8 RADIOLOGY Utilizing new strengths, fixing old weaknesses Ultrasound update for organ imaging personal contact with the patient. Studies have shown that this not only creates a more pleasant at- mosphere during the examination, but also improves the quality of the diagnosis6, she emphasised. ‘This is where we really have a big ad- vantage over other cross-sectional imaging techniques – and this should not be underestimated,’ she concluded. ■ Author: Wolfgang Behrends ) 0 . 4 Y B C C ( 4 2 0 2 s t r o p e R c i f i t n e i c S , W i L , K g n a W , L g n a Y : Illustration of multiple septations in a pleural effusion on ultrasound. Has organ imaging using ultra- sound arrived at the same level as cross-sectional imaging? At the annual conference of the German Society for Internal Medicine (DGIM), PD Dr Corinna Trenker presented new tech- nological developments and their diagnostic significance. Despite numerous innovations such as multiparametric proto- cols and AI support, she made it clear that the human factor re- mains one of the greatest strengths of sonography over other modalities. The examination of neck and soft tissue as well as organs in the near- field – testicles, thyroid or breast – has traditionally been one of the strong suits of ultrasound, reported the expert from the Clinic for Hae- matology, Oncology and Immu- nology at Marburg University Hos- pital: image resolution of up to 0.1 to 0.2 milli- metres, sonography simply ex- ceeds the resolution of CT and lateral ‘With a MRI.’ This is also reflected in the S3 guidelines – ultrasound is the method of choice for both tes- ticular cancer and the detection of thyroid masses. Even though the results of lymph node detection are consistently convincing, the technique has not yet been included in the guide- lines, Dr Trenker summarised. Nevertheless, she said that multi- parametric imaging using elas- tography and contrast-enhanced ultrasound (CEUS) is an exciting new tool for assessing lymph node hardening and vascularisation1 – ‘however, this is not relevant for the primary assessment of malig- nancy,’ the expert added. into play: In thoracic examinations, another advantage of ultrasound imaging comes ‘Sonography clearly has the edge because exam- inations can be performed in real- time,’ explained Dr Trenker. Par- ticularly when assessing thoracic wall infiltration in bronchial carci- also nomas, ‘the decisive factor is the respiratory-dependent fixation of the tumour, and here the sensitivity and specificity of ultrasound is clearly superior to CT imaging.’2 Ultrasound the method of choice for the detection and differentiation of pleural effu- sions in 2025. ‘Not only can pleural effusions of just a few millilitres be detected, but they can also be dif- ferentiated more accurately,’ said the expert.3 remains AI opens new possibilities When it comes to deeper anatomi- cal structures, ultrasound quickly reaches its limits, and the physical limitations of air-filled structures remain unchanged, Dr Trenker pointed out. The high level of examiner dependency also con- tinues to be an issue – but this could change in the future: New publications suggest, for example, that artificial (AI) could support the traditionally de- manding screening for hepatocel- (HCC).4 Con- lular carcinoma intelligence m o r f d e t p a d a sidering the growing issue of staff shortages, which also affects ex- perienced sonographers, this is a promising development, the expert pointed out – even if there are still a lot of unsolved questions regard- ing data protection. Dr Trenker illustrated the amazing diagnostic potential of AI in ultra- sound by looking at a recent meta- analysis from China: here, re- searchers used AI to analyse the sonograms of more than 11,000 breast cancer patients – and, based on this alone, predicted their HER2 receptor status with a high degree of accuracy.5 ‘This is a little hard to grasp, but, at least in initial studies, it is possible. I believe that many more applications in this field will emerge in the future.’ Patient contact: not just ‘nice to have’ Despite the general enthusiasm for technological progress, the expert reminded her audience of one of the key advantages of ultrasound: References: 1Künzel J, Brandenstein M, Zeman F et al.: Multiparametric Ultrasound of Cervical Lymph Node Metastases in Head and Neck Cancer for Planning Non-Surgical Therapy; Diagnostics https://doi. org/10.3390/diagnostics12081842 (Basel) 2022; 2Bandi V, Lunn W, Ernst E et al.: Ultrasound vs. CT in detecting chest wall invasion by tumor: a prospective study; Chest 2008; https://doi.org/10.1378/chest.07–1656 3Yang L, Wang K, Li W, Liu D: Chest ultra- sound is better than CT in identifying sep- tated effusion of patients with pleural dis- ease; 2024; https://doi. org/10.1038/s41598–024–62807–4 Scientific Reports 4Stefanini B, Giamperoli A, Terzi E, Piscaglia F: Artificial intelligence in Ultrasound: Pearls and pitfalls in 2024; Ultraschall in der Medi- zin https://doi. org/10.1055/a-2368–9201 2024; 5Fu Y, Zhou J, Li J.: Diagnostic performance of ultrasound-based artificial intelligence for predicting key molecular markers in breast cancer: A systematic review and meta- analysis; PLoS One 2024; https://doi. org/10.1371/journal.pone.0303669 6Gutzeit A, Sartoretti E, Reisinger C et al.: Di- rect communication between radiologists and patients improves the quality of imaging reports; 2021; https://doi. org/10.1007/s00330–021–07933–7 Radiology European Quantum mechanics to remove noise from medical images The math describing how par- ticles move in space can apply to stray pixels, removing noise from images. Medical imaging methods such as ultrasound and magnetic reson- ance imaging (MRI) are often af- fected by background noise, which can introduce blurring and obscure fine anatomical details in the im- ages. For clinicians who depend on medical images, background noise is a fundamental problem in mak- ing accurate diagnoses. Methods for denoising have been developed with some success, but they struggle with the complexity of noise patterns in medical images and require manual tuning of par- ameters, adding complexity to the denoising process. To solve the denoising problem, some researchers have drawn in- spiration from quantum mech- anics, which describes how matter and energy behave at the atomic scale. Their studies draw an interesting analogy between how particles vibrate and how pixel in- tensity spreads out in images and causes noise. Until now, none of these attempts directly applied the full-scale mathematics of quantum mechanics to image denoising. In a paper in AIP Advances, by AIP Publishing, researchers from Mass- achusetts General Hospital, Har- vard Medical School, Weill Cornell Medicine, GE HealthCare, and Uni- versité de Toulouse took translat- ing the particle-pixel analogy to the next level. “While quantum localization is a well-established phenomenon in physical materials, our key inno- vation was conceptualizing it for noisy images — translating the physics literally, not just meta- phorically,” author Amirreza Has- hemi said. “This foundational anal- ogy didn’t exist before. We’re the first to formalize it.” A central concept in the math de- scribing matter and energy, localiz- ation is used to explain how par- ticles vibrate in a space. Vibrations that stay confined are considered localized, while vibrations that spread out are diffused. Similarly, pixel intensity, or brightness, in a clear image can be considered lo- calized, while noisy patterns in an image can be considered diffused. The authors apply the same mathe- matics that describes the localiz- ation of particle vibrations in the surrounding physical space to de- cipher the localization of pixel in- tensity in images. In this way, they can separate the noise-free “signal” of the anatomical structures in the image from the visual noise of stray pixels. “The main aspect was developing an algorithm that auto- matically separates the localized (signal) and nonlocalized (noise) components of pixels in an image by exploiting their distinct be- haviors,” Hashemi said. The researchers’ direct application of the physics and mathematics of particles also eliminated the need to manually fine-tune parameters in denoising algorithms, which Hashemi said is a major hindrance in traditional approaches. “Our method leverages physics-driven principles, like localization and dif- fusive dynamics, which inherently separate noise from signal without expensive optimization,” Hashemi said. “The algorithm just works by design, avoiding brute-force com- putations.” Their method has applications not only in medical image denoising, but across quantum computing, too. “Our physics-driven frame- work com- putational primitives of quantum systems, offering a potential per- formance advantage as quantum computing scales.” ■ aligns with the Source: American Institute of Physics EUROPEAN HOSPITAL Vol 34 Issue 3/25