Ruddy, Kathryn J.

e13552 Background: Real-world data from Electronic Health Records (EHR) have been widely used for patient identification to build study cohorts for clinical research. Traditionally, diagnosis codes in the EHR, such International Classification of Diseases (ICD), are used to identify the target patients. However, the accuracy of this approach is dependent on the accuracy of ICD coding, with potential errors especially for tumor types that are frequent locations for metastases (which may contribute to mis-coding). In this study, we attempted to develop a Machine Learning (ML) based approach on EHR data to improve the accuracy of identification of patients with lung cancer. Methods: We used survey respondents in the Enhanced, EHR-facilitated Cancer Symptom Control (E2C2, NCT03892967) cluster-randomized trial at Mayo Clinic as our initial pan-cancer cohort. E2C2 includes adults receiving Medical Hematology/Oncology care for a solid or liquid tumor at Mayo Clinic. We collected cancer diagnoses from the individually abstracted Mayo Clinic Cancer Registry to annotate the cancer type for the patients. Lung cancer related ICD-9 (162.X) and ICD-10 (C34.X) codes were used to build a search query on the Mayo’s EHR to find target patients from the E2C2 cohort, and to investigate the ICD-based lung cancer patient identification performance. Diagnosis, radiation oncology treatment (CPT 77261 - 77799), and antineoplastic drug administration data were collected from EHR as variables. Logistic Regression (LR), Support vector machine (SVM), Random Forest (RF), and eXtreme Gradient Boosting (XGB) were selected to build models for lung cancer patient identification. 10-fold cross-validation was implemented to assess the models. Precision, Recall, F1 Score, and area under the curve (AUC) were selected to measure the performance. Results: We collected 13,893 patients with a specific cancer diagnosis, and 1,394 were identified as having lung cancer. The identification performance across different methods is shown. The ICD-based method only had 0.65 precision. It means we collected a lot of false positive cases (other cancer but no lung cancer patient), as we mentioned in the background. SVM gets the best precision results, but its recall and F1 score are not good enough. XGB shows the best F1 Score and AUC, which also means this method achieved the best and most balanced performance. Conclusions: In this study, we found that XGB-based methods achieved the best identification performance for lung cancer. In the future, we will investigate if this is also true for the identification of other cancer types. [Table: see text]

1592 Background: While ongoing efforts to decrease opioid use have dramatically altered the prescribing landscape, the impact on patients with cancer is not well understood. We examined trends in new and persistent opioid use in opioid-naïve older patients with and without cancer in the United States. Methods: This retrospective cohort study used SEER-Medicare linked data for patients with solid tumors malignancies diagnosed from 2012 through 2017, and a random sample of patients without cancer. We included patients who did not receive opioids for 3 months prior to the diagnosis or randomly assigned index date. We identified new opioid use (claim in 0-90 days after diagnosis/index date), early persistent opioid use (claims in days 0-90 and 91-180), and long-term persistent opioid use (claims in days 0-90 and 91-180 and months 13-15). Further analysis focused on 4 clinical strata: non-cancer, metastatic cancer, cancer treated with surgery alone, and cancer treated with surgery plus chemotherapy and/or radiotherapy ("surgery+"). We calculated relative changes in the predicted probability of opioid use from 2012 to 2017 based on a logit model adjusted for patient demographics. Results: We identified 238,470 eligible patients with cancer (mean age 76.3 years). Among patients with cancer, 46.4% had new opioid use; use was higher among those receiving surgery+ (70.6%), those treated with surgery alone (58.8%) and those with metastatic cancer (50.0%), compared to patients without cancer (6.9%). From 2012 to 2017, the predicted probability of new opioid use declined in all groups, with lower relative decline in the cancer subgroups (6.8% surgery+; 9.5% surgery alone; 10.2% metastatic cancer) compared to the non-cancer comparator (24.7%). For instance, in patients with metastatic cancer, predicted probability of new use decreased from 52.4 to 47.1% (10.2%) vs. 8.0 to 6.0% (24.7%) in patients without cancer. Relative decline in predicted probability of long-term persistent opioid use was lower in the surgical cancer strata (44.2% surgery+; 46.1% surgery alone) vs. the non-cancer comparator (54.0%). Conclusions: While opioid use remains common in patients with cancer, new and persistent opioid use declined over time in all patients, though with steeper relative declines in non-cancer compared to cancer populations. Our study suggests important but tempered effects of the changing opioid climate on patients with cancer.[Table: see text]

1509 Background: Medicaid expansion under the Affordable Care Act has been shown to increase insurance coverage and early-stage cancer diagnoses for young adults (YAs), the age group with the highest uninsured rate. We examined whether Medicaid expansion was associated with increased receipt of timely, guideline-concordant treatment and survival among YA women newly diagnosed with breast cancer, the most common YA cancer diagnosis. Methods: Using the National Cancer Database, we identified 51,675 women aged 18-39 years who were diagnosed with breast cancer in 2011-2018. We applied the difference-in-differences (DD) method to estimate outcome changes pre vs. post Medicaid expansion, in expansion- vs. non-expansion states. Linear probability models estimated associations of Medicaid expansion with receipt of guideline-concordant treatment – including any endocrine therapy among women with ER-positive or PR-positive breast cancer, and any chemotherapy or targeted therapy among women with ER-negative and PR-negative breast cancer – and initiation of guideline-concordant treatment (surgery, chemotherapy, or targeted therapy for stage I-III diagnoses, and systemic therapy for stage IV diagnoses) <60 days after diagnosis. Flexible parametric survival models were used to estimate 2-year overall survival. DD models also adjusted for age, race/ethnicity, rurality, zip code-level income, comorbidity, and diagnosis year, with standard errors clustered at the state level. Analyses were stratified by stage at diagnosis (stage I-III vs. IV). Results: Of the subset of women with stage I-III ER-positive or PR-positive breast cancer, the percentage receiving endocrine therapy increased in expansion states (85.20% pre-expansion to 86.46% post-expansion), but decreased in non-expansion states (84.32% to 82.79%), resulting in a net increase of 2.42 percentage points (ppt; 95% CI = 0.56 to 4.28) associated with Medicaid expansion in the adjusted DD model. Among all women with stage I-III breast cancer, the percentage with treatment initiation <60 days after diagnosis decreased less in expansion states (84.10% pre-expansion to 82.36% post-expansion) than in non-expansion states (86.95% to 83.98%), resulting in a net reduction of 1.61 ppt (95% CI = 0.05 to 3.18) in treatment delays associated with Medicaid expansion. Notably, the 2-year overall survival increased in expansion states (96.84% pre-expansion to 96.99% post-expansion), but decreased in non-expansion states (97.24% to 96.30%), resulting in a net survival increase of 1.00 ppt (95% CI = 0.21 to 1.79) associated with Medicaid expansion. Conclusions: Among YA women with stage I-III breast cancer, Medicaid expansion was associated with increased receipt of guideline-concordant treatment, reduced delays in treatment initiation, and improved 2-year survival.

611 Background: Patient reported outcomes (PROs) capture direct feedback from patients in clinical trials. The NIH Patient-Reported Outcomes Measurement Information System (PROMIS) is an open source and validated PRO platform. Data on PROMIS score trajectories over time for breast cancer patients undergoing neoadjuvant chemotherapy (NAC) are limited. Methods: We administered paper surveys to patients enrolled on the I-SPY2 trial, a multicenter, randomized phase 2 trial comparing novel agents to standard NAC in high-risk breast cancer. PRO domains included PROMIS Physical Function, Anxiety, Depression, Applied Cognition and Fatigue. Surveys were administered at baseline (post-consent), C1D1, inter-regimen (~12 wks into NAC), pre-surgery and at 1, 6, 12, and 24 months post-surgery. Responses were scored using the HealthMeasures Scoring Service to generate a standardized T score, with a score of 50 representing the general US population mean. Higher Anxiety, Depression and Fatigue scores represent worse symptoms; higher Physical Function and Applied Cognition scores represent better symptoms. We calculated average scores by time-point and compared changes from baseline to various grouped time points using a paired t-test. Results: From 1/2012 to 7/2020, 1631 patients enrolled in I-SPY 2 with 959 completing the baseline survey. 62% had at least two “on-treatment” surveys, and 52% had at least one follow-up time-point. Average baseline PROMIS scores for Physical Function, Anxiety, Depression, Applied Cognition and Fatigue were 46.7, 55.1, 47.2, 53.3 and 46.0 respectively. Between baseline and the latest on-treatment time point we observed improvement in Anxiety (-5.2) and worsening in Physical Function, Applied Cognition and Fatigue (-5.4, -6.9 and +8.2 respectively). Between baseline and the latest post-surgery time point, we saw improved Anxiety (-5.9) and worsened Physical Function, Applied Cognition and Fatigue (-5.9, -3.6 and +4.3). Depression changed minimally. Conclusions: Using PROMIS, patients on I-SPY2 experienced worse Physical Function, Applied Cognition and Fatigue but improved Anxiety during and after treatment compared to baseline. While changes in Applied Cognition and Fatigue recovered partially after treatment, Physical Function scores did not, suggesting potential for longer term impact on quality of life. Clinical trial information: NCT01042379 . [Table: see text]

505 Background: Increased TILs are associated with better prognosis in triple-negative BC, including in patients (pts) age < 40. However, the role of TILs remains unclear in HR+/HER2- BC and little data exist in young pts, in whom immune microenvironment could be altered by age-related host/tumor differences. We assessed the extent and composition of immune infiltration in HR+ tumors of young women and correlated with clinicopathologic features and survival outcomes. Methods: From a prospective cohort study of women with BC diagnosed age ≤40, we identified those with stage I-III HR+/HER2- BC and available pre-treatment (tx) tumor tissue. Multiplexed immunofluorescence was used to quantify cytotoxic T (CD8+), T helper (Th, CD3+CD8-), T regulatory (Tregs, FOXP3+CD3+) and exhausted T (PD1+CD8+) cells in stroma and tumor. Univariate analyses tested associations between clinicopathologic variables and immune markers by high or low expression, divided based on median. High vs. low TILs were evaluated in Cox regression analyses for invasive breast cancer-free survival (iBCFS), distant disease-free survival (DDFS) and overall survival (OS). Results: In 390 pts, median age was 37 years (21-40), most had grade 2 (51%), T1 (65%), N0 (63%) tumors and 67% received adjuvant chemotherapy. Black pts (n = 17) had higher expression of stromal CD8+ ( P= .010), FOXP3+CD3+ ( P= .027) and PD1+CD8+ TILs ( P= .043); intratumoral TILs did not differ by race. Older age (36-40) was associated with high expression of CD8+ ( P= .033) and PD1+CD8+ TILs ( P= .031) within stroma and CD3+CD8- TILs within tumor ( P= .046). Grade 3 tumors had higher stromal and intratumoral expression of CD3+CD8- ( P= .002; P< .001) and FOXP3+CD3+ TILs ( P= .020; P< .001). No differences in TILs were seen according to recency of pregnancy, BRCA1/2 status or T/N stage. Over a median follow up of 8 years, 85 iBCFS events, 64 DDFS events and 37 deaths occurred. High stromal expression of CD3+CD8- TILs was associated with better iBCFS (HR 0.49, P= .002) and DDFS (HR 0.57, P= .046), which remained significant when adjusted for T/N stage, grade and chemotherapy (iBCFS HR 0.41, P< .001; DDFS HR 0.45, P= .008). High stromal expression of CD3+CD8- and FOXP3+CD3+ TILs was associated with better OS (HR 0.47, P= .038) and iBCFS (HR 0.58, P= .018), respectively, on adjusted analyses only. High intratumoral expression of CD3+CD8- and FOXP3+CD3+ TILs was associated with better iBCFS (HR 0.59, P= .025; HR 0.63, P= .043) after adjustment only. Conclusions: The distribution of TIL subtypes in young women’s HR+ BC varied according to race, age and grade. High stromal and intratumoral expression of Th and Tregs was associated with improved BC outcomes. Characterization of immune cell subsets could help refine the prognostic value of TILs in HR+ BC, particularly in young pts who may benefit from individualized escalated/de-escalated tx strategies.

AbstractBackgroundObesity and inactivity are poor prognostic factors in breast cancer, but less is known regarding physical activity (PA) and weight patterns in young breast cancer survivors.MethodsThe Young and Strong Study was a cluster‐randomized trial evaluating education and support interventions for young women (age <45 years) with newly diagnosed breast cancer. Sites were randomized 1:1 to a Young Women's Intervention (YWI) or a contact‐time control physical activity intervention (PAI). Changes in PA and weight were compared between groups using general estimating equations to evaluate clustered binary and Gaussian data.ResultsA total of 467 patients enrolled between July 2012 and December 2013 across 54 sites. Median age at diagnosis was 40 years (range, 22–45). At baseline, median body mass index (BMI) was 25.4 kg/m2 (range, 16.1–61.1), and participants reported a median of 0 minutes (range, 0–2190) of moderate/vigorous PA/week. PA increased significantly over time in both groups (p < .001), with no difference between groups at any time point. BMI increased modestly but significantly (p < .001) over time in both groups. Provider attention to PA was observed in 74% of participants on PAI and 61% on YWI (p = .145) and correlated with PA at 12 months (median 100 min/week of PA in participants with provider attention to PA vs. 60 min/week in those without, p = .016).ConclusionsIn a cohort of young women with breast cancer, rates of obesity and inactivity were high. PA and BMI increased over time and were not impacted by an educational PA intervention. Findings provide important information for developing lifestyle interventions for young breast cancer survivors.

Abstract Introduction: The utilization of self reported chemotherapy treatment data is common in epidemiologic research but there is little understanding of its accuracy and limitations. Methods: The Mayo Clinic Breast Disease Registry (MCBDR) has over 9,600 participants enrolled within one year of a first breast cancer diagnosis. Participants receive annual follow up surveys through the mail to capture data on breast cancer treatments and outcomes. We compared nurse abstracted medical record chemotherapy data to self reported chemotherapy data from women diagnosed with stage 0 to 3 breast cancer that completed five consecutive annual follow up surveys. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) statistics were assessed for the report of broad chemotherapy receipt (yes, no, DK) within each yearly survey compared to the gold standard medical record data. To assess the accuracy of the self reported use of specific chemotherapy regimens, a sub cohort known as the “congruent sub cohort” was created for each survey year that included subjects that agreed with the nurse abstractors report of their broad chemotherapy usage. Sensitivity, specificity, PPV, NPV, and concordance statistics were calculated for the congruent sub cohorts followed by an assessment of the proportion of subjects providing correct responses for all chemotherapy regimens queried compared to those that had at least one discrepancy with the nurse abstracted data and the proportion of subjects with 0, 1, 2 and 3 or more discrepancies within each survey year. Results: For the 430 survey respondents, the average age was 59.5 years (SD 11.7) with a range of 29 to 94. The mean time between diagnosis and the year 1 follow up survey was 13.5 months, with an average of 61.2 months between diagnosis and the year 5 survey. The respondents were 97% white, 100% female, and 53.5% had a bachelor’s degree or higher. The self-reported receipt of chemotherapy had substantial agreement across all 5 follow up surveys with the nurse abstracted treatment data from the medical record with kappa values ranging from 0.70 (0.63 to 0.76) to 0.64 (0.57 to 0.71). For subjects in the congruent sub cohort (N=125, 120, 122, 121, and 121 subjects across the 5 survey years, respectively), self reported use of specific chemotherapy regimens were reported with an average concordance of 90% across the 5 annual surveys, however only 23% had perfect agreement with nurse abstracted data. The average number of discrepant responses provided by subjects was 1.2, with at least 12% of subjects reporting 3 or more erroneous responses for the specific chemotherapy regimens each survey year. Discussion: This study supports previous findings that self-reported chemotherapy usage may be utilized when resources are unavailable for medical record abstraction, but may have less utility when analyzing specific chemotherapy regimens. Citation Format: Nicole L. Larson, Kathryn J. Ruddy, Feven Abraha, Denise M. Herman, Fergus J. Couch, Janet E. Olson. Accuracy of self reported chemotherapy usage in annual follow up surveys for a breast cancer registry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 748.

Abstract Purpose: Young women’s breast cancers (YBCs), defined herein as breast cancer (BC) diagnosed at age ≤ 40 years, are etiologically and biologically distinct from postmenopausal BCs, demonstrating a higher proportion of estrogen receptor (ER) negative subtypes and a worse prognosis. Studies have not assessed whether risk factors such as recent parity and body mass index (BMI) are linked to protein expression profiles in YBCs. Accordingly, we examined protein expression assessed with GeoMx® Digital Spatial Profiling (DSP) in 640 YBCs prepared as tissue microarrays (TMAs) in the Young Women’s BC Study (YWBCS). Methods: The YWBCS is a multi-institutional cohort of 1,297 women with incident BC diagnosed at age &lt;40 years identified through medical record review, with risk factor data collected via questionnaires. TMAs including 1,505 cores of 640 YBCs that were collected prior to therapy and were suitable for molecular analysis were evaluated for 72 markers with NanoString GeoMx DSP. Associations of log-transformed biomarker expression with BMI and time from last live birth to BC diagnosis were carried out using linear mixed modeling approaches, accounting for multiple tumor tissue cores per individual. Results: Mean age at BC was 35.5 years (range 17-40). Tumor subtypes based on clinical immunostaining were the following: luminal B 46%; luminal A 36%; triple negative 12% and HER2+ 6%. Of 72 biomarkers tested, 53 (74%) were evaluable after QC and normalization. GeoMx ER and PR protein levels were higher in luminal A and luminal B YBCs than in HER2+ and TNBC (p=2x10-94 and 2x10-52, respectively), supporting the accuracy of the GeoMx assay. Compared to nulliparas or those with most recent birth &gt; 3 years before BC, those who gave birth ≤ 3 years prior to BC had tumors with higher expression of PR (8x10-6) and ER (p=0.006), and lower expression of BCL-2 (p=0.006). In contrast, women with last birth ≥ 6 years prior to BC had higher expression of CD8 (p=0.006) and PD-1 (p=0.001) than nulliparas and those with more recent births. Compared to women with BMI between 18.5 and 29.9, those with higher or lower BMIs had higher expression of CD34 (p=0.0003), CD4 (p=0.0009) and CD14 (p=0.001). BCL-2 expression was higher among women with BMI &lt; 18.5 compared to those with higher values (p=0.005). Women with BMI of 30+ had higher β2-microglobulin levels than those with lower BMIs (p=0.003). Inverse dose-response associations (with higher BMI related to lower levels) were observed for expression of ER (p=0.01), pan cytokeratin (p=0.004) and four proteins in the p13K/AKT signaling pathway (p ≤ 0.01 for each). Conclusions: Preliminarily, we identify differences in hormonal and immune markers in YBCs in relation to timing of births and BMI, suggesting that these factors may contribute to etiologic heterogeneity. Multivariate modeling is ongoing to incorporate molecular subtype and other risk factors. Citation Format: Robert A. Vierkant, Laura M. Pacheco-Spann, Stacey J. Winham, Jennifer M. Kachergus, Ji Shi, Craig Snow, Fergus J. Couch, Janet E. Olson, Chen Wang, Derek C. Radisky, Amy C. Degnim, E. Aubrey Thompson, Laura C. Collins, Kathryn J. Ruddy, Ann H. Partridge, Mark E. Sherman. Associations of parity and body mass index with NanoString digital spatial protein profiling in early onset breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1000.

Abstract Purpose: Triple-negative invasive lobular carcinoma (TN-ILC) of breast cancer is a rare disease and the clinical outcomes and prognostic factors are not well-defined. Methods: Women with stage I-III TN-ILC or triple-negative invasive ductal carcinoma (TN-IDC) of the breast undergoing mastectomy or breast-conserving surgery between 2010 and 2018 in the National Cancer Database were included. Kaplan-Meier curves and multivariate Cox proportional hazard regression were used to compare overall survival (OS) and evaluate prognostic factors. Multivariate logistic regression was performed to analyze the factors associated with pathological response to neoadjuvant chemotherapy. Results: The median age at diagnosis for women with TN-ILC was 67 years compared to 58 years in TN-IDC (p<0.001). There was no significant difference in the OS between TN-ILC and TN-IDC in multivariate analysis (HR 0.96, p=0.44). Black race and higher TNM stage were associated with worse OS, whereas receipt of chemotherapy or radiation was associated with better OS in TN-ILC. Among women with TN-ILC receiving neoadjuvant chemotherapy, the 5-year OS was 77.3% in women with a complete pathological response (pCR) compared to 39.8% in women without any response. The odds of achieving pCR following neoadjuvant chemotherapy were significantly lower in women with TN-ILC compared to TN-IDC (OR 0.53, p<0.001). Conclusion: Women with TN-ILC are older at diagnosis but have similar OS compared to TN-IDC after adjusting for tumor and demographic characteristics. Administration of chemotherapy was associated with improved OS in TN-ILC, but women with TN-ILC were less likely to achieve complete response to neoadjuvant therapy compared to TN-IDC.

AbstractBackgroundRisk‐stratified follow‐up guidelines that account for the absolute risk and timing of recurrence may improve the quality and efficiency of breast cancer follow‐up. The objective of this study was to assess the relationship of anatomic stage and receptor status with timing of the first recurrence for patients with local‐regional breast cancer and generate risk‐stratified follow‐up recommendations.MethodsThe authors conducted a secondary analysis of 8007 patients with stage I–III breast cancer who enrolled in nine Alliance legacy clinical trials from 1997 to 2013 (ClinicalTrials.gov identifier NCT02171078). Patients who received standard‐of‐care therapy were included. Patients who were missing stage or receptor status were excluded. The primary outcome was days from the earliest treatment start date to the date of first recurrence. The primary explanatory variable was anatomic stage. The analysis was stratified by receptor type. Cox proportional‐hazards regression models produced cumulative probabilities of recurrence. A dynamic programming algorithm approach was used to optimize the timing of follow‐up intervals based on the timing of recurrence events.ResultsThe time to first recurrence varied significantly between receptor types (p < .0001). Within each receptor type, stage influenced the time to recurrence (p < .0001). The risk of recurrence was highest and occurred earliest for estrogen receptor (ER)‐negative/progesterone receptor (PR)‐negative/Her2neu‐negative tumors (stage III; 5‐year probability of recurrence, 45.5%). The risk of recurrence was lower for ER‐positive/PR‐positive/Her2neu‐positive tumors (stage III; 5‐year probability of recurrence, 15.3%), with recurrences distributed over time. Model‐generated follow‐up recommendations by stage and receptor type were created.ConclusionsThis study supports considering both anatomic stage and receptor status in follow‐up recommendations. The implementation of risk‐stratified guidelines based on these data has the potential to improve the quality and efficiency of follow‐up.