Lapatinib Activity in Premalignant Lesions and HER-2–Positive Cancer of the Breast in a Randomized, Placebo-Controlled Presurgical Trial

To increase the efficiency of drug development in early breast cancer, window-of-opportunity, presurgical models are being used to screen the therapeutic activity of candidate agents and characterize their mechanism of action (1). These studies use tumor Ki-67 labeling index (LI) as the main surrogate biomarker in as much as the decrease in this proliferation-related antigen after short-term presurgical treatment has shown prognostic significance on progression-free and overall survival in several studies (2–4). For instance, our group has used this model to assess the minimal active dose of tamoxifen in ER-positive breast cancer (5) that led to the implementation of phase IIb and phase III prevention trials of tamoxifen at 5 mg/d (6, 7).

Moreover, the presurgical model may provide insight into the preventive potential of drug against tumor-adjacent dysplastic ductal intraepithelial neoplastia (DIN, which comprises ductal carcinoma in situ and atypical ductal hyperplasia) cells and distant ductal hyperplasia (without atypia) cells (the classification DIN is discussed in ref. 8). In other words, the window-of-opportunity trial model can exploit the existence of the field cancerization, which is well documented in breast cancer (9), to get important clues on the preventive effect of the investigational agent. Moreover, recent studies have shown that Ki-67 LI is positively associated with carcinogenesis progression (10) and its level of staining in atypical lesions in core biopsies of nonmalignant lesions may predict subsequent breast cancer risk, going from a 10-year cumulative incidence of breast cancer of 3% for lesions with Ki-67 LI less than 2% to 14% in those with Ki-67 LI 2% or more (11).

Lapatinib is an oral reversible small-molecule tyrosine kinase inhibitor which targets both human epidermal growth factor receptor 2 (HER-2) and epidermal growth factor receptor (EGFR) tyrosine kinases and is currently approved for the treatment of HER-2–positive metastatic breast cancer in combination with capecitabine after trastuzumab progression (12). Lapatinib also showed increased progression-free survival in combination with letrozole over letrozole alone as first line treatment of hormone-sensitive HER-2–positive disease (13) or in combination with trastuzumab over a single agent in patients progressing after trastuzumab (14). A phase III adjuvant trial is ongoing to assess its efficacy alone or in combination with trastuzumab in HER-2–positive breast cancer (15), and several neoadjuvant trials are testing lapatinib in combination with different agents, including trastuzumab, that may provide important insight into the therapeutic activity of drug (16). Given its oral formulation, there is interest to use this agent also in intraepithelial HER-2 disease, but evidence for its preventive potential is lacking. Moreover, different mechanistic aspects still remain unclear, including the influence of hormone receptor and Phosphatase and Tensin homolog (PTEN) expression.

The PTEN tumor suppressor is one of the most frequently mutated genes in human cancer. PTEN is the central negative regulator of the phosphoinositide 3-kinase (PI3K) signal transduction cascade for growth, proliferation, and survival. Despite its well-defined role in signaling at the plasma membrane, PTEN is found in the nucleus in a number of different normal and tumor cell types, which may contribute to its tumor suppressor activity as well (17). The activity of the PI3K/AKT pathway is frequently altered in breast cancer and expression of the tumor suppressor PTEN, which negatively regulates PI3K signaling, is lost in nearly 50% of breast cancers (18) and has been associated with poor prognosis and resistance to trastuzumab (16). A preclinical study suggested that PTEN loss is associated with lapatinib resistance (18), whereas others did not (19, 20).

We tested the antiproliferative activity of lapatinib on hyperplastic, dysplastic, and malignant breast tissue in a window-of-opportunity trial of 3 weeks before surgery in women with HER-2–positive breast cancer.

We conducted a randomized, single institution, phase IIb, window-of-opportunity trial of lapatinib, administered for 3 weeks prior to surgery at 1,500 mg once daily (6 tablets taken in a fasting state in the morning) or matching placebo in women with treatment naive Tis, T1-2, N0-1, HER-2–positive breast cancer (IHC 3⁺ and/or FISH⁺). The study received Institutional Review Board approval (IEO#S291; Eudract code: 2006-000520-14) and all subjects signed an informed consent.

All subjects screened for the study were assigned a 3 digit E-code and randomization numbers were allocated strictly sequentially as subjects entered the study. After randomization, drug was given to the patient by the physician in the outpatient clinic. Patients, care providers, and laboratory staff were blinded to treatment assignment. Toxicity was evaluated using the NCI Common Terminology Criteria for Adverse Events (CTCAE) version 3.0. Compliance was evaluated by pill count and self-administered calendar.

Prior to treatment, a core needle biopsy of the primary tumor was done percutaneously with a 14-gauge needle after the administration of local anesthetic. A single skin incision with the tip of a no. 11 scalpel was done before the biopsy. At the time of surgical removal of the tumor, 3 to 5 specimens of adjacent (within 1 cm from the tumor) and distant (>1 cm form the tumor) grossly normal tissue (i.e., the surgical margins of quadrantectomy or lumpectomy, or the grossly free quadrants from mastectomy specimens) were obtained to assess systematically the prevalence of ductal intraepithelial neoplasia (DIN) and ductal hyperplasia according to Tavassoli et al. classification (21). Specifically, the prevalence of DIN was evaluated in samples both adjacent and distant from the tumor, whereas the prevalence of ductal hyperplasia was evaluated exclusively in samples distant from the tumor. Fasting blood samples were drawn between 8:00 AM and 10:00 AM at baseline prior to treatment initiation (day of randomization) and after 3 weeks of treatment during the day of surgery to assess hematology and biochemistry and to measure circulating biomarkers, that will be the subject of a separate article.

Ki-67 LI, estrogen (ER) and progesterone receptors (PgR), Her2/neu and PTEN were centrally assessed at the Division of Pathology of the European Institute of Oncology (IEO) in all core biopsies and posttreatment surgical samples. Immunostainings were done on 3-μm thick formalin-fixed, paraffin-embedded (FFPE) sections following proper heat-induced antigen retrieval with an automated immunostainer (Autostainer; Dako), using Food and Drug Administration–approved probes (for ER, PgR, and HER-2), and a commercially available detection kit (Dako EnVision Plus-HRP; Dako), according to the manufacturer's instructions, as previously described (22, 23). PTEN immunoreactivity was evaluated using the monoclonal antibody 138G6 (Cell Signaling; 1:25 working dilution), that has recently been proposed as the most reliable probe for the analysis of PTEN immunoreactivity in a CLIA-certified laboratory (24). ER, PgR, Her-2, and PTEN immunoreactivity was evaluated exclusively in invasive cancers, whereas Ki-67 LI was assessed in ductal hyperplasia, DIN, and invasive cancer. Ki-67 immunohistostaining was assessed in randomly selected microscopic fields at the periphery of the invasive tumor, as reported recently (22) and in randomly selected areas of ductal hyperplasia and ductal intraepithelial neoplasia. ER and PgR immunoreactivity was reported according to the recently licensed ASCO/CAP guidelines (25).

For all the probes used, the immunohistochemical results were obtained by counting at least 500 cells at ×400 magnification blinded as to treatment, and by recording the percentage of cells showing any definite nuclear (for ER, PgR, and Ki-67), membranous (for Her-2), and nuclear or cytoplasmic (for PTEN) staining. For PTEN, we recorded the prevalence of neoplastic cells with nuclear or cytoplasmic intensity of immunostaining equal or higher than the built-in positive controls (i.e., endothelial cells).

The primary endpoint was the pre- versus posttreatment difference in the percentage reduction of Ki-67 LI between placebo and lapatinib. We assumed on the basis of a previous trial (5) a 40% reduction of Ki-67 on lapatinib versus placebo with a 50% SD of Ki-67 change, thus requiring 30 subjects per arm for an 80% power, a 5% 2-sided α-error and a 10% drop out rate. A post hoc power calculation on the basis of the observed 30% SD of KI-67 LI change indicates an 87% power to detect the observed 20% reduction of Ki-67 after lapatinib relative to placebo, using linear regression modeling and correcting for baseline Ki-67. A computer-generated randomization list was prepared by an external statistician using the minimization technique and the drug allocation was done using an internet program.

Univariate associations between treatment and covariates were assessed using χ² or Fisher exact test for categorical variables and the Student's t test or the Mann–Whitney U test for continuous variables. A nonparametric test for trend across ordered groups was used when appropriate. The effect of treatment as dummy variable on Ki-67 LI percent change was tested using a linear regression model with Ki-67 LI percent change as response variable and baseline Ki-67 LI as adjustment variable. A full model including age, body mass index (BMI), ER, and PgR status, PTEN overexpression as covariates and their interaction with treatment was fitted, but the results of this analysis closely resembled those of the main analysis. Tumor response was classified using the Response Evaluation Criteria in Solid Tumors (RECIST), although this comparison was hampered by the comparison of different techniques (baseline ultrasound vs. pathologic specimen). The predictive effect of Ki-67 LI change on tumor size change was also assessed by a regression model including the interaction term between treatment and baseline tumor size. Two-tailed probabilities were reported and the P value of 0.05 was used to define significance. All analyses were conducted using Stata software (version 10; StataCorp.).

From November 3, 2006 to May 13, 2009, 285 women with clinical Tis, T1-T2 breast cancer were screened at the IEO. We randomly assigned 29 subjects to lapatinib and 31 subjects to placebo for 3 weeks. Two subjects in each arm were not analyzed for the primary endpoint. The participant flow diagram is summarized in Figure 1. The mean ± SD treatment duration was 20.9 ± 0.7 days. Baseline subject and tumor characteristics were evenly distributed between the 2 arms (Table 1). Nearly half of the women were premenopausal and two thirds had ER-positive disease.

The changes in Ki-67 LI by allocated arm and according to ER and PTEN expression are shown in Table 2. At baseline, the median Ki-67 LI was 30% in both arms. After 3 weeks, the mean ± SD percent change of Ki-67 LI was −9.3% ± 34.2% in the lapatinib arm and +15.1% ± 30.9% in the placebo arm (P = 0.007), with a resulting mean percentage decrease of Ki-67 LI between arms of 20.9% (95% CI: 5.6–36.1), after adjustment for baseline Ki-67 (P = 0.008). Compared with placebo, the Ki-67 LI reduction after lapatinib treatment was significant for ER-negative tumors (−34.8%, 95% CI: −61.0 to −8.6, P = 0.01) but not for ER-positive tumors (−12.3%, 95% CI: −31.8 to 7.2, P = 0.2), although the interaction term was not (P = 0.14). Among the prognostic or predictive factors shown in Table 1, the change in Ki-67 LI after lapatinib was significantly influenced by PTEN immunoreactivity. PTEN overexpressing tumors were nonsignificantly more frequent in ER-negative tumors (7 in nucleus ± cytoplasm and 5 in cytoplasm only in ER negative vs. 16 and 6 in ER-positive tumors, respectively; P = 0.5) and relative to placebo, the decrease in Ki-67 LI on lapatinib was greater in PTEN overexpressing tumors, with the maximal effect in tumors with cytoplasmic PTEN expression (percent change = −31.4, 95% CI: −64.1 to 1.2, P = 0.057, Fig. 2).

The median (range) tumor diameter at surgery was 18 mm (10–37 mm) in the lapatinib arm and 24 mm (11–57 mm) in the placebo arm (P = 0.009). The decrease in tumor size on lapatinib was significantly greater in tumors larger than 20 mm at baseline (P for interaction = 0.018). Partial response was noted in 13.6% versus 3.7%, stable disease in 59.1% versus 40.7%, and progression in 27.3% versus 55.6% of patients on lapatinib and placebo, respectively (P_trend = 0.035).

The prevalence of DIN and ductal hyperplasia and corresponding Ki-67 LI in the surgical specimens by allocated arm are shown in Table 3. The prevalence of DIN was 70% in the lapatinib arm and 76% in the placebo arm (P = 0.8), and the median (range) Ki-67 LI was 15% (5%–35%) in the lapatinib arm versus 20% (5%–60%) in the placebo arm (P = 0.067). Likewise, the prevalence of ductal hyperplasia was 96% and 93%, and the median Ki-67 LI was 1% (1%–7%) versus 3% (1%–5%) in the lapatinib and placebo arms, respectively (P = 0.006). A representation of the pre- versus post-lapatinib decrease in Ki-67 LI in cancer, DIN, and ductal hyperplastic tissue in selected patients is depicted in Figure 3.

Toxicity is illustrated in Table 4. Diarrhea (grade 1: 48% vs. 16%, P = 0.01; grade 2: 14% vs. 0%, P = 0.05) and skin rash (55% vs. 10%, P < 0.001) were more frequent on lapatinib than placebo. There were no serious adverse events, but 3 cases of grade 3 transaminases elevation in the lapatinib arm (P = nonsignificant).

Recent studies have shown that Ki-67 LI in atypical hyperplasia predicts future breast cancer risk (11) and is associated with the progression of carcinogenesis from ductal hyperplasia to DIN (10, 26). Moreover, Ki-67 LI after short-term presurgical treatment predicts recurrence-free and overall survival in women with early breast cancer, supporting the use of this biomarker to tailor adjuvant treatment and to screen drugs in a cost-effective manner (2–4, 27). Our previous study (4) indicated that a 3% absolute decrease in Ki-67 LI (e.g., from 18% to 15%) after 4 weeks of presurgical tamoxifen is associated with a 15% relative decrease of invasive recurrence and a nearly 18% decrease of death, a meaningful clinical accomplishment in women with operable breast cancer. Because the presurgical model may be ideal to unveil the field cancerization effect present in breast surgical specimens of women with breast cancer, we designed the present trial to further characterize the activity of lapatinib in early HER-2–positive breast cancer and to get clues into its preventive potential using adjacent intraepithelial neoplasia and distant proliferative tissue as relevant target tissues.

Compared with the placebo arm, where Ki-67 LI percent change increased by 15%, lapatinib decreased Ki-67 LI percent change by a relative value of 9%, with an absolute difference of approximately 6% between arms, a meaningful clinical effect in the context of previous data (2, 4). Notably, Ki-67 LI in the placebo arm was higher at surgery than at baseline, a common observation in most presurgical studies (5, 28, 29), presumably as a consequence of the higher antigen staining in the tumor edge that cannot be detected in the baseline core biopsy. Although tumor response did not significantly correlate with Ki-67 LI percent change, possibly for the limited study power, lapatinib was able to halve the high progression rate (55.6%) observed in the placebo arm as early as after 3 weeks and to decrease tumor size mainly in large tumors. Quantitatively, the 25% average decrease in tumor diameter with a greater effect in large tumor is a meaningful accomplishment given the prognostic significance of tumor shrinkage in neoadjuvant breast cancer therapy (30). However, our findings should be interpreted with caution given the different methods of pre- versus posttreatment tumor measurements. Moreover, determination of apoptosis might have helped to explain the tumor size reduction. Although phase III trials of lapatinib are already underway, our findings support the use of presurgical models to screen new drugs cost-effectively and to offer an effective treatment against rapidly proliferating tumors during the interval prior to surgery.

Our results indicate that the antiproliferative effect of lapatinib is greater in ER-negative tumors and in tumors overexpressing PTEN, especially those with a cytoplasmic localization of the staining. Laboratory studies have clearly established that HER-2 amplification and overexpression result in hormone therapy resistance (31). On the other hand, ER expression has been associated with lapatinib resistance in preclinical models (32, 33). Here, we provide evidence for this effect, with the hormone receptor expression conferring decreased antiproliferative activity to lapatinib. Moreover, our data provide further clinical evidence that PTEN, a central negative regulator of PI3K-mediated signals for growth, proliferation, and survival, affects lapatinib antiproliferative activity (18). PTEN is a lipid phosphatase that dephosphorylates PIP3 at the plasma membrane and thereby inhibits PI3K. Despite its well-defined role in signaling at the plasma membrane, PTEN is also found in the nucleus of different normal and tumor cell types (17). Although some previous studies suggested that lapatinib activity was not dependent upon PTEN inhibition (20, 34), genome-wide loss-of-function small hairpin RNA screen identified this tumour suppressor as a mediator of lapatinib sensitivity in vitro and in vivo (18), supporting the notion that PTEN expression is a useful predictive marker of response to lapatinib.

Our clinical model shows that the prevalence of the field cancerization effect is high, with proliferative and intraepithelial neoplastic lesions in tissue surrounding invasive cancers ranging from 70% to 75% for DIN to over 90% for ductal hyperplasia, supporting the notion that the presurgical model is a feasible and reliable model to screen compounds' preventive potential in a cost effective manner. Our findings show that Ki-67 LI increased along with the carcinogenesis progression (from ductal hyperplasia to DIN to cancer) and was significantly decreased after lapatinib treatment. In a mouse model, lapatinib was effective in preventing the onset of ER-negative, HER-2–positive mammary tumors (35). A phase III trial is underway to assess the efficacy of trastuzumab and radiotherapy in preventing recurrence after removal of HER-2–positive DIN (36). Altogether, our findings provide the rationale for testing lapatinib in the adjuvant treatment of HER-2–positive DIN after surgical removal.

Toxicity was manageable with 3 cases of grade 3 transaminites, including one SAE which delayed surgery for a few days. Diarrhea and skin rash were relatively frequent but mostly grade 1 or 2 and in line with prior studies (37). Moreover, these effects tend to resolve with conventional approaches and without dose modification in most cases (38). However, the optimal biological dose regimen of lapatinib and the most appropriate way of administration in relation to meal consumption and daily intake frequency have yet to be identified (39). Further studies are necessary to address these issues, particularly in the prevention setting.

In conclusion, lapatinib given for 3 weeks prior to surgery can decrease the proliferative cell fraction in invasive cancer, intraepithelial neoplasia, and hyperplastic lesions and can decrease tumor size and possibly delay tumor progression of HER-2–positive breast cancer. The antiproliferative effect was greater in ER-negative tumors and in tumors overexpressing PTEN in the cytosol. These results provide the rationale for its use before surgery in rapidly aggressive tumors and in a phase III trial in HER-2–positive intraepithelial neoplasia. Moreover, our findings support the use of presurgical models to efficiently assess the preventive potential of drugs by exploiting the high prevalence of the field cancerization effect present in surgical specimens.

No potential conflicts of interest were disclosed.

This work was supported by an unrestricted educational grant from Glaxo Smith Kline and a contract from the Italian Ministry of Health. Drug and placebo were provided at no cost by Glaxo Smith Kline.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.