Variable Penetrance of a Consensus
Classification Scheme for Renal Cell
John M. Hollingsworth, David C. Miller, Stephanie Daignault, Rajal B. Shah, and
Brent K. Hollenbeck
To evaluate the penetrance of the new pathologic standard of care, we characterized the
temporal trends in histologic subtype-specific kidney cancer incidence rates. Molecular genetics
have refined our understanding of kidney cancer, such that kidney cancer is now recognized as
a family of tumors with distinct molecular and clinical characteristics. The histologic classifica-
tion of kidney cancer has been revised to reflect this new paradigm.
Using the Surveillance, Epidemiology, and End Results Program, we identified incident cases
from 1983 to 2002. Tumor histologic types were assigned, using the International Classification of
Disease-Oncology codes. The histologic-specific incidence rates were calculated and directly
age-adjusted to the 2000 U.S. population.
The histologic type was available for 40,813 cases. Subsequent to the Heidelberg consensus
conference, the rate of papillary histologic types rose appropriately from 0.02 in 1998 to 0.89 in
2002 per 100,000 U.S. population, and the incidence of granular cell histologic types remained
relatively stable (0.22 to 0.14 cases per 100,000), despite its exclusion as a unique histologic
subtype. Paradoxically, the incidence of chromophobe tumors decreased during this interval
(0.03 to 0.003 cases per 100,000).
Following the publication of the Heidelberg classification scheme, we have described the
differential changes in incidence rates for newly described histologic variants. Our results suggest
incomplete penetration of these guidelines. The continued reporting of granular cell histologic
types is particularly noteworthy, given that it is no longer recognized as a distinct histologic
subtype. Proper categorization of the histologic subtype (eg, chromophobe, papillary, clear cell)
is imperative, because it may confer useful information regarding the prognosis, response to
adjuvant treatment, and eligibility for clinical trials.
UROLOGY 69: 452–456, 2007. © 2007
renal cell carcinoma is not a single entity, but rather a
family of cancers with distinct biologic and clinical charac-
teristics.1This understanding led to the consensus confer-
ence in Heidelberg, Germany in 1996, in which a refined
histopathologic classification system was adopted by the
experts in the field. To reconcile with contemporary mo-
lecular relationships, some categories (eg, sarcomatoid, tu-
uring the past decade, the genetic underpinnings of
renal cell carcinoma have been further elucidated.
Through this work, it has been recognized that
bulopapillary, and granular cell) were discarded from the
pathology lexicon and others (eg, collecting duct, papillary,
and chromophobe) were confirmed.2The rationale for this
paradigm shift was to develop a morphologic classification
scheme that reflected the distinct genetic abnormalities
underlying these tumors, which are both predictive of pa-
tient outcomes3–7and the targets of emerging molecular
therapies for advanced disease.8,9In this context, we under-
took this study to better understand the dissemination and
MATERIAL AND METHODS
is maintained by the National Cancer Institute. It collects inci-
dence, treatment, and mortality data from 14 population-based
cancer registries and 3 supplemental registries, covering approxi-
mately 26% of the U.S. population. The demographic composi-
This study was supported by NIH-T-32-DK007758 grant to J. M. Hollingsworth and
a grant from the John and Suzanne Munn Endowed Research Fund of the University of
Michigan Comprehensive Cancer Center to B. K. Hollenbeck.
From the Department of Urology, Michigan Urology Center, University of Michigan
Health System, Ann Arbor, Michigan and Department of Pathology, University of
Michigan Medical School, Ann Arbor, Michigan
Reprint requests: Brent K. Hollenbeck, M.D., M.S., Department of Urology,
University of Michigan Health System, Taubman Center 2916H, Box 0330, 1500
East Medical Center Drive, Ann Arbor, MI 48109-0330. E-mail: firstname.lastname@example.org
Submitted: April 30, 2006, accepted (with revisions): November 16, 2006
© 2007 Elsevier Inc.
All Rights Reserved
tion, as well as the cancer incidence and mortality trends, of SEER
Using the original SEER 9 registries, we identified the incident
cases of kidney cancer using the International Classification of Dis-
ease for Oncology, Third Edition (ICD-O3) code (C64.9), histology
codes (clear cell [8310 and 8312]; papillary ; chromophobe
[8317 and 8270]; oncocytoma ; sarcomatoid ; granular
cell ; spindle cell, not otherwise specified [NOS] ;
small cell, NOS ; adenocarcinoma, NOS ; carcinoid
tumor, NOS ); collecting duct ; nephroblastoma,
interstitial tumor ), and behavior codes (2 and 3 for adeno-
carcinoma in situ and adenocarcinoma, NOS, respectively) that
occurred from 1983 to 2002. Because the observed cell frequencies
were low for multiple histology codes, a category “other” was
created, leaving us with the following groupings for our analysis:
clear cell, papillary, chromophobe, granular cell, and “other” renal
cell carcinoma. Additional demographic data abstracted include
the SEER site and year of diagnosis.
In our first analytical step, we calculated the histologic subtype-
specific renal cell carcinoma incidence rates. This was done by first
determining the number of cases diagnosed each year and dividing
by SEER and multiplying by 100,000. The rates were age adjusted
to the 2000 U.S. population using direct adjustment methods. To
examine the relative changes in these rates, we plotted the sub-
type-specific incidence on a logarithmic scale by diagnosis year.
Next, we fit a multivariate logistic regression model to explore
the determinants of obsolete histologic classification following the
Heidelberg conference. To accomplish this, our primary outcome
measure was defined as the persistent use and misclassification of
the granular cell subtype (in lieu of using a more appropriate
subtype such as clear cell, papillary, or chromophobe) in the
post-Heidelberg era. Our primary covariates of interest for this
analysis were the year of diagnosis, SEER site, and an interaction
term between the two. The timeframe for our model was limited to
1998 to 2002 to allow for dissemination of the Heidelberg frame-
All tests were two-tailed and performed at a significance level of
0.05 using the Statistical Analysis System, version 9.1 (SAS In-
stitute, Cary, NC). In accordance with the Code of Federal Reg-
ulations Title 45 Subpart A Section 46.101 paragraph b subpara-
graph 4, institutional review board approval was waived for this
Histologic data were available for 40,813 kidney cancer
cases from 1983 to 2002. A summary of the distribution
of the subset of cases diagnosed after 1997 (n ? 13,196)
by SEER site and diagnosis year is provided in Table 1.
The number of inappropriately classified tumors varied by
site, with the largest number of granular cell tumors
recorded in the Metropolitan Detroit and Iowa cancer
registries. In 2000, nearly 3% of recorded renal cortical
tumors were designated as granular cell tumors. Although
the chromophobe histologic type has been reported to
represent 5% of all kidney cancer cases,11less than 1% of
renal cell carcinomas were classified as such in the SEER
Figure 1 depicts the relative change in age-adjusted
renal cell carcinoma incidence rates by diagnosis year for
three non-clear cell histologic subtypes. Throughout this
interval, clear cell carcinoma remained, by far, the most
frequent histologic subtype (data not shown). Subse-
quent to publication of the Heidelberg classification sys-
tem,2the rate of papillary kidney cancer increased from
0.02 to 0.89 cases per 100,000 U.S. population from 1998
to 2002. Paradoxically, the incidence rates for chromo-
phobe renal cell carcinoma declined in the years after the
Table 1. Distribution of cases by SEER site and diagnosis year (1998–2002)
By SEER site
(% of total for all
By diagnosis year
(% of total in given
SEER ? Surveillance, Epidemiology, End Results.
Granular CellOtherClear Cell Chromophobe
UROLOGY 69 (3), 2007
consensus conference (0.03 to 0.003 cases per 100,000),
even though it was confirmed as a distinct histopatho-
logic entity. At the same time, the incidence of granular
cell renal cell carcinoma remained essentially stable
(0.22 to 0.14 cases per 100,000), despite its exclusion as
a unique subtype.
In our multivariate analysis, both the diagnosis year
and SEER site were independently associated with patho-
logic misclassification (P ? 0.002 and P ?0.001, respec-
tively). We failed to observe a significant interaction
between the year of diagnosis and geographic care set-
ting. Table 2 presents the adjusted odds ratios (ORs) of
pathologic misclassification stratified by year and SEER
site. Relative to the year 1998, the odds of misclassifica-
tion initially increased after publication of the Heidel-
berg scheme, but then gradually decreased. By 2002, the
odds of misclassification were 40% lower (OR 0.6, 95%
confidence interval 0.4 to 0.8). Patients in the Utah and
Metropolitan Detroit registries were more than three
times more likely to be assigned an excluded histologic
subtype compared with those from the Atlanta SEER site
(OR 3.6 and 95% confidence interval 1.8 to 7.4 and OR
3.1 and 95% confidence interval 1.6 to 5.8, respectively).
This study describes the pathologic patterns of care after
the establishment of a new consensus classification sys-
tem for renal cell carcinoma. Notable is our observation
that, despite its exclusion as a distinct histologic type, the
incidence rates for granular cell renal cell carcinoma
remained relatively stable within the SEER Program.
Although the frequency of the granular cell histologic
type varied across geographic care settings, it was reas-
suring that the assignment of this obsolete diagnosis has
declined to some extent during the past several years. At
the same time, 5 years after the publication of the Hei-
delberg classification, the incidence of chromophobe re-
nal cell carcinoma demonstrated a paradoxical decline.
Because chromophobe renal cell carcinoma can have
granular cell features, one possible explanation for this
trend would be the misclassification of some chromo-
phobe tumors with the granular cell histologic type.
Given the role that tumor histologic subtypes play in
patient prognosis, the predicted response to adjuvant
therapy, and the possible enrollment in clinical trials,
persistent pathologic misclassification, in particular the
continued use of an excluded subtype, highlights a po-
tential quality-of-care concern.
Significantly, it has been recognized that kidney can-
cer is not a single disease, but rather a family of distinct
tumors with variable clinical behavior.4The World
Health Organization now acknowledges 12 types of renal
cell tumors on the basis of morphologic, immunohisto-
chemical, and genetic features in its latest classification
system.11Retrospective clinical case series have shown
that patients with chromophobe and papillary kidney
cancer tend to present with lower stages of disease.4,5,7In
addition, patients with these histologic subtypes appear
to have superior cancer-specific survival compared with
those with conventional clear cell renal cell carci-
noma.3–7With refinements in imaging and biopsy tech-
planned surgery may soon be a clinical reality. Recently,
our institution reported a series of 66 patients who un-
derwent image-guided biopsy for evaluation of indeter-
minate renal masses. Among those biopsies adequate for
assessment, a definitive diagnosis could be established in
98% of cases, affecting management.12As such, proper
pathologic classification will be essential for distinguish-
ing potentially indolent and aggressive tumors, thereby
facilitating a tailored treatment approach.
Log rate per 100,000 U.S. population
Figure 1. Age-adjusted renal cell carcinoma incidence rate
stratified by histologic subtype. Rates were age adjusted
and expressed as the number of cases per 100,000 U.S.
population. Papillary histologic type represented by filled
diamonds; chromophobe histologic type by filled triangles;
and granular cell histologic type by open squares.
Table 2. Adjusted odds ratio of histologic misclassifica-
tion for years 1998–2002, controlling for SEER site and
SEER ? Surveillance, Epidemiology, and End Results.
UROLOGY 69 (3), 2007
Misclassification has several other implications, as
well. First, the tumor histologic type has an important
role in counseling patients after local therapy. Although
the extent of disease remains the most important prog-
nostic indicator,14the histologic subtype has been
shown to independently influence the oncologic out-
come.3–7,15,16Thus, proper subtyping may facilitate pa-
tient-centered refinements in surveillance and follow-up
Second, the tumor histologic type may influence the
use of immune modulators in the management of locally
advanced and metastatic disease. These immune modu-
lators are currently limited to the single-agent immuno-
therapies interleukin-2 and interferon alpha-2a. The re-
sponse rates for these agents have been modest (10% to
20% of patients)17,18and largely limited to certain his-
tologic subtypes.19Thus, misclassification could result in
the exposure of patients to ineffective therapies and
potentially lethal side effects.20Finally, molecular profil-
ing techniques are facilitating and motivating the ex-
panded use of targeted therapies for patients with ad-
vanced renal cell carcinoma. Again, proper pathologic
classification is imperative to determine patient eligibility
for these important clinical trials.8,9
Our data have highlighted that consensus opinion does
not necessarily translate into universal adoption and al-
teration of clinical practice. An understanding of diffu-
sion theory helps to explain our observed variations.
Physician adoption of any clinical behavior is a conse-
quence of multiple factors, of which scientific evidence is
only one.21An often overlooked factor is the channel of
communication through which the information is con-
veyed. The Heidelberg classification scheme, published
in a prominent pathology journal, relied on passive dis-
semination of information to effect change. This strategy
is generally an ineffective approach to changing physi-
cian behavior and, therefore, might explain the slow
uptake of the Heidelberg classification scheme.22,23In a
systematic review, Bero and colleagues24reported that
multifaceted interventions are most effective at achiev-
ing changes in practice patterns. This approach often
includes some combination of audit and feedback, re-
minders, local consensus building, and marketing, among
other strategies.24Diffusion theory also posits that clini-
cians must move through a multistep decision process
before they will change their behavior.21Given that
individuals often move through this process at different
rates, a time lag in acceptance between innovators and
late adopters can result.21As a consequence of this lag,
an interval for transition is needed before universal ac-
ceptance can be achieved. With the odds of misclassifi-
cation gradually decreasing, our data might reflect, in
part, this time of transition.
In this study, we have demonstrated the persistent and
inappropriate use of the excluded granular cell histologic
subtype in renal cell carcinoma surveillance systems. As
described, this finding could have multiple quality impli-
cations for patients with both early-stage and advanced
disease. Nonetheless, our observations must be consid-
ered in the context of several limitations. Using the
SEER Program data set, one cannot distinguish between
errors in coding versus errors in pathologic classification.
However, SEER’s data quality is recognized worldwide,
and quality control has been an integral part of SEER’s
activities since the Program’s inception.25To minimize
variation across SEER sites, a standardized protocol for
data abstraction is used. Per SEER’s most recent coding
and staging manual, coders are instructed to “use the
histology stated in the final diagnosis from the (patient’s)
pathology report” or, in the absence of a tumor specimen,
to “code the histology described by the medical practi-
tioner” when assigning a histology code.26Furthermore,
SEER conducts regular case-finding, recoding, and reli-
ability studies, as well as training programs for coders, to
ensure its data quality.25The reliability of the SEER
histologic diagnoses has also been independently con-
firmed. In a subset of lung cancer cases, the reported
histologic types were compared to a review of diagnostic
slides, and the percentage of exact agreement ranged
from 83% to 98%, depending on the subtype.27
An additional limitation was that the SEER data are
available only through 2002. Consequently, we could not
comment on the more recent evolution of the noted
trends. However, our analysis showed fairly constant use
of an excluded subtype 5 years after an international
consensus conference, suggesting incomplete penetration
of the Heidelberg classification guidelines.
Following the Heidelberg consensus conference, we have
described differential changes in the incidence rates for
newly described histologic variants, suggesting incom-
plete penetration of these guidelines. The continued re-
porting of the granular cell histologic type is particularly
noteworthy, given that it is no longer recognized as a
distinct histologic type. As the histologic type of cancer
(eg, chromophobe, papillary, clear cell) may confer ad-
ditional information regarding prognosis, the response to
adjuvant treatment, and eligibility for clinical trials,
proper categorization is imperative. This misclassification
represents a quality-of-care concern. Although these
findings could not distinguish between coding discrepan-
cies and pathologic reporting, they do speak to nonuni-
form implementation of the current classification
scheme. Furthermore, these data have demonstrated the
lag between expert consensus and standard clinical prac-
1. Linehan WM, Walther MM, and Zbar B: The genetic basis of
cancer of the kidney. J Urol 170: 2163–2172, 2003.
2. Kovacs G, Akhtar M, Beckwith BJ, et al: The Heidelberg classifi-
cation of renal cell tumours. J Pathol 183: 131–133, 1997.
UROLOGY 69 (3), 2007
3. Amin MB, Amin MB, Tamboli P, et al: Prognostic impact of Download full-text
histologic subtyping of adult renal epithelial neoplasms: an expe-
rience of 405 cases. Am J Surg Pathol 26: 281–291, 2002.
4. Beck SD, Patel MI, Snyder ME, et al: Effect of papillary and
chromophobe cell type on disease-free survival after nephrectomy
for renal cell carcinoma. Ann Surg Oncol 11: 71–77, 2004.
5. Cheville JC, Lohse CM, Zincke H, et al: Comparisons of outcome
and prognostic features among histologic subtypes of renal cell
carcinoma. Am J Surg Pathol 27: 612–624, 2003.
6. Ljungberg B, Alamdari FI, Stenling R, et al: Prognostic significance
of the Heidelberg classification of renal cell carcinoma. Eur Urol
36: 565–569, 1999.
7. Moch H, Gasser T, Amin MB, et al: Prognostic utility of the
recently recommended histologic classification and revised TNM
staging system of renal cell carcinoma: a Swiss experience with 588
tumors. Cancer 89: 604–614, 2000.
8. Rini BI, Sosman JA, and Motzer RJ: Therapy targeted at vascular
endothelial growth factor in metastatic renal cell carcinoma: biology,
clinical results and future development. BJU Int 96: 286–290, 2005.
9. Yang JC, Haworth L, Sherry RM, et al: A randomized trial of
bevacizumab, an anti-vascular endothelial growth factor antibody,
for metastatic renal cancer. N Engl J Med 349: 427–434, 2003.
10. Hankey BF, Ries LA, and Edwards BK: The Surveillance, Epide-
miology, and End Results program: a national resource. Cancer
Epidemiol Biomarkers Prev 8: 1117–1121, 1999.
11. World Health Organization Classification of Tumours: Pathology and
Genetics of Tumours of the Urinary System and Male Genital Organs.
Lyon, France, IARC Press, 2004.
12. Shah RB, Bakshi N, Hafez KS, et al: Image-guided biopsy in lesions
in contemporary adequacy, clinical impact, pathological diagnosis.
Hum Pathol 36: 1309–1315, 2005.
13. Rybicki FJ, Shu KM, Cibas ES, et al: Percutaneous biopsy of renal
masses: sensitivity and negative predictive value stratified by clin-
ical setting and size of masses. AJR Am J Roentgenol 180: 1281–
14. Skinner DG, Colvin RB, Vermillion CD, et al: Diagnosis and
management of renal cell carcinoma. A clinical and pathologic
study of 309 cases. Cancer 28: 1165–1177, 1971.
15. Patard JJ, Leray E, Rioux-Leclercq N, et al: Prognostic value of
histologic subtypes in renal cell carcinoma: a multicenter experi-
ence. J Clin Oncol 23: 2763–2771, 2005.
16. Zisman A, Pantuck AJ, Dorey F, et al: Improved prognostication of
renal cell carcinoma using an integrated staging system. J Clin
Oncol 19: 1649–1657, 2001.
17. Bukowski RM: Immunotherapy in renal cell carcinoma. Oncology
13: 801–810, 1999.
18. Figlin RA: Renal cell carcinoma: management of advanced disease.
J Urol 161: 381–386, 1999.
19. Motzer RJ, Bacik J, Mariani T, et al: Treatment outcome and
survival associated with metastatic renal cell carcinoma of non-
clear-cell histology. J Clin Oncol 20: 2376–2381, 2002.
20. Coppin C, Porzsolt F, Awa A, et al: Immunotherapy for advanced
renal cell cancer. Cochrane Database of Syst Rev No. CD001425,
21. Rogers E: Diffusion of Innovations. New York, Free Press, 1983.
22. Lomas J: Words without action? The production, dissemination,
and impact of consensus recommendations. Annu Rev Public
Health 12: 41–65, 1991.
23. Lomas J, Anderson GM, Domnick-Pierre K, et al: Do practice
guidelines guide practice? The effect of a consensus statement on
the practice of physicians. N Engl J Med 321: 1306–1311, 1989.
24. Bero LA, Grilli R, Grimshaw JM, et al., for the Cochrane Effective
Practice and Organization of Care Review Group: Closing the gap
between research and practice: an overview of systematic reviews of
interventions to promote the implementation of research findings.
BMJ 317: 465–468, 1998.
25. National Cancer Institute: SEER cancer statistics review, 1975–
2002,2005 update. Available
1975_2002/. Accessed April 30, 2006.
26. National Cancer Institute: The SEER coding and staging manual
2004, 2004 update. Available at: http://seer.cancer.gov/tools/
codingmanuals/. Accessed April 30, 2006.
27. Field RW, Smith BJ, Platz CE, et al: Lung cancer histologic type in
the Surveillance, Epidemiology, and End Results Registry versus
independent review. J Natl Cancer Inst 96: 1105–1107, 2004.
UROLOGY 69 (3), 2007