Clinicopathologic characteristics of 143 patients with synchronous bilateral invasive breast carcinomas treated in a single institution
Article first published online: 9 JUL 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 5, pages 905–912, 1 September 2004
How to Cite
Intra, M., Rotmensz, N., Viale, G., Mariani, L., Bonanni, B., Mastropasqua, M. G., Galimberti, V., Gennari, R., Veronesi, P., Colleoni, M., Tousimis, E., Galli, A., Goldhirsch, A. and Veronesi, U. (2004), Clinicopathologic characteristics of 143 patients with synchronous bilateral invasive breast carcinomas treated in a single institution. Cancer, 101: 905–912. doi: 10.1002/cncr.20452
- Issue published online: 18 AUG 2004
- Article first published online: 9 JUL 2004
- Manuscript Accepted: 24 MAY 2004
- Manuscript Revised: 19 MAY 2004
- Manuscript Received: 5 APR 2004
- breast carcinoma;
- contralateral cancer
Synchronous bilateral invasive breast carcinoma (SBIBC) ranged in incidence from 0.3% to as high as 12%.
Between April 1997 and February 2003, 143 consecutive patients with SBIBC were treated at the European Institute of Oncology (Milan, Italy). Their information was collected prospectively in a database. The bilateral tumors were divded into left and right tumors. Tumor size, histology, grade, lymph node status, estrogen (ER) and progesterone receptor (PgR) status, HER-2 expression, peritumoral vascular invasion (PVI), Ki-67 expression, extensive in situ component (EIC), and multifocality between the two groups were analyzed. During the same time period, 6218 patients with unilateral invasive breast carcinoma (UIBC) were analyzed in the same manner for comparison with the patients with SBIBC.
There were no significant differences between left and right tumors, and the observed histopathologic agreement within the same patient was significantly superior than statistically expected for all characteristics except size, lymph node status, and multifocality. When compared with patients with UIBC, patients with SBIBC were more likely to present with smaller tumors and showed a higher frequency of invasive lobular carcinoma, lower histologic grade, higher rate of ER and PgR positivity, and lower PVI and Ki-67 expression.
The high concordance of histopathologic characteristics between SBIBC within the same patient could reflect a particular hormonal environment that influenced either the initiation and development of these lesions simultaneously and independently from the single or multi-clonal origin, either a less aggressive biological behavior compared with UIBC. In particular, the strong agreement of the observed EIC in SBIBC within the same patient seemed to definitively exclude the metastatic origin of these tumors. Cancer 2004. © 2004 American Cancer Society.
Synchronous bilateral invasive breast carcinoma (SBIBC) is not uncommon, its incidence ranging from 0.3% to 12%.1–3 However, the histopathologic characteristics and the biologic behavior of SBIBC are still unclear. In fact, it is not clear whether SBIBC should be considered a sequential event of a primary tumor or, rather, an independent second primary sharing with the primary localization a biologically predisposing common substratum.4–6 The demonstration of different histopathologic characteristics of synchronous tumors in the left and right breasts and the presence of an in situ component in both types of tumors have been considered to be indicative factors of an independent development.3, 4, 7 Other authors have shown positive correlations in histologic subtype,8, 9 tumor grade,10 and hormone receptor status4, 11 between the two tumors and, based on these similarities, have suggested either a single cell origin with a secondary metastatic spread of cancer cells to the opposite breast12, 13 or a particular hormonal environment favoring similar biologic and pathologic features in different tumor foci of both breasts.11 The current study had two goals. First, the study described the clinicopathologic features of SBIBC in a large cohort of patients consecutively treated in a single institution and compared these characteristics with those of a large cohort of patients with unilateral invasive breast carcinoma (UIBC) treated during the same time period, in the same institution. Second, the study investigated the degree of concordance of pathological tumor characteristics within the same patient.
MATERIALS AND METHODS
Between April 1997 and February 2003, 8378 patients with invasive breast carcinoma were treated at the European Institute of Oncology (Milan, Italy). Patients who received neoadjuvant therapy, had metastatic and advanced disease, and those who had local disease recurrences were excluded from the study. Of the remaining 6361 patients, the characteristics of 143 consecutive female patients with SBIBC were retrieved from the institutional database. The cumulative incidence of SBIBC was 2.2%. Patients with SBIBC were defined as having a contralateral tumor diagnosed simultaneously at the time of the initial diagnosis or within 3 months from the first malignancy. The bilateral tumors were divided into left and right tumors and analyzed for clinicopathologic characteristics including tumor size, histology, grade, axillary lymph node status, estrogen (ER) and progesterone receptor (PgR) status, HER-2 expression, peritumoral vascular invasion (PVI), Ki-67 expression, extensive in situ component (EIC), and multifocality. The mastectomy specimens were examined fresh, and the maximum tumor diameter was recorded. Histologically, the tumors were classified according to the World Health Organization criteria,14 as modified by Rosen and Oberman.15 Tumor grade was evaluated following the recommendations of Elston and Ellis.16 The occurrence of PVI was assessed according to Rosen and Oberman.15 The expression of ER and PgR and the tumor proliferative fraction were evaluated immunohistochemically as previously reported.17 Results were recorded as the percentage of cells showing definite nuclear immunoreactivity over ≥ 2000 neoplastic cells in 10 randomly selected high power fields (HPF; HPF = × 400) at the periphery of the tumor. HER-2/neu overexpression was also investigated immunohistochemically, using a specific polyclonal antiserum (Dako A/S [Carpinteria, CA], working dilution 0.05 mol/L) and the Envision detection reagent (Dako A/S), according to the manufacturer's instructions. The staining results were recorded in a four-tier scale, from 0 to 3+, according to the percentage of immunoreactive cells, and to the intensity and completeness of membrane staining as recommendend by the Food and Drug Administration.18
Data on the patient's age, menopausal status, family history, and type of surgery performed were also collected. For the purpose of the current study, a positive family history was defined as any first or second-degree relative with a history of breast carcinoma. During the same time period, 6218 patients with UIBC were treated in our institution and analyzed in the same manner for comparison with the patients with SBIBC. All tumor specimens were analyzed through the Department of Pathology at the European Institute of Oncology. When surgery was performed elsewhere, all pathologic slides were reviewed at our institution to confirm diagnosis.
Patient and treatment characteristics (Table 1) were compared between patients with SBIBC and patients with UIBC by using the Pearson chi-square test. The pathologic characteristics (Tables 2, 3) of the two patient groups was compared by using generalized linear models, one for each factor, within a generalized estimating equation approach (GEE), so as to account for the positive correlation observed in tumor characteristics of the two SBIBCs within the same patient. As expected, P values obtained with the above approach were always slightly higher than those obtained with the conventional Pearson chi-square test. Nevertheless, the outcome never changed in terms of statistical significance.
|Characteristics||SBIBC (n = 143) (%)||UIBC (n = 6218) (%)||P value|
|Age (range)||56 (33–84 yrs)||53 (20–95 yrs)||0.002|
|Premenopausal||53 (37)||2529 (41)|
|Postmenopausal||90 (63)||3469 (59)|
|Family history of breast carcinoma||0.775|
|None||98 (74)||2259 (74)|
|Yes, first-degree relative||20 (15)||474 (16)|
|Yes, second-degree relative||15 (11)||285 (10)|
|Bilateral BCS||81 (57)||—|
|Bilateral mastectomy||32 (22)||—|
|BCS plus mastectomy||30 (21)|
|Unilateral BCS||—||4856 (78)|
|Unilateral mastectomy||—||1362 (21)|
|Characteristics||Bilateral||Unilateral No. (%)||P value|
|Left tumors No. (%)||Right tumors No. (%)|
|No. of tumors evaluated||143||143||6218|
|Mean size (range)||2.0 (0.03–9 cm)||1.8 (0.04–6.5 cm)||1.9 (0.01–18.0 cm)||< 0.001|
|Invasive ductal carcinoma||100 (70)||99 (69)||4801 (77)|
|Invasive lobular carcinoma||22 (15)||23 (16)||588 (9)|
|Mucinous carcinoma||2 (1)||3 (2)||106 (2)|
|Cribiform carcinoma||4 (3)||5 (3)||195 (3)|
|Mixed||15 (10)||13 (9)||528 (9)|
|1||34 (28)||28 (23)||1059 (18)|
|2||57 (46)||66 (54)||2698 (47)|
|3||32 (26)||29 (24)||2004 (35)|
|Axillary lymph node status||0.676|
|pN−||52 (53)||51 (52)||2820 (50)|
|1–3||27 (28)||31 (32)||1849 (33)|
|4–9||16 (16)||8 (8)||540 (9)|
|≥ 10||3 (3)||8 (8)||445 (8)|
|Negative||19 (13)||14 (10)||1047 (17)|
|Weakly positive (< 10%)||1 (1)||3 (2)||67 (1)|
|Strongly positive (> 10%)||122 (86)||125 (88)||4979 (82)|
|Progesterone receptor||< 0.001|
|Negative||31 (22)||21 (15)||1819 (30)|
|Weakly positive (< 10%)||9 (6)||17 (12)||470 (8)|
|Strongly positive (> 10%)||102 (72)||104 (73)||3808 (62)|
|Not expressed||93 (65)||94 (65)||2625 (60)|
|+||12 (8)||19 (13)||486 (11)|
|++||17 (12)||11 (7)||462 (11)|
|+++||20 (14)||18 (12)||772 (18)|
|Absent||92 (79)||92 (79)||4158 (70)|
|Present||25 (21)||25 (21)||1819 (30)|
|Weakly positive (< 20%)||87 (62)||95 (67)||2977 (49)|
|Strongly positive (> 20%)||54 (38)||46 (33)||3050 (51)|
|Extensive in situ component||0.314|
|Absent||104 (73)||106 (74)||4739 (77)|
|Present||39 (27)||37 (26)||1450 (23)|
|Absent||119 (83)||117 (82)||5245 (85)|
|Present||24 (17)||26 (18)||944 (15)|
|Characteristics||Agreement||Cohen kappa||P value|
|Lymph node status||0.4592||0.3958||0.1049||0.1750|
Restricting the analysis to patients with SBIBC, we first divided the bilateral tumors into left and right tumors and investigated their pathologic characteristics by means of cross-classification contingency tables. Each table was analyzed in two ways. First, the McNemar test for 2 × 2 tables, or its generalization to multinominal responses,19 was performed with the aim of checking whether the investigated characteristic differed on average between right and left tumors. Second, the Cohen kappa value20 was estimated and tested. This statistic is a chance-corrected measure of agreement of the investigated characteristics within the same patient. A value of 1 indicates perfect agreement, whereas a value of 0 denotes the lack of agreement. The associated P value is used for testing whether the estimated kappa value differs from 0. Results of this analysis are shown by reporting the observed agreement (proportion of patientsin whom the two lesions have the same characteristics), the agreement expected by chance, together with the estimated Cohen kappa value and corresponding P value. Computations were made by use of SAS21 and the StatXact software.22 Statistical significance was defined as two-tailed P < 0.05.
Table 1 shows the clinical characteristics of the 143 patients with SBIBC and the control group of 6218 patients with UIBC. In the SBIBC group, the average age of the patients was 56 years (range, 33–84 years; median, 54 years). Fifty-three patients (37%) were premenopausal and 90 (63%) were postmenopausal. Thirty-five patients (26%) had a positive family history for breast carcinoma, whereas no data were available for 10 patients. Most women (57%) received bilateral conservative surgery, whereas the remaining women underwent either bilateral mastectomy (22%) or mastectomy on one side and breast-conserving therapy on the contralateral side (21%). In comparison to the patients with UIBC, there were no significant differences between groups in terms of menopausal status and family history for breast carcinoma, whereas the mean age was higher in the SBIBC group (56 vs. 53 years, P = 0.002). Furthermore, patients with SBIBC tumors were more likely to be treated with mastectomy than patients with UIBC (P < 0.001).
Table 2 shows the tumor pathologic characteristics in the two cohorts. As expected, no difference was observed between left and right tumors as regards pathologic features (P values, not shown, always are well above the 5% significance threshold). In contrast, when compared with the patients with UIBC, patients with SBIBC were more likely to present with smaller tumors (1.90 cm on average for SBIBC vs. 1.98 cm, P < 0.001). Although the most frequent histopathologic subtype was ductal carcinoma in both groups, the distribution of the histologic subtypes was different between the groups as invasive lobular carcinoma was present in a higher percentage of patients with SBIBC than in patients with UIBC (15.5% vs. 9%, P = 0.017). Also, in comparison to patients with UIBC, patients with SBIBC were more likely to have histologic Grade 1 and Grade 2 tumors (Grade 1: 25.5% vs. 18%; Grade 2: 50% vs. 47%; Grade 3: 25% vs. 35%, P = 0.002), to be ER and PgR positive (ER: 88.5% vs. 83%, P = 0.041; PgR: 81.5% vs. 70%, P < 0.001), to lack PVI (79% vs. 70%, P = 0.008), and to have a lower Ki-67 labeling index (77% vs. 73.5, P < 0.001). Finally, there was no significant difference in the percentage of patients with axillary lymph node metastases (P = 0.676), HER-2 expression (P = 0.102), EIC (P = 0.314), and multifocality (P = 0.335) between the two cohorts.
Regarding concordance of tumor characteristics between the two lesions within the same patient, as estimated by the Cohen kappa value, significant results were obtained for the majority of investigated factors, namely, histologic subtype, grade, ER and PgR status, HER-2 expression, PVI, Ki-67 expression, and EIC (Table 3). The Cohen kappa estimates ranged from 0.23 (for Ki-67) to 0.49 (for EIC), figures that denote a moderate degree of agreement. In contrast, no statistically significant associations were obtained for tumor size, axillary lymph node status, and multifocality, for which the estimated Cohen kappa values were 0.09, 0.10, and 0.13, respectively.
The incidence of SBIBC has been shown to range from 0.3% to as high as 12%.1–3 The wide range of reported incidence rates is due, at least in part, to the use of different definitions of SBIBC and to differences in the evaluation of the second breast malignancy. Many authors require both breast carcinomas to be diagnosed within 1 year,23 6 months,6, 8, 24 or simultaneously.7, 9, 25, 26 These different criteria make meaningful comparisons difficult, especially when patient survival is considered.27 At our institution, patients with SBIBC were defined as having a contralateral breast tumor diagnosed at either the time of the initial diagnosis or within 3 months. Using this strict definition, the incidence of SBIBC in this cohort was 2.2% (143 of 6361 patients), a figure in the range of reported incidences of other studies. To our knowledge, this is the largest population of patients with SBIBC ever studied in a single institution.
In our series, the median age of the patients was 56 years (range, 33–84 years). This is in agreement with many studies on SBIBC.28–32 However, a few other studies have reported a higher median age,6, 27, 33, 34 whereas a younger age at diagnosis is rarely observed.35 In most series, the median age of patients with synchronous cancers is not significantly younger than the age of patients with bilateral metachronous and unilateral breast carcinomas, as confirmed in the current series. We did not find a trend toward a young age in the SBIBC group compared with the UIBC cohort (median age, 53 years). Closely related to age is menopausal status, which, however, is rarely reported in the literature. When the data were available, the percentage of postmenopausal women was variable, ranging from 40%29 to 62%.6 In our study, most patients with SBIBC were postmenopausal (63%), and no significant differences were observed among the patients with UIBC (59%, P = 0.121).
Family history is known to be associated with bilateral breast carcinoma. In our series, 26% of patients had a positive family history for breast carcinoma in first or second-degree relatives. A positive family history for breast carcinoma is a well recognized risk factor for bilateral breast carcinoma and many studies show that familial aggregation is frequently associated with a higher than average number of tumors in various organs, including the breast and ovaries.27, 30, 36 These tumors are more frequently bilateral and the age of onset occurs earlier than in women with sporadic tumors. However, in most patients, bilaterality does not necessarily reflect a genetic predisposition,37 unless it is associated with an early age of onset.38 Notably, although some studies support a correlation between family history and bilateral breast carcinoma,3, 28–30, 35, 39 this is not confirmed by others.26, 31, 40 Our results are consistent with the latter studies, as we observed the same proportion (26%) of positive family history in SBIBC and in UIBC, thus supporting the suggestion that interactions between genetic and hormonal environmental factors play a crucial role in the development of both unilateral and bilateral breast carcinoma.
Considerable debate exists regarding the surgical management of patients with SBIBC. Traditionally, most clinicians have approached bilateral disease more aggressively than unilateral disease and most studies have shown a disproportionately higher incidence of bilateral mastectomies performed in this cohort of patients.34 This aggressive approach was adopted based on the suggestion that patients with SBIBC have a worse prognosis than patients with UIBC, according to the concept that patients with tumors with worse prognostic factors need more extensive surgical treatment.25, 35 Conversely, at our institution, the majority of patients underwent bilateral breast conservation treatment (57%) and only as few as 22% of the patients underwent bilateral mastectomy. Although significantly higher than in the UIBC cohort (P < 0.001), the proportion of patients with SBIBC who received mastectomy in our series was much lower than in the majority of published SBIBC series. However, these latter studies covered a long time period during which the use of breast-conserving surgery increased and became the treatment of choice for early-stage breast carcinoma. Heaton et al.35 have shown that bilateral breast carcinoma is amenable to bilateral breast conservative treatment without affecting overall and disease-free survival, while enhancing patient cosmesis. For patients who received mastectomy on one side and quadrantectomy on the contralateral side, to realize the best breast symmetry, the conservative surgery was sometimes performed with the help of a plastic surgeon during a reduction mammaplasty.
The classification of the two lesions in SBIBC is crucial. Many series26, 27, 32 have defined “first” as the tumor initially diagnosed, or the “dominant” as the higher-stage tumor. The “contralateral” tumor is the second lesion diagnosed or the lower-stage tumor when the diagnosis is simultaneous. This arbitrary definition introduces important statistical bias. The first diagnosed tumor is usually the higher-stage tumor by definition and any comparison between the dominant and contralateral tumor with regard to tumor size and lymph node status would be statistically incorrect due to the imbalance of the two groups of patients. Such an imbalance would also bias the estimation of agreement of tumor characteristics within patients, which was an important focus of our analysis. Therefore, in our series, the bilateral tumors were divided into left and right tumors. This definition permits the comparison of two very well balanced and homogeneous groups of lesions and avoids statistical bias. Indeed, as expected, there was no statistically significant difference in the histopathologic characteristics between right and left tumors.
Considering the overall SBIBC characteristics, the average tumor sizes of left and right tumors were 1.8 cm and 2.01 cm with a wide range on both sides (0.03–9 cm on the left side and 0.04–6.5 cm on the right side). This variability reflects the common clinical presentation of SBIBC. In the presence of a large dominant tumor, the contralateral tumor is usually detected earlier. For this reason, many studies have shown that most contralateral tumors are detected by mammography whereas most primary tumors are diagnosed by clinical examination.28 Although the majority of patients had infiltrating ductal carcinoma in both breasts, the proportion of patients with invasive lobular carcinoma was significantly higher in the bilateral breast carcinoma population (15.5%) compared with our unilateral control group (9%) and is consistent with other series.31, 41 The majority of tumors were classified as Grade 2 (50%) with a stronger positive hormone receptor status (88.5% ER positive; 81.5% PgR positive) but the Grade 1 SBIBC rate was higher than Grade 1 UIBC (25.5 vs. 18%, P = 0.017). Similar data were reported by others. Tsuda et al.42 showed that Grade 3 tumors were infrequent in patients with SBIBC. Matsuo et al.31 showed that ER and PgR positivity rates were higher in patients with SBIBC than in patients with UIBC and suggested that the positive hormone receptor status of SBIBC might be related to less aggressive biologic behavior.
In 47.5% of patients with SBIBC, the axillary lymph nodes were metastatic. The high percentage of metastatic lymph nodes is likely to be due to the largeness of the tumors, the maximum diameter of which ranged between 0.03 cm and 6.5 cm. One-third of the tumors in our series overexpressed HER-2. These data do not confirm the high incidence of HER-2 overexpression (71%) in SBIBC reported by Safal et al.33 In contrast, Matsuo et al.31 reported a HER-2 positivity rate (18%) similar to that of unilateral breast carcinoma and consistent with our data. The presence of PVI was identical on each side (21%) and was\significantly lower than in UIBC population. Compared with the UIBC group, the Ki-67 labeling index was lower (i.e., < 20%) in the SBIBC group (64.5% vs. 49%, P < 0.001). One-fifth of our patient population exhibited multifocality on either side (17%). The slightly higher incidence in the SBIBC group, compared with multifocality in the UIBC group (15%), is consistent with other important series. Similar results have been reported by Tulusan et al.13 In that study, the diagnosis of multifocality was based on a random biopsy of the contralateral breast. Newman et al.29 showed that SBIBC was more likely to show multicentric disease. In other studies,39, 40, 43 multicentricity was a significant risk factor for the development of bilateral breast carcinoma. An EIC was present in the same percentage of patients with SBIBC and patients with UIBC. Overall, SBIBC behave less aggressively than UIBC.
The statistically significant correlation among the histologic characteristics of SBIBC within the same patient documented in the current study has been reported by others. Hungness et al.34 showed that the histologic type of breast carcinoma was the same in the two breasts in most patients. Coradini et al.5 reported that ER and PgR levels in the primary and contralateral breast tumors did not differ significantly. All the pathologic characteristics of SBIBC were more frequently concordant than expected and this was statistically significant for all considered items except for size, lymph node status, and multifocality (Table 3). In particular, histologic subtype was the same in both synchronous tumors in 67% of patients, tumor grade in 56% of patients, ER status in 88% of patients, HER-2 expression in 84% of patients, PVI in 74% of patients, Ki-67 labeling index in 64% of patients, and EIC in 80% of patients. The agreement of tumor size and lymph node status was not statistically significant, but this reflects the different clinical presentation and the different stage of SBIBC within the same patients.
A key issue of SBIBC origin is whether the disease is multiclonal or single clonal.4–6, 44 Some authors have considered the contralateral synchronous tumor to be an independent lesion from the primary tumor (multiclonal origin theory),3, 4, 7 whereas others considered SBIBC to be secondary to the metastatic spread from the opposite breast (single-clonal origin).12, 13 Dissimilarity in histology between SBIBC in the same patient was often cited in support of the independent primaries theory. In contrast, positive correlations in histologic subtype, tumor grade, and hormonal receptor status between the two tumors have been considered to be indicative of a single-cell origin with secondary metastatic spread of cancer cells to the opposite breast. Pandis et al.44 showed that SBIBC has the same cytogenetic aberrations, including evidence of polyclonality, as unilateral carcinomas. The cytogenetic findings in their series support the suggestion that the majority of contralateral carcinomas arose independently from the initial tumor and formed completely new primary tumors, but some may result from metastatic spread from one breast to the other. Imyanitov et al.45 and Agelopoulos et al.46 confirmed the clinical, epidemiologic, and genetic observations concerning the independent clonal origin of most, if not all, SBIBC. This distinction is critical because the clinical management of localized breast carcinoma differs from the treatment of metastatic disease.
To conclude, the observed concordance of histopathologic characteristics between SBIBC within the same patient in our series, in particular for histology, tumor grade, hormone receptor status, and expression of HER2 and Ki67, could reflect a particular hormonal environment that influenced either the initiation and development of these lesions simultaneously and independently from the single or multiclonal origin,5 either a less aggressive biological behavior compared with UIBC. In particular, the strong agreement of the observed EIC in SBIBC within the same patient (80% observed vs. 60% expected; P < 0.0001) in our series seems to definitively exclude the metastatic origin of these tumors. Furthermore, the higher incidence of lobular carcinoma in SBIBC associated with a slightly higher incidence of multifocality suggests that SBIBC could represent a special entity of multicentric breast carcinoma44, 46 with similar biologic characteristics of high hormone responsiveness. No comment on overall survival can be made on our patient population until sufficient follow-up has been achieved.
The authors thank Dr. R. Torrisi, E. Millen, J. Rososchansky, and W. Russel Edu for their assistance with the preparation of the article.
- 14World Health Organization. Histological typing of breast tumors, 2nd ed. International histological classification of tumors, no. 2. Geneva: WHO, 1981.
- 15RosenPP, ObermanHA, editors. Tumors of the mammary gland. Atlas of tumor pathology, third series. Washington DC: Armed Forces Instutute of Pathology, 1993.
- 19An introduction to categorical data analysis. New York: John Wiley & Sons, Inc., 1996..
- 21SAS™ technical report P-229, SAS/STAT™ software: changes and enhancements, release 6.07. Cary, NC: SAS Institute, Inc., 1992.
- 22Cytel Software Corporation. StatXact 4 for Windows. User manual. Cambridge, MA: Cytel Software Corporation, 1998.
- 39Special therapeutic problems: management of the contralateral breast. In: HerrisJR, HellmanS, HendersonIC, KinneDW, editors. Breast disease, 2nd ed. Philadelphia: JB Lippincott, 1991: 827–833..