Abstract
Background: Providing the best treatment options and appropriate
prognostic information to patients with cartilaginous neoplasms of long bones
depends on distinguishing benign from malignant lesions. Correlative
interpretation of imaging, histopathology, and clinical information is the
current method for making this distinction, yet the reliability of this
approach has not been critically evaluated. This study quantifies the
interobserver reliability of the determination of grade for cartilaginous
neoplasms among a group of experienced musculoskeletal pathologists and
radiologists.
Methods: Nine recognized musculoskeletal pathologists and eight
recognized musculoskeletal radiologists reviewed forty-six consecutive cases
of cartilaginous lesions in long bones that underwent open biopsy or
intralesional curettage. All diagnosticians had a bulleted history and
preoperative conventional radiographs for review. Pathologists reviewed the
original hematoxylin and eosin-stained glass slides from each case.
Radiologists reviewed any additional imaging that was available, variably
including serial radiographs, magnetic resonance imaging, and computed
tomography scans. Each diagnostician classified a lesion as benign, low-grade
malignant, or high-grade malignant. Kappa coefficients were calculated as a
measure of reliability.
Results: Kappa coefficients for interrater reliability were 0.443
for the pathologists and 0.345 for the radiologists (p < 0.0001 for both).
Kappa coefficients for a subgroup of cases determined to be high risk by
subsequent clinical course were poorer at 0.236 and 0.206, respectively (p
< 0.0001 for both). Slightly improved agreement among radiologists was
noted for the twenty lesions that had magnetic resonance imaging available
(Kappa = 0.437, p < 0.0001), but not for the lesions analyzed with serial
plain radiographs or computed tomography scans.
Conclusions: This study demonstrates low reliability for the grading
of cartilaginous lesions in long bones, even among specialized and experienced
pathologists and radiologists. This included low reliability both in
differentiating benign from malignant lesions and in differentiating
high-grade from low-grade malignant lesions, both of which are critical to the
safe treatment of these neoplasms. This may explain in part the wide variation
in outcomes reported for chondrosarcomas treated in different medical centers.
New diagnostic and grading strategies linked to protocol-driven treatments are
needed, but they must be measured against the long-term gold standard of
patient outcomes.
Distinguishing between enchondroma and chondrosarcoma is one of the most
frequent diagnostic dilemmas facing orthopaedic oncologists and their
colleagues in diagnostic radiology and
pathology1.
Enchondroma is a very common, benign proliferation of cartilaginous tissue
within bone2. The
vast majority of enchondromas outside the tubular bones of the hand are
indolent and asymptomatic. Many orthopaedic oncologists agree that the optimal
treatment for an enchondroma in a long bone is observation with use of serial
imaging alone to rule out progression. The morbidity of curettage or even
biopsy usually outweighs any potential benefit for these indolent lesions.
Chondrosarcoma is the second most common sarcoma of
bone3,4.
Patients with chondrosarcoma experience a wide range of clinical courses, from
slow insidious tumor growth over years to rapid neoplastic progression,
metastasis, and
death5-12.
Because of its recalcitrance to chemotherapy and radiation, chondrosarcoma is
primarily treated with surgery.
Low-grade chondrosarcomas are managed differently in different centers,
with treatment ranging from aggressive intralesional curettage with or without
local adjuvants to en bloc resection and
reconstruction5,6,8-16.
High-grade chondrosarcomas are almost universally treated with wide or radical
resection in the form of amputation or limb-salvage
reconstruction5,6,8-12,17,18.
Such interventions are expensive, fraught with complications, and morbid in
the long term. While such consequences are well tolerated as treatment for an
otherwise potentially fatal condition, they are highly undesirable in the
setting of misinterpreted, nonthreatening disease. A valid, reliable,
sensitive, and specific test to distinguish benign from low-grade malignant
cartilaginous lesions and low-grade from high-grade malignant cartilaginous
lesions is necessary.
While published criteria for pathologic grading of chondrosarcomas exist,
the application of these criteria is difficult even with expert
judgment15,19,20.
If a pathologist encounters difficulty with a cartilaginous lesion, the case
is usually sent for consultative pathologic opinion at a tertiary medical
center. This customary practice unofficially establishes the opinions of such
consulting tertiary pathologists as the best available standard for diagnosis
and grading of cartilaginous neoplasms.
As there is no so-called gold standard for diagnostic truth in
cartilaginous neoplasms beyond such tertiary interpretations, true diagnostic
validity cannot be ascertained. Therefore, as an initial step in the process
of improving diagnostic methods for cartilaginous neoplasms, the purpose of
our study was to experimentally test the reliability, or reproducibility, of
the current methods. The study assessed the agreement among a panel of expert
tertiary pathologists and radiologists, as measured by kappa coefficients,
with regard to distinguishing benign from malignant cartilaginous neoplasms in
long bones.
Cases
With the approval of the institutional review board, the SNOMED (Systemized
Nomenclature of Medicine) diagnostic retrieval database (SNOMED International,
a division of the College of American Pathologists, Northfield, Illinois) for
a single institution was searched for records that contained either
"chondroma" or "chondrosarcoma." A total of 1071
records were identified and reviewed. Only specimens from open biopsy
curettage and open excisional curettage were included. Cases were excluded if
the lesion was located in the pelvis, axial skeleton, hands, or feet. Cases
from patients with a final diagnosis of osteochondroma, periosteal chondroma,
chondroblastoma, or chondromyxoid fibroma were also excluded. Ninety-nine
cartilaginous lesions in long bones met the above criteria. Notably, these
pathology specimens were obtained across a time-interval of fifteen years,
during which a roughly estimated 1500 or more apparently benign enchondromas
would have been clinically evaluated with noninvasive imaging alone and
followed conservatively. Among the ninety-nine long-bone cartilaginous
lesions, cases were chosen only if preoperative imaging and tissue slides were
accessible; this yielded forty-nine cases
(Fig. 1).
For each of the forty-nine cases, the medical record was examined to
identify patient age and gender, indication for biopsy, presence and duration
of pain clinically attributed to the lesion, history of any preoperative
pathologic fracture and its timing with respect to the biopsy or curettage,
pertinent past medical history of the lesion, and any history of prior
malignancy. All available preoperative plain radiographs, serial radiographs,
magnetic resonance images, and computed tomography scans were digitized. All
unique identifiers were removed from the images and the tissue slides to
protect patient anonymity.
Diagnosticians
Case series reviewers were selected by their peers (E.F.M. and B.R.D. for
pathologists and G.Y.E. for radiologists) and were invited to represent the
relatively small group of specialized experts in skeletal neoplastic
diagnosis. Each participating diagnostician had spent a substantial portion of
clinical service time interpreting musculoskeletal radiology or pathology, had
published peer-reviewed articles related to skeletal neoplasia, and regularly
provided consultative opinions on patients with difficult bone tumors referred
from other institutions. Every invitation to participate was willingly
received, but one of the eighteen invited diagnosticians failed to complete
the review in a timely fashion and was therefore not included in the final
study group.
Of the seventeen diagnosticians, nine (A.E.R., B.R.D., C.Y.I., E.F.M.,
F.H.G., G.P.S., K.K.U., M.J.K., and M.K.) were recognized experts in bone
tumor pathology. These pathologists reviewed the bulleted history, digitized
preoperative conventional radiographs, and glass slides from each case. Two
pathologists (E.F.M. and B.R.D.) each reviewed the case set twice, with a year
between interpretations, to provide an estimate of intrarater reliability. Two
of the pathologists elected to review their accompanying set of conventional
radiographs with a radiologist at their institution, rather than alone.
Eight reviewers (D.A.R., D.M.P., D.V., G.Y.E, L.F., M.D.M., M.J.K., and
M.S.) were recognized experts in bone tumor radiology. These radiologists
reviewed the bulleted histories and all available imaging from the cases,
including serial radiographs, magnetic resonance images, and computed
tomography scans when available. One radiologist (G.Y.E.) reviewed the case
set twice, with a year between reviews.
Grading
Pathologists typically classify hyaline cartilage tumors as benign
(chondromas) or malignant (chondrosarcomas). They usually classify malignant
cartilaginous lesions as grade-1, 2 or 3 chondrosarcoma. There are sometimes
uncertainties based on cytology, cellularity, or imaging studies suggesting
more aggressive behavior. Such borderline lesions may be classified as
cellular enchondroma, atypical enchondroma, or chondrosarcoma grade 1/2.
Surgical management decisions often rely on a simpler system. Many surgeons
use some modification of the Musculoskeletal Tumor Society (Enneking) staging
system for clinical decisions; this system relies on the pathologic grades of
benign, low-grade malignant, and high-grade
malignant21. While
this approach remains controversial, for the purpose of this study, the
grading of lesions was limited to these three categories.
Both radiologists and pathologists were asked to grade lesions as benign,
low-grade malignant (grade 1), or high-grade malignant (grade 2 or 3), using
whatever criteria they would use when interpreting cases in consultation from
another institution. Reviewers also rated their confidence that each lesion
was indeed a cartilaginous lesion (0% to 100%, by tens).
Statistical Methods
Since reliability statistics cannot be calculated if data are missing, the
following methods were used to complete the data set. If more than one
diagnostician failed to provide a grade for a lesion, that lesion was removed
from the final analysis. A single diagnostician's missing grade for a given
lesion was estimated by rounding the mean of the other pathologists or
radiologists, respectively, to the nearest integer when grades were
numerically considered as 0, 1, and 2.
Kappa coefficients of interrater
reliability22 and
Kendall coefficients of concordance were calculated with statistical software
(SAS Institute, Cary, North Carolina) to assess agreement between raters for
the pathologists and radiologists separately. Kappa coefficients measure the
percentage of agreement among raters adjusted for the degree of agreement that
would be expected by chance alone. A number of alternative reliability
statistics are available. The difficulty with each is the interpretation of
the numerical output. As the most commonly used reliability statistic in the
medical literature, kappa coefficients were selected because their familiarity
eases interpretation. The Kendall coefficient of concordance nonparametrically
measures the agreement between different sets of rank orderings. Essentially,
the Kendall coefficient was used to measure whether different raters tended to
have the same rank order of lesions from higher to lower grade, even if
stepwise grades differed.
Three subgroup analyses were also performed. First, a subgroup of
clinically high-risk lesions was identified, to see if agreement was improved
in the nonborderline lesions. This subgroup was defined by the development of
metastases within one year of diagnosis or by pathologic evaluation of
subsequent resection specimens that confirmed grossly and histologically
aggressive tumor invasion of surrounding tissues.
Second, the ability of biopsy and pathologic interpretation to reduce
uncertainty in radiographically uncertain lesions was tested by defining a
subgroup according to disagreement among the radiologist reviewers. For this,
a subgroup kappa coefficient was calculated for pathologist agreement on
lesions for which there was more than one discrepant interpretation by a
radiologist.
Third, to see whether reliability among radiologists varied according to
the imaging modalities available, subgroup kappa coefficients for the
radiologists were calculated for the groups of lesions analyzed with computed
tomography scans (fifteen cases), those analyzed with magnetic resonance
imaging scans (twenty), and those analyzed with serial plain radiographs
(seventeen).
Intraobserver reliability was also tested with kappa coefficients for the
three diagnosticians with two reviews of the cases.
Three cases had more than one missing interpretation from the seventeen
diagnosticians. These three were excluded, yielding forty-six cases in the
final analysis. Two other cases had a single interpretation missing. These
were estimated, as noted in the Materials and Methods section, to complete the
data set.
The mean age (and standard deviation) of the forty-six patients comprising
the analyzed case set was 47.7 ± 16.7 years (range, three to eighty-one
years). There were thirty female and sixteen male patients. All but five
patients had pain potentially attributable to the lesion at the time of
presentation. The mean duration of this pain was 12 ± 9 months. Four
patients had pathologic fractures related to the lesions. Two additional
patients had a distant history of a fracture in the same bone as the lesion.
One patient had a history of melanoma, and three had a history of breast
carcinoma. Twenty-three lesions were located in the femur; sixteen, in the
humerus; five, in the tibia; and one each was in the fibula and the ulna.
The indications for biopsy or curettage were an interpretation by a
radiologist that chondrosarcoma could not be ruled out (twelve cases), an
interpretation by a radiologist that a chondrosarcoma was probable (nine
cases), a worrisome pain pattern attributed to the lesion (eight cases), the
combination of a worrisome pattern of pain and an interpretation by a
radiologist of a possible chondrosarcoma (six cases), subjective expansion of
a mass (three cases), insistence on the part of an anxious patient (two
cases), pathologic fracture requiring surgical intervention (one case), and
other miscellaneous indications (five cases).
The kappa coefficient for interrater reliability was 0.443 (standard error,
0.018; p < 0.0001) for the nine pathologists and 0.345 (standard error,
0.021; p < 0.0001) for the eight radiologists
(Figs. 2-A and 2-B). The
Kendall coefficient of concordance for these groups was 0.762 and 0.637,
respectively (p < 0.0001 for both).
Ten cases were designated as lesions with confirmed aggressive behavior.
Six of these patients had metastatic disease develop within one year. The
other four patients had aggressive bone destruction noted on pathologic
evaluation of specimens from a subsequent wide resection. For this subgroup of
outcome-determined high-risk patients, reliability was not improved
(Figs. 2-C and 2-D). The kappa
coefficient was 0.236 (standard error, 0.038; p < 0.0001) for pathologists
and 0.206 (standard error, 0.041; p < 0.0001) for radiologists.
Twenty-seven cases had more than one discrepant interpretation by a
radiologist, and they were subsequently designated as radiographically
uncertain cases. The kappa coefficient for interrater agreement among the
pathologists' interpretations was 0.370 (standard error, 0.023; p < 0.0001)
for this subgroup of lesions.
The analysis of the reliability of the interpretations by radiologists for
subgroups of cases defined according to available imaging modalities showed a
kappa coefficient of 0.437 (standard error, 0.032; p < 0.0001) for the
twenty patients with a magnetic resonance imaging scan available, 0.299
(standard error, 0.036; p < 0.0001) for the fifteen patients with computed
tomography scans, and 0.347 (standard error, 0.034; p < 0.0001) for the
seventeen patients with serial plain radiographs from at least two separate
time-points prior to biopsy or curettage.
Intraobserver reliability was greater than interobserver reliability, with
a kappa coefficient for the two pathologists at 0.725 (95% confidence
interval, 0.560 to 0.890) and 0.714 (95% confidence interval, 0.521 to 0.907),
resulting from eight and six discrepancies, respectively, of a single grade.
The kappa coefficient for intraobserver reliability for the radiologist who
was a repeat reader was 0.601 (95% confidence interval, 0.410 to 0.792),
resulting from eleven discrepancies of a single grade.
Confidence that the lesion was cartilaginous in character was very strong
from the pathologists for nearly every case
(Table I). The two cases with
mean confidence of <90% among the pathologists that the lesion was
cartilaginous were almost universally considered high-grade lesions. One had a
substantial spindle cell infiltrate, which some pathologists interpreted as
evidence of a dedifferentiated chondrosarcoma and others interpreted as an
osteosarcoma. The other had extensive myxoid degeneration, which prompted some
of the pathologists to consider it a myxoid chondrosarcoma, another to
consider it a hemangioendothelioma variant, and another to consider it
metastatic disease from a carcinoma. Confidence of the cartilaginous character
of the lesions was lower among the radiologists, but it was still >80% in
over half of the cases.
These data indicate that the best currently available means of
distinguishing benign from malignant cartilaginous lesions and low-grade from
high-grade malignant cartilaginous lesions in long bones have low reliability.
The difficulties of these grade interpretations for hyaline cartilage
neoplasms have long been suspected, but never previously quantified.
We estimated the interobserver reliability for the grading of cartilaginous
neoplasms in long bones to be 0.443 for pathologists and 0.345 for
radiologists. The magnitudes of these coefficients are very unlikely to be due
to chance alone as the p values were <0.0001 for both. A commonly used
evaluation scheme, by Landis and
Koch23, would judge
the reliability for these two groups as fair to moderate. It is important to
recognize that the Landis and Koch recommendation lacks external validity
beyond frequent use and name recognition. Reliability coefficients must be
interpreted with the meaningful difference between grades in clear
perspective. The decisions at stake in the grade determinations for these
cartilaginous lesions have substantial impact on patient treatment plans,
treatment morbidity, and overall prognosis. Fair-to-moderate reliability
should not be considered adequate for the specific utility of this grading
system.
For comparison, in a study of twelve osteosarcomas interpreted by eight
expert bone pathologists, the Kappa coefficient was 0.547 for determination of
grade, which was a cause for
concern24. The
reliability of grade determination for soft-tissue sarcomas has been reported
at Kappa coefficients of 0.68 and 0.78 in two prior
studies25,26.
These were interpreted as adequate but in need of improvement. Grading of
breast carcinomas has similarly rendered Kappa coefficients in the 0.69 to
0.73 range that were considered
acceptable27-29.
Diagnostic histopathologic analysis of melanomas has shown Kappa coefficients
of >0.75, which were also considered
acceptable30. Thus,
while a Kappa coefficient of 0.443 among the nine study pathologists may be
"moderate" on the Landis scale, it is very low relative to the
accepted reliability of histopathologic grading for other tumors. Further, the
question at hand for these cartilaginous lesions is not only the grade of
malignancy but also the presence or absence of malignancy.
For example, one case was a biopsy specimen of a distal femoral lesion in a
thirty-eight-year-old man who reported that he had had pain for two months
(Figs. 3-A,3-B,
3-C, 3-D). There was no history
of malignancy or prior fracture, but a biopsy of the same lesion had been
performed at an outside hospital eight years earlier, yielding a diagnosis of
"atypical enchondroma." The interpretations of the radiologists in
the study group were split evenly between benign and low-grade malignant,
meaning that in many scenarios this lesion would not have even undergone
biopsy. The interpretations of the biopsy specimen by the pathologists were
divided between one-third who found the lesion to be benign and two-thirds who
found it to be a low-grade malignant tumor. Therefore, some centers would have
done nothing, certain other centers would have done aggressive curettage, and
still others would have done resection and reconstruction. The lesion was
treated as a malignant tumor with resection and allograft reconstruction.
Pathologic assessment of the resected specimen confirmed some areas of host
cortical bone invasion, but no cortical breeches. The patient remained free of
recurrence and metastatic disease after long-term follow-up. Although the
preoperative pain had been alleviated, the patient had development of chronic
pain related to the allograft used in the reconstruction. While this might be
classified as an oncologic success, it is possible that it represents morbid
overtreatment of a lesion that could have been managed less aggressively.
The group of lesions in the present study included all consecutive lesions
with one of two specific diagnoses from one institution and with available
diagnostic materials. Patient selection by surgeons has a powerful influence
on the cases in the study as the vast majority of cartilaginous lesions in
long bones were not biopsied during the period from which the cases were
culled. Whatever theoretical bias this introduces is considered a strength of
the study, as the only reliability of importance is that for lesions with
clinical uncertainty leading to biopsy. Still, some of the included patients
underwent surgery for reasons other than diagnostic uncertainty. A number of
factors conspire to result in biopsy or curettage even in some lesions
suspected to be benign in character, when biopsy or curettage might preferably
be avoided. For example, three children were included and two of them
underwent operations for deformity correction rather than for diagnosis. Two
insistent patients had pain patterns probably not related to tumor, and they
had unremarkable findings on conventional radiographs. A third woman had a
chondroid lesion identified incidentally during workup and staging for breast
carcinoma; her oncologist insisted on biopsy. Not surprisingly, the
interpretations for these lesions all showed strong agreement for being
benign. The phenomenon extends beyond these few patients and is well
demonstrated in Figures 2-A, 2-B, 2-C,
2-D, as the lesions that had been found to have high agreement
among radiologists for being benign on the basis of radiographic imaging alone
usually had similarly strong agreement among pathologists for a benign
grade.
In contrast, the kappa coefficient for interobserver reliability among
pathologists with regard to the grading of lesions that had more than one
discrepant interpretation from a radiologist was 0.37, or fair on the Landis
and Koch scale. Such poor agreement in pathologic interpretation of these
radiographically uncertain lesions seriously compromises the confidence with
which uncertainty may be reduced by invasive procurement of pathologic
tissue.
Intraobserver reliability for the grading of cartilaginous lesions was, as
expected, better than interobserver reliability, but it still represented 15%
to 24% rates of disagreement for the three repeat-readers. The intraobserver
kappa coefficients for the two pathologists who were repeat readers were in
the range of adequate interobserver kappa coefficients for other tumor types,
as discussed above.
Surprisingly, kappa coefficients for the high-risk subset of cartilaginous
lesions showed what would be classified as poor-to-fair reliability in
lesional grading by the Landis and Koch scale. It was anticipated that most of
the disagreement would alternatively cluster at the benign to low-grade
malignant dichotomy and that the higher-risk lesions would demonstrate
improved agreement. The opposite trend was the case. This finding highlights
the possibility not only that some lesions may receive overaggressive
treatment but also that others may receive undertreatment. Given the fact that
nonresection, curettage-based treatment of purportedly low-grade
chondrosarcomas has been used more frequently in recent years, the ability to
distinguish between lesions that can and cannot be treated by less aggressive
means becomes increasingly critical.
The relatively poor interobserver reliability of the histologic grade of
cartilaginous lesions that these data demonstrated tempers any comparative
interpretation of past literature on chondrosarcoma. It may actually explain
the wide variation in mortality rates that has been reported among centers for
grade-specified series of
chondrosarcoma5,6,8-12,15,18,31.
The relatively poor reliability of the histologic grade also calls into
question the use of the histologic grade and even diagnosis as a surrogate for
prognosis, when new cytologic and immunohistochemical tests are
validated9,13,31-35.
The basis for this use of the histologic grade as the surrogate standard
against which other tests are measured has been the strength of the
correlation between histologic grade and
prognosis1,6-8,10-13,18,36.
Poor reliability of grading does not threaten that correlation, but it
questions the discrete steps along the correlation. Even in this series,
Kendall coefficients of concordance were strong for both pathologists and
radiologists. These coefficients of concordance suggest that there is general
agreement across the group with regard to the fact that there are lower and
higher-grade lesions, but specifying where lesions are classified along the
continuum engenders the variability.
These data do not demonstrate that the interpretation of histopathology
from cartilaginous lesions is without merit. Radiologists, tested as a control
comparison, had a slightly poorer interobserver reliability in grading the
behavior of these lesions than did the pathologists. Further, radiologists
have poor confidence in deciding whether the lesions are definitively
cartilaginous in character in many cases. Pathologists interpreting
histopathology slides can confidently determine whether a lesion is
cartilaginous. Further, while improvements are clearly needed, histopathologic
grading remains the only standard available at the present, despite the
difficulties with it that this study demonstrates.
One weakness of the study is that the imaging available for radiologists to
interpret varied. This heterogeneity does not weaken the utility of the
interpretation by radiologists as a control comparison for the pathologists,
but it does raise the question of whether reliability can be increased with
protocol-driven image acquisition for cartilaginous lesions. Interobserver
correlation for radiologists differentiating low-grade chondrosarcoma from
enchondroma using conventional radiographs alone has previously been reported
to be low37.
Secondary analyses of our data attempted to address what additional imaging
modalities may be helpful to improve diagnostic reliability. No dramatic
differences were noted, but reliability for radiologists grading the group of
twenty lesions for which magnetic resonance imaging was available was slightly
higher at 0.437. While this number cannot be directly compared with the kappa
coefficient for the whole case set of lesions, as the prevalence of different
grades can affect kappa coefficients, it does suggest directions for future
study. Murphey et al. previously compared computed tomography, magnetic
resonance imaging, and nuclear medicine modalities in distinguishing
enchondromas from
chondrosarcomas20,38.
Their data suggested that all three of these modalities, particularly computed
tomography, have utility in distinguishing the two. Others have shown that
magnetic resonance imaging can be helpful in determining the cartilaginous
character of intramedullary neoplasms by the gadolinium enhancement
pattern37,39,40.
The data we present quantify and confirm the suspicion that radiologic and
pathologic interpretations of cartilaginous lesions in long bones are not
optimally reliable, even among specialized experts. We hope that this will
raise awareness and encourage efforts to develop better prospective imaging
protocols and pathologic interpretation algorithms to distinguish benign from
malignant cartilaginous lesions as well as identify at presentation patients
with the higher-risk malignant chondrosarcomas with a poor prognosis.
Molecular-based testing may someday supplant our dependence on these
difficult, subjective interpretations of histologic and radiographic patterns
for distinguishing between chondrosarcoma and enchondroma, but no molecular
test has yet proven better than an expert pathologist and his or her
microscope.
The data in the present study also identify a need to improve our
diagnostic decision-making algorithms, but it is not only pathologists and
radiologists who must be diligent in this regard. Our findings demonstrate
that such a search for improved diagnostics must be anchored against clinical
outcomes rather than the previously used surrogate of the histologic grade.
Anchoring diagnostic imaging, histopathology, and molecular tests of the
future to clinical outcomes requires protocol-driven clinical decisions that
can be compared center to center and followed up in agreed-on patterns. This
study did not address the reliability of what we believe is the ideal approach
to these lesions, which includes diagnosis, grading, and management by a team
including a clinician, a radiologist, and a pathologist who are accustomed to
collaboration with each other in the assessment of bone neoplasms. The
reliability of such teams might be profitably studied prospectively in the
future.
Finally, with regard to the relatively poor ability of histopathologic
interpretation to reduce uncertainty in radiographically uncertain lesions,
these data should elevate the caution with which biopsy of borderline
cartilaginous lesions for diagnostic purposes is undertaken. An anxious
patient cannot be promised that examination of the tissue will settle the
issue. ?
Dahlin DC, Henderson ED. Chondrosarcoma,
a surgical and pathological problem; review of 212 cases. J Bone Joint
Surg Am. 1956;38:
1025-38.381025
1956
[PubMed]
Scherer E. Exostosen, Enchondrome und
ihre Beziehung zum Periost [Exostoses, enchondromas and their relationship
with periosteum]. Frankfurt Z Path.
1928;36:
587-605.36587
1928
Larsson SE, Lorentzon R. The incidence
of malignant primary bone tumours in relation to age, sex and site. A study of
osteogenic sarcoma, chondrosarcoma and Ewing's sarcoma diagnosed in Sweden
from 1958 to 1968. J Bone Joint Surg Br.
1974;56:
534-40.56534
1974
Hartley AL, Blair V, Harris M, Birch JM,
Banerjee SS, Freemont AJ, McClure J, McWilliam LJ. Sarcomas in north west
England: II. Incidence. Br J Cancer. 1991;
64: 1145-50.641145
1991
[PubMed][CrossRef]
Rizzo M, Ghert MA, Harrelson JM, Scully
SP. Chondrosarcoma of bone: analysis of 108 cases and evaluation for
predictors of outcome. Clin Orthop Relat Res.
2001;391:
224-33.391224
2001
[PubMed][CrossRef]
Reith JD, Horodyski MB, Scarborough MT.
Grade 2 chondrosarcoma: stage I or stage II tumor? Clin Orthop Relat
Res. 2003;415:
45-51.41545
2003
[CrossRef]
Pritchard DJ, Lunke RJ, Taylor WF,
Dahlin DC, Medley BE. Chondrosarcoma: a clinicopathologic and statistical
analysis. Cancer. 1980;45:
149-57.45149
1980
[PubMed][CrossRef]
Lee FY, Mankin HJ, Fondren G, Gebhardt
MC, Springfield DS, Rosenberg AE, Jennings LC. Chondrosarcoma of bone: an
assessment of outcome. J Bone Joint Surg Am.
1999;81:
326-38.81326
1999
[PubMed]
Kreicbergs A, Boquist L, Borssen B,
Larsson SE. Prognostic factors in chondrosarcoma: a comparative study of
cellular DNA content and clinicopathologic features. Cancer.
1982;50:
577-83.50577
1982
[PubMed][CrossRef]
Henderson ED, Dahlin DC. Chondrosarcoma
of bone—a study of two hundred and eighty-eight cases. J Bone
Joint Surg Am. 1963;45:
1450-8.451450
1963
Gitelis S, Bertoni F, Picci P,
Campanacci M. Chondrosarcoma of bone. The experience at the Istituto
Ortopedico Rizzoli. J Bone Joint Surg Am.
1981;63:
1248-57.631248
1981
[PubMed]
Fiorenza F, Abudu A, Grimer RJ, Carter
SR, Tillman RM, Ayoub K, Mangham DC, Davies AM. Risk factors for survival and
local control in chondrosarcoma of bone. J Bone Joint Surg Br.
2002;84:
93-9.8493
2002
[PubMed][CrossRef]
Sanerkin NG. The diagnosis and grading
of chondrosarcoma of bone: a combined cytologic and histologic approach.
Cancer. 1980;45:
582-94.45582
1980
[PubMed][CrossRef]
Ozaki T, Lindner N, Hillmann A,
Rödl R, Blasius S, Winkelmann W. Influence of intralesional surgery on
treatment outcome of chondrosarcoma. Cancer. 1996;
77: 1292-7.771292
1996
[PubMed][CrossRef]
Evans HL, Ayala AG, Romsdahl MM.
Prognostic factors in chondrosarcoma of bone: a clinicopathologic analysis
with emphasis on histologic grading. Cancer.
1977;40:
818-31.40818
1977
[PubMed][CrossRef]
Bauer HC, Brosjö O, Kreicbergs A,
Lindholm J. Low risk of recurrence of enchondroma and low-grade chondrosarcoma
in extremities. 80 patients followed for 2-25 years. Acta Orthop
Scand. 1995;66:
283-8.66283
1995
[CrossRef]
Weiner SD. Enchondroma and
chondrosarcoma of bone: clinical, radiologic, and histologic differentiation.
Instr Course Lect. 2004;53:
645-9.53645
2004
[PubMed]
Ozaki T, Hillmann A, Lindner N, Blasius
S, Winkelmann W. Metastasis of chondrosarcoma. J Cancer Res Clin
Oncol. 1996;122:
625-8.122625
1996
[CrossRef]
Bertoni F, Bacchini P, Hogendoorn PC.
Chondrosarcoma. In: Fletcher CDM, Unni KK, Mertens F, editors. WHO
classification of tumours. Pathology and genetics of tumours of soft tissue
and bone. Vol 5. Lyon, France: IARC Press;
2002. p 247-51.5247
2002
Murphey MD, Flemming DJ, Boyea SR,
Bojescul JA, Sweet DE, Temple HT. Enchondroma versus chondrosarcoma in the
appendicular skeleton: differentiating features. Radiographics.
1998;18:
1213-37;1244-5.181213
1998
[PubMed]
Enneking WF. A system of staging
musculoskeletal neoplasms. Clin Orthop Relat Res.
1986;204:
9-24.2049
1986
[PubMed]
Cohen J. A coefficient of agreement for
nominal scales. Educ Psychol Meas.
1960;20:
37-46.2037
1960
[CrossRef]
Landis JR, Koch GG. The measurement of
observer agreement for categorical data. Biometrics.
1977;33:
159-74.33159
1977
[PubMed][CrossRef]
Kilpatrick SE, Abdul-Karim FW, Renner
JB, King TS, Klein MJ, Rosenberg AE. Interobserver variability among expert
orthopedic pathologists for diagnosis, histologic grade, and determination of
the necessity for chemotherapy in osteosarcoma. Pediatr Pathol Mol
Med. 2000;19:
337-58.19337
2000
[CrossRef]
Hasegawa T, Yamamoto S, Nojima T, Hirose
T, Nikaido T, Yamashiro K, Matsuno Y. Validity and reproducibility of
histologic diagnosis and grading for adult soft-tissue sarcomas. Hum
Pathol. 2002;33:
111-5.33111
2002
[CrossRef]
Coindre JM, Trojani M, Contesso G, David
M, Rouesse J, Bui NB, Bockaert A, De Mascarel I, De Mascarel A, Goussot JF.
Reproducibility of a histopathologic grading system for adult soft tissue
sarcoma. Cancer. 1986;58:
306-9.58306
1986
[PubMed][CrossRef]
Harvey JM, de Klerk NH, Sterrett GF.
Histological grading in breast cancer: interobserver agreement, and relation
to other prognostic factors including ploidy. Pathology.
1992;24:
63-8.2463
1992
[PubMed][CrossRef]
Robbins P, Pinder S, de Klerk N, Dawkins
H, Harvey J, Sterrett G, Ellis I, Elston C. Histological grading of breast
carcinomas: a study of interobserver agreement. Hum Pathol.
1995;26:
873-9.26873
1995
[PubMed][CrossRef]
Theissig F, Kunze KD, Haroske G, Meyer
W. Histological grading of breast cancer. Interobserver, reproducibility and
prognostic significance. Pathol Res Pract.
1990;186:
732-6.186732
1990
[PubMed]
Urso C, Rongioletti F, Innocenzi D,
Saieva C, Batolo D, Chimanti S, Filotico R, Gianotti R, Lentini M, Tomasini C,
Rebora A, Pippione M. Interobserver reproducibility of histological features
in cutaneous malignant melanoma. J Clin Pathol.
2005;58:
1194-8.581194
2005
[PubMed][CrossRef]
Aigner T, Müller S, Neureiter D,
Illstrup DM, Kirchner T, Björnsson J. Prognostic relevance of cell
biologic and biochemical features in conventional chondrosarcomas.
Cancer. 2002;94:
2273-81.942273
2002
[PubMed][CrossRef]
Coughlan B, Feliz A, Ishida T, Czerniak
B, Dorfman HD. p53 expression and DNA ploidy of cartilage lesions. Hum
Pathol. 1995;26:
620-4.26620
1995
[CrossRef]
Kusuzaki K, Murata H, Takeshita H,
Hirata M, Hashiguchi S, Tsuji Y, Nakamura S, Ashihara T, Hirasawa Y.
Usefulness of cytofluorometric DNA ploidy analysis in distinguishing benign
cartilaginous tumors from chondrosarcomas. Mod Pathol.
1999;12:
863-72.12863
1999
[PubMed]
Larramendy ML, Mandahl N, Mertens F,
Blomqvist C, Kivioja AH, Karaharju E, Valle J, Böhling T, Tarkkanen M,
Rydholm A, Akerman M, Bauer HC, Anttila JP, Elomaa I, Knuutila S. Clinical
significance of genetic imbalances revealed by comparative genomic
hybridization in chondrosarcomas. Hum Pathol.
1999; 30:
1247-53.301247
1999
[PubMed][CrossRef]
Sawyer JR, Swanson CM, Lukacs JL,
Nicholas RW, North PE, Thomas JR. Evidence of an association between 6q13-21
chromosome aberrations and locally aggressive behavior in patients with
cartilage tumors. Cancer.
1998;82:
474-83.82474
1998
[PubMed][CrossRef]
Sheth DS, Yasko AW, Johnson ME, Ayala
AG, Murray JA, Romsdahl MM. Chondrosarcoma of the pelvis. Prognostic factors
for 67 patients treated with definitive surgery. Cancer.
1996;78:
745-50.78745
1996
[PubMed][CrossRef]
Geirnaerdt MJ, Hermans J, Bloem JL,
Kroon HM, Pope TL, Taminiau AH, Hogendoorn PC. Usefulness of radiography in
differentiating enchondroma from central grade 1 chondrosarcoma. AJR Am
J Roentgenol. 1997;169:
1097-104.1691097
1997
Flemming DJ, Murphey MD. Enchondroma and
chondrosarcoma. Semin Musculoskelet Radiol.
2000;4:
59-71.459
2000
[PubMed][CrossRef]
Geirnaerdt MJ, Bloem JL, Eulderink F,
Hogendoorn PC, Taminiau AH. Cartilaginous tumors: correlation of
gadolinium-enhanced MR imaging and histopathologic findings.
Radiology. 1993;186:
813-7.186813
1993
[PubMed]
Geirnaerdt MJ, Hogendoorn PC, Bloem JL,
Taminiau AH, van der Woude HJ. Cartilaginous tumors: fast contrast-enhanced MR
imaging. Radiology. 2000;
214: 539-46.214539
2000
[PubMed]