Development and validation of an AI-enabled digital breast cancer assay to predict early-stage breast cancer recurrence within 6 years.

BACKGROUND: Breast cancer (BC) grading plays a critical role in patient management despite the considerable inter- and intra-observer variability, highlighting the need for decision support tools to improve reproducibility and prognostic accuracy for use in clinical practice. The objective was to evaluate the ability of a digital artificial intelligence (AI) assay (PDxBr) to enrich BC grading and improve risk categorization for predicting recurrence. METHODS: In our population-based longitudinal clinical development and validation study, we enrolled 2075 patients from Mount Sinai Hospital with infiltrating ductal carcinoma of the breast. With 3:1 balanced training and validation cohorts, patients were retrospectively followed for a median of 6 years. The main outcome was to validate an automated BC phenotyping system combined with clinical features to produce a binomial risk score predicting BC recurrence at diagnosis. RESULTS: The PDxBr training model (n = 1559 patients) had a C-index of 0.78 (95% CI, 0.76-0.81) versus clinical 0.71 (95% CI, 0.67-0.74) and image feature models 0.72 (95% CI, 0.70-0.74). A risk score of 58 (scale 0-100) stratified patients as low or high risk, hazard ratio (HR) 5.5 (95% CI 4.19-7.2, p < 0.001), with a sensitivity 0.71, specificity 0.77, NPV 0.95, and PPV 0.32 for predicting BC recurrence within 6 years. In the validation cohort (n = 516), the C-index was 0.75 (95% CI, 0.72-0.79) versus clinical 0.71 (95% CI 0.66-0.75) versus image feature models 0.67 (95% CI, 0.63-071). The validation cohort had an HR of 4.4 (95% CI 2.7-7.1, p < 0.001), sensitivity of 0.60, specificity 0.77, NPV 0.94, and PPV 0.24 for predicting BC recurrence within 6 years. PDxBr also improved Oncotype Recurrence Score (RS) performance: RS 31 cutoff, C-index of 0.36 (95% CI 0.26-0.45), sensitivity 37%, specificity 48%, HR 0.48, p = 0.04 versus Oncotype RS plus AI-grade C-index 0.72 (95% CI 0.67-0.79), sensitivity 78%, specificity 49%, HR 4.6, p < 0.001 versus Oncotype RS plus PDxBr, C-index 0.76 (95% CI 0.70-0.82), sensitivity 67%, specificity 80%, HR 6.1, p < 0.001. CONCLUSIONS: PDxBr is a digital BC test combining automated AI-BC prognostic grade with clinical-pathologic features to predict the risk of early-stage BC recurrence. With future validation studies, we anticipate the PDxBr model will enrich current gene expression assays and enhance treatment decision-making.

Correlation of the Ki67 Working Group prognostic risk categories with the Oncotype DX Recurrence Score in early breast cancer.

BACKGROUND: The relationship between Ki67 assessed by immunohistochemistry (IHC) and the Oncotype DX Recurrence Score (RS) is unclear. The objective of this study was to determine the correlation between the 21-gene RS and IHC-measured Ki67 with the prognostic classification groups recommended by the International Ki67 Working Group (IKWG). METHODS: The authors performed a retrospective chart review of women who had hormone receptor (HR)-positive, human epidermal growth factor receptor 2-negative early breast cancer with zero to three positive lymph nodes and both Ki67 and the 21-gene RS performed at their institution from 2013 to 2021. Patients were categorized into low (≤5%), intermediate (6%-29%), and high Ki67 groups (≥30%) according to IKWG recommendations. Overall agreement and risk-stratified agreement between Ki67 and RS were assessed with the proportion of agreement and the κ statistic. RESULTS: The study included 525 patients with HR-positive breast cancer. Among the 49% of patients with intermediate Ki67 values of 6%-29%, the distribution of low (0-10), intermediate (11-25), and high RS (26-100) was 19%, 66%, and 15%, respectively. There was slight agreement (κ = 0.01-0.20) between Ki67 and RS (κ = 0.027) in the overall population, although this was not significant (p = .1985). There was fair agreement (κ = 0.21-0.40) between high Ki67 and RS values (κ = 0.280; p < .0001). A higher progesterone receptor percentage was associated with lower RS values (p > .0001) but not lower Ki67 values. A positive nodal status and a larger tumor size were associated with higher Ki67 values (p = .0059 and p < .0001) but not with RS. CONCLUSIONS: In this group of patients selected to have a 21-gene RS, there was no significant correlation between Ki67 and RS in the overall population, and there was fair agreement between high Ki67 and high RS values. LAY SUMMARY: In patients with early-stage, hormone receptor-positive breast cancer, decisions on adjuvant chemotherapy are based on certain biological features of the cancer and genomic assays such as the Oncotype DX Recurrence Score (RS). The goal of this study was to determine the correlation between Ki67, a marker of proliferation, and the Oncotype DX RS, a 21-gene assay demonstrated to be predictive of an adjuvant chemotherapy benefit in patients with early-stage breast cancer. In 525 patients, the authors did not find a significant correlation between Ki67 and RS.

Prognostic Factors and Survival Outcomes among Patients with Breast Cancer and Brain Metastases at Diagnosis: A National Cancer Database Analysis.

INTRODUCTION: The aim of this study was to assess for clinicopathologic and socioeconomic features that predict improved survival for patients with advanced breast cancer with synchronous brain metastases at diagnosis. METHODS: We utilized the National Cancer Database (NCDB) to identify all patients with brain metastases present at diagnosis, with adequate information on receptor status (ER, PR, Her2), clinical T stage of cT1-4, clinical M1, with 3,943 patients available for analysis. The association between brain metastases patterns and patient/disease variables was examined by robust Poisson regression model. Cox proportional hazards model was used to quantify the associations between overall survival (OS) and these variables. RESULTS: In univariable analysis, OS was significantly associated with the number of sites of metastases (p < 0.0001). Patients with 2 or more additional extracranial sites of metastases had significantly worse OS (median 8.8 months, 95% confidence interval [CI] 7.8, 9.9) than patients with brain metastases only (median OS 10.6 months, 95% CI 9.4, 12.9) or brain metastases plus one other extracranial site of metastases (median OS 13.1 months, 95% CI 11.8, 14.4). Risk factors which predicted poor prognosis included triple-negative disease, high comorbidity score, poorly differentiated tumors, invasive lobular histology, multi-organ involvement of metastases, and government or lack of insurance. Factors which improve survival include younger age and Hispanic race. DISCUSSION/CONCLUSION: Using a large NCDB, we identified various factors associated with prognosis for patients with brain metastases at the time of breast cancer diagnosis. Insurance status and related socioeconomic challenges provide potential areas for improvement in care for these patients. This information may help stratify patients into prognostic categories at the time of diagnosis to improve treatment plans.

Secondary Invasive Breast Events among Patients with Hormone-Positive Breast Cancer and High-Risk Oncotype DX Recurrence Scores 26-30 and ≥31.

BACKGROUND: The Oncotype DX Recurrence Score (ODx RS) is the most widely adopted genomic assay used to guide treatment for patients with early-stage, hormone-positive (HR+) breast cancer (BC), with higher scores predicting greater risk of recurrence and benefit from chemotherapy. Patients with ODx RS >25 typically recieve adjuvant chemotherapy; however, data regarding efficacy of chemotherapy for reducing recurrence in this population have been mixed. OBJECTIVES: This study aimed to evaluate outcomes of patients with early-stage HR+ BC with high-risk ODx RS (26-30 and ≥31) in order to assess treatment patterns and outcomes. We hypothesized that the benefit of chemotherapy in these groups may be minimal and that select patients may forgo chemotherapy in favor of more aggressive endocrine therapy and ovarian suppression. METHODS: We performed a retrospective analysis of 515 patients with early-stage, HR+ BC with high-risk ODx RS 26-30 and ≥31 treated between 2006 and 2018. Patients were stratified by RS: low-risk (≤10), intermediate-risk (11-25), and high-risk (≥26). The Kaplan-Meier method was used to estimate the time to secondary invasive breast events (SIBE) or distributions overall and among different RS groups with the log rank test used to compare distributions between groups. RESULTS: Rates of chemotherapy administration were 7% among the low-risk group, 18% among the intermediate-risk group, and 83% among high-risk patients with 41 SIBE (8%) reported. When stratified by ODx RS, 5-year rates of SIBE were 4%, 6%, and 16% for low-risk, intermediate-risk, and high-risk RS, respectively. Among the 27 lymph node (LN)-negative patients with ODx RS 26-30, 74% received chemotherapy. The 5-year rate of SIBE was 25% among patients who received chemotherapy and 33% among those who did not (p = 0.5489). Among the 23 LN-negative patients with ODx RS ≥31, 91% of patients received chemotherapy. The 5-year rate of SIBE was 0% both with and without chemotherapy. CONCLUSIONS: There was no statistically significant difference in SIBE for patients with high-risk ODx RS based on chemotherapy treatment. More aggressive endocrine therapy with ovarian suppression has become an alternative to chemotherapy among patients with intermediate-risk ODx RS (16-25). This approach may be useful among patients with high-risk ODx RS, with additional studies needed in this patient population.

Diabetes and Metformin Association with Recurrence Score in a Large Oncotype Database of Breast Cancer Patients.

BACKGROUND: The Oncotype DX® (ODX) is a genomic assay that provides clinicians with a prediction of benefit of chemotherapy in node-negative, tamoxifen-treated breast cancer. However, the relationship between ODX recurrence score (RS) and diabetes, a common comorbidity in breast cancer patients, has been inadequately described in the literature. Specifically, the association of diabetes treatment with metformin and RS is inconclusive, with different studies reporting conflicting results. Because diabetes has been associated with higher RS, it has been suggested that management of diabetes with metformin in breast cancer patients may be associated with a lower RS. OBJECTIVES: We studied a large cohort of early-stage, hormone-positive breast cancer patients to determine if there is an association between RS and metformin treatment. METHODS: In this study, we retrospectively examined the medical records of 514 early-stage, hormone-positive breast cancer patients who had oncotype testing performed between 2007 and 2017. Number (%) or median were used to describe the patients' characteristics between groups and were compared by the Kruskal-Wallis test at a significance level of 5%. RESULTS: Of this cohort, 67 (13%) had a diabetes diagnosis at the time of breast cancer diagnosis, including both diabetes mellitus and pre-diabetes. The median RS for non-diabetic patients was 16 and the median RS for diabetic patients was 15. This difference was not significant, nor was there a statistical difference in RS between diabetic patients taking metformin (median RS = 15) and diabetic patients not taking metformin (median RS = 15). These results held true even when controlling for BMI. CONCLUSIONS: We conclude that neither diabetes diagnosis nor metformin use is associated with a difference in oncotype RS in this population of diabetic patients.

BRCA Mutation Association with Recurrence Score and Discordance in a Large Oncotype Database.

BACKGROUND: The Oncotype DX Breast Cancer Assay is a 21-gene assay used to predict the likelihood of distant recurrence and the benefit of chemotherapy in patients with node-negative, tamoxifen-treated breast cancer. Prior studies demonstrated 7-19% discordance, or a difference between the recurrence score (RS) and tumor grade (TG) in breast cancer patients. BRCA mutated tumors (BRCA+) have been shown to be associated with higher RS as compared to BRCA-negative patients (BRCA-). OBJECTIVES: We developed a large Oncotype RS database to determine if the BRCA mutation status is associated with discordance. METHODS: We identified 723 patients (32 [4%] mutation-positive and 691 [96%] mutation-negative patients) with early-stage, hormone-positive breast cancer treated between 2006 and 2018, with tumor characteristics available for analysis. Discordance was defined as one- or two-step difference between RS (low, intermediate, high risk) and TG (well [WD], moderately [MD], and poorly [PD] differentiated). Mutation positive was defined as BRCA1 deleterious mutation, BRCA2 deleterious mutation, BRCA mutation of unknown type, BRCA variant of undetermined significance (VUS) or other mutation (classified as other VUS). Number (%) or median were used to describe patient characteristics between groups and were compared by the Kruskal-Wallis test at a significance level of 5%. RESULTS: Among these patients, there were 32 (4% of total) who were identified as mutation-positive. Of those patients, 16% had a documented deleterious mutation in BRCA1, 22% in BRCA2, 6% had a BRCA mutation of unknown type (either 1 or 2), 25% were BRCA VUS, and 31% other VUS (most commonly CHEK2 and ATM). The median RS was 23.5 in patients with deleterious BRCA mutations (1, 2 or unknown) versus 16 in patients in the BRCA-negative database, which was statistically significant (p < 0.01). One- and two-step discordance was present in 46 and 8%, respectively, of patients with deleterious BRCA mutations versus 53 and 11%, respectively, in the BRCA-negative database. CONCLUSIONS: Patients with deleterious BRCA mutations demonstrated no difference in rates of discordance as compared to BRCA-negative patients. We further demonstrated that patients with BRCA-positive tumors display higher RS than patients with BRCA-negative tumors.

Correlating Predicted Adjuvant Therapy Benefit and Risk of Recurrence Between Breast Cancer Index (BCI) and the 21-Gene Oncotype DX Recurrence Score (RS).

INTRODUCTION: Breast Cancer Index (BCI) is a genomic assay that evaluates the benefit of extending endocrine therapy (ET) from 5 to 10 years and predicts recurrence risk (RR). We evaluated the association between BCI and Oncotype DX (ODX). PATIENTS: Women with hormone receptor (HR)-positive early-stage breast cancer (EBC) who had BCI and ODX performed were included. METHODS: We performed a retrospective review of women with HR-positive EBC. BCI was categorized as predictive of extended ET versus not and ODX recurrence score (RS) as low (0-10), intermediate (11-25), and high (26-100). Univariate and multivariable logistic and linear regression models assessed the relationship between BCI and ODX, factors associated with each, and discordance between scores. RESULTS: We identified 153 women, 22% were premenopausal and 18% were lymph node positive. The univariate logistic and linear models revealed an association between BCI predictive score and ODX RS (OR 7.84, CI, 2.63-23.36, P < .001) and log of BCI RR (Beta 0.04, CI, 0.02-0.06, P < .001). Seventy-four percent of BCI predictive scores were concordant with ODX RS and 83% of BCI RR was concordant with ODX RR. In a univariate logistic regression model, BCI predictive of ET benefit was associated with discordance (OR 28.00, CI, 10.58-74.02, P < .001). Higher ODX RR was associated with discordance (OR 1.92, CI, 1.42-2.59, P < .001). CONCLUSION: We found a significant association between ODX and BCI predictive and prognostic scores. BCI predictive of extended ET benefit was associated with discordance with ODX RS. Higher predicted RR on ODX was associated with discordance with BCI predicted RR.

Impact of the COVID-19 Pandemic on Cancer Care and Quality of Life for Patients With Breast and Gynecologic Malignancies: A Single-Center Survey-Based Study.

This study evaluated the mental health and cancer treatment-related impact of the first wave of the COVID-19 pandemic on patients with breast and gynecologic cancers. An 18-question survey was administered in June 2020 at a New York City-based cancer center to assess the quality of life (QOL) and overall health (OH) during both the pandemic time period from March 1, 2020, through June 30, 2020, and the pre-pandemic period (prior to March 1, 2020). Survey questions were answered on a 5-point Likert scale and a 7-point EORTC QLQ-C30 QOL scale. Differences in mean QOL and OH scores were evaluated using a paired t-test. QOL and OH were significantly worsened by the pandemic, with significant increases in anxiety, depression, and mood swings.

Patient Perception of Telehealth Services for Breast and Gynecologic Oncology Care during the COVID-19 Pandemic: A Single Center Survey-based Study.

Prior to the coronavirus disease 2019 (COVID-19) pandemic, telehealth was rarely utilized for oncologic care in metropolitan areas. Our large New York City based outpatient breast/gynecologic cancer clinic administered an 18-question survey to patients from March to June 2020, to assess the perceptions of the utility of telehealth medicine. Of the 622 patients, 215 (35%) completed the survey, and of the 215 respondents, 74 (35%) had participated in a telehealth visit. We evaluated the use of telehealth services using the validated Service User Technology Acceptability Questionnaire. Sixty-eight patients (92%) reported that telehealth services saved them time, 54 (73%) reported telehealth increased access to care, and 58 (82%) reported telehealth improved their health. Overall, 67 (92%) of patients expressed satisfaction with the use of telehealth services for oncologic care during the COVID-19 pandemic. Telehealth services should be carefully adopted as an addition to in-person clinical care of patients with cancer.