. J Clin Pathol 1991;44:3533–7.
Ramsay AD, Gallagher PJ. Local audit of surgical pathology:
18 months’ experience of peer review-based quality assessment in an English teaching hospital. Am J Surg Pathol
1992;16:476–82.
Rosai J. Borderline epithelial lesions of the breast. Am J Surg
Pathol 1991;15:209–21.
Schnitt SJ, Connolly JL, Tavassoli FA, Fechner RE, Kempson
RL, Gelman R, et al. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria. Am J Surg Pathol 1992;16:1133–43.
Zarbo RJ, Hoffman GG, Howanitz PJ. Interinstitutional comparison of frozen-section consultation: a College of American Pathologists Q-Probe study of 79,647 consultations in
297 North American institutions. Arch Pathol Lab Med
1991;115:1187–94.
Ferreiro JA, Myers JL, Bostwick DG. Accuracy of frozen section diagnosis in surgical pathology: review of a 1-year experience with 24,880 cases at Mayo Clinic Rochester. Mayo
Clin Proc 1995;70:1137–41.
Whitehead ME, Fitzwater JE, Kindley SK, Kern SB, Ulrisch
RC, Winecoff WF III. Quality assurance of histopathologic
diagnoses: a prospective audit of 3,000 cases. Am J Clin Pathol 1984;81:487–91.
Safrin RE, Bark CJ. Surgical pathology sign-out: routine review of every case by a second pathologist. Am J Surg Pathol
1993;17:1190–2.
Lind AC, Bewtra C, Healy JC, Sims KL. Prospective peer review in surgical pathology. Am J Clin Pathol 1995;104:560–
6.
Scott CB, Nelson JS, Farnan NC, Curran WJ Jr., Murray KJ,
Fischbach AJ, et al. Central pathology review in clinical trials
for patients with malignant glioma: a report of Radiation
Therapy Oncology Group 83-02. Cancer 1995;76:307–13.
Glantz MJ, Burger PC, Herndon JE, Friedman AH, Cairncross
JG, Vick NA, et al. Influence of the type of surgery on the
histologic diagnosis in patients with anaplastic gliomas.
Neurology 1991;41:1741–4.
McCormack BM, Miller DC, Budzilovich GN, Voorhees GJ,
Ransohoff J. Treatment and survival of low-grade astrocytoma in adults: 1977–1988. Neurosurgery 1992;31:636–42.
W: Cancer
Документ
Категория
Без категории
Просмотров
4
Размер файла
101 Кб
Теги
346
1/--страниц
Пожаловаться на содержимое документа