CODEX Digest - 12.11.25

1)  A comprehensive, multidisciplinary approach to improving lung cancer screening. (subscription required)

Fortuna RJ, Wandtke B, Nead MA, et al. NEJM Catal. 2025;6(11).  

Lung cancer leads U.S. cancer deaths, which is related to the underuse of screening leading to diagnostic delays. This commentary reports on a program that raised screening rates through education, primary care integration, recalls, and centralized support. Screening rates more than doubled, and most cancers were found early, showing the success of this coordinated approach. 

2) Detecting depression through speech and text from casual talks with fully automated virtual humans. 

Gómez-Zaragozá L, Altozano A, Llanes-Jurado J, et al. Artif Intell Med. Epub 2025 Nov 13.  

Current diagnostic methods often fail to detect depression. This study examines voice markers in conversations with virtual humans, finding that speech rhythm and tone combined with semantics may help to identify depressive symptoms. This paper raises a potential use case for AI-supported diagnosis utilizing casual speech.   

3) Sex disparities in acute myocardial infarction diagnosis and treatment. (subscription required) 

Imboden MT, Koltner E, Bryant JH, et al. Am J Cardiol.  2026;259(1):179-186. 

Gender bias may delay cardiac diagnosis and treatment. This study finds that female acute myocardial infarction (AMI) patients in the emergency department (ED) experience significantly longer time-to-ECGs than male patients. These delays hinder timely diagnosis and revascularization, increasing adverse outcomes. Recognizing these disparities can drive quality improvements for more prompt care and better patient results. 

4) Application of artificial intelligence in cervical cancer diagnosis using risk factors: A systematic review. 

Kondo TS, Ngondya D, Rusheke H. Telematics Informatics Reps  2025;20:100250. 

Cervical cancer screening can be difficult to access, particularly for women in low- and middle-income countries.  This review identifies research gaps in representation of sub-Saharan African populations and urges better policies for diagnostic tool performance and data processing.  

5) Clinical reasoning and diagnostic errors. (subscription required) 

McGervey M, Olson A, Mohmand M, et al. Med Clin North Am. 2025;109(5):997-1008. 

Accurate diagnosis is crucial for patient safety, yet errors remain common. This commentary highlights that while no single solution exists to improve diagnosis, progress requires individual, team, and systems-based strategies, and that teamwork and information technologies are key to realizing diagnostic excellence. 

6) Surveying patients to assess timeliness of cancer diagnosis.

O’Hanlon CE, Berdahl C, Ye F, et al. Diagnosis (Berl). Epub 2025 Oct 27.  

Patient experiences are key to understanding factors affecting the timeliness of cancer diagnosis. This study surveys newly diagnosed cancer patients to evaluate how their responses can measure care process effectiveness. The findings add to the growing existing literature that patients can identify obstacles that contribute to longer diagnostic intervals for cancer. 

7) When did “I don’t know” become a diagnosis? A conceptual review on diagnoses of exclusion. 

Papandreas M, Carrasco-Labra A, Glick M. Oral Surg Oral Med Oral Pathol Oral Radiol. Epub 2025 Sep 5.  

There are currently no standard criteria for making a diagnosis of exclusion, which is when no definitive diagnosis is determined after a full differential workup. This article proposes a framework to help clinicians use probabilistic diagnostics to improve diagnostic certainty when arriving at a condition or reaching a diagnosis of exclusion. 

8) Evaluating a disease-specific look-back trigger methodology vs. traditional screening for diagnostic errors in the emergency department. (subscription required)

Pavuluri SK, Sangal RB, Rothenberg C, et al. Jt Comm J Qual Patient Saf. Epub October 17, 2025.

Triggers can help target priorities for diagnostic improvement work. This study compares using specific high-risk diagnoses (diagnosis-based triggers) against traditional quality assurance methods. For appendicitis and neurologic hemorrhage, lookback review found more diagnostic errors from prior ED visits which quality assurance processes missed.  

9) Beyond single systems: how multi-agent AI is reshaping ethics in radiology. 

Salehi S, Singh Y, Habibi P, et al. Bioengineering (Basel). 2025;12(10):1100.  

Advanced computing will significantly influence radiology, though AI's impact on clinical and ethical practice remains unclear. This paper discusses how agentic AI systems add complexity to radiology practice through the actions of multiple interacting agents. The piece discusses how as AI grows more capable, it becomes less interpretable, challenging trust and oversight, and suggests approaches for using AI responsibly while supporting transparency and accountability. 

10) Missing the needle in the haystack: diagnostic errors in neurological emergencies within Canadian emergency departments.  (subscription required) 

Skoblenick K, Finestone PJ, Perron D, et al. CJEM. Epub 2025 Nov 6. 

Misdiagnosis of neurological conditions in EDs can cause serious harm, including disability or death. This study of closed medical malpractice cases from Canadian EDs identified stroke, traumatic injuries, and infection as the most frequently missed. Importantly, the documented patient concerns matched expert peer review.  

11) Analysis of patient safety event report to understand the contribution of health IT to diagnostic error. 

Spaar P, Krevat SM, Boxley CL, et al. J Patient Saf. Epub 2025 Nov 11.  

Identifying health IT's role in diagnostic error can guide improvement efforts. This study of patient safety reports shows that health IT—especially EHRs—contribute to diagnostic missteps, with testing-related issues most frequent and IT errors often causing more harm than those unrelated to health IT. Addressing these issues requires better reporting processes and enhancements in health IT design. 

12) A quantitative study of pathologists’ perceptions towards artificial intelligence-assisted diagnostic system. 

Ye Z, Lu Q, Wang J, Jiang Y, et al. PLOS Digit Health. 2025;4(10):e0001052.  

AI acceptance as a diagnostic tool is vital for its successful integration into care delivery. This study examines pathologists' perspectives on AI. Over 80% support using AI for clinical diagnostics due to faster results and reduced workload but have concerns about its accuracy.