© 2007 Foundation for Promotion of Cancer Research
Which is the Better Pathological Prognostic Factor, the Nottingham Histological Grade or the Japanese Nuclear Grade? A Large Scale Study with a Long-term Follow-up
1 Departments of Pathology
2 Breast Surgery, Seirei Hamamatsu General Hospital, Hamamatsu, Shizuoka
3 Kanzaki Clinic, Hamamatsu, Shizuoka, Japan
For reprints and all correspondence: Hiroshi Kobayashi, Department of Pathology, Seirei Hamamatsu General Hospital, 2-12-12 Sumiyoshi, Hamamatsu, 430-8558, Japan. E-mail: kobayashi{at}sis.seirei.or.jp
Received October 3, 2006; accepted November 13, 2006
 |
Abstract |
|---|
 TOP Abstract INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
Background: We compared the Nottingham histological grade (H-grade) and the Japanese nuclear grade (N-grade) to select the better prognostic factor for breast cancers.
Methods: The series included 1786 patients with breast cancers with the exception of non-invasive and stage 4 cancers. They were classified according to the H- and N-grade. We analyzed their survival curves and also performed multivariate Cox regression analyses.
Results: According to the H-grade classification, 476 cases were grade 1, 647 cases were grade 2 and 663 cases were grade 3. According to the N-grade, 381 cases were grade 1, 215 cases were grade 2, and 1129 cases were grade 3. In the survival curves of those cases with lymph node metastases (N(+)) and recurrent cases, there were statistically significant differences in different categories of the H-grades, but not in the N-grades. The survival curves of all the cases and those cases without lymph node metastases (N(–)) always exhibited statistically significant differences. According to the 2003 St Gallen consensus, the N(–) group was classified as a minimal risk and an average risk groups. Both H- and N-grade exhibited statistically significant differences between the minimal risk and the average risk groups in the disease-free survival. The multivariate analyses proved that the H-grade was a statistically significant prognostic factor in all the cases and N(+) group, but the N-grade was not significant in any of the studies.
Conclusions: The H-grade is clearly proved to be a more significant prognostic factor for wider stage cases than the N-grade.
Key Words: breast cancer • histological grade • pathological grade • prognostic factor
|
INTRODUCTION |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
Breast cancer has become one of the most important areas of research among cancers in Japan, because of its increasing mortality rate that was estimated as 14.9 per 100,000 women in 2002 (1). The pathological tumor grade is well recognized as one of the most significant prognostic factors for breast cancers. The tumor grade, along with tumor size, patient age, estrogen receptor (ER)/progesterone receptor (PgR) expression, and Her2/neu expression, is used to select a patient for adjuvant systemic therapy after surgical tumor resection. The St Gallen conference for the international expert consensus of the primary therapy for early breast cancer has adopted the tumor grade as one of several prognostic features to define the risk category of breast cancer (2, 3). In Japan, the Japanese Breast Cancer Society recommends two tumor grades (4), one of which is the Nottingham histological grade (H-grade) (5) while the other is the nuclear grade (N-grade), which was recommended in a multi-institutional protocol study, the National Surgical Adjuvant Study of Breast Cancer (NSAS-BC), to select high-risk, node-negative breast cancers in Japan (6). In the current large scale study with a long term follow up, we compared the H-grade and the N-grade to decide which is the more statistically significant prognostic factor for breast cancers. Furthermore, we also compared the two grading methods by categorizing the risk groups for recurrence following the 2003 St Gallen recommendation (2).
|
PATIENTS AND METHODS |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
Case Series
The series included 1786 patients with all types of breast cancers with the exception of non-invasive and stage 4 cancers. All patients were Japanese women (ages ranging from 19 to 96 years; mean 52.1 years) who had attended Seirei Hamamatsu General Hospital from April 1976 to December 2004 and who had been treated with surgical excision, comprising 323 partial mastectomies and 1463 total mastectomies. Following the surgical intervention, 1,353 patients were treated with hormone therapy, 1,285 patients with chemotherapy and 319 patients with irradiation therapy. The tumor recurrence and patients' death were checked in our hospital during the following up term.
Pathology
In 1,274 patients who underwent resection from May 1981 to July 2000, small fragments from the resected specimens were prepared for enzyme immunoassay for ER and PgR, and the rest of the specimens were fixed in 10% phosphate buffered formalin (pH 7.4). In the other 512 patients, the entire resected specimens were fixed in formalin. They were routinely processed in paraffin blocks and the 4 µm sections were prepared. In all cases specimens were stained with hematoxylin and eosin (HE) and immunohistochemical staining for ER and PgR were performed using the labeled streptavidin-biotin system (VENTANA NX system: Ventana Medical Systems, Arizona, USA). Both anti-ER and anti-PgR antibodies (Ventana Medical Systems, Arizona, USA) were diluted to 1/2. The antigen was retrieved by boiling in a citrated buffer (pH 7.0) for 15 min.
Tumor Grading
All the breast cancers were graded by three pathologists (Y.O., S.S. and H.K.) in the single institute. Olympus BX41 microscopes with WH10X for ocular lenses and UPlanApo 4 x, 10 x, 20 x and 40 x for objective lenses were used to assign the grading (Olympus, Tokyo, Japan). At the beginning of this study, these three pathologists participated in a single demonstration and discussed how to correctly examine and assign the grading for enough cases to reach consensus. After that, evaluation was performed independently without bias by each pathologist. We classified all the invasive breast cancers according to the H- and N-grades (5, 6). The Japanese Breast Cancer Society originally analyzed the N-grade, which was adopted only for invasive ductal carcinoma, but the H-grade was internationally adopted for any invasive cancers. In this study, therefore, we decided to adopt both the H- and N-grade for any invasive breast cancers (5, 6). The H-grade has three parameters, namely tubule formation, nuclear pleomorphism and the number of mitoses. Each parameter is graded into three categories to which a score of 1 to 3 is assigned as follows: tubule formation (1: > 75%, 2: 10–75%, 3: < 10%); nuclear pleomorphism (1: none, 2: moderate, 3: pronounced); and the number of mitoses/10 high power fields (HPF) (x 40 objective lens) (1: 0–9 mitoses; 2: 10–19 mitoses; and 3: > 19 mitoses). Finally the three H-grades are determined by the sum of the scores of the three parameters (3, 4, or 5 = H-grade 1; 6 or 7 = H-grade 2; and 8 or 9 = H-grade 3). The N-grade has two parameters, namely nuclear pleomorphism and the number of mitoses. Each parameter is graded into three categories to which a score 1 to 3 is assigned as follows: nuclear pleomorphism (1: none, 2: moderate, 3: pronounced), and the number of mitoses/10 HPF (x 40 objective lens) (1: 0–4 mitoses; 2: 5–10 mitoses; and 3: > 10 mitoses). Finally the three N-grades are determined by the sum of the scores of the two parameters (2 or 3 = N-grade 1; 4 = N-grade 2; and 5 or 6 = N-grade 3).
Statistical Analysis
The overall survival curves and disease-free survival curves were constructed using the Kaplan–Meier method (7). The differences among the groups were assessed with the log-rank test (8). To determine the statistically significant risk of relapse or death in breast cancer, multivariate analysis by Cox proportional hazards regression was used. The level of statistical significance was defined as P < 0.05. When the survival curves and the risk of relapse or death were analyzed, deaths from other diseases were excluded from the case series.
Steroid Hormone Receptor Status
An enzyme immunoassay technique was used to quantify the expression of ER and PgR in 1274 cases. In this assay, the value > 5 fmol/mg of ER concentrations in the diluted cytosols was regarded as positive. Immunohistochemical assays for ER and PgR were used to detect endocrine responsiveness in 512 cases. Immunohistochemically, positive expression of ER and PgR meant that more than 10% of the nuclei of all the tumor cells were positive for ER and PgR.
|
RESULTS |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
Distribution of Patients
Under the H-grade classification, 476 cases (26.7%) were grade 1 (H-grade 1), 647 cases (36.2%) were grade 2 (H-grade 2) and 663 cases (37.1%) were grade 3 (H-grade 3). Under the N-grade classification, 381 cases (22.1%) were grade 1 (N-grade 1), 215 cases (12.5%) were grade 2 (N-grade 2) and 1129 cases (65.4%) were grade 3 (N-grade 3). A group of the cases without axillary lymph node metastases was named N(–), and the other group of the cases with axillary lymph node metastases was named N(+). When the N(–) group was classified with the H-grade, H-grade 1 accounted for 351 cases (33.8%), H-grade 2 for 364 cases (35.0%), and H-grade 3 for 325 cases (31.3%). When the N(–) group was classified with the N-grade, N-grade 1 accounted for 266 cases (26.8%), N-grade 2 for 135 cases (13.6%), and N-grade 3 for 591 cases (59.6%). When the N(+) group was classified with the H-grade, H-grade 1 accounted for 113 cases (15.7%), H-grade 2 for 278 cases (38.5%) and H-grade 3 for 331 cases (45.8%). When the N(+) group was classified with the N-grade, N-grade 1 accounted for 109 cases (15.3%), N-grade 2 for 76 cases (10.7%) and N-grade 3 for 527 cases (74.0%). This series had some cases with recurrence of cancer. These recurrent cases were also classified under the H- and N-grades. In the recurrent cases, H-grade 1 accounted for 42 cases (13.6%), H-grade 2 for 124 cases (40.1%), and H-grade 3 for 143 cases (46.3%), whereas N-grade 1 accounted for 44 cases (14.2%), N-grade 2 for 37 cases (12.0%) and N-grade 3 for 228 cases (73.8%). These results are summarized in Tables 1 and 2.
|
|
Comparison of the Survival Curves under the H-grade and N-grade
The overall survival curves and disease-free survival curves of all the cases, the N(–) group, and the N(+) group were statistically analyzed to assess the efficacy of the H-grade and N-grade as prognostic factors (Figs 1–4). When all cases and the N(–) group were classified under the H-grade and N-grade, the cases with higher grade tumors exhibited statistically significantly worse overall survival and worse disease-free survival (Figs 1 and 2). In both the overall survival and disease-free survival of N(+) cases, however, there were no statistically significant differences in the different N-grade categories, although there were statistically significant differences in the different H-grade categories (Fig. 3). The overall survival curves of the recurrent cases were also analyzed to allow comparison between H-grade and N-grade subjects (Fig. 4). In this study, there were statistically significant differences in the different H-grade categories. In contrast, there were no significant differences in the different N-grade categories (Fig. 4).
|
|
|
|
Switch from H-grade to N-grade
In this study, all the H-grade 1 cases (396 cases) were classified as N-grade 1 (286 cases) or N-grade 2 (110 cases). All the H-grade 2 cases (601 cases) were classified as N-grade 1 (71 cases), N-grade 2 (91 cases), or N-grade 3 (439 cases). All the H-grade 3 cases were always classified as N-grade 3. Although we compared each separate group of N-grade subjects that was switched from H-grade 1 or H-grade 2, there were no significant differences in either overall survival or disease-free survival in this study (Figs 5 and 6).
|
|
St Gallen Risk Categories
According to the 2003 St Gallen consensus report as summarized in Table 3 (2), the N(–) group was classified as the minimal risk group and the average risk group. Either the H-grade or N-grade was used for tumor grade in this classification. We constructed and analyzed the survival curves of the minimal risk group and the average risk group, which were classified under either the H-grade or N-grade (Fig. 7). Both the H-grade and N-grade exhibited statistically significant differences between the minimal risk group and the average risk group in the disease-free survival (Fig. 7c,d). But statistically significant differences in overall survival were not exhibited in either group (Fig. 7a,b).
|
|
Multivariate Analyses
To examine the independent prognostic significance of the H-grade and the N-grade, multivariate Cox regression analyses were performed in all the cases, N(–) group, N(+) group, and recurrent cases. In these analyses, the H-grade emerged as an independent prognostic factor for overall survival and disease-free survival in all the cases and N(+) group, but the H-grade did not emerge as an independent prognostic factor for the other groups, and the N-grade did not for any of the groups. We also performed the multivariate analyses for the survival rates of the minimal risk group and the average risk group following the 2003 St Gallen consensus, which were classified under either the H-grade or N-grade. In these analyses, the average risk group did not exhibit a statistically significant hazard ratio against the minimal risk group whenever groups were classified under both the H- and N-grade. The results of these multivariate analyses are summarized in Table 4 .
|
|
DISCUSSION |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
Recently, post-operative adjuvant systemic therapies have been selectively applied to breast cancer patients for whom such therapeutic regimens have been considered as being possibly effective. Therefore, specific criteria are needed to facilitate the identification of high risk breast cancers which show a significantly higher recurrence rate and therefore require adjuvant systemic therapies. Breast cancers with a higher recurrence rate have been selected according to the tumor grade, tumor size, endocrine responsiveness, and so on. In the tumor grade, the H-grade is one of the most widely used histological grading systems, although other grading systems have also proved to be statistically significant (5, 6, 9–11). In Japan, the Japanese Breast Cancer Society recommends two tumor grading systems (the H-grade and N-grade) as described above (4). However, no report has to date compared H-grade and N-grade, as far as we know.
The guideline for the primary therapy of early breast cancer was updated at the 2005 St Gallen conference (3). A global consensus has been reached that appropriate adjuvant therapies should be applied according to the evaluation of hormone receptor status (responsive) and recurrence rate, according to the consensus report arrived at during this conference. This consensus report also recommended both the histological grade and nuclear grade for tumor grading to classify the breast cancer into different risk categories. However, even at the latest conference, the most useful tumor grading system has not yet been decided (2, 3). In this study, the H-grade and N-grade systems were used to determine risk categories for early breast cancers according to the 2003 St Gallen consensus report. The results showed that the H-grade was not so different from the N-grade with respect to survival curves. The analyses by log-rank test showed that both the H- and N-grade could be useful for tumor grading following the 2003 St Gallen consensus report, but multivariate analyses did not exhibit in our study. We should therefore continue to make further analyses according to the latest risk categories recommended by the latest 2005 St Gallen consensus report in the future. It is to be noted that there was a difference between the H- and the N-grade. The number of average risk patients classified under the N-grade was 16 patients more than that under the H-grade in terms of disease-free survival. These 16 patients, who need not receive chemotherapy in the H-grade classification, need it according to the N-grade classification.
In this study, we were able to reveal some noticeable differences between the H-grade and N-grade. The analyses of the survival curves prepared under the H-grade and the N-grade systems showed that the H-grade could more effectively categorize the breast cancer cases into three different prognostic groups (grades 1–3) than the N-grade. The reason was that the H-grade exhibited significant differences in each category in every analysis by log-rank test, but the N-grade failed to show such differences in the N(+) group and recurrent cases. The H-grade seemed to be available to predict prognoses both in the early stage cases such as the N(–) group and in the progressive cases such as the N(+) group and recurrent cases. However the N-grade was available only for the progressive cases. We should however notice some inconsistent results in the survival curves of the H-grade. When all the cases were classified under the H-grade, H-grade 3 reversely exhibited better prognosis than H-grade 2 in overall survival rate over 15 years and in disease-free survival rate over 10 years. These inconsistencies were not seen in N-grade 2 and 3. However, we think these are minor faults in the H-grade, because all the cases with higher grade tumors have exhibited statistically significantly worse overall survival and worse disease-free survival in the H-grade. The multivariate Cox regression analyses proved that the H-grade was a statistically significant prognostic factor in all the cases and N(+) group, but the N-grade was not significant in any of the studies. These multivariate analyses could therefore exhibit the H-grade as a significant prognostic factor for wider stage cases than the N-grade.
In the distribution of cases, all the groups [all the cases in the N(–) and N(+) groups] included around 60%, or more than 70%, of N-grade 3 in each group and some of the cases were classified into a higher grade under the N-grade system compared with the H-grade. This switch is mainly due to the fact that the number of mitoses corresponding to an N-grade score 1 was quite different from the H-grade (0–4 and 0–9, respectively) and the number of mitoses corresponding to an N-grade score 2 was also quite different from the H-grade (5–10 and 10–19, respectively). Since the N-grade exhibited an extremely high rate of grade 3 and produced a more unbalanced distribution of cases than the H-grade, H-grade could also be a better grading system than N-grade in the distribution of cases.
In our laboratory, we routinely use the H-grade as the grading system for breast cancers. Therefore, in this study we reclassified all cases from the H-grade under the N-grade. After the switch of the cases of H-grade 1 and 2 into the N-grade, no significant difference was seen in any of the grades under the N-grading system and we thus considered it was not necessary to reclassify the cases from H-grade into N-grade.
We have shown here the results of a large scale study with a long-term follow-up. Based on the results of this study, both the H- and N-grade could predict the prognoses of the breast cancers. However, we believe that the H-grading system should be more strongly recommended for the grading of breast cancer than the N-grade, because the former was proved to be significantly prognostic for wider stage cases than the latter and also it exhibited a much more balanced distribution of cases in its three grades than the latter.
|
Conflict of interest statement |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
None declared.
|
Acknowledgments |
|---|
We thank Dr Hitoshi Tsuda, Department of Pathology II, National Defense Medical College for his valuable advice and suggestions. We also thank all the staff of the Department of Clinical Laboratory, Seirei Hamamatsu General Hospital for expert technical assistance and Dr Toru Nakamura, Department of Surgery, Seirei Hamamatsu General Hospital, and Mr Masaya Suzuki, 1st Department of Pathology, Hamamatsu University School of Medicine for helpful suggestions for the statistical analyses.
|
References |
|---|
|
TOPAbstractINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONConflict of interest statementReferences |
|---|
1 Kokumin Eisei no Doukou. Health and Welfare Statistics Association. (2004) 389. (in Japanese).
2 Goldhirsch A, Wood WC, Gelber RD, Coates AS, Thurlimann B, Senn HJ. Meeting highlights: updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol (2003) 21:3357–65.
3 Goldhirsch A, Glick JH, Gelber RD, Coates AS, Thurlimann B, Senn HJ, et al. Meeting highlights: international expert consensus on the primary therapy of early breast cancer 2005. Ann Oncol (2005) 16:1569–83.
4 General Rules for Clinical and Pathological Recording of Breast Cancer. (2004) 15, 15th edn. The Japanese Breast Cancer Society. (in Japanese).
5 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 (1991) 19:403–10.[ISI][Medline]
6 Tsuda H, Akiyama F, Kurosumi M, Sakamoto G, Watanabe T. Establishment of histological criteria for high-risk node-negative breast carcinoma for a multi-institutional randomized clinical trial of adjuvant therapy. Japan National Surgical Adjuvant Study of Breast Cancer (NSAS-BC) Pathology Section. Jpn J Clin Oncol (1998) 28:486–91.
7 Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc (1958) 53:457–81.[CrossRef][ISI]
8 Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. analysis and examples. Br J Cancer (1977) 35:1–39.[ISI][Medline]
9 Jonjic N, Mustac E, Krstulja M, Zamolo G, Valkovic T, Sasso F, et al. Nuclear grade in invasive ductal breast carcinoma. Comparative study on nuclear grade and morphometric parameters, hormonal receptor status, flow cytometric analysis and lectin histochemistry. Gen Diagn Pathol (1996) 142:83–8.[Medline]
10 Lagios MD. Duct carcinoma in situ. Pathology and treatment. Surg Clin North Am (1990) 70:853–71.[ISI][Medline]
11 Fisher ER, Redmond C, Fisher B. Histologic grading of breast cancer. Pathol Annu (1980) 15:239–51.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||