Artificial neural networks and computer vision in medicine and surgery

Introduction: Artificial neural networks are becoming an essential technology in data analysis, and their influence is starting to permeate the field of medicine. Experimental surgery has been a long-term subject of study of our lab; this is naturally reflected in our interest in other areas of modern technologies including artificial neural networks and their advancements. In the current issue, we would like to explore this aspect of technical progress. The main goal is to critically evaluate the strengths and weaknesses of artificial neural network technology concerning its use in clinical and experimental surgery. Methods: The article is focused on in-silico modeling, particularly on the potential of neural networks in terms of image data processing in medicine. The text briefly summarizes the historical development of deep learning neural networks and their basic principles. Furthermore, basic taxonomy tasks are presented. Finally, potential learning problems and possible solutions are also mentioned. Results: The article points out various possible uses of artificial neural networks in biological applications. Several biomedical applications of artificial neural networks are used to describe the division and principles of the most common tasks of machine learning and deep learning such as classification, detection, and segmentation. Conclusion: The application of artificial neural network methods in medicine and surgery offers a considerable potential; by learning directly from the data, they make it possible to avoid lengthy and subjective setting of parameters by an expert engineer. Nevertheless, the use of an unbalanced dataset can lead to unexpected, although traceable errors. The solution is to collect a dataset large enough to enable both learning and verification of proper functionality.

[Differential diagnosis and treatment of pituitary adenomas]. / Differenzialdiagnostik und Therapie der Hypophysenadenome.

Pituitary adenomas are among the most common primary brain tumors. These tumors can produce all hormones of the anterior pituitary and thus cause endocrine diseases. Compression of the pituitary gland, the surrounding cranial nerves, or brain structures can lead to hypopituitarism, cranial nerve deficits, or diverse neurological symptoms. Visual impairment, typically with bitemporal hemianopsia, is the most common cardinal symptom. The diagnostic workup requires broad interdisciplinary cooperation. With the exception of prolactinoma, the treatment of choice for symptomatic pituitary adenoma is transnasal transsphenoidal resection. For prolactinoma, dopamine agonistic therapy is the primary treatment. Adequate hormone replacement therapy is essential in cases of hypopituitarism. Long-term follow-up is a vital part of the treatment concept.

Intraoperative magnetic resonance imaging.

Intraoperative magnetic resonance imaging is a widely accepted method for resection control of glial tumors. Increasingly, it is also used during the resection of skull base tumors. Several studies have independently demonstrated an increase in the extent of resection in these tumors with improved prognosis for the patients. Technical innovations combined with the easier operation of this imaging modality have led to its widespread implementation. The development of digital image processing has also brought other modalities such as ultrasound and computed tomography to the focus of skull base surgery.

[Intraoperative magnetic resonance imaging. Fifteen years' experience in the neurosurgical hybrid operating suite]. / Intraoperative Magnetresonanztomographie. Erfahrungen aus 15 Jahren Hybridoperationssaal in der Neurochirurgie.

Intraoperative magnetic resonance imaging was established 15 years ago due to special requirements for the resection of cerebral gliomas. Several studies have independently shown an increase of the extent of resection of the tumor and also an improved survival of the patients. Technical innovations combined with an easier operation of this imaging modality led to widespread implementation of this method. The introduction of functional and metabolic imaging opened up new prospects of further improving the therapeutic outcome.

Lows and highs: 15 years of development in intraoperative magnetic resonance imaging.

Intraoperative magnetic resonance imaging (ioMRI) during neurosurgical procedures was first implemented in 1995. In the following decade ioMRI and image guided surgery has evolved from an experimental stage into a safe and routinely clinically applied technique. The development of ioMRI has led to a variety of differently designed systems which can be basically classified in one- or two-room concepts and low- and high-field installations. Nowadays ioMRI allows neurosurgeons not only to increase the extent of tumor resection and to preserve eloquent areas or white matter tracts but it also provides physiological and biological data of the brain and tumor tissue. This article tries to give a comprehensive review of the milestones in the development of ioMRI and neuronavigation over the last 15 years and describes the personal experience in intraoperative low and high-field MRI.

MR monitoring of focused ultrasound surgery in a breast tissue model in vivo.

The objective of this study was to investigate MRI methods for monitoring focused ultrasound surgery (FUS) of breast tumors. To this end, the mammary glands of sheep were used as tissue model. The tissue was treated in vivo with numerous single sonications which covered extended target volumes by employing a scanning technique. The ultrasound focus position was controlled by online temperature mapping based on the temperature dependence of the relaxation time T(1). This approach proved to be reliable and offers thus an alternative to proton resonance frequency methods, whose application is hampered in fatty tissues. FUS-induced tissue changes were visible on T(2)- as well as on pre- and post-contrast T(1)-weighted images. According to our initial experience, noninvasive MRI-guided FUS of breast tumors is feasible.

[Magnetic resonance-guided therapy with focused ultrasound. Non-invasive surgery of breast carcinoma?]. / Kernspingesteuerte Therapie mit fokussiertem Ultraschall. Nichtinvasive Chirurgie des Mammakarzinoms?

Magnetic resonance guided focused ultrasound surgery (MRgFUS) has the potential to become an important therapy modality in the adjuvant, neoadjuvant or palliative cancer treatment. All ultrasound accessible regions are possible target areas, especially breast tumors. Ultrasound propagation is well predictable. The ultrasound energy can be focused to a defined spot through the intact skin, and temperatures of 60 degrees C to 85 degrees C can be induced locally for a few seconds that instantaneously necrose biological tissues, while sparing surrounding healthy tissue. In addition, MRI is sensitive to temperature allowing for online monitoring of the temperature focus. In this work we demonstrate our Heidelberg experiments from basic research and animal studies towards the clinical realization of MRgFUS in breast cancer patients. The most important of these experiments involved sheep as an appropriate model for the human breast. A new therapy setup is designated to treat human breast patients in a clinical 1.5 T MRI scanner. While the therapies have been successful so far without any side effects, the future clinical role of noninvasive MRgFUS has to be defined by clinical studies.