Good Space a ranking of the record

TheGood surrounding AI

574 documented gains are on the board. 82% rest on at least one peer-reviewed study, and 8% clear the high-evidence bar. Health carries 49% of it. The latest was filed Oct 16, 2026.

Every gain on the board, one square each, shaded by the strength of its evidence. High 45 Moderate 529 Emerging 0

The board · page 12 of 20

331

Evidence-backed gainPeer-reviewedHealth

Clinical-parameter XGBoost model stratified patients into low-risk and high-risk groups with 94.0% vs 65.0% 2-year local control after carbon-ion radiotherapy for early-stage peripheral NSCLC.

Source article: Local Recurrence Prediction After Carbon-ion Radiotherapy for Early-stage Non-small Cell Lung Cancer Using Machine Learning

Anticancer Research

Reader signal

How should this claim be treated?

332

Evidence-backed gainPeer-reviewedHealth

START-AI improved prediction of inpatient admission from the emergency department from AUROC 0.78 to 0.90 when ensemble and transformer features were added.

Source article: The Sydney Triage to Admission Risk Tool With Artificial Intelligence (START-AI) to Support Decision Making in Emergency Departments: Model Explainability and Feature Importance Analysis

Emergency Medicine Australasia

Reader signal

How should this claim be treated?

333

Evidence-backed gainPeer-reviewedHealth

A point-of-care exhaled breath condensate device combined with physiological parameters and ensemble machine learning identified early-stage lung cancer in a real-world screening cohort with 85.7% accuracy and 100% specificity on held-out test data.

Source article: Harnessing Exhaled Breath for Lung Cancer Early Detection-Results From the ExPeL Study

Clinical and Translational Science

Reader signal

How should this claim be treated?

334

Evidence-backed gainPeer-reviewedHealth

Integrating large-scale transcriptomic profiling with machine learning identified a three-gene blood signature that enables early sepsis diagnosis and risk stratification across severity levels.

Source article: Multiomics Profiling Identifies Blood-Based Diagnostic Markers for Sepsis

Journal of Cellular Physiology

Reader signal

How should this claim be treated?

335

Evidence-backed gainPeer-reviewedEducation

AI-generated subtitles for ECFS e-learning modules enabled quick and affordable creation of multilingual education packages in Ukrainian, Romanian, and Turkish that achieved high accuracy after expert editing.

Source article: Bridging the Gap: Translated Medical Education to Support Cystic Fibrosis Centers From Non-English Speaking Countries

Pediatric Pulmonology

Reader signal

How should this claim be treated?

336

Evidence-backed gainPeer-reviewedHealth

Machine learning models using structural MRI, amyloid PET, and demographic features can classify tau positivity in the Braak III/IV region in amyloid-positive cohorts, achieving AUC 0.92 and 85% accuracy on external validation.

Source article: Classification of tau status with machine learning models in amyloid-positive cohorts

Alzheimer's & Dementia

Reader signal

How should this claim be treated?

337

Evidence-backed gainPeer-reviewedMedia & Arts

Adoption of text-to-image generative AI increased human creative productivity and the likelihood of peer favorites per view among artists in a large online art community.

Source article: Generative artificial intelligence, human creativity, and art

PNAS Nexus

Reader signal

How should this claim be treated?

338

Evidence-backed gainPeer-reviewedHealth

Survival models including machine learning approaches predicted imminent 30-day opioid overdose following a first opioid-related diagnosis with C-indices up to 0.745, identifying prior overdose and opioid misuse as strongest predictors for proactive clinical intervention.

Source article: Imminent opioid overdose risk prediction using classical and machine learning survival models following first recorded opioid-related diagnosis: a prospective cohort study from the <i>All of Us</i> research program

The American Journal of Drug and Alcohol Abuse

Reader signal

How should this claim be treated?

339

Evidence-backed gainPeer-reviewedHealth

ConvNeXt-based AI trained on 3676 smartphone images from 1016 patients achieved AUCs of 0.825-0.860 for EASI component scores >=2 and >0.900 for score 3, supporting use as an educational tool to standardize EASI scoring.

Source article: Educational Artificial Intelligence Software to Support Assessment of Atopic Dermatitis Severity

The Journal of Dermatology

Reader signal

How should this claim be treated?

340

Evidence-backed gainPeer-reviewedHealth

AI with multi-omics and systems-level frameworks was used to support target identification, microbial network reconstruction, biomarker discovery, and therapeutic prioritization for anti-virulence strategies against Salmonella Typhi.

Source article: Quorum-sensing, microbiome interactions, and emerging artificial intelligence-assisted anti-virulence strategies in Salmonella Typhi: a critical review of translational opportunities and challenges

Archives of Microbiology

Reader signal

How should this claim be treated?

341

Evidence-backed gainPeer-reviewedHealth

Survey-weighted Explainable Boosting Machine achieved strong temporal stability for severe tooth loss prediction on US BRFSS cohorts, supporting transparent population-level risk stratification.

Source article: Interpretable Machine Learning for Population-Level Tooth Loss Prediction

Journal of Dental Research

Reader signal

How should this claim be treated?

342

Evidence-backed gainPeer-reviewedHealth

A two-stage CNN pipeline using Faster R-CNN with EfficientNet-B7 detected dental implants on panoramic radiographs and classified brand and prosthetic platform size with high accuracy, supporting faster and more standardized clinical workflows when records are missing.

Source article: Automated classification of dental implant brands and prosthetic platform sizes on panoramic radiographs using deep learning

The Journal of Prosthetic Dentistry

Reader signal

How should this claim be treated?

343

Evidence-backed gainPeer-reviewedScience

LLM-driven adaptive level modification framework classified players by skill with 97.82% accuracy and generated modified levels that remained traversable at 74.1% full-level and 83.5% chunk-level rates.

Source article: Adaptive level modification via player skill classification and large language models

Scientific Reports

Reader signal

How should this claim be treated?

344

Evidence-backed gainPeer-reviewedHealth

A machine learning-derived methionine metabolism-related risk score stratified colorectal cancer patients into high- and low-risk groups with significantly different overall survival in training and external validation cohorts.

Source article: Identification of Methionine Metabolism-Driven Heterogeneous Subtypes in Colorectal Cancer and Their Associated Immune Microenvironment

Asia-Pacific Journal of Clinical Oncology

Reader signal

How should this claim be treated?

345

Evidence-backed gainPeer-reviewedHealth

Deep learning models using orbital MRI predicted continuous clinical activity score with low error and inferred patient characteristics including age, sex, and smoking status.

Source article: Deep learning models for predicting clinical activity and characteristics in thyroid eye disease using magnetic resonance imaging

Orbit

Reader signal

How should this claim be treated?

346

Evidence-backed gainPeer-reviewedHealth

Machine learning models trained on TriNetX data stratified patients receiving immune checkpoint inhibitors into risk tiers for cardiac immune-related adverse events within 90 days, achieving moderate discrimination.

Source article: Machine learning approaches to predict early cardiac immune-related adverse events in patients receiving immune checkpoint inhibitors

Supportive Care in Cancer

Reader signal

How should this claim be treated?

347

Evidence-backed gainPeer-reviewedHealth

A LightGBM model trained on metabolomic data predicted incident major depressive disorder among obese participants with AUCs around 0.82-0.84 over 3 to 9 years and outperformed existing clinical models.

Source article: Metabolites stratify future major depressive disorder risk in obese population: a longitudinal machine learning analysis

Psychological Medicine

Reader signal

How should this claim be treated?

348

Evidence-backed gainPeer-reviewedHealth

Advances in AI have enabled opportunistic identification of osteoporosis and other pathologies on imaging performed for other indications.

Source article: Clinical justification for osteoporosis investigation: transitioning from opportunistic to diagnostic referrals from alternative forms of imaging

Journal of Radiological Protection

Reader signal

How should this claim be treated?

349

Evidence-backed gainPeer-reviewedHealth

In a retrospective audit of 5,000 women, the LBC-NET pipeline detected HSIL+ with higher sensitivity than senior cytopathologists and reduced review time while increasing daily capacity.

Source article: An explainable machine learning approach for cervical cancer screening: decoding morphological diagnostic drivers

Acta Clinica Belgica

Reader signal

How should this claim be treated?

350

Evidence-backed gainPeer-reviewedScience

Release of SynTabFall synthetic dataset with 745,380 samples allows training fall-risk prediction models without original patient data while achieving performance on par with models trained on real data.

Source article: Anonymized but Useful Synthetic Tabular Health Data for AI based Fall Risk Assessment

Scientific Data

Reader signal

How should this claim be treated?

351

Evidence-backed gainPeer-reviewedHealth

An XGBoost model trained on routinely collected prehospital vitals and demographics reduced under-triage and over-triage compared to a surrogate of the 2021 National Field Triage Guidelines in 1.2M trauma patients.

Source article: Beyond Thresholds: Can Machine Learning Improve Trauma Field Triage?

The American Surgeon™

Reader signal

How should this claim be treated?

352

Evidence-backed gainPeer-reviewedHealth

Clinicians assisted by Retina4IRD achieved higher top-5 genetic diagnosis accuracy and better downstream management scores for suspected inherited retinal disease compared to specialist-only care.

Source article: AI-based clinician decision support system for diagnosis of inherited retinal diseases: a multicenter, randomized trial

Nature Medicine

Reader signal

How should this claim be treated?

353

Evidence-backed gainPeer-reviewedHealth

A machine learning model using bedside fNIRS functional connectivity during audio movie clips predicted 6-month functional outcome in ICU patients with acute brain injury with 81.3% balanced accuracy, outperforming clinical models.

Source article: Early Identification of Recovery Potential After Acute Brain Injury Using Functional Near-Infrared Spectroscopy

Neurocritical Care

Reader signal

How should this claim be treated?

354

Evidence-backed gainPeer-reviewedHealth

Monthly allocation of Medicaid care-management outreach by predicted individualized treatment effect prevented substantially more ED visits or hospital admissions than risk-based allocation at the same 10% capacity.

Source article: Treatment-Effect-Based Versus Risk-Based Targeting of Care Management Outreach in Medicaid: A Retrospective Cohort Study with Machine Learning

Population Health Management

Reader signal

How should this claim be treated?