Comparison of risk assessment in 1652 early ER positive, HER2 negative breast cancer in a real-world data set: classical pathological parameters vs. 12-gene molecular assay (EndoPredict)

Background Risk assessment on the molecular level is important in predictive pathology to determine the risk of metastatic disease for ERpos, HER2neg breast cancer. The gene expression test EndoPredict (EP) was trained and validated for prediction of a 10-year risk of distant recurrence to support therapy decisions regarding endocrine therapy alone or in combination with chemotherapy. The EP test provides the 12-gene Molecular Score (MS) and the EPclin-Score (EPclin), which combines the molecular score with tumor size and nodal status. In this project we investigated the correlation of 12-gene MS and EPclin scores with classical pathological markers. Methods EndoPredict-based gene expression profiling was performed prospectively in a total of 1652 patients between 2017 and 2020. We investigated tumor grading and Ki67 cut-offs of 20% for binary classification as well as 10% and 30% for three classes (low, intermediate, high), based on national and international guidelines. Results 410 (24.8%) of 1652 patients were classified as 12-gene MS low risk and 626 (37.9%) as EPclin low risk. We found significant positive associations between 12-gene MS and grading (p < 0.001), EPclin and grading (p = 0.001), 12-gene MS and Ki67 (p < 0.001), and EPclin and Ki67 (p < 0.001). However, clinically relevant differences between EP test results, Ki67 and tumor grading were observed. For example, 118 (26.3%) of 449 patients with Ki67 > 20% were classified as low risk by EPclin. Same differences were seen comparing EP test results and tumor grading. Conclusion In this study we could show that EP risk scores are distributed differentially among Ki67 expression groups, especially in Ki67 low and high tumors with a substantial proportion of patients with EPclin high risk results in Ki67 low tumors and vice versa. This suggests that classical pathological parameters and gene expression parameters are not interchangeable, but should be used in combination for risk assessment.

FRI0057 A MODEL FOR QUANTIFYING THE EFFECT OF INFLAMMATION ON CARDIOVASCULAR DISEASE RISK PREDICTION IN RA PATIENTS

Background:Patients with rheumatoid arthritis (RA) are at increased risk for cardiovascular disease (CVD)[1]. Quantifying the effect of inflammation on CVD risk is important because rheumatologists can reduce inflammation with effective RA medications. A new score has been developed for predicting the risk for a CVD event (MI, stroke or CV death) in RA patients. It combines serological measures of inflammation (the multi-biomarker disease activity [MBDA] score, a measure of RA disease activity; and three individual biomarkers [TNF-RI, MMP-3 and leptin]), with age and four conventional CVD risk factors (smoking, hypertension, diabetes and history of a high- risk CVD condition)[2]. To gain insight into the potential effect that treating inflammation may have on the CVD risk score, it would be useful to know how the score is affected by the level of inflammation.Objectives:Explore the quantitative contribution of inflammation to CVD risk score in individual RA patients.Methods:To quantify the effect of inflammation on the CVD risk score across a range of MBDA scores, a commercial dataset of 177,486 RA patients with ≥2 MBDA tests between October 2010 and June 2019 was split 2:1 into training and validation datasets. Curves showing variation in the CVD risk score across the spectrum of all possible MBDA scores (1-100) were generated for canonical patient types differing in the number of conventional risk factors (0 to 4) and age (45, 55, 65, 75, 85 years). To generate these curves, the contributions of TNF-RI, MMP-3 and leptin to the CVD risk score were treated in aggregate (denoted the molecular score) and estimated using a linear regression model of the difference in molecular scores vs. the difference in MBDA scores. This model for the molecular score was fit in the training dataset, then in the full dataset, with dataset (training or validation) and the interaction between dataset and change in MBDA score included as additional predictor variables. The method was considered validated if the F-test for the interaction variable was not significant at the 0.05 level.Results:The model for estimating the molecular score from the MBDA scores was validated and shown to fit the data well (Figure 1). The estimated molecular score was applied to the CVD risk score algorithm to generate curves that show how CVD risk score varies with MBDA score for several distinct patient types. These curves demonstrate that the predicted 3-year CVD risk increases continuously and markedly with increasing level of inflammation, as represented by the MBDA score (Figure 2). Age and the number of conventional risk factors also affected the predicted CVD risk, with older patients (Figure 2a) and those with more conventional risk factors (Figure 2b) being at higher risk for a CVD event.Conclusion:The level of CVD risk predicted by a new prognostic test for RA patients depends not only on conventional risk factors, which are relatively time invariant, but also varies greatly due to inflammation, which can potentially be reduced with RA treatment.References:[1]Agca et al (2017).Ann Rheum Dis.76(1):17-28. doi: 10.1136/annrheumdis-2016-209775.[2]Curtis JR, Xie F, Crowson CS et al. (2019) ACR meeting abstract #446Disclosure of Interests:Rotem Ben-Shachar Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Darl Flake Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Richard Bamford Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Brent Mabey Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Eric Sasso Shareholder of: Myriad Genetics, Inc., Employee of: Myriad Genetics, Inc., Jeffrey Curtis Grant/research support from: AbbVie, Amgen, Bristol-Myers Squibb, Corrona, Janssen, Lilly, Myriad, Pfizer, Regeneron, Roche, UCB, Consultant of: AbbVie, Amgen, Bristol-Myers Squibb, Corrona, Janssen, Lilly, Myriad, Pfizer, Regeneron, Roche, UCB

Evaluation of the 12-Gene Molecular Score and the 21-Gene Recurrence Score as Predictors of Response to Neo-adjuvant Chemotherapy in Estrogen Receptor-Positive, HER2-Negative Breast Cancer

Background Neo-adjuvant chemotherapy (NaCT) facilitates complete surgical resection in locally advanced breast cancer. Due to its association with improved outcome, complete pathologic response (pCR) to neo-adjuvant treatment has been accepted as a surrogate for long-term outcome in clinical trials of human epidermal growth factor receptor 2 (HER2)-positive, triple-negative, or luminal B breast cancer patients. In contrast, NaCT is effective in only ~ 7–10% of estrogen receptor (ER)-positive, HER2-negative disease. Response biomarkers would enable such patients to be selected for NaCT. Methods Two commercially available breast cancer prognostic signatures [12-gene molecular score (MS) and the 21-gene Recurrence Score (RS)] were compared in their ability to predict pCR to NaCT in ER-positive, HER2-negative breast cancer in six public RNA expression microarray data sets. Scores were approximated according to published algorithms and analyzed by logistic regression. Results Expression data were available for 764 ER-positive, HER2-negative breast cancer samples, including 59 patients with pCR. The two scores were well correlated. Either score was a significant predictor of pCR (12-gene MS p = 9.4 × 10−5; 21-gene RS p = 0.0041). However, in a model containing both scores, the 12-gene MS remained significant (p = 0.0079), while the 21-gene RS did not (p = 0.79). Conclusions In this microarray study, two commercial breast cancer prognostic scores were significant predictors of response to NaCT. In direct comparison, the 12-gene MS outperformed the 21-gene RS as a predictive marker for NaCT. Considering pCR as surrogate for improved survival, these results support the ability of both scores to predict chemotherapy sensitivity.

Clinical outcomes in men with prostate cancer who selected active surveillance using a clinical cell cycle risk score

Aim: To evaluate active surveillance (AS) selection, safety and durability among men with low-risk prostate cancer assessed using the clinical cell cycle risk (CCR) score, a combined clinical and molecular score. Patients & methods: Initial treatment selection (AS vs treatment) and duration of AS were evaluated for men with low-risk prostate cancer according to the CCR score and National Comprehensive Cancer Network guidelines. Adverse events included biochemical recurrence and metastasis. Results: 82.4% (547/664) of men initially selected AS (median follow-up: 2.2 years), 0.4% (2/547) of whom experienced an adverse event. Two-thirds of patients remained on AS for more than 3 years; patient choice was the most common reason for leaving AS. Conclusion: The CCR score may aid in the identification of men who can safely defer prostate cancer treatment.

High Prognostic Value of a 3-Genes Molecular Score in Non-High Risk Acute Promyelocytic Leukemia Treated with AIDA 2000 Protocol: A Retrospective Study from a Regional Register

BACKGROUND AND AIMS Non high risk acute promyelocytic leukemia (APL) patients, as defined by WBC count at diagnosis, have nowadays a very good prognosis when treated with ATRA plus chemotherapy based protocols, with high rate of complete molecular remission after consolidation therapy. However, a small proportion of patients (roughly 10-15%) will eventually relapse, despite the achievement of molecular CR. In the past years, some groups showed that relapse risk could be predicted by testing for some gene alterations, such as FLT3-ITD, or by break point cluster region (BCR) analysis. However, those results were not confirmed in prospective trials. We retrospectively applied a simple 3 gene based molecular panel to low-intermediate risk acute promyelocytic patients, alongside with BCR analysis, in order to develop a risk score. MATERIALS AND METHODS Fifty-nine low/intermediate risk APL patients, treated with the Italian age-adapted AIDA protocol from January 1st 2004 to December 31st 2014 in our center were retrospectively included in this study. Median age was 47 years (range 19-88), 16 patients were older than 60 years (27%). BCR analysis was available in all patients, 41 patients showed BCR1/2, whereas 18 patients had BCR3 (30%). Molecular profile included determination of FLT3-ITD mutation, WT1 and BAALC gene expression levels and was available in 38/59 patients (64%); bone marrow sample of diagnosis is available for all other 21 patients for further analysis. Cut-off values for WT1 (20000/abl x 104) and BAALC (450/abl x 104) were arbitrarily chosen after pre-analysis and comparison with our published data. Five patients had FLT3-ITD mutation (13%), 18 (45%) patients had higher WT1 expression levels and BAALC was overexpressed in 9 (24%) patients. Pre-analysis showed that FLT3-ITD, low WT1 levels, high BAALC levels but not the presence of BCR3 were associated with higher relapse risk. A molecular score including those 3 genes was built: 16 patients had no risk factors (42%), 17 patients had one risk factor (45%), 5 had two risk factors (13%). RESULTS Fifty-four patients survived induction, all of them achieved CR, in 52 PML/RARα transcript was undetectable after last consolidation therapy. The two patients with molecular persistence of disease received further therapy and then achieved molecular CR. Nine patients (15%) relapsed after a median follow-up of 56 months, 45 patients are alive at the time of analysis (76%). Relapse risk was influenced only by the presence of at least one molecular risk factor (p<0.05), whereas age at diagnosis older than 60 years had only a borderline impact (p=0.105); none of the molecular alterations reached statistical significance if taken alone. Disease free survival (DFS) duration was significantly higher in patients who had none of the molecular risk factor, when compared to patients with at least one (3-years DFS of 100% and 70.8%, respectively, p<0.05, Figure 1). Younger patients had also a longer DFS, with a 3-years DFS of 90.7% and 72% for patients younger or older than 60 years, respectively (p<0.03). Overall survival (OS) analysis led to similar results: patients with no molecular risk factors had a very good outcome when compared to patients with at least one alteration (3-years OS of 100% and 71.5%, respectively, p<0.03). Survival was also influenced by age at diagnosis, as younger patients did significantly better, with 3-years OS of 90% (p<0.05). Again, none of the molecular alterations, if taken alone, had a significant impact on DFS or OS. CONCLUSIONS Even in our small cohort, molecular profile could identify a subset of low-intermediate APL patients at higher risk of relapse, who may take benefit from an intensified consolidation therapy, as it is currently applied to high risk patients defined by WBC at diagnosis. Figure 1. DFS according to molecular risk score Figure 1. DFS according to molecular risk score Disclosures No relevant conflicts of interest to declare.

Molecular stratification model for prognosis in cytogenetically normal acute myeloid leukemia

We have evaluated 9 new molecular markers (ERG, EVI1, MLL-PTD, MN1, PRAME, RHAMM, and WT1 gene-expression levels plus FLT3 and NPM1 mutations) in 121 de novo cytogenetically normal acute myeloblastic leukemias. In the multivariate analysis, high ERG or EVI1 and low PRAME expressions were associated with a shorter relapse-free survival (RFS) and overall survival (OS). A 0 to 3 score was given by assigning a value of 0 to favorable parameters (low ERG, low EVI1, and high PRAME) and 1 to adverse parameters. This model distinguished 4 subsets of patients with different OS (2-year OS of 79%, 65%, 46%, and 27%; P = .001) and RFS (2-year RFS of 92%, 65%, 49%, and 43%; P = .005). Furthermore, this score identified patients with different OS (P = .001) and RFS (P = .013), even within the FLT3/NPM1 intermediate-risk/high-risk subgroups. Here we propose a new molecular score for cytogenetically normal acute myeloblastic leukemias, which could improve patient risk-stratification.

Estimation of the proportion of genetic variance explained by molecular markers

Estimation of the proportion of genetic variance explained by molecular markers (p) plays an important role in basic studies of quantitative traits, as well as in marker-assisted selection (MAS), if the selection index proposed by Lande and Thompson (Genetics 124: 743-756, 1990) is used. Frequently, the coefficient of determination (R2) is used to account for this proportion. In the present study, a simple estimator of p is presented, which is applicable when a multiple regression approach is used, and progenies are evaluated in replicated trials. The associated sampling distribution was obtained and compared with that of R2. Simulations indicated that, when the number of evaluated progenies is small, the statistics are not satisfactory, in general, due to bias and/or low precision. Coefficient R2 was found adequate in situations where p is high. If a large number of progenies is evaluated (say, a few hundreds), then the proposed estimator <img src="http:/img/fbpe/gmb/v21n4/1974f1.jpg" alt="1974f1.jpg (1159 bytes)" align="middle"> appears to be better, with acceptable precision and considerably lower bias than R2. A normal approximation to the sampling distribution of <img src="http:/img/fbpe/gmb/v21n4/1974f1.jpg" alt="1974f1.jpg (1159 bytes)" align="middle"> is given, using Taylor's expansion of the expectation and variance of this statistic. Approximate confidence intervals for p, based on normal distribution, are reasonable, if the number of progenies is large. The use of <img src="http:/img/fbpe/gmb/v21n4/1974f1.jpg" alt="1974f1.jpg (1159 bytes)" align="middle"> in MAS is illustrated for estimation of the weight given to the molecular score, when a selection index is used.