Do online prognostication tools represent a valid alternative to genomic profiling in the context of adjuvant treatment of early breast cancer? A systematic review of the literature

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.

Correlation Between Oncotype DX, PREDICT and the Nottingham Prognostic Index: Implications for the Management of Early Breast Cancer

Introduction Breast cancer remains the most common cancer diagnosis in the UK. The current clinical practice utilises two different types of modalities to estimate the prognosis, risk of recurrence and benefit from adjuvant chemotherapy treatment in patients with early breast cancer. The first set of modalities includes risk calculators based on clinicopathological features, e.g. PREDICT or the Nottingham Prognostic Index (NPI); the second includes genetic profiling of tumour tissue using Oncotype DX (ODX; Genomic Health, Redwood City, CA) testing. PREDICT, NPI and ODX stratify breast cancers into high-, intermediate- and low-risk categories to help guide adjuvant chemotherapy treatment decisions. This study compares PREDICT, NPI and ODX Recurrence Scores (RS), with the aim of assessing 1) the correlation between the RS for PREDICT, NPI and ODX and 2) whether early breast cancer patients are stratified into similar risk categories by all three modalities. Methods This retrospective study included early breast cancer patients treated at a National Health Service (NHS) hospital over a 12-month period (October 1, 2017 to September 30, 2018). Inclusion criteria: consecutive patients with estrogen receptor (ER)-positive, human epidermal growth factor receptor 2 (HER2)-negative and lymph node-negative breast cancer. All patients were discussed at the local multidisciplinary team (MDT) meeting and underwent ODX testing. Exclusion criteria: patients without ODX test scores; patients with an in-breast recurrence; patients who did not undergo a sentinel lymph node biopsy (SLNB); and patients with ductal carcinoma in situ (DCIS) only. NPI and PREDICT scores were calculated for each patient using online tools, and ODX data was obtained through Genomic Health and MDT records. Patients were risk-stratified into high, intermediate and low risk of recurrence groups based on their PREDICT, NPI and ODX scores. The thresholds for risk stratification were based on current practice, which is evidence-based. Correlations between PREDICT, NPI and ODX scores were analysed using Spearman's correlation coefficient. Results Forty-six patients (mean age: 56 years), with a total of 57 early breast cancers, underwent ODX testing. Risk categories generated by PREDICT very strongly correlated with NPI for all patients (r=0.92; P<0.0001). However, the RS generated by ODX testing only strongly correlated for patients with low-risk PREDICT scores (r=0.51; P=0.0134), while no correlation between RS and PREDICT was observed for patients with intermediate- or high-risk PREDICT scores (r=-0.0064; P=0.9767). Similar results were seen between NPI and RS. Overall, only 19/46 (41.3%) patients had an RS which corresponded to PREDICT risk category, while 18/46 (39.1%) patients had an RS that indicated a higher risk of recurrence than PREDICT, and 9/46 (19.6%) patients had an RS indicating a lower risk of recurrence than PREDICT. Similar results were found when comparing RS and NPI. Conclusion The risk of recurrence estimated by ODX in patients deemed low risk by PREDICT or NPI highly correlated, while no such correlation existed in patients with an estimated intermediate- or high-risk breast cancer. In PREDICT- or NPI-estimated intermediate- and high-risk patients, ODX provided valuable additional prognostic information to guide adjuvant treatment, while the potential avoidance of ODX testing in low-risk patients presents significant cost-savings.

The Utility of Oncotype DX for Adjuvant Chemotherapy Treatment Decisions in Estrogen Receptor-positive, Human Epidermal Growth Factor Receptor 2-negative, Node-negative Breast Cancer

Introduction Breast cancer is the most common cancer diagnosis in the UK. Recently, there has been a reduction in breast cancer-specific mortality and recurrence attributed, in part, to the delivery of adjuvant chemotherapy. The National Institute for Health and Care Excellence (NICE) recommends the use of genetic profiling with Oncotype DX (ODX) to guide decisions to offer adjuvant chemotherapy after surgery in intermediate-risk early breast cancer patients. This study aimed to evaluate the utility of ODX testing in routine clinical practice in a National Health Service (NHS) hospital.  Methods Consecutive early breast cancer patients, identified through the multidisciplinary team (MDT) records, treated in our institution over 12 months (October 2017-September 2018) were included. PREDICT and Nottingham prognostic index (NPI) scores (from online clinicopathological recurrence risk tools) were calculated for each patient, and ODX data obtained through Genomic Health, Inc. (Redwood City, California). Patients were divided into two groups, those that underwent ODX testing (ODX group) and those that did not (non-ODX group). Descriptive statistics were used to analyse patient and tumour characteristics. The Gaussian distribution of each data set was assessed using the Anderson-Darling test. For comparisons between patient groups, the non-parametric equivalent of the two-tailed t-test (Mann-Whitney) was used. Dichotomous variables (e.g. chemotherapy decisions) were compared using chi-squared tests. Results One-hundred thirty-three patients (mean age 62 years) treated for 152 early breast cancers, were included in the final analysis. Breast cancers in the ODX group were of greater median tumour size (24 vs 16 mm; P<0.0001) and higher median tumour grade (3 vs 2; P<0.0001). PREDICT scores (3 vs 1, P<0.0001) and NPI scores (3.40 vs 2.30, P<0.0001) for the ODX group were also significantly higher than the non-ODX group. A greater proportion of patients were offered chemotherapy in the ODX group (39.9% vs 6.9%, P<0.001). However, for the PREDICT-calculated intermediate-risk patients, ODX testing resulted in a lower proportion of patients being offered chemotherapy compared to the intermediate-risk patients who were not genetically profiled (54.5% vs 83.3%, P=0.3547), although this result was not statistically significant. Conclusions Patients selected for ODX testing were younger, with significantly higher-grade and larger-sized tumours compared to patients not selected for genetic profiling. ODX testing significantly impacted the delivery of chemotherapy, as the recurrence score generated through ODX testing guided the final decision.

Communicating prognosis to women with early breast cancer - overview of prediction tools and the development and pilot testing of a decision aid

Background

Shared decision-making in oncology requires information on individual prognosis. This comprises cancer prognosis as well as competing risks of dying due to age and comorbidities. Decision aids usually do not provide such information on competing risks. We conducted an overview on clinical prediction tools for early breast cancer and developed and pilot-tested a decision aid (DA) addressing individual prognosis using additional chemotherapy in early, hormone receptor-positive breast cancer as an example.

Methods

Systematic literature search on clinical prediction tools for the effects of drug treatment on survival of breast cancer. The DA was developed following criteria for evidence-based patient information and International Patient Decision Aids Standards. We included data on the influence of age and comorbidities on overall prognosis. The DA was pilot-tested in focus groups. Comprehension was additionally evaluated through an online survey with women in breast cancer self-help groups.

Results

We identified three prediction tools: Adjuvant!Online, PREDICT and CancerMath. All tools consider age and tumor characteristics. Adjuvant!Online considers comorbidities, CancerMath displays age-dependent non-cancer mortality. Harm due to therapy is not reported.

Twenty women participated in focus groups piloting the DA until data saturation was achieved. A total of 102 women consented to participate in the online survey, of which 86 completed the survey. The rate of correct responses was 90.5% and ranged between 84 and 95% for individual questions.

Conclusions

None of the clinical prediction tools fulfilled the requirements to provide women with all the necessary information for informed decision-making. Information on individual prognosis was well understood and can be included in patient decision aids.

Electronic supplementary material

The online version of this article (10.1186/s12913-019-3988-2) contains supplementary material, which is available to authorized users.

Impact of Micrometastatic Axillary Nodes on Survival of Breast Cancer Patients with Tumors ≤2 cm

BACKGROUND

This study investigated the impact of pN1mi disease on the survival of T1 breast cancer patients and examined the clinical usefulness of the online PREDICT tool and updated staging system.

METHODS

The node stages of 2344 patients were divided into pN0, pN1mi, and pN1a. Clinicopathological parameters and survival outcomes were retrospectively analyzed. Data for 111 micrometastatic diseases were applied to the PREDICT version 2.0 and re-classified using the 8th edition of the cancer staging manual.

RESULTS

Univariable analyses demonstrated worse disease-free and overall survival rates for patients with node-positive cancer; however, the significance was not maintained in multivariable analyses. Chemotherapy improved outcomes in patients with node-positive and non-luminal A-like subtype cancers. The PREDICT tool demonstrated good performance when estimating the 5-year overall survival for pN1mi disease (area under the receiver operating characteristic curve, 0.834). According to the updated staging system, 74% of cases were down-staged to IA, and clearly splitting survival curves were identified.

CONCLUSION

pN1mi disease alone did not adversely affect survival outcomes. Biologic and treatment factors determined outcomes in cases of small-volume node micrometastasis. The PREDICT tool or new staging classification could help predict the survival of patients with micrometastatic sentinel nodes.

Can Complete Axillary Node Dissection Be Safely Omitted in Patients with Early Breast Cancer When the Sentinel Node Biopsy Is Positive for Malignancy? An Update for Clinical Practice

The sentinel lymph node biopsy (SLNB) has become the new standard-of-care for patients with clinically node-negative invasive breast cancer. The focused examination of fewer lymph nodes in addition to improvements in histopathological and molecular analysis have increased the rate at which micrometastases and isolated tumor cells are identified. We reviewed the literature and summarized the evidence regarding the need for complete axillary lymph node dissection (ALND) following the finding of a positive sentinel node biopsy through the identification of the most important outcomes and evaluation of quality of evidence. The article focuses on the safe omission of complete ALND when the axillary lymph nodes contain macrometastases and provides an overview of the topic primarily based on level 1 evidence derived from randomized clinical trials with a critical appraisal of the ACOSOG Z0011 trial.

Prognostic & predictive factors for planning adjuvant chemotherapy of early-stage breast cancer

Breast cancer is a heterogeneous disease and may present with different clinical and biological characteristics. At present, breast cancer is divided into molecular subgroups besides its histopathological classification. Decision for adjuvant chemotherapy is made based on not only histopathological characteristics but also molecular and genomic characteristics using indices, guidelines and calculators in early-stage breast cancer. Making a treatment plan through all these prognostic and predictive methods according to risk categories aims at preventing unnecessary or useless treatments. In this review, an attempt to make a general assessment of prognostic and predictive methods is made which may be used for planning individualized therapy and also the comments of the guidelines used by the oncologists worldwide on these methods.

Optimizing adjuvant endocrine therapy for early ER+ breast cancer: An update for surgeons

INTRODUCTION

The optimal duration of adjuvant endocrine therapy in early ER + breast cancer has been controversial. This article aims to provide an overview of the evidence.

METHODS

A search of the literature was conducted via MEDLINE using appropriate keywords. Eligible studies were screened and relevant articles were selected for this report.

RESULTS

Studies investigating the role of extended adjuvant tamoxifen beyond 5 years have revealed mixed results depending on the proportion of node positivity. In postmenopausal women, aromatase inhibitors (AIs) for 5 years are superior to tamoxifen. Extending the use of AIs beyond 5 years seem to reduce the risk of relapse in postmenopausal women with node positive disease. The addition of bisphosphonates to counteract AI-related osteopenia may further improve overall and disease-free survival. Women younger than 40 years seem to benefit from ovarian suppression combined with tamoxifen or exemestane.

CONCLUSIONS

An individualised approach is required for every patient. The adverse effects of endocrine therapy should be weighed against the potential benefits of extended therapy to better inform decision-making.