Marlene Lin

Marlene Lin

Biostatistics & Epidemiology, University of California, San Francisco, San Francisco, California, USA
+1
orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 July 2023 in Metabolites
Verified
Journal Article
Open access
PDF file
Ajin JoyAjin Joy
Ajin Joy
Marlene LinMarlene Lin
Marlene Lin
MJMelissa Joines
Melissa Joines
ASAndres Saucedo
Andres Saucedo
SLStephanie Lee-Felker
Stephanie Lee-Felker
michael albert thomasmichael albert thomas
michael albert thomas
The main objective of this work was to evaluate the application of individual and ensemble machine learning models to classify malignant and benign breast masses using features from two-dimensional (2D) correlated spectroscopy spectra extracted from five-dimensional echo-planar correlated spectroscopic imaging (5D EP-COSI) and diffusion-weighted imaging (DWI). Twenty-four different metabolite and lipid ratios with respect to diagonal fat peaks (1.4 ppm, 5.4 ppm) from 2D spectra, and water and fat peaks (4.7 ppm, 1.4 ppm) from one-dimensional non-water-suppressed (NWS) spectra were used as the features. Additionally, water fraction, fat fraction and water-to-fat ratios from NWS spectra and apparent diffusion coefficients (ADC) from DWI were included. The nine most important features were identified using recursive feature elimination, sequential forward selection and correlation analysis. XGBoost (AUC: 93.0%, Accuracy: 85.7%, F1-score: 88.9%, Precision: 88.2%, Sensitivity: 90.4%, Specificity: 84.6%) and GradientBoost (AUC: 94.3%, Accuracy: 89.3%, F1-score: 90.7%, Precision: 87.9%, Sensitivity: 94.2%, Specificity: 83.4%) were the best-performing models. Conventional biomarkers like choline, myo-Inositol, and glycine were statistically significant predictors. Key features contributing to the classification were ADC, 2D diagonal peaks at 0.9 ppm, 2.1 ppm, 3.5 ppm, and 5.4 ppm, cross peaks between 1.4 and 0.9 ppm, 4.3 and 4.1 ppm, 2.3 and 1.6 ppm, and the triglyceryl–fat cross peak. The results highlight the contribution of the 2D spectral peaks to the model, and they demonstrate the potential of 5D EP-COSI for early breast cancer detection.

26 November 2025 in NEJM AI
Verified
Journal Article
Closed access
ABAndrew Bishara
Andrew Bishara
Marlene LinMarlene Lin
Marlene Lin
JDJean Digitale
Jean Digitale
JYJoanne W. L. Yim
Joanne W. L. Yim
HSHossein Soleimani
Hossein Soleimani
PRPriya Ramaswamy
Priya Ramaswamy
Current preoperative blood ordering relies on outdated, procedure-based maximum surgical blood order schedule (MSBOS) that overlook individual patient risk. Smart Match, a machine-learning (ML) tool that predicts patient-specific transfusion needs, was developed and silently validated. Integrated into real-time workflows, Smart Match demonstrates the potential of ML to replace MSBOS with personalized, data-driven approaches to perioperative blood management. Using 82 variables and 1921 features from the electronic medical record of a tertiary academic medical center, a custom extreme gradient boosting (XGBoost) model was optimized to predict perioperative red blood cell (RBC) transfusions before performing adult elective surgeries, aligning sensitivity with native MSBOS thresholds. The retrospective cohort was split into training, validation, and test sets. Silent prospective validation was conducted in real-world elective cases, comparing Smart Match predictions with both MSBOS and clinician RBC preordering behaviors at multiple time points preceding surgery. Model discrimination, calibration, and clinical utility were assessed. In addition, a hybrid model to estimate daily blood bank needs was developed. Retrospective data consisted of 235,054 cases with a 3.04% (95% confidence interval [CI], 2.97 to 3.11) transfusion rate. The model predictors included medical history, laboratory results, demographics, surgery information, medications, transfusion history, and MSBOS recommendations. The test data achieved an area under the receiver-operating-characteristic curve (AUROC) of 0.94 (95% CI, 0.93 to 0.95) and an area under the precision–recall curve (AUPRC) of 0.57 (95% CI, 0.53 to 0.61). Silent prospective validation (n=24,003, transfusion rate 2.18%, 95% CI, 2.07 to 2.46) maintained a 0.94 (95% CI, 0.92 to 0.95) AUROC and a 0.55 (95% CI, 0.51 to 0.60) AUPRC. Prospectively, the model’s sensitivity was 0.72 (95% CI, 0.67 to 0.75) and its positive predictive value was 0.34 (95% CI, 0.31 to 0.37), surpassing both MSBOS and clinician behavior. Our hybrid model outperformed MSBOS for daily RBC needs with a mean absolute error of 12.86 (95% CI, 11.86 to 13.84) versus 13.34 (95% CI, 12.14 to 14.51) on the test set. Smart Match reliably outperforms MSBOS and clinician orders, works in real time, and will next be assessed through a randomized control trial to gauge its usability and clinical impact. (Funded by the National Institute of General Medical Sciences and others.)

6 October 2026 in medRxiv
Verified
Preprint
Open access
PDF file
RJRenaud La Joie
Renaud La Joie
Marlene LinMarlene Lin
Marlene Lin
YWYingbing Wang
Yingbing Wang
DSDavid N. Soleimani-Meigooni
David N. Soleimani-Meigooni
TETara Ellingson
Tara Ellingson
GRGil D. Rabinovici
Gil D. Rabinovici
Amyloid-targeting therapies (ATTs) substantially reduce amyloid PET signal and slow clinical decline in early symptomatic Alzheimer disease. As ATTs enter clinical practice, PET may assess target engagement and guide treatment management, but there is no consensus on how to measure treatment-related amyloid clearance (TRAC). We compared Centiloid quantitation across pipelines and visual reads for interpreting pre– and post-ATT amyloid PET. Methods : We analyzed 18F-florbetapir or 18F-florbetaben PET from 101 patients treated with lecanemab or donanemab: 88 pre-ATT and 54 post-ATT scans, including 41 paired scans. Centiloids were derived using FDA-cleared MIMneuro and two research pipelines (rPOP, CapAIBL) and compared using intraclass correlation coefficients (ICC). Clearance rates were estimated using linear mixed-effects models. Full TRAC was defined on post-ATT scans quantitatively (<11 Centiloids) or based on visual interpretation by 3 expert readers. Results : Treated patients were 74±7 years old; 57% were female, 59% were apolipoprotein E ε4 carriers, and 78% had mild cognitive impairment (22% mild dementia). Centiloid agreement between pipelines was equally high pre– and post-ATT (both ICCs=0.90). Among paired scans (mean treatment duration, 16.7±4.6 months), clearance estimates were similar across pipelines and greater with donanemab (n=15 patients; between –62 and –60 Centiloids/year depending on the pipeline) than lecanemab (n=26; –38 to –36 Centiloids/year; all P<0.005). Among 54 post-ATT scans, between 44% and 57% were <11 Centiloids depending on pipeline, with 20 (37%) meeting this criterion across all 3 pipelines. In contrast, only 11%–19% were visually negative across readers, with 4 (7%) unanimously read as negative. Of the 20 post-ATT scans <11 Centiloids across all pipelines, 13 (65%) were considered positive by all readers, commonly with residual occipital binding. Conclusion : Centiloid quantitation was robust across pipelines before and after ATT initiation, yielding clearance rates consistent with clinical trials. However, quantitative and visual assessments were markedly discordant after ATT, with an unusually high proportion of scans <11 Centiloids remaining visually positive, warranting cautious interpretation of post-ATT Centiloid values. Larger studies should assess the clinical relevance of these findings and further develop robust approaches to establish full TRAC.

8 May 2023 in Preprints.org
Verified
Preprint
Open access
PDF file
Ajin JoyAjin Joy
Ajin Joy
Marlene LinMarlene Lin
Marlene Lin
MJMelissa Joines
Melissa Joines
ASAndres Saucedo
Andres Saucedo
SLStephanie Lee-Felker
Stephanie Lee-Felker
michael albert thomasmichael albert thomas
michael albert thomas
The main objective of this work was to evaluate the application of individual and ensemble machine learning models to classify malignant and benign breast masses using features from two-dimensional (2D) correlated spectroscopy spectra extracted from five-dimensional Echo Planar-Correlated Spectroscopic Imaging (5D EP-COSI) and Diffusion-weighted imaging (DWI). Twenty-four different metabolite and lipid ratios with respect to 2D diagonal peaks at 1.4ppm and 5.4ppm, and water from one-dimensional non-water-suppressed (NWS) spectra were used as the features. Additionally, water fraction, fat fraction and water-to-fat ratios from NWS spectra and apparent diffusion coefficients (ADC) from DWI were included. Nine most important features were identified using recursive feature elimination. XGBoost (AUC:93.0%, Accuracy:85.7%, F1-score:87.6%), GradientBoost (AUC:94.4%, Accuracy:87.0%, F1-score:89.4%), CatBoost (AUC:95.2%, Accuracy:86.9%, F1-score:88.4%) and RandomForest (AUC:92.2%, Accuracy:85.3%, F1-score:87.6%) were the best performing models. While the conventional biomarkers like choline, myo-Inositol, and glycine were statistically significant predictors, the key features contributing to the classification were ADC, 2D diagonal peaks at 0.9ppm, 2.1ppm and 2.3ppm, cross peaks between 1.4 and 0.9ppm, 4.3 and 4.1ppm, 2.1 and 1.4ppm and the triglyceryl-fat cross peak. The results highlight the contribution of the 2D spectral peaks to the model, and they demonstrate the potential of 5D EP-COSI for early breast cancer detection.