Patterns in Immunohistochemical Usage in Extended Core Prostate Biopsies Comparisons Among Genitourinary Pathologists and Nongenitourinary Pathologists Anna Plourde, MD; Alden Gross, PhD; Zhong Jiang, MD; Christopher L. Owens, MD
Context.—Immunohistochemical (IHC) stains have known utility in prostate biopsies and are widely used to augment routine staining in difficult cases. Patterns in IHC utilization and differences based on pathologist training and experience is understudied in the peer-reviewed literature. Objectives.—To compare the rates of IHC usage between specialized (genitourinary; [GU]) and nonspecialized (non-GU) pathologists in extended core prostate biopsies (ECPBs) and the effects of diagnosis; and in cancer cases Gleason grade, disease extent, and perineural invasion on the rate. Design.—Consecutive ECPBs from 2009–2011 were identified and billing data were used to determine the number of biopsies and IHC stains per case. Diagnoses were mapped and in cancer cases, Gleason grade, extent of disease, and perineural invasion were recorded. Pathologists were classified as GU or non-GU on the basis of training and experience. Results.—A total of 618 ECPBs were included in the
study. Genitourinary pathologists ordered significantly fewer IHC tests per case and per biopsy than non-GU pathologists. The rate of ordering was most disparate for biopsies of cancerous and benign lesions. For biopsies of cancerous lesions, high-grade cancer, bilateral disease, and perineural invasion decreased the rate of ordering in both groups. In cancer cases, GU pathologists ordered significantly fewer stain tests for highest Gleason grade of 3 þ 3 ¼ 6, for patients with focal disease and for patients with multiple positive bilateral cores. The effect of the various predictors on IHC ordering rates was similar in both groups. Conclusions.—Genitourinary pathologists ordered significantly fewer IHC stain tests than non-GU pathologists in ECPBs. Guidelines to define when IHC workup is necessary and not necessary may be helpful to guide workups. (Arch Pathol Lab Med. 2013;137:1630–1634; doi: 10.5858/arpa.2012-0517-OA)
I
stain test for diagnostic purposes is based on the judgment of the individual pathologist. Unlike the situation that many of our clinical colleagues face when ordering diagnostic tests, there are no established criteria that define situations where an IHC workup is not justified or conversely, when it should be considered mandatory. This may lead to excessive or unnecessary IHC workups or omissions of critical IHC staining. In recent years, there has been increasing recognition of the need for evidence-based approaches in all fields of medicine, including pathology.1–5 With our nation’s health care costs rapidly rising, the use of efficient, evidence-based, and cost-effective diagnostic approaches is of utmost importance. Evidence-based guidelines can help direct physicians toward tests that have substantial impact on patient diagnosis and treatment and steer physicians away from those that do not. They can also direct physicians as to when to use particular tests. Unfortunately, firmly established guidelines for the use of IHC stains are lacking. In this study, we investigated factors that may influence the usage of IHC stains in extended core prostate biopsies (ECPBs). In this specific setting the most common IHC
mmunohistochemical (IHC) stains are important ancillary tests frequently used by anatomic pathologists to complement findings on routine staining. Immunohistochemical stains are useful for diagnostic purposes (differentiating benign from malignant processes, subtyping tumors, etc) and therapeutic purposes (identifying tumors appropriate for targeted therapeutic agents). However, IHC stains increase costs and turnaround time and therefore should be used judiciously. The decision to order an IHC Accepted for publication December 11, 2012. From the Department of Pathology, University of California San Francisco, San Francisco (Dr Plourde); the Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School and the Institute for Aging Research, Hebrew SeniorLife, Boston, Massachusetts (Dr Gross); and the Department of Pathology, University of Massachusetts Medical School, Worcester (Dr Jiang and Dr Owens). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Christopher L. Owens, MD, Department of Pathology, University of Massachusetts Medical School, Three Biotech, One Innovation Drive, Worcester, MA 01605 (e-mail: Christopher. [email protected]). 1630 Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Table 1.
Summary of Descriptive Statistics of Patients (N ¼ 618) GU (n ¼ 364)
Diagnosis, n (%) Benign HGPIN ASAP Cancer No. of stains per ECPB, mean (SD) No. of biopsies per ECPB, mean (SD) Rate of stains/biopsy, mean (SD)
142 15 28 179 1.3 9.1 0.1
(39.0) (4.1) (7.7) (49.2) (2.3) (2.5) (0.3)
Non-GU (n ¼ 254) 133 10 9 102 2.2 7.7 0.3
(52.4) (3.9) (3.5) (40.2) (5.1) (2.1) (0.7)
P Value .01
.004 ,.001 ,.001
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; ECPB, extended core prostate biopsy; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. Bolded values are statistically significant.
stains ordered are stains for basal cells, including 34bE12 (cytokeratin 903, CK903) and p63 protein, as well as amethylacyl coenzyme A racemase (AMACR, p504s), a biomarker overexpressed in prostate cancer. These stains have known utility in complementing routine stains to aid in confirming or refuting small atypical glandular foci as cancer.6–9 The 3 stains are commercially available as a cocktail with 2 chromagens, to facilitate evaluation of small foci with 1 tissue level, and are widely used by pathologists who interpret prostate needle biopsy findings. Less commonly, other IHC stains may be used in poorly differentiated tumors to aid in classification or to exclude/confirm secondary malignancy in ECPBs. For this study we divided the pathologists into 2 groups, genitourinary specialists and nonspecialists, by background and training. Immunohistochemical stain utilization in ECPBs was compared between the 2 groups. The effects of diagnosis and tumor grade, tumor extent, and perineural invasion in cancer cases were also analyzed in this study.
number of ordered stain tests by the number of biopsies for each ECPB. The rate was regressed separately on predictors, including diagnosis, number of biopsies, Gleason grade, extent of cancer, and presence of perineural invasion in cancer cases, by using Poisson regression with a log-linear link function to account for nonnormal distribution of rates. To test whether associations differed by specialty status, we included statistical interactions between specialty status and each variable in separate regressions. Associations with P values ,.05 were considered significantly different. Analyses were conducted with Stata statistical software, version 12.1 (StataCorp, College Station, Texas).
We calculated descriptive characteristics of the sample by using v2 tests for categorical variables and t tests for continuous variables. We calculated the rate of IHC stain tests per biopsy by dividing the
RESULTS In total, 618 ECPBs were identified and included in the study. Of these, 364 (58.9%) were interpreted by a GU pathologist and 254 (41.1%) were interpreted by a non-GU pathologist (Table 1). The distribution of diagnoses differed between the GU pathologists and the non-GU pathologists (P ¼ .01): GU pathologists diagnosed cancer, ASAP, HGPIN, and benign lesion in 49.2%, 7.7%, 4.1%, and 39.0% of cases, respectively. Non-GU pathologists diagnosed cancer, ASAP, HGPIN, and benign lesion in 40.2%, 3.5%, 3.9%, and 52.4% of cases, respectively. Among patients with cancer there was no difference for the highest Gleason grade (P ¼ .11) or extent of cancer (P ¼ .30), but perineural invasion was identified in a significantly higher percentage of cancer cases by GU pathologists (39.1% versus 25.5%, P ¼ .02) (Table 2). The number of biopsy samples per ECPB was significantly higher in the GU group (9.1 versus 7.7, P , .001) (Table 1). The GU pathologists ordered significantly fewer stain tests per ECPB (1.3 versus 2.2, P ¼ .004) and stain tests per biopsy (0.1 versus 0.3, P , .001). Table 1 summarizes the distribution of diagnoses, numbers of stain tests and biopsy samples per ECPB, and the rate of stains per biopsy by specialty status. Table 2 summarizes the profile of patients with cancer by specialty type. Table 3 summarizes the percentage of patients with IHC workup by diagnosis, Table 4 summarizes the rate of ordering by diagnosis, and Table 5 summarizes the rate of ordering in cancer cases stratified by Gleason grade, disease extent, and perineural invasion. The percentage of patients for whom any IHC stain tests were ordered was similar in all diagnostic categories between the 2 groups (Table 3); however, GU pathologists ordered IHC stain tests at a significantly lower rate in benign (0.06 versus 0.19, P ¼ .02) and cancer (0.16 versus 0.38, P , .001) cases than did nonGU pathologists (Table 4). The rate of ordering was lowest in benign cases for both GU and non-GU pathologists and highest in ASAP cases for both GU and non-GU pathologists. Rate of ordering was not different in HGPIN
Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al 1631
MATERIALS AND METHODS Consecutive cases of ECPBs at the University of Massachusetts Medical School (UMASS) in Worcester, Massachusetts, between the years 2009–2011 were identified by computer search. UMASS is a large tertiary academic medical center with a partially specialized pathology sign-out model. Some ECPBs are signed out as part of the general service and some are signed out as part of a specialized genitourinary service. Consultation cases and cases with fewer than 6 biopsies were excluded. The number of biopsies for each case and the number of IHC stain tests ordered for each case were determined by billing data. For each case the diagnosis and the pathologist rendering the diagnosis were recorded. The diagnoses were mapped to 1 of 4 categories: cancer, atypical or suggestive of cancer (ASAP), high-grade prostatic intraepithelial neoplasia (HGPIN), and benign. Each ECPB was classified according to the highest lesion diagnosed in any of the parts of the ECPB specimen (cancer . ASAP . HGPIN . benign). For cancer cases, the highest Gleason grade, extent of disease, and presence of perineural invasion was recorded. Disease extent was mapped into the following categories: unilateral 1 (1 positive core, 10% involvement), unilateral 2 (1 positive core, .10% involvement), unilateral 3 (multiple positive, ipsilateral cores), bilateral 1 (bilateral positive cores, at least 1 side with only 1 positive core), and bilateral 2 (bilateral positive cores, both sides with multiple positive cores). Pathologists were grouped into 2 groups: GU specialists and nonspecialists (non-GU). Genitourinary pathologists had either extra training in GU pathology, participated in GU tumor board, or both, and participated in specialized GU sign-out.
Statistical Analyses
Table 2.
Summary of Profile of Patients With Cancer (N ¼ 281) GU (n ¼ 179)
Gleason grade, n (%) 3þ3¼6 3þ4¼7 4þ3¼7 Gleason grade 8–10 Not graded Extent of cancer,a n (%) Unilateral 1 Unilateral 2 Unilateral 3 Bilateral 1 Bilateral 2 Perineural invasion, n (%) Maximum involvement,b %
Non-GU (n ¼ 102)
P Value
68 37 29 43 2
(38.0) (20.7) (16.2) (24.3) (1.1)
54 21 14 13 0
(52.9) (20.6) (13.7) (12.7) (0.0)
.11
26 14 46 34 59 70
(14.5) (7.8) (25.7) (19.0) (33.0) (39.1) 49.0
17 7 37 17 24 26
(16.7) (6.9) (36.3) (16.7) (23.5) (25.5) 41.0
.30
.02; .05
Abbreviations: GU, genitourinary pathologist; non-GU, general pathologist. a Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 1 ¼ bilateral positive cores, at least 1 side with only 1 core involved; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores. b Maximum involvement is the percent involvement of the core with the highest percentage of cancer for each patient with cancer.
and ASAP cases between the 2 groups. Among cancer cases the rate of ordering was highest for Gleason score 3 þ 3 ¼ 6 relative to all other grades of cancers. The rate of ordering IHC stain tests was significantly lower for GU pathologists in cases with highest Gleason grade 3 þ 3 ¼ 6 (0.28 versus 0.55, P ¼ .01). Rates were not significantly different between the 2 groups in other Gleason grades of cancer. Genitourinary pathologists had a significantly lower ordering rate in unilateral 1 (0.33 versus 0.66, P ¼ .01) and bilateral 2 (0.05 versus 0.26, P ¼ .02) cancers. Genitourinary pathologists had a significantly lower ordering rate in cancer cases without perineural invasion (0.22 versus 0.44, P ¼ .01). Poisson regressions for the rate of stain usage per biopsy on the various predictors are shown in Table 6. The number of biopsy samples per ECPB was not associated with the rate of ordering in either group. The effect of the various predictors on the rate of IHC ordering was similar in direction and magnitude for both groups. COMMENT In our single-institution study of 618 consecutive ECPBs, we found significant differences in utilization of IHC stains by genitourinary pathologists as compared to general pathologists. Overall, the number of stain tests ordered per case and the rate of IHC stains per biopsy was significantly lower in the GU group. The rate of ordering was most disparate between the 2 groups in benign and cancerous cases, while cases reported as HGPIN or
Table 3. Percentage of Patients With Immunohistochemical Workup by Diagnosis GU, % Diagnosis Benign HGPIN ASAP Cancer
13 60 96 33
Non-GU, % 13 60 89 43
Table 4.
P Value .98 ..99 .40 .09
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. 1632 Arch Pathol Lab Med—Vol 137, November 2013
suspicious had similarly high rates of IHC usage in the 2 groups. In cancer cases, GU pathologists ordered stain tests significantly less often in cases with highest Gleason score of 3 þ 3 ¼ 6, focal disease (1 positive core, ,10% involvement), and for patients with bilateral disease with multiple positive cores on both sides. The difference in ordering between the 2 groups in 3 þ 3 ¼ 6 disease and in focal disease probably represents greater confidence of GU pathologists in diagnosing small foci of prostate cancer on needle biopsy. The difference between the 2 groups, observed with multiple positive cores bilaterally (eg, extensive disease), probably reflects a greater awareness by GU pathologists of clinical settings where workup of a small focus is not necessary owing to minimal impact on clinical decision making. Several of the parameters we measured influenced the rate of ordering in a predictable fashion. Not surprisingly, the diagnosis of ASAP had the highest rate of ordering in both groups and the diagnosis of benign lesion had the lowest rate of ordering in both groups. In cancer cases, highest Gleason score 3 þ 3 ¼ 6 had the highest rate of ordering in both groups. Interestingly, Gleason score 4 þ 3 ¼ 7 had the lowest rate of ordering in both groups, lower even than Gleason scores 8 through 10. This probably reflects the process of interrogating poorly differentiated tumors with neuroendocrine markers or with markers to exclude secondary carcinoma of prostate. Bilateral cancers had lower rates of IHC usage than unilateral cancers in both groups.
Diagnosis Benign HGPIN ASAP Cancer
Rate of Immunohistochemical Usage by Diagnosis GU
Non-GU
P Value
0.06 0.34 0.49 0.16
0.19 0.39 0.58 0.38
.02 .76 .65 ,.001
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Table 5.
Summary of Rate of Immunohistochemical Usage in Cancer Cases
Gleason grade 3þ3¼6 3þ4¼7 4þ3¼7 8–10 Extent of cancera Unilateral 1 Unilateral 2 Unilateral 3 Bilateral 1 Bilateral 2 All unilateral cases All bilateral cases Perineural invasion Not present Present
GU
Non-GU
P Value
0.28 0.10 0.03 0.08
0.55 0.32 0.03 0.14
.01 .11 .86 .41
0.33 0.31 0.15 0.15 0.05 0.23 0.09
0.66 0.18 0.38 0.34 0.26 0.44 0.30
.01 .42 .08 .19 .02 .02 .01
0.22 0.05
0.44 0.20
.01 .07
Abbreviations: GU, genitourinary pathologist; non-GU, general pathologist. a Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 1 ¼ bilateral positive cores, at least 1 side with only 1 core involved; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores.
Poisson regression analyses demonstrated that in both groups the effect of the various measured predictors on the rate was similar in both direction and magnitude. Our results add to the literature in several important ways. Although the utility of the IHC stains in ECPBs is well known, there are very few published data on frequency of usage and/or patterns of usage in specific settings. A recent abstract comparing IHC utilization in prostate needle biopsies between community practices, private laboratories, and academic settings reported that community and private laboratory pathologists order more IHC stain tests than
Table 6.
academic pathologists.10 These findings are interesting but limited by including only cancer cases and by the unavoidable bias of a study of nonconsecutive cases. Our results provide a published rate of IHC usage for a large number of consecutive cases of ECPB at an academic center in the prostate-specific antigen screening era. Our center is also unique in that we have a partially specialized sign-out system: some ECPBs are part of specialized genitourinary sign-out and some are part of a general service. Our study period also follows the publication of seminal and confirmatory studies demonstrating the utility of basal cell markers, AMACR staining, and combination cocktails; thus, our study falls in the ‘‘triple-stain’’ era. The average number of IHC stain tests ordered per case was 1.3 and 2.2 for GU and non-GU pathologists, respectively. Considering there are 3 IHC units billed per triple staining, on average, less than 1 triple staining was ordered per ECPB over the course of the entire study at our institution. Note that our study does not separate triple-stain orders from other immunostain orders; thus, our rate slightly overestimates, but generates a reasonable proxy for, triple-stain usage. Our study also suggests that guidelines or criteria to define when an IHC workup is indicated, and when it is not indicated, would be useful. Anecdotally, we see many cases of ECPBs in our GU subspecialty sign-out that are referred for confirmation before surgery, with IHC workups of equivocal foci that in our opinion are not necessary owing to findings in other cores. Perhaps if there were clear guidelines defining when an equivocal focus in ECPBs does not require workup, owing to lack of clinical impact, unnecessary IHC workups would be reduced. In summary, our single-institution study of consecutive extended core prostate biopsies showed that significant differences exist in the usage of immunohistochemical stains between genitourinary and general pathologists in extended core prostate biopsies. Continuing and targeted education, as well as specific guidelines on best usage of ancillary studies, could be of benefit in the future.
Poisson Regression for the Rate of Immunohistochemical Usage on Predictors (N ¼ 618)
Variable
GU (n ¼ 254) IRR (95% CI)
Non-GU (n ¼ 364) IRR (95% CI)
P Valuea
No. of biopsies Diagnosis Benign HGPIN ASAP Cancer Gleason grade 3þ3¼6 3þ4¼7 4þ3¼7 8–10 Extent of diseaseb Unilateral 1 Unilateral 2 Unilateral 3 All unilateralc Bilateral 2d Perineural invasion
0.97 (0.87–1.09)
1.01 (0.90–1.12)
.69
1.0 6.01 8.68 2.77
(reference) (1.98–18.31) (3.62–20.79) (1.26–6.09)
1.0 2.01 3.03 1.98
(reference) (0.69–5.87) (1.18–7.76) (1.20–3.26)
.16 .11 .48
3.35 1.21 0.33 1.0
(1.08–10.35) (0.29–5.16) (0.03–3.73) (reference)
3.86 2.25 0.23 1.0
(0.87–17.04) (0.44–11.43) (0.01–5.81) (reference)
1.0 0.94 0.46 2.72 0.31 0.23
(reference) (0.30–2.98) (0.17–1.24) (1.20–6.21) (0.07–1.32) (0.08–0.70)
1.0 0.27 0.57 1.47 0.78 0.44
(reference) (0.04–1.71) (0.26–1.25) (0.75–2.91) (0.25–2.41) (0.17–1.12)
.88 .58 .87
.26 .73 .26 .32 .39
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; CI, confidence interval; GU, genitourinary pathologist; HGPIN, highgrade prostatic intraepithelial neoplasia; IRR, incidence rate ratio; non-GU, general pathologist. a P value for group difference between GU and non-GU pathologist. b Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores. c Compared with all bilateral. d Compared with bilateral 1. Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al 1633
References 1. Chawla A, Gunderman RB. Defensive medicine: prevalence, implications, and recommendations. Acad Radiol. 2008;15(7):948–949. 2. Marchevsky AM. Evidence-based medicine in pathology: an introduction. Semin Diagn Pathol. 2005;22(2):105–115. 3. Marchevsky AM, Wick MR. Evidence-based medicine, medical decision analysis, and pathology. Hum Pathol. 2004;35(10):1179–1188. 4. Wick MR, Marchevsky AM. Evidence-based principles in pathology: existing problem areas and the development of ‘‘quality’’ practice patterns. Arch Pathol Lab Med. 2011;135(11):1398–1404. 5. Marchevsky AM, Wick MR. Evidence-based guidelines for the utilization of immunostains in diagnostic pathology: pulmonary adenocarcinoma versus mesothelioma. Appl Immunohistochem Mol Morphol. 2007;15(2):140–144.
1634 Arch Pathol Lab Med—Vol 137, November 2013
6. Wojno KJ, Epstein JI. The utility of basal cell-specific anti-cytokeratin antibody (34 beta E12) in the diagnosis of prostate cancer: a review of 228 cases. Am J Surg Pathol. 1995;19(3):251–260. 7. Hedrick L, Epstein JI. Use of keratin 903 as an adjunct in the diagnosis of prostate carcinoma. Am J Surg Pathol. 1989;13(5):389–396. 8. Jiang Z, Woda BA, Rock KL, et al. P504S: a new molecular marker for the detection of prostate carcinoma. Am J Surg Pathol. 2001;25(11):1397–1404. 9. Rubin MA, Zhou M, Dhanasekaran SM, et al. alpha-Methylacyl coenzyme A racemase as a tissue biomarker for prostate cancer. JAMA. 2002;287(13):1662– 1670. 10. Robinson B, Yantiss RK. Practice-based differences in ancillary stain usage when evaluating prostate-needle core biopsies. Mod Pathol. 2012;25:237A.
Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Copyright of Archives of Pathology & Laboratory Medicine is the property of College of American Pathologists and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Context.—Immunohistochemical (IHC) stains have known utility in prostate biopsies and are widely used to augment routine staining in difficult cases. Patterns in IHC utilization and differences based on pathologist training and experience is understudied in the peer-reviewed literature. Objectives.—To compare the rates of IHC usage between specialized (genitourinary; [GU]) and nonspecialized (non-GU) pathologists in extended core prostate biopsies (ECPBs) and the effects of diagnosis; and in cancer cases Gleason grade, disease extent, and perineural invasion on the rate. Design.—Consecutive ECPBs from 2009–2011 were identified and billing data were used to determine the number of biopsies and IHC stains per case. Diagnoses were mapped and in cancer cases, Gleason grade, extent of disease, and perineural invasion were recorded. Pathologists were classified as GU or non-GU on the basis of training and experience. Results.—A total of 618 ECPBs were included in the
study. Genitourinary pathologists ordered significantly fewer IHC tests per case and per biopsy than non-GU pathologists. The rate of ordering was most disparate for biopsies of cancerous and benign lesions. For biopsies of cancerous lesions, high-grade cancer, bilateral disease, and perineural invasion decreased the rate of ordering in both groups. In cancer cases, GU pathologists ordered significantly fewer stain tests for highest Gleason grade of 3 þ 3 ¼ 6, for patients with focal disease and for patients with multiple positive bilateral cores. The effect of the various predictors on IHC ordering rates was similar in both groups. Conclusions.—Genitourinary pathologists ordered significantly fewer IHC stain tests than non-GU pathologists in ECPBs. Guidelines to define when IHC workup is necessary and not necessary may be helpful to guide workups. (Arch Pathol Lab Med. 2013;137:1630–1634; doi: 10.5858/arpa.2012-0517-OA)
I
stain test for diagnostic purposes is based on the judgment of the individual pathologist. Unlike the situation that many of our clinical colleagues face when ordering diagnostic tests, there are no established criteria that define situations where an IHC workup is not justified or conversely, when it should be considered mandatory. This may lead to excessive or unnecessary IHC workups or omissions of critical IHC staining. In recent years, there has been increasing recognition of the need for evidence-based approaches in all fields of medicine, including pathology.1–5 With our nation’s health care costs rapidly rising, the use of efficient, evidence-based, and cost-effective diagnostic approaches is of utmost importance. Evidence-based guidelines can help direct physicians toward tests that have substantial impact on patient diagnosis and treatment and steer physicians away from those that do not. They can also direct physicians as to when to use particular tests. Unfortunately, firmly established guidelines for the use of IHC stains are lacking. In this study, we investigated factors that may influence the usage of IHC stains in extended core prostate biopsies (ECPBs). In this specific setting the most common IHC
mmunohistochemical (IHC) stains are important ancillary tests frequently used by anatomic pathologists to complement findings on routine staining. Immunohistochemical stains are useful for diagnostic purposes (differentiating benign from malignant processes, subtyping tumors, etc) and therapeutic purposes (identifying tumors appropriate for targeted therapeutic agents). However, IHC stains increase costs and turnaround time and therefore should be used judiciously. The decision to order an IHC Accepted for publication December 11, 2012. From the Department of Pathology, University of California San Francisco, San Francisco (Dr Plourde); the Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School and the Institute for Aging Research, Hebrew SeniorLife, Boston, Massachusetts (Dr Gross); and the Department of Pathology, University of Massachusetts Medical School, Worcester (Dr Jiang and Dr Owens). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Christopher L. Owens, MD, Department of Pathology, University of Massachusetts Medical School, Three Biotech, One Innovation Drive, Worcester, MA 01605 (e-mail: Christopher. [email protected]). 1630 Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Table 1.
Summary of Descriptive Statistics of Patients (N ¼ 618) GU (n ¼ 364)
Diagnosis, n (%) Benign HGPIN ASAP Cancer No. of stains per ECPB, mean (SD) No. of biopsies per ECPB, mean (SD) Rate of stains/biopsy, mean (SD)
142 15 28 179 1.3 9.1 0.1
(39.0) (4.1) (7.7) (49.2) (2.3) (2.5) (0.3)
Non-GU (n ¼ 254) 133 10 9 102 2.2 7.7 0.3
(52.4) (3.9) (3.5) (40.2) (5.1) (2.1) (0.7)
P Value .01
.004 ,.001 ,.001
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; ECPB, extended core prostate biopsy; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. Bolded values are statistically significant.
stains ordered are stains for basal cells, including 34bE12 (cytokeratin 903, CK903) and p63 protein, as well as amethylacyl coenzyme A racemase (AMACR, p504s), a biomarker overexpressed in prostate cancer. These stains have known utility in complementing routine stains to aid in confirming or refuting small atypical glandular foci as cancer.6–9 The 3 stains are commercially available as a cocktail with 2 chromagens, to facilitate evaluation of small foci with 1 tissue level, and are widely used by pathologists who interpret prostate needle biopsy findings. Less commonly, other IHC stains may be used in poorly differentiated tumors to aid in classification or to exclude/confirm secondary malignancy in ECPBs. For this study we divided the pathologists into 2 groups, genitourinary specialists and nonspecialists, by background and training. Immunohistochemical stain utilization in ECPBs was compared between the 2 groups. The effects of diagnosis and tumor grade, tumor extent, and perineural invasion in cancer cases were also analyzed in this study.
number of ordered stain tests by the number of biopsies for each ECPB. The rate was regressed separately on predictors, including diagnosis, number of biopsies, Gleason grade, extent of cancer, and presence of perineural invasion in cancer cases, by using Poisson regression with a log-linear link function to account for nonnormal distribution of rates. To test whether associations differed by specialty status, we included statistical interactions between specialty status and each variable in separate regressions. Associations with P values ,.05 were considered significantly different. Analyses were conducted with Stata statistical software, version 12.1 (StataCorp, College Station, Texas).
We calculated descriptive characteristics of the sample by using v2 tests for categorical variables and t tests for continuous variables. We calculated the rate of IHC stain tests per biopsy by dividing the
RESULTS In total, 618 ECPBs were identified and included in the study. Of these, 364 (58.9%) were interpreted by a GU pathologist and 254 (41.1%) were interpreted by a non-GU pathologist (Table 1). The distribution of diagnoses differed between the GU pathologists and the non-GU pathologists (P ¼ .01): GU pathologists diagnosed cancer, ASAP, HGPIN, and benign lesion in 49.2%, 7.7%, 4.1%, and 39.0% of cases, respectively. Non-GU pathologists diagnosed cancer, ASAP, HGPIN, and benign lesion in 40.2%, 3.5%, 3.9%, and 52.4% of cases, respectively. Among patients with cancer there was no difference for the highest Gleason grade (P ¼ .11) or extent of cancer (P ¼ .30), but perineural invasion was identified in a significantly higher percentage of cancer cases by GU pathologists (39.1% versus 25.5%, P ¼ .02) (Table 2). The number of biopsy samples per ECPB was significantly higher in the GU group (9.1 versus 7.7, P , .001) (Table 1). The GU pathologists ordered significantly fewer stain tests per ECPB (1.3 versus 2.2, P ¼ .004) and stain tests per biopsy (0.1 versus 0.3, P , .001). Table 1 summarizes the distribution of diagnoses, numbers of stain tests and biopsy samples per ECPB, and the rate of stains per biopsy by specialty status. Table 2 summarizes the profile of patients with cancer by specialty type. Table 3 summarizes the percentage of patients with IHC workup by diagnosis, Table 4 summarizes the rate of ordering by diagnosis, and Table 5 summarizes the rate of ordering in cancer cases stratified by Gleason grade, disease extent, and perineural invasion. The percentage of patients for whom any IHC stain tests were ordered was similar in all diagnostic categories between the 2 groups (Table 3); however, GU pathologists ordered IHC stain tests at a significantly lower rate in benign (0.06 versus 0.19, P ¼ .02) and cancer (0.16 versus 0.38, P , .001) cases than did nonGU pathologists (Table 4). The rate of ordering was lowest in benign cases for both GU and non-GU pathologists and highest in ASAP cases for both GU and non-GU pathologists. Rate of ordering was not different in HGPIN
Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al 1631
MATERIALS AND METHODS Consecutive cases of ECPBs at the University of Massachusetts Medical School (UMASS) in Worcester, Massachusetts, between the years 2009–2011 were identified by computer search. UMASS is a large tertiary academic medical center with a partially specialized pathology sign-out model. Some ECPBs are signed out as part of the general service and some are signed out as part of a specialized genitourinary service. Consultation cases and cases with fewer than 6 biopsies were excluded. The number of biopsies for each case and the number of IHC stain tests ordered for each case were determined by billing data. For each case the diagnosis and the pathologist rendering the diagnosis were recorded. The diagnoses were mapped to 1 of 4 categories: cancer, atypical or suggestive of cancer (ASAP), high-grade prostatic intraepithelial neoplasia (HGPIN), and benign. Each ECPB was classified according to the highest lesion diagnosed in any of the parts of the ECPB specimen (cancer . ASAP . HGPIN . benign). For cancer cases, the highest Gleason grade, extent of disease, and presence of perineural invasion was recorded. Disease extent was mapped into the following categories: unilateral 1 (1 positive core, 10% involvement), unilateral 2 (1 positive core, .10% involvement), unilateral 3 (multiple positive, ipsilateral cores), bilateral 1 (bilateral positive cores, at least 1 side with only 1 positive core), and bilateral 2 (bilateral positive cores, both sides with multiple positive cores). Pathologists were grouped into 2 groups: GU specialists and nonspecialists (non-GU). Genitourinary pathologists had either extra training in GU pathology, participated in GU tumor board, or both, and participated in specialized GU sign-out.
Statistical Analyses
Table 2.
Summary of Profile of Patients With Cancer (N ¼ 281) GU (n ¼ 179)
Gleason grade, n (%) 3þ3¼6 3þ4¼7 4þ3¼7 Gleason grade 8–10 Not graded Extent of cancer,a n (%) Unilateral 1 Unilateral 2 Unilateral 3 Bilateral 1 Bilateral 2 Perineural invasion, n (%) Maximum involvement,b %
Non-GU (n ¼ 102)
P Value
68 37 29 43 2
(38.0) (20.7) (16.2) (24.3) (1.1)
54 21 14 13 0
(52.9) (20.6) (13.7) (12.7) (0.0)
.11
26 14 46 34 59 70
(14.5) (7.8) (25.7) (19.0) (33.0) (39.1) 49.0
17 7 37 17 24 26
(16.7) (6.9) (36.3) (16.7) (23.5) (25.5) 41.0
.30
.02; .05
Abbreviations: GU, genitourinary pathologist; non-GU, general pathologist. a Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 1 ¼ bilateral positive cores, at least 1 side with only 1 core involved; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores. b Maximum involvement is the percent involvement of the core with the highest percentage of cancer for each patient with cancer.
and ASAP cases between the 2 groups. Among cancer cases the rate of ordering was highest for Gleason score 3 þ 3 ¼ 6 relative to all other grades of cancers. The rate of ordering IHC stain tests was significantly lower for GU pathologists in cases with highest Gleason grade 3 þ 3 ¼ 6 (0.28 versus 0.55, P ¼ .01). Rates were not significantly different between the 2 groups in other Gleason grades of cancer. Genitourinary pathologists had a significantly lower ordering rate in unilateral 1 (0.33 versus 0.66, P ¼ .01) and bilateral 2 (0.05 versus 0.26, P ¼ .02) cancers. Genitourinary pathologists had a significantly lower ordering rate in cancer cases without perineural invasion (0.22 versus 0.44, P ¼ .01). Poisson regressions for the rate of stain usage per biopsy on the various predictors are shown in Table 6. The number of biopsy samples per ECPB was not associated with the rate of ordering in either group. The effect of the various predictors on the rate of IHC ordering was similar in direction and magnitude for both groups. COMMENT In our single-institution study of 618 consecutive ECPBs, we found significant differences in utilization of IHC stains by genitourinary pathologists as compared to general pathologists. Overall, the number of stain tests ordered per case and the rate of IHC stains per biopsy was significantly lower in the GU group. The rate of ordering was most disparate between the 2 groups in benign and cancerous cases, while cases reported as HGPIN or
Table 3. Percentage of Patients With Immunohistochemical Workup by Diagnosis GU, % Diagnosis Benign HGPIN ASAP Cancer
13 60 96 33
Non-GU, % 13 60 89 43
Table 4.
P Value .98 ..99 .40 .09
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. 1632 Arch Pathol Lab Med—Vol 137, November 2013
suspicious had similarly high rates of IHC usage in the 2 groups. In cancer cases, GU pathologists ordered stain tests significantly less often in cases with highest Gleason score of 3 þ 3 ¼ 6, focal disease (1 positive core, ,10% involvement), and for patients with bilateral disease with multiple positive cores on both sides. The difference in ordering between the 2 groups in 3 þ 3 ¼ 6 disease and in focal disease probably represents greater confidence of GU pathologists in diagnosing small foci of prostate cancer on needle biopsy. The difference between the 2 groups, observed with multiple positive cores bilaterally (eg, extensive disease), probably reflects a greater awareness by GU pathologists of clinical settings where workup of a small focus is not necessary owing to minimal impact on clinical decision making. Several of the parameters we measured influenced the rate of ordering in a predictable fashion. Not surprisingly, the diagnosis of ASAP had the highest rate of ordering in both groups and the diagnosis of benign lesion had the lowest rate of ordering in both groups. In cancer cases, highest Gleason score 3 þ 3 ¼ 6 had the highest rate of ordering in both groups. Interestingly, Gleason score 4 þ 3 ¼ 7 had the lowest rate of ordering in both groups, lower even than Gleason scores 8 through 10. This probably reflects the process of interrogating poorly differentiated tumors with neuroendocrine markers or with markers to exclude secondary carcinoma of prostate. Bilateral cancers had lower rates of IHC usage than unilateral cancers in both groups.
Diagnosis Benign HGPIN ASAP Cancer
Rate of Immunohistochemical Usage by Diagnosis GU
Non-GU
P Value
0.06 0.34 0.49 0.16
0.19 0.39 0.58 0.38
.02 .76 .65 ,.001
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; GU, genitourinary pathologist; HGPIN, high-grade prostatic intraepithelial neoplasia; non-GU, general pathologist. Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Table 5.
Summary of Rate of Immunohistochemical Usage in Cancer Cases
Gleason grade 3þ3¼6 3þ4¼7 4þ3¼7 8–10 Extent of cancera Unilateral 1 Unilateral 2 Unilateral 3 Bilateral 1 Bilateral 2 All unilateral cases All bilateral cases Perineural invasion Not present Present
GU
Non-GU
P Value
0.28 0.10 0.03 0.08
0.55 0.32 0.03 0.14
.01 .11 .86 .41
0.33 0.31 0.15 0.15 0.05 0.23 0.09
0.66 0.18 0.38 0.34 0.26 0.44 0.30
.01 .42 .08 .19 .02 .02 .01
0.22 0.05
0.44 0.20
.01 .07
Abbreviations: GU, genitourinary pathologist; non-GU, general pathologist. a Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 1 ¼ bilateral positive cores, at least 1 side with only 1 core involved; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores.
Poisson regression analyses demonstrated that in both groups the effect of the various measured predictors on the rate was similar in both direction and magnitude. Our results add to the literature in several important ways. Although the utility of the IHC stains in ECPBs is well known, there are very few published data on frequency of usage and/or patterns of usage in specific settings. A recent abstract comparing IHC utilization in prostate needle biopsies between community practices, private laboratories, and academic settings reported that community and private laboratory pathologists order more IHC stain tests than
Table 6.
academic pathologists.10 These findings are interesting but limited by including only cancer cases and by the unavoidable bias of a study of nonconsecutive cases. Our results provide a published rate of IHC usage for a large number of consecutive cases of ECPB at an academic center in the prostate-specific antigen screening era. Our center is also unique in that we have a partially specialized sign-out system: some ECPBs are part of specialized genitourinary sign-out and some are part of a general service. Our study period also follows the publication of seminal and confirmatory studies demonstrating the utility of basal cell markers, AMACR staining, and combination cocktails; thus, our study falls in the ‘‘triple-stain’’ era. The average number of IHC stain tests ordered per case was 1.3 and 2.2 for GU and non-GU pathologists, respectively. Considering there are 3 IHC units billed per triple staining, on average, less than 1 triple staining was ordered per ECPB over the course of the entire study at our institution. Note that our study does not separate triple-stain orders from other immunostain orders; thus, our rate slightly overestimates, but generates a reasonable proxy for, triple-stain usage. Our study also suggests that guidelines or criteria to define when an IHC workup is indicated, and when it is not indicated, would be useful. Anecdotally, we see many cases of ECPBs in our GU subspecialty sign-out that are referred for confirmation before surgery, with IHC workups of equivocal foci that in our opinion are not necessary owing to findings in other cores. Perhaps if there were clear guidelines defining when an equivocal focus in ECPBs does not require workup, owing to lack of clinical impact, unnecessary IHC workups would be reduced. In summary, our single-institution study of consecutive extended core prostate biopsies showed that significant differences exist in the usage of immunohistochemical stains between genitourinary and general pathologists in extended core prostate biopsies. Continuing and targeted education, as well as specific guidelines on best usage of ancillary studies, could be of benefit in the future.
Poisson Regression for the Rate of Immunohistochemical Usage on Predictors (N ¼ 618)
Variable
GU (n ¼ 254) IRR (95% CI)
Non-GU (n ¼ 364) IRR (95% CI)
P Valuea
No. of biopsies Diagnosis Benign HGPIN ASAP Cancer Gleason grade 3þ3¼6 3þ4¼7 4þ3¼7 8–10 Extent of diseaseb Unilateral 1 Unilateral 2 Unilateral 3 All unilateralc Bilateral 2d Perineural invasion
0.97 (0.87–1.09)
1.01 (0.90–1.12)
.69
1.0 6.01 8.68 2.77
(reference) (1.98–18.31) (3.62–20.79) (1.26–6.09)
1.0 2.01 3.03 1.98
(reference) (0.69–5.87) (1.18–7.76) (1.20–3.26)
.16 .11 .48
3.35 1.21 0.33 1.0
(1.08–10.35) (0.29–5.16) (0.03–3.73) (reference)
3.86 2.25 0.23 1.0
(0.87–17.04) (0.44–11.43) (0.01–5.81) (reference)
1.0 0.94 0.46 2.72 0.31 0.23
(reference) (0.30–2.98) (0.17–1.24) (1.20–6.21) (0.07–1.32) (0.08–0.70)
1.0 0.27 0.57 1.47 0.78 0.44
(reference) (0.04–1.71) (0.26–1.25) (0.75–2.91) (0.25–2.41) (0.17–1.12)
.88 .58 .87
.26 .73 .26 .32 .39
Abbreviations: ASAP, atypical acinar proliferation or suspicious for cancer; CI, confidence interval; GU, genitourinary pathologist; HGPIN, highgrade prostatic intraepithelial neoplasia; IRR, incidence rate ratio; non-GU, general pathologist. a P value for group difference between GU and non-GU pathologist. b Unilateral 1 ¼ 1 positive core, 10% involved; Unilateral 2 ¼ 1 positive core, .10% involved; Unilateral 3 ¼ multiple positive ipsilateral cores; Bilateral 2 ¼ bilateral positive cores, both sides with multiple positive cores. c Compared with all bilateral. d Compared with bilateral 1. Arch Pathol Lab Med—Vol 137, November 2013
Immunohistochemical Usage in Prostate Biopsies––Plourde et al 1633
References 1. Chawla A, Gunderman RB. Defensive medicine: prevalence, implications, and recommendations. Acad Radiol. 2008;15(7):948–949. 2. Marchevsky AM. Evidence-based medicine in pathology: an introduction. Semin Diagn Pathol. 2005;22(2):105–115. 3. Marchevsky AM, Wick MR. Evidence-based medicine, medical decision analysis, and pathology. Hum Pathol. 2004;35(10):1179–1188. 4. Wick MR, Marchevsky AM. Evidence-based principles in pathology: existing problem areas and the development of ‘‘quality’’ practice patterns. Arch Pathol Lab Med. 2011;135(11):1398–1404. 5. Marchevsky AM, Wick MR. Evidence-based guidelines for the utilization of immunostains in diagnostic pathology: pulmonary adenocarcinoma versus mesothelioma. Appl Immunohistochem Mol Morphol. 2007;15(2):140–144.
1634 Arch Pathol Lab Med—Vol 137, November 2013
6. Wojno KJ, Epstein JI. The utility of basal cell-specific anti-cytokeratin antibody (34 beta E12) in the diagnosis of prostate cancer: a review of 228 cases. Am J Surg Pathol. 1995;19(3):251–260. 7. Hedrick L, Epstein JI. Use of keratin 903 as an adjunct in the diagnosis of prostate carcinoma. Am J Surg Pathol. 1989;13(5):389–396. 8. Jiang Z, Woda BA, Rock KL, et al. P504S: a new molecular marker for the detection of prostate carcinoma. Am J Surg Pathol. 2001;25(11):1397–1404. 9. Rubin MA, Zhou M, Dhanasekaran SM, et al. alpha-Methylacyl coenzyme A racemase as a tissue biomarker for prostate cancer. JAMA. 2002;287(13):1662– 1670. 10. Robinson B, Yantiss RK. Practice-based differences in ancillary stain usage when evaluating prostate-needle core biopsies. Mod Pathol. 2012;25:237A.
Immunohistochemical Usage in Prostate Biopsies––Plourde et al
Copyright of Archives of Pathology & Laboratory Medicine is the property of College of American Pathologists and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
