J Clin Oncol 26: by American Society of Clinical Oncology INTRODUCTION

VOLUME 26 NUMBER 19 JULY 1 2008 JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Prognostic Significance of Nottingham Histologic Grade in Invasive Breast Carcinoma Emad A. Rakha, Maysa E. El-Sayed, Andrew H.S. Lee, Christopher W. Elston, Matthew J. Grainge, Zsolt Hodi, Roger W. Blamey, and Ian O. Ellis From the Department of Histopathology, School of Molecular Medical Sciences, Division of Epidemiology and Public Health, Department of Surgery, Nottingham University Hospitals NHS Trust, University of Nottingham, United Kingdom. Submitted December 3, 2007; accepted March 14, 2008; published online ahead of print at www.jco.org on May 19, 2008. Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Corresponding authors: Emad A. Rakha and Ian O. Ellis, MD, PhD, Molecular Medical Sciences, Department of Histopathology, University of Nottingham, Nottingham City Hospital NHS Trust, Hucknall Rd, Nottingham, NG5 1PB United Kingdom; e-mail: emadrakha@yahoo.com or ian.ellis@nottingham.ac.uk. 2008 by American Society of Clinical Oncology 0732-183X/08/2619-3153/$20.00 DOI: 10.1200/JCO.2007.15.5986 A B S T R A C T Purpose The three strongest prognostic determinants in operable breast cancer used in routine clinical practice are lymph node (LN) stage, primary tumor size, and histologic grade. However, grade is not included in the recent revision of the TNM staging system of breast cancer as its value is questioned in certain settings. Materials and Methods This study is based on a large and well-characterized consecutive series of operable breast cancer (2,219 cases), treated according to standard protocols in a single institution, with a long-term follow-up (median, 111 months) to assess the prognostic value of routine assessment of histologic grade using Nottingham histologic grading system. Results Histologic grade is strongly associated with both breast cancer specific survival (BCSS) and disease-free survival (DFS) in the whole series as well as in different subgroups based on tumor size (pt1a, pt1b, pt1c, and pt2) and LN stages (pn0 and pn1 and pn2). Differences in survival were also noted between different individual grades (1, 2, and 3). Multivariate analyses showed that histologic grade is an independent predictor of both BCSS and DFS in operable breast cancer as a whole as well as in all studied subgroups. Conclusion Histologic grade, as assessed by the Nottingham grading system, provides a strong predictor of outcome in patients with invasive breast cancer and should be incorporated in breast cancer staging systems. J Clin Oncol 26:3153-3158. 2008 by American Society of Clinical Oncology INTRODUCTION Current routine clinical management of breast cancer relies on availability of robust clinical and pathologic prognostic and predictive factors to support clinical and patient decision making where potentially suitable treatment options are increasingly available. The three strongest prognostic determinants in operable breast cancer used in routine clinical practice internationally are lymph node (LN) stage, primary tumor size, and tumor histologic grade. The most widely used histologic grading system of breast cancer is the Nottingham combined histologic grade (Elston-Ellis modification of Scarff- Bloom-Richardson grading system), also known as the Nottingham grading system. 1,2 Multiple studies have shown an independent prognostic significance of grade in breast cancer 3-5 and improved interobserver agreement with the Nottingham histologic grade method compared with other grading systems. 6-8 In multivariate analysis, histologic grade has equivalent prognostic value to that of LN stage, 9 and it can be combined with LN stage and tumor size in the form of a prognostic index. One example, the Nottingham prognostic index (NPI), 10 has become widely used for the treatment of patients with breast cancer in the United Kingdom, 11 and its validity has been confirmed in other independent studies. 12-14 The NPI has been recognized as the only appropriately validated prognostic index in breast cancer. 15 In NPI, grade and LN stage has equal weighting. However in another prognostic index (Kalmar prognostic index 5 ) histologic grade is given a high weighting value (1.57), which is higher than that for LN stage (0.79) or size (0.31). In addition, the clinical contribution of grade has become more important as a consequence of earlier detection of breast cancer through mammographic screening and increase self awareness, which have resulted in a shift in stage distribution with lower median size of 2 2008 by American Society of Clinical Oncology 3153

Rakha et al cm or smaller 16,17 and greater proportion of LN negative tumors at presentation. 18 Histologic grading is now part of the minimum data set for breast cancer pathology reporting produced by the United Kingdom Royal College of Pathologists 19 and European Commission, 20 and is endorsed by the WHO 21 and the College of American Pathologists. 22 However, despite this evidence, the latest Breast Task Force of the American Joint Committee on Cancer did not include histologic tumor grade in its staging criteria. 23 One of the main reasons for omitting grade in the staging criteria includes the interaction between tumor size and histologic grade and lack of clear evidence for the role of grade in small tumors (TNM stage pt1 and pt2) as a result of difficulties in comparisons among previous studies because of the variety of follow-up times, grading systems, patient samples, and measured outcomes. 23 They also reported that in studies that used Nottingham combined histologic grade, grade 2 either clustered with grade 3 or else was intermediate between grade 1 and grade 3. 24,25 Studies that looked only at the smallest tumors (pt1a, pt1b) 25,26 tended to show somewhat smaller outcome differences between grade 1 and grade 3 than studies that included larger tumors. Therefore, in this study, we performed a retrospective analysis of a large and well-characterized series of operable breast cancers (pt1 and pt2 tumors) with long-term follow-up comprising clinicopathologic and outcome information. Cases entered into this single institution tumor series have been routinely graded using the Nottingham grading system and treated in a uniform and conventional manner. Our aim was to assess the prognostic value of routine application of the Nottingham grading system in patients presenting with operable breast as a whole (pt1 and pt2) as well as in the different size and LN subgroups to address issues of its potential value for inclusion (or not) in breast cancer staging system. METHODS The study population was derived from the Nottingham Tenovus Primary Breast Carcinoma Series of women age 70 years or younger, who presented with primary operable invasive breast carcinomas (with tumors of 5cm diameter on clinical/preoperative measurement and/or on operative histology) between 1990 and 1999. Women older than 70 years were not included because of the increased confounding factor of death from other causes and because primary treatment protocols for these patients often differed from those for younger women. Similarly, the majority of women with tumors of larger than 5 cm diameter (locally advanced primary tumors) were treated with different protocols. Patients diagnosed before 1990 and after 1999 have not been included because major changes in diagnosis and treatment were applied to patients before and after these time points. Patient s clinical history and tumor characteristics including method of referral (screening v symptomatic), patients age, menopausal status, family history, bilaterality, primary tumor size, histologic tumor type, histologic grade, 2,4 degree of tubule formation, nuclear pleomorphism and mitosis, LN stage, vascular invasion (VI), and NPI 10 and estrogen receptor (ER) status were obtained from the database. Tumor grading was assessed, using the Nottingham grading system, 2 in a consistent way using uniform methods of specimen fixation, sampling, and processing. NPI was calculated using the following equation: NPI 0.2 tumor size (cm) grade (1 to 3) lymph node score (1 to 3). 10 This is a well-characterized series of patients treated uniformly in a single institution. Adjuvant systemic therapies were offered according to the NPI group. No systemic therapy was offered to patients in the good (NPI 3.4) prognostic groups. 10 Patients in the moderate I group (3.41 to 4.4) with ER positive tumors were offered hormone therapy. Patients in the moderate II (4.41 to 5.4) and poor ( 5.41) groups received hormone therapy for ERpositive tumors and cytotoxic therapy for ER-negative tumors. Survival data including survival time, disease-free survival (DFS), and development of distant metastasis, locoregional recurrence was maintained on a prospective basis. Patients were followed up at 3-month intervals initially, then 6-month intervals, and then annually for a median period of 111 months (range, 4 to 211). At death, the hospital notes are examined and deaths allocated to with/ from breast cancer or to without known breast cancer. Breast cancer specific survival (BCSS) was defined as the interval between the operation and death from breast cancer, death being scored as an event, and patients who died from other causes or were still alive were censored at the time of last follow-up. DFS was also calculated from the date of first operation, with first recurrence, local, regional, or distant, being scored as an event, and with censoring of other patients at the time of last follow-up or death. Local recurrence was defined as tumor arising in the treated breast or chest wall. Regional recurrence was defined as tumor arising in the axillary or internal mammary LNs. Statistical Analysis Statistical analysis was performed using SPSS 15.0 statistical software (SPSS Inc, Chicago, IL). BCSS and DFS curves were drawn using Kaplan- Meier estimates, and were compared using log-rank tests. Survival rates are presented with their 95% CIs. Multivariate analyses of DFS and BCSS, with stepwise variable selection, were conducted using Cox proportional hazards regression models. The clinicopathologic variables were compared using contingency tables and 2 tests. A P value lower than.05 was considered significant. RESULTS Of the whole series (2,608 cases), 2,219 cases had complete data on grade, LN stage, size, VI, and outcome data. These 2,219 cases formed the basis of this study. Of these cases, 412 cases (18.6%) were grade 1, 790 were grade 2 (35.6%), and 1,017 cases (45.6%) were grade 3. The median size was 1.8 cm (range, 0.2 to 5 cm). ER status was available in 2,158 cases; of those, 1,556 were ER positive (72%). Hormone therapy was given to 889 patients (41%) and chemotherapy to 414 patients (19%); 52 patients (2.5%) received both hormone therapy and chemotherapy. Recurrences developed in 745 patients (33.6%) including 220 with local recurrence (10%), 152 with regional recurrences (7%), and 541 with distant metastasis (24.4%). Of all patients, 458 died of breast cancer (20.6%), 300 died of other causes (13.6%), while 1,461 were either still alive at the end of follow-up or lost to followup (65.8%). Association Between Histologic Grade and Other Clinicopathologic Variables Table A1 (online only) summarizes the associations between histologic grade and other clinicopathologic variables. There was a positive association between low histologic grades and screening method of referral, elderly postmenopausal patients, smaller primary tumor size, absence of LN metastasis, or definite VI and ER positivity. No association was found between grade and family history or tumor bilaterality. Association With Patients Outcome In the whole series, survival analyses showed an association between higher histologic grade and poorer patients outcome for both BCSS and DFS (Table 1; Figs 1A and 1B). In addition, when individual grade categories were compared, differences in patients outcome was found with the most significant difference between grade 1 versus grade 3 (log-rank 104; P.001; and log-rank 44; P.001 for 3154 2008 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Prognostic Significance of Histologic Grade in Invasive Breast Cancer Table 1. Association Between Histologic Grade and Patients Outcome in the Whole Breast Cancer Series and in the Different Pathologic Subsets Breast Cancer Related Deaths Disease-Free Survival Patient Group No. Log-Rank P Log-Rank P Whole series 2,219 172.001 62 <.001 Size subsets, tumor (cm) 1 (T1a, T1b) 398 22.001 3.24 1-2 cm (T1c) 1,064 60.001 18 <.001 2-5 cm (T2) 757 46.001 21 <.001 Lymph node subsets Negative (N0) 1,432 77.001 12.002 Positive (N1, N2) 787 62 <.001 39 <.001 NOTE. Bold text indicates statistical significance. BCSS and DFS, respectively), grade 2 versus 3 (log-rank 90; P.001; and log-rank 30; P.001 for BCSS and DFS, respectively), and then between 2 versus 1 (log-rank 18; P.001; and log-rank 9.3; P.002 for BCSS and DFS, respectively). Multivariate analyses, after adjustment for other prognostic indicators, including LN stage, tumor size, and VI were performed to determine whether histologic grade was an independent prognostic factor for BCSS and DFS. This showed that grade was an independent predictor of survival (Table 2). The same results were obtained after addition of ER status and adjuvant systemic treatment to the previous model. In addition, when we repeated the multivariate analysis but with addition of the interaction between histologic grade (as a three-level category) and LN stage and size (both fitted as linear terms), P interaction was not significant (P.73; HR 0.96 [95% CI, 0.76 to 1.2]; and P.33; HR 1.1 [95% CI 0.9 to 1.3] for LN stage and size, respectively), indicating that the effect of grade on patients outcome is not significantly affected by LN stage or tumor size. Grade in Relation to Tumor Size and LN Status Patients were stratified according to primary tumor size into three groups (according to TNM classification): group A (pt1a and pt1b), 1 cm; group B (pt1c), larger than 1 to 2 cm; and group C (pt2), larger than 2 cm to 5 cm. Associations between higher grade and poorer outcome were noted in all three groups (Table 1 and Appendix Table A4 [online only]; Appendix Fig A1A and Table A5, Fig A1B and Table A6, and Fig A1C and Table A7, online only). Although Table 1 shows lack of significant association between grade and development of recurrence in group A patients, when we analyzed the association between grade and different events of recurrence in this group, we found an association with development of distant metastasis (log-rank 21; P.001), but not with locoregional recurrence (log-rank 1.9; P.38). To further assess the survival difference between individual grades in the different size groups, we repeated the analyses between grade 1 and 2 tumors and between grade 2 and 3 tumors and statistical differences in patients outcome between individual grades were found. Similar to the whole series, multivariate analyses, after adjustment for other prognostic variables showed that grade is an independent prognostic factor in the different size groups. Grade and LN Stage Survival analyses showed that high histologic grade is associated with both shorter BCSS and DFS in the LN-negative (pn0) as well as in the LN-positive (pn1 and pn2) patients (Table 1; Appendix Fig A2A and Table A8, and Fig A2B and Table A9, online only). Interestingly these associations were independent of the size. Multivariate analyses showed that grade is an independent prognostic factor in both LNnegative and LN-positive groups. Because LN stage is the most recognized prognostic factor in operable breast cancer, we compared the prognostic significance of grade as compared with LN stage. We analyzed the prognostic significance of LN stage (LN negative [stage 1/pN0] v LN positive [stages 2 and 3/pN1 and pn2]) in the different grades. We found that LN stage A 1.0 G1 B 1.0 Probability of Survival 0.8 0.6 0.4 0.2 G2 G3 Probability of Survival 0.8 0.6 0.4 0.2 G3 G1 G2 Fig 1. Relation between histologic grade and (A) breast cancer specific survival (for the number of patients at risk at each time period, refer to Appendix Table A2, online only) and (B) disease-free survival in months (Appendix Table A3, online only). 0 60 120 180 240 0 60 120 180 240 Breast Cancer Specific Survival (months) Disease-Free Survival (months) www.jco.org 2008 by American Society of Clinical Oncology 3155

Rakha et al Table 2. Multivariate Cox Regression Analysis of Factors Associated With Cancer Specific Survival and Disease-Free Survival the Whole Series Cancer-Specific Survival Disease-Free Survival Predictor Hazard Ratio 95% CI P Hazard Ratio 95% CI P Lymph node status 1.8 1.5 to 2.0.001 1.5 1.4 to 1.7.001 Vascular invasion, yes v no 1.6 1.3 to 1.9.001 1.4 1.2 to 1.6.001 Size, 2 v 2 cm 1.4 1.1 to 1.6.003 1.1 0.9 to 1.3.115 Grade.001.001 2 v 1 1.6 1.1 to 2.5.022 1.1 0.9 to 1.4.29 3 v 1 3.9 2.6 to 5.8.001 1.6 1.3 to 2.0.001 NOTE. Bold text indicates statistical significance. Hazard ratio represents change in risk per change one unit change in status category. Lymph node status is coded as 1, 2 or 3. was associated with patients outcome in grade 3 (log-rank 50.3; P.001) and grade 2 (log-rank 22.3; P.001) tumors but lost its prognostic significance in grade 1 tumors (log-rank 1.4; P.24). We assessed the prognostic significance of individual patients based on a combination of LN stage and grade (Fig 2). This survival curve shows that outcome in patients with grade 1/LN-negative (pn0) and -positive (pn1) tumors is similarly very good and increasingly adverse outcome was observed for grade 2/pN0, grade 2/pN1, grade 3/pN0, and grade 3 pn1, respectively. Interestingly, we found that grade 3/pN0 tumors were associated with a significantly worse outcome when compared to grade 2/LN-positive (pn1) tumors (log-rank 6; P.02) and grade 1/LN positive (pn1 and pn2) tumors (log-rank 7.1; P.008) and these associations were independent of tumor size. A difference which failed to achieve significance was observed between grade 2/LN negative (worse outcome) and grade 1/LN positive (pn1) tumors (log-rank 2.6; P.11). Probability of Survival 1.0 0.8 0.6 0.4 0.2 G1/LN2 G2/LN2 G3/LN1 G3/LN2 G1/LN1 G2/LN1 0 60 120 180 240 Breast Cancer Specific Survival (months) Fig 2. Association between combined histologic grade and lymph node (LN) stage (stage 1 negative and stage 2 one to three positive nodes) and breast cancer specific survival (BCSS; in months). This graph shows grade 1/LN 1 and grade 1/LN 2 at top, grade 2/LN 1 and grade 2/LN 2 in the middle, while grade 3/LN 1 and grade 3/LN 2 are in the bottom of the graph (log-rank 183; P.001); Appendix Table A10, online only. DISCUSSION Recent high profile molecular profiling studies of breast cancer have demonstrated the importance of the genetic makeup of tumors and showed associations between molecular portrait of breast cancer and tumor biologic and clinical behavior. 27 Furthermore, they have demonstrated that histologic grade, when compared to LN stage and tumor size, represents the predominant contributor to the molecular makeup of breast cancer, 28 and showed an association between genetic grade signature of breast cancer and patients outcome independent of LN status or tumor size. 29,30 Although histologic grade of breast cancer has been recognized for a long period of time, 2 and its prognostic value has been validated in multiple independent studies, 3,4,6-8 there are still some concerns regarding the incorporation of grade into routine breast cancer staging systems. 23 In this study, we assessed the prognostic significance of histologic grade in a large series of operable breast cancer graded in routine practice using the Nottingham histologic grading system, uniformly treated in a single institution and with available long-termterm follow-up data. Our results demonstrated that histologic grade is associated with other well established prognostic variables and with patients outcome. Grade was associated with shorter BCSS and DFS in the whole series as well as in the different subgroups including small size tumors (T1a, T1b, and T1c), LN-negative and -positive tumors. Differences in survival were also noted between individual grades in the whole series as well as in tumors of different size categories. Importantly, this prognostic value of histologic grade is independent of other prognostic variable, such as size or LN stage. Furthermore, to assess the value of grade as compared with LN stage, which is currently widely regarded as the most important prognostic variable in operable breast cancer, we combined grade and LN stage and assessed the prognostic significance of the different combinations. Our results demonstrate that grade is an important determinant of breast cancer outcome and complimentary to LN stage. For example, we found that the outcome of patients with grade 2/LNpositive tumors (N1 one to three positive nodes) is better than that of patients with grade 3/LN-negative (N0) tumors. Therefore, grade will influence the outcome of patients in different LN stage categories. As a consequence, we believe the treatment of patients should be based on 3156 2008 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

Prognostic Significance of Histologic Grade in Invasive Breast Cancer both grade and stage in addition to tumor size (components of NPI 10 ) and not restricted to information on size and LN stage (TNM system). Another reason for the reluctance to rely on histologic grading is the subjective nature of grade and a perceived poor reproducibility and consistency 31 with a subsequent difference in the proportion of tumors assigned to each grade. This study shows a high interobserver reproducibility of the Nottingham combined histologic grade because the proportion of the different grade in this series is almost identical to the grade assignment in the original series (625 patients) published in 1984 (grade 1, 17%; grade 2, 37%; and grade 3, 46%). 32 This observation is consistent with and supports the findings of the previous studies that reported acceptable levels of inter- and intraobserver variability of histologic grade. 2,4,33 We have consistently applied and endorsed rigorous optimized and standardized methods for tissue handling, fixation, and preparation. Differences between centers can in many cases be attributed to differences in quality of tissue preparation. Suboptimal tissue fixation has been clearly demonstrated to impact adversely on ability to assess mitotic frequency resulting in a systematic downgrading of a proportion of cases. This typically results in a reduction in the proportion of cases assigned to grade 3 with resultant increase in grade 2 cases. 19 Critical evaluation of these issues with recommendations for good practice have been provided by professional organizations. 19 Significant improvements in consistency of histologic grading have been observed on a national basis in the United Kingdom through publication of guidelines with linked educational activity and associated external quality assurance. 34 These guidelines not only provide information on histologic grading methodology but also recommendations of application of these methods and guidance on tissue handling. It is well recognized that assessment of critical biomarkers in breast cancer, such as ER, progesterone receptor, and HER2, require use of standardized validated reproducible laboratory methods linked to use of internal and external control systems and monitored through audit and participation in external quality control. We would endorse similar diligence and rigor in use of histologic grading. From our experience and results identified in this study, we believe that histologic grade, when assessed in such a rigorous fashion is a valuable prognostic factor in breast cancer, which can be assessed reproducibly and cost efficiently in routine clinical practice, to give consistent results complimentary to assessment of LN stage and tumor size. In conclusion, our results are consistent with previous studies which showed an independent prognostic significance for histologic grade, 2,4,33 including small size tumors, 25,35 providing evidence for the importance of routine assessment of histologic grade in breast cancer in addition to LN stage and size of the primary tumor. We recommend routine assessment of histologic grade of breast cancer using the Nottingham consensus criteria 19-21 linked with good practice in tumor tissue handling and preparation to overcome the problems of consistency and reproducibility of histologic grading and to provide a consistent prognostic significance and the incorporation of grade in breast cancer staging system. AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. AUTHOR CONTRIBUTIONS Conception and design: Emad A. Rakha, Maysa E. El-Sayed Administrative support: Christopher W. Elston, Ian O. Ellis Provision of study materials or patients: Christopher W. Elston, Roger Blamey, Ian O. Ellis Collection and assembly of data: Emad A. Rakha, Maysa E. El-Sayed, Christopher W. Elston, Roger Blamey, Ian O. Ellis Data analysis and interpretation: Emad A. Rakha, Maysa E. El-Sayed, Matthew Grainge, Ian O. Ellis Manuscript writing: Emad A. Rakha, Maysa E. El-Sayed, Andrew H. S Lee, Roger Blamey, Ian O. Ellis Final approval of manuscript: Emad A. Rakha, Maysa E. El-Sayed, Andrew H. S Lee, Christopher W. Elston, Matthew Grainge, zsolt hodi, Roger Blamey, Ian O. Ellis REFERENCES 1. Bloom HJ, Richardson WW: Histological grading and prognosis in breast cancer: A study of 1409 cases of which 359 have been followed for 15 years. Br J Cancer 11:359-377, 1957 2. Elston CW, Ellis IO: Pathological prognostic factors in breast cancer I: The value of histological grade in breast cancer: Experience from a large study with long-term follow-up. Histopathology 19: 403-410, 1991 3. Symmers WC: Assessment of Histological Grade. Edinburgh, United Kingdom, Churchill Livingstone, 1998 4. Pereira H, Pinder SE, Sibbering DM, et al: Pathological prognostic factors in breast cancer. IV: Should you be a typer or a grader? A comparative study of two histological prognostic features in operable breast carcinoma. Histopathology 27:219-226, 1995 5. Sundquist M, Thorstenson S, Brudin L, et al: Applying the Nottingham Prognostic Index to a Swedish breast cancer population: South East Swedish Breast Cancer Study Group. Breast Cancer Res Treat 53:1-8, 1999 6. Dalton LW, Page DL, Dupont WD: Histologic grading of breast carcinoma: A reproducibility study. Cancer 73:2765-2770, 1994 7. Frierson HF Jr, Wolber RA, Berean KW, et al: Interobserver reproducibility of the Nottingham modification of the Bloom and Richardson histologic grading scheme for infiltrating ductal carcinoma. Am J Clin Pathol 103:195-198, 1995 8. Robbins P, Pinder S, de Klerk N, et al: Histological grading of breast carcinomas: A study of interobserver agreement. Hum Pathol 26:873-879, 1995 9. Walker RA: Prognostic and Predictive Factors in Breast Cancer (ed 1). New York, NY, Informa Health Care, 2003 10. Galea MH, Blamey RW, Elston CE, et al: The Nottingham Prognostic Index in primary breast cancer. Breast Cancer Res Treat 22:207-219, 1992 11. Elston CW, Ellis IO, Pinder SE: Pathological prognostic factors in breast cancer. Crit Rev Oncol Hematol 31:209-223, 1999 12. Balslev I, Axelsson CK, Zedeler K, et al: The Nottingham Prognostic Index applied to 9,149 patients from the studies of the Danish Breast Cancer Cooperative Group (DBCG). Breast Cancer Res Treat 32:281-290, 1994 13. D Eredita G, Giardina C, Martellotta M, et al: Prognostic factors in breast cancer: The predictive value of the Nottingham Prognostic Index in patients with a long-term follow-up that were treated in a single institution. Eur J Cancer 37:591-596, 2001 14. Okugawa H, Yamamoto D, Uemura Y, et al: Prognostic factors in breast cancer: The value of the Nottingham Prognostic Index for patients treated in a single institution. Surg Today 35:907-911, 2005 15. Clark GM: Do we really need prognostic factors for breast cancer? Breast Cancer Res Treat 30:117-126, 1994 16. Anttinen J, Kautiainen H, Kuopio T: Role of mammography screening as a predictor of survival in postmenopausal breast cancer patients. Br J Cancer 94:147-151, 2006 17. Shen Y, Yang Y, Inoue LY, et al: Role of detection method in predicting breast cancer survival: Analysis of randomized screening trials. J Natl Cancer Inst 97:1195-1203, 2005 18. Hanrahan EO, Valero V, Gonzalez-Angulo AM, et al: Prognosis and management of patients with node-negative invasive breast carcinoma that is 1 cm or smaller in size (stage 1; T1a,bN0M0): A review of the literature. J Clin Oncol 24:2113-2122, 2006 www.jco.org 2008 by American Society of Clinical Oncology 3157

Rakha et al 19. NHS Breast Screening Programme: Guidelines for Pathology Reporting in Breast Cancer Screening and the Second Edition of The Royal College of Pathologists Minimum Dataset for Breast Cancer Histopathology. Sheffield, United Kingdom, NHS, 2005 20. European Commission: European Guidelines for Quality Assurance in Mammography Screening. Luxembourg: Office for Official Publications of the European Communities 1996 21. WHO: Classification of Tumours: Pathology and Genetics. Lyon, France, IARC Press, 2003 22. Fitzgibbons PL, Page DL, Weaver D, et al: Prognostic factors in breast cancer: College of American Pathologists Consensus Statement 1999. Arch Pathol Lab Med 124:966-978, 2000 23. Singletary SE, Allred C, Ashley P, et al: Revision of the American Joint Committee on Cancer staging system for breast cancer. J Clin Oncol 20:3628-3636, 2002 24. Genestie C, Zafrani B, Asselain B, et al: Comparison of the prognostic value of Scarff-Bloom- Richardson and Nottingham histological grades in a series of 825 cases of breast cancer: Major importance of the mitotic count as a component of both grading systems. Anticancer Res 18:571-576, 1998 25. Kollias J, Murphy CA, Elston CW, et al: The prognosis of small primary breast cancers. Eur J Cancer 35:908-912, 1999 26. Rosner D, Lane WW: Should all patients with node-negative breast cancer receive adjuvant therapy? Identifying additional subsets of low-risk patients who are highly curable by surgery alone. Cancer 68:1482-1494, 1991 27. Perou CM, Sorlie T, Eisen MB, et al: Molecular portraits of human breast tumours. Nature 406:747-752, 2000 28. Yu K, Lee CH, Tan PH, et al: A molecular signature of the Nottingham prognostic index in breast cancer. Cancer Res 64:2962-2968, 2004 29. Ivshina AV, George J, Senko O, et al: Genetic reclassification of histologic grade delineates new clinical subtypes of breast cancer. Cancer Res 66: 10292-10301, 2006 30. Sotiriou C, Wirapati P, Loi S, et al: Gene expression profiling in breast cancer: Understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst 98:262-272, 2006 31. Gilchrist KW, Kalish L, Gould VE, et al: Interobserver reproducibility of histopathological features in stage II breast cancer: An ECOG study. Breast Cancer Res Treat 5:3-10, 1985 32. Elston CW: The assessment of histological differentiation in breast cancer. Aust N Z J Surg 54:11-15, 1984 33. Theissig F, Kunze KD, Haroske G, et al: Histological grading of breast cancer: Interobserver, reproducibility and prognostic significance. Pathol Res Pract 186:732-736, 1990 34. Ellis IO, Coleman D, Wells C, et al: Impact of a national external quality assessment scheme for breast pathology in the UK. J Clin Pathol 59:138-145, 2006 35. Sundquist M, Mitchell M, Blamey R, et al: The prognosis of small breast cancers and selection for omission of adjuvant chemotherapy. Eur J Cancer 5:14-15, 2007 Acknowledgment This study was approved by Nottingham Research Ethics Committee 2 under the title: development of a molecular genetic classification of breast cancer. Appendix The Appendix is included in the full-text version of this article, available online at www.jco.org. It is not included in the PDF version (via Adobe Reader ). 3158 2008 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY