Suboptimal Echocardiographic Image Quality Predicts in-Hospital Mortality Across the Adult Age Spectrum, Independent of Admission Type: A Large-Scale NLP Study
Echocardiography, cilt.43, sa.9, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 43 Sayı: 9
- Basım Tarihi: 2026
- Doi Numarası: 10.1111/echo.70613
- Dergi Adı: Echocardiography
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, CINAHL, EMBASE, MEDLINE, Academic Search Ultimate (EBSCO), Biomedical Reference Collection: Corporate Edition (EBSCO)
- Anahtar Kelimeler: aging, echocardiography, mortality, natural language processing, prognosis
- Hacettepe Üniversitesi Adresli: Evet
Özet
Purpose: Suboptimal echocardiographic image quality is common and usually treated as a reason for study exclusion, yet discarded studies come disproportionately from older patients. We examined whether it independently predicts in-hospital mortality across the adult age spectrum. Methods: We analyzed 32,813 reports from 19,731 patients (ECHO-NOTE2NUM/MIMIC-III, 2001–2012). A validated natural language processing (NLP) classifier (κ = 0.996) identified suboptimal studies. Multivariable logistic regression, GEE, and IPTW assessed mortality across five age groups (18–49, 50–64, 65–74, 75–84, ≥85 years). Results: Overall, 9,150 studies (27.9%) were suboptimal, with similar rates across age groups but differing causes (tachycardia and body habitus in younger versus poor acoustic windows in older patients). Suboptimal imaging independently predicted mortality (IPTW OR 1.47, 95% CI 1.41–1.54). Age modified this association (interaction p = 0.04): the effect was consistently elevated across 18–74 and attenuated to 1.20 (1.06–1.35) in patients aged ≥85. Comorbidity adjustment only partly attenuated the age gradient (Charlson-adjusted interaction p = 0.12), while documented goals-of-care limitations (Do Not Resuscitate / Do Not Intubate; DNR/DNI) rose from 7.9% (18–49) to 32.0% (≥85), peaking in the same stratum in which the mortality association disappeared. Conclusion: Suboptimal image quality independently predicts in-hospital mortality across a broad 18–74 vulnerability window, with attenuation in patients aged ≥85 that is only partly explained by comorbidity and parallels a steep rise in goals-of-care limitations. Image quality is best read as an age-dependent prognostic signal whose actionability declines as care shifts toward comfort in the oldest old, supporting active escalation in adults aged 18–74 and goals-concordant interpretation thereafter.