Growth fraction as a predictor of response to chemotherapy in node-negative breast cancer

Authors

  • Mohammed A. Aleskandarany,

    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    2. Pathology Department, Faculty of Medicine, Menoufyia University, Menoufyia, Egypt
    Search for more papers by this author
  • Andrew R. Green,

    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    Search for more papers by this author
  • Emad A. Rakha,

    1. Pathology Department, Faculty of Medicine, Menoufyia University, Menoufyia, Egypt
    2. Department of Pathology, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom
    Search for more papers by this author
  • Rabab A. Mohammed,

    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    2. Pathology Department, Faculty of Medicine, Assiut University, Assiut, Egypt
    Search for more papers by this author
  • Somaia E. Elsheikh,

    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    2. Pathology Department, Faculty of Medicine, Menoufyia University, Menoufyia, Egypt
    Search for more papers by this author
  • Desmond G. Powe,

    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    Search for more papers by this author
  • Emma C. Paish,

    1. Department of Pathology, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom
    Search for more papers by this author
  • R. Douglas Macmillan,

    1. The Breast Unit, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom
    Search for more papers by this author
  • Steve Chan,

    1. Department of Oncology, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom
    Search for more papers by this author
  • Samreen I. Ahmed,

    1. Department of Oncology, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom
    Search for more papers by this author
  • Ian O. Ellis

    Corresponding author
    1. Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, United Kingdom
    • Department of Histopathology, Molecular Medical Sciences, Nottingham City Hospital NHS Trust, University of Nottingham, Hucknall Road, Nottingham NG5 1PB, United Kingdom
    Search for more papers by this author
    • Fax: +44-0115-9627768


Abstract

Accurate predictive markers of chemotherapeutic response in early breast cancer are still lacking. The role of tumour growth fraction as a predictor of response to chemotherapy was assessed in early breast cancer. In this study, immunohistochemical expression of MIB1 was studied in a well-characterised series of early (Stages I and II) node-negative breast carcinoma cases (n = 100) with long-term follow-up that have received adjuvant chemotherapy (cyclophosphamide/methotrexate/5-fluorouracil regimen). In addition, 728 cases who did not receive adjuvant chemotherapy were used as a control group. Increased tumour growth fraction was associated with a better response to adjuvant chemotherapy in terms of longer breast cancer specific survival and disease-free interval [hazard ratio (HR) = 0.354, 95% CI = 0.177–0.688, p = 0.003 and HR = 0.396, 95% CI = 0.205–0.768, p = 0.006, respectively]. In contrast to the control group, patients with high growth fraction tumour (>70%) showed an excellent outcome with infrequently reported events during the period of follow-up. Importantly, patients with a low growth fraction (≤10%) showed frequent recurrences and shorter survival time with outcome comparable to those of high growth fraction who did not receive chemotherapy. Therefore, tumour growth fraction can be used to assign patients into distinct groups showing differential response to adjuvant chemotherapy. Patients with a high growth fraction appear to be ideal candidates for adjuvant chemotherapy while those with low growth fraction are less likely to benefit and are prone to the potential serious side effects of adjuvant chemotherapy.

Breast cancer is regarded as a heterogeneous group of tumours with diverse behaviour, outcome, and response to therapy.1 Much effort has been made to identify useful prognostic and predictive factors to support patient care. These factors not only help in assessing the overall prognosis, but also serve a crucial role in decision making for therapeutic strategies. They are able to identify those patients with an aggressive phenotype who potentially may benefit most from adjuvant therapy, and those with indolent tumours with minimal risk of recurrence who may therefore not necessarily need adjuvant treatment and its associated risk.2, 3

Adjuvant systemic therapy is increasingly used in the management of patients with breast carcinomas (BCs) to reduce the incidence of relapse, prolong the patient's survival by the early treatment of occult metastases and to palliate symptoms related to the disease.4, 5 The introduction of such therapies besides the early detection conferred by the screening programs has contributed to the improvements in breast cancer survival achieved so far.6, 7 In early-stage breast cancer, it has been estimated that chemotherapy can achieve 20–30% improvement in disease-free survival and around 15% or greater increase in overall survival rates.8, 9 However, resistance to therapy, a challenge in which the patient does not respond to therapy either partially or absolutely, is observed in a significant subset of patients, leading to subsequent disease progression.10 Moreover, systemic chemotherapy has the disadvantage of a broad range of side effects, and candidate patients should be carefully selected to avoid overtreatment and potentially serious side effects.11, 12 Accordingly, much effort has been made to develop consensus treatment guidelines for early breast cancer through analysis of the results of the randomized clinical trials and data from evidence-based literature and their implementation according to their relevance to individual patients with BC. A simple example of these guidelines are those reached by the National Cancer Institute Consensus panel.13 A more comprehensive statement has been published by the National Comprehensive Cancer Network (NCCN),14 whereas the most detailed set of guidelines was produced by the St Gallen's Consensus Panels.15, 16 Consequently, the concept of tailored therapy is evolving aiming at assigning patients with early stage breast cancer into specific therapeutic regimens based on risk assessment and prediction of their potential responsiveness.17 Patients with lymph node (LN)-negative (LN−) disease are managed based on primary tumour size and tumour grade in addition to other variables such as age and oestrogen receptor (ER) status. A subset of these patients with unfavourable prognosis are offered systemic therapy in the form of hormone therapy and/or chemotherapy with or without targeted therapy (i.e., Herceptin for HER2-amplified tumours).14

In contrast to predictive factors for hormonal therapy (HT), which are well established, predictive markers for chemosensitivity are less well defined.18, 19 Furthermore, it is difficult to recognize those patients who are more likely to respond to a particular chemotherapy regimen based on traditional prognostic factors such as LN status and tumour size alone.20 It is now acknowledged that increased cell proliferation is a key determinant of clinical outcome in patients with breast cancer.21 Moreover, genes involved in cell cycle regulation are highly expressed in the poor outcome subsets of breast cancer in almost all tumour microarray gene expression data sets suggesting that cell proliferation is the prime driving mechanism associated with aggressive phenotype and poor outcome.22

Ki-67, a labile non-histone nuclear protein of ∼395 kDa encoded for by almost 30,000 nucleotide base pairs within the human genome, is present exclusively in proliferating cells yet absent in quiescent or resting cells, as revealed by detailed cell cycle analysis.23 This protein undergoes phosphorylation and dephosphorylation during mitosis and is eliminated by proteases, implying a tightly regulated pathway ensuring rapid degradation, with an estimated half life of 60 ± 90 min.24–26 Thus, it has been used as a bio-marker to assay the growth fraction of a given cell population. Previous studies have demonstrated that assessment of Ki-67/MIB1 using immunohistochemistry (IHC) represents an easy, cost-effective, reproducible and reliable method of assessing tumour cell proliferation in breast cancer.23 In addition, extensive studies have demonstrated the prognostic significance of growth fraction assessment in patients with BC using Ki-67/MIB1.27–29 However, one of the limitations in using Ki-67/MIB1 in routine practice is the diversity in cut-off points used in research studies that can separate low from highly proliferative tumours.29

In this study, we have assessed the usefulness of growth fraction assessment, using the Ki-67-specific antibody MIB1, in predicting response to chemotherapy in patients with primary operable (Stages I and II) LN− invasive BC.

Material and Methods

Patients selection

In this study, we investigated a consecutive series of 828 invasive operable LN− BC cases equivalent to TNM Stages I and II, presenting between 1988 and 1998, obtained from the Nottingham Primary Breast Carcinoma Series for patients aged 70 years or less. This is a well-characterised series of primary BC cases with a long-term follow-up, and has been treated in a uniform way and used to study a broad range of biomarkers.30, 31 All patients who received chemotherapy were included (as a test group) to study the predictive utility of MIB1 Labelling Index (LI), while the remaining LN− patients who did not receive chemotherapy were used in determination of MIB1 LI optimal cut-point to overcome the previously reported varied MIB1 LI cut-off points.29 For all patients, a standardised approach for diagnosis was followed, and the treatment given accordingly. Adjuvant treatment was scheduled on the basis of patients' tumour prognostic and predictive factor status including Nottingham Prognostic Index (NPI), ER status, and menopausal status. Patients within the good prognostic group (NPI < 3.4) did not receive adjuvant therapy. HT was prescribed to patients with ER-positive tumours and NPI scores of >3.4 (moderate and poor prognostic groups). Premenopausal patients within the moderate and poor prognostic groups were candidates for CMF (cyclophosphamide, methotrexate and 5-flourouracil) chemotherapy, and those ER-positive LN-positive patients were offered HT in addition to CMF. Conversely, postmenopausal patients with moderate or poor NPI and ER-positive were offered HT, while ER-negative patients received CMF if fit. Systemic chemotherapy in this patients' series was used only in the adjuvant setting.

Clinical history including age, menopausal status and family history, and tumour characteristics including tumour type, tumour size, histologic grade, vascular invasion, LN status and NPI were available. Information on patients' therapies and clinical outcomes are maintained on a prospective basis. These include survival status, survival time, cause of death, disease-free interval (DFI), time to loco-regional recurrence and/or distant metastasis (DM). The Breast Cancer Specific Survival (BCSS) is defined as the time in months from the date of the primary surgery to the date of breast cancer-related death. DFI is defined as the duration (in months) from the date of primary surgery to the appearance of loco-regional recurrence or DM.

This study was approved by the Nottingham Research Ethics Committee 2 under the title “Development of a molecular genetic classification of breast cancer”.

Immunohistochemistry

Four-micrometer formalin-fixed paraffin-embedded tissue sections of full face breast cancer tissue were immunohistochemically stained, employing the standard streptavidin–biotin complex method. Primary mouse monoclonal antibody against Ki-67 antigen (clone MIB1; DAKO, Denmark) diluted 1:100 in normal swine serum was applied to each slide and incubated for 60 min at room temperature. 3,3′-Diaminobenzidine tetrahydrochloride (Dako liquid DAB Plus, K3468) was used as a chromogen. The sections were counterstained with Mayer's haematoxylin. Human tonsillar sections were used with each run as a positive control, while negative controls were performed by omitting the application of primary antibody.

MIB1 scoring

The MIB1 LI was determined using a semi-quantitative visual approach. Scoring was performed by 2 of the authors independently (MAA and SEE) blinded to patients' information and outcomes. The entire slide was scanned for immunostaining evaluation using light microscope at low-power magnification (×100). All tumour cell nuclei with homogenous granular staining, multiple speckled staining or nucleolar staining were regarded as positively stained, regardless of intensity, while any cytoplasmic immunoreactivity was considered non-specific and hence not taken into consideration. Scoring was performed in the areas with highest number of positive nuclei (hot spot) within the invasive component of the tumour. The MIB1 LI (tumour growth fraction) was expressed as the percentage of MIB1-positive malignant cells among a total number of 1,000 malignant cells, at high-power magnification (×400).

Statistical analysis

The relationship between categorised MIB1 LI and various pathobiological variables was investigated using the Chi square (χ2) test. Kappa agreement coefficient test was used to assess agreements between the 2 readings of MIB1 LI read by the 2 authors.

Determination of the optimal MIB1 LI cut-offs were performed using distribution histogram of the MIB1 LI of the training set. The results were cross-validated using X-tile bioinformatics software, version 3.6.1, 2003–2005, Yale University, USA).32 Relation between LI and patients' outcome measures were performed using the Kaplan–Meier method and differences tested for significance using the log-rank (LR) test. Multivariate analyses were performed using Cox regression proportional hazard analysis. All p-values were two-sided with statistical significance set at p < 0.05. All statistical analyses were performed using SPSS version 15 software (SPSS, Chicago, IL).

Results

In the current study, tumour growth fraction was assessed using MIB1 IHC in full-face tissue sections from 828 tumour tissues of patients with LN− primary operable invasive BC. For the purpose of this study, 728 cases that did not receive chemotherapy were used as a training set to determine the optimal cut-off point of MIB1 LI and as a control group for outcome analysis. A further 100 cases that received chemotherapy were used as a test set to assess the efficacy of MIB1-determined growth fraction in prediction of patients' response to chemotherapy. Sections scoring performed semi-quantitatively with a very good agreement between MIB1 reactivity scored by MAA and SEE (kappa coefficient = 0.84, p < 0.001). The clinico-pathological characteristics of the study population are summarised in Table 1.

Table 1. Summary of the characteristics of the study population and their specimens
inline image

For the test set, the patients had a median age of 45 years (range 28–70 years), with 79% of patients being premenopausal. Of all tumours, 83% were ductal carcinoma of no special type (duct/NST), with the remainder consisting of tumours of other histologic types including medullary, tubular mixed, lobular mixed, and mixed NST and lobular. The majority of tumours (83%) were Grade III, ER-negative (80%) and HER2-negative (83%). The NPI in the studied cases ranged from 3.41 to 5 (mean, 4.4). Regarding the adjuvant systemic therapy used, all 100 treated patients received classic chemotherapy CMF; 5 of these (5%) received additional tamoxifen as HT. The median follow-up time was 124 months (range 10–196 months). Loco-regional recurrence occurred in 13 patients (13%), while 18 patients (18%) experienced DM (Table 1).

Determination of MIB1 LI cut-point(s)

A total of 728 patients with LN− breast cancer who did not receive chemotherapy were used as training set to determine the optimal cut-points for MIB1 LI. The distribution of MIB LI within this group was found to be non-normal, with a 10% median expression (range, 0–99). Initial analysis of the MIB1 LI as continuous variable using correlation statistics revealed significant association between MIB1 LI and standard clinico-pathologic parameters. However, a biologically and clinically relevant cut-point has to be determined to study the MIB1 LI implications on patients' outcomes. The cut-off points at 10 and 70% were taken at the mid-point of the observed troughs in the distribution histogram of MIB1 LI, which may represent different populations of cases (Fig. 1). Cross-validation was performed using X-tile bio-informatics software, which randomly divides the total patient cohort into 2 separate equal training and validation sets by producing separate lists of “censored” and “uncensored” observations, ranked by patients' follow-up time. Those died from breast cancer were considered as an event (uncensored), while patients died of other causes or lost to follow-up during this study were censored during the analysis at the time of that event. The optimal cut-points were (10 and 70%), determined by locating the brightest pixel on the X-tile plot diagram of the training set. Statistical significance was tested by validating the obtained cut-point to the validation set.32 This resulted in classifying cases into 3 groups: low proliferative (0–10%), moderately proliferative (11–70%) and highly proliferative (71–100%). Highly statistically significant differences was observed between these different proliferative groups regarding BCSS and DFI (LR = 29.5, p < 0.001; LR = 12.7, p = 0.002) (Fig. 2).

Figure 1.

Distribution histogram of MIB LI within the training set of LN patients that did not receive chemotherapy (=728) with cut-points set at 10 and 70%.

Figure 2.

Kaplan–Meier survival plot for BCSS and recurrence-free survival at 10 and 70% MIB1 LI cut-off points: (a, b) Patients that did not receive chemotherapy (training set); (c, d) Patients who received chemotherapy (test set). Low proliferative (0–10%), moderately proliferative (11–70%) and highly proliferative (71–100%)

Assessment of MIB1 expression in patients with LN− BC who received chemotherapy

MIB1 expression was detected in all patients with LN− BC that received chemotherapy (n = 100) with a median expression of 70% (range 1–100%). The correlation between MIB1 LI in these patients and the clinico-pathologic variables was assessed using the determined cut-off points (10 and 70%). Table 2 shows the association between MIB1 LI and other clinico-pathological variables. Higher MIB1 LI were associated with Grade III tumours, invasive carcinoma NST, EGFR-positive and ER- and HER2-negative tumours. On the contrary, no differences were observed between the MIB1 LI and patients' age, menopausal status, tumour size, Bax and p53 expression.

Table 2. Correlation between MIB1 li and clinicopathologic parameters in lymph node negative invasive breast carcinoma patients that received chemotherapy
inline image

Patients' outcome

Survival analyses showed an association between tumour growth fraction and patients' outcome in terms of BCSS and DFI (Figs. 2 and 3 and Table 3). Patients within the low and moderately proliferative groups showed 5/12 and 8/38 deaths due to breast cancer, respectively, within 10 years of follow-up. Contrasting this, in the highly proliferative group (47 cases), only 4 deaths occurred. Interestingly, we noticed that no deaths were recorded in those having a growth fraction equal to or more than 90% (Fig. 3a).

Figure 3.

(a) Kaplan–Meier survival plot for patients' BCSS within the test set at 10 MIB1 LI above and below 90%. (b) Kaplan–Meier survival plot for patients' recurrence-free survival within the test set at 10 MIB1 LI above and below 90%.

Regarding the DFI, patients within the low proliferative group experienced significantly more recurrences than moderately and highly proliferative groups within 10 years of follow-up, indicating better response to chemotherapy (5/12, 13/38 and 6/47, respectively). Similar to BCSS, the recurrences within the highly proliferative group occurred in those having a growth fraction from 71 to 89%, with no recurrences noticed in those ≥90% (Fig. 3b).

In addition, this pattern of survival was the reverse of that seen in the training set where highly proliferative tumours were associated with a worse outcome while low proliferative tumours showed the best outcome. Table 3 displays a summary of outcome analysis in the training and test sets.

Table 3. BCSS and DFS (disease free survival) analysis at 10 years follow-up for the training and test sets
inline image

To test the impact of combined effect of proliferation and apoptosis on patients' response to CMF chemotherapy, cases were stratified into 4 groups based on MIB1 LI and expression of the pro-apoptotic Bax (MIB1 low/Bax low, MIB1 high/Bax low, MIB1 low/Bax high and MIB1 high/Bax high). There were no statistically significant differences between these groups regarding their BCSS and DFI (LR = 2.78 and 1.414, p = 0.427 and 0.702, respectively).

Multivariate analysis

A multivariate Cox hazard model analysis of predictors of BCSS for the test set patients was performed including MIB1 LI, tumour grade, histologic tumour type, tumour size, ER status and menopausal status. This analysis demonstrated that high MIB LI (>70%) is an independent predictor of longer BCSS in chemotherapy-treated LN− patients (HR = 0.354, 95 CI = 0.177–0.688, p = 0.003) (Table 4).

Table 4. Cox proportional hazards analysis for predictors of bcss: effect of mib1 li, tumour grade, size, histologic type, and er status in the test set
inline image

Discussion

Breast cancer is now recognised as a systemic disease with an inherent tendency to disseminate to remote sites even in apparently indolent disease.33 This concept forms the basis of prescription of adjuvant chemotherapy, based on primary tumour attributes and patients' menopausal status. Conventional prognostic parameters, including tumour size and LN status for instance, are good indicators for probability of cancer relapse and survival, but are not predictive of sensitivity to a particular therapeutic regimen.34 Moreover, adjuvant chemotherapy carries the risk of potentially serious side effects on different body systems and organs; thus, valid predictive factors according to which patients could be enrolled or excluded from adjuvant chemotherapeutic regimens are currently extensively scrutinised.35, 36

In this study, the primary tumour growth fraction using MIB1 was investigated as a potential predictor for the response of combination chemotherapy on DFI and BCSS in a subset of patients with operable invasive LN−BC. We selected the LN− patients to avoid the confounding effect of nodal status on patients' outcome37 and hence on the predictive ability of MIB1. Moreover, as the reported tumour growth fraction thresholds vary widely in the literature,29 we included an observational training cohort of LN− BC cases that did not receive chemotherapy for optimal determination of growth fraction cut-off points. Within the constraints of current clinical management of breast cancer, it is clearly not possible ethically to identify a control group of patients who did not receive adjuvant chemotherapy with comparable tumour characteristics to those who were offered chemotherapy and vice versa. Therefore, a large tissue archive such as ours with long clinical follow-up is invaluable for performing this type of study.

The association between MIB1 LI and different clinico-pathologic parameters at optimised MIB1-determined growth fraction revealed a statistically significant difference with reference to tumour grade, histologic tumour type, ER, HER2/neu and EGFR status. These findings underscore the prognostic utility of MIB1-determined proliferative fraction in patients with breast cancer. Previous reports demonstrated comparable findings in premenopausal,27 postmenopausal38 and node-negative breast cancer.39 The Kaplan–Meier and LR test were used to evaluate the value of MIB1-determined primary tumour growth fraction as a predictor of patients' response to adjuvant chemotherapy, in terms of BCSS, DFI and time to loco-regional or distant recurrence. The results revealed that the higher the tumour growth fraction, the longer the BCSS, the less the susceptibility to loco-regional and distant recurrence, elucidating a better response to chemotherapy than those tumours of low proliferative fractions, independently of menopausal status, tumour size, grade, histologic tumour type and ER status. Bearing in mind that MIB1 is positive only in cells during the active phases of the cell cycle while being absent in quiescent cells and the biologic evidence that increased tumour cell proliferation is a driving factor for the tumour growth, the better outcome of patients harbouring highly proliferative tumours signifies better response to cytotoxic chemotherapy that target primarily dividing cell population.9, 40 In addition, the extent of treatment benefit from adjuvant endocrine therapy has recently been reported to be greater among patients with high MIB1-IHC-determined tumour growth fraction than those with low tumour growth fractions.41

Importantly, the outcome analysis at different MIB1 LI cut-off points revealed 2 interesting findings: first, tumours of 90% LI or more suffered no events in terms of loco-regional or distant recurrence and breast cancer-related deaths. Second, tumours of 10% LI or less suffered the majority of deaths in a steady fashion up to 100 months of follow-up, after which no events were reported; thus, the patients either suffered a recurrence, died early or survived event-free thereafter. Tumours of growth fractions intermediate between these 2 extremes (10 and 90%) suffered recurrences/deaths modestly less than the former but greater than the later. These findings might demonstrate the ability of MIB1-determined tumour growth fraction in predicting responders from non-responders to CMF chemotherapy, underscoring its utility as a potential additional biomarker to assist in allocating patients to chemotherapeutic regimens. In other words, patients having low proliferative tumours benefited the least from such systemic therapy and suffer a comparably poor outcome, besides the potential serious side effect that must in no means outweigh the potential therapeutic benefits. These findings back other authors' opinions that 10% MIB1-determined growth fraction threshold at or below which the potential benefits of chemotherapy should be deliberated to avoid patients' overtreatment.29, 42

As the total tumour growth rate is determined by the balance between proliferation and apoptosis,43 we used the combined IHC profile of MIB1 LI and the apoptosis regulator Bax in the test set. This combination, however, did not confer additional significance in predicting patients' response to CMF chemotherapy, probably due to limited number of cases in the resulting groups.

It is also important to mention that global gene expression analysis studies have demonstrated the role of proliferation signature in breast cancer prognostication and prediction to therapy. A recent meta-analysis of publicly available breast cancer gene expression studies showed that the key biological drivers in 9 prognostic signatures are proliferation, in addition to ER signalling and ERBB2 amplification,44 results that have been reinforced by other authors.45 Therefore, emphasising the need to integrate data on proliferation assessment to other molecular (e.g., molecular classes such as basal-like, HER2+ and luminal classes, and molecular prognostic signatures such as recurrence score) and clinico-pathological markers already used in routine breast cancer management. Assessment of MIB1 as shown in this study provides a simple and reliable method of assessment of tumour proliferative activity that can be used in routine practice further improving the prognostic and predictive value of the combinatorial biomarker sets that are currently used (ER, PgR and HER2).

In conclusion, MIB1-determined primary tumour growth fraction is a valuable predictive tool for the patients' response to chemotherapy in patients with operable LN− BC. The findings of this study, as well as the evidences of others, rationalise its routine use in the adjuvant setting of breast cancer. Patients with highly proliferative tumours (>70%) appear to benefit most from chemotherapy and should be offered systemic chemotherapy, to which their response would expectedly be excellent. On the contrary, those with 10% growth fraction or less benefit the least and have a higher risk of relapse; thus, the decision of their allocation into the chemotherapeutic regimens should be thoroughly weighed against the potentially serious side effects of chemotherapy.

Acknowledgements

MAA, SEE and RAM are funded by the Ministry of High Education (Egypt). ARG is funded by the Breast Cancer Campaign.

Ancillary