Concordance with breast cancer pathology reporting practice guidelines

Concordance with Breast Cancer Pathology Reporting Practice Guidelines Neal W Wilkinson, MD, FACS, Azin Shahryarinejad, MD, Janet S Winston, MD, Nancy Watroba, MPA, Stephen B Edge, MD, FACS Accurate pathology reporting is important for treatment of breast cancer. The College of American Pathologists (CAP) distributed guidelines for reporting cancer specimens in 1998. The aim of this study was to determine community-wide concordance with CAP breast cancer reporting guidelines. STUDY DESIGN: Pathology reporting of stage I and II breast cancers was examined for adherence to CAP guidelines. Pathology reports were reviewed from 100 consecutive cases of invasive breast cancers referred to Roswell Park Cancer Institute in 1998 to 1999 from community hospitals after excisional breast biopsy and 20 consecutive cases with excisional biopsy at RPCI. Adherence to CAP guidelines for clinically relevant items was determined from the original pathology report in each case. RESULTS: One hundred one cases met the inclusion criteria. Most reports did not include at least one of the guideline required elements. Surgical margins were inked in only 77%, and the margins oriented in only 25% of patients. Many specimens were not oriented by the surgeon. Grade was reported in most cases, but the Bloom Scarf Richardson grade was reported in only 6%. The presence or absence of lymphovascular invasion, and of coexisting in situ disease, was reported in 57% and 71%, respectively. The extent and type of in situ disease was reported in 47% and 49%, respectively. CONCLUSIONS: Breast cancer pathology reporting varies widely. Key elements that affect treatment are often omitted. These include gross description and size, orientation and involvement of surgical margins, and description of histologic features, including Bloom Scarf Richardson reporting of grade and the extent of an in situ component. Passive distribution of CAP practice guidelines might be insufficient to accomplish community-wide quality improvement in breast pathology reporting. ( J Am Coll Surg 2003;196:38-43. © 2003 by the American College of Surgeons) BACKGROUND:

Proper medical and surgical treatment of cancer requires a clear understanding of the extent of disease at the time of initial presentation. Although initial physical examination and radiographic findings provide information for clinical staging, most treatment decisions are based on pathologic staging with information derived from the pathology report. Breast cancer treatment decisions require accurate and complete information on specimen

and tumor description, orientation and analysis of surgical margins, and full reporting of histologic features. Unfortunately, not all pathology reports contain the information needed for clinical decision-making. To improve the quality of pathology reporting, the College of American Pathologists (CAP) developed and published guidelines designed to standardize the reporting of surgical specimens for all malignancies.1 The guidelines provide templates by organ site and by type of surgical specimen. The guidelines were published in September 1998 and widely distributed throughout the pathology community in America, but there was no requirement for their mandatory use. To date, there has been no evaluation of the extent to which these guidelines have been implemented in practice. The purpose of this study was to determine the degree to which breast cancer pathology reporting adheres to the CAP practice guideline recommendations in western New York.

No competing interests declared.

Abstract and poster presented at the 55th Annual Cancer Symposium of the Society of Surgical Oncology, Denver, CO, March 2002. Received June 26, 2002; Revised August 7, 2002; Accepted August 7, 2002. From the Department of Surgical Oncology (Wilkinson, Watroba, Edge), and Department of Pathology (Winston), Roswell Park Cancer Institute, Buffalo, NY and the State University of New York at Buffalo, Buffalo, NY (Shahryarinejad). Correspondence address: Stephen B Edge, MD, FACS, Surgical Oncology, Roswell Park Cancer Institute, Elm & Carlton St, Buffalo, NY, 14263.

© 2003 by the American College of Surgeons Published by Elsevier Science Inc.

38

ISSN 1072-7515/03/$21.00 PII S1072-7515(02)01627-7

Vol. 196, No. 1, January 2003

METHODS Patients for whom breast cancer was diagnosed using excisional breast biopsy as the first procedure were identified from the breast cancer database in the Department of Surgery at Roswell Park Cancer Institute (RPCI). One hundred consecutive patients in 1998 and 1999 were identified for whom the excisional biopsy was performed before referral to RPCI for further treatment. These patients are defined as community hospital (CH) patients. All patients underwent open excisional biopsy before a known diagnosis of cancer. Patients were excluded if simultaneous axillary staging or mastectomy were performed. Patients were excluded if a preceding fine-needle aspiration or a core-needle biopsy provided a cancer diagnosis. To allow review of as homogeneous a population as possible, the study was limited to patients whose diagnosis showed infiltrating carcinoma, excluding those with pure in situ carcinoma. The study was approved by the RPCI institutional review board. RPCI established a standardized pathology-reporting template for breast cancer in 1994. The template was developed and implemented before the release of the CAP guideline recommendations, but contains many of the key features in the guideline. For comparison, 20 similar consecutive patients were identified for whom the excisional biopsy was performed at RPCI during the same timeframe. We chose to examine our own pathology reporting to determine whether it was concordant with the CAP guidelines and to serve as a cancer center comparison with community practice. Completeness of reporting of invasive breast cancer in excisional biopsy specimens was examined for items identified in the applicable CAP guideline on breast cancer entitled, “Complete excision of tumor less than total mastectomy with or without axillary contents.”1 Each CAP guideline has three sections: clinical information, macroscopic examination, and microscopic examination. The elements in each section for the breast cancer excision CAP guideline are shown in Table 1. For this article, concordance was evaluated only for those elements felt to be clinically important in the management of breast cancer (denoted by an asterisk in Table 1). Each item on each report was scored as guideline-concordant (present on pathology report), not concordant (absent from pathology report), or not applicable. Although the results of special studies (estrogen receptor, progesterone receptor, and Her-2) impact on patient care, analysis of

Wilkinson et al

Breast Cancer Reporting Guidelines

39

these data was not included in this study because many pathology departments send specimens to reference laboratories and the results are not always recorded in the pathology report. A single reviewer compiled data from the original pathology reports and the results were validated by the other investigators. Macroscopic examination Size of specimen and tumor

For gross description of the surgical specimen and the tumor, CAP guidelines require reporting size in three dimensions. For this element, the report was scored as showing size in three dimensions, two dimensions, one dimension, or not applicable if the report stated that there was no gross tumor identified. Margin evaluation

For margin evaluation, the guidelines require reporting orientation by the surgeon and the pathologist. Pathology reports were considered concordant with guidelines regarding margin orientation if the margins were properly oriented by the pathologist, recognizing that the pathologist cannot orient the specimen if the surgeon has not properly labeled it. Guidelines require reporting of gross margin status. In cases in which the tumor was not detected grossly, gross margin status and gross size were coded as “not applicable.” For microscopic reporting of margins, pathology reports were considered guideline-concordant if the involved or close margins were mentioned, even in the absence of the mention of orientation in the gross description. Microscopic evaluation

Microscopic description of the tumor entails complete histologic description, grading, and reporting associated subtypes, such as in situ components. For microscopic reporting of tumor size, the terms “microscopic finding” and “extensive multifocal” were considered sufficient for guideline concordance. For histologic reporting, the guidelines require reporting grade using the Bloom Scarf Richardson (BSR) scale. Because this was not used in most cases, we also recorded reporting of nuclear or histologic grade. The histologic description for type and grade were coded as concordant with guidelines even if not recorded in cases of lobular, mucinous, medullary, papillary, or tubular subtypes. The guidelines require reporting of the extent and histologic subtype (eg, solid, cribiform, comedo, etc) of any in situ component. If the in situ component was reported as only LCIS, the histo-

40

Wilkinson et al

Breast Cancer Reporting Guidelines

Table 1. College of American Pathologists Guidelines for Reporting Breast Cancer Specimens: Complete Excision of Tumor Less than Total Mastectomy With or Without Axillary Contents Clinical information Patient identification Responsible physician Date of procedure Relevant history Clinical findings Clinical diagnosis Operative findings Anatomic site of specimen Macroscopic examination Specimen fixation Tissue size in three dimensions* Location of biopsy site Orientation by the surgeon* Results of intraoperative consultation Tumor size in three dimensions* Tumor descriptive features* Correlation with imaging studies Gross margin designation* Margin inking* Orientation* Status of surgical margins* Distance of closest margins* Frozen section Microscopic evaluation Tumor size* Histologic type* Histologic grade* Ductal carcinoma in situ pattern and extent* Microcalcifications Extent of invasion* Bloom Scarf Richardson scale* Blood/lymphatic vessel invasion* Tumor, node, metastasis staging* Margins* Results of special studies *Reviewed items.

logic subtype and extent were considered guideline-concordant. Data are represented using descriptive statistics comparing CH with RPCI. No statistical comparisons were performed. RESULTS Of 100 consecutive CH patients, 83 met the inclusion criteria (excluded cases: 15—DCIS; 1—LCIS;

J Am Coll Surg

1—pathology report not available). The pathology reports came from 33 different CHs. The median number of reports per hospital was two (range 1 to 7). Of the 20 RPCI patients, 18 met the inclusion criteria (excluded: 1—DCIS; 1—not primary excisional biopsy). These reports were examined for concordance with CAP guideline standards for the guideline elements highlighted in Table 1. Size of specimen and tumor

The guidelines require that the gross specimen and tumor sizes be described in three dimensions. Of the 101 pathology reports, the specimen size was reported in three dimensions in 91% (CH 74 of 83 patients, or 89%, and RPCI 18 of 18 patients, or 100%). Reports included the tumor size in three dimensions, or reported “no gross tumor” in 40% (CH 28 of 83 patients, or 34%, and RPCI 12 of 18 patients, or 66%). Many of the nonconcordant reports included the tumor size in either one or two dimensions (Fig. 1). In most of the cases marked as “other,” the maximum tumor dimension was provided only in the microscopic description with no reporting of the gross tumor size. All reports allowed for subsequent tumor, node, metastasis (TNM) staging of tumor size, even when failing to meet CAP guidelines. Orientation of the specimen

Orientation of the specimen for margin analysis, and examination of the margins for involvement with cancer are CAP guideline requirements. The absence of margin orientation resulted from either the surgeon’s failure to orient the specimen, the pathologist’s failure to maintain orientation provided by the surgeon, or could not be attributed to either specialist because of a lack of documentation. The margins of the gross specimen were oriented in only 25% of patients (CH 13 of 83 patients, or 16%, and RPCI 13 of 18 patients, or 72%) (Table 2). The inking standard at RPCI is used when performing multicolor inking on an oriented specimen. In the nonoriented RPCI cases, failure of the surgeon to orient the specimen was reported as such and inked using a single color. Margin evaluation

In the absence of margin orientation, defining concordance with microscopic margin reporting guidelines was problematic. We scored as guideline-concordant microscopic margin reporting irrespective of orientation. The gross margin reporting and specimen inking results are

Wilkinson et al

Vol. 196, No. 1, January 2003

Breast Cancer Reporting Guidelines

41

summarized in Table 2. The status of the microscopic margin analysis was reported in 94% of the total patients (CH 77 of 83 patients, or 93%, and RPCI 18 of 18 patients, or 100%). The distance to the closest margin was reported in 69% of the total patients (CH 58 of 83 patients, or 67%, and RPCI 12 of 18 patients, or 70%). Microscopic evaluation

Adequate microscopic description was reported in the large majority of cases: histology—100%, grade—90%, and size—90%. Grade was only reported as Grade 1, 2, or 3 and was not reported using the guideline-required BSR scale in the large majority of cases. The BSR grading scale was reported in 6% of the total patients (CH 6 of 83 patients, or 7%, and RPCI 0 of 18 patients, or 0%). The presence or absence of lymphovascular invasion (LVI) was reported in 47% of total patients (CH 37 of 83 patients, or 45%, and RPCI 11 of 18 patients, or 61%) (Table 3). The reporting of the in situ component of the cancers varied widely. There was mention of the presence or absence of an in situ component in only 71% of patients (CH 55 of 83 patients, or 67%, and RPCI 17 of 18 patients, or 94%). In those patients with an in situ component reported on CH pathology reports, the extent of the in situ component was described in only 37% and the histologic subtype of DCIS was reported in only 41% of patients (Table 4). The large majority of RPCI reports included an adequate description of the in situ component. DISCUSSION The choice of treatment for solid tumors depends in large part on the pathologic features of the primary tumor, surrounding tissues, and regional lymph nodes. Consistent, complete, and unambiguous pathology reporting is a key component of delivering quality cancer care. There is little data that quantifies the extent and clinical impact of variation in pathology reporting. Our

Figure 1. The guidelines require that the specimen and tumor sizes need to be described in three dimensions. Many of the nonconcordant reports included the tumor size in either one or two dimensions. In most of the cases marked as “other,” the maximum tumor dimension was provided only in the microscopic description with no reporting of the gross tumor size. All reports allowed for subsequent TNM staging despite failing to meet College of American Pathologists guidelines.

study examined the variation in pathology reporting of factors that affect the treatment of breast cancer in one region of the United States. Pathology reports from excision of a primary breast cancer from 100 consecutive patients referred to one cancer center after the excision at a CH were compared with the 1998 pathology reporting guidelines of the CAP. The pathology reports from similar patients who underwent the cancer excision at the cancer center were reviewed for comparison. Each report was examined for

Table 2. Margin Evaluation

Total (101) CH (83) RPCI (18)

Gross examination (%)

Specimen inked (%)

Margin orientation (%)

Micoscopic margin status reported (%)

Distance to closest margin reported (%)

52 49 67

77 75 89

25 16 72

94 93 100

69 67 70

CH, community hospital; RPCI, Roswell Park Cancer Institute.

42

Wilkinson et al

Breast Cancer Reporting Guidelines

J Am Coll Surg

Table 3. Microscopic Evaluation of Primary Tumor

Total (101) CH (83) RPCI (18)

Histologic type (%)

Tumor grade BSR (%)

Tumor grade (%)

Tumor size (%)

Presence/absence LVI (%)

Reporting of TNM staging (%)

100 100 100

6 7 0

90 88 100

90 88 100

47 45 61

9 0 50

BSR, Bloom Scarf Richardson; CH, community hospital; LVI, lymphovascular invasion; RPCI, Roswell Park Cancer Institute.

the presence or absence of data elements required by the CAP guideline, including macroscopic and microscopic descriptions of size, histologic type, grade, margin orientation, margin status, and TNM classification. Most pathology reports lacked at least some of the key clinically relevant information. It is important to recognize that complete reporting of pathologic information is a shared responsibility of the surgeon and pathologist. For example, failure of the surgeon to mark the surgical specimen prohibits the pathologist from orienting margins. The CAP guidelines require documentation of the specimen marking by the surgeon and the orientation by the pathologist in the macroscopic description of the specimen. Nonconcordance with guidelines occurred from both failure to orient the specimen (surgeon) and failure to utilize surgeon orientation (pathologist). Also, the CAP guidelines require reporting of the TNM stage of the tumor. This presents a problem for excisional breast biopsies as the TNM stage remains unknown (pNxMx). The clinical examination (cTNM) is not reported to the pathologist and inclusion of the staging information might be inappropriate. Subsequent surgical treatments (reexcision or axillary staging) might not reveal any local or regional disease. So, the initial procedure and pathology report must provide all the information required to properly stage the disease. Unfortunately, if the initial report lacks needed information, the subsequent final staging might omit key treatment items. In the current American Joint Commission (AJCC) on Table 4. Description of In Situ Component (DCIS/LCIS) Reported presence/ Reported absence of extent of Reported histologic in situ in situ type of in situ component (%) component (%) disease (%)

Total (101) CH (83) RPCI (18)

71 67 94

47 37 94

49 41 88

CH, community hospital; DCIS, ductal carcinoma in situ; LCIS, lobular carcinoma in situ; RPCI, Roswell Park Cancer Institute.

Cancer Staging Handbook, correct classification of T status is dependent on clear reporting of invasive and noninvasive components and clear margin reporting, both macroscopic and microscopic. In addition, the AJCC staging handbook strongly recommends using the BSR histologic grading system at present and have indicated that grade might be incorporated into the AJCC staging system in the near future.2 The BSR grading system has been shown to provide prognostic information and has less intra- and interpathologist variability.3 Only a small number of cases included grading by the BSR scale. The BSR grading system was not routinely reported at RPCI in 1998 to 1999 but was fully implemented in 2001. There is only a limited amount of data to quantify the degree of variation in pathology reporting in the United States.4 In New York State, a Quality Improvement Study released in 2001 documented low performance levels for reporting of mastectomy specimens. The following categories fell within compliance: margin status (69%), tumor size (63%), and histologic grade (59%).5 In a recent European Organization for Research and Treatment of Cancer (EORTC) trial involving DCIS, pathology reports described the lesion diameter in 25% of cases and margins were reported as “free,” without further specifications, in 49% of cases.6 There is literature demonstrating similar findings in the United Kingdom, Australia, and Europe.7,8,9 In these studies, the most consistently reported element for breast cancer was histologic type. Other elements, including tumor size, grade, and margins, were reported with less consistency. The use of standardized pathology reporting is not a new concept. In 1973, the Breast Cancer Task Force of the Pathology Working Group recommended standardized reporting of breast specimens.10 In the early 1990s, the Association of Directors of Anatomic and Surgical Pathology recommended that pathologists use a “checklist” approach for recording information affecting treatment and prognosis.11 The CAP Cancer Committee organized disease-site task forces to develop guidelines for the reporting of cancer specimens. This effort culmi-

Vol. 196, No. 1, January 2003

nated in publication of comprehensive guidelines for over 40 disease sites, with many specimen-specific guidelines in each disease site. For example, for breast cancer there are separate guidelines for reporting breast biopsy (incisional and core needle), complete excision (less than total mastectomy), and mastectomy.1 Each guideline identifies the necessary data elements that should be included in a pathology report to provide clinicians with complete and unambiguous information. Although the guidelines were widely distributed in 1998, there was no requirement that they be used and no data on their impact on practice. Recently, the American College of Surgeons Commission on Cancer decided to require that approved cancer programs use CAP guidelines for pathology reporting. There are about 1,400 hospitals with American College of Surgeons-approved programs in the United States that collectively treat about 70% of cancer patients in the country. Use of CAP guidelines will be required for all cancer cases at these hospitals, effective January 1, 2004. A barrier to full implementation of the CAP guidelines is the concern that the guidelines are too detailed and require elements that are not generally available to the pathologist, or that might not have direct bearing on prognosis of treatment. Examples include demographic and clinical information, such as family history of cancer and physical examination findings. These concerns have led CAP to reconsider some of these elements and to revise the guidelines to identify “required” and “optional” elements. In addition, the guideline revision will ease requirement for some data elements, such as the requirement of reporting tumor size in three dimensions. These revised guidelines will be available by fall 2002 (C Compton, personal communication). Although these results clearly illustrate the problem and support the concept of standardized pathology reporting, it is possible that the findings from this study are not fully representative of pathology reporting in Western New York or the United States. We did not attempt statistical comparisons between CH and RPCI because this convenience cohort is not an unbiased representative sample and because the number of cases from each hospital was small. Also, this study only examines reporting for breast cancer. Another limitation of this study is that the nature of the data process required subjective interpretation of the text of the report. The CAP guidelines provide suggested content for each item, but do not require use of a specific template. It is recog-

Wilkinson et al

Breast Cancer Reporting Guidelines

43

nized that the terminology to describe the same pathologic characteristics varies and can be ambiguous.1 CAP is currently working to clarify the intended meaning of ambiguous terminology and to improve the interpretation of reports (C Compton, personal communication). This study shows that most breast specimen reports in 1998 to 1999 did not meet CAP standards. These findings suggest that passive distribution of guidelines can be insufficient to improve practice, and supports the implementation of mandatory use of CAP guidelines for reporting of cancer specimens. Author Contributions

Study conception and design: Edge, Shahryarinejad Acquisition of data: Shahryarinejad, Watroba Analysis and interpretation of data: Shahryarinejad, Watroba Drafting of manuscript: Wilkinson, Edge Critical revisions: Edge, Winston, Wilkinson REFERENCES 1. Compton CC, Henson DE, Hammond EH, Schramm JB. Reporting on cancer specimens: protocols and case summaries. Northfield, IL: College of American Pathologists; 1998. 2. American Joint Commission on Cancer. AJCC Staging handbook. 6th ed. New York: Springer-Verlag; 2002:255–281. 3. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991;19:403–410. 4. Rosai J. Standardized reporting of surgical pathology diagnosis for the major tumor types: a proposal. Am J Clin Pathol 1993; 100:240–255. 5. Imperato PJ, Waisman J, Wallen M, et al. Breast cancer pathology practices among Medicare patients undergoing unilateral extended simple mastectomy in New York State. Lake Success, NY: Island Peer Review Organization; 2001. 6. Bijker N, Peterse JL, Duchateau L, et al. Risk factors for recurrence and metastasis after breast-conserving therapy for ductal carcinoma-in-situ: analysis of European Organization for Research and Treatment of Cancer Trial 10853. J Clin Oncol 2001; 19:2263–2271. 7. Ma M, Bell J, Campbell S, et al. Breast cancer management: is volume related to quality? Clinical Advisory Panel. Br J Cancer 1997;75:1652–1659. 8. Kricker A, Armstrong B, Smith C, et al. An audit of breast cancer pathology reporting in Australia in 1995. Br J Cancer 1999;80: 563–568. 9. Faverly D, Renard F, Drijkoningen M, Scheiden R. Breast cancer screening pathology: an assessment of the practise and needs in Belgium and Luxembourg. Virchows Arch 2000;437:354–359. 10. Standardized management of breast specimens recommended by Pathology Working Group, Breast Cancer Task Force. Am J Clin Pathol 1973;60:789–798. 11. Rosai J, Bonfiglio TA, Corson JM, et al. Standardization of the surgical pathology report. Mod Pathol 1992;5:197–199.