DATA / BIOLOGY
RETROSPECTIVE STUDY
Association of Age, Axillary Lymph Node Involvement, and Survival in Breast Cancer
An analysis of age, positive lymph node count and survival using SEER-derived patient records.

I am collecting my research work here with the questions, evidence and methods that shaped each project.
Four research records / 2026.
Mean count in each ten-year age group
RETROSPECTIVE STUDY
An analysis of age, positive lymph node count and survival using SEER-derived patient records.

EXPERIMENTAL DESIGN
Contamination-aware planning for suspected-particle assessment with filtration and optical microscopy.

LONGITUDINAL RESEARCH DESIGN
Studying how changes in speech over time could help estimate clinical progression in Huntington’s disease.

PAPER ANALYSIS / PROPOSED STUDY
A study of early MRI-detected brain atrophy and a proposed one-year comparison of preHD gene carriers with healthy controls.

These records document completed analyses alongside work still in development. Each entry includes its research question, methods, evidence and sources.
BREAST CANCER RESEARCH
SEER-DERIVED DATA
A retrospective analysis of 4,024 patient records, ages 30 to 69.
I. THE QUESTION
How do age at diagnosis, the number of positive axillary lymph nodes and recorded survival months relate to one another among patients with breast cancer?
II. THE APPROACH
The analysis used a SEER-derived public dataset with 4,024 records for women aged 30 to 69. Spearman rank correlation was used to evaluate age and node count, and node count and survival months. A separate categorical test assessed node-severity distribution across four age groups.
III. ANATOMICAL REFERENCE

Axillary lymph nodes are important in breast cancer staging and prognosis. The anatomy helps explain why nodal involvement was included as a variable.
IV. RESULTS
AGE AND NODE COUNT
−0.020Spearman ρNODE COUNT AND SURVIVAL
−0.110Spearman ρThere was no statistically significant rank correlation between age and positive node count in the sample. Positive node count showed a weak negative association with survival months. The chi-square test of node-severity categories across age groups was significant, although the effect size was negligible (Cramér's V = 0.041).
Figure 1
Age at diagnosis and positive lymph nodesFigure 2
Lymph node severity by age group| Age group | n | Mean nodes | Median nodes | SD |
|---|---|---|---|---|
| 30–39 | 230 | 4.57 | 3 | 5.24 |
| 40–49 | 1,124 | 3.78 | 2 | 4.53 |
| 50–59 | 1,390 | 4.42 | 2 | 5.44 |
| 60–69 | 1,280 | 4.13 | 2 | 5.18 |
| All patients | 4,024 | 4.16 | 2 | 5.11 |
Figure 3
Positive lymph nodes and survival monthsV. RESTRICTIONS
The studied dataset did not contain tumor size, receptor status, histologic grade or patients over age 69. The results are associations, not evidence of causation. A nonsignificant correlation cannot establish that the true effect is exactly zero.
VI. MY CONTRIBUTION
I developed the research question, selected the SEER-derived dataset, organized the variables and conducted the statistical comparisons. I used Excel to prepare tables and figures and worked through the interpretation with feedback from PSI mentors.
I presented the breast cancer analysis at the 2026 OHSU Partnership for Scientific Inquiry Research Symposium.
VII. SOURCES
Association of Age, Axillary Lymph Node Involvement, and Survival in Breast Cancer: A SEER-Based Analysis (2026), poster and research report.
Dataset: SEER-derived Breast Cancer Dataset. Anatomy: National Cancer Institute, SEER.
POLYTRACK
LABORATORY METHOD
A proposal for contamination-aware visible-particle screening with filtration and optical microscopy.
I. PURPOSE
PolyTrack explores whether a low-cost optical microscopy and image-analysis workflow can count and size suspected particles in water exposed to polypropylene (PP #5) food containers. The method needs procedural blanks, calibrated imaging and repeatability checks before measurements can be interpreted.
II. INITIAL PROTOCOL
The current proposed feasibility stage consists of a water-only procedural blank and one PP room-temperature control. Both are described with a proposed 200 mL water contact volume and 30-minute interval. Execution remains subject to supervisor approval.
Water-only run to check what the workflow contributes without a PP sample.
One new PP food-contact container under the documented control conditions.
Check filter condition, handling, background contamination and imaging readiness.
III. REFERENCE IMAGES
These scientific references show the types of filtration and microscopic examination informing PolyTrack's measurement design.
Reference images illustrate the methods. Polymer identity and particle counts for PolyTrack remain subjects of experimental validation.
IV. ANALYTICAL LIMITS
The proposed measurement pipeline requires calibrated magnification, an understanding of size-dependent detection, contamination controls and a repeatability check. Optical counting alone cannot establish polymer identity.
V. TWO-PART PROJECT
PolyTrack also has a separate student awareness component. The planned two-group survey compares participants receiving a short “CHECK → CONFIRM → CHOOSE” educational activity with a comparison group. The design includes an initial survey, an immediate follow-up and a further check approximately two weeks later.
This survey intervention is separate from the laboratory experiments. No survey outcomes are reported in this portfolio.
VI. STAGED PILOT
Stage 1 is limited to two separately processed runs: a water-only procedural blank (B1) and one room-temperature PP container control (C1). The working plan specifies 200 mL per run, 30 minutes per run, and separate labelled filters, with review by the teacher or scientist before later stages. Heated conditions belong to the deferred protocol and are not results of this initial test.
VII. MY CONTRIBUTION AND STATUS
I reviewed published PP-container studies, compared potential food-contact samples, worked through blank and filtration requirements, and drafted the experimental and awareness-survey designs. The project remains in planning and feasibility review; validated particle counts, polymer identities and treatment comparisons are not reported yet.
VIII. REFERENCES
Hu et al. (2022). Microplastics in Widely Used Polypropylene-Made Food Containers. Toxics, 10(12), 762.
Photographs: USGS filtration apparatus, Jamie Velkoverh and USGS microplastic fragments.
HUNTINGTON’S DISEASE
LONGITUDINAL RESEARCH
Exploring whether changes in speech can offer a practical way to monitor motor, cognitive and functional decline.
RESEARCH QUESTION
Can a short speech assessment reveal how Huntington’s disease changes within the same person over time?
2026 / Research proposal and study design
I. BACKGROUND
Huntington’s disease is an inherited neurological disorder that affects movement, thinking and everyday functioning. Speech production depends on coordinated motor and cognitive processes. Changes in its timing, rhythm and articulation may therefore provide useful information about disease severity.
My research question focuses on the progression of these changes. A model trained to estimate clinical scores at one visit is not necessarily able to track how an individual’s condition changes between visits.
II. NEUROLOGICAL CONTEXT
The striatum, particularly the caudate nucleus and putamen, is affected in Huntington’s disease. Alongside clinical measures, the proposal considers whether longitudinal speech changes might be related to changes in MRI-derived striatal volume.
III. RESEARCH DESIGN
The proposal outlines a longitudinal observational study of approximately 100 Huntington’s disease gene carriers, including premanifest and manifest stages, with follow-up over two to three years and assessments approximately every six to twelve months. These are proposed targets, not completed recruitment.
Use short standardized forward and backward counting recordings at repeated visits.
Examine speaking rate, pause duration, rhythm and articulation-related errors.
Compare speech-derived estimates with observed UHDRS-related clinical scores, including cUHDRS, TMS and TFC.
Assess whether an existing prediction pipeline reflects within-person change over time.
IV. RESEARCH FOUNDATION
Riad and colleagues (2022) analyzed 126 recordings from 103 Huntington’s disease gene carriers, using forward and backward counting tasks to estimate motor, cognitive and functional scores. This published approach informs the proposed longitudinal extension; their sample and prediction results are not outcomes of my project.
The aim is to apply an established model across repeated observations, not to claim a new machine learning model was developed.
V. QUESTIONS TO TEST
Do changes in timing and articulation follow clinical decline between visits?
Do speech-derived estimates track observed motor and functional scores?
Where MRI data are available, are speech changes associated with striatal volume changes?
VI. RESEARCH STAGE
I developed a research proposal to apply an existing speech-prediction approach across multiple visits and compare estimated clinical change with observed assessments. A second proposed aim examines links between speech change and MRI-derived striatal volume.
No new patient recordings, model-training results or independently validated longitudinal predictions are presented here. Research with participants would require ethics approval and appropriate clinical-data access.
VII. SOURCES
Riad et al. (2022). Predicting clinical scores in Huntington’s disease: a lightweight speech test. Journal of Neurology.
National Institute of Neurological Disorders and Stroke. Huntington’s Disease.
Image sources: NIST CAG repeat illustration; Gaillard MRI; Riad et al. Figure 1.
PREMANIFEST HD
SCIENTIFIC POSTER
Examining structural brain changes before clinical onset through an original-paper analysis and proposed follow-up study.
WHY THIS MATTERS
Mean percentage brain volume change over one year, measured with SIENA. Between-group comparison: p = 0.031.
I. ORIGINAL STUDY
Majid and colleagues (2011) investigated whether automated structural MRI analysis could detect brain atrophy in premanifest Huntington’s disease. Their one-year study compared 35 preHD gene carriers and 22 age-matched healthy controls.
Participants underwent structural T1-weighted MRI at two time points, approximately one year apart. The images were analyzed with SIENA to estimate percentage brain volume change and voxel-level patterns of atrophy.
II. KEY FINDING
The original study found more whole-brain volume loss in the preHD group than in controls. Atrophy was particularly evident around the ventricles and basal ganglia. The subgroup closer to predicted clinical onset had the largest mean loss.
The reported preHD versus control difference was statistically significant (p = 0.031). These findings belong to the original published study and are presented here as part of my paper analysis.
| Group | n | Mean PBVC | SD |
|---|---|---|---|
| Healthy controls | 22 | −0.149% | 0.348 |
| Premanifest HD | 35 | −0.388% | 0.425 |
| Far-from-onset preHD | 17 | −0.259% | 0.390 |
| Close-to-onset preHD | 18 | −0.510% | 0.431 |
Published comparisons: preHD versus controls, p = 0.031, Cohen's d = 0.617; close-to-onset preHD versus controls, p = 0.019, Cohen's d = 0.920. Source: Majid et al. (2011), Table 3 and Results.
The MRI edge-displacement panels used in the original poster come from Majid et al. (2011), Figure 2. The published figure is linked directly to preserve its source and publisher attribution.
III. IDENTIFIED GAP
Although the original study detected atrophy over one year, this observation period could not show whether brain volume loss continues at a similar rate, accelerates or plateaus over the longer term.
That limitation led me to consider how follow-up imaging could be used to track disease progression before clinical symptoms appear.
IV. PROPOSED QUESTION AND HYPOTHESIS
Do people with premanifest Huntington’s disease lose more brain volume over one year than healthy people?
Individuals with premanifest Huntington’s disease will show greater brain volume loss over one year than healthy control participants.
V. PROPOSED STUDY DESIGN
The proposed study compares preHD gene carriers (CAG ≥ 38) with age-matched controls without the Huntington’s disease gene mutation. Both groups would undergo structural T1-weighted MRI at baseline and again one year later. SIENA would be used to compare percentage brain volume change.
Primary outcome: between-group difference in whole-brain volume change over one year. The participant numbers above refer to Majid et al.’s study, not a recruited cohort for this proposal.
VI. DESIGN CONSIDERATION
The proposed one-year comparison tests the type of group difference described in the original paper. A longer follow-up with more than two scans would be needed to answer the identified question about whether atrophy accelerates, slows or remains steady over time.
My contribution to this poster was reviewing the published study, identifying that limitation and proposing the comparison of preHD carriers with matched controls. The percentages shown are findings reported by Majid et al. (2011), not outcomes of a new MRI experiment.
VII. REFERENCES & IMAGE CREDITS
Majid, D. S. A., Stoffers, D., Sheldon, S., Hamza, S., Thompson, W. K., Goldstein, J., Corey-Bloom, J., & Aron, A. R. (2011). Automated structural imaging analysis detects premanifest Huntington’s disease neurodegeneration within 1 year. Movement Disorders, 26(8), 1481–1488.
Publication record · Full study and original MRI figures.
Kinnunen, K. M., et al. (2021). Volumetric MRI-Based Biomarkers in Huntington's Disease: An Evidentiary Review. Frontiers in Neurology, 12, 712555. DOI. Figure 2, CC BY 4.0.
Additional MRI: Frank Gaillard, Huntington.jpg, Radiopaedia / Wikimedia Commons, CC BY-SA 3.0.
ABOUT & RESEARCH EXPERIENCE
I'm a high school student interested in medical research, data analysis and finding practical ways to communicate science. This archive brings together completed analyses and studies I am still developing.
Through PSI, I learned from OHSU physicians, professors, and researchers in lectures and workshops covering biomedical research, genetics, cancer, neuroscience, and medical imaging. The program also introduced me to reading scientific papers, research methods, data analysis, and communicating scientific work. I developed a research proposal and presented my breast cancer analysis at the 2026 PSI research symposium.
I founded the Franklin High School HOSA chapter and serve as its student president, working on organizing meetings and introducing students to health-science opportunities.
I volunteered with OSBHA in health education outreach, helping research, edit and share social media materials about youth health and well-being. I also translated health information from English into Vietnamese to make these resources easier for Vietnamese-speaking students and community members to understand.
I attended an OSBHA Student Health Advocates conference and learned more about the health information and support available to young people.
While volunteering at a hospital in Vietnam, I helped distribute food to patients and people with disabilities. Speaking with patients encouraged me to listen more carefully and understand the person behind each health condition.
APPEARANCE
Choose a layout and color treatment. Your selection applies across the archive and is saved in this browser.
Which balance of historical archive and contemporary research publication best reflects your taste? Font combinations and colors remain editable. Scientific results and citations stay consistent across the three layouts.