The variability and biases in the real-world performance benchmarking of deep learning models for medical imaging compromise their trustworthiness for real-world deployment. The common approach of holding out a single fixed test set fails to quantify the variance in the estimation of test performance metrics. This study introduces NACHOS (Nested and Automated Cross-validation and Hyperparameter Optimization using Supercomputing) to reduce and quantify the variance of test performance metrics of deep learning models. NACHOS integrates Nested Cross-Validation (NCV) and Automated Hyperparameter Optimization (AHPO) within a parallelized high-performance computing (HPC) framework. NACHOS was demonstrated on a chest X-ray repository and an Optical Coherence Tomography (OCT) dataset under multiple data partitioning schemes. Beyond performance estimation, DACHOS (Deployment with Automated Cross-validation and Hyperparameter Optimization using Supercomputing) is introduced to leverage AHPO and cross-validation to build the final model on the full dataset, improving expected deployment performance. The findings underscore the importance of NCV in quantifying and reducing estimation variance, AHPO in optimizing hyperparameters consistently across test folds, and HPC in ensuring computational feasibility. By integrating these methodologies, NACHOS and DACHOS provide a scalable, reproducible, and trustworthy framework for DL model evaluation and deployment in medical imaging. To maximize public availability, the full open-source codebase is provided at https://github.com/thepanlab/NACHOS.

Epidural anesthesia requires injection of anesthetic into the epidural space in the spine. Accurate placement of the epidural needle is a major challenge. To address this, we developed a forward-view endoscopic optical coherence tomography (OCT) system for real-time imaging of the tissue in front of the needle tip during the puncture. We tested this OCT system in porcine backbones and developed a set of deep learning models to automatically process the imaging data for needle localization. A series of binary classification models were developed to recognize the five layers of the backbone, including fat, interspinous ligament, ligamentum flavum, epidural space, and spinal cord. The classification models provided an average classification accuracy of 96.65%. During puncture, it is important to maintain a safe distance between the needle tip and the dura mater. Regression models were developed to estimate that distance based on the OCT imaging data. Based on the Inception architecture, our models achieved a mean absolute percentage error of 3.05% ± 0.55%. Overall, our results validated the technical feasibility of using this novel imaging strategy to automatically recognize different tissue structures and measure the distances ahead of the needle tip during the epidural needle placement.

Percutaneous renal access is the critical initial step in many medical settings. In order to obtain the best surgical outcome with minimum patient morbidity, an improved method for access to the renal calyx is needed. In our study, we built a forward-view optical coherence tomography (OCT) endoscopic system for percutaneous nephrostomy (PCN) guidance. Porcine kidneys were imaged in our experiment to demonstrate the feasibility of the imaging system. Three tissue types of porcine kidneys (renal cortex, medulla, and calyx) can be clearly distinguished due to the morphological and tissue differences from the OCT endoscopic images. To further improve the guidance efficacy and reduce the learning burden of the clinical doctors, a deep-learning-based computer aided diagnosis platform was developed to automatically classify the OCT images by the renal tissue types. Convolutional neural networks (CNN) were developed with labeled OCT images based on the ResNet34, MobileNetv2 and ResNet50 architectures. Nested cross-validation and testing was used to benchmark the classification performance with uncertainty quantification over 10 kidneys, which demonstrated robust performance over substantial biological variability among kidneys. ResNet50-based CNN models achieved an average classification accuracy of 82.6%±3.0%. The classification precisions were 79%±4% for cortex, 85%±6% for medulla, and 91%±5% for calyx and the classification recalls were 68%±11% for cortex, 91%±4% for medulla, and 89%±3% for calyx. Interpretation of the CNN predictions showed the discriminative characteristics in the OCT images of the three renal tissue types. The results validated the technical feasibility of using this novel imaging platform to automatically recognize the images of renal tissue structures ahead of the PCN needle in PCN surgery.

The three-dimensional (3D) tumor spheroid model is a critical tool for high-throughput ovarian cancer research and anticancer drug development in vitro. However, the 3D structure prevents high-resolution imaging of the inner side of the spheroids. We aim to visualize and characterize 3D morphological and physiological information of the contact multicellular ovarian tumor spheroids growing over time. We intend to further evaluate the distinctive evolutions of the tumor spheroid and necrotic tissue volumes in different cell numbers and determine the most appropriate mathematical model for fitting the growth of tumor spheroids and necrotic tissues. A label-free and noninvasive swept-source optical coherence tomography (SS-OCT) imaging platform was applied to obtain two-dimensional (2D) and 3D morphologies of ovarian tumor spheroids over 18 days. Ovarian tumor spheroids of two different initial cell numbers (5,000- and 50,000- cells) were cultured and imaged (each day) over the time of growth in 18 days. Four mathematical models (Exponential-Linear, Gompertz, logistic, and Boltzmann) were employed to describe the growth kinetics of the tumor spheroids volume and necrotic tissues. Ovarian tumor spheroids have different growth curves with different initial cell numbers and their growths contain different stages with various growth rates over 18 days. The volumes of 50,000-cells spheroids and the corresponding necrotic tissues are larger than that of the 5,000-cells spheroids. The formation of necrotic tissue in 5,000-cells numbers is slower than that in the 50,000-cells ones. Moreover, the Boltzmann model exhibits the best fitting performance for the growth of tumor spheroids and necrotic tissues. Optical coherence tomography (OCT) can serve as a promising imaging modality to visualize and characterize morphological and physiological features of multicellular ovarian tumor spheroids. The Boltzmann model integrating with 3D OCT data of ovarian tumor spheroids provides great potential for high-throughput cancer research in vitro and aiding in drug development.

Optical coherence tomography (OCT) can be used to image microstructures of human kidneys. However, current OCT probes exhibit inadequate field-of-view, leading to potentially biased kidney assessment. Here we present a robotic OCT system where the probe is integrated to a robot manipulator, enabling wider area (covers an area of 106.39 mm by 37.70 mm) spatially-resolved imaging. Our system comprehensively scans the kidney surface at the optimal altitude with preoperative path planning and OCT image-based feedback control scheme. It further parameterizes and visualizes microstructures of large area. We verified the system positioning accuracy on a phantom as 0.0762 ± 0.0727 mm and showed the clinical feasibility by scanning ex vivo kidneys. The parameterization reveals vasculatures beneath the kidney surface. Quantification on the proximal convoluted tubule of a human kidney yields clinical-relevant information. The system promises to assess kidney viability for transplantation after collecting a vast amount of whole-organ parameterization and patient outcomes data.

Understanding the functional wiring of neural circuits and their patterns of activation following sensory stimulations is a fundamental task in the field of neuroscience. Furthermore, charting the activity patterns is undoubtedly important to elucidate how neural networks operate in the living brain. However, optical imaging must overcome the effects of light scattering in the tissue, which limit the light penetration depth and affect both the imaging quantitation and sensitivity. Laminar optical tomography (LOT) is a three-dimensional (3-D) in-vivo optical imaging technique that can be used for functional imaging. LOT can achieve both a resolution of 100 to $200μm$ and a penetration depth of 2 to 3 mm based either on absorption or fluorescence contrast, as well as large field-of-view and high acquisition speed. These advantages make LOT suitable for 3-D depth-resolved functional imaging of the neural functions in the brain and spinal cords. We review the basic principles and instrumentations of representative LOT systems, followed by recent applications of LOT on 3-D imaging of neural activities in the rat forepaw stimulation model and mouse whisker-barrel system.