Prospective Multicenter Validation of the Independent Prognostic Value of the Mitotic Activity Index in Lymph Node–Negative Breast Cancer Pa
http://www.100md.com
《临床肿瘤学》
the Departments of Pathology and Epidemiology and Biostatistics, and Medical Oncology, Vrije Universiteit Medical Center, Amsterdam
University Medical Center Nijmegen, Nijmegen, the Netherlands
Departments of Pathology, Stavanger University Hospital, Stavanger
the Gade Institute, University of Bergen, Norway
University Hospital, Antwerpen, Belgium
ABSTRACT
PURPOSE: To validate the independent strong prognostic value of mitotic activity index (MAI) in lymph node (LN) –negative invasive breast cancer patients younger than 55 years in a nationwide multicenter prospective study.
PATIENTS AND METHODS: Analysis of routinely assessed MAI and other prognosticators in 516 patients (median follow-up, 118 months; range, 8 to 185 months), without systemic adjuvant therapy or previous malignancies.
RESULTS: Distant metastases occurred in 127 patients (24.6%); 90 (17.4%) died as a result of metastases. MAI (< 10, 10) showed strong association with recurrence (hazard ratio [HR], 3.12; 95% CI, 2.17 to 4.50; P .0001) and mortality (HR, 4.42; 95% CI, 2.79 to 7.01; P < .0001). The absolute difference in 10-year Kaplan-Meier estimates of time to distant recurrence as well as survival was 22% between MAI less than 10 versus 10. This effect was independent of age, estrogen receptor (ER) status, and tumor diameter (which were significant prognosticators). In multivariate analysis with regard to patient age, tumor diameter, grade, ER status, and the St Gallen criterion, MAI proved to be an independent and the strongest prognosticator. Tubular formation (TF) and nuclear atypia (NA), as constituents of (expert revised) grade, had no (for TF) or limited (for NA, P = .048) additional prognostic value to the MAI. In the group with MAI less than 10, MAI less than 3 versus more than 3 had additional value but the classical threshold of 0 to 5 v 6 to 10 did not. With this additional subdivision of MAI as less than 3, 3 to 9, and more than 9, NA lost its additive prognostic value.
CONCLUSION: The MAI is the strongest, most widely available, easily assessable, inexpensive, well-reproducible prognosticator and is well suited to routinely differentiate between high- and low-risk LN-negative breast cancer patients younger than 55 years.
INTRODUCTION
Trials have shown that both prolonged hormonal and adjuvant systemic treatment (AST) can improve prognosis in subsets of lymph node (LN) –negative breast cancers. The absolute benefits of systemic adjuvant chemotherapy were most apparent in younger women and in women with a worse prognosis, whereas tamoxifen appeared to be effective irrespective of age and menopausal status.1,2 Clearly, in patients with a better prognosis, the absolute improvements in recurrence and survival will be less substantial and the benefits need to be weighed against the adverse effects of therapy. Any easily assessable, reproducible, and reliable prognostic marker that will add information about primary tumor characteristics in relation to the follow-up (in addition to the well-established prognosticators such as nodal status, tumor size, and age) would be useful for this purpose. Predictive factors such as estrogen receptor (ER) status and HER-2 status are well established for treatment decision making.
The St Gallen classification3 recommends AST for all LN-negative premenopausal patients except those at minimal/low risk, which was defined as having postsurgical size of the tumor 2 cm, grade 1, ER-positive and/or progesterone receptor (PR) –positive status, and age older than 35 years. However, general application of this guideline means that approximately 85% of all premenopausal LN-negative patients would be treated with AST, whereas only a relatively small number would develop distant metastases without AST. It has also been argued that the St Gallen classification is an inaccurate measure of prognosis for patients with LN-negative breast cancer and should be used with caution.4 Moreover, the reproducibility of grade, an important discriminator in LN-negative patients in this risk classification, is far from perfect (low value for grade 1 v 2 among different observers).5
Proliferation variables such as the tritiated thymidine labeling index (TLI)6,7 and percentage of S-phase cells assessed by flow cytometry8,9 are good prognostic factors. In one study, patients with LN-negative, rapidly proliferating tumors, defined according to TLI, significantly benefited from systemic adjuvant chemotherapy.10 However, TLI is technically cumbersome and not widely available. One of the strongest yet simple and well-reproducible proliferation-associated prognostic factors is the mitotic activity index (MAI). Its prognostic value has been shown in many retrospective and prospective studies using a fixed threshold (MAI < 10, favorable; MAI 10, unfavorable).11-19 The MAI is not sensitive to fixation delay20,21 and is already part of different grading systems.22-24 Despite proof that the MAI is an independent strong prognostic factor in many previous studies, a routine laboratory test for therapeutic decision requires evaluation in a prospective multicenter study.25,26 Often, such validations are lacking because they are expensive, time consuming, and sometimes regarded as unoriginal. To validate the MAI with respect to its reproducibility, accessibility, and prospective prognostic power in relation to other classical prognostic factors, the nationwide Multicenter Mammary Carcinoma Project (MMMCP) was set up in the Netherlands in 1987, including 34 hospitals.27 A total of 3,479 consecutive primary breast cancer patients were enrolled from October 1, 1987, through December 31, 1989. It has been reported previously that the MAI was well reproducible among the collaborating laboratories.28
This report describes the long-term prognostic value of different features in the 516 women with LN-negative invasive breast cancer who were younger than 55 years of age and without previous malignancies. In accordance with the general policy during the enrollment period in the Netherlands, none of the patients were receiving or had received adjuvant systemic treatment.
PATIENTS AND METHODS
Patients
Details of the MMMCP protocol have been described elsewhere.27 The essential details will be presented here. After pathologists and technicians received training in mitosis counting using the MMMCP protocol and reproducibility was proven among the participants,28 all consecutive primary breast cancer patients diagnosed in the collaborating hospitals from October 1, 1987, through January 1, 1990, were enrolled. Follow-up was updated annually and data collection closed in 2003. Of the 3,479 MMMCP patients registered, 603 patients (17.3%) were females younger than 55 years of age with LN-negative breast cancer. Of these, 14 had carcinoma-in-situ and 31 had a carcinoma-in-situ with a microinvasive component less than 1 mm, which is ineligible for mitoses counting; these patients were not considered for the study, leaving 558 patients. Of these, 27 had a history of previous malignancies (breast, n = 21; other organs, n = 6), four had poor-quality sections preventing adequate MAI assessment, and 11 patients were lost to follow-up, leaving 516 patients for further analysis. In total, 591 LN-positive patients younger than 55 years were recruited in the MMMCP.
All patients were treated with modified radical mastectomy (n = 174) or breast-conserving therapy (n = 342); adequate axillary LN dissection was always performed. Locoregional radiotherapy was given in patients who underwent breast-conserving therapy or had medially localized tumors. None of the LN-negative patients received any form of adjuvant systemic therapy (in agreement with the usual therapeutic regimens of many hospitals in the Netherlands in the 1980s). The survival results of the LN-negative patients with MAI less than 10 and 10 were compared with those of 451 patients younger than 55 years with LN-positive MMMCP. These 451 patients, a subset of all 591 LN-positive patients younger than 55 years, had not received any form of adjuvant systemic therapy (patients, including most LN-positive patients, did not commonly receive adjuvant chemotherapy in many hospitals in the Netherlands in the 1980s).
Postsurgical size of the tumor was measured in the fresh specimens; the tumors were cut in slices 0.5 cm thick, fixed in buffered 4% formaldehyde, and embedded in paraffin. At least 10 (median, 14) LNs were detected in the axillary LN dissection specimens. Paraffin sections 4 μm thick were cut and stained with hematoxylin and eosin. Histologic type was assessed afterward according to the WHO criteria.22 Grade was assessed during careful review by two pathologists with considerable experience in breast pathology, according to the Nottingham modification,23,24 using MAI 0 to 5 = 1, 6 to 10 = 2, and more than 10 = 3; nuclear atypia as mild = 1, moderate = 2, and marked = 3; and tubular formation as majority (> 75%) = 1, moderate (10% to 75%) = 2, and little or none (< 10%) = 3. The grade is the sum of tubular formation + nuclear atypia + MAI class, where the sum of 3, 4, or 5 is grade 1; the sum of 6 or 7 is grade 2; and the sum of 8 or 9 is grade 3. ER value was assessed in reference laboratories with the ligand-binding charcoal technique (cutoff, 10 fmol/mg protein).
The MAI was assessed in each of the participating laboratories by a total of 27 different technicians who were instructed at four training sessions held during the first 3 months. For the assessment of the MAI, the MMMCP protocol27,28 was used, which is as follows.
First, with a black marker on the glass slide, the most poorly differentiated peripheral area of the tumor is demarcated. Necrotic, heavily inflamed, or benign areas are avoided. This area, called the measurement area, is minimally 1 x 1 mm and maximally 5 x 5 mm. Second, in the measurement area, at x400 magnification (objective 40, field diameter 450 μm at the specimen level), mitoses are counted in 10 consecutive neighboring fields of vision in the most cellular area (representing a total area of 1.59 mm2 at the specimen level). Third, only certain mitoses are counted; doubtful structures and apoptotic bodies are ignored. The total resulting number of well-defined mitotic figures counted in the 10 fields of vision is the MAI. Fourth, if the MAI is between 5 and 15, it is assessed once more, and the highest number of the two counts is taken as the MAI.
An accurate MAI assessment takes 3 to 5 minutes. Correction of the MAI for the percentage of tissue occupied by stroma or the number of tumor cells was not applied because it was shown previously that this does not substantially improve the prognostic value of the MAI and is substantially time consuming.29 The MAI is a continuous variable. According to many previous studies, the most important prognostic threshold is 10, with MAI less than 10 indicating favorable prognosis and MAI 10 indicating poor prognosis11-19,29 (this threshold is denoted as MAI10), whereas patients with MAI less than 3 have an especially favorable prognosis.11
Statistical Analysis
Main end points were recurrence and mortality. In analyzing the probability that patients would remain free of distant metastases, we defined recurrence as any first recurrence at distant sites. All other patients were censored on the date of the last follow-up visit; deaths as a result of causes other than breast cancer, local or regional recurrences, or the development of a secondary primary cancer (including contralateral breast cancer) were also censored. Mortality was defined as any death as a result of distant metastases (as evident from clinical, radiologic, histologic, or autopsy data; no patients died as a result of locoregional disease). If the cause of death was unknown, but a metastasis was detected previously, then death was considered to related breast cancer unless explicitly stated otherwise (in line with other studies).1 If the status during follow-up indicated a confirmed metastasis without a date of recurrence, the date of that follow-up visit was used.
Age, time to first recurrence, and survival time were calculated relative to the date of primary diagnosis. Median follow-up of the 516 patients was 118 months (range, 8 to 185 months). Distant metastases occurred in 127 patients (24.6%) and 90 patients (17.4%) died as a result of disease. Survival curves were constructed using the Kaplan-Meier technique.30 Differences between groups were tested by log-rank tests or tests for trend. The relative importance of potential prognostic variables was tested using Cox proportional hazards analysis and expressed as hazard ratios (HRs) with 95% CIs.31 All variables were tested for proportionality and continuous variables were checked for linearity or nonlinearity and transformed or recoded if necessary (or useful).
RESULTS
Table 1 shows the 10-year recurrence and disease-related mortality estimates and HRs for the most relevant patient and tumor characteristics. The MAI showed a prognostically optimal cutoff value at 10 and tumor size at 2 cm.
Grade was a strong prognostic factor, both for recurrence and mortality (Figs 1A and 1B). When the relation between the different grade components (tubulus formation, nuclear atypia, and MAI) and survival was studied, the difference in survival between favorable and unfavorable values of any of these features was largest with MAI (92% for MAI < 10 v 70% if MAI 10, a difference of 22%; Figs 2A and 2B). Figure 3 compares the survival curves of LN-negative MAI less than 10 and 10 (10-year recurrence-free survival rates of 85% and 63%, respectively) with one to three and more than three positive LNs in patients with breast cancer (survival rates for one to three positive LNs, 65%; survival rates for > three positive LNs, 46%; data are all from women younger than 55 years from the MMMCP project, none of whom received systemic adjuvant therapy). The survival difference of LN-negative patients with MAI 10 and LN-positive patients is negligible.
With multiple regression, MAI10 and all other non-MAI features that were significant with univariate analysis, with the exception of grade (Table 1), were included. Two models were tested (one with and the other without ER, given that ER was lacking in 43 patients) and the results were the same. With both forward and backward analysis, MAI10 was selected (P < .00001), and none of the other features had independent additive prognostic value, with the exception of nuclear atypia (P = .048). Additional analysis of many different other dichotomous MAI thresholds showed that MAI10 was the strongest prognostic threshold (in agreement with previous studies). The models were repeated with the other MAI categories (MAI < 3, 3 to 9, > 9; MAI < 5, 5 to 10, > 10). MAI < 3 v 3 to 9 had the strongest prognostic value, both for recurrence and survival (HR MAI < 3 v 10, 8.64; HR MAI < 3 v 3, 3.11). Once this additional subdivision of MAI10 was carried out, none of the other features were selected; nuclear atypia also lost its additional prognostic value. The number of patients with MAI less than 3 was clinically relevant (n = 158; 31% of all LN-negative patients younger than 55 years) and had a recurrence-free and overall survival of 91% and 96%, in contrast with 78% and 87% for MAI between 3 and 10, and 63% and 70% for MAI 10, respectively. Table 2 illustrates the prognostic significance of MAI in subgroups of patients according to tumor diameter, ER status, and the St Gallen classification. Note the consistent significant prognostic effect of MAI on those prognostically important subgroups for both recurrence and mortality.
DISCUSSION
The aim of this prospective multicenter study was to validate the prognostic value of the MAI in relation to other widely used prognosticators in a large group of LN-negative invasive breast cancer patients aged younger than 55 years. Univariate and multivariate overall survival analysis showed, in accordance with many previous smaller retrospective and prospective studies on premenopausal LN-negative breast cancers, that MAI10 is the strongest independent dichotomous prognosticator relative to other tumor and cell features evaluated. Consequently, the MAI with a cutoff value of 10 is well suited to differentiate routinely between LN-negative patients younger than 55 years with low risk (58% of the patients; 8% mortality without adjuvant treatment at 10 years follow-up) and high risk (42% of the LN-negative patients younger than 55 years; 30% mortality). The prognosis of LN-negative patients with a MAI 10 is similar to that seen in patients with one to three positive LNs not treated with systemic adjuvant therapy. The use of the MAI allows a more restricted use of adjuvant therapy in LN-negative breast cancers when compared with the St Gallen classification.
Grade 1 cancer patients have a somewhat better overall survival than patients with MAI less than 10 (94% and 91%, respectively). However, the subgroup of MAI less than 10 is much larger than that with grade 1 (299 v 161 patients). For an adequate comparison of grade 1 with MAI, one should consider using MAI less than 3. These patients were shown previously to have an excellent prognosis,11 as confirmed in the current study (Table 1), and their number is nearly the same as those with grade 1 (n = 158 v n = 160). Mortality for MAI less than 3 and grade 1 is comparable (4% and 6%, respectively; Table 1). Comparing the other grade components, tubular formation and nuclear atypia, shows that only nuclear atypia has some additional prognostic value to MAI10, but only for patients with MAI less than 10. The additional value of nuclear atypia in this subgroup with MAI less than 10 was due to the MAI divided as less than 3 v 3 to 9. When the MAI was included as such, nuclear atypia was no longer significant. These results are in agreement with recent findings of Volpi et al.32
It thus appears that the MAI is the dominant prognostic factor in breast cancer and can be used to determine accurately the outcome in LN-negative invasive breast cancer patients younger than 55 years. This also confirms results of studies showing that other proliferation markers such as percentage of Ki67, percentage of cells in S phase, or TLI are prognostically important in breast cancer.6-8 Moreover, high proliferation may also be useful for treatment selection, as it identifies those LN-negative breast cancer patients who may benefit from adjuvant chemotherapy.10 The same may hold true for trastuzumab treatment, given that HER-2 expression was not related to relapse-free survival in an overall series of breast cancers but was a significant discriminant of prognosis in the subgroup of patients with rapidly proliferating tumors.33 In breast cancers detected during population screening, interval cancers also have a high MAI.18
Not only is the MAI prognostically strong, it is also easy to assess, and if its protocol is carefully followed, MAI is well reproducible.28 This reproducibility was proven in a large, prospective, multicenter evaluation on 2,469 breast cancers, thus contrasting the more routinely used grade assessment.5 It should be emphasized that in the current study MAI was prospectively and routinely assessed during a period of more than 2 years by many different technicians using a strict MAI assessment protocol, whereas grade was the revised opinion of one single expert. This results in a negative bias for the MAI; nonetheless, MAI was still a stronger prognostic indicator than grade. It is of practical importance that the MAI can be assessed in any pathology laboratory because it requires only a conventional light microscope and a good-quality section stained with hematoxylin and eosin, and the assessment takes no more than 5 minutes per sample. Fixation delay for as long as 24 hours after extirpation of the tumor does not decrease the MAI.21
Unfortunately, some investigators do not always follow a strict mitoses counting protocol, but rather estimate the mitotic activity as described previously28 and prescribed by the WHO.23,24 They seem to believe that tubular formation and nuclear atypia have such strong prognostic value that the exact value of the MAI has little additional value, and that a rough impression of the number of mitoses gives equally strong prognostic information as the MAI. The current results show that this opinion is not correct. In contrast, an accurately assessed MAI overshadows all prognostic information contained within the other grade components. MAI protocol violations may result in too low or too high grades, and hence differences in the selection of patients for adjuvant therapy. It could also result in wrong prognostic impressions in studies evaluating prognostic factors, particularly mitotic activity.
Regarding the molecular-biologic explanation of the prognostic value of the MAI in LN-negative breast cancer, a specific chromosomal aberration (1p31 loss) was strongly correlated with high MAI values, as a recent study has shown.34 A candidate tumor suppressor gene is ARHI/Noey2 on 1p31,35 a member of the Ras super family of small G proteins that is frequently inactivated in ovarian and breast tumors. Loss of heterozygosity studies36 found that ARHI/Noey2 is the center of allelic deletion within chromosome 1p31. Loss of Noey2 expression has been found to correlate with a high Ki67 proliferation index in a small group of ductal carcinoma in situ patients.37 Another interesting aberration is gains found at 3q26, which were strong predictors of clinical outcome.30 Genes located in this region include the human telomerase RNA component38 and PIK3CA39 (the p110alpha catalytic subunit of a PI3-kinase class 1a). In breast cancer, telomerase positivity is associated with lymphovascular invasion and decreased disease-free survival.40 PIK3CA is an important component for phosphoinositide 3-kinase (PI3-kinase), a growth factor-activated transforming lipid (and protein) kinase. PI3-kinases of the 1a class have multiple effectors and are involved in cell motility and invasion. The protein of the suppressor gene PTEN, phosphoinositide 3-phosphatase, is an antagonist of the PI3-kinases and loss of PTEN (like gain of PIK3CA) will thus induce the effectors of PI3-kinase. A loss of 10q or a gain of 3q was found in 11 of 15 ductal carcinomas with recurrences, which could mean that PI3-kinase and its effectors (PKB/AKT and others) play a major role in the process of cancer recurrence.34
A recent gene-expression signature study accurately predicted survival in LN-negative breast cancer, with overall 10-year survival in the good- and poor-prognosis signature patients of 97% and 50%,41 respectively, compared with overall 10-year survival in patients with MAI10 of 92% and 70%, respectively). For recurrence, these figures were 87% and 44%, respectively (compared with recurrence in patients with MAI10 of 85% and 63%, respectively). The discrimination is therefore better than with the MAI but the difference is mainly in the unfavorable patients; the MAI does not require fresh frozen tissue, can be done on sections stained with standard hematoxylin and eosin, is widely available, and is much less expensive. It would be interesting to study the two methods in parallel on the same material.
In conclusion, the prognostic strength, wide applicability, reproducibility, simplicity, low cost, and the favorable balance between expected lives saved and patients overtreated argue for a more intensified application of the MAI in the selection of AST in LN-negative breast cancers.
Appendix
The following specialists, in alphabetical order of hospital, institute, or organization, participated in this study: Antoni van Leeuwenhoekhuis, Amsterdam, the Netherlands: J.A. van Dongen, J.L. Peterse. Comprehensive Cancer Center Amsterdam: J.T. Benraadt. Onze Lieve Vrouwe Gasthuis, Amsterdam: J. van de Haar, M.E.I. Schipper, W.F.H.L. Breuker, K.J. Roozendaal. Academisch Ziekenhuis VU Medisch Centrum, Amsterdam: H.M. Pinedo, S. Meyer, J. Th. Plukker, P.D. de Rooy, R. Bleichroth. St Lucas Andreas Ziekenhuis, Amsterdam: S.M. Bellot, G. Groot. Stichting Laboratorium Pathologie Oost Nederland (Enschede Pathologisch Anatomisch Laboratorium), Enschede: J.G. Somsen, F. de La Fuente. St Antonius Ziekenhuis, Nieuwegein: L.H.M. van Gorp. Diakonessen Ziekenhuis, Utrecht: H.M. Ruitenberg. St Ignatius Ziekenhuis, Breda: J. Los, M. Snieders, J. Peters. Stichting Pathol. Anat. en Cytol. Lab. West Brabant, Bergen op Zoom: J. Fijnheer, D.E. Isra?l, A.G.M.J. Elias. Pathologisch Anatomisch Laboratorium Stichting de PAMM, Eindhoven: M.W.P.M. van Beek. St Laurentius Ziekenhuis, Pathologisch en Anatomisch Lab., Roermond, St Jans Gasthuis, Weert: W.S. Kwee, A.P. Willig. St Maartens Gasthuis, Pathologisch Anatomisch Lab., Venlo: R.F.M. Schapers. Maasland Ziekenhuis, Pathologisch Anatomisch Lab., Sittard: A. Th. Ariens. Pathologisch Anat. Lab., Martini Ziekenhuis (RK. Ziekenverpleging "Onze Lieve Vrouw Behoudenis der Kranken"), Groningen: E.J. Ebels, H. Dop. Medisch Spectrum Twente (Ziekenhuis De Stadsmaten), Enschede: P.W.A.A. Koch, M.C.O. Fehmers. Medisch Spectrum Twente (Ziekenhuis Ziekenzorg), Enschede: H.J. Mulders, J.H. van Lijf, M. van Reyn. Stichting StreekZiekenhuizen Midden-Twente Hengelo OV (Koningin Juliana Ziekenhuis, Stichting StreekZiekenhuizen Midden Twente, Gerardus Majella Ziekenhuis), Hengelo: B.A. van Driel, H. Dankbaar. Twenteborg Ziekenhuis (Stichting Streekziekenhuis), Almelo: D.A. Buys. Stichting Ziekenhuisvoorzieningen Oostachterhoek, Streekziekenhuis, Winterswijk: A.A. Fierstra. Medisch Spectrum Twente, Oldenzaal: T.L. van de Sluis. St Antonius Ziekenhuis, Nieuwegein: R. Sybrandy. Ziekenhuis Overvecht, Utrecht: J.M. Wakelkamp. Medisch Centrum Berg en Bosch, Bilthoven: S.H.M. Vrind. Hofpoort Ziekenhuis, Woerden: W. van Eesteren, W. Eijlers, J.P. Vente. Lorentz Ziekenhuis, Zeist: W.J. van de Ven, P.D. Warners. Stichting Inrichting van Diakonessen, Utrecht: H. Nortier, P. Leguit. Ziekenhuis Oudenrijn, Utrecht: W. J. Geurts. St Ignatius Ziekenhuis, Breda: E.J. van Herk, R. Krause, H.M. Mendelaar, A.C.J.M. Holdrinet. Interconfessioneel Ziekenhuis, De Baronie Diaconessenziekenhuis, Breda: F.J.M. van Geloven, F.J. van Ijzeren, A.M. van Gent. St Laurens Ziekenhuis, Breda: G.A.M. Kokke, P.J. Stijnen. St Ziekenhuis Lievensberg, Bergen op Zoom: E.J.R. Slingenberg, P.M.R. Versteegh, F. Vaas. Ziekenhuis St. Franciscus, Roozendaal: J.L.M. Nooren, C.P. de Kraker, E.H.M. Kamphuys. Mammawerkgroep IKZ/SOOZ: H.F. P. Hillen. Catharina Ziekenhuis, Eindhoven. St Joseph Ziekenhuis, Eindhoven. Diaconnessenhuis, Eindhoven. St Anna Ziekenhuis, Geldrop. St Laurentius Ziekenhuis, Roermond: A.P. M Rutten, P.J.H. Sikkenk, C.M. Nuyens, R.J.A. Estourgie, J.A.J.M. Wils. St. Jans Gasthuis, Weert: W.A. van Deurzen, F.M. Lasisang, H.A.J.M. Scheerder. St Maartens Gasthuis, Venlo: F.G.J. Laudy, H.A.J.M. Scheerder. St. Elisabeth Ziekenhuis, Venray: A.L. Eekhout. Stichting Ziekenzorg Westelijke Mijnstreek, locatie Sittard: H.J.J. Stroehen. Stichting Ziekenzorg Westelijke Mijnstreek, locatie Geleen: F.W.C. van de Ent.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank J. Littooij, L. Schuurmans, M. Broeckaert, E. Matze-Cok, H. de Zeeuw, E. Wisse-Brekelmans, J. van Eijk, C. van Galen, J. Konneman, J. Brugghe, B. van Diermen, and S.M. Allen-de Jong for their skilled technical assistance.
NOTES
Supported by grant No. 28-1398 of the National Health and Research Council of The Netherlands, Zorg Onderzoek Nederland Medische Wetenschappen (ZonMw), and grant Nos. 95-112 and 98-136 of the Stichting Bevordering Diagnostiche Morfometrie (SBDM).
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Early Breast Cancer Trialists' Collaborative Group: Tamoxifen for early breast cancer: An overview of the randomised trials. Lancet 351:1451-1467, 1998
Early Breast Cancer Trialists' Collaborative Group: Polychemotherapy for early breast cancer: An overview of the randomised trials. Lancet 352:930-942, 1998
Goldhirsch A, Wood WC, Gelber RD, et al: Meeting highlights: Updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol 21:3357-3365, 2003
Boyages J, Chua B, Taylor R, et al: Use of the St Gallen classification for patients with node-negative breast cancer may lead to overuse of adjuvant chemotherapy. Br J Surg 89:789-796, 2002
European Commission Group on Breast Screening Pathology: Consistency achieved by 23 European pathologists from 12 countries in diagnosing breast disease and reporting prognostic features of carcinomas. Virchows Arch 434:3-10, 1999
Meyer JS, Friedman E, McCrate MM, et al: Prediction of early course of breast carcinoma by thymidine labeling. Cancer 51:1879-1886, 1983
Silvestrini R, Daidone MG, Luisi A, et al: Cell proliferation in 3,800 node-negative breast cancers: Consistency over time of biological and clinical information provided by 3H-thymidine labelling index. Int J Cancer 74:122-127, 1997
Sigurdsson H, Baldetorp B, Borg A, et al: Indicators of prognosis in node-negative breast cancer. N Engl J Med 322:1045-1053, 1990
Bergers E, van Diest PJ, Baak JPA: Prognostic implications of different cell cycle analysis models of flow cytometric DNA histograms of 1301 breast cancer patients: Results from the Multicenter Morphometric Mammary Carcinoma Project. Int J Cancer 74:260-269, 1997
Amadori D, Nanni O, Marangolo M, et al: Disease-free survival advantage of adjuvant cyclophosphamide, methotrexate, and fluorouracil in patients with node-negative, rapidly proliferating breast cancer: A randomized multicenter study. J Clin Oncol 18:3125-3134, 2000
Baak JPA, van Dop H, Kurver PHJ, et al: The value of morphometry to classic prognosticators in breast cancer. Cancer 56:374-382, 1985
Tosi P, Luzi P, Sforza V, et al: Correlation between morphometrical parameters and disease-free survival in ductal breast cancer treated only by surgery. Appl Pathol 4:33-42, 1986
van der Linden JC, Baak JP, Lindeman J, et al: Prospective evaluation of prognostic value of morphometry in patients with primary breast cancer. J Clin Pathol 40:302-306, 1987
Uyterlinde AM, Baak JPA, Schipper NW, et al: Prognostic value of morphometric and DNA flow-cytometry features of invasive breast cancers detected by population screening: Comparison with control group of hospital patients. Int J Cancer 48:173-181, 1991
Clayton F: Pathologic correlates of survival in 378 lymph node-negative infiltrating ductal breast carcinomas: Mitotic count is the best single predictor. Cancer 68:1309-1317, 1991
Makkink-Nombrado SV: Quantitative Pathological, Cell and Molecular Biological Markers in Breast Cancer. Amsterdam, the Netherlands, Free University Press, 1995, pp 149-163
Thor AD, Liu S, Moore DH II, et al: Comparison of mitotic index, in vitro bromodeoxyuridine labeling, and MIB-1 assays to quantitate proliferation in breast cancer. J Clin Oncol 17:470-477, 1999
Groenendijk RP, Bult P, Noppen CM, et al: Mitotic activity index in interval breast cancers. Eur J Surg Oncol 29:29-31, 2003
Manders P, Bult P, Sweep CG, et al: The prognostic value of the mitotic activity index in patients with primary breast cancer who were not treated with adjuvant systemic therapy. Breast Cancer Res Treat 77:77-84, 2003
Donhuijsen K, Schmidt U, Hirsche H, et al: Changes in mitotic rate and cell cycle fractions caused by delay in fixation. Hum Pathol 21:709-714, 1990
Bergers E, Jannink I, van Diest PJ, et al: Influence of fixation delay on mitotic activity and flow cytometric %S-phase. Hum Pathol 28:95-100, 1997
World Health Organisation.Histologic Typing of Breast Tumors (ed 2). Geneva, Switzerland, World Health Organisation, 1981
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
Ellis IO, Schnitt SJ, Sastre-Garau X, et al: Invasive breast carcinoma, in Tavassoli FA, Devilee P (eds): Pathology and Genetics of Tumours of the Breast and Female Genital Tract Organs. IARC Press, Lyon, France, 2003, pp 18-19
Hall PA, Going JJ: Predicting the future: A critical appraisal of cancer prognosis studies. Histopathology 35:489-494, 1999
Baak JPA: The framework of pathology: Good laboratory practice by quantitative and molecular methods. J Pathol 198:277-283, 2002
Baak JPA, van Diest PJ, Ariens A, et al: The multicenter morphometric mammary carcinoma project (MMMMCP): A nationwide prospective study on reproducibility and prognostic power of routine quantitative assessments in The Netherlands. Pathol Res Pract 185:664-670, 1989
van Diest PJ, Baak JPA, Matze-Cok P, et al: Reproducibility of mitosis counting in 2469 breast cancer specimens: Results from the Multicenter Morphometric Mammary Carcinoma Project. Hum Pathol 23:603-607, 1992
Jannink I, Diest PJ van, Baak JPA: Comparison of the prognostic value of mitotic frequency and mitotic activity index in breast cancer. Breast 5:31-36, 1996
Kaplan EL, Meier P: Long parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Cox DR: Regression models and life tables. J R Stat Soc B 34:187-220, 1972
Volpi A, Bacci F, Paradiso A, et al: Prognostic relevance of histological grade and its components in node-negative breast cancer patients. Mod Pathol 17:1038-1044, 2004
Volpi A, Nanni O, De Paola F, et al: HER-2 expression and cell proliferation: Prognostic markers in patients with node-negative breast cancer. J Clin Oncol 21:2708-2712, 2003
Janssen EA, Baak JPA, Guervos MA, et al: In lymph node-negative invasive breast carcinomas, specific chromosomal aberrations are strongly associated with mitotic activity and predict outcome more accurately than, grade, tumour diameter and oestrogen receptor. J Pathol 201:555-561, 2003
Yu Y, Xu F, Peng H, et al: Noey2 (ARH1), an imprinted putative tumour suppressor gene in ovarian and breast carcinomas. Proc Natl Acad Sci U S A 96:214-219, 1999
Hisatomi H, Nagao K, Wakita K, et al: ARH1/NOEY2 inactivations maybe important in breast tumour pathogenesis. Oncology 62:136-140, 2002
Wang L, Hoque A, Luo RZ, et al: Loss of the expression of the tumor suppressor gene ARHI in associated with progression of breast cancer. Clin Cancer Res 9:3660-3666, 2003
Soder AI, Hoare SF, Muir S, et al: Amplification, increased dosage and in situ expression of the telomerase RNA gene in human cancer. Oncogene 14:1013-1021, 1997
Ma YY, Wei SJ, Lin YC, et al: PIK3CA as an oncogene in cervical cancer. Oncogene 19:2739-2744, 2000
Mokbel KM, Parris CN, Ghilchilk M, et al: Telomerase activity and lymphovascular invasion in breast cancer. Eur J Surg Oncol 26:30-33, 2000
van de Vijver MJ, He YD, van't Veer LJ, et al: A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999-2009, 2002(Jan P.A. Baak, Paul J. va)
University Medical Center Nijmegen, Nijmegen, the Netherlands
Departments of Pathology, Stavanger University Hospital, Stavanger
the Gade Institute, University of Bergen, Norway
University Hospital, Antwerpen, Belgium
ABSTRACT
PURPOSE: To validate the independent strong prognostic value of mitotic activity index (MAI) in lymph node (LN) –negative invasive breast cancer patients younger than 55 years in a nationwide multicenter prospective study.
PATIENTS AND METHODS: Analysis of routinely assessed MAI and other prognosticators in 516 patients (median follow-up, 118 months; range, 8 to 185 months), without systemic adjuvant therapy or previous malignancies.
RESULTS: Distant metastases occurred in 127 patients (24.6%); 90 (17.4%) died as a result of metastases. MAI (< 10, 10) showed strong association with recurrence (hazard ratio [HR], 3.12; 95% CI, 2.17 to 4.50; P .0001) and mortality (HR, 4.42; 95% CI, 2.79 to 7.01; P < .0001). The absolute difference in 10-year Kaplan-Meier estimates of time to distant recurrence as well as survival was 22% between MAI less than 10 versus 10. This effect was independent of age, estrogen receptor (ER) status, and tumor diameter (which were significant prognosticators). In multivariate analysis with regard to patient age, tumor diameter, grade, ER status, and the St Gallen criterion, MAI proved to be an independent and the strongest prognosticator. Tubular formation (TF) and nuclear atypia (NA), as constituents of (expert revised) grade, had no (for TF) or limited (for NA, P = .048) additional prognostic value to the MAI. In the group with MAI less than 10, MAI less than 3 versus more than 3 had additional value but the classical threshold of 0 to 5 v 6 to 10 did not. With this additional subdivision of MAI as less than 3, 3 to 9, and more than 9, NA lost its additive prognostic value.
CONCLUSION: The MAI is the strongest, most widely available, easily assessable, inexpensive, well-reproducible prognosticator and is well suited to routinely differentiate between high- and low-risk LN-negative breast cancer patients younger than 55 years.
INTRODUCTION
Trials have shown that both prolonged hormonal and adjuvant systemic treatment (AST) can improve prognosis in subsets of lymph node (LN) –negative breast cancers. The absolute benefits of systemic adjuvant chemotherapy were most apparent in younger women and in women with a worse prognosis, whereas tamoxifen appeared to be effective irrespective of age and menopausal status.1,2 Clearly, in patients with a better prognosis, the absolute improvements in recurrence and survival will be less substantial and the benefits need to be weighed against the adverse effects of therapy. Any easily assessable, reproducible, and reliable prognostic marker that will add information about primary tumor characteristics in relation to the follow-up (in addition to the well-established prognosticators such as nodal status, tumor size, and age) would be useful for this purpose. Predictive factors such as estrogen receptor (ER) status and HER-2 status are well established for treatment decision making.
The St Gallen classification3 recommends AST for all LN-negative premenopausal patients except those at minimal/low risk, which was defined as having postsurgical size of the tumor 2 cm, grade 1, ER-positive and/or progesterone receptor (PR) –positive status, and age older than 35 years. However, general application of this guideline means that approximately 85% of all premenopausal LN-negative patients would be treated with AST, whereas only a relatively small number would develop distant metastases without AST. It has also been argued that the St Gallen classification is an inaccurate measure of prognosis for patients with LN-negative breast cancer and should be used with caution.4 Moreover, the reproducibility of grade, an important discriminator in LN-negative patients in this risk classification, is far from perfect (low value for grade 1 v 2 among different observers).5
Proliferation variables such as the tritiated thymidine labeling index (TLI)6,7 and percentage of S-phase cells assessed by flow cytometry8,9 are good prognostic factors. In one study, patients with LN-negative, rapidly proliferating tumors, defined according to TLI, significantly benefited from systemic adjuvant chemotherapy.10 However, TLI is technically cumbersome and not widely available. One of the strongest yet simple and well-reproducible proliferation-associated prognostic factors is the mitotic activity index (MAI). Its prognostic value has been shown in many retrospective and prospective studies using a fixed threshold (MAI < 10, favorable; MAI 10, unfavorable).11-19 The MAI is not sensitive to fixation delay20,21 and is already part of different grading systems.22-24 Despite proof that the MAI is an independent strong prognostic factor in many previous studies, a routine laboratory test for therapeutic decision requires evaluation in a prospective multicenter study.25,26 Often, such validations are lacking because they are expensive, time consuming, and sometimes regarded as unoriginal. To validate the MAI with respect to its reproducibility, accessibility, and prospective prognostic power in relation to other classical prognostic factors, the nationwide Multicenter Mammary Carcinoma Project (MMMCP) was set up in the Netherlands in 1987, including 34 hospitals.27 A total of 3,479 consecutive primary breast cancer patients were enrolled from October 1, 1987, through December 31, 1989. It has been reported previously that the MAI was well reproducible among the collaborating laboratories.28
This report describes the long-term prognostic value of different features in the 516 women with LN-negative invasive breast cancer who were younger than 55 years of age and without previous malignancies. In accordance with the general policy during the enrollment period in the Netherlands, none of the patients were receiving or had received adjuvant systemic treatment.
PATIENTS AND METHODS
Patients
Details of the MMMCP protocol have been described elsewhere.27 The essential details will be presented here. After pathologists and technicians received training in mitosis counting using the MMMCP protocol and reproducibility was proven among the participants,28 all consecutive primary breast cancer patients diagnosed in the collaborating hospitals from October 1, 1987, through January 1, 1990, were enrolled. Follow-up was updated annually and data collection closed in 2003. Of the 3,479 MMMCP patients registered, 603 patients (17.3%) were females younger than 55 years of age with LN-negative breast cancer. Of these, 14 had carcinoma-in-situ and 31 had a carcinoma-in-situ with a microinvasive component less than 1 mm, which is ineligible for mitoses counting; these patients were not considered for the study, leaving 558 patients. Of these, 27 had a history of previous malignancies (breast, n = 21; other organs, n = 6), four had poor-quality sections preventing adequate MAI assessment, and 11 patients were lost to follow-up, leaving 516 patients for further analysis. In total, 591 LN-positive patients younger than 55 years were recruited in the MMMCP.
All patients were treated with modified radical mastectomy (n = 174) or breast-conserving therapy (n = 342); adequate axillary LN dissection was always performed. Locoregional radiotherapy was given in patients who underwent breast-conserving therapy or had medially localized tumors. None of the LN-negative patients received any form of adjuvant systemic therapy (in agreement with the usual therapeutic regimens of many hospitals in the Netherlands in the 1980s). The survival results of the LN-negative patients with MAI less than 10 and 10 were compared with those of 451 patients younger than 55 years with LN-positive MMMCP. These 451 patients, a subset of all 591 LN-positive patients younger than 55 years, had not received any form of adjuvant systemic therapy (patients, including most LN-positive patients, did not commonly receive adjuvant chemotherapy in many hospitals in the Netherlands in the 1980s).
Postsurgical size of the tumor was measured in the fresh specimens; the tumors were cut in slices 0.5 cm thick, fixed in buffered 4% formaldehyde, and embedded in paraffin. At least 10 (median, 14) LNs were detected in the axillary LN dissection specimens. Paraffin sections 4 μm thick were cut and stained with hematoxylin and eosin. Histologic type was assessed afterward according to the WHO criteria.22 Grade was assessed during careful review by two pathologists with considerable experience in breast pathology, according to the Nottingham modification,23,24 using MAI 0 to 5 = 1, 6 to 10 = 2, and more than 10 = 3; nuclear atypia as mild = 1, moderate = 2, and marked = 3; and tubular formation as majority (> 75%) = 1, moderate (10% to 75%) = 2, and little or none (< 10%) = 3. The grade is the sum of tubular formation + nuclear atypia + MAI class, where the sum of 3, 4, or 5 is grade 1; the sum of 6 or 7 is grade 2; and the sum of 8 or 9 is grade 3. ER value was assessed in reference laboratories with the ligand-binding charcoal technique (cutoff, 10 fmol/mg protein).
The MAI was assessed in each of the participating laboratories by a total of 27 different technicians who were instructed at four training sessions held during the first 3 months. For the assessment of the MAI, the MMMCP protocol27,28 was used, which is as follows.
First, with a black marker on the glass slide, the most poorly differentiated peripheral area of the tumor is demarcated. Necrotic, heavily inflamed, or benign areas are avoided. This area, called the measurement area, is minimally 1 x 1 mm and maximally 5 x 5 mm. Second, in the measurement area, at x400 magnification (objective 40, field diameter 450 μm at the specimen level), mitoses are counted in 10 consecutive neighboring fields of vision in the most cellular area (representing a total area of 1.59 mm2 at the specimen level). Third, only certain mitoses are counted; doubtful structures and apoptotic bodies are ignored. The total resulting number of well-defined mitotic figures counted in the 10 fields of vision is the MAI. Fourth, if the MAI is between 5 and 15, it is assessed once more, and the highest number of the two counts is taken as the MAI.
An accurate MAI assessment takes 3 to 5 minutes. Correction of the MAI for the percentage of tissue occupied by stroma or the number of tumor cells was not applied because it was shown previously that this does not substantially improve the prognostic value of the MAI and is substantially time consuming.29 The MAI is a continuous variable. According to many previous studies, the most important prognostic threshold is 10, with MAI less than 10 indicating favorable prognosis and MAI 10 indicating poor prognosis11-19,29 (this threshold is denoted as MAI10), whereas patients with MAI less than 3 have an especially favorable prognosis.11
Statistical Analysis
Main end points were recurrence and mortality. In analyzing the probability that patients would remain free of distant metastases, we defined recurrence as any first recurrence at distant sites. All other patients were censored on the date of the last follow-up visit; deaths as a result of causes other than breast cancer, local or regional recurrences, or the development of a secondary primary cancer (including contralateral breast cancer) were also censored. Mortality was defined as any death as a result of distant metastases (as evident from clinical, radiologic, histologic, or autopsy data; no patients died as a result of locoregional disease). If the cause of death was unknown, but a metastasis was detected previously, then death was considered to related breast cancer unless explicitly stated otherwise (in line with other studies).1 If the status during follow-up indicated a confirmed metastasis without a date of recurrence, the date of that follow-up visit was used.
Age, time to first recurrence, and survival time were calculated relative to the date of primary diagnosis. Median follow-up of the 516 patients was 118 months (range, 8 to 185 months). Distant metastases occurred in 127 patients (24.6%) and 90 patients (17.4%) died as a result of disease. Survival curves were constructed using the Kaplan-Meier technique.30 Differences between groups were tested by log-rank tests or tests for trend. The relative importance of potential prognostic variables was tested using Cox proportional hazards analysis and expressed as hazard ratios (HRs) with 95% CIs.31 All variables were tested for proportionality and continuous variables were checked for linearity or nonlinearity and transformed or recoded if necessary (or useful).
RESULTS
Table 1 shows the 10-year recurrence and disease-related mortality estimates and HRs for the most relevant patient and tumor characteristics. The MAI showed a prognostically optimal cutoff value at 10 and tumor size at 2 cm.
Grade was a strong prognostic factor, both for recurrence and mortality (Figs 1A and 1B). When the relation between the different grade components (tubulus formation, nuclear atypia, and MAI) and survival was studied, the difference in survival between favorable and unfavorable values of any of these features was largest with MAI (92% for MAI < 10 v 70% if MAI 10, a difference of 22%; Figs 2A and 2B). Figure 3 compares the survival curves of LN-negative MAI less than 10 and 10 (10-year recurrence-free survival rates of 85% and 63%, respectively) with one to three and more than three positive LNs in patients with breast cancer (survival rates for one to three positive LNs, 65%; survival rates for > three positive LNs, 46%; data are all from women younger than 55 years from the MMMCP project, none of whom received systemic adjuvant therapy). The survival difference of LN-negative patients with MAI 10 and LN-positive patients is negligible.
With multiple regression, MAI10 and all other non-MAI features that were significant with univariate analysis, with the exception of grade (Table 1), were included. Two models were tested (one with and the other without ER, given that ER was lacking in 43 patients) and the results were the same. With both forward and backward analysis, MAI10 was selected (P < .00001), and none of the other features had independent additive prognostic value, with the exception of nuclear atypia (P = .048). Additional analysis of many different other dichotomous MAI thresholds showed that MAI10 was the strongest prognostic threshold (in agreement with previous studies). The models were repeated with the other MAI categories (MAI < 3, 3 to 9, > 9; MAI < 5, 5 to 10, > 10). MAI < 3 v 3 to 9 had the strongest prognostic value, both for recurrence and survival (HR MAI < 3 v 10, 8.64; HR MAI < 3 v 3, 3.11). Once this additional subdivision of MAI10 was carried out, none of the other features were selected; nuclear atypia also lost its additional prognostic value. The number of patients with MAI less than 3 was clinically relevant (n = 158; 31% of all LN-negative patients younger than 55 years) and had a recurrence-free and overall survival of 91% and 96%, in contrast with 78% and 87% for MAI between 3 and 10, and 63% and 70% for MAI 10, respectively. Table 2 illustrates the prognostic significance of MAI in subgroups of patients according to tumor diameter, ER status, and the St Gallen classification. Note the consistent significant prognostic effect of MAI on those prognostically important subgroups for both recurrence and mortality.
DISCUSSION
The aim of this prospective multicenter study was to validate the prognostic value of the MAI in relation to other widely used prognosticators in a large group of LN-negative invasive breast cancer patients aged younger than 55 years. Univariate and multivariate overall survival analysis showed, in accordance with many previous smaller retrospective and prospective studies on premenopausal LN-negative breast cancers, that MAI10 is the strongest independent dichotomous prognosticator relative to other tumor and cell features evaluated. Consequently, the MAI with a cutoff value of 10 is well suited to differentiate routinely between LN-negative patients younger than 55 years with low risk (58% of the patients; 8% mortality without adjuvant treatment at 10 years follow-up) and high risk (42% of the LN-negative patients younger than 55 years; 30% mortality). The prognosis of LN-negative patients with a MAI 10 is similar to that seen in patients with one to three positive LNs not treated with systemic adjuvant therapy. The use of the MAI allows a more restricted use of adjuvant therapy in LN-negative breast cancers when compared with the St Gallen classification.
Grade 1 cancer patients have a somewhat better overall survival than patients with MAI less than 10 (94% and 91%, respectively). However, the subgroup of MAI less than 10 is much larger than that with grade 1 (299 v 161 patients). For an adequate comparison of grade 1 with MAI, one should consider using MAI less than 3. These patients were shown previously to have an excellent prognosis,11 as confirmed in the current study (Table 1), and their number is nearly the same as those with grade 1 (n = 158 v n = 160). Mortality for MAI less than 3 and grade 1 is comparable (4% and 6%, respectively; Table 1). Comparing the other grade components, tubular formation and nuclear atypia, shows that only nuclear atypia has some additional prognostic value to MAI10, but only for patients with MAI less than 10. The additional value of nuclear atypia in this subgroup with MAI less than 10 was due to the MAI divided as less than 3 v 3 to 9. When the MAI was included as such, nuclear atypia was no longer significant. These results are in agreement with recent findings of Volpi et al.32
It thus appears that the MAI is the dominant prognostic factor in breast cancer and can be used to determine accurately the outcome in LN-negative invasive breast cancer patients younger than 55 years. This also confirms results of studies showing that other proliferation markers such as percentage of Ki67, percentage of cells in S phase, or TLI are prognostically important in breast cancer.6-8 Moreover, high proliferation may also be useful for treatment selection, as it identifies those LN-negative breast cancer patients who may benefit from adjuvant chemotherapy.10 The same may hold true for trastuzumab treatment, given that HER-2 expression was not related to relapse-free survival in an overall series of breast cancers but was a significant discriminant of prognosis in the subgroup of patients with rapidly proliferating tumors.33 In breast cancers detected during population screening, interval cancers also have a high MAI.18
Not only is the MAI prognostically strong, it is also easy to assess, and if its protocol is carefully followed, MAI is well reproducible.28 This reproducibility was proven in a large, prospective, multicenter evaluation on 2,469 breast cancers, thus contrasting the more routinely used grade assessment.5 It should be emphasized that in the current study MAI was prospectively and routinely assessed during a period of more than 2 years by many different technicians using a strict MAI assessment protocol, whereas grade was the revised opinion of one single expert. This results in a negative bias for the MAI; nonetheless, MAI was still a stronger prognostic indicator than grade. It is of practical importance that the MAI can be assessed in any pathology laboratory because it requires only a conventional light microscope and a good-quality section stained with hematoxylin and eosin, and the assessment takes no more than 5 minutes per sample. Fixation delay for as long as 24 hours after extirpation of the tumor does not decrease the MAI.21
Unfortunately, some investigators do not always follow a strict mitoses counting protocol, but rather estimate the mitotic activity as described previously28 and prescribed by the WHO.23,24 They seem to believe that tubular formation and nuclear atypia have such strong prognostic value that the exact value of the MAI has little additional value, and that a rough impression of the number of mitoses gives equally strong prognostic information as the MAI. The current results show that this opinion is not correct. In contrast, an accurately assessed MAI overshadows all prognostic information contained within the other grade components. MAI protocol violations may result in too low or too high grades, and hence differences in the selection of patients for adjuvant therapy. It could also result in wrong prognostic impressions in studies evaluating prognostic factors, particularly mitotic activity.
Regarding the molecular-biologic explanation of the prognostic value of the MAI in LN-negative breast cancer, a specific chromosomal aberration (1p31 loss) was strongly correlated with high MAI values, as a recent study has shown.34 A candidate tumor suppressor gene is ARHI/Noey2 on 1p31,35 a member of the Ras super family of small G proteins that is frequently inactivated in ovarian and breast tumors. Loss of heterozygosity studies36 found that ARHI/Noey2 is the center of allelic deletion within chromosome 1p31. Loss of Noey2 expression has been found to correlate with a high Ki67 proliferation index in a small group of ductal carcinoma in situ patients.37 Another interesting aberration is gains found at 3q26, which were strong predictors of clinical outcome.30 Genes located in this region include the human telomerase RNA component38 and PIK3CA39 (the p110alpha catalytic subunit of a PI3-kinase class 1a). In breast cancer, telomerase positivity is associated with lymphovascular invasion and decreased disease-free survival.40 PIK3CA is an important component for phosphoinositide 3-kinase (PI3-kinase), a growth factor-activated transforming lipid (and protein) kinase. PI3-kinases of the 1a class have multiple effectors and are involved in cell motility and invasion. The protein of the suppressor gene PTEN, phosphoinositide 3-phosphatase, is an antagonist of the PI3-kinases and loss of PTEN (like gain of PIK3CA) will thus induce the effectors of PI3-kinase. A loss of 10q or a gain of 3q was found in 11 of 15 ductal carcinomas with recurrences, which could mean that PI3-kinase and its effectors (PKB/AKT and others) play a major role in the process of cancer recurrence.34
A recent gene-expression signature study accurately predicted survival in LN-negative breast cancer, with overall 10-year survival in the good- and poor-prognosis signature patients of 97% and 50%,41 respectively, compared with overall 10-year survival in patients with MAI10 of 92% and 70%, respectively). For recurrence, these figures were 87% and 44%, respectively (compared with recurrence in patients with MAI10 of 85% and 63%, respectively). The discrimination is therefore better than with the MAI but the difference is mainly in the unfavorable patients; the MAI does not require fresh frozen tissue, can be done on sections stained with standard hematoxylin and eosin, is widely available, and is much less expensive. It would be interesting to study the two methods in parallel on the same material.
In conclusion, the prognostic strength, wide applicability, reproducibility, simplicity, low cost, and the favorable balance between expected lives saved and patients overtreated argue for a more intensified application of the MAI in the selection of AST in LN-negative breast cancers.
Appendix
The following specialists, in alphabetical order of hospital, institute, or organization, participated in this study: Antoni van Leeuwenhoekhuis, Amsterdam, the Netherlands: J.A. van Dongen, J.L. Peterse. Comprehensive Cancer Center Amsterdam: J.T. Benraadt. Onze Lieve Vrouwe Gasthuis, Amsterdam: J. van de Haar, M.E.I. Schipper, W.F.H.L. Breuker, K.J. Roozendaal. Academisch Ziekenhuis VU Medisch Centrum, Amsterdam: H.M. Pinedo, S. Meyer, J. Th. Plukker, P.D. de Rooy, R. Bleichroth. St Lucas Andreas Ziekenhuis, Amsterdam: S.M. Bellot, G. Groot. Stichting Laboratorium Pathologie Oost Nederland (Enschede Pathologisch Anatomisch Laboratorium), Enschede: J.G. Somsen, F. de La Fuente. St Antonius Ziekenhuis, Nieuwegein: L.H.M. van Gorp. Diakonessen Ziekenhuis, Utrecht: H.M. Ruitenberg. St Ignatius Ziekenhuis, Breda: J. Los, M. Snieders, J. Peters. Stichting Pathol. Anat. en Cytol. Lab. West Brabant, Bergen op Zoom: J. Fijnheer, D.E. Isra?l, A.G.M.J. Elias. Pathologisch Anatomisch Laboratorium Stichting de PAMM, Eindhoven: M.W.P.M. van Beek. St Laurentius Ziekenhuis, Pathologisch en Anatomisch Lab., Roermond, St Jans Gasthuis, Weert: W.S. Kwee, A.P. Willig. St Maartens Gasthuis, Pathologisch Anatomisch Lab., Venlo: R.F.M. Schapers. Maasland Ziekenhuis, Pathologisch Anatomisch Lab., Sittard: A. Th. Ariens. Pathologisch Anat. Lab., Martini Ziekenhuis (RK. Ziekenverpleging "Onze Lieve Vrouw Behoudenis der Kranken"), Groningen: E.J. Ebels, H. Dop. Medisch Spectrum Twente (Ziekenhuis De Stadsmaten), Enschede: P.W.A.A. Koch, M.C.O. Fehmers. Medisch Spectrum Twente (Ziekenhuis Ziekenzorg), Enschede: H.J. Mulders, J.H. van Lijf, M. van Reyn. Stichting StreekZiekenhuizen Midden-Twente Hengelo OV (Koningin Juliana Ziekenhuis, Stichting StreekZiekenhuizen Midden Twente, Gerardus Majella Ziekenhuis), Hengelo: B.A. van Driel, H. Dankbaar. Twenteborg Ziekenhuis (Stichting Streekziekenhuis), Almelo: D.A. Buys. Stichting Ziekenhuisvoorzieningen Oostachterhoek, Streekziekenhuis, Winterswijk: A.A. Fierstra. Medisch Spectrum Twente, Oldenzaal: T.L. van de Sluis. St Antonius Ziekenhuis, Nieuwegein: R. Sybrandy. Ziekenhuis Overvecht, Utrecht: J.M. Wakelkamp. Medisch Centrum Berg en Bosch, Bilthoven: S.H.M. Vrind. Hofpoort Ziekenhuis, Woerden: W. van Eesteren, W. Eijlers, J.P. Vente. Lorentz Ziekenhuis, Zeist: W.J. van de Ven, P.D. Warners. Stichting Inrichting van Diakonessen, Utrecht: H. Nortier, P. Leguit. Ziekenhuis Oudenrijn, Utrecht: W. J. Geurts. St Ignatius Ziekenhuis, Breda: E.J. van Herk, R. Krause, H.M. Mendelaar, A.C.J.M. Holdrinet. Interconfessioneel Ziekenhuis, De Baronie Diaconessenziekenhuis, Breda: F.J.M. van Geloven, F.J. van Ijzeren, A.M. van Gent. St Laurens Ziekenhuis, Breda: G.A.M. Kokke, P.J. Stijnen. St Ziekenhuis Lievensberg, Bergen op Zoom: E.J.R. Slingenberg, P.M.R. Versteegh, F. Vaas. Ziekenhuis St. Franciscus, Roozendaal: J.L.M. Nooren, C.P. de Kraker, E.H.M. Kamphuys. Mammawerkgroep IKZ/SOOZ: H.F. P. Hillen. Catharina Ziekenhuis, Eindhoven. St Joseph Ziekenhuis, Eindhoven. Diaconnessenhuis, Eindhoven. St Anna Ziekenhuis, Geldrop. St Laurentius Ziekenhuis, Roermond: A.P. M Rutten, P.J.H. Sikkenk, C.M. Nuyens, R.J.A. Estourgie, J.A.J.M. Wils. St. Jans Gasthuis, Weert: W.A. van Deurzen, F.M. Lasisang, H.A.J.M. Scheerder. St Maartens Gasthuis, Venlo: F.G.J. Laudy, H.A.J.M. Scheerder. St. Elisabeth Ziekenhuis, Venray: A.L. Eekhout. Stichting Ziekenzorg Westelijke Mijnstreek, locatie Sittard: H.J.J. Stroehen. Stichting Ziekenzorg Westelijke Mijnstreek, locatie Geleen: F.W.C. van de Ent.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank J. Littooij, L. Schuurmans, M. Broeckaert, E. Matze-Cok, H. de Zeeuw, E. Wisse-Brekelmans, J. van Eijk, C. van Galen, J. Konneman, J. Brugghe, B. van Diermen, and S.M. Allen-de Jong for their skilled technical assistance.
NOTES
Supported by grant No. 28-1398 of the National Health and Research Council of The Netherlands, Zorg Onderzoek Nederland Medische Wetenschappen (ZonMw), and grant Nos. 95-112 and 98-136 of the Stichting Bevordering Diagnostiche Morfometrie (SBDM).
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Early Breast Cancer Trialists' Collaborative Group: Tamoxifen for early breast cancer: An overview of the randomised trials. Lancet 351:1451-1467, 1998
Early Breast Cancer Trialists' Collaborative Group: Polychemotherapy for early breast cancer: An overview of the randomised trials. Lancet 352:930-942, 1998
Goldhirsch A, Wood WC, Gelber RD, et al: Meeting highlights: Updated international expert consensus on the primary therapy of early breast cancer. J Clin Oncol 21:3357-3365, 2003
Boyages J, Chua B, Taylor R, et al: Use of the St Gallen classification for patients with node-negative breast cancer may lead to overuse of adjuvant chemotherapy. Br J Surg 89:789-796, 2002
European Commission Group on Breast Screening Pathology: Consistency achieved by 23 European pathologists from 12 countries in diagnosing breast disease and reporting prognostic features of carcinomas. Virchows Arch 434:3-10, 1999
Meyer JS, Friedman E, McCrate MM, et al: Prediction of early course of breast carcinoma by thymidine labeling. Cancer 51:1879-1886, 1983
Silvestrini R, Daidone MG, Luisi A, et al: Cell proliferation in 3,800 node-negative breast cancers: Consistency over time of biological and clinical information provided by 3H-thymidine labelling index. Int J Cancer 74:122-127, 1997
Sigurdsson H, Baldetorp B, Borg A, et al: Indicators of prognosis in node-negative breast cancer. N Engl J Med 322:1045-1053, 1990
Bergers E, van Diest PJ, Baak JPA: Prognostic implications of different cell cycle analysis models of flow cytometric DNA histograms of 1301 breast cancer patients: Results from the Multicenter Morphometric Mammary Carcinoma Project. Int J Cancer 74:260-269, 1997
Amadori D, Nanni O, Marangolo M, et al: Disease-free survival advantage of adjuvant cyclophosphamide, methotrexate, and fluorouracil in patients with node-negative, rapidly proliferating breast cancer: A randomized multicenter study. J Clin Oncol 18:3125-3134, 2000
Baak JPA, van Dop H, Kurver PHJ, et al: The value of morphometry to classic prognosticators in breast cancer. Cancer 56:374-382, 1985
Tosi P, Luzi P, Sforza V, et al: Correlation between morphometrical parameters and disease-free survival in ductal breast cancer treated only by surgery. Appl Pathol 4:33-42, 1986
van der Linden JC, Baak JP, Lindeman J, et al: Prospective evaluation of prognostic value of morphometry in patients with primary breast cancer. J Clin Pathol 40:302-306, 1987
Uyterlinde AM, Baak JPA, Schipper NW, et al: Prognostic value of morphometric and DNA flow-cytometry features of invasive breast cancers detected by population screening: Comparison with control group of hospital patients. Int J Cancer 48:173-181, 1991
Clayton F: Pathologic correlates of survival in 378 lymph node-negative infiltrating ductal breast carcinomas: Mitotic count is the best single predictor. Cancer 68:1309-1317, 1991
Makkink-Nombrado SV: Quantitative Pathological, Cell and Molecular Biological Markers in Breast Cancer. Amsterdam, the Netherlands, Free University Press, 1995, pp 149-163
Thor AD, Liu S, Moore DH II, et al: Comparison of mitotic index, in vitro bromodeoxyuridine labeling, and MIB-1 assays to quantitate proliferation in breast cancer. J Clin Oncol 17:470-477, 1999
Groenendijk RP, Bult P, Noppen CM, et al: Mitotic activity index in interval breast cancers. Eur J Surg Oncol 29:29-31, 2003
Manders P, Bult P, Sweep CG, et al: The prognostic value of the mitotic activity index in patients with primary breast cancer who were not treated with adjuvant systemic therapy. Breast Cancer Res Treat 77:77-84, 2003
Donhuijsen K, Schmidt U, Hirsche H, et al: Changes in mitotic rate and cell cycle fractions caused by delay in fixation. Hum Pathol 21:709-714, 1990
Bergers E, Jannink I, van Diest PJ, et al: Influence of fixation delay on mitotic activity and flow cytometric %S-phase. Hum Pathol 28:95-100, 1997
World Health Organisation.Histologic Typing of Breast Tumors (ed 2). Geneva, Switzerland, World Health Organisation, 1981
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
Ellis IO, Schnitt SJ, Sastre-Garau X, et al: Invasive breast carcinoma, in Tavassoli FA, Devilee P (eds): Pathology and Genetics of Tumours of the Breast and Female Genital Tract Organs. IARC Press, Lyon, France, 2003, pp 18-19
Hall PA, Going JJ: Predicting the future: A critical appraisal of cancer prognosis studies. Histopathology 35:489-494, 1999
Baak JPA: The framework of pathology: Good laboratory practice by quantitative and molecular methods. J Pathol 198:277-283, 2002
Baak JPA, van Diest PJ, Ariens A, et al: The multicenter morphometric mammary carcinoma project (MMMMCP): A nationwide prospective study on reproducibility and prognostic power of routine quantitative assessments in The Netherlands. Pathol Res Pract 185:664-670, 1989
van Diest PJ, Baak JPA, Matze-Cok P, et al: Reproducibility of mitosis counting in 2469 breast cancer specimens: Results from the Multicenter Morphometric Mammary Carcinoma Project. Hum Pathol 23:603-607, 1992
Jannink I, Diest PJ van, Baak JPA: Comparison of the prognostic value of mitotic frequency and mitotic activity index in breast cancer. Breast 5:31-36, 1996
Kaplan EL, Meier P: Long parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Cox DR: Regression models and life tables. J R Stat Soc B 34:187-220, 1972
Volpi A, Bacci F, Paradiso A, et al: Prognostic relevance of histological grade and its components in node-negative breast cancer patients. Mod Pathol 17:1038-1044, 2004
Volpi A, Nanni O, De Paola F, et al: HER-2 expression and cell proliferation: Prognostic markers in patients with node-negative breast cancer. J Clin Oncol 21:2708-2712, 2003
Janssen EA, Baak JPA, Guervos MA, et al: In lymph node-negative invasive breast carcinomas, specific chromosomal aberrations are strongly associated with mitotic activity and predict outcome more accurately than, grade, tumour diameter and oestrogen receptor. J Pathol 201:555-561, 2003
Yu Y, Xu F, Peng H, et al: Noey2 (ARH1), an imprinted putative tumour suppressor gene in ovarian and breast carcinomas. Proc Natl Acad Sci U S A 96:214-219, 1999
Hisatomi H, Nagao K, Wakita K, et al: ARH1/NOEY2 inactivations maybe important in breast tumour pathogenesis. Oncology 62:136-140, 2002
Wang L, Hoque A, Luo RZ, et al: Loss of the expression of the tumor suppressor gene ARHI in associated with progression of breast cancer. Clin Cancer Res 9:3660-3666, 2003
Soder AI, Hoare SF, Muir S, et al: Amplification, increased dosage and in situ expression of the telomerase RNA gene in human cancer. Oncogene 14:1013-1021, 1997
Ma YY, Wei SJ, Lin YC, et al: PIK3CA as an oncogene in cervical cancer. Oncogene 19:2739-2744, 2000
Mokbel KM, Parris CN, Ghilchilk M, et al: Telomerase activity and lymphovascular invasion in breast cancer. Eur J Surg Oncol 26:30-33, 2000
van de Vijver MJ, He YD, van't Veer LJ, et al: A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999-2009, 2002(Jan P.A. Baak, Paul J. va)