Histologic grade, stage, and survival in breast carcinoma

Comparison of African American and Caucasian women

Abstract

BACKGROUND

African American women have lower breast carcinoma survival rates than do Caucasian women. African American women often present with advanced-stage disease and more aggressive tumors as shown by histologic and laboratory-based prognostic factors. Aggressive tumor behavior may be responsible, at least in part, for the advanced stage and reduced survival rates.

METHODS

The authors investigated the correlation between survival and histologic grade, stage of disease, and tumor size for both African American and Caucasian women who were younger than age 50 years and age 50 years and older. The authors also investigated the distribution of grade within each stage group and the distribution of grade by tumor size. African American and Caucasian women were matched by stage, tumor size, and histologic grade. Survival was represented by 6-year breast carcinoma–specific survival rates.

RESULTS

Compared with Caucasian women, African American women, regardless of age, had proportionally more Grade III tumors and fewer Grade I and II tumors for all stages combined and for each individual stage group. Similarly, matched for tumor size, African American women had more Grade III tumors and fewer Grade I and II tumors compared with Caucasian women, except for tumors smaller than 1.0 cm. For nearly all combinations of stage and grade regardless of age, the 6-year breast carcinoma–specific survival rate was lower for African American women than for Caucasian women, although it did not always reach statistical significance.

CONCLUSIONS

Compared with Caucasian women, African American women, regardless of age, presented with proportionally more aggressive tumors for each stage of disease and for each tumor size above 1.0 cm as revealed by the histologic grade. Higher histologic grade may be a significant contributing factor to survival disadvantage for African American women. Cancer 2003;98:908–17. Published 2003 by the American Cancer Society.

DOI 10.1002/cncr.11558

Although African American women have a lower incidence of breast carcinoma than Caucasian women (100.2 cases per 100,000 and 114.0 cases per 100,000, respectively), African American women have a less favorable outcome than Caucasian women stage for stage. According to the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute, the 5-year survival rates, all stages combined, for African American and Caucasian women are 69.0% and 84.4%, respectively.1 Further evidence for this survival disadvantage is based on institutional studies of racial differences in outcome.2, 3

Socioeconomic status (SES), lack of access to medical care, less utilization of screening, variations in treatment, comorbid conditions, lack of compliance with treatment recommendations, and late stage presentation have all been offered as explanations for the lower survival rates among African American women.4–8 Although SES, even more than race, is considered by many the most important predictor of outcome,5, 9, 10 studies continue to show that African American women present with more aggressive breast tumor types compared with Caucasian women.11–13 In the current study, we investigated histologic grade as it relates to stage of disease, tumor size, and survival between African American and Caucasian women using data from the SEER program. The results indicate that aggressive tumor behavior as revealed by histologic grade should also be taken into account in studies on disparity.

MATERIALS AND METHODS

Data were obtained from the SEER program for the period January 1992 to December 1999. SEER covers 11 geographic areas: Atlanta, GA; Connecticut; Detroit, MI; Hawaii; Iowa; Los Angeles, CA; New Mexico; San Francisco–Oakland, CA; San Jose–Monterey, CA; Seattle–Puget Sound; and Utah. These areas include approximately 12% of the U.S. population and represent national demographic patterns. Population based from inception, SEER has accrued data since 1973 in most survey areas through regional and statewide tumor registries.1

Six-year cause-specific survival rates were calculated using death certificate data that listed breast carcinoma as the underlying cause of death. These disease-specific survival rates and standard errors were calculated for African American and Caucasian women using the SEER Public Use CD-ROM, 1973–1999, with its SEER*Stat 4.2 computer program. In situ cases (Stage 0) were excluded from the analysis. The number of cases recorded for each combination of grade and stage is listed in Tables 1–4. In the current study, Caucasian is used synonymously with white. All other racial/ethnic groups were omitted except for African American women.

Table 1. Comparison of 6-Year Breast Carcinoma Survival Rates by Race, Grade, and American Joint Committee on Cancer Stage for Women Younger than Age 50 Years

	No. of patients		6-yr survival rate		Black/White survival rate difference
	No. of White	Black	White	Black
AJCC: American Joint Committee on Cancer; na: not available. a P value (two-sided) < 0.01 (0.05/5).
AJCC Stages I–IV
Grade I–IV	20,064	2828	0.84	0.68	0.17a
Grade I	2238	177	0.98	0.96	0.02
Grade II	7407	776	0.90	0.77	0.13a
Grade III	9516	1771	0.78	0.61	0.17a
Grade IV	903	104	0.75	0.59	0.16
Unknown	3600	661	0.85	0.74
AJCC Stage I
Grade I–IV	7702	782	0.96	0.91	0.05a
Grade I	1492	98	0.99	0.99	0.00
Grade II	3287	283	0.98	0.93	0.05
Grade III	2668	380	0.92	0.87	0.04
Grade IV	255	21	0.91	0.93	0.02
Unknown	1567	218	0.97	0.93
AJCC Stage II
Grade I–IV	9987	1534	0.84	0.70	0.14a
Grade I	672	70	0.98	0.95	0.03
Grade II	3495	400	0.87	0.74	0.13a
Grade III	5339	1008	0.80	0.67	0.13a
Grade IV	481	56	0.78	0.66	0.12
Unknown	1490	318	0.87	0.79
AJCC Stage III
Grade I–IV	1777	363	0.62	0.36	0.25a
Grade I	61	5	0.88	1.00	−0.12
Grade II	469	68	0.72	0.53	0.19
Grade III	1119	276	0.58	0.31	0.27a
Grade IV	128	14	0.53	0.41	0.12
Unknown	299	54	0.58	0.42
AJCC Stage IV
Grade I–IV	598	149	0.24	0.11	0.14a
Grade I	13	4	na	na
Grade II	156	25	0.32	0.25	0.07
Grade III	390	107	0.23	0.08	0.16a
Grade IV	39	13	na	na
Unknown	244	71	0.21	0.17

Table 2. Comparison of 6-Year Breast Carcinoma Survival Rates by Race, Grade, and American Joint Committee on Cancer Stage for Women Age 50 Years and Older

	No. of patients		6-yr survival rate		Black/White survival rate difference
	White	Black	White	Black
AJCC: American Joint Committee on Cancer; na: not available. a P value (two-sided) < 0.01 (0.05/5).
AJCC Stages I–IV
Grade I–IV	58,963	4769	0.86	0.74	0.12a
Grade I	11,182	631	0.97	0.92	0.05a
Grade II	26,100	1728	0.90	0.82	0.09a
Grade III	19,936	2280	0.75	0.64	0.11a
Grade IV	1745	130	0.76	0.69	0.07
Unknown	14,382	1576	0.83	0.71
AJCC Stage I
Grade I–IV	31,226	1858	0.96	0.93	0.04a
Grade I	8520	414	0.99	0.98	0.01
Grade II	14,793	821	0.97	0.92	0.05a
Grade III	7246	588	0.93	0.90	0.03
Grade IV	667	35	0.93	1.00	−0.07a
Unknown	7379	596	0.97	0.95
AJCC Stage II
Grade I–IV	21,834	2099	0.82	0.75	0.07a
Grade I	2358	186	0.94	0.92	0.02
Grade II	9385	718	0.87	0.81	0.06a
Grade III	9321	1130	0.75	0.69	0.06a
Grade IV	770	65	0.78	0.74	0.04
Unknown	4596	580	0.86	0.77
AJCC Stage III
Grade I–IV	3754	500	0.55	0.40	0.15a
Grade I	192	17	0.88	0.35	0.53a
Grade II	1254	121	0.65	0.55	0.10
Grade III	2111	348	0.46	0.36	0.10
Grade IV	197	14	0.52	na	na
Unknown	939	162	0.51	0.33
AJCC Stage IV
Grade I–IV	2149	312	0.16	0.10	0.06
Grade I	112	14	0.30	na	na
Grade II	668	68	0.23	0.14	0.09
Grade III	1258	214	0.12	0.10	0.02
Grade IV	111	16	0.06	na	na
Unknown	1468	238	0.15	0.12

Table 3. Comparison of 6-Year Breast Carcinoma Survival Rates by Race, Grade, and Tumor Size for Women Younger than Age 50 Years

	No. of patients		6-yr survival rate		Black/White survival rate difference
	White	Black	Whites	Blacks
na: not available. a P value (two-sided) < 0.0125 (0.05/4).
Tumor size < 1.0 cm
Grade I–IV	3435	291	0.95	0.89	0.06
Grade I	888	54	0.99	0.98	0.01
Grade II	1501	120	0.96	0.87	0.09
Grade III	947	106	0.91	0.86	0.05
Grade IV	99	11	0.91	na	na
Unknown	896	120	0.96	0.91
Tumor size 1.1–2.0 cm
Grade I–IV	7412	890	0.91	0.81	0.10a
Grade I	915	75	0.98	0.95	0.03
Grade II	3094	288	0.95	0.83	0.12a
Grade III	3115	503	0.86	0.78	0.08a
Grade IV	288	24	0.86	0.87	−0.01
Unknown	1111	188	0.93	0.89
Tumor size 2.1–5.0 cm
Grade I–IV	6964	1154	0.80	0.66	0.14a
Grade I	342	36	0.96	0.94	0.02
Grade II	2200	276	0.84	0.75	0.09
Grade III	4059	800	0.77	0.61	0.16a
Grade IV	363	42	0.72	0.61	0.11
Unknown	986	229	0.83	0.72
Tumor size ≥ 5.1 cm
Grade I–IV	1286	312	0.65	0.42	0.23a
Grade I	53	6	0.86	0.83	0.02
Grade II	331	50	0.72	0.59	0.12
Grade III	805	243	0.63	0.39	0.24a
Grade IV	97	13	0.53	0.26	0.27
Unknown	275	59	0.65	0.34

Table 4. Comparison of 6-Year Breast Carcinoma Survival Rates by Race, Grade, and Tumor Size for Women Age 50 Years and Older

	No. of patients		6-yr survival rate		Black/White survival rate difference
	White	Black	White	Black
na: not available. a P value (two-sided) < 0.0125 (0.05/4).
Tumor size < 1.0 cm
Grade I–IV	15,460	872	0.97	0.92	0.05a
Grade I	5298	249	0.99	0.98	0.01
Grade II	7086	406	0.97	0.94	0.03
Grade III	2780	204	0.93	0.81	0.12a
Grade IV	296	13	0.92	0.91	0.01
Unknown	4021	313	0.97	0.95
Tumor size 1.1–2.0 cm
Grade I–IV	22,817	1570	0.92	0.86	0.06a
Grade I	4239	226	0.98	0.93	0.05
Grade II	11,139	655	0.94	0.87	0.07a
Grade III	6875	654	0.86	0.82	0.04
Grade IV	564	35	0.84	0.82	0.02
Unknown	4616	439	0.93	0.85
Tumor size 2.1–5.0 cm
Grade I–IV	15,985	1664	0.76	0.69	0.07a
Grade I	1339	129	0.92	0.91	0.01
Grade II	6391	513	0.82	0.74	0.08
Grade III	7619	963	0.70	0.63	0.06a
Grade IV	636	59	0.74	0.68	0.06
Unknown	3534	459	0.78	0.69
Tumor size ≥ 5.1 cm
Grade I–IV	2626	421	0.50	0.39	0.11a
Grade I	185	17	0.73	na	na
Grade II	859	89	0.60	0.52	0.08
Grade III	1440	304	0.41	0.35	0.06
Grade IV	142	11	0.56	na	na
Unknown	836	166	0.53	0.39

Staging was based on the TNM classification as published by the American Joint Committee on Cancer (AJCC) with four stages (Stages I–IV).14 Stage IV survival rates could not be determined for each grade for African American women. Histologic grade and tumor size were recorded from pathology reports. There were 4 categories for tumor size (i.e., < 1.0 cm, 1.1–2.0 cm, 2.1–5.0 cm, and ≥ 5.1 cm, which corresponds to AJCC staging). To evaluate the effect of age, the study population was divided into 2 groups (younger than age 50 years and age 50 years and older).

Methods for the statistical analyses of differences in stage distributions and differences in survival rates between Caucasian and African American women have been reported.15 Corrections for multiple comparisons were used.

RESULTS

Distribution of Histologic Grade by American Joint Committee on Cancer Stage

Figure 1 compares the grade distribution for African American and Caucasian women (younger than age 50 years) for Stages I–IV combined, and for each individual stage, excluding cases of unknown grade. In general, Caucasian women had proportionally more Grade I and II tumors whereas African American women had a higher proportion of Grade III tumors for all stages combined and for each stage group. Figure 2 shows the distribution of grade for women (age 50 years and older) for Stages I–IV combined and for each stage. Again, African American women had proportionally more Grade III tumors and proportionally fewer Grade I and II tumors than Caucasian women did.

Figure 1. Comparison of the proportional distribution of histologic grade for Caucasian women (white bars) and African American women (blue bars) younger than age 50 years, for all stages combined and for each stage. Patients were staged according to the American Joint Committee on Cancer TNM classification.

Figure 2. Comparison of the proportional distribution of histologic grade for Caucasian women (white bars) and African American women (blue bars) age 50 years and older, for all stages combined and for each stage grouping. Patients were staged according to the American Joint Committee on Cancer TNM classification.

For all stages combined (Figs. 1, 2), African American women had a significantly higher proportion of Grade III tumors than Caucasian women in both age groups. For Caucasian women younger than age 50 years, the proportion of Grade III tumors was significantly higher than the proportion of Grade II tumors. In contrast, for Caucasian women age 50 years and older, the proportion of Grade III tumors was significantly less than the proportion of Grade II tumors. However, this pattern was not seen among African American women. Among African American women age 50 years and older, the proportion of Grade III tumors was significantly higher than the proportion of Grade II tumors.

Stage, Histologic Grade, and Survival

For each combination of grade and stage, the 6-year disease-specific survival rate for African American women younger than age 50 years was consistently lower than that for Caucasian women (Table 1). Although this observation did not reach statistical significance for every combination of grade and stage, it was observed in 11 of the 14 combinations that were analyzed (Table 1). The difference in survival was significantly smaller for African American women for Grade II and III tumors in Stages I–IV combined (Table 1). It also was significantly smaller for all grades and stages combined (Table 1). African American women had significantly lower survival rates for Grade II and III tumors in Stage II, Grade III tumors in Stage III, and for Grade III tumors in Stage IV.

For each combination of grade and stage, the 6-year disease-specific survival rate for African American women age 50 years and older was consistently lower than that for Caucasian women (Table 2). This difference did not reach statistical significance for every combination of grade and stage, but it was observed in 12 of the 13 combinations that were analyzed (Table 2). The difference in survival was significantly less for African American women for Grade II tumors in Stage I, for Grade II and III tumors in Stage II, and for Grade I tumors in Stage III (Table 2). It was also significant for all grades and stages combined.

Distribution of Histologic Grade by Tumor Size

The distribution of grade evaluated in correlation to tumor size is shown in Figures 3 and 4. For every size except tumors smaller than 1.0 cm, African American women had significantly more Grade III tumors in both age groups than did Caucasian women. Figures 3 and 4 also indicate that grade increases with tumor size in both African American and Caucasian women, regardless of age. For instance, for tumor size 1.1–2.0 cm in women younger than age 50 years, 42% and 57% of Caucasian and African American women, respectively, had Grade III tumors. For tumor sizes 5.1 cm or larger, 63% of Caucasian women had Grade III tumors, compared with 78% of African American women. A similar observation was made for women age 50 years and older.

Figure 3. Comparison of the proportional distribution of histologic grade for Caucasian women (white bars) and African American women (blue bars) younger than age 50 years, according to the size of the primary tumor.

Figure 4. Comparison of the proportional distribution of histologic grade for Caucasian women (white bars) and African American women (blue bars) age 50 years and older, according to the size of the primary tumor.

Tumor Size, Histologic Grade, and Survival

Tables 3 and 4 show, respectively, the 6-year breast carcinoma–specific survival rates for each combination of tumor size and grade for women younger than age 50 years and for women age 50 years and older. For all but 1 combination of size and grade (Grade IV, 1.1–2.0 cm) in women younger than age 50 years, African American women had a less favorable outcome.

DISCUSSION

It has been known for years that histologic grade is correlated with survival in women with breast carcinoma. Greenough,16 who defined three grades based on his analysis of eight histologic variables, reported the first studies in 1925. These variables included cell size, glandular formation, epithelial secretion, nuclear size, hyperchromatism, variation in size of the cell and nucleus, and mitotic activity. Since then, many investigators have confirmed the observations of Greenough.17–19 Clinical trials and population studies have further confirmed the utility of grade as a prognostic factor.20–22 The SEER data show that African American women have a less favorable 6-year cause-specific survival than Caucasian women for nearly every combination of stage and grade, regardless of age. Although these differences did not reach statistical significance for most combinations, they were observed in 11 of the 14 combinations of stage and grade in women younger than age 50 years and in 13 of 14 combinations in women age 50 years and older. These results indicate that breast carcinoma survival rates for African American women are not only less favorable for all stages, they are also less favorable for nearly all combinations of stage and grade.

For all stages, African American women, regardless of age, presented with proportionally more Grade III and fewer Grade I or II tumors compared with Caucasian women. This finding indicates that histologic grade, which is a reflection of aggressive tumors, may be an important factor in racial/ethnic disparity. Socioeconomic factors, access to care, less utilization of screening, treatment differences, and late-stage presentation have been proposed as major causes of disparity. However, the results of the current study suggest that histologic grade also needs to be considered. Grade is evaluated at time of diagnosis and therefore reflects events occurring in the tumor before discovery and treatment. In malignant tumors, grade often correlates with aggressive behavior because higher-grade tumors usually show higher mitotic rates and, by definition, less differentiation. Previous studies have revealed that African American women are more likely to present with estrogen receptor–negative tumors, higher S-phase fractions, and higher mitotic activity12, 13 factors usually associated with a less favorable outcome and more aggressive disease. These factors also correlate with higher-grade tumors.

When the patterns of grade distribution were considered for all stages combined (Figs. 1, 2), differences were observed between African American women and Caucasian women. Combining stage groups merges all cases into a single group and enables an evaluation of the distribution of grade alone in the study population. For Caucasian women age 50 years and older, the proportion of Grade III tumors was significantly lower than the proportion of Grade II tumors, which differs from the finding for women younger than age 50 years. However, for African American women in both age groups, the proportion of Grade III tumors was significantly higher than the proportion of Grade II tumors. It probably is premature to speculate on the reasons for these different patterns, but they reflect the grade distribution in all stages.

Similarly, for each tumor size, African American women consistently had reduced survival compared with Caucasian women. Although tumor size is part of the definition of stage, the data corroborate the reduced survival by stage in African American women.

Additional comments should be made. Because the SEER data are derived from widely different areas of the country, pathology practice patterns (including the assignment of grade) will vary in these geographic areas. The assessment of grade is subjective and is a function of education and practice experience, which varies among institutions. However, despite these variations and lack of standardized grading criteria, the black-white differences in outcome were consistent within the combinations of stage and grade.

Improved grading systems, such as the Nottingham Combined Histological Grade, have been recommended.23 The Nottingham system not only provides uniform and quantitative data, it also provides measurements of differentiation, nuclear grade, and mitotic count, which are important parameters in the assessment of tumor aggressiveness.

Published studies have shown that survival rates for Grade III and IV tumors are not statistically different.24 As a result, most pathologists use only three grades for breast carcinoma, combining grades III–IV. For this reason, the term ‘grades III–IV’ often is used in the literature to indicate that some pathologists still use Grade IV. Nonetheless, the data clearly indicate that African American women have a less favorable grade distribution at diagnosis for all stages compared with Caucasian women, even when Grade III and Grade IV tumors were combined.

Whereas low SES and/or lack of screening may account for late-stage disease and larger tumor size at presentation, it is not so obvious that low SES or lack of screening accounts for the higher-grade tumors seen in African American women compared with Caucasian women. Although all pathologists will agree that high tumor grade indicates a more aggressive type of tumor, grade and other prognostic factors have not been considered a function of delayed diagnosis or lack of screening in contrast to stage and tumor size.

Others have observed a higher proportion of Grade III–IV breast tumors among African American women compared with Caucasian women.11, 12, 25–27 Our data not only confirm these observations but also indicate that higher-grade tumors are found in all stages of breast carcinoma among African American women. For this reason, efforts to detect early-stage disease through screening, although improving survival, may not eliminate the survival disparity because of the higher-grade disease already present in Stage I and II among African American women.