William Barlow · University of Washington · PhdFit
Resume-aware faculty matching
Find professors who actually fit you
Upload your resume. PhdFit's six research agents compare your background with faculty profiles, recent publications, lab focus, and outreach opportunities, then rank professors with evidence you can review.
Resume-aware matchingEvidence from recent researchOutreach drafts included
Advisor Match Brief Resume uploaded
Resume signals5 found
NLPHCIFairnessGraph MLPyTorch
Six-agent analysisRunning
Axel · Candidate AnalystResume parsed
Nova · Professor ResearcherRecent papers
Echo · Match ExplainerEvidence linked
Atlas · Strategy AdvisorFit ranked
Lyra · Outreach WriterDraft queued
Juno · Meeting Prep CompilerNext step ready
SC
Dr. Sarah Chen
Stanford · NLP & interpretability
91
fit
Fairness + NLP projects overlap with her recent work.
Strong methods fit: model evaluation and interpretability.
Clear outreach angle for current trustworthy AI research.
Outreach angle
Ask how her lab is extending interpretability methods into fairness audits for real-world AI systems.
William E. Barlow is a Research Professor in the Department of Biostatistics at the University of Washington School of Public Health. He holds a BA in Psychology from Western Washington University, an MA in Psychology from the University of Toronto, and both MS and PhD degrees in Biostatistics from the University of Washington. His research interests include biomarkers, clinical trials, survival analysis, cancer research, epidemiology, medical diagnostics, ophthalmology, and sports statistics. Barlow's work focuses on applying biostatistical methods to various medical and health-related fields, contributing to scientific collaborations and consulting services within the university. His extensive background in biostatistics supports his role in advancing research and education in public health.
BACKGROUND: S0221 investigated weekly vs every 2 weeks dosing of doxorubicin (A) and cyclophosphamide (C) followed by paclitaxel in patients with high-risk early breast cancer. After an interim analysis, random assignment to the 2 AC arms was stopped for futility, and the trial was modified to study only the paclitaxel schedules. METHODS: Between December 2003 and November 2010, a total of 2716 patients were randomly assigned in a 2 × 2 factorial design to 15 weeks of weekly A and daily C vs 6 cycles of every 2 weeks AC; and weekly paclitaxel for 12 weeks vs 6 cycles of every 2 weeks paclitaxel. Between January 2011 and January 2012, an additional 578 patients were assigned to 4 cycles of every 2 weeks AC and randomly assigned to weekly vs every 2 weeks paclitaxel. Updated survival was assessed using log-rank tests and Cox regression models. We compared outcomes by breast cancer subtype as well. RESULTS: At a median follow-up of 12.1 years, there were no statistically significant differences among the 4 treatment arms in disease-free survival (DFS) (P = .91) or overall survival (P = .34) in the original protocol. Among the 578 patients assigned AC for 4 cycles and randomly assigned to paclitaxel weekly vs every 2 weeks paclitaxel, there were no overall differences in DFS (P = .32) or overall survival (P = .42). CONCLUSION: As there were no statistically significant outcome differences in DFS or overall survival between the studied schedules of AC and paclitaxel with extended follow-up in the original or revised protocol, either paclitaxel schedule may be recommended, with selection based on toxicity, cost, or patient preference.
Importance: Breast cancer treatment is associated with cancer-related cognitive impairment (CRCI). However, the association of endocrine therapy (ET) vs chemotherapy plus endocrine therapy (CET) with CRCI is poorly understood. Objective: To compare patient-reported CRCI between women with breast cancer treated with ET vs CET and to consider whether menopausal status may be associated. Design, Settings, and Participants: This was a prespecified secondary analysis of RxPONDER (SWOG S1007), a multinational phase 3 randomized clinical trial of more than 5000 women with hormone receptor-positive ERBB2-negative (formerly HER2-negative) breast cancer with 1 to 3 involved lymph nodes and Oncotype DX (21-gene recurrence score) of 25 or less. Participants were enrolled from February 2011 to September 2017, with results first reported in December 2020. Participants were randomly assigned to CET or ET, with ongoing follow-up. This secondary analysis assessed cognitive function using the Patient-Reported Outcomes Measurement Information System Perceived Cognitive Function Concerns (PCF) questionnaire at baseline, 6, 12, and 36 months. Data were analyzed from July 2022 to August 2025. Intervention: Random assignment to CET or ET. Main Outcomes and Measures: Mean PCF standardized (T) scores by menopausal status over time using generalized estimating equations analysis for continuous outcomes. Results: Of the 568 patients who completed the baseline questionnaire and were included in the analysis, 139 (24%) were premenopausal (median [range] age, 47.8 [28.0-56.3] years) and 429 (76%) were postmenopausal (median [range] age, 62.3 [37.3-87.6] years). Among the 274 (48%) who received CET and the 294 (52%) who received ET alone, CET was determined to have a greater negative association with patient-reported CRCI in both the pre- and postmenopausal participants during the 36-month follow-up. In the ET alone group, PCF scores for premenopausal participants decreased from baseline to 6 and 12 months (53.53, 51.51, and 51.72, respectively) but recovered to baseline (54.36) at 36 months. For postmenopausal participants, mean PCF scores were essentially stable (51.72, 51.13, 51.11, and 51.70, respectively); however, in the CET group, PCF scores for both pre- and postmenopausal participants decreased from baseline to 6 and 12 months (premenopausal, 52.84, 49.27, 48.04; postmenopausal, 50.65, 48.39, 47.13, respectively) and did not return to baseline at 36 months (premenopausal, 49.25; postmenopausal, 48.44). The difference in longitudinal mean PCF scores over time between CET and ET groups was -3.02 (95% CI, -5.33 to -0.72; P = .01) for premenopausal and -2.37 (95% CI, -3.92 to -0.82; P = .003) for postmenopausal participants. Conclusions and Relevance: This secondary analysis of the RxPONDER found that CET had a greater negative association with patient-reported CRCI compared to ET alone in both pre- and postmenopausal participants over a 36-month follow-up period. Interventions to prevent or treat CRCI are needed to improve the long-term quality of life of these patients treated with chemotherapy. Trial Registration: ClinicalTrials.gov Identifier: NCT01272037.
<p>Supplementary Figure 4. Kaplan-Meier plot of overall survival according to change in CTC status from baseline to first follow-up (cutoff for CTC positivity ≥1 CTC). (A) HER2-positive/hormone receptor-positive tumors; (B) HER2-positive/hormone receptor-negative tumors. Survival time is given as time since baseline CTC assessment in months. The P-values refers to the global log rank test with 3 degrees of freedom.</p>
<p>Supplementary Figure 3: Kaplan-Meier plots of overall survival (OS) according to CTC status at baseline for HER2-positive/hormone receptor-positive tumors(A, B) and HER2-positive/hormone receptor-negative tumors(C, D). Survival time is given as time since baseline CTC assessment in months. (A, C) cutoff for CTC positivity ≥1 CTC; (B, D) cutoff for CTC positivity ≥5 CTCs. The P-values refer to pairwise comparisons with the log rank test.</p>
<p>Supplementary Table 4: Comparisons of the probabilities of being CTC positive betweenHER2- positive/hormone receptor-positive and HER2-positive/hormone receptor-negative tumors according to time point of CTC assessment (baseline, first follow-up) and cutoff for CTC positivity (≥1 CTC, ≥5 CTCs). Probabilities of being CTC positive are given in terms of percentages and odds ratios with 95% confidence intervals (CI) obtained using binary logistic regression models with the response variable CTC positive (yes vs no).</p>
Abstract Background: While the landscape of metastatic breast cancer treatment has evolved over the years with incorporation of targeted agents and antibody drug conjugates, chemotherapy continues to be the mainstay of adjuvant treatment for high-risk early breast cancer. S0221 was a previously reported phase III randomized trial performed by the North American Breast Cancer Intergroup (now known as the National Clinical Trials Network [NCTN]) investigating alternative dosing schedules of chemotherapy for early breast cancer. Methods: S0221 investigated weekly (arms 2 and 4) doxorubicin (A)/cyclophosphamide (C) + granulocyte colony stimulating factor (GCSF; filgrastim or pegfilgrastim) versus (vs.) every 2 weeks (Q2W) (arms 1 and 3) schedule AC, followed by paclitaxel (P) given Q2W or weekly for 12 weeks as post-operative adjuvant therapy in node-positive or high-risk node-negative breast cancer. Weekly AC was given as doxorubicin 24mg/m2 IV once per week and cyclophosphamide 60mg/m2 orally once per day with GCSF. Between December 2003 and November 2010, 2716 patients were randomized in a 2x2 factorial design to: 1) 15 weeks of weekly AC vs. 6 cycles of Q2W AC and 2) P weekly vs. P Q2W with GCSF support and this accounted for arms 1-4. After enrollment of 2716 patients, randomization to the two AC arms was stopped for futility and the trial was modified to study only the P schedules. Between January 2011 and January 2012, an additional 578 patients were assigned to 4 cycles of Q2W AC and randomized to P weekly vs. Q2W accounting for arms 5-6. Here, we present updated survival outcomes of four arms on the original protocol and report, for the first time, analysis of 578 patients from two additional arms on the revised protocol. Updated survival was assessed using log-rank tests and Cox regression models. Results: At a median follow-up of 12.1 years, among the patients treated in the original protocol, there were no significant differences among the four treatment arms for disease free survival [DFS] (p=0.91) or overall survival [OS] (p=0.34). When stratified by disease subtype, the human epidermal receptor-2 (HER2) positive cohort had the highest 10-year DFS rate (77.7%) compared to HR-positive/HER2-negative (70.6%) or HR-negative/HER2-negative (70.3%) cohorts (p value = 0.0005). The HER2-positive cohort also had a superior 10-year OS rate of 82.3% compared to HR-positive/HER2-negative (78.1%) or HR-negative/HER2-negative (74.9%) cohorts (p=0.0044). Among the 578 patients assigned AC for 4 cycles and randomized to P weekly vs. Q2W P, there were no overall differences in DFS (p=0.32) or OS (p=0.42). There was no difference in 10-year DFS rate between original (71.3%) or revised (74.4%) protocol (HR 0.80; 95% CI 0.75-1.05). Patients were also stratified in terms of sex (23 men enrolled) and race (379 Blacks enrolled). While women have superior DFS and OS compared to men, due to small number of men and wide CI, these data should be interpreted with caution. Black patients had worse DFS and OS (10-year DFS rate of 64.2% vs. 72.8%; 10-year OS rate of 70.4% vs. 79.0%) compared to non-Blacks. In terms of toxicity, cardiac toxicity profile was more favorable in arms 2 and 4 (0.5-0.7%) that used weekly AC compared to the other four arms that incorporated AC Q2W (1.1-3.3%). There was more skin toxicity with weekly AC schedule with 15.7-16.1% events compared to 2.4-4.0% events in the other four arms. There was no difference in infectious risk or changes in metabolic profile but there were more neurological AEs, and more pain with arms that used Q2W paclitaxel compared to weekly paclitaxel. Conclusion: As there were no significant outcome differences in DFS or OS between the studied schedules with extended follow-up in the original cohort or in the additional 578 patients treated on the revised protocol, either paclitaxel schedule may be recommended, with selection based on toxicity, cost, or patient preference rather than efficacy. Citation Format: Azka Ali, William E Barlow, Halle CF Moore, Timothy J Hobday, Claudine Isaacs, Muhammad Salim, Jonathan K Cho, Kristine J Rinn, Kathy S Albain, Helen K Chew, Gary V Burton, Timothy D Moore, Gordan Srkalovic, Bradley A McGregor, Lawrence E Flaherty, Danika. Long-Term Follow-up and updated analysis from S0221, Comparing Alternative Dose-Schedules of Adjuvant Anthracycline/Taxane Therapy in High-Risk Early Breast Cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr RF1-04.
<p>Supplementary Table 3: Comparisons of the probabilities of being CTC positive among tumor subtypes (hormone receptor-positive/HER2-negative, HER2-positive, triple-negative) according to time point of CTC assessment (baseline, first follow-up) and cutoff for CTC positivity (≥1 CTC, ≥5 CTCs). Probabilities of being CTC positive are given in terms of percentages and odds ratios with 95% confidence intervals (CI) obtained using binary logistic regression models with the response variable CTC positive (yes vs no)</p>
<p>Supplementary Table 6: Comparisons of the probabilities of patients that are CTC positive at baseline being a CTC responder (i.e., switching from CTC positive at baseline to CTC negative at first follow-up) between HER2-positive/hormone receptor-negative and HER2-positive/hormone receptor-positive tumors according to cutoff for CTC positivity (≥1 CTC, ≥5 CTCs). Probabilities of being a CTC responder are given in terms of percentages and odds ratios with 95% confidence intervals (CI) obtained using binary logistic regression models with the response variable CTC responder (yes vs no).</p>
36 Background: Primary prophylactic colony stimulating factor (PP-CSF) is recommended for patients at high risk of febrile neutropenia (FN) receiving chemotherapy. Delivery options for PP-CSF include in-clinic administration (IC), home delivery via an on-body device (OBD), or home self-injection without a device (SI). OBDs are marketed as more convenient and as potentially improving adherence, but are associated with higher costs. The impact of PP-CSF delivery method on FN risk is unknown. SWOG S1415CD (TrACER) was a pragmatic, cluster-randomized trial evaluating a guideline informed standing order system to improve PP-CSF use and reduce FN. We conducted a secondary analysis of TrACER to evaluate associations between PP-CSF delivery method and FN. Methods: Female patients with breast cancer enrolled in TrACER who received PP-CSF with a known delivery method were included. The primary outcome was FN within 6 months by delivery method (IC, OBD, SI). Multivariable logistic regression adjusted for delivery method, PP-CSF agent, FN risk category of the chemotherapy regimen, and performance status (PS). Secondary outcomes included adherence to chemotherapy and PP-CSF. Results: Of 1713 patients, 773 (45%) received IC, 862 (50%) received OBD, and 78 (5%) received SI. Most received high-risk FN chemotherapy regimens (IC 86%, OBD 87%, SI 91%). Patients in the IC and OBD groups received pegfilgrastim almost exclusively (98% and 99%, respectively), compared to 29% in the SI group. IC patients were more likely to receive PP-CSF on the same day as chemotherapy rather than the recommended next-day administration (p < 0.001). IC and OBD groups were demographically similar (white 86% and 82%; Medicaid 12% and 9%) while the SI group was more diverse (white 56%, Medicaid 40%). FN incidence was low: IC 3.8% (95% CI 2.5–5.3%), OBD 5.0% (3.6–6.6%), SI 6.4% (2.1–14.3%) (p = 0.29). In multivariable analysis, FN was associated with high-risk FN chemotherapy regimens (OR 2.97, 95% CI 1.2–9.9, p = 0.038) and PS ≥2 (OR 4.99, 95% CI 1.12–15.9, p = 0.014); neither delivery method nor PP-CSF agent was significantly associated with FN. Chemotherapy adherence (dose intensity and duration) did not differ by delivery method. Conclusions: FN rates were low across all PP-CSF delivery methods. There were no significant differences in FN based on PP-CSF delivery method or agent. Given the similar outcomes, delivery decisions may be guided by patient-centered factors such as cost, convenience, and access. These findings may inform future prescribing practices and payer coverage policies. Clinical trial information: NCT02728596 .
