Karen J. Taylor
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Karen J. Taylor has a total of 115 co-authors. Here are the top co-authors by number of shared publications:
- Jane Bayani
Diagnostic Development, Ontario Institute for Cancer Research, Toronto, ON, Canada
15 shared publications - John M. S. Bartlett
Cancer Research UK Scotland Centre, University of Edinburgh, Edinburgh, United Kingdom
15 shared publications - Elizabeth Mallon
Department of Pathology, NHS Greater Glasgow and Clyde, Glasgow, Scotland, United Kingdom
14 shared publications - Annette Hasenburg
Department of Obstetrics and Gynecology, University Medical Center of the Johannes Gutenberg University Mainz, Langenbeckstraße, 1, 55131 Mainz, Germany
13 shared publications - Christos Markopoulos
Medical School, National and Kapodistrian University of Athens, Athens, Greece
13 shared publications - Cornelis J. H. van de Velde
Department of Surgery, Leiden University Medical Center, 2333 ZA Leiden, The Netherlands
13 shared publications
Latest Publications
- Open Access
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Primary Analysis of the OPTIMA Phase III Randomised Non-Inferiority Trial of Test-Directed Chemotherapy in Patients With High Clinical Risk ER-Positive HER2-Negative Early Breast Cancer
Background: Evidence supporting tumour gene expression test use to assist adjuvant chemotherapy decisions in high-risk early breast cancer is limited. We investigated the safety of treatment decisions made using a 50-gene test in this population.Methods: OPTIMA [ISRCTN42400492], an international randomised trial, recruited patients aged 40 or older with oestrogen receptor-positive HER2-negative early breast cancer and 0-9 involved axillary lymph nodes. Randomisation was performed centrally (1:1) between standard chemo-endocrine therapy and test-directed treatment; participants with a 50-gene tumour Risk of Recurrence (ROR) score >60 were allocated chemo-endocrine therapy whilst those with lower scores received endocrine therapy alone; for premenopausal women this included ovarian function suppression. The trial was analysed for non-inferiority of Invasive Breast Cancer Free Survival (IBCFS) in the per-protocol population with a 3% margin. Patients were recruited from 171 hospitals in 6 countries. Recruitment of an expanded premenopausal cohort is ongoing.Findings: 4429 patients underwent randomisation between January 16, 2017, and December 12, 2025; 4153 (93.8%) were included in the per-protocol analysis, 2059 in the control arm and 2094 in the test-directed arm. With 301 IBCFS events, the 5-year rate in the control arm was 91.8% [95% confidence interval (CI) 90.1-93.2%] and 90.3% [95% CI 88.5-91.8%] in the test-directed arm, adjusted hazard ratio (HR) 1.03 [90% CI 0.85-1.25], p(non-inferiority)=0.006. Similar results were seen in the 2779 (66.9%) per-protocol patient subgroup with ROR score ≤60 tumours (HR 1.06 [90% CI 0.81-1.40), p(non-inferiority]=0.003) and all other predefined subgroups including menopausal status, and lymph node and TNM stage and in the ITT population.Interpretation: The 50-gene test can be used to make safe adjuvant chemotherapy decisions, including for premenopausal women treated with ovarian function suppression and patients with high lymph node burden or stage IIIA disease. Funding UK National Institute for Health and Care Research Health Technology Assessment Programme and others.
- 2 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPrimary Analysis of the OPTIMA Phase III Randomised Non-Inferiority Trial of Test-Directed Chemotherapy in Patients With High Clinical Risk ER-Positive HER2-Negative Early Breast Cancer
Background: Evidence supporting tumour gene expression test use to assist adjuvant chemotherapy decisions in high-risk early breast cancer is limited. We investigated the safety of treatment decisions made using a 50-gene test in this population.Methods: OPTIMA [ISRCTN42400492], an international randomised trial, recruited patients aged 40 or older with oestrogen receptor-positive HER2-negative early breast cancer and 0-9 involved axillary lymph nodes. Randomisation was performed centrally (1:1) between standard chemo-endocrine therapy and test-directed treatment; participants with a 50-gene tumour Risk of Recurrence (ROR) score >60 were allocated chemo-endocrine therapy whilst those with lower scores received endocrine therapy alone; for premenopausal women this included ovarian function suppression. The trial was analysed for non-inferiority of Invasive Breast Cancer Free Survival (IBCFS) in the per-protocol population with a 3% margin. Patients were recruited from 171 hospitals in 6 countries. Recruitment of an expanded premenopausal cohort is ongoing.Findings: 4429 patients underwent randomisation between January 16, 2017, and December 12, 2025; 4153 (93.8%) were included in the per-protocol analysis, 2059 in the control arm and 2094 in the test-directed arm. With 301 IBCFS events, the 5-year rate in the control arm was 91.8% [95% confidence interval (CI) 90.1-93.2%] and 90.3% [95% CI 88.5-91.8%] in the test-directed arm, adjusted hazard ratio (HR) 1.03 [90% CI 0.85-1.25], p(non-inferiority)=0.006. Similar results were seen in the 2779 (66.9%) per-protocol patient subgroup with ROR score ≤60 tumours (HR 1.06 [90% CI 0.81-1.40), p(non-inferiority]=0.003) and all other predefined subgroups including menopausal status, and lymph node and TNM stage and in the ITT population.Interpretation: The 50-gene test can be used to make safe adjuvant chemotherapy decisions, including for premenopausal women treated with ovarian function suppression and patients with high lymph node burden or stage IIIA disease. Funding UK National Institute for Health and Care Research Health Technology Assessment Programme and others.
- Open Access
- 14 April 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28, e88665
Transfer Learning and Machine Learning for Training Five-Year Survival Prognostic Models in Early Breast Cancer: Development and Validation Study
Background: Prognostic information is essential for decision-making in breast cancer management. In recent years, trials and clinical practice have emphasized genomic prognostication tools, despite clinicopathological methods being more affordable and accessible. PREDICT v3 is one such tool with promising results across cohorts. Advances in machine learning (ML), transfer learning, and ensemble methods provide opportunities to enhance these approaches, especially where missing data and model assumptions differ across diverse populations. Objective: This study evaluates the potential to improve survival prognostication in breast cancer. More precisely, we compare de novo ML, transfer learning from the pretrained prognostication model PREDICT v3, and a stacked ensemble approach. Methods: Data from the MA.27 trial (NCT00066573) were used for model training, with external validation on data from the Tamoxifen Exemestane Adjuvant Multinational trial (NCT00279448 and NCT00032136) and a US Surveillance, Epidemiology, and End Results cohort. Transfer learning was applied by re-estimating the parameters of the pretrained prognostic tool PREDICT v3. De novo ML included random survival forests and extreme gradient boosting, and the ensemble was implemented using weighted linear stacking of model predictions. Internal and external validation was assessed in terms of the integrated calibration index and discrimination. Shapley Additive Explanations values were used to explain model predictions and decision-curve analysis to facilitate the interpretation of performance differences. Results: Transfer learning, de novo random survival forest, and the stacked ensemble improved calibration in MA.27 over the pretrained model (integrated calibration index reduced from 0.042 in PREDICT v3 to ≤0.007) while discrimination remained comparable (AUROC increased from 0.738 in PREDICT v3 to 0.744-0.799). In decision-curve analysis, these approaches demonstrated consistently positive net benefit across clinically relevant thresholds, while PREDICT v3 lost net benefit beyond 7.5% predicted risk. Invalid PREDICT v3 predictions were observed in 23.8% to 25.8% of MA.27 individuals due to missing information. In contrast, ML models and the stacked ensemble predicted survival despite missing data. Across all models, patient age, nodal status, pathological grading, and tumor size had the highest Shapley Additive Explanations values, indicating their importance for survival prognostication. External validation in the US Surveillance, Epidemiology, and End Results cohort confirmed the benefits of transfer learning, RSF, and ensemble in terms of calibration while maintaining discrimination at comparable levels. In contrast, generalizability was limited in the Tamoxifen Exemestane Adjuvant Multinational trial, a cohort with a substantially different distribution of clinicopathological characteristics. Conclusions: This study demonstrates that transfer learning, de novo RSF, and a stacked ensemble can improve prognostication compared with the pretrained PREDICT v3, particularly in the presence of missing or uncertain inputs. Transportability may be limited in cohorts with different clinicopathological profiles, requiring local validation before clinical deployment. Ultimately, better survival estimation can provide more meaningful guidance in breast cancer care. Trial Registration: ClinicalTrials.gov NCT00066573; https://clinicaltrials.gov/study/NCT00066573, NCT00279448; https://clinicaltrials.gov/study/NCT00279448, NCT00032136; https://clinicaltrials.gov/study/NCT00032136
- 14 April 2026
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 28, e88665
Open AccessTransfer Learning and Machine Learning for Training Five-Year Survival Prognostic Models in Early Breast Cancer: Development and Validation Study
Background: Prognostic information is essential for decision-making in breast cancer management. In recent years, trials and clinical practice have emphasized genomic prognostication tools, despite clinicopathological methods being more affordable and accessible. PREDICT v3 is one such tool with promising results across cohorts. Advances in machine learning (ML), transfer learning, and ensemble methods provide opportunities to enhance these approaches, especially where missing data and model assumptions differ across diverse populations. Objective: This study evaluates the potential to improve survival prognostication in breast cancer. More precisely, we compare de novo ML, transfer learning from the pretrained prognostication model PREDICT v3, and a stacked ensemble approach. Methods: Data from the MA.27 trial (NCT00066573) were used for model training, with external validation on data from the Tamoxifen Exemestane Adjuvant Multinational trial (NCT00279448 and NCT00032136) and a US Surveillance, Epidemiology, and End Results cohort. Transfer learning was applied by re-estimating the parameters of the pretrained prognostic tool PREDICT v3. De novo ML included random survival forests and extreme gradient boosting, and the ensemble was implemented using weighted linear stacking of model predictions. Internal and external validation was assessed in terms of the integrated calibration index and discrimination. Shapley Additive Explanations values were used to explain model predictions and decision-curve analysis to facilitate the interpretation of performance differences. Results: Transfer learning, de novo random survival forest, and the stacked ensemble improved calibration in MA.27 over the pretrained model (integrated calibration index reduced from 0.042 in PREDICT v3 to ≤0.007) while discrimination remained comparable (AUROC increased from 0.738 in PREDICT v3 to 0.744-0.799). In decision-curve analysis, these approaches demonstrated consistently positive net benefit across clinically relevant thresholds, while PREDICT v3 lost net benefit beyond 7.5% predicted risk. Invalid PREDICT v3 predictions were observed in 23.8% to 25.8% of MA.27 individuals due to missing information. In contrast, ML models and the stacked ensemble predicted survival despite missing data. Across all models, patient age, nodal status, pathological grading, and tumor size had the highest Shapley Additive Explanations values, indicating their importance for survival prognostication. External validation in the US Surveillance, Epidemiology, and End Results cohort confirmed the benefits of transfer learning, RSF, and ensemble in terms of calibration while maintaining discrimination at comparable levels. In contrast, generalizability was limited in the Tamoxifen Exemestane Adjuvant Multinational trial, a cohort with a substantially different distribution of clinicopathological characteristics. Conclusions: This study demonstrates that transfer learning, de novo RSF, and a stacked ensemble can improve prognostication compared with the pretrained PREDICT v3, particularly in the presence of missing or uncertain inputs. Transportability may be limited in cohorts with different clinicopathological profiles, requiring local validation before clinical deployment. Ultimately, better survival estimation can provide more meaningful guidance in breast cancer care. Trial Registration: ClinicalTrials.gov NCT00066573; https://clinicaltrials.gov/study/NCT00066573, NCT00279448; https://clinicaltrials.gov/study/NCT00279448, NCT00032136; https://clinicaltrials.gov/study/NCT00032136
- Open Access
- 29 November 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Transfer Learning and Machine Learning for Training Five-Year Survival Prognostic Models in Early Breast Cancer: Development and Validation Study (Preprint)
BACKGROUND Prognostic information is essential for decision-making in breast cancer management. In recent years, trials and clinical practice have emphasized genomic prognostication tools, despite clinicopathological methods being more affordable and accessible. PREDICT v3 is one such tool with promising results across cohorts. Advances in machine learning (ML), transfer learning, and ensemble methods provide opportunities to enhance these approaches, especially where missing data and model assumptions differ across diverse populations. OBJECTIVE This study evaluates the potential to improve survival prognostication in breast cancer. More precisely, we compare de novo ML, transfer learning from the pretrained prognostication model PREDICT v3, and a stacked ensemble approach. METHODS Data from the MA.27 trial (NCT00066573) were used for model training, with external validation on data from the Tamoxifen Exemestane Adjuvant Multinational trial (NCT00279448 and NCT00032136) and a US Surveillance, Epidemiology, and End Results cohort. Transfer learning was applied by re-estimating the parameters of the pretrained prognostic tool PREDICT v3. De novo ML included random survival forests and extreme gradient boosting, and the ensemble was implemented using weighted linear stacking of model predictions. Internal and external validation was assessed in terms of the integrated calibration index and discrimination. Shapley Additive Explanations values were used to explain model predictions and decision-curve analysis to facilitate the interpretation of performance differences. RESULTS Transfer learning, de novo random survival forest, and the stacked ensemble improved calibration in MA.27 over the pretrained model (integrated calibration index reduced from 0.042 in PREDICT v3 to ≤0.007) while discrimination remained comparable (AUROC increased from 0.738 in PREDICT v3 to 0.744-0.799). In decision-curve analysis, these approaches demonstrated consistently positive net benefit across clinically relevant thresholds, while PREDICT v3 lost net benefit beyond 7.5% predicted risk. Invalid PREDICT v3 predictions were observed in 23.8% to 25.8% of MA.27 individuals due to missing information. In contrast, ML models and the stacked ensemble predicted survival despite missing data. Across all models, patient age, nodal status, pathological grading, and tumor size had the highest Shapley Additive Explanations values, indicating their importance for survival prognostication. External validation in the US Surveillance, Epidemiology, and End Results cohort confirmed the benefits of transfer learning, RSF, and ensemble in terms of calibration while maintaining discrimination at comparable levels. In contrast, generalizability was limited in the Tamoxifen Exemestane Adjuvant Multinational trial, a cohort with a substantially different distribution of clinicopathological characteristics. CONCLUSIONS This study demonstrates that transfer learning, de novo RSF, and a stacked ensemble can improve prognostication compared with the pretrained PREDICT v3, particularly in the presence of missing or uncertain inputs. Transportability may be limited in cohorts with different clinicopathological profiles, requiring local validation before clinical deployment. Ultimately, better survival estimation can provide more meaningful guidance in breast cancer care. CLINICALTRIAL ClinicalTrials.gov NCT00066573; https://clinicaltrials.gov/study/NCT00066573, NCT00279448; https://clinicaltrials.gov/study/NCT00279448, NCT00032136; https://clinicaltrials.gov/study/NCT00032136
- 29 November 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessTransfer Learning and Machine Learning for Training Five-Year Survival Prognostic Models in Early Breast Cancer: Development and Validation Study (Preprint)
BACKGROUND Prognostic information is essential for decision-making in breast cancer management. In recent years, trials and clinical practice have emphasized genomic prognostication tools, despite clinicopathological methods being more affordable and accessible. PREDICT v3 is one such tool with promising results across cohorts. Advances in machine learning (ML), transfer learning, and ensemble methods provide opportunities to enhance these approaches, especially where missing data and model assumptions differ across diverse populations. OBJECTIVE This study evaluates the potential to improve survival prognostication in breast cancer. More precisely, we compare de novo ML, transfer learning from the pretrained prognostication model PREDICT v3, and a stacked ensemble approach. METHODS Data from the MA.27 trial (NCT00066573) were used for model training, with external validation on data from the Tamoxifen Exemestane Adjuvant Multinational trial (NCT00279448 and NCT00032136) and a US Surveillance, Epidemiology, and End Results cohort. Transfer learning was applied by re-estimating the parameters of the pretrained prognostic tool PREDICT v3. De novo ML included random survival forests and extreme gradient boosting, and the ensemble was implemented using weighted linear stacking of model predictions. Internal and external validation was assessed in terms of the integrated calibration index and discrimination. Shapley Additive Explanations values were used to explain model predictions and decision-curve analysis to facilitate the interpretation of performance differences. RESULTS Transfer learning, de novo random survival forest, and the stacked ensemble improved calibration in MA.27 over the pretrained model (integrated calibration index reduced from 0.042 in PREDICT v3 to ≤0.007) while discrimination remained comparable (AUROC increased from 0.738 in PREDICT v3 to 0.744-0.799). In decision-curve analysis, these approaches demonstrated consistently positive net benefit across clinically relevant thresholds, while PREDICT v3 lost net benefit beyond 7.5% predicted risk. Invalid PREDICT v3 predictions were observed in 23.8% to 25.8% of MA.27 individuals due to missing information. In contrast, ML models and the stacked ensemble predicted survival despite missing data. Across all models, patient age, nodal status, pathological grading, and tumor size had the highest Shapley Additive Explanations values, indicating their importance for survival prognostication. External validation in the US Surveillance, Epidemiology, and End Results cohort confirmed the benefits of transfer learning, RSF, and ensemble in terms of calibration while maintaining discrimination at comparable levels. In contrast, generalizability was limited in the Tamoxifen Exemestane Adjuvant Multinational trial, a cohort with a substantially different distribution of clinicopathological characteristics. CONCLUSIONS This study demonstrates that transfer learning, de novo RSF, and a stacked ensemble can improve prognostication compared with the pretrained PREDICT v3, particularly in the presence of missing or uncertain inputs. Transportability may be limited in cohorts with different clinicopathological profiles, requiring local validation before clinical deployment. Ultimately, better survival estimation can provide more meaningful guidance in breast cancer care. CLINICALTRIAL ClinicalTrials.gov NCT00066573; https://clinicaltrials.gov/study/NCT00066573, NCT00279448; https://clinicaltrials.gov/study/NCT00279448, NCT00032136; https://clinicaltrials.gov/study/NCT00032136
- 23 June 2025
- Published by Springer Nature in Breast Cancer Research and Treatment
- Vol. 213 (1) , 51-61
Adjunctive statistical standardization of quantitated adjuvant HER2 and ultra-low HER2 in Canadian Cancer Trials Group MA.27 trial of exemestane versus anastrozole
Statistically standardized estrogen receptor (ER) and progesterone receptor (PgR) differentiated prognosis. Here we examined statistically standardized human epidermal growth receptor 2 (HER2). CCTG MA.27 (NCT00066573) was an adjuvant phase III trial of exemestane versus anastrozole in postmenopausal women with ER + and/or PgR + tumors. We centrally quantitated machine-image immunohistochemical HER2, defined American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) dual-probe FISH HER2/CEP17 categories, determined ultra-low HER2 (IHC 0 with (0,10%] 1 + stain), and standardized HER2 to mean 0, standard deviation 1. Univariate distant disease-free survival (DDFS) was described with Kaplan-Meier plots and examined with Wilcoxon (Peto-Prentice) test statistic. Adjusted Cox multivariable regressions 2-sided Wald tests had nominal significance p < 0.05. Of 7576 women, 2900 had ER results; 2726, PgR; 2680, HER2; and 2325, ER/PgR/HER2 for multivariable investigations. ASCO/CAP categorization significantly differentiated univariate DDFS (p = 0.01), although not values of IHC 0 (N = 864) and ultra-low HER2 (N = 1143). Statistical standardization did not differentiate univariate DDFS (p = 0.08-0.27); however, (natural logarithm-) standardized values ≤ - 1.0 (ultra-low 1 + /2 + /3 + HSCORE, or % + , < 0.1) were similar to > 1.0 (HSCORE > 19; % + > 14). Neither ASCO/CAP, nor statistically standardized, ER (p = 0.65-0.94) or HER2 (p = 0.20-0.97) were associated with DDFS in models with PgR; higher PgR had better DDFS (p ≤ .003). ASCO/CAP categories significantly differentiated DDFS, while statistical standardization did not. Patients with ultra-low HER2 and IHC 0 without stain had similar 5-year DDFS, while standardization indicated similar prognosis for very low 1 + /2 + /3 + and highest HER2 stain. We caution about assessment of ultra-low, or very low, HER2 due to HER2 assay dynamic range.
- 23 June 2025
- Published by Springer Nature in Breast Cancer Research and Treatment
- Vol. 213 (1) , 51-61
Adjunctive statistical standardization of quantitated adjuvant HER2 and ultra-low HER2 in Canadian Cancer Trials Group MA.27 trial of exemestane versus anastrozole
Statistically standardized estrogen receptor (ER) and progesterone receptor (PgR) differentiated prognosis. Here we examined statistically standardized human epidermal growth receptor 2 (HER2). CCTG MA.27 (NCT00066573) was an adjuvant phase III trial of exemestane versus anastrozole in postmenopausal women with ER + and/or PgR + tumors. We centrally quantitated machine-image immunohistochemical HER2, defined American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) dual-probe FISH HER2/CEP17 categories, determined ultra-low HER2 (IHC 0 with (0,10%] 1 + stain), and standardized HER2 to mean 0, standard deviation 1. Univariate distant disease-free survival (DDFS) was described with Kaplan-Meier plots and examined with Wilcoxon (Peto-Prentice) test statistic. Adjusted Cox multivariable regressions 2-sided Wald tests had nominal significance p < 0.05. Of 7576 women, 2900 had ER results; 2726, PgR; 2680, HER2; and 2325, ER/PgR/HER2 for multivariable investigations. ASCO/CAP categorization significantly differentiated univariate DDFS (p = 0.01), although not values of IHC 0 (N = 864) and ultra-low HER2 (N = 1143). Statistical standardization did not differentiate univariate DDFS (p = 0.08-0.27); however, (natural logarithm-) standardized values ≤ - 1.0 (ultra-low 1 + /2 + /3 + HSCORE, or % + , < 0.1) were similar to > 1.0 (HSCORE > 19; % + > 14). Neither ASCO/CAP, nor statistically standardized, ER (p = 0.65-0.94) or HER2 (p = 0.20-0.97) were associated with DDFS in models with PgR; higher PgR had better DDFS (p ≤ .003). ASCO/CAP categories significantly differentiated DDFS, while statistical standardization did not. Patients with ultra-low HER2 and IHC 0 without stain had similar 5-year DDFS, while standardization indicated similar prognosis for very low 1 + /2 + /3 + and highest HER2 stain. We caution about assessment of ultra-low, or very low, HER2 due to HER2 assay dynamic range.
- 7 May 2025
- Published by Oxford University Press (OUP) in JNCI: Journal of the National Cancer Institute
- Vol. 117 (8) , 1745-1746
Response to Li, Liu, Duan, et al.
- 7 May 2025
- Published by Oxford University Press (OUP) in JNCI: Journal of the National Cancer Institute
- Vol. 117 (8) , 1745-1746
Response to Li, Liu, Duan, et al.
- Open Access
- 18 October 2024
- Published by Oxford University Press (OUP) in JNCI: Journal of the National Cancer Institute
- Vol. 117 (3) , 486-495
Validation of a breast cancer assay for radiotherapy omission: an individual participant data meta-analysis
There are currently no molecular tests to identify individual breast cancers where radiotherapy (RT) offers no benefit. Profile for the Omission of Local Adjuvant Radiotherapy (POLAR) is a 16-gene molecular signature developed to identify low risk cancers where RT will not further reduce recurrence rates. An individual participant data meta-analysis was performed in 623 cases of node-negative ER+/HER2-negative early breast cancer enrolled in three RT randomized trials for whom primary tumor material was available for analysis. A Cox proportional hazards model on time to locoregional recurrence (LRR) was used to test the interaction between POLAR score and RT. 429 (69%) patients’ tumors had a high POLAR score and 194 (31%) had a low score. Patients with high POLAR score had, in the absence of RT, a 10-year cumulative incidence of LRR: 20% (15%-26%) vs 5% (2%-11%) for those with a low score. Patients with a high POLAR score had a large benefit from RT (hazard ratio [HR] for RT vs no RT: 0.37 [0.23-0.60], p < .001). In contrast, there was no evidence of benefit from RT for patients with a low POLAR score (HR: 0.92 [0.42-2.02], p = .832). The test for interaction between RT and POLAR was statistically significant (p = .022). POLAR is not only prognostic for locoregional recurrence but also predictive of benefit from radiotherapy in selected patients. Patients ≥ 50 years with ER+/HER2-negative disease and a low POLAR score could consider omitting adjuvant RT. Further validation in contemporary clinical cohorts is required.
- 18 October 2024
- Published by Oxford University Press (OUP) in JNCI: Journal of the National Cancer Institute
- Vol. 117 (3) , 486-495
Open AccessValidation of a breast cancer assay for radiotherapy omission: an individual participant data meta-analysis
There are currently no molecular tests to identify individual breast cancers where radiotherapy (RT) offers no benefit. Profile for the Omission of Local Adjuvant Radiotherapy (POLAR) is a 16-gene molecular signature developed to identify low risk cancers where RT will not further reduce recurrence rates. An individual participant data meta-analysis was performed in 623 cases of node-negative ER+/HER2-negative early breast cancer enrolled in three RT randomized trials for whom primary tumor material was available for analysis. A Cox proportional hazards model on time to locoregional recurrence (LRR) was used to test the interaction between POLAR score and RT. 429 (69%) patients’ tumors had a high POLAR score and 194 (31%) had a low score. Patients with high POLAR score had, in the absence of RT, a 10-year cumulative incidence of LRR: 20% (15%-26%) vs 5% (2%-11%) for those with a low score. Patients with a high POLAR score had a large benefit from RT (hazard ratio [HR] for RT vs no RT: 0.37 [0.23-0.60], p < .001). In contrast, there was no evidence of benefit from RT for patients with a low POLAR score (HR: 0.92 [0.42-2.02], p = .832). The test for interaction between RT and POLAR was statistically significant (p = .022). POLAR is not only prognostic for locoregional recurrence but also predictive of benefit from radiotherapy in selected patients. Patients ≥ 50 years with ER+/HER2-negative disease and a low POLAR score could consider omitting adjuvant RT. Further validation in contemporary clinical cohorts is required.
- 20 August 2024
- Published by American Society of Clinical Oncology (ASCO) in Journal of Clinical Oncology
- Vol. 42 (24) , 2887-2898
Adjunctive Statistical Standardization of Adjuvant Estrogen Receptor and Progesterone Receptor in Canadian Cancer Trials Group MA.27 Postmenopausal Breast Cancer Trial of Exemestane Versus Anastrozole
PURPOSE: ASCO/College of American Pathologists guidelines recommend reporting estrogen receptor (ER) and progesterone receptor (PgR) as positive with (1%-100%) staining. Statistically standardized quantitated positivity could indicate differential associations of positivity with breast cancer outcomes. METHODS: MA.27 (ClinicalTrials.gov identifier: NCT00066573 ) was a phase III adjuvant trial of exemestane versus anastrozole in postmenopausal women with early-stage breast cancer. Immunochemistry ER and PgR HSCORE and % positivity (%+) were centrally assessed by machine image quantitation and statistically standardized to mean 0 and standard deviation (SD) 1 after Box-Cox variance stabilization transformations of square for ER; for PgR, (1) natural logarithm (0.1 added to 0 HSCOREs and 0%+) and (2) square root. Our primary end point was MA.27 distant disease-free survival (DDFS) at a median 4.1-year follow-up, and secondary end point was event-free survival (EFS). Univariate survival with cut points at SDs about a mean of 0 (≤–1; (–1, 0]; (0, 1]; >1) was described with Kaplan-Meier plots and examined with Wilcoxon (Peto-Prentice) test statistic. Adjusted Cox multivariable regressions had two-sided Wald tests and nominal significance P < .05. RESULTS: Of 7,576 women accrued, 3,048 women's tumors had machine-quantitated image analysis results: 2,900 (95%) for ER, 2,726 (89%) for PgR, and 2,582 (85% of 3,048) with both ER and PgR. Higher statistically standardized ER and PgR HSCORE and %+ were associated with better univariate DDFS and EFS ( P < .001). In multivariable assessments, ER HSCORE and %+ were not significantly associated ( P = .52-.88) with DDFS in models with PgR, whereas higher PgR HSCORE and %+ were significantly associated with better DDFS ( P = .001) in models with ER. CONCLUSION: Adjunctive statistical standardization differentiated quantitated levels of ER and PgR. Patients with higher ER- and PgR-standardized units had superior DDFS compared with those with HSCOREs and %+ ≤–1.
- 20 August 2024
- Published by American Society of Clinical Oncology (ASCO) in Journal of Clinical Oncology
- Vol. 42 (24) , 2887-2898
Adjunctive Statistical Standardization of Adjuvant Estrogen Receptor and Progesterone Receptor in Canadian Cancer Trials Group MA.27 Postmenopausal Breast Cancer Trial of Exemestane Versus Anastrozole
PURPOSE: ASCO/College of American Pathologists guidelines recommend reporting estrogen receptor (ER) and progesterone receptor (PgR) as positive with (1%-100%) staining. Statistically standardized quantitated positivity could indicate differential associations of positivity with breast cancer outcomes. METHODS: MA.27 (ClinicalTrials.gov identifier: NCT00066573 ) was a phase III adjuvant trial of exemestane versus anastrozole in postmenopausal women with early-stage breast cancer. Immunochemistry ER and PgR HSCORE and % positivity (%+) were centrally assessed by machine image quantitation and statistically standardized to mean 0 and standard deviation (SD) 1 after Box-Cox variance stabilization transformations of square for ER; for PgR, (1) natural logarithm (0.1 added to 0 HSCOREs and 0%+) and (2) square root. Our primary end point was MA.27 distant disease-free survival (DDFS) at a median 4.1-year follow-up, and secondary end point was event-free survival (EFS). Univariate survival with cut points at SDs about a mean of 0 (≤–1; (–1, 0]; (0, 1]; >1) was described with Kaplan-Meier plots and examined with Wilcoxon (Peto-Prentice) test statistic. Adjusted Cox multivariable regressions had two-sided Wald tests and nominal significance P < .05. RESULTS: Of 7,576 women accrued, 3,048 women's tumors had machine-quantitated image analysis results: 2,900 (95%) for ER, 2,726 (89%) for PgR, and 2,582 (85% of 3,048) with both ER and PgR. Higher statistically standardized ER and PgR HSCORE and %+ were associated with better univariate DDFS and EFS ( P < .001). In multivariable assessments, ER HSCORE and %+ were not significantly associated ( P = .52-.88) with DDFS in models with PgR, whereas higher PgR HSCORE and %+ were significantly associated with better DDFS ( P = .001) in models with ER. CONCLUSION: Adjunctive statistical standardization differentiated quantitated levels of ER and PgR. Patients with higher ER- and PgR-standardized units had superior DDFS compared with those with HSCOREs and %+ ≤–1.
- Open Access
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Supplementary Table S2 from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Supplemental Table 2. Distribution of patients as defined by BCI prognostic (BCI/BCIN+) and BCI predictive (BCI (H/I)).
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Table S2 from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Supplemental Table 2. Distribution of patients as defined by BCI prognostic (BCI/BCIN+) and BCI predictive (BCI (H/I)).
- Open Access
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Data from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Purpose:Patients with early-stage hormone receptor–positive (HR+) breast cancer face a prolonged risk of recurrence even after adjuvant endocrine therapy. The Breast Cancer Index (BCI) is significantly prognostic for overall (0–10 years) and late (5–10 years) distant recurrence (DR) risk in N0 and N1 patients. Here, BCI prognostic performance was evaluated in HR+ postmenopausal women from the Tamoxifen and Exemestane Adjuvant Multinational (TEAM) trial.Experimental Design:3,544 patients were included in the analysis (N = 1,519 N0, N = 2,025 N+). BCI risk groups were calculated using pre-specified cutoff points. Kaplan–Meier analyses and log-rank tests were used to assess the prognostic significance of BCI risk groups based on DR. Hazard ratios (HR) and confidence intervals (CI) were calculated using Cox models with and without clinical covariates.Results:For overall 10-year DR, BCI was significantly prognostic in Ni0 (N = 1,196) and N1 (N = 1,234) patients who did not receive prior chemotherapy (P < 0.001). In patients who were DR-free for 5 years, 10-year late DR rates for low- and high-risk groups were 5.4% and 9.3% (N0 cohort, N = 1,285) and 4.8% and 12.2% (N1 cohort, N = 1,625) with multivariate HRs of 2.25 (95% CI, 1.30–3.88; P = 0.004) and 2.67 (95% CI, 1.53–4.63; P < 0.001), respectively. Late DR performance was substantially improved using previously optimized cutoff points, identifying BCI low-risk groups with even lower 10-year late DR rates of 3.8% and 2.7% in N0 and N1 patients, respectively.Conclusions:The TEAM trial represents the largest prognostic validation study for BCI to date and provides a more representative assessment of late DR risk to guide individualized treatment decision-making for HR+ patients with early-stage breast cancer.
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessData from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Purpose:Patients with early-stage hormone receptor–positive (HR+) breast cancer face a prolonged risk of recurrence even after adjuvant endocrine therapy. The Breast Cancer Index (BCI) is significantly prognostic for overall (0–10 years) and late (5–10 years) distant recurrence (DR) risk in N0 and N1 patients. Here, BCI prognostic performance was evaluated in HR+ postmenopausal women from the Tamoxifen and Exemestane Adjuvant Multinational (TEAM) trial.Experimental Design:3,544 patients were included in the analysis (N = 1,519 N0, N = 2,025 N+). BCI risk groups were calculated using pre-specified cutoff points. Kaplan–Meier analyses and log-rank tests were used to assess the prognostic significance of BCI risk groups based on DR. Hazard ratios (HR) and confidence intervals (CI) were calculated using Cox models with and without clinical covariates.Results:For overall 10-year DR, BCI was significantly prognostic in Ni0 (N = 1,196) and N1 (N = 1,234) patients who did not receive prior chemotherapy (P < 0.001). In patients who were DR-free for 5 years, 10-year late DR rates for low- and high-risk groups were 5.4% and 9.3% (N0 cohort, N = 1,285) and 4.8% and 12.2% (N1 cohort, N = 1,625) with multivariate HRs of 2.25 (95% CI, 1.30–3.88; P = 0.004) and 2.67 (95% CI, 1.53–4.63; P < 0.001), respectively. Late DR performance was substantially improved using previously optimized cutoff points, identifying BCI low-risk groups with even lower 10-year late DR rates of 3.8% and 2.7% in N0 and N1 patients, respectively.Conclusions:The TEAM trial represents the largest prognostic validation study for BCI to date and provides a more representative assessment of late DR risk to guide individualized treatment decision-making for HR+ patients with early-stage breast cancer.
- Open Access
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Data from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Purpose:Patients with early-stage hormone receptor–positive (HR+) breast cancer face a prolonged risk of recurrence even after adjuvant endocrine therapy. The Breast Cancer Index (BCI) is significantly prognostic for overall (0–10 years) and late (5–10 years) distant recurrence (DR) risk in N0 and N1 patients. Here, BCI prognostic performance was evaluated in HR+ postmenopausal women from the Tamoxifen and Exemestane Adjuvant Multinational (TEAM) trial.Experimental Design:3,544 patients were included in the analysis (N = 1,519 N0, N = 2,025 N+). BCI risk groups were calculated using pre-specified cutoff points. Kaplan–Meier analyses and log-rank tests were used to assess the prognostic significance of BCI risk groups based on DR. Hazard ratios (HR) and confidence intervals (CI) were calculated using Cox models with and without clinical covariates.Results:For overall 10-year DR, BCI was significantly prognostic in Ni0 (N = 1,196) and N1 (N = 1,234) patients who did not receive prior chemotherapy (P < 0.001). In patients who were DR-free for 5 years, 10-year late DR rates for low- and high-risk groups were 5.4% and 9.3% (N0 cohort, N = 1,285) and 4.8% and 12.2% (N1 cohort, N = 1,625) with multivariate HRs of 2.25 (95% CI, 1.30–3.88; P = 0.004) and 2.67 (95% CI, 1.53–4.63; P < 0.001), respectively. Late DR performance was substantially improved using previously optimized cutoff points, identifying BCI low-risk groups with even lower 10-year late DR rates of 3.8% and 2.7% in N0 and N1 patients, respectively.Conclusions:The TEAM trial represents the largest prognostic validation study for BCI to date and provides a more representative assessment of late DR risk to guide individualized treatment decision-making for HR+ patients with early-stage breast cancer.
- 15 April 2024
- Published by American Association for Cancer Research (AACR)
Open AccessData from Validation of the Prognostic Performance of Breast Cancer Index in Hormone Receptor–Positive Postmenopausal Breast Cancer Patients in the TEAM Trial
Purpose:Patients with early-stage hormone receptor–positive (HR+) breast cancer face a prolonged risk of recurrence even after adjuvant endocrine therapy. The Breast Cancer Index (BCI) is significantly prognostic for overall (0–10 years) and late (5–10 years) distant recurrence (DR) risk in N0 and N1 patients. Here, BCI prognostic performance was evaluated in HR+ postmenopausal women from the Tamoxifen and Exemestane Adjuvant Multinational (TEAM) trial.Experimental Design:3,544 patients were included in the analysis (N = 1,519 N0, N = 2,025 N+). BCI risk groups were calculated using pre-specified cutoff points. Kaplan–Meier analyses and log-rank tests were used to assess the prognostic significance of BCI risk groups based on DR. Hazard ratios (HR) and confidence intervals (CI) were calculated using Cox models with and without clinical covariates.Results:For overall 10-year DR, BCI was significantly prognostic in Ni0 (N = 1,196) and N1 (N = 1,234) patients who did not receive prior chemotherapy (P < 0.001). In patients who were DR-free for 5 years, 10-year late DR rates for low- and high-risk groups were 5.4% and 9.3% (N0 cohort, N = 1,285) and 4.8% and 12.2% (N1 cohort, N = 1,625) with multivariate HRs of 2.25 (95% CI, 1.30–3.88; P = 0.004) and 2.67 (95% CI, 1.53–4.63; P < 0.001), respectively. Late DR performance was substantially improved using previously optimized cutoff points, identifying BCI low-risk groups with even lower 10-year late DR rates of 3.8% and 2.7% in N0 and N1 patients, respectively.Conclusions:The TEAM trial represents the largest prognostic validation study for BCI to date and provides a more representative assessment of late DR risk to guide individualized treatment decision-making for HR+ patients with early-stage breast cancer.