Reproducibility of Deep Learning Algorithms Developed for Medical Imaging Analysis: A Systematic Review

Moassefi, Mana; Rouzrokh, Pouria; Conte, Gian Marco; Vahdati, Sanaz; Fu, Tianyuan; Tahmasebi, Aylin; Younis, Mira; Farahani, Keyvan; Gentili, Amilcare; Kline, Timothy L.; Kitamura, Felipe; Huo, Yuankai; Kuanar, Shiba; Younis, Khaled; Erickson, Bradley J.; Faghani, Shahriar

doi:10.1007/s10278-023-00870-5

Cited by 5 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Es besteht ein hohes Risiko der Verzerrung, insbesondere aufgrund fehlender externer Validierung. Außerdem fehlt in den meisten Fällen eine klare und dem Anwender verständliche Erklärung, wie das System arbeitet und wo die Limitationen liegen, die zu beachten sind [4,15].…”

Section: Optimierung Der Ai Und Minimierung Der Potenziellen Schwachs...unclassified

“…The systems and especially primary system testing and the corresponding adaptation of the AI systems during the software training phase typically ensure that the corresponding software functions reasonably in a narrow application field, i. e., in the framework of the learned parameters [14,15].…”

Section: Explainable Aimentioning

confidence: 99%

“…There is a high risk of distortion, par-ticularly due to the lack of external validation. Moreover, a clear and understandable explanation of how the system works and which limitations need to be taken into consideration is missing in most cases [4,15].…”

Section: Optimization Of Ai and Minimization Of Potential Weaknessesmentioning

confidence: 99%

“…[4]. Die Systeme und vor allem die primäre Testung der Systeme und die entsprechende Adaption der KI während der Trainingsphase der Software sorgen in aller Regel dafür, dass die entsprechende Software in einem engen Anwendungsfeld vernünftig, also im Rahmen der erlernten Parameter, funktioniert [14,15].…”

Section: Explainable Aierklärbare Künstliche Intelligenzunclassified

See 3 more Smart Citations

The radiologist as a physician – artificial intelligence as a way to overcome tension between the patient, technology, and referring physicians – a narrative review

Stueckle,

Haage

2024

Rofo

View full text Add to dashboard Cite

Background Large volumes of data increasing over time lead to a shortage of radiologistsʼ time. The use of systems based on artificial intelligence (AI) offers opportunities to relieve the burden on radiologists. The AI systems are usually optimized for a radiological area. Radiologists must understand the basic features of its technical function in order to be able to assess the weaknesses and possible errors of the system and use the strengths of the system. This “explainability” creates trust in an AI system and shows its limits. Method Based on an expanded Medline search for the key words “radiology, artificial intelligence, referring physician interaction, patient interaction, job satisfaction, communication of findings, expectations”, subjective additional relevant articles were considered for this narrative review. Results The use of AI is well advanced, especially in radiology. The programmer should provide the radiologist with clear explanations as to how the system works. All systems on the market have strengths and weaknesses. Some of the optimizations are unintentionally specific, as they are often adapted too precisely to a certain environment that often does not exist in practice – this is known as “overfitting”. It should also be noted that there are specific weak points in the systems, so-called “adversarial examples”, which lead to fatal misdiagnoses by the AI even though these cannot be visually distinguished from an unremarkable finding by the radiologist. The user must know which diseases the system is trained for, which organ systems are recognized and taken into account by the AI, and, accordingly, which are not properly assessed. This means that the user can and must critically review the results and adjust the findings if necessary. Correctly applied AI can result in a time savings for the radiologist. If he knows how the system works, he only has to spend a short amount of time checking the results. The time saved can be used for communication with patients and referring physicians and thus contribute to higher job satisfaction. Conclusion Radiology is a constantly evolving specialty with enormous responsibility, as radiologists often make the diagnosis to be treated. AI-supported systems should be used consistently to provide relief and support. Radiologists need to know the strengths, weaknesses, and areas of application of these AI systems in order to save time. The time gained can be used for communication with patients and referring physicians. Key Points Citation Format

show abstract

Section: Optimierung Der Ai Und Minimierung Der Potenziellen Schwachs...unclassified

Section: Explainable Aimentioning

confidence: 99%

Section: Optimization Of Ai and Minimization Of Potential Weaknessesmentioning

confidence: 99%

Section: Explainable Aierklärbare Künstliche Intelligenzunclassified

See 2 more Smart Citations

The radiologist as a physician – artificial intelligence as a way to overcome tension between the patient, technology, and referring physicians – a narrative review

Stueckle,

Haage

2024

Rofo

View full text Add to dashboard Cite

show abstract

“…Deep learning algorithms have revolutionized disease diagnostics by enabling highly accurate and efficient analysis of medical imaging data (Chen et al, 2023). Deep learning has been utilized in tasks like a diagnosis of disease from microscopic images, identifying tumors in radiological images, detecting abnormalities in mammograms, and diagnosing various conditions from X-rays (Chen et al, 2023;Hasan et al, 2019;Moassefi et al, 2023). For instance, Irmak (2021) proposed a novel deep-learning method for malaria parasite diagnosis from blood cell images.…”

Section: Domain Of Medical Imagingmentioning

confidence: 99%

Exploring the Potential of Deep Learning in Healthcare: A perspective

Lufyagila,

Ruambo

2024

East Afr. j. interdiscip. stud.

View full text Add to dashboard Cite

Deep learning has received much interest in the field of healthcare in recent years. Health care plays a significant role in delivering services and practices that promote, maintain, and restore health of an individual. However, applying deep learning in healthcare is still an exciting area of research. This paper explores the application of deep learning, and henceforth, it highlights new perspectives in healthcare by reviewing published state-of-the-art research works from four scholarly databases, including Scopus, Web of Science, Pubmed, and Google Scholar. The selected studies were from April 2014 to April 2024, and based on the predefined quality assessment criteria, 16 articles were thoroughly reviewed after the preliminary extraction, review, and screening phases. The study’s findings indicate that deep learning has been applied in healthcare, particularly in medical images, digital consultation, Electronic medical records, and genomics. Furthermore, challenges such as deep learning cannot replicate the human touch and emotional connection that patients often seek in their healthcare journey, and data privacy are highlighted. Lastly, new perspectives, such as leveraging emerging technologies like Augmented Reality (AR), Virtual Reality (VR), and federated learning, are suggested to address these challenges

show abstract