- Email: [email protected]
Accurate Prediction and Validation of Response to Endocrine Therapy in Breast Cancer
PMRT improved the 5-year LRRFS marginally from 95% to 98% (not statistically significant), and there was no improvement in the propensity score analysis. The major drawbacks of this study are its retrospective, non-randomized nature and the small number of patients available for subgroup analysis. Randomized studies have shown both a locoregional control and overall survival benefit of PMRT in appropriately selected patients.2 Subset analyses, particularly with small numbers of patients in the subgroups, can be misleading. Therefore, although the improvement in overall survival with trastuzumab is consistent with previous studies, the lack of effect of PMRT on LRRFS in patients receiving trastuzumab may
not be reproducible. Based on this study alone, it would be premature to conclude that LRRFS is high and the benefit of PMRT absent in patients receiving trastuzumab. R. Gopal, MD, PhD
References 1. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology. Breast cancer. Version 1.2016. http://www. nccn.org/professionals/physician_gls/ pdf/breast.pdf. Accessed March 16, 2016. 2. Overgaard M, Christensen JJ, Johansen H, et al. Evaluation of radiotherapy in high-risk breast
cancer patients: report from the Danish Breast Cancer Cooperative Group (DBCG 82) Trial. Int J Radiat Oncol Biol Phys. 1990;19:1121-1124. 3. Buchholz TA, Tucker SL, Masullo L, et al. Predictors of local-regional recurrence after neoadjuvant chemotherapy and mastectomy without radiation. J Clin Oncol. 2002; 20:17-23. 4. Gianni L, Pienkowski T, Im YH, et al. The efficacy and safety of neoadjuvant pertuzumab and trastuzumab in women with locally advanced, inflammatory or early HER2-positive breast cancer (NeoSphere): a randomised multicentre, open label, phase 2 trial. Lancet Oncol. 2012;13: 25-32.
HORMONAL THERAPY Accurate Prediction and Validation of Response to Endocrine Therapy in Breast Cancer Turnbull AK, Arthur LM, Renshaw L, et al (Univ of Edinburgh Cancer Res UK Centre; et al) J Clin Oncol 33:2270-2278, 2015
Purpose.dAromatase inhibitors (AIs) have an established role in the treatment of breast cancer. Response rates are only 50% to 70% in the neoadjuvant setting and lower in advanced disease. Accurate biomarkers are urgently needed to predict response in
these settings and to determine which individuals will benefit from adjuvant AI therapy. Patients and Methods.dPretreatment and on-treatment (after 2 weeks and 3 months) biopsies were obtained from 89 postmenopausal women who had estrogen receptorealpha positive breast cancer and were receiving neoadjuvant letrozole for transcript profiling. Dynamic clinical response was assessed with use of three-dimensional ultrasound measurements. Results.dThe molecular response to letrozole was characterized and a four-gene classifier of clinical response was established (accuracy of 96%) on the basis of the level of
two genes before treatment (one gene [IL6ST] was associated with immune signaling, and the other [NGFRAP1] was associated with apoptosis) and the level of two proliferation genes (ASPM, MCM4) after 2 weeks of therapy. The four-gene signature was found to be 91% accurate in a blinded, completely independent validation data set of patients treated with anastrozole. Matched 2-week on-treatment biopsies were associated with improved predictive power as compared with pretreatment biopsies alone. This signature also significantly predicted recurrence-free survival (P ¼ .029) and breast cancere specific survival (P ¼ .009). We
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
161
demonstrate that the test can also be performed with use of quantitative polymerase chain reaction or immunohistochemistry. Conclusion.dA four-gene predictive model of clinical response to AIs by 2 weeks has been generated and validated. Deregulated immune and apoptotic responses before treatment and cell proliferation that is not reduced 2 weeks after initiation of treatment are functional characteristics of breast tumors that do not respond to AIs. The development, validation, and routine clinical use of prognostic and predictive gene expression profiles in the care of patients with early-stage breast cancer has resulted in the delivery of personalized treatment, allowing many women who either do not need or are unlikely to benefit from chemotherapy to avoid its toxic effects.1 In this regard, perhaps the best predictive factor in all of oncology is the estrogen receptor (ER) for benefit from endocrine therapy (ET). We know that patients with ERnegative tumors have almost no chance of this benefit; therefore, we recommend with great confidence against ET for such patients.2 However, ET is effective for only approximately half of all ERpositive cancers. Moreover, in metastatic disease, even if one ET does not work, others may. Because ET is the cornerstone of treatment in both the adjuvant and the metastatic settings for hormone receptorepositive disease, biomarkers that predict benefit from specific ET modalities are urgently needed. In this article, Turnbull and colleagues aimed to establish the analytic validity and clinical validity of a 4-gene signature predictive of response to neoadjuvant letrozole. The
162
signature included pretreatment expression levels of IL6ST and NGFRAP1 and expression levels of ASPM and MCM4 after 2 weeks of letrozole therapy. The authors reported that this 4-gene signature was 91% accurate in predicting treatment response. Because AIs are a principal component of therapy for patients with hormone receptorpositive disease, the development of an assay that would predict benefit from an AI would be of tremendous clinical value. We want to highlight several interesting observations from this study. The authors discovered that a strong predictor of response to ET was high expression of IL6ST, an immunerelated gene. Indeed, the pretreatment level of IL6ST alone was predictive of response to neoadjuvant letrozole with 85% accuracy. Notably, this gene is also included in the EndoPredict assay, which has demonstrated analytic and clinical validity as a prognostic tool in patients with early-stage breast cancer. Second, applying this particular assay in the clinic would be challenging because 2 of the genes included in this 4-gene signature require measurement 2 weeks after initiation of AI therapy. This necessitates an on-treatment biopsy, which is cumbersome for patients and difficult to implement in practice. The authors did demonstrate that 3 of the 4 genes included in the signature can be assessed via immunohistochemical analysis; whereas immunohistochemical analyses are most certainly easier to use in the clinical setting than a quantitative polymerase chain reaction-based test, this still does not obviate the need for an on-treatment biopsy. Finally, while a gene expression profile predictive of response to neo-
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
adjuvant AI therapy would certainly be of clinical use, an assay predictive of response to AI therapy in the metastatic setting would also be of great value. We suspect that a genomic assay predictive of benefit from AI therapy in metastatic disease would need to incorporate additional or alternative genes as well as the mutational status of estrogen receptor alpha (ESR1). Numerous studies have now demonstrated that mutations in the ligand-binding domain of ESR1 are present at a high frequency in ETrefractory, hormone receptorepositive metastatic disease and that these mutations may confer resistance to AI therapy.3-8 These mutations are not found at a high frequency in primary breast tumors; therefore, the mutational status of ESR1 is not informative when assessing the potential for response to neoadjuvant AI treatment. E. F. Cobain, MD D. F. Hayes, MD
References 1. Hayes DF. Targeting adjuvant chemotherapy: a good idea that needs to be proven! J Clin Oncol. 2012;30: 1264-1267. 2. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Davies C, Godwin J, Gray R, et al. Relevance of breast cancer hormone receptors and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis of randomised trials. Lancet. 2011;378: 771-784. 3. Robinson DR, Wu YM, Vats P, et al. Activating ESR1 mutations in hormone-resistant metastatic breast cancer. Nat Genet. 2013;45: 1446-1451. 4. Li S, Shen D, Shao J, et al. Endocrine-therapy-resistant ESR1
variants revealed by genomic characterization of breast-cancerderived xenografts. Cell Rep. 2013; 26:1116-1130. 5. Merenbakh-Lamin K, Ben-Baruch N, Yeheskel A, et al. D538G mutation in estrogen receptor-a: a novel mechanism for acquired endocrine
resistance in breast cancer. Cancer Res. 2013;73:6856-6864. 6. Toy W, Shen Y, Won H, et al. ESR1 ligand-binding domain mutations in hormone-resistant breast cancer. Nat Genet. 2013;45:1439-1445. 7. Segal CV, Dowsett M. Estrogen receptor mutations in breast
cancerdnew focus on an old target. Clin Cancer Res. 2014;20:1724-1726. 8. Fuqua SA, Gu G, Rechoum Y. Estrogen receptor (ER) a mutations in breast cancer: hidden in plain sight. Breast Cancer Res Treat. 2014;144: 11-19.
SURVIVORSHIP The Long-Term Risk of Upper-Extremity Lymphedema is Two-Fold Higher in Breast Cancer Patients Than in Melanoma Patients Voss RK, Cromwell KD, Chiang Y-J, et al (Univ of Texas MD Anderson Cancer Ctr, Houston; et al) J Surg Oncol 112:834-840, 2015
Background and Objectives.dWe assessed the cumulative incidence, symptoms, and risk factors for upperextremity lymphedema in breast cancer and melanoma patients undergoing sentinel lymph node biopsy or axillary lymph node dissection. Methods.dPatients were recruited preoperatively (time 0) and assessed at 6, 12, and 18 months postoperatively. Limb volume change (LVC) was measured by perometry. Lymphedema was categorized as none, mild (LVC 5e9.9%), or moderate/severe (LVC$10%). Symptoms were assessed with a validated lymphedema instrument. Longitudinal logistic regression analyses were conducted to identify risk factors associated with moderate/ severe lymphedema.
Results.dAmong 205 breast cancer and 144 melanoma patients, the cumulative incidence of moderate/ severe lymphedema at 18 months was 36.5% and 35.0%, respectively. However, in adjusted analyses, factors associated with moderate/severe lymphedema were breast cancer (OR 2.0, P ¼ 0.03), body mass index $30 kg/m2 (OR 1.6, P ¼ 0.04), greater number of lymph nodes removed (OR 1.05, P < 0.01), and longer interval since surgery (OR 2.33 at 18 months, P < 0.01). Conclusions.dLymphedema incidence increased over time in both cohorts. However, the adjusted risk of moderate/severe lymphedema was two-fold higher in breast cancer patients. These results may be attributed to surgical treatment of the primary tumor in the breast and more frequent use of radiation. In this study by Voss and colleagues, the authors applied a method of assessing for secondary lymphedema to melanoma patients that was previously proven in breast cancer patients. Comparing melanoma and breast cancer patients unveils unique features of each subset of secondary
lymphedema. The true incidence of lymphedema in melanoma patients is currently unknown. The study methods included 2 important aspects that are imperative when assessing for edema. First, the researchers obtained baseline measurements on all patients prior to surgery. Because asymmetry naturally exists between limbs, this measurement should be documented and taken into consideration when assessing for increases in limb volumes.1 Second, the use of perometry for the measurements in this study further enhances the results, as the perometer has been proven to be equal in accuracy to the “gold standard,” water displacement, and superior to other volume measurement techniques.2 This study also had quite a large cohort, with 205 breast cancer patients and 144 melanoma patients. However, the discrepancy in the number of participants for each disease subset could skew the overall results of the comparison: the breast cancer cohort likely represents a greater portion of the general population, whereas the melanoma cohort may not. This further becomes problematic when accounting for the attrition rate.
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
163
not be reproducible. Based on this study alone, it would be premature to conclude that LRRFS is high and the benefit of PMRT absent in patients receiving trastuzumab. R. Gopal, MD, PhD
References 1. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology. Breast cancer. Version 1.2016. http://www. nccn.org/professionals/physician_gls/ pdf/breast.pdf. Accessed March 16, 2016. 2. Overgaard M, Christensen JJ, Johansen H, et al. Evaluation of radiotherapy in high-risk breast
cancer patients: report from the Danish Breast Cancer Cooperative Group (DBCG 82) Trial. Int J Radiat Oncol Biol Phys. 1990;19:1121-1124. 3. Buchholz TA, Tucker SL, Masullo L, et al. Predictors of local-regional recurrence after neoadjuvant chemotherapy and mastectomy without radiation. J Clin Oncol. 2002; 20:17-23. 4. Gianni L, Pienkowski T, Im YH, et al. The efficacy and safety of neoadjuvant pertuzumab and trastuzumab in women with locally advanced, inflammatory or early HER2-positive breast cancer (NeoSphere): a randomised multicentre, open label, phase 2 trial. Lancet Oncol. 2012;13: 25-32.
HORMONAL THERAPY Accurate Prediction and Validation of Response to Endocrine Therapy in Breast Cancer Turnbull AK, Arthur LM, Renshaw L, et al (Univ of Edinburgh Cancer Res UK Centre; et al) J Clin Oncol 33:2270-2278, 2015
Purpose.dAromatase inhibitors (AIs) have an established role in the treatment of breast cancer. Response rates are only 50% to 70% in the neoadjuvant setting and lower in advanced disease. Accurate biomarkers are urgently needed to predict response in
these settings and to determine which individuals will benefit from adjuvant AI therapy. Patients and Methods.dPretreatment and on-treatment (after 2 weeks and 3 months) biopsies were obtained from 89 postmenopausal women who had estrogen receptorealpha positive breast cancer and were receiving neoadjuvant letrozole for transcript profiling. Dynamic clinical response was assessed with use of three-dimensional ultrasound measurements. Results.dThe molecular response to letrozole was characterized and a four-gene classifier of clinical response was established (accuracy of 96%) on the basis of the level of
two genes before treatment (one gene [IL6ST] was associated with immune signaling, and the other [NGFRAP1] was associated with apoptosis) and the level of two proliferation genes (ASPM, MCM4) after 2 weeks of therapy. The four-gene signature was found to be 91% accurate in a blinded, completely independent validation data set of patients treated with anastrozole. Matched 2-week on-treatment biopsies were associated with improved predictive power as compared with pretreatment biopsies alone. This signature also significantly predicted recurrence-free survival (P ¼ .029) and breast cancere specific survival (P ¼ .009). We
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
161
demonstrate that the test can also be performed with use of quantitative polymerase chain reaction or immunohistochemistry. Conclusion.dA four-gene predictive model of clinical response to AIs by 2 weeks has been generated and validated. Deregulated immune and apoptotic responses before treatment and cell proliferation that is not reduced 2 weeks after initiation of treatment are functional characteristics of breast tumors that do not respond to AIs. The development, validation, and routine clinical use of prognostic and predictive gene expression profiles in the care of patients with early-stage breast cancer has resulted in the delivery of personalized treatment, allowing many women who either do not need or are unlikely to benefit from chemotherapy to avoid its toxic effects.1 In this regard, perhaps the best predictive factor in all of oncology is the estrogen receptor (ER) for benefit from endocrine therapy (ET). We know that patients with ERnegative tumors have almost no chance of this benefit; therefore, we recommend with great confidence against ET for such patients.2 However, ET is effective for only approximately half of all ERpositive cancers. Moreover, in metastatic disease, even if one ET does not work, others may. Because ET is the cornerstone of treatment in both the adjuvant and the metastatic settings for hormone receptorepositive disease, biomarkers that predict benefit from specific ET modalities are urgently needed. In this article, Turnbull and colleagues aimed to establish the analytic validity and clinical validity of a 4-gene signature predictive of response to neoadjuvant letrozole. The
162
signature included pretreatment expression levels of IL6ST and NGFRAP1 and expression levels of ASPM and MCM4 after 2 weeks of letrozole therapy. The authors reported that this 4-gene signature was 91% accurate in predicting treatment response. Because AIs are a principal component of therapy for patients with hormone receptorpositive disease, the development of an assay that would predict benefit from an AI would be of tremendous clinical value. We want to highlight several interesting observations from this study. The authors discovered that a strong predictor of response to ET was high expression of IL6ST, an immunerelated gene. Indeed, the pretreatment level of IL6ST alone was predictive of response to neoadjuvant letrozole with 85% accuracy. Notably, this gene is also included in the EndoPredict assay, which has demonstrated analytic and clinical validity as a prognostic tool in patients with early-stage breast cancer. Second, applying this particular assay in the clinic would be challenging because 2 of the genes included in this 4-gene signature require measurement 2 weeks after initiation of AI therapy. This necessitates an on-treatment biopsy, which is cumbersome for patients and difficult to implement in practice. The authors did demonstrate that 3 of the 4 genes included in the signature can be assessed via immunohistochemical analysis; whereas immunohistochemical analyses are most certainly easier to use in the clinical setting than a quantitative polymerase chain reaction-based test, this still does not obviate the need for an on-treatment biopsy. Finally, while a gene expression profile predictive of response to neo-
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
adjuvant AI therapy would certainly be of clinical use, an assay predictive of response to AI therapy in the metastatic setting would also be of great value. We suspect that a genomic assay predictive of benefit from AI therapy in metastatic disease would need to incorporate additional or alternative genes as well as the mutational status of estrogen receptor alpha (ESR1). Numerous studies have now demonstrated that mutations in the ligand-binding domain of ESR1 are present at a high frequency in ETrefractory, hormone receptorepositive metastatic disease and that these mutations may confer resistance to AI therapy.3-8 These mutations are not found at a high frequency in primary breast tumors; therefore, the mutational status of ESR1 is not informative when assessing the potential for response to neoadjuvant AI treatment. E. F. Cobain, MD D. F. Hayes, MD
References 1. Hayes DF. Targeting adjuvant chemotherapy: a good idea that needs to be proven! J Clin Oncol. 2012;30: 1264-1267. 2. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG), Davies C, Godwin J, Gray R, et al. Relevance of breast cancer hormone receptors and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis of randomised trials. Lancet. 2011;378: 771-784. 3. Robinson DR, Wu YM, Vats P, et al. Activating ESR1 mutations in hormone-resistant metastatic breast cancer. Nat Genet. 2013;45: 1446-1451. 4. Li S, Shen D, Shao J, et al. Endocrine-therapy-resistant ESR1
variants revealed by genomic characterization of breast-cancerderived xenografts. Cell Rep. 2013; 26:1116-1130. 5. Merenbakh-Lamin K, Ben-Baruch N, Yeheskel A, et al. D538G mutation in estrogen receptor-a: a novel mechanism for acquired endocrine
resistance in breast cancer. Cancer Res. 2013;73:6856-6864. 6. Toy W, Shen Y, Won H, et al. ESR1 ligand-binding domain mutations in hormone-resistant breast cancer. Nat Genet. 2013;45:1439-1445. 7. Segal CV, Dowsett M. Estrogen receptor mutations in breast
cancerdnew focus on an old target. Clin Cancer Res. 2014;20:1724-1726. 8. Fuqua SA, Gu G, Rechoum Y. Estrogen receptor (ER) a mutations in breast cancer: hidden in plain sight. Breast Cancer Res Treat. 2014;144: 11-19.
SURVIVORSHIP The Long-Term Risk of Upper-Extremity Lymphedema is Two-Fold Higher in Breast Cancer Patients Than in Melanoma Patients Voss RK, Cromwell KD, Chiang Y-J, et al (Univ of Texas MD Anderson Cancer Ctr, Houston; et al) J Surg Oncol 112:834-840, 2015
Background and Objectives.dWe assessed the cumulative incidence, symptoms, and risk factors for upperextremity lymphedema in breast cancer and melanoma patients undergoing sentinel lymph node biopsy or axillary lymph node dissection. Methods.dPatients were recruited preoperatively (time 0) and assessed at 6, 12, and 18 months postoperatively. Limb volume change (LVC) was measured by perometry. Lymphedema was categorized as none, mild (LVC 5e9.9%), or moderate/severe (LVC$10%). Symptoms were assessed with a validated lymphedema instrument. Longitudinal logistic regression analyses were conducted to identify risk factors associated with moderate/ severe lymphedema.
Results.dAmong 205 breast cancer and 144 melanoma patients, the cumulative incidence of moderate/ severe lymphedema at 18 months was 36.5% and 35.0%, respectively. However, in adjusted analyses, factors associated with moderate/severe lymphedema were breast cancer (OR 2.0, P ¼ 0.03), body mass index $30 kg/m2 (OR 1.6, P ¼ 0.04), greater number of lymph nodes removed (OR 1.05, P < 0.01), and longer interval since surgery (OR 2.33 at 18 months, P < 0.01). Conclusions.dLymphedema incidence increased over time in both cohorts. However, the adjusted risk of moderate/severe lymphedema was two-fold higher in breast cancer patients. These results may be attributed to surgical treatment of the primary tumor in the breast and more frequent use of radiation. In this study by Voss and colleagues, the authors applied a method of assessing for secondary lymphedema to melanoma patients that was previously proven in breast cancer patients. Comparing melanoma and breast cancer patients unveils unique features of each subset of secondary
lymphedema. The true incidence of lymphedema in melanoma patients is currently unknown. The study methods included 2 important aspects that are imperative when assessing for edema. First, the researchers obtained baseline measurements on all patients prior to surgery. Because asymmetry naturally exists between limbs, this measurement should be documented and taken into consideration when assessing for increases in limb volumes.1 Second, the use of perometry for the measurements in this study further enhances the results, as the perometer has been proven to be equal in accuracy to the “gold standard,” water displacement, and superior to other volume measurement techniques.2 This study also had quite a large cohort, with 205 breast cancer patients and 144 melanoma patients. However, the discrepancy in the number of participants for each disease subset could skew the overall results of the comparison: the breast cancer cohort likely represents a greater portion of the general population, whereas the melanoma cohort may not. This further becomes problematic when accounting for the attrition rate.
Breast Diseases: A Year BookÒ Quarterly Vol 27 No 2 2016
163