Failure event types and prognostic factors after node-positive breast cancer in patients treated by adjuvant chemotherapy: impact on follow-up

ARTICLE

bdc.2012.1592

Auteur(s) : Thomas Filleron¹ filleron.thomas@claudiusregaud.fr, Andrew Kramar^2,3, Florence Dalenc¹, Marc Spielmann⁴, Pierre Fumoleau⁵, Pierre Kerbrat⁶, Anne-Laure Martin⁷, Henri Roché¹

¹ Institut Claudius-Regaud, 20-24, rue du Pont-Saint-Pierre, 31052 Toulouse, France

² CRLC Val-d’Aurelle, unité de biostatistiques, parc Euromédecine, 34298 Montpellier Cedex 5, France

³ Centre Oscar-Lambret, unité de méthodologie et biostatistique, 3, rue Frédéric-Combemale, BP 307, 59020 Lille Cedex, France

⁴ Institut Gustave-Roussy, 94800 Villejuif, France

⁵ Centre Georges-François-Leclerc, 21000 Dijon, France

⁶ Centre Eugène-Marquis, 35042 Rennes, France

⁷ Fédération nationale des centres de lutte contre le cancer, 101, rue de Tolbiac, 75654 Paris, France

Reprints: T. Filleron

Introduction

Breast cancer, with an estimated 1.15 million new cases each year in the world, is the most common cancer in women and it has become a major public health problem. Due to this high incidence and relatively good prognosis related to programs of mass screening and improvements in adjuvant treatments, it is the most prevalent cancer in the world. Consequently, the number of patients attending follow-up visits after curative intent is on the rise.

The main objective of post-therapeutic follow-up is to detect local recurrence or second primary cancer at an early stage [1]. Other objectives are the diagnostic of symptomatic distant metastasis, detecting delayed side effects of treatment, as well as the provision for psychological security and the collection of data for research and quality assurance purposes [2, 3]. Intensive surveillance was a common practice in the 1970s and 1980s, but two large randomized trials have demonstrated that early detection of metastatic disease offers no benefit in terms of long-term survival in comparison to disease, which is discovered by patient symptoms or physical examination [4, 5]. It was important to remark that these two trials were conducted before the recent advances in the treatment of metastatic disease [6, 7]. Based on the results of these two trials, a majority of surveillance programs proposes more frequent examinations during the first three or five first years and annually thereafter. For example, the American Society of Clinical Oncology and the European Society of Medical Oncology suggest physical examination three- to six-monthly for three years, then six- to 12-monthly for two years followed by an indefinite period of annual follow-up with recommended annual mammography [8, 9].

Actually, there is no available evidence that early detection and treatment of recurrence has a favorable impact on prognosis. Different studies have investigated the role of routine follow-up to detect locoregional recurrence and the impact on prognosis [3, 10-14]. Only one recent meta-analysis of 13 retrospective studies supports the hypothesis that the detection of isolated locoregional or contralateral breast cancer recurrences in patients without symptoms has a beneficial impact on survival of breast cancer patients when compared to late symptomatic detection [15]. A systematic review was performed whether routine clinical assessment including clinical examination, surveillance mammograms or breast self-examination affects the method of detection of locoregional relapse or contralateral new primaries [16] and does strengthen the argument for the benefit of routine surveillance mammograms. Consequently, in absence of prospective studies, surveillance programs recommend to detect locoregional recurrence.

Concerning distant recurrence, in the absence of data showing improved survival or quality of life with an early detection, recommendations from national oncology societies remain conservative, calling for routine clinic visits but not for any radiologic and/or biological tests for detection of metastases in asymptomatic patients [17].

On the other hand, it has been demonstrated that scheduled clinic visits induce anxiety associated with the risk of detecting tumor relapse [18]. With the cost of complementary investigation and the limited resources of health care systems, the cost-effectiveness of frequent follow-up in terms of survival benefit and quality of life are highly questionable.

Actual post-treatment follow-up does not take into account any prognostics factors. However breast cancer is a heterogeneous disease, whose prognosis and clinical course may be dependent on clinical factors and molecular subtype [19, 20]. Not all patients have the same risk of developing locoregional recurrences, distant metastasis or contralateral breast cancer. Age is one of the most established risk factors for local recurrence after breast conservation [21]. Nodal status does not appear to be associated with an increased risk of local recurrence after either breast conservative surgery or mastectomy, but results from different series are contradictory [21, 22]. Patients who underwent breast conservative surgery in comparison with patients treated by mastectomy and patients with a higher tumor stage were at an increased risk of locoregional recurrence. Concerning distant metastases, the two majors prognostic factors identified in the literature were histologic tumor size and lymph node involvement. Hormonal receptors have been widely analyzed as prognostic factors; their significance has been variable according to different series. More recently, two studies demonstrated important differences in metastatic and locoregional recurrence risk, between breast cancer subtypes as defined by a panel of six-marker immunohistochemical, suggesting different program surveillance according to tumor biology [23]. The main aim of this paper is to identify prognostic factors associated with different types of first events and overall survival post-relapse.

Patients and methods

Study population

Three adjuvant chemotherapy clinical trials sponsored by the Fédération nationale des centres de lutte contre le cancer (FNCLCC) for node-positive breast cancer patients were included in this project: PEGASE-01, PACS-01, and only the over-expressed HER2 subgroup of patients in the PACS-04 trial (i.e., arms C and D). Major inclusion criteria of these three trials were presented in table 1. Among the 2,841 patients included in these trials, 21 patients were not analyzed for the following reasons: relapse before end of chemotherapy or radiotherapy, and lost to follow-up after completion of treatment. The remaining 2,820 patients are the subjects of this report. A summary of the three trials is presented below.

Table 1 Major inclusion criteria of the three adjuvant studies.

The PEGASE-01 protocol was designed to assess the value of one terminal high-dose regimen following conventional chemotherapy in a high risk (>8 involved nodes) population [24]. The standard arm (arm A) used four cycles of FEC100 (500 mg/m² of fluorouracil, 100 mg/m² of epirubicin [E], and 500 mg/m² of cyclophosphamide every three weeks) (n = 155). The experimental arm (arm B, n = 159) received after four cycles of the same regimen, one cycle of CMA (120 mg/kg of cyclophosphamide, 45 mg/m² of mitoxantrone, and 140 mg/m² of melphalan). At three years, disease-free survival was significantly better in Arm B (71% versus 55%, P = 0.002), as was event-free survival (EFS) (68% versus 53%, P = 0.006). No statistical difference was shown between arms for overall survival at that time.

In the PACS-01 trial, 1,999 patients were randomized to receive six cycles of FEC100 (996 patients) or a sequential regimen of three cycles of FEC100 followed by three cycles of docetaxel (D) (FEC-D) (1,003 patients) [25]. With a median follow-up of 60 months, the results showed that sequential adjuvant chemotherapy significantly improves disease-free and overall survival.

The PACS-04 trial was designed to compare six cycles of concomitant D and E versus six FEC100 in the adjuvant treatment of node-positive early breast cancer. Three thousand and ten patients were randomized between six cycles of adjuvant FEC100 (arm A: n = 1,515) and six cycles of concomitant ED (E and D 75 mg/m²) every three weeks (arm B: n = 1,495). HER2-positive patients (n = 528) were then randomised to observation only (arm C, n = 260) or to either receive trastuzumab during one year (arm D, n = 268) [26]. After a median follow-up of 48 months, no statistical difference was shown between arms C and D (HR = 0.86; 95%CI = [0.61-1.22]). For all of these three trials, five-year hormonal treatment was given when estrogen and/or progesterone receptor was expressed (>10% positive cells by an HIC method).

Protocols of surveillance used in the three trials were summarized in table 2. Between PACS-04 and the two other trials, the frequency of visits was different during the first two years. This was due to the fact that patients of the PACS-04 included in the present study, over-expressed HER2 and were randomized between arms C and D. The others patients of the PACS-04 trial had the same frequency of visits as the patients in the other two trials. During the early follow-up, the rhythm of surveillance of the PEGASE-01 and PACS-01 trials proposed respectively three-monthly and four-monthly visits and for the PACS-04 visits were planned at six, eight, nine and 12 months and six-monthly during the second year. After the second year, six-monthly visits were planned for all trials. A physical examination was performed at each visit. Imaging studies (i.e., mammography, chest X-Ray, liver ultrasound and bone scan) were performed annually for five years.

Table 2 Surveillance protocols used in the three clinical trials^a.

	Year 1						Year 2						Year 3		Year 4		Year 5
	M3	M4	M6	M8	M9	M12	M15	M16	M18	M20	M21	M24	M30	M36	M42	M48	M54	M60
Physical examination	#		#		#	#	#		#		#	#	#	#	#	#	#	#
Chest X-ray						#						#		#		#		#
Liver ultrasound						#						#		#		#		#
Bone scan						#						#		#		#		#
Mammography						#						#				#		#

# PEGASE-01; PACS-01; PACS-04 (subgroup of patients with over-expression of HER2+ and randomized between arms C and arm D).

^a All times were computed from date of surgery.

Nottingham Prognostic Index

The Nottingham Prognostic Index (NPI) was defined by Galea et al. as a function of tumor size, number of lymph nodes and grade [27]: NPI = 0.2 × T (cm) + N + SBR where T is the tumor size in cm, N the lymph node stage (1: node-negative; 2: 1-3 positive lymph nodes; 3: >3 positive lymph nodes), and the Scarf-Bloom and Richardson (SBR) grade (1: I, 2: II, 3: III) (SBR prognostic grading system). This index allows to define three subgroups with different prognosis for overall and EFS: low risk (NPI < 3.4), intermediate risk (3.4 ≤ NPI ≤ 5.4) and high risk (NPI > 5.4).

Statistical analysis

Data is summarized by frequency and percentage for categorical variables and by median and range for continuous variables. EFS is defined as the time from end of treatment (chemotherapy or radiotherapy if applicable) until first relapse (local, regional or distant), or contralateral breast cancer. Patients alive and who never relapsed were censored at last follow-up news. Overall survival post-relapse was defined for patients who relapsed as the time from first relapse to death or last follow-up news (censored).

EFS and post-relapse overall survival rates were estimated using the Kaplan-Meier method and univariate analyses were performed using the Logrank test to identify associated prognostic factors. All factors considered significant at the P < 0.10 level were included in a Cox multivariate analysis to identify the major independent prognostic factors [28].

Cumulative incidence associated with each type of first event was estimated using competing risks methodology [29]. Multivariate analysis were also conducted using the Fine and Gray model [30] in order to estimate the potential effects of different factors on the cumulative incidence in the presence of competing risks (i.e., locoregional relapse, contralateral and distant metastasis). This model does not make the strong assumption of independence between events and covariate effects can be interpreted directly in terms of the cumulative incidence function. In fact, a subdistribution hazard ratio (sHR) for a covariate greater than 1 implies a constant relative increase of the subdistribution hazard and hence a higher cumulative incidence.

All P-values reported are two-sided. For all statistical tests, differences were considered significant at the 5% level. Stata was used for all statistical analyses and the cmprsk R package was used for the Fine and Gray model.

Results

Patients and treatment characteristics

Main clinical and pathological characteristics of the 2,820 patients were reported in table 3. Median age of the patient population is 49 years (range: [22; 66]). The median follow-up was 53.1 months (95%CI = [52.8; 53.3]). The number of patients who died was 357 (12.7%), and the overall survival at two and five years were 95.4 and 85.0%, respectively. Patients alive at last follow-up news were followed in majority between 24 and 60 months (n = 1,786; 72.5%) and only 93 patients (3.8%) had less than 24 months of follow-up.

Table 3 Population characteristics.

SBR: Scarf-Bloom and Richardson prognostic grading system.

First event analysis

On the overall population, 732 patients (26%) had a disease-related event. The two- and five-year EFS rates are 85.1 and 71.4%, respectively.

The results of univariate analyses were shown in table 4. The following six covariates had a significant effect on EFS: age (P < 0.001), grade (P < 0.001), histologic tumor size (P < 0.001), number of involved lymph nodes (P < 0.001), estrogen (P < 0.001), progesterone receptors status (P < 0.001) and type of surgery (P < 0.001).

Table 4 Results of univariate analysis event-free survival (EFS).

SBR: Scarf-Bloom and Richardson prognostic grading system.

In the multivariate analysis (table 5), increased histologic tumor size, increased number of involved lymph nodes, high grade, negativity of estrogen receptors and negativity of progesterone receptors and age lower than 35 years were associated with an increased risk of relapse.

Table 5 Multivariate analysis – Cox versus Fine and Gray (n = 2,455).

EFS: event-free survival; NPI: Nottingham Prognostic Index; SBR: Scarf-Bloom and Richardson prognostic grading system.

Competing risks analysis

Distant metastasis was the most frequent first event type (n = 560, 19.9%), locoregional relapse and contralateral breast cancer occurred respectively in 114 patients (4.1%) and 58 patients (2%). The five-year cumulative incidence rates for distant metastasis, locoregional relapse and contralateral breast cancer were estimated at 22.1, 4.2 and 2.3%, respectively (figure 1). Results of the competing risks regression using the Fine and Gray model are summarized in table 5.

Locoregional recurrence

The major prognostic factor associated with locoregional recurrence was grade III (sHR = 9.64, P = 0.025). Age lower than 35 years showed a significantly higher risk of locoregional recurrence (sHR = 2.37, P = 0.007) compared to women of 50 years or older, and there was a trend for patients aged between 35 to 50 (sHR = 1.49, P = 0.079). Patients with more than eight involved lymph nodes were associated with a high probability of locoregional relapse (8-11: sHR = 2.74, P < 0.001; >11: sHR = 2.94, P < 0.001). Negativity of progesterone receptors was not statistically associated with an increased rate of locoregional recurrence, but there was a tendency (sHR = 1.66, P = 0.053). Histologic tumor size and negativity of estrogen receptors were not associated with an increased probability of locoregional failure. Patients who underwent mastectomy were associated with a lower probability of locoregional recurrence (sHR = 0.62; P = 0.037) compared to patients with breast conservative surgery.

Distant metastasis

In the multivariate analysis, major factors associated with an increased probability of distant metastasis were grade (grade II: sHR = 2.99, P < 0.001, grade III: sHR = 3.44, P < 0.001 compared to grade I) and the number of involved lymph nodes (N4-7: sHR = 2.05, P < 0.001; N8-11: sHR = 2.56, P < 0.001; n > 11: sHR = 3.94, P < 0.001 compared to N1-3). Histologic tumor size and negativity of progesterone receptors were also related to the rate of distant metastasis. The probability of distant metastasis is increased for women less than 35 years compared to women of 50 years or older (sHR = 1.91, P < 0.001). Age between 35 and 50 years was not associated with an increased risk of distant metastasis compared to older women (P = 0.37). No association was found between negativity of estrogen receptors and distant metastasis rate (P = 0.206). Patients who underwent mastectomy were associated with a higher rate of distant metastasis as first events (sHR = 1.26, P = 0.187) compared to women who underwent breast conservative surgery.

Contralateral breast cancer

Estrogen receptor negativity was associated with an increased probability of contralateral breast cancer (sHR = 3.97, P < 0.001). There was no other factor on multivariate analysis significantly associated with contralateral breast cancer.

Overall survival post-first event

For the 732 patients who had a first event according to the definition, the median post-event follow-up was 27 months (95%CI = [24.5; 30.1]) of whom 336 (45.9%) died. Overall survival following a first event was 56.8% (95%CI = [52.6; 60.8]) at two years.

On univariate analysis, we compared groups according to usual prognostic factors and type of first events (locoregional relapse, contralateral and distant metastasis). In order to explore if early and late recurrences have potentially different natural histories, we also studied the relapse-free interval (<12 months, 12-24 months, >24 months).

Results of univariate analysis are provided in table 6. The type of first events influenced overall survival following post-first event (P < 0.001, figure 2A). As expected, patients with distant metastasis had poorer survival compared to patients with other first events. Two-year overall survival rates following the first event were estimated as 87.7, 80.7 and 49% for contralateral breast cancer, locoregional recurrence and distant metastasis respectively. Three periods of relapse with different prognosis on overall survival post-relapse were identified: <12 months, 12 to 24 months and >24 months (P < 0.001, figure 2B). Other factors influencing post-relapse overall survival were age (P < 0.001), grade (P < 0.001), number of lymph node involved (P = 0.002), estrogen receptor status (P < 0.001) and progesterone receptor status (P < 0.001). There was no difference for tumor size (P = 0.199) nor histologic subtype (P = 0.159).

Table 6 Results of univariate and multivariate analysis – Overall survival post-first event.

SBR: Scarf-Bloom and Richardson prognostic grading system.

Results of multivariate analysis were shown in table 6. The major prognostic factor influencing post-first event overall survival was the type of first event with an increased risk of death after locoregional relapse (HR = 2.51; 95%CI = [1.10; 5.72]) and distant metastasis (HR = 6.80, 95%CI = [2.10; 13.30]) compared to contralateral breast cancer. Relapse-free interval was associated with an increased risk of death. Age more than 50 (HR = 1.31; 95%CI = [1.04; 1.65] P = 0.019), grade 3 tumors (HR = [1.56; 95%CI = [1.22; 1.9] P < 0.001), negativity of estrogen receptors (HR = 1.44; 95%CI = [1.10; 1.87] P = 0.007) and number of involved lymph nodes was also significantly associated with an increased risk of death. Negativity of progesterone receptors was not associated with overall survival post-relapse.

Prognostic index and competing risks

The NPI defined three groups of patients at different risks of first event (table 5), five year EFS rates were respectively estimated as 90.6, 71.0 and 52.5% for good, moderate and poor risk (figure 3A). Five-year estimates of cumulative incidence associated with locoregional recurrence were 0.9, 4.5 and 7.4% for respectively good, moderate and poor risk (figure 3B,C). At five years, the rates of distant metastasis were 6.4% 21.9% and 38.3% respectively. Results of the Fine and Gray model are provided in table 5. Patients in the moderate and poor risk groups were associated with an increased probability of distant metastasis and locoregional recurrence compared to patients in the good risk group.

The rates of relapse at 24 months were respectively estimated as 3% (locoregional recurrence: 0.2%, contralateral: 1%, distant metastasis: 1.8%) for patients classified at good prognosis, 14.6% (locoregional recurrence: 2.3%, contralateral: 0.8%, distant metastasis: 11.5%) in the intermediate prognostic group and as 28.9% (locoregional recurrence: 5.1%, contralateral: 1.3%, distant metastasis: 22.5%) in the poor prognostic group.

Discussion

We have analyzed data of 2,820 breast cancer patients with lymph node involvement treated in FNCLCC adjuvant chemotherapy trials. We first evaluated patterns of recurrence without taking into account the type of first event. We used competing risk methodology with three types of events (locoregional recurrence, contralateral breast cancer and distant metastasis) to study prognostic factors associated with each type of event in order to better adapt protocols of surveillance. This model distinguished between patients who were still alive and those who had already failed from competing causes; for example, a high risk of distant metastasis reduces the probability of observing locoregional recurrence or contralateral breast cancer. We also studied prognosis of overall survival following the first event. And finally, using competing risk methodology, the capacity of discrimination with the NPI score according to the type of first events was examined.

Contrary to other publications, the date of origin for EFS and competing risks analysis was defined as the date of end of treatment (chemotherapy or radiotherapy, if applicable) as opposed to date of surgery. Since we are in an adjuvant setting oriented to adapting follow-up, we considered that patients were at risk of relapse from the end of chemotherapy and radiotherapy treatment.

Unfortunately, our study had several limitations. First, the follow-up period of this study was short in relation to the natural history of breast cancer, but the majority of events occurred in the first five years post-treatment and clinicians need such information to adjust the timing of follow-up visits. Besides, the modalities of diagnosis (symptomatic or asymptomatic) and treatment at relapse were not available in these trials, consequently their respective impacts on survival could not be evaluated. Since patients in this study were included in three randomized trials with intensive follow-up, the rate of distant metastasis diagnosed asymptomatically during the first two years could be increased. The survival from time of detection of metastasis until death could be longer for these patients compared to those with symptomatic detection. The experimental arm of the PACS and PEGASE-01 trials were statistically associated with an improvement in disease-free survival. So, patients treated in the control arm of these two trials received a suboptimal treatment and the rate of relapse might be overestimated in this study population. Moreover, HER2 status and other components of molecular subtypes were not available for all patients, so their influence compared to conventional factors could not be reliably assessed. Furthermore, only node-positive women up to about 65 years of age who satisfied strict inclusion criteria (table 1) were included in these trials. Due to these important limitations, we are unable to suggest precise follow-up schedules and programs.

Consistently with the literature, patients with a high lymph node involvement recurred earlier [31, 32]. In our study, the rate of recurrence was estimated at 20.4% at three years and most of them were systemic ones. The cumulative incidences of distant metastasis and potentially curative events (i.e., locoregional recurrence or contralateral breast cancer) were 15.9 and 4.5%, respectively at three years. Among the 266 patients with grade I tumors, only seven potential curative events occurred (one locoregional recurrence at 47 months, and six contralateral breast cancers). This finding confirms results from other studies [33]. So, there seems to be no reason to propose frequent follow-ups during the first years in this good prognosis sub group of patients. As a result, we could suggest that this good prognosis group of patients (grade I and pN+) need essentially locoregional and annual surveillance during the first five years after radiotherapy, a similar result was available on the literature [33]. On the contrary, intensive and global (locoregional and distant) follow-up should be more appropriate for grade III tumors, patients younger than 35 years, women with more than eight involved lymph nodes and negative progesterone receptor status. In fact, these factors correspond to the major prognostic factors associated with locoregional recurrence. This last finding need to be imbalance with the fact that this population is also at higher risk of distant metastasis. Nevertheless, in absence of recent data showing a potential benefit to treat low burden asymptomatic metastatic disease, there is no recommendation made to perform systematically detection of metastasis in asymptomatic patients, [1, 17]. As a result, there is no patent reason to propose today, more frequent follow-ups to screen distant metastasis, including population of patients at high risk. New innovative and prospective studies will be done. Another interesting result was the capacity of the NPI score to discriminate three populations of patients with different risks of locoregional recurrence and distant metastasis. During the first year post-treatment, only two distant metastasis and one contralateral breast cancer were observed in the NPI good risk group. In addition, only four patients in this group had locoregional recurrences during longer follow-up. Using this index, it was possible to identify a population of patients for which the risk of distant metastasis was not low. In fact, early local recurrences reflect aggressive disease and more frequent development of metastases compared with delayed recurrences [34].

The prognosis of patients after recurrence depends partially on the relapse-free interval. Three periods with different prognosis of relapse were identified: early and intermediate recurrences which correspond respectively to the first and the second year post-treatment and late relapses after two years post-treatment. Other studies considered only two periods of relapse, early failure associated with an event in the first two years post-surgery and late recurrence with relapse after two years [31, 32]. In our study, as expected, the major prognostic factor associated with an increased risk of post-relapse death was the type of first event: groups with local recurrence or contralateral breast cancer having the most favorable post-relapse survival. According to the literature, other prognostic factors associated with an increased risk of death were linked to the aggressiveness of the initial tumor.

In conclusion, our findings in this set of lymph node-positive breast cancer patients confirm that frequency and perhaps modality of follow-up during the first five years can be adapted using prognostic factors associated with relapse. For patients with a low risk of locoregional recurrence, it seems reasonable to limit the frequency of routine follow-up during the first years. In order to adapt the frequency according to the risk of relapse, its seems favorable to propose one or two visits by years during the first five years, but optimal interval and total duration of follow-up remain unknown. This strategy will reduce costs and optimize staff resources [1]. For patients at a high risk of recurrence, regular follow-up should be maintained in order to detect potential curative events. Concerning screening for distant metastasis, there are no arguments at the present time to propose more intensive follow-up, thus adding systematic work-up in asymptomatic patients. The case of overexpressing HER2+ tumors should be specifically evaluated due to the efficacy of targeted anti HER2 agents even in the metastatic setting. Future research should focus on the adaptation of post-therapeutic follow-up schedules to molecular subtypes in order to take into account the dynamic process of the events. So, it seems important to evaluate the impact of early detection in these different subgroups on overall survival and quality of life data.

Acknowledgements

T.F. received a grant from the Institut national du cancer (INCa: 07/3D1317/66-PDOC-RC-GSO-005/NG-LC). All the investigators of these three trials are warmly thanked.

Support: T.F. received a grant from the Institut national du cancer (INCa: 07/3D1317/66-PDOC-RC-GSO-005/NG-LC).

Conflicts of interests

P. Fumoleau is PACS 01-04 investigator.

References

1. Kimman ML, Voogd AC, Dirksen CD, et al. Follow-up after curative treatment for breast cancer: why do we still adhere to frequent outpatient clinic visits?. Eur J Cancer 2007 ; 43 : 647-653.

2. Loong S, Wilkins S, Bliss J.M. The effectiveness of the routine clinic visit in the follow-up of breast cancer patients: analysis of a defined patient cohort. Clin Oncol 1998 ; 10 : 103-106.

3. te Boekhorst DS, Peer NG, van der Sluis RF, Wobbes T, Ruers T.J. Periodic follow-up after breast cancer and the effect on survival. Eur J Surg 2001 ; 167 : 490-496.

4. GIVIO Impact of follow-up testing on survival and health related quality of life in breast cancer patients. A multicenter randomized controlled trial. The GIVIO investigators. J Am Med Assoc 1994 ; 271 : 1587-1592.

5. Del Turco MR, Palli D, Cariddi A, Ciatto S, Pacini P, Distante V. Intensive diagnostic follow-up after treatment of primary breast cancer. A randomized trial. National Research Council Project on Breast Cancer follow-up. J Am Med Assoc 1994 ; 271 : 1593-1597.

6. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001 ; 344 : 783-792.

7. Hortobagyi G.N. Can we cure limited metastatic breast cancer?. J Clin Oncol 2001 ; 20 : 620-623.

8. Khatcheressian JL, Wolff AC, Smith TJ, et al. American Society of Clinical Oncology 2006. Update of the breast cancer follow-up and management guidelines in the adjuvant setting. J Clin Oncol 2006 ; 24 : 5091-5097.

9. Pestalozzi BC, Luporsi-Gely E, Jost LM, Bergh J ESMO Guidelines Task ForceESMO minimum clinical recommendations for diagnosis, adjuvant treatment and follow-up of primary breast cancer. Ann Oncol 2005 ; 16 (suppl. 1) : i7-i9.

10. Brada M. Is there a need to follow-up cancer patients?. Eur J Cancer 1995 ; 31 : 655-657.

11. Peppercorn J, Partridge A, Burstein HJ, Winer E.P. Standards for follow-up care of patients with breast cancer. Breast 2005 ; 14 : 500-508.

12. Grunfeld E. Optimizing follow-up after breast cancer treatment. Curr Opin Obstet Gynecol 2009 ; 21 : 92-96.

13. Jacobs HJ, van Dijck JA, de Kleijn EM, Kiemeney LA, Verbeek A.L. Routine follow-up examinations in breast cancer patients have minimal impact on life expectancy: a simulation study. Ann Oncol 2001 ; 12 : 1107-1113.

14. de Bock GH, Bonnema J, van der Hage J, Kievit J, van de Velde C.J. Effectiveness of routine visits and routine tests in detecting isolated locoregional recurrences after treatment for early-stage invasive breast cancer: a meta-analysis and systematic review. J Clin Oncol 2004 ; 22 : 4010-4018.

15. Lu WL, Jansen L, Post WJ, Bonnema J, Van de Velde JC, De Bock G.H. Impact on survival of early detection of isolated breast recurrences after the primary treatment for breast cancer: a meta-analysis. Breast Cancer Res Treat 2009 ; 114 : 403-412.

16. Montgomery DA, Krupa K, Cooke T.G. Follow-up in breast cancer: does routine clinical examination improve outcome? A systematic review of the literature. Br J Cancer 2007 ; 97 : 1632-1641.

17. Hayes D.F. Follow-up of patients with early breast cancer. N Engl J Med 2007 ; 356 : 2505-2513.

18. Allen A. The meaning of the breast cancer follow-up experience for the women who attend. Eur J Oncol Nurs 2002 ; 6 : 155-161.

19. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000 ; 406 : 747-752.

20. Dent R, Trudeau M, Pritchard KI, et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Res 2007 ; 1 : 4429-4434.

21. Voogd AC, Nielsen M, Peterse JL, et al. Differences in risk factors for local and distant recurrence after breast-conserving therapy or mastectomy for stage I and II breast cancer: pooled results of two large European randomized trials. J Clin Oncol 2001 ; 19 : 1688-1697.

22. Dalberg K, Mattsson A, Rutqvist LE, Johansson U, Riddez L, Sandelin K. Breast conserving surgery for invasive breast cancer: risk factors for ipsilateral breast tumor recurrences. Breast Cancer Res Treat 1997 ; 43 : 73-86.

23. Voduc D, Cheang M, Tyldesley S, Gelmon K, Nielsen T, Kennecke H. Breast cancer subtypes and the risk of local and regional relapse. J Clin Oncol 2010 ; 28 : 1684-1691.

24. Roché H, Viens P, Piron P, Lotz JP, Asselain B. High-dose chemotherapy for breast cancer: the French Pegase experience. Cancer Control 2003 ; 10 : 42-47.

25. Roché H, Fumoleau P, Spielmann P, et al. Sequential adjuvant epirubicin-based and docetaxel chemotherapy for node-positive breast cancer patients: the FNCLCC PACS 01 trial. J Clin Oncol 2006 ; 24 : 5667-5671.

26. Spielmann M, Roché H, Delozier T, et al. Trastuzumab for patients with axillary-node-positive breast cancer: results of the FNCLCC-PACS 04 trial. J Clin Oncol 2009 ; 27 : 6129-6134.

27. Galea MH, Blamey RW, Elston CE, Ellis I.O. The Nottingham prognostic index in primary breast cancer. Breast Cancer Res Treat 1992 ; 22 : 207-219.

28. Cox D.R. Regression model and life tables (with discussion). J R Stat Soc B 1972 ; 34 (suppl. 1) : 187-220.

29. Kalbfleish JD, Prentice R.L. The statistical analysis of failure time data. Wiley, 2^nd edThe statistical analysis of failure time data : 2002.

30. Fine J, Gray R. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999 ; 94 : 496-509.

31. GoldHirsch A, Gelber RD, Castiglione M. Relapse of breast cancer after adjuvant treatment in premenopausal and perimenopausal women: patterns and prognoses. J Clin Oncol 1988 ; 6 : 89-97.

32. Faneyte IF, Peterse JL, van Tinteren H, et al. Predicting early failure after adjuvant chemotherapy in high-risk breast cancer patients with extensive lymph node involvement. Clin Cancer Res 2004 ; 10 : 4457-4463.

33. Churn M, Kelly V. Outpatient follow-up after treatment for early breast cancer: updated results after 5 years. Clin Oncol 2001 ; 13 : 187-194.

34. Punglia RS, Morrow M, Winer EP, Harris J.R. Local therapy and survival in breast cancer. N Engl J Med 2007 ; 7 : 2399-2405.