Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Surgery Surgery

Machine Learning and Artificial Intelligence in Surgical Research.

In The Surgical clinics of North America

Machine learning, a subtype of artificial intelligence, is an emerging field of surgical research dedicated to predictive modeling. From its inception, machine learning has been of interest in medical and surgical research. Built on traditional research metrics for optimal success, avenues of research include diagnostics, prognosis, operative timing, and surgical education, in a variety of surgical subspecialties. Machine learning represents an exciting and developing future in the world of surgical research that will not only allow for more personalized and comprehensive medical care.

Srinivas Shruthi, Young Andrew J

2023-Apr

Artificial intelligence, Machine learning, Surgical research

Surgery Surgery

Generation Learning Differences in Surgery: Why They Exist, Implication, and Future Directions.

In The Surgical clinics of North America

The evolution of the knowledge economy and technology industry have fundamentally changed the learning environments occupied by contemporary surgical trainees and created pressures that will force the surgical community to consider. Although some learning differences are intrinsic to the generations themselves, these differences are primarily a function of the environments in which surgeons of different generations trained. Acknowledgment of the principles of connectivism and thoughtful integration of artificial intelligence and computerized decision support tools must play a central role in charting the future course of surgical education.

Weykamp Mike, Bingham Jason

2023-Apr

Artificial intelligence, Asynchronous learning, Connectivism, Generational learning, Learning disparities, Learning theory, Machine learning, Medical education

General General

Dynamic 3D imaging of cerebral blood flow in awake mice using self-supervised-learning-enhanced optical coherence Doppler tomography.

In Communications biology

Cerebral blood flow (CBF) is widely used to assess brain function. However, most preclinical CBF studies have been performed under anesthesia, which confounds findings. High spatiotemporal-resolution CBF imaging of awake animals is challenging due to motion artifacts and background noise, particularly for Doppler-based flow imaging. Here, we report ultrahigh-resolution optical coherence Doppler tomography (µODT) for 3D imaging of CBF velocity (CBFv) dynamics in awake mice by developing self-supervised deep-learning for effective image denoising and motion-artifact removal. We compare cortical CBFv in awake vs. anesthetized mice and their dynamic responses in arteriolar, venular and capillary networks to acute cocaine (1 mg/kg, i.v.), a highly addictive drug associated with neurovascular toxicity. Compared with awake, isoflurane (2-2.5%) induces vasodilation and increases CBFv within 2-4 min, whereas dexmedetomidine (0.025 mg/kg, i.p.) does not change vessel diameters nor flow. Acute cocaine decreases CBFv to the same extent in dexmedetomidine and awake states, whereas decreases are larger under isoflurane, suggesting that isoflurane-induced vasodilation might have facilitated detection of cocaine-induced vasoconstriction. Awake mice after chronic cocaine show severe vasoconstriction, CBFv decreases and vascular adaptations with extended diving arteriolar/venular vessels that prioritize blood supply to deeper cortical capillaries. The 3D imaging platform we present provides a powerful tool to study dynamic changes in vessel diameters and morphology alongside CBFv networks in the brain of awake animals that can advance our understanding of the effects of drugs and disease conditions (ischemia, tumors, wound healing).

Pan Yingtian, Park Kicheon, Ren Jiaxiang, Volkow Nora D, Ling Haibin, Koretsky Alan P, Du Congwu

2023-Mar-21

Pathology Pathology

Medical diffusion on a budget: textual inversion for medical image generation

ArXiv Preprint

Diffusion-based models for text-to-image generation have gained immense popularity due to recent advancements in efficiency, accessibility, and quality. Although it is becoming increasingly feasible to perform inference with these systems using consumer-grade GPUs, training them from scratch still requires access to large datasets and significant computational resources. In the case of medical image generation, the availability of large, publicly accessible datasets that include text reports is limited due to legal and ethical concerns. While training a diffusion model on a private dataset may address this issue, it is not always feasible for institutions lacking the necessary computational resources. This work demonstrates that pre-trained Stable Diffusion models, originally trained on natural images, can be adapted to various medical imaging modalities by training text embeddings with textual inversion. In this study, we conducted experiments using medical datasets comprising only 100 samples from three medical modalities. Embeddings were trained in a matter of hours, while still retaining diagnostic relevance in image generation. Experiments were designed to achieve several objectives. Firstly, we fine-tuned the training and inference processes of textual inversion, revealing that larger embeddings and more examples are required. Secondly, we validated our approach by demonstrating a 2\% increase in the diagnostic accuracy (AUC) for detecting prostate cancer on MRI, which is a challenging multi-modal imaging modality, from 0.78 to 0.80. Thirdly, we performed simulations by interpolating between healthy and diseased states, combining multiple pathologies, and inpainting to show embedding flexibility and control of disease appearance. Finally, the embeddings trained in this study are small (less than 1 MB), which facilitates easy sharing of medical data with reduced privacy concerns.

Bram de Wilde, Anindo Saha, Richard P. G. ten Broek, Henkjan Huisman

2023-03-23

Surgery Surgery

Big Data in Surgery.

In The Surgical clinics of North America

The emergence of Big Data has been facilitated by technological advancements in the processing, storage, and analysis of large quantities of data. Its strength is derived from its size, ease of access, and speed of analysis, and it has enabled surgeons to investigate areas of interest that traditional research models have historically been unable to address. In the future, Big Data will likely assist in the incorporation of more advanced technologies into surgical practice, including artificial intelligence and machine learning to realize the full potential of Big Data in Surgery.

Prien Christopher, Lincango Eddy P, Holubar Stefan D

2023-Apr

Artificial intelligence, Big data, Machine learning, Natural language processing, Quality, Registries, Surgery, outcomes research

Radiology Radiology

Radiologist's Disease: Imaging for Renal Cancer.

In The Urologic clinics of North America

There is a clear benefit of imaging-based differentiation of small indeterminate masses to its subtypes of clear cell renal cell carcinoma (RCC), chromophobe RCC, papillary RCC, fat poor angiomyolipoma and oncocytoma because it helps determine the next step options for the patients. The work thus far in radiology has explored different parameters in computed tomography, MRI, and contrast-enhanced ultrasound with the discovery of many reliable imaging features that suggest certain tissue subtypes. Likert score-based risk stratification systems can help determine management, and new techniques such as perfusion, radiogenomics, single-photon emission tomography, and artificial intelligence can add to the imaging-based evaluation of indeterminate renal masses.

Chung Alex, Raman Steven S

2023-May

CT, MRI, Multiphasic contrast imaging, Ultrasound