Validation of the online prediction tool PREDICT v. 2.0 in the Dutch breast cancer population

Evaluating PREDICT and developing outcome prediction models in early-onset breast cancer using data from Alberta, Canada

INTRODUCTION

Outcome prediction research in early-onset breast cancer (EoBC) is limited. This study evaluated the predictive performance of NHS PREDICT v2.1 and developed two prediction models for 5-year and 10-year all-cause mortality in a cohort of EoBC patients in Alberta, Canada.

METHODS

Adults < 40 years diagnosed with invasive breast cancer in Alberta, Canada from 2004 to 2020 were included. Patient data were entered into PREDICT v2.1 and mortality estimates at 5 and 10 years were extracted. Two prediction models were developed for all-cause mortality: multivariable Cox regression with LASSO penalization (LASSO Cox) and random survival forests (RSF). Internal validation of the developed models was performed using nested tenfold cross-validation repeated 200 times. Model performance was assessed using receiver operator characteristic and calibration curves for mortality at 5 and 10 years.

RESULTS

In total, 1827 patients with EoBC were eligible for inclusion. At 5 years, PREDICT had an area under the curve of 0.78 (95%CI 0.74-0.82) and overestimated mortality by 2.4% (95%CI 0.70-4.33) in the overall cohort. No differences in observed and predicted mortality by PREDICT were observed at 10 years. The LASSO Cox model showed better discrimination at 5 and 10 years than the RSF model, but both had poor calibration and underestimated mortality.

CONCLUSION

PREDICT v2.1 tended to overestimate 5-year mortality in those with > 30% predicted risks and 10-year mortality in those with > 50% predicted risks for EoBC in Alberta, Canada. We did not identify additional models that would be clinically useful by applying machine learning. More follow-up data and emerging systemic treatment variables are required to study outcome prediction in modern cohorts.

Development of radiomics-based models on mammograms with mass lesions to predict prognostically relevant characteristics of invasive breast cancer in a screening cohort

BACKGROUND

Optimizing breast-screening performance involves minimizing overdiagnosis of prognostically favorable invasive breast cancer (IBC) that does not need immediate recall and underdiagnosis of prognostically unfavorable IBC that is not recalled timely. We investigated whether mammographic features of masses predict prognostically relevant IBC characteristics.

METHODS

In a screening cohort, we obtained pathological information of 1587 IBCs presenting as a mass through the nationwide cancer registry and pathology databank. We developed models based on mammographic tumor appearance to predict whether IBC was prognostically favorable (T1N0M0 luminal A-like) or unfavorable. Models were based on 1095 positive screening mammograms (possible overdiagnosis), or on 603 last negative mammograms with in retrospect visible masses (possible underdiagnosis). We calculated performance metrics using cross-validation.

RESULTS

23.5% of masses were prognostically favorable IBC. Using 1095 positive mammograms, the model's predictions to have prognostically favorable IBC (10th-90th percentile range 8.7-47.0%) yielded AUC 0.75 (SD across repeats 0.01), slope 1.16 (SD 0.07). Performance in 603 last negative screening mammograms with masses was poor: AUC 0.60 (SD 0.02), slope 0.85 (SD 0.28).

CONCLUSIONS

Mammography-based models from masses representing IBC at time of recall (possible overdiagnosis) predict prognostically relevant characteristics of IBC. Models based on in retrospect visible masses (possible underdiagnosis) performed poorly.

Insights into the performance of PREDICT tool in a large Mainland Chinese breast cancer cohort: a comparative analysis of versions 3.0 and 2.2

BACKGROUND

PREDICT is a web-based tool for forecasting breast cancer outcomes. PREDICT version 3.0 was recently released. This study aimed to validate this tool for a large population in mainland China and compare v3.0 with v2.2.

METHODS

Women who underwent surgery for nonmetastatic primary invasive breast cancer between 2010 and 2020 from the First Affiliated Hospital of Wenzhou Medical University were selected. Predicted and observed 5-year overall survival (OS) for both v3.0 and v2.2 were compared. Discrimination was compared using receiver-operator curves and DeLong test. Calibration was evaluated using calibration plots and chi-squared test. A difference greater than 5% was deemed clinically relevant.

RESULTS

A total of 5424 patients were included, with median follow-up time of 58 months (IQR 38-89 months). Compared to v2.2, v3.0 did not show improved discriminatory accuracy for 5-year OS (AUC: 0.756 vs 0.771), same as ER-positive and ER-negative patients. However, calibration was significantly improved in v3.0, with predicted 5-year OS deviated from observed by -2.0% for the entire cohort, -2.9% for ER-positive and -0.0% for ER-negative patients, compared to -7.3%, -4.7% and -13.7% in v2.2. In v3.0, 5-year OS was underestimated by 9.0% for patients older than 75 years, and 5.8% for patients with micrometastases. Patients with distant metastases postdiagnosis was overestimated by 10.6%.

CONCLUSIONS

PREDICT v3.0 reliably predicts 5-year OS for the majority of Chinese patients with breast cancer. PREDICT v3.0 significantly improved the predictive accuracy for ER-negative groups. Furthermore, caution is advised when interpreting 5-year OS for patients aged over 70, those with micrometastases or metastases postdiagnosis.

Independent Validation of the BRENDA-Score Breast Cancer Prognosis Prediction Tool In Chinese Patients

BACKGROUND

The BRENDA-Score was developed and used to predict the prognosis of patients with breast cancer (BC). This study was performed to validate the use of this tool in Chinese patients with primary invasive BC patients.

METHODS

Patients underwent surgery for BC from January 2009 to December 2016. Discrimination was assessed by the area under the receiver operating characteristic (ROC) curve (AUC). Calibrations were assessed by comparing predicted and observed 5-year and 10-year metastasis-free survival (MFS) in the overall cohort and patient subgroups.

RESULTS

A total of 2029 BC patients were enrolled. Kaplan-Meier analysis revealed significant differences in MFS risk groups (log-rank test P < .01). ROC analysis showed good accuracy for 5-year MFS (AUC 0.779) and fair accuracy for 10-year MFS (AUC 0.728). The BRENDA-Score accurately predicted 5-year and 10-year MFS in the entire cohort and in all other predefined subgroups, except for the 5-year MFS in the subgroup aged<40 years, which was overestimated (differences between the predicted and observed MFS were 6.7%, P < .05). The 5-year MFS rates of ER- positive and ER-negative patients were 90.9% and 80.6%, respectively (P < .05). The 10-year MFS rates of ER-positive and ER-negative patients were 78.0% and 73.7%, respectively (P = .25).

CONCLUSIONS

The BRENDA-Score accurately predicted 5-year and 10-year MFS. The results showed good validity, transportability, and potential clinical value. However, the results for 5-years MFS should be interpreted carefully in patients aged <40 years. After 10 years the value of the ER as a prognostic factor was less important.

An updated PREDICT breast cancer prognostic model including the benefits and harms of radiotherapy

PREDICT Breast ( www.breast .predict.nhs.uk ) is a prognostication tool for early invasive breast cancer. The current version was based on cases diagnosed in 1999-2003 and did not incorporate the benefits of radiotherapy or the harms associated with therapy. Since then, there has been a substantial improvement in the outcomes for breast cancer cases. The aim of this study was to update PREDICT Breast to ensure that the underlying model is appropriate for contemporary patients. Data from the England National Cancer Registration and Advisory Service for invasive breast cancer cases diagnosed 2000-17 were used for model development and validation. Model development was based on 35,474 cases diagnosed and registered by the Eastern Cancer Registry. A Cox model was used to estimate the prognostic effects of the year of diagnosis, age at diagnosis, tumour size, tumour grade and number of positive nodes. Separate models were developed for ER-positive and ER-negative disease. Data on 32,408 cases from the West Midlands Cancer Registry and 100,551 cases from other cancer registries were used for validation. The new model was well-calibrated; predicted breast cancer deaths at 5-, 10- and 15-year were within 10 per cent of the observed validation data. Discrimination was also good: The AUC for 15-year breast cancer survival was 0.809 in the West Midlands data set and 0.846 in the data set for the other registries. The new PREDICT Breast model outperformed the current model and will be implemented in the online tool which should lead to more accurate absolute treatment benefit predictions for individual patients.

Gene expression profiles in clinically T1-2N0 ER+HER2- breast cancer patients treated with breast-conserving therapy: their added value in case sentinel lymph node biopsy is not performed

PURPOSE

Omitting sentinel lymph node biopsy (SLNB) in breast cancer treatment results in patients with unknown positive nodal status and potential risk for systemic undertreatment. This study aimed to investigate whether gene expression profiles (GEPs) can lower this risk in cT1-2N0 ER+ HER2- breast cancer patients treated with BCT.

METHODS

Patients were included if diagnosed between 2011 and 2017 with cT1-2N0 ER+ HER2- breast cancer, treated with BCT and SLNB, and in whom GEP was applied. Adjuvant chemotherapy recommendations based on clinical risk status (Dutch breast cancer guideline of 2020 versus PREDICT v2.1) with and without knowledge on SLNB outcome were compared to GEP outcome. We examined missing adjuvant chemotherapy indications, and the number of GEPs needed to identify one patient at risk for systemic undertreatment.

RESULTS

Of 3585 patients, 2863 (79.9%) had pN0 and 722 (20.1%) pN + disease. Chemotherapy was recommended in 1354 (37.8% guideline-2020) and 1888 patients (52.7% PREDICT). Eliminating SLNB outcome (n = 722) resulted in omission of chemotherapy recommendation in 475 (35.1% guideline-2020) and 412 patients (21.8% PREDICT). GEP revealed genomic high risk in 126 (26.5% guideline-2020) and 82 patients (19.9% PREDICT) in case of omitted chemotherapy recommendation in the absence of SLNB. Extrapolated to the whole group, this concerns 3.5% and 2.3%, respectively, resulting in the need for 28-44 GEPs to identify one patient at risk for systemic undertreatment.

CONCLUSION

If no SLNB is performed, clinical risk status according to the guideline of 2020 and PREDICT predicts a very low risk for systemic undertreatment. The number of GEPs needed to identify one patient at risk for undertreatment does not justify its standard use.

External validation and clinical utility assessment of PREDICT breast cancer prognostic model in young, systemic treatment-naïve women with node-negative breast cancer

BACKGROUND

The validity of the PREDICT breast cancer prognostic model is unclear for young patients without adjuvant systemic treatment. This study aimed to validate PREDICT and assess its clinical utility in young women with node-negative breast cancer who did not receive systemic treatment.

METHODS

We selected all women from the Netherlands Cancer Registry who were diagnosed with node-negative breast cancer under age 40 between 1989 and 2000, a period when adjuvant systemic treatment was not standard practice for women with node-negative disease. We evaluated the calibration and discrimination of PREDICT using the observed/expected (O/E) mortality ratio, and the area under the receiver operating characteristic curve (AUC), respectively. Additionally, we compared the potential clinical utility of PREDICT for selectively administering chemotherapy to the chemotherapy-to-all strategy using decision curve analysis at predefined thresholds.

RESULTS

A total of 2264 women with a median age at diagnosis of 36 years were included. Of them, 71.2% had estrogen receptor (ER)-positive tumors and 44.0% had grade 3 tumors. Median tumor size was 16 mm. PREDICT v2.2 underestimated 10-year all-cause mortality by 33% in all women (O/E ratio:1.33, 95%CI:1.22-1.43). Model discrimination was moderate overall (AUC10-year:0.65, 95%CI:0.62-0.68), and poor for women with ER-negative tumors (AUC10-year:0.56, 95%CI:0.51-0.62). Compared to the chemotherapy-to-all strategy, PREDICT only showed a slightly higher net benefit in women with ER-positive tumors, but not in women with ER-negative tumors.

CONCLUSIONS

PREDICT yields unreliable predictions for young women with node-negative breast cancer. Further model updates are needed before PREDICT can be routinely used in this patient subset.

Evaluation of Predict, a prognostic risk tool, after diagnosis of a second breast cancer

BACKGROUND

The UK National Health Service's Predict is a clinical tool widely used to estimate the prognosis of early-stage breast cancer. The performance of Predict for a second primary breast cancer is unknown.

METHODS

Women 18 years of age or older diagnosed with a first or second invasive breast cancer between 2000 and 2013 and followed for at least 5 years were identified from the US Surveillance, Epidemiology, and End Results (SEER) database. Model calibration of Predict was evaluated by comparing predicted and observed 5-year breast cancer-specific mortality separately by estrogen receptor status for first vs second breast cancer. Receiver operating characteristic curves and areas under the curve were used to assess model discrimination. Model performance was also evaluated for various races and ethnicities.

RESULTS

The study population included 6729 women diagnosed with a second breast cancer and 357 204 women with a first breast cancer. Overall, Predict demonstrated good discrimination for first and second breast cancers (areas under the curve ranging from 0.73 to 0.82). Predict statistically significantly underestimated 5-year breast cancer mortality for second estrogen receptor-positive breast cancers (predicted-observed = ‒6.24%, 95% CI = ‒6.96% to ‒5.49%). Among women with a first estrogen receptor-positive cancer, model calibration was good (predicted-observed = ‒0.22%, 95% CI = ‒0.29% to ‒0.15%), except in non-Hispanic Black women (predicted-observed = ‒2.33%, 95% CI = ‒2.65% to ‒2.01%) and women 80 years of age or older (predicted-observed = ‒3.75%, 95% CI = ‒4.12% to ‒3.41%). Predict performed well for second estrogen receptor-negative cancers overall (predicted-observed = ‒1.69%, 95% CI = ‒3.99% to 0.16%) but underestimated mortality among those who had previously received chemotherapy or had a first cancer with more aggressive tumor characteristics. In contrast, Predict overestimated mortality for first estrogen receptor-negative cancers (predicted-observed = 4.54%, 95% CI = 4.27% to 4.86%).

CONCLUSION

The Predict tool underestimated 5-year mortality after a second estrogen receptor-positive breast cancer and in certain subgroups of women with a second estrogen receptor-negative breast cancer.

Validation of PREDICT Version 2.2 in a Retrospective Cohort of Indian Women With Operable Breast Cancer

PURPOSE

Online prediction models that use known prognostic factors in breast cancer (BC) are routinely used to assist in decisions for adjuvant therapy. PREDICT Version 2.2 (P2.2) is one such online tool, which uses tumor size, lymph node involvement, grade, age, hormone receptor status, human epidermal growth factor receptor 2 (HER2) status, and Ki67. We performed an external validation in a retrospective cohort of patients treated at a tertiary center in India.

METHODS

Women with operable BC between 2008 and 2016 with nonmetastatic, T1-T2 invasive, and HER2 receptor-negative BC and with available 5-year overall survival (OS) data were selected. Median predicted 5-year OS rates were used to calculate predicted events for the whole cohort and subgroups. The chi-square test was used to evaluate the goodness of fit of the tool.

RESULTS

Of 11,760 cases registered between 2008 and 2016, 2,783 (23.66%) eligible patients with a median age of 50 (26-70) years and a median pT size of 2.5 (0.1-5) cm, 2,037 (73.19%) with grade 3 tumors, 1,172 (42.11%) with node-positive disease, 817 (29.35%) with triple-negative breast cancer, and 1,966 (70.64%) with HR-positive BC were included in the analysis. The observed 5-year OS and predicted 5-year OS in the whole cohort were 94.8% and 90.00%, respectively, with an absolute difference of 4.8% (95% CI, 3.417 to 6.198, P < .001). The observed 5-year OS and predicted 5-year OS were also different in various subgroups.

CONCLUSION

PREDICT version 2.2 overestimated the number of deaths, with lower predicted 5-year OS compared with the observed value, in this retrospective Indian cohort. The reasons for this discrepancy could be differing biologic characteristics and possible selection bias in our cohort. We recommend a prospective validation of PREDICT in Indian patients and advocate caution in its use until such validation is achieved.

Adjuvant therapeutic strategy decision support for an elderly population with localized breast cancer: A monocentric cohort retrospective study

Guidelines for the management of elderly patients with early breast cancer are scarce. Additional adjuvant systemic treatment to surgery for early breast cancer in elderly populations is challenged by increasing comorbidities with age. In non-metastatic settings, treatment decisions are often made under considerable uncertainty; this commonly leads to undertreatment and, consequently, poorer outcomes. This study aimed to develop a decision support tool that can help to identify candidate adjuvant post-surgery treatment schemes for elderly breast cancer patients based on tumor and patient characteristics. Our approach was to generate predictions of patient outcomes for different courses of action; these predictions can, in turn, be used to inform clinical decisions for new patients. We used a cohort of elderly patients (≥ 70 years) who underwent surgery with curative intent for early breast cancer to train the models. We tested seven classification algorithms using 5-fold cross-validation, with 80% of the data being randomly selected for training and the remaining 20% for testing. We assessed model performance using accuracy, precision, recall, F1-score, and AUC score. We used an autoencoder to perform dimensionality reduction prior to classification. We observed consistently better performance using logistic regression and linear discriminant analysis models when compared to the other models we tested. Classification performance generally improved when an autoencoder was used, except for when we predicted the need for adjuvant treatment. We obtained overall best results using a logistic regression model without autoencoding to predict the need for adjuvant treatment (F1-score = 0.869).

PREDICT validity for prognosis of breast cancer patients with pathogenic BRCA1/2 variants

We assessed the PREDICT v 2.2 for prognosis of breast cancer patients with pathogenic germline BRCA1 and BRCA2 variants, using follow-up data from 5453 BRCA1/2 carriers from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA) and the Breast Cancer Association Consortium (BCAC). PREDICT for estrogen receptor (ER)-negative breast cancer had modest discrimination for BRCA1 carrier patients overall (Gönen & Heller unbiased concordance 0.65 in CIMBA, 0.64 in BCAC), but it distinguished clearly the high-mortality group from lower risk categories. In an analysis of low to high risk categories by PREDICT score percentiles, the observed mortality was consistently lower than the expected mortality, but the confidence intervals always included the calibration slope. Altogether, our results encourage the use of the PREDICT ER-negative model in management of breast cancer patients with germline BRCA1 variants. For the PREDICT ER-positive model, the discrimination was slightly lower in BRCA2 variant carriers (concordance 0.60 in CIMBA, 0.65 in BCAC). Especially, inclusion of the tumor grade distorted the prognostic estimates. The breast cancer mortality of BRCA2 carriers was underestimated at the low end of the PREDICT score distribution, whereas at the high end, the mortality was overestimated. These data suggest that BRCA2 status should also be taken into consideration with tumor characteristics, when estimating the prognosis of ER-positive breast cancer patients.

Development and internal-external validation of statistical and machine learning models for breast cancer prognostication: cohort study

OBJECTIVE

To develop a clinically useful model that estimates the 10 year risk of breast cancer related mortality in women (self-reported female sex) with breast cancer of any stage, comparing results from regression and machine learning approaches.

DESIGN

Population based cohort study.

SETTING

QResearch primary care database in England, with individual level linkage to the national cancer registry, Hospital Episodes Statistics, and national mortality registers.

PARTICIPANTS

141 765 women aged 20 years and older with a diagnosis of invasive breast cancer between 1 January 2000 and 31 December 2020.

MAIN OUTCOME MEASURES

Four model building strategies comprising two regression (Cox proportional hazards and competing risks regression) and two machine learning (XGBoost and an artificial neural network) approaches. Internal-external cross validation was used for model evaluation. Random effects meta-analysis that pooled estimates of discrimination and calibration metrics, calibration plots, and decision curve analysis were used to assess model performance, transportability, and clinical utility.

RESULTS

During a median 4.16 years (interquartile range 1.76-8.26) of follow-up, 21 688 breast cancer related deaths and 11 454 deaths from other causes occurred. Restricting to 10 years maximum follow-up from breast cancer diagnosis, 20 367 breast cancer related deaths occurred during a total of 688 564.81 person years. The crude breast cancer mortality rate was 295.79 per 10 000 person years (95% confidence interval 291.75 to 299.88). Predictors varied for each regression model, but both Cox and competing risks models included age at diagnosis, body mass index, smoking status, route to diagnosis, hormone receptor status, cancer stage, and grade of breast cancer. The Cox model's random effects meta-analysis pooled estimate for Harrell's C index was the highest of any model at 0.858 (95% confidence interval 0.853 to 0.864, and 95% prediction interval 0.843 to 0.873). It appeared acceptably calibrated on calibration plots. The competing risks regression model had good discrimination: pooled Harrell's C index 0.849 (0.839 to 0.859, and 0.821 to 0.876, and evidence of systematic miscalibration on summary metrics was lacking. The machine learning models had acceptable discrimination overall (Harrell's C index: XGBoost 0.821 (0.813 to 0.828, and 0.805 to 0.837); neural network 0.847 (0.835 to 0.858, and 0.816 to 0.878)), but had more complex patterns of miscalibration and more variable regional and stage specific performance. Decision curve analysis suggested that the Cox and competing risks regression models tested may have higher clinical utility than the two machine learning approaches.

CONCLUSION

In women with breast cancer of any stage, using the predictors available in this dataset, regression based methods had better and more consistent performance compared with machine learning approaches and may be worthy of further evaluation for potential clinical use, such as for stratified follow-up.

A Comparison between the Online Prognostic Tool PREDICT and myBeST for Women with Breast Cancer in Malaysia

The PREDICT breast cancer is a well-known online calculator to estimate survival probability. We developed a new prognostic model, myBeST, due to the PREDICT tool’s limitations when applied to our patients. This study aims to compare the performance of the two models for women with breast cancer in Malaysia. A total of 532 stage I to III patient records who underwent surgical treatment were analysed. They were diagnosed between 2012 and 2016 in seven centres. We obtained baseline predictors and survival outcomes by reviewing patients’ medical records. We compare PREDICT and myBeST tools’ discriminant performance using receiver-operating characteristic (ROC) analysis. The five-year observed survival was 80.3% (95% CI: 77.0, 83.7). For this cohort, the median five-year survival probabilities estimated by PREDICT and myBeST were 85.8% and 82.6%, respectively. The area under the ROC curve for five-year survival by myBeST was 0.78 (95% CI: 0.73, 0.82) and for PREDICT was 0.75 (95% CI: 0.70, 0.80). Both tools show good performance, with myBeST marginally outperforms PREDICT discriminant performance. Thus, the new prognostic model is perhaps more suitable for women with breast cancer in Malaysia.

myBeST-A Web-Based Survival Prognostic Tool for Women with Breast Cancer in Malaysia: Development Process and Preliminary Validation Study

Women with breast cancer are keen to know their predicted survival. We developed a new prognostic model for women with breast cancer in Malaysia. Using the model, this study aimed to design the user interface and develop the contents of a web-based prognostic tool for the care provider to convey survival estimates. We employed an iterative website development process which includes: (1) an initial development stage informed by reviewing existing tools and deliberation among breast surgeons and epidemiologists, (2) content validation and feedback by medical specialists, and (3) face validation and end-user feedback among medical officers. Several iterative prototypes were produced and improved based on the feedback. The experts (n = 8) highly agreed on the website content and predictors for survival with content validity indices ≥ 0.88. Users (n = 20) scored face validity indices of more than 0.90. They expressed favourable responses. The tool, named Malaysian Breast cancer Survival prognostic Tool (myBeST), is accessible online. The tool estimates an individualised five-year survival prediction probability. Accompanying contents were included to explain the tool's aim, target user, and development process. The tool could act as an additional tool to provide evidence-based and personalised breast cancer outcomes.

The majority of 922 prediction models supporting breast cancer decision-making are at high risk of bias

OBJECTIVES

To systematically review the currently available prediction models that may support treatment decision-making in breast cancer.

STUDY DESIGN AND SETTING

Literature was systematically searched to identify studies reporting on development of prediction models aiming to support breast cancer treatment decision-making, published between January 2010 and December 2020. Quality and risk of bias were assessed using the Prediction model Risk Of Bias (ROB) Assessment Tool (PROBAST).

RESULTS

After screening 20,460 studies, 534 studies were included, reporting on 922 models. The 922 models predicted: mortality (n = 417 45%), recurrence (n = 217, 24%), lymph node involvement (n = 141, 15%), adverse events (n = 58, 6%), treatment response (n = 56, 6%), or other outcomes (n = 33, 4%). In total, 285 models (31%) lacked a complete description of the final model and could not be applied to new patients. Most models (n = 878, 95%) were considered to contain high ROB.

CONCLUSION

A substantial overlap in predictor variables and outcomes between the models was observed. Most models were not reported according to established reporting guidelines or showed methodological flaws during the development and/or validation of the model. Further development of prediction models with thorough quality and validity assessment is an essential first step for future clinical application.

A comparative analysis of recurrence risk predictions in ER+/HER2- early breast cancer using NHS Nottingham Prognostic Index, PREDICT, and CanAssist Breast

AIMS

Clinicians use multi-gene/biomarker prognostic tests and free online tools to optimize treatment in early ER+/HER2- breast cancer. Here we report the comparison of recurrence risk predictions by CanAssist Breast (CAB), Nottingham Prognostic Index (NPI), and PREDICT along with the differences in the performance of these tests across Indian and European cohorts.

METHODS

Current study used a retrospective cohort of 1474 patients from Europe, India, and USA. NPI risk groups were categorized into three prognostic groups, good (GPG-NPI index ≤ 3.4) moderate (MPG 3.41-5.4), and poor (PPG  > 5.4). Patients with chemotherapy benefit of < 2% were low-risk and ≥ 2% high-risk by PREDICT. We assessed the agreement between the CAB and NPI/PREDICT risk groups by kappa coefficient.

RESULTS

Risk proportions generated by all tools were: CAB low:high 74:26; NPI good:moderate:poor prognostic group- 38:55:7; PREDICT low:high 63:37. Overall, there was a fair agreement between CAB and NPI[κ = 0.31(0.278-0.346)]/PREDICT [κ = 0.398 (0.35-0.446)], with a concordance of 97%/88% between CAB and NPI/PREDICT low-risk categories. 65% of NPI-MPG patients were called low-risk by CAB. From PREDICT high-risk patients CAB segregated 51% as low-risk, thus preventing over-treatment in these patients. In cohorts (European) with a higher number of T1N0 patients, NPI/PREDICT segregated more as LR compared to CAB, suggesting that T1N0 patients with aggressive biology are missed out by online tools but not by the CAB.

CONCLUSION

Data shows the use of CAB in early breast cancer overall and specifically in NPI-MPG and PREDICT high-risk patients for making accurate decisions on chemotherapy use. CAB provided unbiased risk stratification across cohorts of various geographies with minimal impact by clinical parameters.

Population-based estimates of overtreatment with adjuvant systemic therapy in early breast cancer patients with data from the Netherlands and the USA

Purpose

Although adjuvant systemic therapy (AST) helps increase breast cancer-specific survival (BCSS), there is a growing concern for overtreatment. By estimating the expected BCSS of AST using PREDICT, this study aims to quantify the number of patients treated with AST without benefit to provide estimates of overtreatment.

Methods

Data of all non-metastatic unilateral breast cancer patients diagnosed in 2015 were retrieved from cancer registries from The Netherlands and the USA. The PREDICT tool was used to estimate AST survival benefit. Overtreatment was defined as the proportion of patients that would have survived regardless of or died despite AST within 10 years. Three scenarios were evaluated: actual treatment, and recommendations by the Dutch or USA guidelines.

Results

59.5% of Dutch patients were treated with AST. 6.4% (interquartile interval [IQI] = 2.5, 8.2%) was expected to survive at least 10 years due to AST, leaving 93.6% (IQI = 91.8, 97.5%) without AST benefit (overtreatment). The lowest expected amount of overtreatment was in the targeted and chemotherapy subgroup, with 86.5% (IQI = 83.4, 89.6%) overtreatment, and highest in the only endocrine treatment subgroup, with 96.7% (IQI = 96.0, 98.1%) overtreatment. Similar results were obtained using data from the USA, and guideline recommendations.

Conclusion

Based on PREDICT, AST prevents 10-year breast cancer death in 6.4% of the patients treated with AST. Consequently, AST yields no survival benefit to many treated patients. Especially improved personalization of endocrine therapy is relevant, as this therapy is widely used and is associated with the highest amount of overtreatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10549-022-06550-2.

A scoping review of interactive and personalized web-based clinical tools to support treatment decision making in breast cancer

The increasing attention on personalized breast cancer care has resulted in an explosion of new interactive, tailored, web-based clinical decision tools for guiding treatment decisions in clinical practice. The goal of this study was to review, compare, and discuss the clinical implications of current tools, and highlight future directions for tools aiming to improve personalized breast cancer care. We searched PubMed, Embase, PsychInfo, Cochrane Database of Systematic Reviews, Web of Science, and Scopus to identify web-based decision tools addressing breast cancer treatment decisions. There was a total of 17 articles associated with 21 unique tools supporting decisions related to surgery, radiation therapy, hormonal therapy, bisphosphonates, HER2-targeted therapy, and chemotherapy. The quality of the tools was assessed using the International Patient Decision Aid Standard instrument. Overall, the tools considered clinical (e.g., age) and tumor characteristics (e.g., grade) to provide personalized outcomes (e.g., survival) associated with various treatment options. Fewer tools provided the adverse effects of the selected treatment. Only one tool was field-tested with patients, and none were tested with healthcare providers. Future studies need to assess the feasibility, usability, acceptability, as well as the effects of personalized web-based decision tools on communication and decision making from the patient and clinician perspectives.

Validity of the prognostication tool PREDICT version 2.2 in Japanese breast cancer patients

Introduction

PREDICT is a prognostication tool that calculates the potential benefit of various postsurgical treatments on the overall survival (OS) of patients with nonmetastatic invasive breast cancer. Once patient, tumor, and treatment details have been entered, the tool will show the estimated 5‐, 10‐, and 15‐year OS outcomes, both with and without adjuvant therapies. This study aimed to conduct an external validation of the prognostication tool PREDICT version 2.2 by evaluating its predictive accuracy of the 5‐ and 10‐year OS outcomes among female patients with nonmetastatic invasive breast cancer in Japan.

Methods

All female patients diagnosed from 2001 to 2013 with unilateral, nonmetastatic, invasive breast cancer and had undergone surgical treatment at Kyushu University Hospital, Fukuoka, Japan, were selected. Observed and predicted 5‐ and 10‐year OS rates were analyzed for the validation population and the subgroups. Calibration and discriminatory accuracy were assessed using Chi‐squared goodness‐of‐fit test and area under the receiver operating characteristic curve (AUC).

Results

A total of 636 eligible cases were selected from 1, 213 records. Predicted and observed OS differed by 0.9% (p = 0.322) for 5‐year OS, and 2.4% (p = 0.086) for 10‐year OS. Discriminatory accuracy results for 5‐year (AUC = 0.707) and 10‐year (AUC = 0.707) OS were fairly well.

Conclusion

PREDICT tool accurately estimated the 5‐ and 10‐year OS in the overall Japanese study population. However, caution should be used for interpretation of the 5‐year OS outcomes in patients that are ≥65 years old, and also for the 10‐year OS outcomes in patients that are ≥65 years old, those with histologic grade 3 and Luminal A tumors, and in those considering ETx or no systemic treatment.

Do histopathological features of breast cancer in Australian Indigenous women explain the survival disparity? A two decade long study in the Northern Territory

AIMS

In the Northern Territory (NT) of Australia, Indigenous women have a lower incidence of breast cancer, but a higher mortality than Non-indigenous women. The aim of this study was to describe and compare breast cancer pathological features related to stage and biological aggression between the two groups.

METHODS

Subjects were identified by extract from the NT Cancer Registry in two separate cohorts, cohort 1 (1991-2000) and cohort 2 (2001-2010). Data from cohort 1 included age, stage, tumor grade and estrogen receptor status (ER) and treatment completion. Additional pathological variables including tumor size, HER2 status, lymphovascular invasion and derived tumor phenotype were available for cohort 2. Bivariate P values for categoric variables were calculated using Fisher's exact tests.  The Wilcoxon rank-sum test was used to compare cohort 2. Logistic regression was used to calculate odds ratios.

RESULTS

There were 359 (44 indigenous) eligible women in cohort 1 and 526 (100 indigenous) for cohort 2. In cohort 1, in both cohorts, indigenous women were more likely to present at an advanced stage, but there was no difference in ER status or tumor grade. When derived phenotypes were compared, indigenous women were less likely to have better prognosis luminal A tumors, and more likely to have HER2-enriched tumors.

CONCLUSION

This two decade long comparison of the pathological features of breast cancer between indigenous and nonindigenous women of the NT has confirmed that Indigenous women not only present at a later stage than NI women but are also afflicted by poorer prognosis tumors, particularly HER2 enriched.

Multidimensional Machine Learning Personalized Prognostic Model in an Early Invasive Breast Cancer Population-Based Cohort in China: Algorithm Validation Study

Background

Current online prognostic prediction models for breast cancer, such as Adjuvant! Online and PREDICT, are based on specific populations. They have been well validated and widely used in the United States and Western Europe; however, several validation attempts in non-European countries have revealed suboptimal predictions.

Objective

We aimed to develop an advanced breast cancer prognosis model for disease progression, cancer-specific mortality, and all-cause mortality by integrating tumor, demographic, and treatment characteristics from a large breast cancer cohort in China.

Methods

This study was approved by the Clinical Test and Biomedical Ethics Committee of West China Hospital, Sichuan University on May 17, 2012. Data collection for this project was started in May 2017 and ended in March 2019. Data on 5293 women diagnosed with stage I to III invasive breast cancer between 2000 and 2013 were collected. Disease progression, cancer-specific mortality, all-cause mortality, and the likelihood of disease progression or death within a 5-year period were predicted. Extreme gradient boosting was used to develop the prediction model. Model performance was assessed by calculating the area under the receiver operating characteristic curve (AUROC), and the model was calibrated and compared with PREDICT.

Results

The training, test, and validation sets comprised 3276 (499 progressions, 202 breast cancer-specific deaths, and 261 all-cause deaths within 5-year follow-up), 1405 (211 progressions, 94 breast cancer-specific deaths, and 129 all-cause deaths), and 612 (109 progressions, 33 breast cancer-specific deaths, and 37 all-cause deaths) women, respectively. The AUROC values for disease progression, cancer-specific mortality, and all-cause mortality were 0.76, 0.88, and 0.82 for training set; 0.79, 0.80, and 0.83 for the test set; and 0.79, 0.84, and 0.88 for the validation set, respectively. Calibration analysis demonstrated good agreement between predicted and observed events within 5 years. Comparable AUROC and calibration results were confirmed in different age, residence status, and receptor status subgroups. Compared with PREDICT, our model showed similar AUROC and improved calibration values.

Conclusions

Our prognostic model exhibits high discrimination and good calibration. It may facilitate prognosis prediction and clinical decision making for patients with breast cancer in China.

The Mitotic Activity Index in combination with Her2neu: a strong prognosticator in breast cancer

PURPOSE

The aim of this study is to evaluate the prognostic value of the Mitotic Activity Index (MAI) in combination with the human epidermal growth factor receptor (Her2) for distant metastases-free survival (DMFS) and disease-specific survival (DSS) in breast cancer and compare it with the immunohistochemically (IHC) profile types.

METHODS

Analyses were based on 2.923 breast-conserving breast cancer specimens with known MAI, Her2 status, and hormone receptor status, resulting in 2.678 Her2MAI combinations, MAI ≤ 12/Her2negative, MAI > 12/Her2negative, MAI > 12/Her2positive, and MAI ≤ 12/Her2positive, and 2.560 IHC profile types, luminal A, luminal B, triple negative, and non-luminal Her2positive.

RESULTS

For DMFS, the MAI > 12/Her2negative combination showed a significantly worse outcome in multivariate analyses compared to the MAI ≤ 12/Her2negative combination. None of the IHC profile types showed significantly different outcomes for DMFS and DSS as compared to luminal A. We performed a separate analysis on age and lymph node status. The significance of MAI > 12/Her2negative seems to be limited to women ≤ 55 years for both DMFS and DSS. However, with respect to DSS, this seems to be limited to node negative cases. The IHC profile types for DSS, luminal B showed a significantly worse outcome for women > 55 years compared to that for luminal A, although it showed rather wide confidence interval.

CONCLUSION

The MAI > 12/Her2negative combination seems to be a strong prognosticator for DMFS and DSS, particularly for women ≤ 55 years. However, none of the IHC profile types seems to be a prognosticator in breast cancer.

A literature review of treatment-specific clinical prediction models in patients with breast cancer

Despite advances in the breast cancer treatment, significant variability in patient outcomes remain. This results in significant stress to patients and clinicians. Treatment-specific clinical prediction models allow patients to be matched against historical outcomes of patients with similar characteristics; thereby reducing uncertainty by providing personalised estimates of benefits, harms, and prognosis. To achieve this objective, models need to be clinical-grade with evidence of accuracy, reproducibility, generalizability, and be user-friendly. A structured search was undertaken to identify treatment-specific clinical prediction models for therapeutic or adverse outcomes in breast cancer using clinicopathological data. Significant gaps in the presence of validated models for available treatments was identified, along with gaps in prediction of therapeutic and adverse outcomes. Most models did not have user-friendly tools available. With the aim being to facilitate the selection of the best medicine for a specific patient and shared-decision making, future research will need to address these gaps.

Winter is over: The use of Artificial Intelligence to individualise radiation therapy for breast cancer

Artificial intelligence demonstrated its value for automated contouring of organs at risk and target volumes as well as for auto-planning of radiation dose distributions in terms of saving time, increasing consistency, and improving dose-volumes parameters. Future developments include incorporating dose/outcome data to optimise dose distributions with optimal coverage of the high-risk areas, while at the same time limiting doses to low-risk areas. An infinite gradient of volumes and doses to deliver spatially-adjusted radiation can be generated, allowing to avoid unnecessary radiation to organs at risk. Therefore, data about patient-, tumour-, and treatment-related factors have to be combined with dose distributions and outcome-containing databases.

The intra-tumoural stroma in patients with breast cancer increases with age

PURPOSE

The tumour microenvironment in older patients is subject to changes. The tumour-stroma ratio (TSR) was evaluated in order to estimate the amount of intra-tumoural stroma and to evaluate the prognostic value of the TSR in older patients with breast cancer (≥ 70 years).

METHODS

Two retrospective cohorts, the FOCUS study (N = 619) and the Nottingham Breast Cancer series (N = 1793), were used for assessment of the TSR on haematoxylin and eosin stained tissue slides.

RESULTS

The intra-tumoural stroma increases with age in the FOCUS study and the Nottingham Breast Cancer series (B 0.031, 95% CI 0.006-0.057, p = 0.016 and B 0.034, 95% CI 0.015-0.054, p < 0.001, respectively). Fifty-one per cent of the patients from the Nottingham Breast Cancer series < 40 years had a stroma-high tumour compared to 73% of the patients of ≥ 90 years from the FOCUS study. The TSR did not validate as an independent prognostic parameter in patients ≥ 70 years.

CONCLUSIONS

The intra-tumoural stroma increases with age. This might be the result of an activated tumour microenvironment. The TSR did not validate as an independent prognostic parameter in patients ≥ 70 years in contrast to young women with breast cancer as published previously.

Validation of the online prediction model CancerMath in the Dutch breast cancer population

PURPOSE

CancerMath predicts the expected benefit of adjuvant systemic therapy on overall (OS) and breast cancer-specific survival (BCSS). Here, CancerMath was validated in Dutch breast cancer patients.

METHODS

All operated women diagnosed with stage I-III primary invasive breast cancer in 2005 were identified from the Netherlands Cancer Registry. Calibration was assessed by comparing 5- and 10-year predicted and observed OS/BCSS using χ² tests. A difference > 3% was considered as clinically relevant. Discrimination was assessed by area under the receiver operating characteristic (AUC) curves.

RESULTS

Altogether, 8032 women were included. CancerMath underestimated 5- and 10-year OS by 2.2% and 1.9%, respectively. AUCs of 5- and 10-year OS were both 0.77. Divergence between predicted and observed OS was most pronounced in grade II, patients without positive nodes, tumours 1.01-2.00 cm, hormonal receptor positive disease and patients 60-69 years. CancerMath underestimated 5- and 10-year BCSS by 0.5% and 0.6%, respectively. AUCs were 0.78 and 0.73, respectively. No significant difference was found in any subgroup.

CONCLUSION

CancerMath predicts OS accurately for most patients with early breast cancer although outcomes should be interpreted with care in some subgroups. BCSS is predicted accurately in all subgroups. Therefore, CancerMath can reliably be used in (Dutch) clinical practice.

Facilitating validation of prediction models: a comparison of manual and semi-automated validation using registry-based data of breast cancer patients in the Netherlands

Background

Clinical prediction models are not routinely validated. To facilitate validation procedures, the online Evidencio platform (https://www.evidencio.com) has developed a tool partly automating this process. This study aims to determine whether semi-automated validation can reliably substitute manual validation.

Methods

Four different models used in breast cancer care were selected: CancerMath, INFLUENCE, Predicted Probability of Axillary Metastasis, and PREDICT v.2.0. Data were obtained from the Netherlands Cancer Registry according to the inclusion criteria of the original development population. Calibration (intercepts and slopes) and discrimination (area under the curve (AUC)) were compared between semi-automated and manual validation.

Results

Differences between intercepts and slopes of all models using semi-automated validation ranged from 0 to 0.03 from manual validation, which was not clinically relevant. AUCs were identical for both validation methods.

Conclusions

This easy to use semi-automated validation option is a good substitute for manual validation and might increase the number of validations of prediction models used in clinical practice. In addition, the validation tool was considered to be user-friendly and to save a lot of time compared to manual validation. Semi-automated validation will contribute to more accurate outcome predictions and treatment recommendations in the target population.

Electronic supplementary material

The online version of this article (10.1186/s12874-019-0761-5) contains supplementary material, which is available to authorized users.

Ultrasound for Breast Cancer Screening in High-Risk Women: Results From a Population-Based Cancer Screening Program in China

Background: Ultrasound is an important modality for breast cancer screening. However, the evidence on the effectiveness of ultrasound screening in population-based cancer screening program in lacking. We aimed to evaluate the diagnostic yield of ultrasound screening in a population-based breast cancer screening in China.

Methods: The analyses were conducted in the context of the Cancer Screening Program in Urban China, which recruited 1,938,996 eligible participants aged 40–69 years from 16 provinces in China from 2012 to 2016. We included 72,250 women assessed to be high-risk for breast cancer who undertook ultrasound screening per study protocol. Diagnostic yield according to the Breast Imaging Reporting and Data System (BI-RADS) was evaluated. Risk factors associated with the positive findings of ultrasound were also explored by univariate and multivariable logistic regression analyses.

Results: Overall, there were 9,765 (13.51%) women had positive findings of ultrasound screening, including 8,487 (11.75%), 1,210 (1.67%), and 68 (0.09%) of BI-RADS categories of III, IV, and V, respectively. Younger ages, late age of 1st live birth and short-term breast feeding were found to be positively associated with positive findings under ultrasound in multivariate analyses stratified by menopause status and family history of breast cancer. Multivariable prediction models were constructed and yielded only modest prediction accuracy, with AUCs around 0.55.

Conclusions: We found the diagnostic yield of ultrasound screening for breast cancer in high-risk population was satisfactory. Prediction models based on environmental risk factors had limited prediction accuracy and need to be improved in the future.