KAKatrina Armstrong

Katrina Armstrong

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Katrina Armstrong has a total of 22 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Impact of PIK3CA tumor mutation on the association of aspirin or NSAID use and time to breast cancer recurrence.

    by Anne Marie McCarthy, Wei He, Susan Regan, Andrew T. Chan, Beverly Moy, A. John Iafrate, Aditya Bardia, Katrina Armstrong

    1521 Background: Aspirin or NSAID (A/N) use post diagnosis is associated with lower risk of breast cancer recurrence and mortality in cohort studies. A potential mechanism is that A/Ns may suppress cell growth and induce apoptosis in tumors driven by phosphatidylinositol 3 kinase ( PIK3CA), the most common oncogene mutation in breast cancer. An interaction of A/Ns and PIK3CA mutation has been observed for colorectal cancer prognosis, but has not been studied in breast cancer. The objective was to assess time to breast cancer recurrence (TTR) with respect to A/N use and PIK3CAmutation. Methods: Patients with HR+/HER2- breast cancer treated at Massachusetts General Hospital in 2009-2014 who received tumor genotyping were included. PIK3CA mutations, including 8 common hotspot mutations, were assessed by a high-throughput tumor genotyping assay using DNA from formalin-fixed, paraffin-embedded tumor tissue. A/N use beginning 6 months post diagnosis through metastasis was extracted from electronic medical records using coded data and natural language processing. Patients with de novo metastatic disease or progressive disease within 6 months of primary diagnosis were excluded. TTR was estimated using Cox proportional hazards models. Results: Among breast cancer patients (N=212), 60 (28%) used A/Ns and 69 (33%) had PIK3CA mutation (see Table). After adjusting for age, stage, adjuvant endocrine therapy, radiation, and chemotherapy, A/N users had significantly longer TTR (HR=0.65 p=0.01). The association was similar for wild type (HR=0.58 p=0.01) and PIK3CA mutated tumors (HR=0.60 p=0.06), with no significant interaction of A/N use and PIK3CA (p=0.34). Conclusions: Among HR+ breast cancer patients, those who used A/Ns following primary diagnosis had longer TTR than non-users, regardless of tumor PIK3CA mutation status. The study provides a model for how tumor genomics could be integrated into secondary chemoprevention studies. [Table: see text]


  • Open Access

    Using lessons from breast, cervical, and colorectal cancer screening to inform the development of lung cancer screening programs

    by Katrina Armstrong, Jane J. Kim, Ethan A. Halm, Rachel M. Ballard, Mitchell D. Schnall

    Multiple advisory groups now recommend that high‐risk smokers be screened for lung cancer by low‐dose computed tomography. Given that the development of lung cancer screening programs will face many of the same issues that have challenged other cancer screening programs, the National Cancer Institute‐funded Population‐based Research Optimizing Screening through Personalized Regimens (PROSPR) consortium was used to identify lessons learned from the implementation of breast, cervical, and colorectal cancer screening that should inform the introduction of lung cancer screening. These lessons include the importance of developing systems for identifying and recruiting eligible individuals in primary care, ensuring that screening centers are qualified and performance is monitored, creating clear communication standards for reporting screening results to referring physicians and patients, ensuring follow‐up is available for individuals with abnormal test results, avoiding overscreening, remembering primary prevention, and leveraging advances in cancer genetics and immunology. Overall, this experience emphasizes that effective cancer screening is a multistep activity that requires robust strategies to initiate, report, follow up, and track each step as well as a dynamic and ongoing oversight process to revise current screening practices as new evidence regarding screening is created, new screening technologies are developed, new biological markers are identified, and new approaches to health care delivery are disseminated. Cancer 2016;122:1338–1342. © 2016 American Cancer Society.


  • Cost Differences After Initial CT Colonography Versus Optical Colonoscopy in the Elderly

    by Hanna M. Zafar, Jianing Yang, Katrina Armstrong, Peter Groeneveld


  • Open Access

    The use of the Gail model, body mass index and SNPs to predict breast cancer among women with abnormal (BI-RADS 4) mammograms

    by Anne Marie McCarthy, Brad Keller, Despina Kontos, Leigh Boghossian, Erin McGuire, Mirar Bristol, Jinbo Chen, Susan Domchek, Katrina Armstrong

    Introduction: Mammography screening results in a significant number of false-positives. The use of pretest breast cancer risk factors to guide follow-up of abnormal mammograms could improve the positive predictive value of screening. We evaluated the use of the Gail model, body mass index (BMI), and genetic markers to predict cancer diagnosis among women with abnormal mammograms. We also examined the extent to which pretest risk factors could reclassify women without cancer below the biopsy threshold. Methods: We recruited a prospective cohort of women referred for biopsy with abnormal (BI-RADS 4) mammograms according to the American College of Radiology’s Breast Imaging-Reporting and Data System (BI-RADS). Breast cancer risk factors were assessed prior to biopsy. A validated panel of 12 single-nucleotide polymorphisms (SNPs) associated with breast cancer were measured. Logistic regression was used to assess the association of Gail risk factors, BMI and SNPs with cancer diagnosis (invasive or ductal carcinoma in situ). Model discrimination was assessed using the area under the receiver operating characteristic curve, and calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test. The distribution of predicted probabilities of a cancer diagnosis were compared for women with or without breast cancer. Results: In the multivariate model, age (odds ratio (OR) = 1.05; 95% confidence interval (CI), 1.03 to 1.08; P < 0.001), SNP panel relative risk (OR = 2.30; 95% CI, 1.06 to 4.99, P = 0.035) and BMI (≥30 kg/m2 versus 2; OR = 2.20; 95% CI, 1.05 to 4.58; P = 0.036) were significantly associated with breast cancer diagnosis. Older women were more likely than younger women to be diagnosed with breast cancer. The SNP panel relative risk remained strongly associated with breast cancer diagnosis after multivariable adjustment. Higher BMI was also strongly associated with increased odds of a breast cancer diagnosis. Obese women (OR = 2.20; 95% CI, 1.05 to 4.58; P = 0.036) had more than twice the odds of cancer diagnosis compared to women with a BMI 2. The SNP panel appeared to have predictive ability among both white and black women. Conclusions: Breast cancer risk factors, including BMI and genetic markers, are predictive of cancer diagnosis among women with BI-RADS 4 mammograms. Using pretest risk factors to guide follow-up of abnormal mammograms could reduce the burden of false-positive mammograms.


  • The Role of Testing forBRCA1andBRCA2Mutations in Cancer Prevention

    by Anne Marie McCarthy, Katrina Armstrong

    In May 2013, Angelina Jolie, the actress and director, announced in an op-ed in the New York Times that she carried a BRCA1 mutation and had had a preventive do